REANIMATED VOICES
Pragmatics & Beyond New Series Editor: Andreas H. Jucker (Justus Liebig University, Giessen) Associate Editors: Jacob L. Mey (Odense University) Herman Parret (Belgian National Science Foundation, Universities of Louvain and Antwerp) Jef Verschueren (Belgian National Science Foundation, Univ. of Antwerp) Editorial Address: Justus Liebig University Giessen, English Department Otto-Behaghel-Strasse 10, D-35394 Giessen, Germany e-mail:
[email protected] Editorial Board: Shoshana Blum-Kulka (Hebrew University of Jerusalem) Chris Butler (University College of Ripon and York); Jean Caron (Université de Poitiers) Robyn Carston (University College London); Bruce Fraser (Boston University) Thorstein Fretheim (University of Trondheim); John Heritage (University of California at Los Angeles) Susan Herring (University of Texas at Arlington); Masako K. Hiraga (St. Paul’s (Rikkyo) University) David Holdcroft (University of Leeds); Sachiko Ide (Japan Women’s University) Catherine Kerbrat-Orecchioni (University of Lyon 2) Claudia de Lemos (University of Campinas, Brazil); Marina Sbisà (University of Trieste) Emanuel Schegloff (University of California at Los Angeles) Deborah Schiffrin (Georgetown Univ.); Paul O. Takahara (Kobe City Univ. of Foreign Studies) Sandra Thompson (University of California at Santa Barbara) Teun A. Van Dijk (University of Amsterdam); Richard J. Watts (University of Berne)
85 Daniel E. Collins Reanimated Voices Speech reporting in a historical-pragmatic perspective
REANIMATED VOICES SPEECH REPORTING IN A HISTORICAL-PRAGMATIC PERSPECTIVE
DANIEL E. COLLINS The Ohio State University
JOHN BENJAMINS PUBLISHING COMPANY AMSTERDAM/PHILADELPHIA
8
TM
The paper used in this publication meets the minimum requirements of American National Standard for Information Sciences — Permanence of Paper for Printed Library Materials, ANSI Z39.48-1984.
Library of Congress Cataloging-in-Publication Data Collins, Daniel E. Reanimated voices : speech reporting in a historical-pragmatic perspective / Daniel E. Collins. p. cm. -- (Pragmatics & beyond, ISSN 0922-842X ; new ser. 85) Includes bibliographical references and index. 1. Grammar, Comparative and general--Indirect discourse. 2. Pragmatics. 3. Trial transcripts. I. Title. II. Series. P301.5.I53 C65 2000 415--dc21 ISBN 90 272 5104 5 (Eur.) / 1 58811 023 0 (US) (alk. paper)
00-051913
© 2001 – John Benjamins B.V. No part of this book may be reproduced in any form, by print, photoprint, microfilm, or any other means, without written permission from the publisher. John Benjamins Publishing Co. • P.O.Box 75577 • 1070 AN Amsterdam • The Netherlands John Benjamins North America • P.O.Box 27519 • Philadelphia PA 19118-0519 • USA
To Henrik Birnbaum
Table of contents
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
xi
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii Conventions for citing Cyrillic sources . . . . . . . . . . . . . . . . . . . . . . . . xix Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxi C 1 The pragmatics of reported speech . . . . . . . . . . . . . . . . . . . . 1.1 Reported speech as intention and as perception . . . . . . . . . 1.2 A “thick description” of reporting . . . . . . . . . . . . . . . . . . 1.3 Some shortcomings of nonpragmatic reductionist approaches 1.4 A method for historical-pragmatic analysis . . . . . . . . . . . . 1.5 Previous applications of the “Method of Residual Forms” . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. 1 . 1 . 3 . 10 . 16 . 23
C 2 The text-kind: A pragmaphilological overview . . . . . . . . . . . . . . . . . . 27 C 3 Testimony . . . . . . . . . . . . . . . . . . . . . . . . . 3.1 The standard reporting strategy . . . . . . . . 3.2 Strategic choices in the standard tag . . . . . 3.3 Direct speech: Verbatimness or relevance? 3.4 A convention of explicit typicality . . . . . . 3.5 From conduit to collaborator . . . . . . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
35 35 42 49 59 64
viii
T C
C 4 Residual forms in testimony . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1 Participial tags and foregrounding/backgrounding effects . . . . . 4.2 Alternative tag verbs and different participant status . . . . . . . . 4.3 Free direct speech and cohesion effects . . . . . . . . . . . . . . . . . 4.4 Complementized reports: presupposition and grounding effects 4.5 Fused reported speech and streamlining effects . . . . . . . . . . . 4.6 Narrative reports of speech acts . . . . . . . . . . . . . . . . . . . . . . 4.7 Free indirect speech and the boundary with narrative . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
75 75 85 92 103 116 124 133
C 5 The question framework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1 The general function of the judges’ reported speech . . . . . . . . . . 5.2 Dyadic questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3 Reduction of the tag clause and rapid-fire exchanges . . . . . . . . . 5.4 Verba-dicendi tags: turning the tables on the judge . . . . . . . . . . . 5.5 Free direct speech and crossplay within a dominant encounter . . . 5.6 Stage-managing: The dative with infinitive after speech-act verbs
. . . . . . .
. . . . . . .
. . . . . . .
151 151 156 162 165 167 172
C 6 Reporting from judicial-referral hearings 6.1 The nature of judicial-referral hearings 6.2 Verifications . . . . . . . . . . . . . . . . . . 6.3 Falsifications . . . . . . . . . . . . . . . . . . 6.4 Preliminary reports . . . . . . . . . . . . . .
. . . . .
. . . . .
. . . . .
177 177 180 192 197
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . . . . .
. . . . .
. . . . .
C 7 Layered reports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.1 Preliminary remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.2 Additional layers of direct speech: narratives and dyads . . . . . . . . . . 7.3 Complementized indirect speech: rejoinders, hearsay, and hypothetical reports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.4 Additional layers of uncomplementized indirect speech . . . . . . . . . . . 7.5 Additional layers of fused reported speech . . . . . . . . . . . . . . . . . . . 7.6 Two functions of intercalated tags: Cohesion and disclaimers . . . . . . 7.7 Narrative reports of speech acts . . . . . . . . . . . . . . . . . . . . . . . . . . .
203 203 205 212 228 233 237 241
T C
ix
C 8 Reporting the verdict . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.1 Preliminary remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.2 Reported speech in the adjudication . . . . . . . . . . . . . . . . . . . 8.3 Reported speech in the judges’ ratio decidendi . . . . . . . . . . . 8.3.1 Indirect and nondirect speech . . . . . . . . . . . . . . . . . . 8.3.2 Fused reported speech . . . . . . . . . . . . . . . . . . . . . . . 8.3.3 Direct speech . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.3.4 Intercalated verba dicendi and other quotation markers 8.3.5 Narrative reports of speech acts . . . . . . . . . . . . . . . . 8.3.6 Free reported speech in the ratio decidendi . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
245 245 248 252 253 259 261 265 270 274
C 9 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.1 The purposiveness of reporting in trial transcripts . . . . . . . 9.2 The main functions of the reporting strategies in the corpus 9.3 Some methodological implications . . . . . . . . . . . . . . . . . .
. . . .
. . . .
. . . .
. . . .
. . . .
285 285 288 298
. . . .
. . . .
Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 303 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 319 Appendix: Text-kind and date of the investigated trial transcripts . . . . 343 Name index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 349 Subject index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 355
Acknowledgments
I would like to thank the many people who supported me in the writing of this book. I owe special thanks to two department chairs — Stephen Summerhill, who helped me to understand that a book of this kind was possible; and Irene Masing-Delic, who showed constant solicitude about the manuscript and provided me with travel support during the revision process. I am also grateful to my friend Rafael A. Madan, who arranged for me to use the Washington house of the Agrupación Católica Universitaria as a research retreat, and to all the members of the Agrupación and residents of the Washington house for their hospitality during my three visits. My thanks also to the scholars who have commented on various portions of the book — in particular, my teachers Henrik Birnbaum and Henning Andersen; my colleague Brian Joseph; and Andreas H. Jucker, Richard J. Watts, Leslie K. Arnovick, and other participants of the Historical Pragmatics panels at the 1998 International Pragmatics Association Conference and the 1999 International Conference on Historical Linguistics. I am especially grateful to Andreas and Leslie for inviting me to present papers in these panels and to the Ohio State University College of Humanities, Center for Medieval and Renaissance Studies, and Department of Slavic and East European Languages and Literatures, which provided financial support to enable me to participate. The anonymous referees of Pragmatics and Beyond gave me valuable advice on improving the manuscript. Marika Whaley deserves special recognition for her work in editing the manuscript and for the advice that she provided during the revision process. I would like to thank Predrag Matejic for allowing me to use the font OCS/PM, which he designed. Last but not least, let me express my thanks and love to my family for bearing with me during the process of writing and revising.
Preface
Recent scholarship in historical pragmatics has shown that analytic methods devised to explain usage in modern languages can also shed light on diachronic developments and earlier language states.1 In this study, I hope to contribute to this burgeoning field by demonstrating how a long-standing method of historical-comparative linguistics can be adapted for pragmaphilology and used to account for synchronic patterns of lexical and syntactic variation in a corpus of premodern writings. Applying the method of residual forms, I examine the distribution of speech-reporting strategies in a text-kind in which they are privileged — trial transcripts written in the chancery variety of Old Russian during the early Muscovite period (ca. 1410–1505). My chief goal is to discover the factors that motivated medieval writers to choose a particular form of reported speech — understood as any means of representing spoken or written discourse, not just indirect speech — in a specific context. (Reported thought is not discussed because it is not attested in my corpus.) Function-to-form matching is possible here because the communicative purposes of the different contextualizations can be inferred either from internal evidence or from the socially institutionalized function(s) of the text-kind. The results of my investigation thus validate a genre-based method for studying patterns of syntactic and lexical usage. The title of this book, Reanimated Voices, alludes to three activities that may be seen as the overarching themes of this study of speech reporting. The first and most evident is the activity of reporters, who, for purposes of their own, choose to evoke distal speech events for their audience to imagine. The second is the activity of that audience — the interpreters who, in order to understand the reporters’ communicative intention, must construct (or reconstruct) a mental image of distal speech events, and who thus become collaborators and co-authors in the act of reporting. The third (and the most
xiv
P
remote from everyday experience) is the activity of the historical pragmatician studying reported speech, an eavesdropper in time who must find a way to reconstruct the language behavior of long-silenced reporters and interpreters, to reanimate their voices for purposes that they never intended or envisioned. One of the goals of this work is to show that this reconstructive endeavor is methodologically feasible. Reported speech has received a great deal of attention in recent typological and functional (pragmatic, sociolinguistic, and anthropological-linguistic) scholarship.2 One of the reasons why it is of significance for pragmatics is that the differences between its formal varieties cannot be understood in any meaningful sense without reference to pragmatic (contextual) factors. The various forms of reported speech are not synonyms but rather instruments appropriate for different kinds of tasks; the choice of a given strategy is determined by the larger structure of the discourse and by the communicative intentions of the speaker or writer.3 In this study I provide further evidence for the context-sensitivity of reported speech, which up to now has been argued primarily for modern languages; at the same time, I show how the choice of strategy is oriented to the intended interpreter in the reporting situation. Speakers and writers choose the form that they perceive as potentially most effective for what they want to communicate and, concomitantly, for how they intend to organize their texts. Their perceptions are socially grounded — based on their experiential knowledge of how a specific kind of audience goes about interpreting the discourse in particular contextualizations and in particular genres. The functionalist current of research on reported speech was brilliantly anticipated in Bakhtin’s pioneering study (published under the name of his friend Vološinov) of the “basic and constant tendencies in the active reception of other speakers’ speech” (1929/1986: 116–17).4 In this long-neglected, now much-cited work, Bakhtin/Vološinov shifted the focus from the syntax to what in current linguistic terminology would be called the pragmatics of reported speech; he emphasized the decisive role that the reporter’s intention (“the teleology of the authorial context”) plays in the choice of “stabilized constructional patterns” (ibid.: 116, 122). Bakhtin/Vološinov was truly a voice crying in the wilderness; he remains, with good reason, a fundamental source for many function-oriented works on reported speech, including this one. The continuing influence of Bakhtin/Vološinov’s classic study gives particular urgency to the investigation of reported speech in premodern
P
xv
languages and Old Russian in particular — paradoxically, because of a shortcoming; for in his discussion of that subject Bakhtin/Vološinov departs markedly from the context-sensitive methods that he uses elsewhere in his work. On the basis of a limited reading of medieval texts (primarily chronicles and The Song of Prince Igor’s Campaign), he declares that reported speech was essentially monolithic in Old Russian, with direct speech predominant and indirect speech virtually nonexistent. He sees this as a reflection of a larger cultural tendency towards “authoritarian dogmatism”, which he also finds in medieval French: “If, at some given stage in its development, a language habitually perceives another’s utterance as a compact, indivisible, fixed, impenetrable whole, then that language will command no other pattern than that of primitive, inert direct discourse…” (ibid.: 128; see also 119–20, 123). Bakhtin/Vološinov’s assessment of reported speech in medieval Russian is based on false premises; it ignores the selectiveness intrinsic to every form of reporting — even direct speech. Moreover, it is empirically wrong; it ignores several well-attested patterns of nondirect speech, including varieties no longer available in the modern language. In reality there was a great variety of reporting strategies, with pronounced differences in how they were distributed both among and within diverse text-kinds. Bakhtin/Vološinov’s conclusion could have been reached only by extrapolating from the modern language; indeed, the words “primitive, inert” (“primitivnoj, inertnoj”, Vološinov 1929/1993: 138) betray an a priori, anti-uniformitarian assumption, also seen in other historical studies of Russian, that reported speech must be evolving toward greater diversity and expressiveness — i.e., that it must be more developed (however that is to be measured) in the modern language than in premodern texts (see D. Collins 1996 for discussion). In fact, Bakhtin/Vološinov did not take the necessary step of looking for “the teleology of the authorial context” in his medieval sources; he neglected to consider their individual traits and the kinds of contexts that tend to occur in them, all the while conjecturing large-scale tendencies that essentially obviated individual intentions. Similar methodological errors may be found in most other studies of Old Russian reported speech (even those not influenced by Bakhtin/Vološinov’s work). To be methodologically valid, a functionalist/pragmatic approach to reported speech (or any other complex of syntactic and lexical alternatives) in premodern texts must examine the usage in several synchronic slices
xvi
P
thoroughly before advancing hypotheses about large-scale diachronic developments. This, in turn, requires painstaking attention to particular contexts (linguistic, textual, and social/institutional), unbiased by the modern state of affairs or by premature panchronic generalizations. As in the New Philology, one must “recontextualize the texts as acts of communication” (Fleischman 1990: 37). However, the need to consider authorial intentions and communicative purposes in premodern texts encounters a substantial methodological difficulty. With only the written texts as observables, can one really recover all or any of the factors that motivated variation? Can one interpret patterns of speech behavior that cannot be observed directly (or introspected) but must be inferred from very partial context clues? These barriers are endemic to historical pragmatics, as a ramification of the general “Data Problem” (see Jacobs and Jucker 1995). I offer one solution to difficulties of this kind in the method of analysis demonstrated in this study. My investigation has three main goals, corresponding to three consecutive stages in my analysis. First, I set out to establish the norms of distribution for the various reporting strategies in a corpus of utilitarian texts; in particular, I try to determine which patterns were preferred in specific recurring contextualizations with known or inferable functions. This distinguishes my investigation from many other studies of reported speech, which concentrate on belletristic texts that do not feature recurring contextualizations of this kind. Second, I undertake to identify the pragmatic factors that, given the formal properties of the reporting strategies, could have justified these conventional preferences as the most effective means of accomplishing the communicative goals of the texts, which were, at least in part, socially institutionalized. Third, where there are departures from the conventions for a given context, I try to detect atypical features that could have motivated the scribes to choose unconventional strategies, again as a way of promoting optimal communication. All three of these goals serve a broader purpose — to explore how contextualization conditions reflect collectively and individually purposive use of speech-reporting strategies. The methodology and results of my study will, I hope, have relevance both for the discipline of historical pragmatics and for further functionalist research on reported speech. The investigation also serves to bring Slavic data, which have not been readily accessible to broader scholarship, to historical pragmatics, a field that up to now has focused mainly on Western
P
xvii
European languages. I hope, conversely, that it will begin a new direction of research in Slavistics by demonstrating the need for and advantages of function-oriented approaches to medieval Slavic texts, which have been little studied within the pragmatic framework.
Conventions for citing Cyrillic sources
Excerpts from editions of medieval texts are transliterated except in a few cases where the discussion focuses on spelling. Modern Cyrillic is transliterated according to the ISO/R9 (diacritic) system. The following system is employed for premodern Cyrillic: Transliteration a (not after j) b c c¦ d e e¦ f g i ja ju k ks l m
Cyrillic
Transliteration
Cyrillic
n o p r s š šc¦ t u (not after j) v x y z z¦ ’ ”
In excerpts from diplomatic editions, superscript letters are brought down to the line; abbreviations are expanded, with the tilde (titlo, a diacritic indicating abbreviations) omitted and the supplied letters enclosed in parentheses. For example, the ubiquitous abbreviation (gsne) ‘lord-’ is cited as g(o)s(podi)ne. Where the editions are not diplomatic, I follow, of necessity,
xx
C C C S
the editors’ conventions. I also retain the capitalization, punctuation, and word-spacing conventions of the editions; however, I ignore their paragraphing conventions, which are not particularly consistent and are not needed for comprehension. Line and page breaks in the manuscripts are not indicated. Where the editors of the documents cited draw attention to mistakes or omissions in the originals, [sic] is inserted; any corrections that I supply is indicated by [sic — DEC] or, when further clarification is needed, by [sc.… — DEC]. A list of abbreviations used in the glosses and elsewhere may be found in the front matter. The glosses indicate case and number for nouns and adjectives, and tense, mood, person, and number for verbs; gender, aspect, and voice are only mentioned when they are directly relevant to the discussion. Singular number and indicative mood may be assumed unless otherwise indicated. Contiguous agreeing elements in noun phrases are grouped together inside square brackets. Mention-forms are cited in italics, in a normalized transcription based on the main entries in Sreznevskij’s dictionary (see the References). For example, the form s”kazati ‘say, tell’ is employed rather than skazati or skazat’. While Sreznevskij’s entry forms are, on occasion, archaic for the fifteenth century, I prefer this to the anachronistic modernizing found in the other major dictionary, SRJa. When the mention-form is of a lexeme attested in both an ecclesiastical (Church Slavonic) and a vernacular variant, e.g., rešcˇi and recˇi ‘speak, say’, respectively, the latter is preferred. Verbs are generally mentioned in their infinitive form. Variable elements in phraseologisms are indicated by the symbols X, Y, and Z. Grammatical information is given in abbreviated form (see the List of Abbreviations); generally speaking, the grammatical tags refer only to inflections and ignore inherent categories, except when they are directly pertinent. In citing primary sources in the main text, I provide brief bibliographical information. This is keyed to more detailed information about the transcripts, including the dates of composition and copying, which can be found in the Appendix. Other conventions are explained when they first occur in the discussion.
C 1 The pragmatics of reported speech
1.1
Reported speech as intention and as perception
Reported speech () is both a universal of the language capacity and a pervasive phenomenon in ordinary language use.1 Every language has some explicit means, and generally more than one, of encoding messages that evoke (“represent”) other messages, whether fictive or actual. It is hardly surprising that such reports should be ubiquitous in actual discourse, given the nature of linguistic socialization and the central role of communication (and dialogue in particular) in human society; so much of our experience consists of speech events that “talk about talk” may even predominate in language activity.2 For the same reasons, many statements that are not explicitly marked as are nevertheless far from being unique creations; they turn out to be inherently heterogeneous (more precisely, heteroglossic): “Each utterance is filled with the echoes and reverberations of other utterances…” (Bakhtin 1952–53/1986: 91). Speakers cannot be Adam, able to utter a word in isolation from the speech of others (ibid.: 93–94).3 As a natural-language model for conceptualizing pragmatic phenomena — verbal behavior and heteroglossia — performs what has been aptly termed a metapragmatic function of language. Reporters and interpreters act as naive pragmaticians, analyzing and classifying speech events in accordance with culturally specific conventions (cf. J. Collins 1987: 71; Silverstein 1985: 134–35, 1993: 55).4 Indeed, encapsulates three of the major concerns of the pragmatic enterprise, since it represents the functioning of language in particular contexts, as seen from particular points of view. There is another, perhaps more tangible reason why is of particular significance for pragmatics: the very category cannot be defined without reference to the pragmatic factors of intention and perception. One of the intrinsic features of is its perceptual autonomy; reporters intend reports to
2
R V
be interpreted as heterogeneous with the surrounding co-text — as insets in frames (see Sternberg 1982b: 108–9). Though much of the information communicated in everyday conversation comes at second or third hand, often it is neither meant nor perceived as . This can be recognized by folk wisdom; thus statements that are not ostensibly reported can be challenged with rejoinders such as “Who says (so)?” or “Where did you hear about that?”, which violate conversational cooperation. Conversely, information that is not formally marked as can be intended as such, as in free indirect speech () and free direct speech (). (Positing a deleted verb of saying here is, for pragmatic purposes, a non-explanation that mistakes the very nature of these strategies.) Thus the communicative success of a report depends in part on whether the interpreter perceives it as heteroglossic, as intended by the reporter; nonrecognition can lead to miscommunications, as is depicted in the following dangling conversation: (1) “ ‘Hullo! hullo — ullo! oh, operator, shall I call thee bird or but a wandering voice?… Not at all, I had no intention of being rude, my child, that was a quotation from the poetry of Mr. Wordsworth…’ ” (Sayers 1927/1987: 77).
A second pragmatic aspect to the definition of is that reports are mediated by mental images of speech events in both the production and the interpretation processes; they do not come directly from the represented speech event. This is true even when there is an actual anterior utterance and even when the mode of reporting is direct speech (). Reporting always has an intentional and creative character. No matter how closely a report may approximate some anterior utterance, it is never mechanical reproduction of form and content (see Chapter 3); rather, it is a token that evokes a mental image — an ideal type — for both the reporter and the audience (Fludernik 1993: 17; Sternberg 1982b: 108). The mediation inherent in representation leads to a third pragmatic aspect of . Given that reports are not just fragments of inviolable prior text “repeated”, with or without paraphrase, in a parrot-like, unintentional, decontextualized manner, their meaning must necessarily be constrained by their contextualization. In other words, they are of necessity communicatively subordinated (Sternberg 1982b: 109) to the enframing discourse and to the illocutionary goals of the reporter in the ongoing speech event (cf. Bakhtin 1929/1971: 177–78). Thus, for example, a sound recording of a human voice
T P R S
3
is not until someone forms a mental image of its meaning and chooses to present it to interpreters in a specific context, for his own communicative purposes. The goals of the reportee (attributed speaker), if any, are most often irrelevant. The act of representation that mediates between reports and their anterior utterances (real or projected) not only allows but even compels reporters to impose their will upon both form and content through acts of “responsive understanding” (Vološinov 1929/1986: 122–23), which can include selection, choice of reporting strategy and contextualization, condensation or amplification, and evaluation. The pragmatic factors that are integral to the very definition of play a central role in each individual act of reporting. Thus, in examining the patterns of in a given language, it is not enough to treat them (as is done in most of the previous studies of Old Russian) in a decontextualized manner as an inventory of mere syntactic constructions, as if they were functionally equivalent and randomly distributed (see 1.3). Rather, one must establish the contextualizations of the various strategies and determine how they relate to the intentions of the reporters. The goal should be to achieve an understanding of that is as contextual and as close to the “emic” interpretive framework of the participants as possible — that is, to create a “thick description” of reporting, one that does justice to the “multiplicity of complex conceptual structures, many of them superimposed upon or knotted into one another”, which are involved in any social activity (Geertz 1973: 10).5
1.2
A “thick description” of reporting
The fact that reporting is an intentional activity, an exercise of the will, does not imply that every step in the process is necessarily conscious. Speakers acquire an ability to reason from ends to means as part of their pragmatic competence (cf. Brown and Levinson 1987: 58, 61, 64, 85); this ability is automatized to such an extent that it is ordinarily not a subject of introspection.6 Thus intention can, but need not, imply deliberation or calculation; it should be understood as a mental state oriented to a goal in the world — an “operation-order” that motivates action (the means to the end), often on a subconscious level (Austin 1966/1989: 274–77, 283–86).7 Intentional acts are by nature purposive, directed at effecting a desired state of affairs or avoiding an undesired one.
4
R V
Individual acts of reporting are the product of a series of choices (cf. Leech and Short 1981; Thompson 1996), each of which is intentional action. In selecting a particular strategy, reporters act on a sense of its appropriateness for the context, in anticipation of some desired effect on the audience; their own communicative goals — their speech wills or speech plans (Bakhtin 1952–53/1986: 77) — are a privileged factor. The reporters’ choice is informed by their knowledge of the typical distribution and functions of the strategy, which they acquire as part of their general pragmatic competence and their competence in particular genres; it is validated by their own experience as interpreters. As will be discussed in 1.4, one way in which the historical pragmatician can reconstruct this competence is to uncover the patterns of that are conventional in specific genres and correlate them with the social and textual functions of those genres. The first choice that reporters face in presenting information that they perceive as heteroglossic is whether to make reportedness (the separate “voice” or viewpoint) an issue in the discourse; that is, they must decide whether to single the information out as , attributable — though not necessarily attributed — to themselves or to some other speaker, or else cue the interpreters to treat it as part of the authorial (non-reportive) discourse. As noted in 1.1, much of our everyday discourse is objectively heteroglossic but never marked or even perceived as such; the essential issue is thus what may be called subjective heteroglossia — whether the speakers or writers proceed with a mental image of a separate source. The degree of explicitness, i.e., the extent to which reporters leave the burden of inference to the interpreters, is determined by their communicative goals and judgments about relevance. This preliminary decision is by no means trivial; it depends in large part on the reporters’ ideas of what is relevant for the interpretation process (cf. Sperber and Wilson 1986). It is not just a matter of attribution as opposed to “the text… averring anything which is not specifically attributed to another source” (Thompson 1996: 506); reporters can intend for information that is not explicitly indexed to be perceived as heteroglossic, as in much and , as well as allusive quotations like (1) in 1.1. Reportedness can be signalled overtly, e.g., through tags, nonnarrative features within the reports, intonation, or graphic devices. Alternatively, it can be accessible to the interpreters through ordinary inferential channels, e.g., shared knowledge about probable information sources, recurrent discourse frames, or genrebased “reading conventions that trigger an interpretation in terms of speech
T P R S
5
or thought representation” (Fludernik 1993: 7; see also 15–16; Fónagy 1986: 284). The ability to recognize and utilize such pragmatic entailments develops in the course of socialization (see Goffman 1981: 150–51; Hickmann 1993). Having undertaken to report explicitly rather than “aver” (Sinclair 1988), reporters proceed to select a formal strategy — a decision that comprises several steps.8 Analyzing the steps inevitably distorts them into seeming deliberate; in fact, like other intentional acts, they are usually automatic. There are two main clusters of choices, one involving the tag (how the report is indexed or attributed in the authorial context), and the other the mode of reporting (how the report itself is configured). The possibility of alternative strategies is undoubtedly universal, though some languages, or synchronic states of a single language, have a larger repertory than others. The first decision in the tagging process is, obviously, whether or not to have any explicit attribution (formal indexing) — a choice that, as mentioned above, does not constitute the difference between and unacknowledged prior text. The second decision is what kind of tag to use (if any). Crosslinguistically, the most widespread and important kind of tags are specialized verba dicendi (’s).9 In addition to such semantically faded verbs (e.g., English say, Old Russian s(”)kazati/s(”)kazyvati ‘say-/’), reporters in many languages can use lexically more specific verbs from semantic classes broadly connected with communication, including phase verbs, speech-act verbs (’s), and manner-of-speaking verbs (“graphic introducers”, Tannen 1986: 322; see also Fónagy 1986: 264–75, Silverstein 1985: 137). For example, in a recent novel in Contemporary Standard Russian (), six different verbs or verb phrases are used to attribute as many instances of indirect speech () in a single paragraph — skazat’ ‘say’, sprosit’ sebja ‘wonder aloud’, predpoložit’ ‘surmise’, vydvinut’ versiju ‘advance a scenario’, zajavit’ ‘declare’, and zasvidetel’stvovat’ ‘bear witness’ (P’ecux 1990: 249). In the same chapter (ibid.: 244–53), there are 49 cases of attributed direct speech (); 18 are tagged by the default verb skazat’, and 4 by less common ’s — soobšcˇit’ ‘report, inform’ (2×), molvit’ ‘say (obsolete)’ and progovorit’ ‘say, utter.’ There are 16 instances of ’s used to attribute : sprosit’ ‘inquire’ (7×), otvetit’ ‘answer (3×), otozvat’sja ‘respond’ (2×), soglasit’sja ‘agree’, zajavit’ ‘declare’, vozrazit’ ‘object’, and raz”jasnit’ ‘explain.’ Five of the verbs tagging denote phases or realignments in the dialogue: vstupit’ ‘enter in’ (2×), nacˇat’ ‘begin’, prodolžat’
6
R V
‘continue’, and obratit’sja ‘turn to.’ Finally, there are verbs of mental activity — predpoložit’ ‘surmise’ (2×) and zakljucˇit’ ‘conclude’; manner-ofspeaking verbs — vskricˇat’ and voskliknut’ ‘exclaim’; and an attitudinal verb — nadut’sja ‘pout, sulk.’ It is inadequate to explain diversity of this kind as mere “elegant variation” (to use Page’s term, 1988: 27). In using a wide variety of tags, authors are trying to narrow the readers’ range of interpretive possibilities in order to further their own communicative goals. Such use of nuanced vocabulary, which is especially though not exclusively typical of modern literary languages, is “a speaker-based strategy” (Lakoff 1984: 483–84), in which the author gives the readers relatively little of the responsibility for sense-making — in this case, evaluation of the represented speech events. The repertory of tagging devices is not limited to verbs. In Old Russian (), for example, could be indexed by adjuncts (2a), citation particles or quotative markers (’s), textual conveyor nouns, or nominal labels (2b).10 (2) a.
A po skaske [služylyx ljudej [tot and by deposition- [of.service people]-. [that Semejka ubit v [Pegoj orde Semejka]- kill- in [Skewbald Horde]- And, according to the deposition of the service people, the aforementioned Semejka was killed in the Skewbald Horde (1648; Tokarev (ed.) 1970: 896).
b. [V”spros” Inokentiev: “Gosudar’ Pafnotej! [question Innokentij-]- [lord Pafnutij]- Poveli… napisati zaveˇšcˇanie o [monastyr’skom bid- write- testament- about [of.monastery stroenii…” order]- Innokentij’s question: “Lord Pafnutij! Bid [someone] write down [your] testament about the monastic rule…” (ca. 1478; Dmitriev and Lixacˇev (eds.) 1982: 496).
Another arena of choice is found in the composition of the tag clause apart from the actual quotative device. In , reporters were faced, inter alia, with decisions about tense and aspect, which could be varied for different effects,
T P R S
7
and about the order of elements within the tag, when it was not dictated by rules of information structuring. A further choice — one with ramifications for the salience of the report in the discourse — involves the position of the tag vis-à-vis the report. For example, in , tags can be preposed (3a), postposed (3b), or intercalated (interposed/medial) (3c). (3) a.
I igumen mitropolitu tak rek”: Jaz… and abbot- metropolitan- thus speak- I- xožu po [staroi pošline… go-.1 by [old custom]- And the abbot spoke to the metropolitan in this way: “I… govern according to the old custom…” (1391; ASÈI 3: 16, no. 5).
b. i vy budeˇte mneˇ v” s(y)ny i and you-. be-.2. me- in sons-. and dšcˇeri g(lago)let’ [g(ospod)’ vsedr”žitel’ daughters-. say-.3 [Lord Almighty]- And ye shall be my sons and daughters, saith the Lord Almighty (2 Corinthians 6: 18; Gennadij 1499/1992: 247). c.
i [t”t” pop” s [temi tvoimi gorodskymi and [that priest]- with [those your- of.city ljudmi… [dvorjan moix pere[bi]li: a people]-. [servants my]-. beat-. and bili, skazyvajut, na sm(e)rt’ beat-. say-.3. to death- And that priest with those subjects of yours from the city… beat up my servants; and beat [them], they say, [almost] to death (ca. 1451; ASÈI 3: 25, no. 9).
I treat intercalated reports as a separate strategy rather than a subtype of postposition; this contrasts with the way the phenomenon is viewed by some other scholars (e.g., Hermon 1979; Green 1980 with reference to English). My reasons for this relate specifically to premodern Slavic languages, though I suspect that they have a broader typological bearing. First, conflation of intercalation and postposition obscures a striking difference in distribution: as in English (see Partee 1973: 411; Hermon
8
R V
1979), postposition is rare and “literary” in , as in excerpt (3b), a biblical text in the ecclesiastical (Church Slavonic) register. (One must except postposed tags after complementized , which are common in some textkinds, including legal-administrative writing, when the complements are presupposed — e.g., from preceding stimuli-questions.) By contrast, intercalation is common and found in texts of all registers. Indeed, it is fairly frequent in everyday speech in . Second, like other parentheticals, intercalated tags in tend to appear in second position — after the first phonological word (Wackernagel’s position) or, in later texts, after the first major constituent (not in any position, as claimed by Molotkov (1958: 31)). Thus, when first position is occupied by subordinators, intercalated tags can precede reports, as in (4): (4) Jaz, gospodine, pomnju za sorok let, I- lord- remember-.1 for forty- years-. cˇto, gospodine, kažet, [to mesto vyprosil [otec’ that lord- say-.3 [that place]- request- [father Ivanov u Anny… Ivan-]- at Anna- “I, lord, remember for forty years that, lord, ([he] says) Ivan’s father obtained that place from Anna…” (ASÈI 3: 291, no. 276).
Here the intercalated verb kažet” lies outside the syntactic structure of the complement clause, like the vocative that precedes it; the fact of the is not being asserted as part of the represented speaker’s forty-year-old recollection. (In (4), it is ambiguous whether the subject of the intercalated verb is Ivan or the represented speaker of the , i.e., whether the tag belongs to the or the authorial narrative.) Third, there is a strong tendency in for intercalated ’s to have implicit subjects — “the ultimate ‘backgrounding’” (Chvany 1973/1996: 122) — whereas postposed ’s, at least in third-person narratives, generally have explicit, asserted subjects. (This is mandatory for postposed ’s in ; see ibid.: 121–22.) Finally, the discourse function of intercalated reports may differ from that of postposed ones. In this study, I present evidence that intercalation was favored in certain contextualizations precisely because of its interruptive nature (see 7.6). The reporters’ choice of tagging strategies faces an additional complication in reports that contain more than one predicate; here the reporters must
T P R S
9
decide whether each noninitial clause requires supplementary tagging. Thus in multiple attributions are attested, with preposed tags reinforced by intercalated ones (5): (5) [Knjaz’ Velikij Ivan” Vasil’evicˇ’ skazyvaet”: na cˇem” k” [prince grand Ivan Vasil’evicˇ]- say-.3 on what- to vam”, k” [svoej otcˇineˇ, rekl”, po tomu you-. to [. patrimony]- speak- by that- vas”, skazyvaet”, žaloval”… you-. say-.3 favor- Grand Prince Ivan Vasil’evicˇ says, “On [the basis of] what I have said to you, my patrimony, in accordance with that”, he says, “I have granted you the boon…” (1471; AI 1: 512–13, no. 280).
Cross-linguistically, there is considerable diversity in reporting modes.11 However, certain patterns are typologically widespread, and the choice between tagged and untagged (free) , in particular, is probably universal. Within a single language, the range of formal possibilities can be quite broad. For example, had both , as in (2b), (3a), (3b), and (5), and , as in (3c). Both of these modes are found in tagged and (as I shall argue) free varieties, and, when tagged, in complementized and uncomplementized forms. Thus even the use of one of the cardinal modes of reporting involved at least three choices. Also attested are deictically ambiguous reports (often treated, imprecisely, as ), as in (2a), (4), and (6) [1], below; intermediate forms reflecting slipping from one category to another; and various kinds of narrative reports of speech acts (’s). Both ’s and ’s could take reports that were highly reduced and integrated into the main-clause predicate, as in the small-clause structure seen in (6) [2], or the accusative with infinitive or with participle found in ecclesiastical (Church Slavonic) texts.12 (6) skazal brat ego… cˇto [siju duxovnuju Aleksandr say- brother- his that [this will]- Aleksandr- pisal, a ego skazal bolnogo write- and he- say- ill- [1] His brother said… that Aleksandr wrote this will, and [2] he said that he [Aleksandr] was ill [at that time] (1472; ASÈI 3: 100, no. 67).
The inventory of report-clause types found in was thus larger than in ,
10
R V
which reflects a long-term tendency to eliminate the more bound forms (except after volitional ’s). It is important to keep in mind that formally distinct reporting strategies are not, as a rule, equivalent in function, even when they appear synonymous. While they all make reference to the occurrence of a speech act, each conventionally emphasizes different aspects; in Bakhtin/Vološinov’s classic formulation (1929/1986: 129), each “‘hears’ a message differently; it actively receives and brings to bear in transmission different factors, different aspects of the message than do the other patterns.” To express it less anthropomorphically, each strategy allows the reporter to guide or manipulate the interpretation process in a different way: “Different ways of framing cue listeners that such has different functions or meanings in the discourse, and is to be interpreted variably” (Philips 1985: 168). This principle may be illustrated by ’s such as English assert and affirm, which cannot be used interchangeably despite their near-synonymy (Wierzbicka 1987: 322). Even coexistent ’s that are semantically faded differ from one another in function; each has a distinct perspective on the “scene of linguistic action”, as may be seen by examining their contextualizations (see Dirven et al. 1982, especially 166–67, 169; Goosens 1987). For instance, in colloquial American English, the default verb say tends to be used with reports of others’ speech, the emergent verbial be like with selfquotations and reported thoughts, and go with sounds (Romaine and Lange 1991: 237–38, 240, 243). Likewise, in Navajo, one of the two most frequent tags has a “discourse-segmenting function, focused on the speaker and the ‘expressive’ aspect of speech”, while the other has “a discourse-continuing function, focused on the speaker-hearer dyad and the ‘interpersonal’ aspect of speech” (J. Collins 1987: 83). The particular implications that the individual reporting strategies are suited to convey by virtue of their form determine both their syntactic combinability and the contexts in which they will be preferred in a purposive use of language — a fact that allows the important methodological step of contextualization-to-function mapping (see 1.4).
1.3
Some shortcomings of nonpragmatic reductionist approaches
While the distribution of forms is crucial in the contextualization-to-function methodology outlined below, one must bear in mind that the consideration
T P R S
11
of formal distinctions is only a preliminary to explaining the functional differences among the strategies. However, many previous studies of in and other languages have not moved beyond this first step. This is the case, for example, with the many reductionist syntactic approaches, which have been chiefly concerned with positing rules for converting one construction into another (“grammar-as-usual”, Silverstein 1985: 143). Such analyses are operating, in effect, on the basis of a covert assumption that the constructions are equivalent in any linguistically interesting aspect. Though still widespread, this notion was debunked already by Bakhtin/Vološinov (1929/1986: 128): “This sort of implementation of the patterns of speech reporting has nothing even remotely to do with their real existence…. Each pattern treats the message to be reported in its own creative fashion, following the specific direction proper to that pattern alone.”13 In this light, I think it is important to point out that exclusively syntactic, clause- or sentence-level approaches to are incapable of dealing with the entire phenomenon, even in its formal aspects, for the simple reason that is a category of discourse analysis rather than syntax. It does not form a coherent group of syntactic constructions; the features that distinguish the various modes often relate to the level of text rather than to that of sentence or clause. For instance, in many languages, there is no properly syntactic difference between and non-reportative narrative (cf. Hagenaar 1996; Paducˇeva 1996: 343–44, 347).14 Likewise, in languages like and , where is not marked by backshifting or mood changes, it is doubtful that complementized and are really distinct syntactic constructions, despite the long tradition of treating them that way; the complementizer has the same explicative function in both cases. In fact, the two strategies, like and in general, are differentiated by features that are not syntactic, given that the deictic orientation point can only be determined by reference to the larger discourse. Though frequently treated as object clauses,15 reports are entities of a different order than syntactic constructions; this is shown, inter alia, by the fact that reported information can be the sole content of clauses belonging to other, well-defined syntactic categories. For example, topicalizing clauses in , as in (7), often consist entirely of plus a citation particle — approximately, “in re ‘X’”.
12
R V
(7) cˇto dei po teˇx po ix xrest’jan… that for those-. for their peasants-. priezdjat [pristavove moi… i jaz ix… come-.3. [constables my]-. and I- them- požaloval… favor- [As for the fact] that []… my constables come after those peasants of theirs… I have granted them a favor… (1462; ASÈI 1: 215, no. 304).
In (7), the only explicit signal of reportedness is the particle deˇ(i) (in the reduced form dei), a grammaticalized third-person singular nonpast or imperative of the archaic deˇti or deˇjati.16 The information topicalized in the cˇ’to (’that’) clause is the content of the report, not the fact that the report occurred. The same use of can be found in causal and relative clauses. Likewise, in Hungarian, “there are as many categories of reported sentences as subordinate clauses” (A. Dömötör, cited in Fónagy 1986: 260). Obviously it would not be desirable to double the inventory of clause types by distinguishing reportative from non-reportative varieties. A further problem for a purely syntactic approach is the fact that the varieties of form a continuum, with indeterminate boundaries between the individual types.17 Without the benefit of intonational or graphic cues, thirdperson reports that lack deictic elements coreferential with the ongoing speech event cannot be classified in a principled manner as either or . This is illustrated in the example in (8), given in phonetic transcription to avoid graphic signals of or : (8) [m%rjíj6 sk%zál6 sjlj7´ dujuw jw j6j6 stantsGj6 plów jw j6tj Marija- say- [next station]- square- j n6g iná] Nogin- Marija said the next station [is/was] Nogin Square. “The next station [is] Nogin Square.”
The ambiguity exemplified in (8) can exist even in the presence of intonational cues, since the prosodic features that are often said to characterize are optional and can actually accompany as well (Klewitz and Couper-Kuhlen
T P R S
13
1999). Such indeterminate cases are not amenable to a purely syntactic treatment. The same is true of deictically ambiguous clauses or sentences within reports that otherwise belong to a clear-cut mode (e.g., ). It would be uneconomical to treat the indeterminate portion as anything other than ; yet this classification depends entirely on the larger discourse context. A third problem is that reporting strategies are not coextensive with clauses or sentences, for all that they are treated under the rubric of compound sentences in many grammars. On the one hand, a single report can contain an indefinite, potentially large number of sentences. The main narrative in I. S. Turgenev’s novella First Love (1860) is one long report — memoirs “quoted” in their entirety and framed by a brief outer narrative ending in a colon. On the other hand, complete reports or stretches of a particular reporting strategy can be smaller than a clause, as in the “incorporated” or “embedded” quotation in (9) [1] (see Clark and Gerrig 1990: 791; Semino, Short, and Culpeper 1997: 31): (9) [1] The notion that [Solzhenitsyn] is “anti-Western” is wrong, and [2] “arose out of the inordinate sensitivity and superficiality of Western correspondents”, he said. “My speeches to the A.F.L.-C.I.O. in 1975…” (Remnick 1994: 74).
This excerpt illustrates a further way in which strategies are not coextensive with syntactic units — the phenomenon of slipping from one reporting mode to another in midstream (“fadein, fadeout”, Tannen 1989: 117–18; Schuelke 1958). In (9), the (with partial ) of the first clause [1] — with a thirdperson subject referring to the reported speaker, Solzhenitsyn — yields to unambiguous (“my speeches”) [2]. A more clear-cut instance of slipping may be seen in (10), from an Old Novgorodian (northwestern Russian) birchbark letter of ca. 1200–1220, where the speaker, urging her brother to deliver a message, suddenly switches from oriented on his perspective to dominated by hers: (10) ty že [brace gospodine molovi emo [sic] tako ože you- [brother lord]- speak- him- thus if budu ljudi na [moju s’tru ože be-.3. people-. against [my sister]- if budu ljudi pri komo budu dala be-.3. people-. before whom- ..1 give-
14
R V
ruku za zjate to te ja vo vine hand- for son-in-law- I- in fault- But you, lord brother, speak to him in this manner: “If there are people [to witness] against my sister”, if there are people before whom I shall have given surety for [my] son-in-law, then I am at fault (Zaliznjak 1995: 344, no. 531).
In a purely syntactic approach, cases of this would probably have to be treated as ungrammatical (anacoluthic). Partee (1973: 411) excludes examples of slipping from her generative analysis on the basis of “the admittedly prejudiced belief that such sentences do not occur in ordinary spoken language.” In fact, slipping and other “mixed” forms of are common in ordinary discourse in every language with which I am familiar.18 It is beside the point that they do not seem to be syntactically well-formed; they are phenomena of discourse, of text at large, and cannot be explained away as contaminations. Moreover, slipping can be both purposive and intentional (cf. Aaron 1992; Yule 1993). A final difficulty for exclusively syntactic approaches is the fact that the tag and the report can be in entirely different codes, as in the example in (11): (11) — Mais c’est un palais, — skazala ona mužu, say- she- husband- ogljadyvajas’ krugom… look- around “Mais c’est un palais!” she told her husband, looking around… (Tolstoj 1978: 95).
There can even be changes of medium. For example, in a recent English version of “Goldilocks and the Three Bears”, the attributions, including that of the narrator (“story-teller”) are presented as consistent visual icons in the left margins (Pearce-Higgins 1989). It would be unsatisfactory to fragment the category of by treating “amenable” varieties under a syntactic rubric, separately from “intractable” ones; indeed, such a stratagem could lead to inconsistencies. An example of this may be seen in an authoritative grammar of , which declares that with the deictic orientation of the represented speaker — the hallmark of
T P R S
15
oratio recta — is not if it is “incorporated in the narrative without introductory words” in some position (Vinogradov and Istrina (eds.) 1960: 403). Yet the tag clause is surely not constitutive of , since it also occurs with nondirect reporting modes, while the configuration of features that does define tagged is shared with the untagged variety. Moreover, the tag clause itself does not have a clearly syntactic function, as opposed to textual ones such as promoting coherence. Many of the syntactically problematic characteristics of discussed above also prove to be stumbling-blocks for lexical approaches, in which the choice of complement type is explained by the existence of a certain semantic component in the matrix verb (see Silverstein 1985: 143). Approaches of this kind have considerable explanatory power in dealing with the clausal complements of ’s.19 However, their applicability to as a whole is less clear. They would be unable to deal with as a unified phenomenon unless they posited underlying tags to account for the untagged varieties (cf. Wierzbicka 1974: 294–97); even then, another stratagem would have to be found to account for unbound forms attributed by adjuncts — According to X, In X’s view, etc. In addition, it is problematic to treat reports in , which are open-ended in principle and frequently lengthy, as “complement types” of the same order as the clausal complements of ’s such as order or persuade. It is also unclear how such lexical approaches could account for the multiplicity of complement types that can be taken by ’s in and other languages. One would not want to posit polysemy in the matrix verbs, nor could one correlate the choice of the alternative complement types with different semantic components, at least without drastic oversimplification. Thus some other methodology is needed to account for the reporting process in all its intricacy and richness. Conceivably, a reductionist approach could try to account for why reporters choose one configuration over another by positing one of three alternatives: randomness; syntactic change reflected in synchronic dynamism; or stylistic differences. The first of these is a patent nonexplanation (and perhaps a straw man). If anything seems random in the way that pragmatically competent speakers use higher-level units, apart from genuine mistakes, it is a matter of patterns that have not been detected, either because of their complexity or because of methodological inadequacies (cf. Diver 1969). The appearance of could indeed be taken as random in the trivial sense that other constructions could appear in its stead. However, actual instances of
16
R V
occur only within discourses — contexts that contain motivations for the choice of one strategy over another. The second alternative, syntactic change, can undoubtedly affect . For example, long-term drift with continuous synchronic dynamism may be seen in the gradual decrease in the use of bound reporting modes after ’s in the history of Russian (see 1.2), or in German , where there is a tendency for the indicative to replace the subjunctive and, where the latter is maintained, for the present tense to replace the preterit (Hammer 1983: 265–66; Jespersen 1924/1992: 295–99). However, it must be emphasized that the extension of one strategy at the expense of another generally has pragmatic causes; it is a diachronic reflection of speakers’ synchronic choices, which are based on socially-motivated preferences (cf. Keller 1985). Likewise, new strategies develop as the result of individual acts that are exploratory and purposive: “As they select novel and traditional expressions in accordance with their individual hypotheses about their appropriateness… the speakers in effect negotiate the norms that they look upon as their community norms” (Andersen 1989: 25).20 Innovative strategies are acceptable to the community for pragmatic reasons, e.g., because they are communicatively effective or aesthetically pleasing. Thus it is not enough to offer exclusively historicalsyntactic explanations for the coexistence of reporting strategies (as is done in much of the literature on ); one must also uncover the factors that motivate the distribution in a given synchronic slice. The third alternative, stylistic differences, can likewise play a role in the reporter’s choices; however, they cannot account for every kind of variation. It is almost always the case that various reporting strategies coexist within the same styles, genres, and texts; the prevalence of one variety or other is motivated by the needs of the discourse. Unless mere taxonomy is the goal, one must propose some functional explanation to account for the overlapping distribution of reporting strategies or indeed any other stylistic phenomenon.
1.4
A method for historical-pragmatic analysis
Historical pragmatics is a comparatively new field of linguistics and hence one for which methodological issues are of great importance. One of the chief obstacles that must be overcome is what has been called “the Data Problem” — the need to draw conclusions about language behavior and
T P R S
17
conventions on the basis of written artifacts rather than direct observation (Jacobs and Jucker 1995: 6). Generally speaking, the only observables in premodern texts are the forms themselves. There tends to be only scanty information, if any, about the writers’ intentions, participation frameworks, and other factors of the speech situations in which the texts were produced; often such information has to be inferred from internal evidence alone. Except in rare cases of limited utility (e.g., early grammars and conversation books), native-speaker intuitions are not available as a check on explanations, and they cannot be elicited in any event. While typological data can shed light on some issues, they must be used with caution — ideally, only as a verification procedure — to avoid universalist overgeneralizations and petitio principii. Several solutions to particular aspects of the Data Problem are found in the literature (see Jacobs and Jucker 1995: 7–10); I offer another in the methodology employed in this investigation, which adapts the “method of residual forms” of phonological reconstruction (see 1.5, below), and which, I think, can have broader applicability in the field of historical pragmatics. The goal of this methodology is to establish conventions of usage and to reconstruct motivations for the choice between “alternative” — semantically similar but formally diverse — strategies. It is important to note that the methodology is intended to reconstitute conventions and motivations operative in written language, without reference to the norms of speech. This is in line with the direction of historical-pragmatic research that treats written texts not just as “imperfect renderings of the real thing” (i.e., the spokenlanguage data preferred in general pragmatics) but as “communicative manifestations in their own right and as such… amenable to pragmatic analyses” (ibid.: 9). The methodology outlined here has three prerequisites, two of which concern quantity of data, and the third quality. First, it must be applied to a reasonably ample corpus of texts, which should be homogeneous in genre in order to facilitate form-to-function matching. Second, the corpus must supply a large number of tokens of the strategies under investigation, which must appear in diverse recurrent contextualizations. This is a general requirement for drawing valid conclusions about grammatical and discourse phenomena in premodern languages; as Lyons observes (1968: 138), “The adequacy of the description [of rules in a dead language] will be proportionate to the amount and variety of the material upon which it is based.” Third, it must be
18
R V
possible to link the strategies with functions, and more than one function must be operative; that is, the contextualizations of the strategies must have known or inferable functions, which generally will reflect socially institutionalized speech events or complexes of speech events. Clearly, no methodology can make it possible to recover every possible motivation for the selection of a strategy — e.g., the contextualization of a reporting device — in a corpus of premodern writings; the real question is whether it is valid and feasible to offer even partial matches between form and function. The immediate answer is that one could not read and understand premodern texts at all without inferring, however imperfectly, some matches of this kind.21 Yet having enough knowledge of grammar and lexicon to do accurate close readings still does not guarantee that one can make form-to-function pairings that are “emic” — i.e., that even roughly approximate the pragmatic competence of the original writers. However, as when one is analyzing a modern foreign language, the risk of interference diminishes in proportion as one becomes familiar with the cultural setting and with a broad range of representative speech genres, which contributes to intuitions about typical cooccurrence patterns and contextualizations. In other words, the ability to make probable form-to-function matches depends, in large part, on the “capacity for a truly empathetic reading” — the ability “to transport oneself, as it were, into the very milieu in which [the texts] were produced, reproduced, and read” (Birnbaum 1985: 171–72). While conclusive verification may ultimately be impossible, the validity of an analysis can be appraised, at the very least, for its plausibility on culture-specific and typological grounds. The first step in the analysis is pragmaphilological, in that it relates to “the contextual aspects of historical texts” (Jacobs and Jucker 1995: 11–12). Its aim is to establish the diplomatic structure, real-world function(s), and reading conventions of the text-kind that furnishes the corpus. This information is indispensable for “empathetic reading”, for acquiring a sense of the knowledge and inferential patterns of the intended audience. The medieval Russian trial transcripts used in this study furnish relatively rich internal evidence for their situational context, and external evidence of somewhat later provenience may also be found, e.g., in law codes and in travel writings by foreigners; information of this kind is often not available in premodern sources (ibid.: 6–7). Trial transcripts will be described in some detail in the next chapter; for present purposes, it is enough to note that they typically
T P R S
19
consist of several distinct sections or divisions, which can be identified by their form, content, and relation to the different phases of the real-world trial. These are primary speech genres within the secondary or complex speech genre (Bakhtin 1952–53/1986: 61–62) of the transcripts as a whole. Each section has a broad institutionalized function that can be identified on the basis of internal evidence and other sources of information about medieval Russian law. For example, the first section, the trial record (sudnyj spisok) was intended to provide an impartial record of undigested testimony for later review; it was used to jog the memory of judges if the trial was delayed, and it served as the primary source of information for judges of higher instance. The second step in the analysis is to determine which kinds of represented speech events or situations predominate in each of the natural sections (form- and content-based divisions) of the text-kind — the incipit, trial record, judicial-referral record, verdict, and end protocols — and in any coherent subsections or recurring speech situations. For example, in the trial record section, the prevalent speech event is testimony that serves to establish the litigants’ claims and occurs, elicited by the judge, in a dyadic question-and-answer framework (see Chapters 2–3). Proceeding section by section or contextualization by contextualization makes it possible to match the form of the with the form and function of the surrounding discourse in a fine-grained manner. I define the predominant speech event or situation in a given section or contextualization as its prototype. To be sure, mere frequency is not a sound method for defining prototypes, since entities can be perceived as the most prominent or representative members of their categories without constituting a majority of the tokens (Givón 1989: 38–43). Nevertheless, the results of my investigation show that a majority — sometimes an overwhelming one — of the represented speech events in a specific contextualization do cluster around what I define as the prototype core (cf. ibid.: 42). While there is obviously no direct means of testing for perceptual prominence, the posited prototypes generally match the kinds of speech events that, on the basis of institutional function, one would expect to be central in the given phase of the trial. The third step in the analysis is to determine how the prototypical speech event in a given contextualization tends to be represented — that is, to match it with a preferred reporting strategy, determined, again, on the basis of frequency. The norms established in this manner are the “stabilized constructional patterns”, the “basic and constant tendencies in the active reception of
20
R V
other speakers’ speech [author’s emphasis]” that Bakhtin/Vološinov (1929/ 1986: 116–17) considers to be the chief object in the study of . To avoid overgeneralizations and concomitant loss of potentially significant information, I first identify the patterns in each individual transcript and then collate the results to determine the norms of the text-kind as a whole. The results justify my prior hypothesis: in each of the major contextualizations there is a distinct preference for a specific reporting strategy — a convention that differs from section to section. In referring to convention, I do not mean it dismissively, as if the strategies were merely inflexible, socially prescribed clichés. Conventions typically have a rational basis as mutually accepted precedents for coordinating interactions in a group; they arise when agents perceive the given action as an effective way of accomplishing some intention (cf. Lewis 1969; Brown and Levinson 1987: 56, 86). In linguistic terms, the group-coordination problem that conventions address is interpretation; besides facilitating the process of production, conventions serve as contextualization cues for the interpreter. In this light, a given reporting strategy could only have become conventional if it were perceived as appropriate (felicitous). That is, the writers, relying on their experience as interpreters — their responsive understanding (Bakhtin 1952–53/1986: 68–69, 78–79) — would prefer it only if they found it more effective than the other available forms of in carrying out the institutional function of the given section or in conveying salient pragmatic aspects of the messages that were prototypical for that context. Rather than being “bound” by the conventions, the scribes chose to use them as a proven means of expression (and perhaps also as a sign of their own enfranchisement in the culture of judicial writing). It would be missing the point to associate conventionality with non-purposive mechanization, especially in the case of recurrent compositional frameworks that can have varying content. The prevalent reporting strategies had a social purposiveness; each felicitous use, each successful perlocution validated and re-established the convention or norm. As Coulmas observes (1986a: 3), “In the enactment of verbal routines the creativity of language is socially canalized according to successful solutions of recurring verbal tasks, fixed by functional appropriateness and tradition.” Thus, instead of saying that individual writers “followed” the conventions, it would be more accurate to state that they participated in the conventions as a social institution and so re- or co-created them. To be sure, the very existence of a convention can be strong motivation
T P R S
21
in and of itself for following that convention (Lewis 1969). Hence a choice that is intentional for some agents can be made automatically by others. At first glance, this possibility seems to cast suspicion on the assumption that any individual report reflects a context-sensitive use of discourse strategies for specific effects — that is, effects other than the face-oriented one of being seen to conform to convention.22 To a large extent, this problem is resolved by locating the text-kind in its social context. Each section of trial transcripts had an institutional function that the preferred reporting strategies were proven to subserve efficiently. This rational motivation must have been accessible to the pragmatic competence of the writers who formulated the convention. Since, in order to persist, the convention must have been validated continually by felicitous uses, its motivation must also have been accessible to the enfranchised scribes who perpetuated it and who were imitated by less competent writers unable to perceive its rational base. Thus even scribes who were “merely” following convention were partaking, albeit at one remove, in institutionalized intentions, and the result — the effect on the intended interpreter — was the same. However, one may well doubt that a scribe could be competent enough to follow conventions appropriately while being utterly insensitive to the linguistic context. The fourth step is form-to-function mapping; the goal is to identify factors that could have motivated the correlations between reporting strategies and contextualizations established in the third step. This procedure entails a minute search for contextualization clues in the discourse, including, in particular, the known or inferable function(s) of the given section (speech genre) and the real-world speech events represented therein. Once I isolate these potential communicative factors, I examine how the formal properties of the strategies could have promoted the effective encoding of the message — the represented speech act — in the given contextualization, from the inferred standpoints of the writer and the intended interpreter. The pertinent formal properties (see 1.3) include whether there is a tag clause and, if so, how it is placed; what verbs of saying or quotative particles are used in it; whether there is a complementizer; and what deictic pivot is chosen for the report. Typological data on the communicative functions of analogous strategies in other languages often prove relevant in this phase of the investigation. It turns out that the majority — sometimes the overwhelming majority — of reports in each contextualization are presented by strategies that are not
22
R V
only conventional but also motivated. However, there is often a small residue of deviant or problematic instances. It has been said that pragmatic analysis does not need to account for 100% of the tokens in a corpus to be valid, given the inconsistencies of human behavior (cf. Hopper 1979: 221, Note 5); however, while this observation is valid on the whole, it does not absolve one from trying to account for the exceptions in a fine-grained analysis. One of the advantages of operating with prototypes is that they establish a framework in which to explain many of the exceptions in a non-random manner, so that the results in fact come closer to accounting for 100% of the examples in the corpus. The fifth step in the analysis is to account for the problematic tokens by means of the “method of residual forms” mentioned above. I examine the context and content of the exceptions (the residual forms) to isolate any features that distinguished the represented speech events from the prototypes. It is almost always possible to find atypical features in most of the deviant cases — a fact that validates both the posited prototypes and the methodology itself. Identifying additional contextualizations of this kind as potential conditioning factors leads to a second phase of form-to-function mapping — comparable to the fourth step, above — with the aim of discovering the properties of the unconventional strategies that would made them appropriate for such “highly entailing effects” (Silverstein 1985: 142–43). This accounts for nearly all of the apparent exceptions. Many of the deviations fall into classes that may be viewed as speech genres in their own right — e.g., choral reports (see 4.3) or boundary-setting descriptions with a mix of gestural testimony and spoken commentary (see 4.1–4.3). Other deviations are unique but still amenable to classification, like the case in which a litigant calls one of the judges as a witness — a striking change in footing (in the sense of Goffman 1981: 137). Indeed, many of the exceptions appear to index changes in the participation status or configuration of the participants in the represented trials. By now it will have become clear that the methodology outlined above presupposes not only that the writers of the investigated documents were rational in their choices but also that they were pragmatically competent in the speech genre. The latter assumption may not be self-evident; given the large percentage of conventional tokens, one might suspect that the lowfrequency residual forms were not purposive but rather inappropriate usages. While this cannot be ruled out, there is no reason to ascribe residual forms
T P R S
23
to incompetence if they can be linked with motivations that could conceivably be rational in the context.23 Indeed, generally speaking there would seem to be no particular justification for blaming residual forms on incompetence unless there are other, independent signs of it, e.g., startling departures from the diplomatic norms of the text-kind or (perhaps) grammatical mistakes that are not the result of recopying. In any case, incompetence is not likely to be an adequate explanation for deviant usages in the trial transcripts used in this study, since the majority of the documents show evidence of having been read and approved by their intended interpreters, the judges (see Chapter 2). This is not to deny that some usages may have been less optimal than others; however, degrees of effectiveness cannot be recovered on the basis of the given methodology (or any other, to my knowledge). The low frequency associated with residual forms cannot be used by itself as evidence of lesser effectiveness. To do so would conflate inept and successful creative effects; that is, it would confuse the pragmatically weakest strategies with the pragmatically strongest ones, those that are “foregrounded” in the Praguean sense. The line between these two types of unexpected usage can only be drawn on the basis of a deep knowledge of the milieu in which the writing took place.
1.5
Previous applications of the “Method of Residual Forms”
Though new to historical-pragmatic research, the “method of residual forms” has been applied in a few recent anthropological-linguistic studies involving limited corpora of Amerind texts (Silverstein 1985; J. Collins 1987). It derives from a key principle of phonetic reconstruction codified by Bloomfield (1933/1984: 351–60), among others: before rejecting exceptions to postulated sound-laws as false cognates, one should determine whether they are members of some subsidiary correspondence that has yet to be identified. The results of further analysis have often validated this second alternative: “The fact that residues have again and again revealed new correlations is a strong confirmation of [this] method” (ibid.: 351). In the approach outlined by Silverstein (1985: 142–43), which I follow to a large extent, the basis for adapting the method of residual forms to the analysis of discourse is the fact that “there are normal, ‘appropriate’,
24
R V
presupposed contexts for use of particular forms… and highly marked contexts, unexpected for the occurrence of such forms, though occur they do with a ‘foregrounded’…, highly entailing or creative effect” (ibid.: 142). Accordingly, the first step in understanding the distribution of strategies is to determine “the basic usages of particular forms as appropriate to certain discourse contexts, and formulate an hypothesis about pragmatic (indexical) meaning in text as a representation of something.” Subsequently the unexpected forms in a given context can be explained as “more marked versions of those in the first, more ‘automatized’ set.” This program calls for a few words of commentary. The idea that the basic usages are “automatized” seems problematic, at first glance; it could be construed as reducing the basic usages to mere defaults; which would beg the question of whether they are used purposively. This obstacle is only apparent, given that automatized usages can be intentional (see 1.2). However, it is necessary to make a proviso here: even though the basic usages — in my terminology, the preferred strategies — can indeed be automatized, it is important to remember they are not necessarily so. Paradoxically, in the analysis of premodern discourse, the only way to gauge whether strategies were automatized is to find what seem to be inappropriate uses (infelicitous extensions or contaminations), which, in accordance with the method of residual forms, must be first treated as possible reflections of an additional distinction. Another potential problem in the methodology, as sketched by Silverstein, is its use of quasi-Praguean binary oppositions (“marked” vs. “unmarked” forms), even though this is phrased in relative (scalar) terms that bring it closer to prototype classification. To begin with, the concept of markedness, as it is usually understood, presupposes some “mark” or feature that defines the marked category; but complex strategies like reporting methods tend to be characterized by a number of features, any one or any combination of which may make them appropriate for a specific discourse contextualization. In addition, unmarked categories are generally simpler than marked ones (see Battistella 1990: 27); however, it is far from clear that this is true of the categories identified as unmarked (“basic usages”) in Silverstein’s approach. I know of no way in which to measure the relative simplicity of different ’s, i.e., verbs indicating utterances without any further lexical specification; in any case, there can be contexts that favor greater specificity without any foregrounding effect — e.g., in my data, tags for the judges’ reports (see Chapter 5). As for reporting modes, it is debatable — to
T P R S
25
take only one example — that , which is typically preferred, e.g., with reports that are foregrounded in narrative (see 3.5), is actually simpler than alternative strategies like , since it involves “decentering” or displacement of indexical ground (Hanks 1990: 206). Assuming for the sake of argument that it is, one can still find many contexts in which putatively more complex reporting methods are “basic usages” (cf. D. Collins 1996; Ebert 1986). (This, indeed, is a general problem for the concept of markedness in grammar, epitomized by the preference for the marked perfective in narrative, the prototypical past-tense contextualization.) Finally, the results of my investigation do not bear out the notion (Silverstein 1985: 142–43) that residual forms invariably have the foregrounding effect predicted for marked forms, whether in the Praguean sense of “actualization” (Havránek 1932, 1964) or in any other sense. While they undoubtedly have “highly entailing effects”, these do not necessarily equate with foregrounding. The results of my analysis suggest that an element can be contextually unexpected without being prominent; this is particularly the case when the expected element is itself associated with a relatively high degree of foregrounding. A program of analysis that is similar to the method of residual forms has been outlined by Brown and Levinson (1987: 282), who observe that “nonoccurrence of expected forms, ‘exceptions’ and ‘breeches’, can be made sense of by analyzing the rational sources for those occurrences in an otherwise homogeneous style with other rational sources.” A further kindred method has been proposed by Derjagin (1971: 151–52), whose approach derives from the ancillary historical discipline of diplomatics. Derjagin’s methodology is interesting in the present context primarily because he applies it to medieval Russian legal/administrative texts. His first step is to establish the syntactic and lexical patterns that are standard in a given textkind. Next he identifies features characteristic of individual writers by comparative analysis of a large corpus of texts. The final step is to juxtapose the works of a single scribe to isolate any peculiarities that may be ascribed to the principal for whom the document was written (the person who commissioned the document and who is explicitly represented in the first person). In Derjagin’s view, the usage of individual authors becomes a factor only in the last two phases of analysis. This approach has three major shortcomings. First, it provides no explanation for residual forms other than idiosyncracy or individual whim. Second, it fails to account for the fact that
26
R V
there can be variation even within individual usage. Third, it sets up a model of the writer as a slave. In separating convention from deviation (“individuality”), it does not do justice to the fact that conventional strategies can be, and usually are, intentional — in other words, that they are an aspect of individual usage.
C 2 The text-kind A pragmaphilological overview
Medieval Russian trial transcripts are detailed protocols of trial hearings that include extensive representations of dialogue — the testimony of litigants and witnesses interspersed with the questions and statements of the judges — set in a third-person narrative. While these documents adhere to the general norm seen in narrative in that they recount sequentially ordered past actions, they differ from it in two important respects. First, there is usually little apparent causality in the chain of narrated actions, except in the representations of adjacency pairs (dialogic exchanges); the question-and-answer dyad is the central organizational principle of the text-kind. Second, the main focus of interest or information load is the ; much of the narrative framework consists of tags (attributions), while motions and nonverbal behavior are often treated as incidentals or (if one can draw any conclusions from silence) not reported at all. Moreover, even as “dialogue discourses” (Longacre 1983: 44), the transcripts are peculiar in that the does not so much tell a story as deliver raw information; the writers leave their intended interpreters the task of sifting through the data and determining their ultimate relevance. Most of the extant trial transcripts are records of real-estate lawsuits; a few are accounts of criminal proceedings involving property violations such as theft, arson, or abetting runaway serfs.1 The text-kind is thought to have arisen around the mid-thirteenth century (Leont’ev 1969: 45–46); the oldest surviving examples, which are discussed in this study, date from the first decades of the fifteenth century. While there are some earlier documents relating to lawsuits, they are not trial transcripts in a strict sense; this is the case, for example, with a verdict record of 1284, which contains no reports
28
R V
of testimony and is framed from the first person of the judge (Avanesov (ed.) 1963: 62). The oldest surviving document with protocols of dialogic testimony embedded in narrative is a “monastic-rule charter” of 1391 (ASÈI 3: 16–17, no. 5); this edict has a different structure than trial transcripts, and part of the narrative is presented from the judge’s first-person viewpoint. The corpus used in this study comprises 154 trial transcripts (some extant in more than one copy) from the territories that consolidated to form the Muscovite state.2 To my knowledge, this is an exhaustive corpus for the period from 1400–1505 — that is, the period prior to the official review of all charters in 1506, at the beginning of Vasilij III’s reign. While many of the documents are originals or contemporary copies, some are later copies with lost protographs; these evidently reflect the medieval legal custom of wordfor-word copying — verifiably so, in some cases where there are multiple copies from a single protograph (see also 3.4). I make special note of instances where there are traces of editing that are pertinent to reported speech. The trial transcripts in the corpus are attested in three main subtypes — trial records, judgment charters, and referred trial records.3 These subtypes are not distinguished consistently in medieval sources. In addition, copyists sometimes confused trial transcripts with similar text-kinds such as records of amicable settlements (e.g., ASÈI 1: 235, no. 326; 288, no. 397; 320, no. 431). Trial records (sudnye spiski — literally, “trial copies”) are transcripts of judicial hearings held on disputed properties, which follow the order of the real-world events. Their writers, who are usually anonymous but were apparently professional scribes, took notes of some kind during the hearings and then drew up polished versions in a conventional format prior to the judges’ deliberations. The purpose of trial records was evidently to enable the judges to review — indeed, reexperience (see 3.5) — the testimony when this was necessitated by the complexity of the evidence or delays in the trial. Some forms of delay were institutionalized. Real-estate trials were generally held on the disputed property, and the on-site judges did not always have the legal authority to hand down verdicts. In such cases they referred the trials to courts of higher instance presided over by grand princes or princesses, high-ranking boyars, or metropolitans. The higher judges then held hearings in which the trial records were read aloud, and the litigants were asked to verify that they were accurate (see Chapter 6).4 This procedure was a safeguard against errors and falsifications; if one of the litigants
T T-K
29
disputed the record, it was submitted to the “men of court” (sudnye muži) — official observers of the on-site hearings — for corroboration. After establishing that the records were accurate, the higher judges used them to review the testimony prior to issuing their verdicts (see Chapter 8). It must be stressed that the primary motivation of the verification process was not to ensure accuracy in the written document, although that was certainly a by-product. By custom, the higher judges and litigants listened to the transcripts, which were read aloud by an officer of the court. In addition, the trial judges sometimes made preliminary reports, again in oral form, at the beginning of the referral hearings. Medieval Rus’ was a “mixed-orality” culture, in which “the influence of writing remain[ed]… partial” (Fleischman 1990: 20). Trials, in particular, were an institution that was still largely oral in its basis.5 Thus the written aspect of trial records was incidental; the documents were, in their essence, a means of reanimating oral situations (see 3.5). Relatively few trial records have survived in separate form, probably because they reflect an unresolved stage in the judicial process. The majority of the extant transcripts include both a trial record and a narrative account of the verdict. Like the judicial referral, the verdict seems to have been primarily an oral institution. The judges apparently announced their decision aloud; court scribes then drew up a conventionalized verdict record in third-person narrative, which was appended to the trial record and the judicial-referral record, if any, to make up the completed trial transcript. There are several facts that provide evidence for this two-tiered process, including a verdict record written in the first person from Pskov (Northwest Rus’; GVNP no. 340); occasional cases of slipping into the first person in verdicts from Northeast Rus’ (see Chapter 8); and specific references to oral verdicts in seventeenth-century trial dossiers (see the texts in Jakovlev 1943/1970). Trial transcripts with verdict records fall into two subtypes, which seem to have been equivalent in their force, though administratively distinct.6 Judgment charters — Kleimola’s (1975) translation of pravye gramoty — were awarded to the victorious party as proof of ownership; they gave their bearers the legal leverage to implement the judgment and to protect themselves against future infringements. Referred trial records (dokladnye (sudnye) spiski) are documents issued by courts of higher instance, with accounts of the judicial-referral hearing and verdict appended to the trial record. These were usually given to the judges of the on-site trial, who executed the judgments of the higher courts and issued judgment charters to
30
R V
the victors; however, they were apparently sometimes given directly to the victorious litigants in lieu of a special judgment charter. The trial transcripts investigated are relatively uniform in structure. They typically consist of three or four sections, each with its own characteristic formulae. They begin with an incipit, which provides orientation by identifying the main participants and, in some cases, the location of the trial — often nothing more specific than “having stood on the [disputed] land” (see Kleimola 1972: 357–58). Many transcripts open with a brief formula indicating that some ruler has authorized the trial (cf. Kaštanov 1970: 37–38); this generally takes the form po X- slovu “in accordance with X’s command” (by X- word-). While the authorization formula allows the documents to be dated to a range of years (sometimes more precisely, with the help of other internal evidence), the date itself is never included in the setting information. The explanation for this absence — an omission, by modern expectations — is that the documents themselves “did not yet fully constitute the legal act which they set down”, as Kittay (1988: 211) notes of medieval charters from Western Europe. Prior to the verdict, the written transcripts were used only in the context of “personal attendance and verbal confirmation” (ibid.); this can be seen in the procedures at judicial-referral hearings, where the validity of the documents depended on the memory of participants — the trial judge, litigants, and men of court — who carried knowledge of the date and location within them. Only with the issuing of the verdict, when the transcripts were cut loose from this context of active, verbalized (or verbalizable) remembering, do their dates become potentially relevant as orientation for readers/interpreters with no involvement in the trial process. Even so, relatively few of the documents are internally dated, as compared with other chancery text-kinds; of the 152 transcripts investigated, only one (AFZX no. 248) has the date within the verdict, while 34 others have it accompanying the seals and/or signatures in the end protocols.7 Only one of the 13 transcripts from before 1462 is dated (ASÈI 3: no. 32). Significantly, this document comes from the chancery of Metropolitan Photios, a Greek, and is dated in the ecclesiastical (originally Byzantine) style, by indiction; in other words, it reflects, in part, legal institutions in which writing played a greater role than in ordinary medieval Russian law. Of the remaining transcripts with internal dates, only 4 (out of a possible 57) are from before the 1490s; 11 (out of a possible 53) come from the period 1491–1500, and 19 (out of a
T T-K
31
possible 29) are from 1501–5.8 These figures suggest that the date was gradually assuming greater importance; however, during the period investigated it remained optional, of uncertain relevance for any of the conceivable interpreters. In most incipits, the judge is introduced after the authorization formula (if any) with the set phrase exemplified in (1a), from a judgment charter of ca. 1448–52, which I chose as the main illustration in this chapter for the sake of its brevity: (1) a.
†Si sud sudil [Mixailo Fedorovicˇ(’). [this trial]- judge- [Mixajlo Fedorovicˇ]- † Mixailo Fedorovicˇ [Saburov] judged this trial (ASÈI 3: 58, no. 35).
The word order in this prepatterned sentence is, as expected, “presentational” — inverted, as compared with the neutral Subject-Verb order — because the subject is asserted; the topic is the direct object “this trial”, which is presupposed from the very existence of the record. The formula exemplified in (1a) can be elaborated with a participial (gerund) clause specifying the venue of the trial — stav” na/v” X- “having appeared at/in X” (stand- on/in X-). The next sentence is normally a set phrase, again in presentational word order, that introduces the litigants, with plaintiff first (1b): (1) b. Tjagalsja [arximandrit Feofan s Vašutoju, da… sue- [abbot Feofan]- with Vašuta- and Abbot Feofan sued Vašuta and [three other respondents are named] (ibid.).
This formula has several possible elaborations, which need not be discussed here. The incipit is followed by the trial record proper, the transcript of the lower-court hearing (see Chapters 3–5). Typically this opens with the plaintiff’s charge against the respondent (2a); the judge is then depicted asking the respondent for a rejoinder (2b): (2) a.
Tak rek [arximandrit Feofan: Žaloba thus speak- [abbot Feofan]- complaint- mi, g(ospodi)ne, na [tog(o) Vašutu, da… me- lord- against [that Vašuta]- and
32
R V
vstupajutsja, g(ospodi)ne, v [ozero monastyr’skoe trespass-.3. lord-3 in [lake of.monastery]- da v reˇcˇku v Vašku, lovjat, g(ospodi)ne, and in stream- in Vaška- fish-.3. lord- i… saki v reˇcˇkeˇ sacˇjat, a menja and nets-. in stream- drag-.3. and me- ne dokladaa, ni [moeg(o) poselskog(o). inform- nor [my manager]- Abbot Feofan spoke in this way: “I have a complaint, lord, against that Vašuta and [the other respondents]… [They] are trespassing, lord, on the monastery’s lake and on Vaška Stream; they are fishing, lord, and dragging the stream with… nets, without informing me or my steward” (ibid.) b. I [Mixailo Fedorovicˇ(’) vsprosil Vašuty i eg(o) and [Mixajlo Fedorovicˇ]- ask- Vašuta- and his tovarišcˇov: Otveˇcˇaite! I Vašuta tak rek: fellows-. respond-. and Vašuta- thus speak- Doseleˇ nam, g(ospodi)ne, za to arximandrit ne to.now us- lord- for that- abbot- stojal, ni eg(o) poselskyi, i my, g(ospodi)ne, stand- nor his manager- and we- lord- lovili na ozereˇ i na reˇcˇkeˇ; a nynicˇa, fish-. on lake- and on stream- and now g(ospodi)ne, ne velit nam arximandrit loviti, lord- bid-.3 us- abbot- fish- i my, g(ospodi)ne, ne lovim, a v ozero and we- lord- fish-.1. and in lake- eg(o) ni v reˇcˇku v Vašku ne vstupaemsja; his nor in stream- in Vaška- trespass-.1. a to, g(ospodi)ne, veˇdaem, cˇto [to ozero and that- lord- know-.1. that [that lake]- da reˇcˇka… c(e)rkovnoe… and stream- church-
T T-K
33
And Mixailo Fedorovicˇ asked Vašuta and his co-respondents, “Respond!” And Vašuta spoke in this way: “Until now, lord, the abbot has not confronted us about this, nor [has] his steward, and we, lord, have fished on the lake and on the stream; but now, lord, the abbot bids us not to fish, and we, lord, are not fishing, nor are [we] trespassing on his lake or on Vaška Stream; and we know, lord, that that lake and… stream are church [property]…” (ibid.)
Though typical in form, the trial record cited in these examples — a nolo contendere plea, in effect — is unusually brief. Ordinarily the respondents put forth counterarguments, the plaintiffs offer rebuttals, and both sides present witnesses. In some cases, litigants request adjournments to procure new evidence, or they challenge their opponents to take “the field” (a judicial duel) or kiss the cross (a solemn oath) to prove their ownership. The judges act as moderators, eliciting each act of testimony with questions and commands. Thus the minimal dyadic structure illustrated in (2b) can go on for many folia. The trial record can conclude in two different ways. In referred trial records and in trial records without a verdict, there is a formula involving a narrative report of a speech act (), in which the judge promises to refer the case to a court of higher instance (3): (3) I sud’ja rekl”sja/jal”sja doložiti X- and judge- promise- report.to- X- And the judge promised to report to X [about the case].
This is followed in referred trial records by the account of the judicialreferral hearing. In judgment charters, the testimony is followed directly by the verdict of the on-site judge. The verdict section itself (4) consists of prepatterned narrative reports of the judge’s speech acts, finding one of the litigants right [1] and the other wrong [2] and, in some cases, imposing specific financial penalties or conveying other instructions [3] (see Chapter 8). (4) I po tomu [Mixailo Fedorovicˇ(‘) [arximandrita Feofana and by that- [Mixajlo Fedorovicˇ]- [abbot Feofan]- opravil, a Vašutu… obinil i prisudil find.right- and Vašuta- find.wrong- and award-
34
R V
ozero i [reˇcˇku Vašku [s(vja)t(o)mu Mixailovu lake- and [stream Vaška]- [holy Michael- ˇ Cjudu [arximandritu Feofanu. miracle]- [abbot Feofan]- And, in accordance with that, Mixailo Fedorovicˇ [1] found in favor of Abbot Feofan and [2] found against Vašuta… and [3] awarded the lake and Vaška Stream to the Holy Miracle of Michael, to Abbot Feofan (ibid.)
The three past-tense verbs in (4) are all terms for legal speech acts — [1] opraviti ‘pass judgment that someone is innocent/right’; [2] obiniti ‘pass judgment that someone is guilty/wrong’; and [3] prisuditi ‘pass judgment awarding a property to someone’ (see SRJa, s.vv.). The verdict in (4) is likewise comparatively brief; in many other documents the ’s are elaborated by a statement of the judge’s ratio decidendi. Referred trial records can have two verdict sections, one relating the higher judge’s instructions to the trial judge (the “directed verdict”), and the other conveying the trial judge’s identical decision (the “executed verdict”). Many trial transcripts, especially those that survive in their originals, cap off the verdict with a list of the “men of court” (official witnesses, see above). Referred trial records can feature two such lists — one after the record of the lower-court hearing, the other after the verdict(s). The basic structure exemplified in (1)–(4), with various elaborations, can be found in virtually all northeastern Russian trial transcripts from the fifteenth and early sixteenth centuries. Each of the major divisions mentioned — the incipit, trial record, judicial-referral record, and verdict — forms a context with its own prototypical speech events and its own norms of reporting. I shall comment on variations in the diplomatic structure of these sections where appropriate in the following chapters.
C 3 Testimony
3.1
The standard reporting strategy
When the scribes of fifteenth- and early-sixteenth-century trial records needed to report the testimony of the litigants and witnesses, they usually chose , in accordance with what was undoubtedly a convention of the textkind. In , as in other languages, is characterized by elements oriented to the represented speech event rather than the context of reporting — in particular, first-person forms and spatio-temporal deictics referring to the represented speaker rather than the reporter (author or narrator), and secondperson forms and vocatives referring to the represented addressee. From a typological standpoint, the sine qua non of the category is its orientation to the perspective of the represented speaker.1 Thus, when referentially shifting elements are missing, the mode of reporting is inherently indeterminate (Worth 1985: 234–35); however, if it is methodologically convenient, one can find probable cause for treating even noncommittal reports as direct — for example, when they are surrounded by clear-cut . Another feature sometimes said to be characteristic of is the presence of emotive/expressive elements that depart from the style of the authorial/ narratorial discourse. In fact, this criterion is neither necessary nor sufficient. There can be reports in clear-cut that lack distinctive emotive elements because they are stylized to conform to the narrative; this is typically the case in trial transcripts, as in many other medieval writings. Moreover, emotive elements oriented to represented speakers can be found in nondirect modes of reporting — e.g., the “texture-analyzing” or “colored” of modern literature (Vološinov 1929/1986: 130–32; Page 1988: 33, 37); they can even penetrate authorial discourse for focalization effects (cf. Fludernik 1993: 423). I shall return to the question of why was the conventional mode of reporting in 3.3. First it is necessary to discuss the structure and function of
36
R V
the frame in which direct reports of testimony were most frequently set. In the transcripts investigated, one may identify a single standard reporting strategy, which accounts for the majority of the reports attributed to the litigants and witnesses. This strategy is attested in virtually all the cases of what may be termed prototypical testimony, the speech event that was perceptually central in the trial institution. Prototypical testimony — a concept that will recur throughout this study — is an uninterrupted statement made by one of the principals (the plaintiff or respondent), which provides information intended to establish the speaker’s rights to the disputed property; it is elicited by the judge(s) in a face-to-face exchange, presumably in some standard configuration determined by judicial custom. The prototypicality of this speech event stems from the fact that, under ordinary circumstances, it was better suited than any other kind of utterance to supply information that could be probative for the verdict — proofs of ownership offered by the parties who were interested and legally liable in the lawsuits. Further evidence that testimony in this strict sense was perceptually central may be found in certain abridged transcripts (e.g., ASÈI 2: 504–5, no. 465), where the only statements reported in are those that conform to the prototype, along with the questions that elicited them. (The latter reflect the prototyical speech event for the trial judges — requests for information about the disputed property, addressed to and cooperatively answered by individual participants, again in a face-to-face exchange.) The standard formula for introducing testimony is given in two predictable positional variants in (1a–b).2 This tag, which forms a link in the polysyndetic chain of sequenced narrative clauses in the authorial framework (see 3.2), always precedes the report, without any intervening complementizer. (1) a.
Initial report: tak(o) rek(l”)/rekli X- … thus spoke-./ X- X spoke in this way: []
b. Elsewhere: i X- tak(o) rek(l”)/rekli… and X- thus spoke-./ And X spoke in this way: []
This reporting strategy is otherwise characteristic of only a few of the textkinds that are known to have existed in legal-administrative writing — in
T
37
particular, default-judgment charters (bessudnye pravye gramoty), and records of amicable settlements (otvodnye, razvodnye, and raz”ezžie gramoty).3 All of these less common text-kinds recount legal or quasi-legal hearings; they are similar to trial records in form and function, although they exhibit more variation in reporting methods. Other varieties of legal-administrative writing show strong preferences for less-than-direct and especially grammatically integrated forms of . The formula given in (1b) is occasionally also found in narrative text-kinds of the ecclesiastical/literary (Slavonic) register, such as saints’ lives or chronicle entries; however, other tag clauses are more frequent there. The preposing of the tag clause to the report in (1a–b) must have been intentional, given that there were several alternatives — postposed and intercalated tags and nonclausal or subordinated adjuncts (see 1.2). The subject of the tags in (1a–b) is a noun phrase denoting the represented speaker(s). The predicate consists of the cataphoric adverb tak(o) ‘in this/ that/such a manner’ (spelled variously tak, tak”, tako) followed by the preterit of the recˇi ‘speak, say ()’, which agrees with the subject in number and gender.4 The word-order variation in (1a–b) does not imply a difference in reporting strategy. Word order in discourse was determined by pragmatic factors such as information structuring; there was a general tendency for non-thematized information to follow thematized (see Ickler 1978). This explains the “presentational” word order, with subject postposed to predicate, which is typical for the opening tag (1a). In subsequent tags (1b), the subject is normally not postposed, since its referent, the represented speaker, is usually thematized as the explicit or implicit addressee of the preceding report, as in (2): (2) I sud(’)i sprosili Saltyka da Visla… I and judges-. ask-. Saltyk- and Vislo- and Saltyk da Vislo tak rkli… I sud(’)i Saltyk- and Vislo- thus speak-. and judges-. sprosili [starca Kirila… I Kirilo tak rek… ask-. [elder Kirilo]- and Kirilo- thus speak- And the judges asked Saltyk and Vislo… And Saltyk and Vislo spoke in this way… And the judges asked the elder Kirilo… And Kirilo spoke in this way… (ASÈI 2: 315, no. 333).
38
R V
Non-responses, which are rare, can be signalled in unusual ways (see 4.6). As in narratives of dialogue in general, it was evidently expected that questions would be answered by their addressees, in accordance with the norms of cooperative dyadic conversation. The intended interpreters of trial transcripts had good reason to infer that this convention was in effect, given the relative power of the judge in the real-world situation (cf. Brown and Levinson 1987). The predictability of speakers in the question-and-answer structure explains, in part, sporadic cases in which the tag clause is omitted (see also 4.3). Likewise, because the plaintiff was always the first speaker, the attribution of the opening complaint could occasionally be omitted (3); interpreters with the expected background knowledge of the text-kind could infer it from the incipit, in which the plaintiff is conventionally the subject. (3) Tjagalsja Savka Mandakov… s Ysacˇkom s litigate- [Savko Mandakov]- with Isacˇko- with [Vaskovym s(y)n(o)m” Obnorin(a)… To, g(ospodi)ne, [otvotcˇiki [Vas’ko- son]- Obnorin- lord- [rentmen kirilovskie… otveli u menja, u Savki, of.Cyril]-. take.away-. at me- at Savko- pocˇinok k” [svoei pustoši… x Kocˇevinskoi… clearing- to [. empty.land]- to Kocˇevinskaja Savka Mandakov… sued Isacˇko, son of Vasko Obnorin… “The rentmen of Cyril [Monastery]… took away from me, Savka, a clearing [and annexed it] to their uncultivated property of Kocˇevinskaja…” (ASÈI 2: 192, no. 285).
However, null tag clauses, like other forms of ellipsis, are not typical in trial transcripts; the scribes aimed at maximal clarity and, in particular, tried to make the relationship between the authorial context and the explicit to preempt misinterpretations. This effort to be explicit, at least with information potentially relevant for the judgment, is reflected in the typical ways of referring to the represented speaker. As a rule, the reported speaker is identified briefly by a unique name, nickname, or title — e.g., Saltyk da Vislo and Kirilo in (2) or poselskoj (“estate manager”; Kaštanov 1970: 387, no. 27). When witnesses are presented as a group, in many cases only one or two of them — possibly
T
39
the spokesmen in the real-world trial — are named individually; the others, mentioned by name in the litigants’ testimony, are referred to by the noun phrase i ego/ix” tovarišcˇi “and his companions” (and his/their fellows.) or variant (e.g., ASÈI 2: 409, no. 401). Such generic labels are hardly ever used to present multiple litigants, who were more individuated in the scribe’s perception because of their prominence in the trial. The preferred referential strategy in the tag clauses is thus to use noun phrases with unique referents, even though the referents are usually already thematized in the preceding narrative. Pronominal anaphora scarcely ever occurs in the tag clauses, even though it is common within the ; indeed, it is rarely used with the subjects of other narrative clauses. This avoidance of anaphora is a convention specific to trial transcripts; it is not typical of other kinds of narrative or legal-administrative writing. One of the only cases in the corpus (4) appears in a passage that exhibits several other unusual features as well (see also 4.2): (4) I Mixal’ tako rek: Jaz, gospodine, zemli ne and Mixal’- thus speak- I- lord- land- znaju, mne sja zemlja dostala novo, a know-.1 me- land- obtain- newly and znaet, gospodine, zemlju [Fedor Kokoš, poselskoj know-.3 lord- land- [Fedor Kokoš steward]- [velikogo knjazja. I sud’ja vsprosil Kokoša, i [grand prince]- and judge- ask- Kokoš- and Kokoš molvit: Jaz, gospodine, mež ne Kokoš- say-.3 I- lord- boundaries-. znaju. I sud’ja tako rekl: Daj že know-.1 and judge- thus speak- give- ty mne starožilcev, xto znaet [zemle you-. me- old.residents- know-.3 [land Cˇekmakovskoj s [Putilovskoju zemleju meži. Cˇekmakovskaja]- with [Putilovskaja land]- borders-. I on postavil [Ontipu desjatskogo, da Frola… and he- put.up- [Ontipa tenthman]- and Frol- And Mixal’ spoke in this way: “I don’t know the land, lord, I just obtained it, but Fedor Kokoš, lord, the grand prince’s steward, knows
40
R V
the land.” And the judge asked Kokoš, and Kokoš says, “I don’t know the boundaries, lord.” And the judge spoke in this way: “Then give me witnesses who know the boundaries of the Cˇekmakovskaja property and the Putilovskaja property.” And he put up Ontip the Tenthman and Frol… (ASÈI 1: 321, no. 431).
This passage provides a clear illustration of why the scribes avoided pronominal subjects. The referent of the pronoun in the underlined clause is ambiguous; it could be interpreted either as Mixal’, the respondent, or as Fedor Kokoš, the last speaker before the judge’s command, whom Mixal’ names as co-respondent. As the nearest possible referent other than the judge (who is excluded because of his role in the trial), Kokoš is the most likely referent of the pronoun, but Mixal’ is actually more probable in view of the participation framework of the speech event. The interpreter would have to fall back on background knowledge for the information that it was the litigants who were ordinarily responsible for producing witnesses. Even this does not exclude Kokoš, since Mixal’ has tried to transfer responsibility to him.5 As the estate manager, Kokoš would certainly be expected to know the lands under his care or, at least, to be able to name knowledgeable locals as witnesses. Such potential difficulties in interpretation were undesirable in a text-kind whose purpose was, in large part, to provide enough information to ensure a fair judgment. As a rule, the scribes strived to made their texts as explicit and intelligible as possible and accordingly avoided anaphora and other potentially ambiguous devices. (However, null reference is often employed in one particular unambiguous context; see 4.3.) There are three chief ways in which the standard tag formula can be elaborated. First, the predicate can include an adjunct identifying the attributed speaker as the spokesman for a group: (5) a.
tak rek Stepanko i v” [vseˇx” kr(e)st’jan thus speak- Stepanko- in [all peasants]-. meˇsto… place- Stepanko spoke in this way on behalf of all the peasants… (ASÈI 1: 478, no. 589).
Second, the entire tag can be preceded (5b) or, less often, followed (5c) by a clause consisting of a past gerund (deverbal adverb) plus one or more
T
41
’s. This kind of elaboration specifies the location of the trial or given segment of the trial: (5) b. Stav na mežeˇ, tak rek Davydko… stand- on boundary- thus speak- Davydko- Having stood on the boundary, Davydko spoke in this way… (ASÈI 1: 246, no. 340). c.
I Okiš, da i Larivon… tak rkli, and Okiš- and Larivon-… thus speak-. stav na sudeˇ… stand- at court- And Okiš and Larion…, having stood [or appeared] in the place of judgment, spoke in this way… (ibid.: 247).
As orientations, such gerund clauses often occur either with the tags for the opening turns, as in (5b), or at episode boundaries, as in (5c), where witnesses mentioned previously in — given but not fully activated in the discourse — are introduced for the first time in the narrative. In a few cases, new participant configurations are conveyed by other devices — for example, the prepositional phrase pered” sud’eju “before the judge” (before judge-),6 or the adverbial clause i vystupjasja “and, having come forward” (and step.forth-).7 Third, the standard tag could be elaborated by a noun phrase identifying the represented addressee. There is evidence from other text-kinds that the recˇi could take an indirect object (or, in ecclesiastical writing, a prepositional phrase) referring to the addressee; however, this is rare in trial transcripts. Omission of this information could have little effect on the interpretability of the text; the judge was the only ratified addressee in the represented speech event, and trial procedure (if the internal evidence of the written transcripts is any guide) required the participants to speak only when so instructed by the judge. In most transcripts, economy was thus more important than explicitness in this matter; there is only one transcript in which a reference to the represented addressee (sud’i judge-) appears repeatedly in the tag clause (14 out of 31 tokens). Significantly, in 12 of these cases, the participants seem to have undergone a change in footing. They are depicted speaking without prompting or in other contexts that
42
R V
depart from the conventional question-and-answer structure — unsolicited offers to produce additional witnesses or show the judge the boundary; running commentaries during boundary-setting procedures (see also 4.1); and challenges to the preceding testimony.8 These examples bear out Longacre’s suggestion (1994: 132) that “variations… in regard to mention/non-mention of Speaker and/or Addressee — beyond the basic need of participant identification and reference — are indexical of the intensity of participant interaction in reported dialogue”; in particular, as is evidently the case in the given instance, they can signal “dominance or attempted dominance of one participant over the other.” The two examples that do occur in predictable dyadic contexts (responses to questions, ASÈI 2: 434, 440, no. 411) suggest that the scribe had a general proclivity for the device; nevertheless, he showed this proclivity most consistently where it was most motivated, in signalling unexpected changes in footing (cf. Goffman 1981) as a precaution against misinterpretation.
3.2
Strategic choices in the standard tag
Looking at the standard reporting strategy, we find a series of choices, none of which were trivial.9 First, the scribes chose not to aver the testimony (cf. Sinclair 1988) or summarize it but to present it as sequences of reports. Second, they chose to tag the reports in narrative frameworks, and they adopted a specific style of narration. Third, they chose as the mode of reporting. Fourth, they chose to include specific kinds of information in the tag clauses. Fifth, they chose to tag with independent clauses that were preposed to independent reports. If we ask what informed or motivated the particular choices, we find that, even if they were conventional (cf. Lewis 1969), the choices were not only intentional but also purposive. They were all aimed at specific needs of the judges, the intended interpreters; hence they all made specific contributions to coordinating the sense-making process, in accordance with the purposes of the text-kind. The institutional task of the scribes was to communicate information about the trial in a way that would facilitate the judges’ decisions. This role did not — indeed, could not — discourage them from exercising their speech wills in the matter of what or how to report; however, it did impose certain standards of cooperative behavior, which
T
43
would be flouted if they were to analyze or evaluate the information themselves. This made averrals, summaries, and out-of-sequence presentations infelicitous. The scribe’s task favored narration presented from the vantage point of a witness rather than an interested or otherwise active participant, and in a covert style (see Rimmon-Kenan 1983: 97–98) that permitted orientation information and attribution but drew the line at judgments and commentary. This purposive suppression of evaluation was part of the motivation for the use of (see 3.5). It also explains why the scribes generally refrained from explicitly evaluative devices such as lexically specific ’s, manner adverbs, and nouns or adjectives that went beyond identifying the speakers; they preferred instead a minimal tag with a bleached that asserted nothing more than that an act of speaking occurred (cf. Hermon 1979: 27 on English say). The other elements of choice in the standard reporting strategy also served to further the work of the interpreter. For example, the preposing of an independent (main-clause) tag to an independent (unbound, clausal) mode of reporting gave both the attribution and the report contextual prominence. Green (1980: 583, 592), writing on English, asserts the opposite: since old information tends to precede new, preposing puts the attribution in “presupposed” (backgrounded) position. In fact, this word-order principle, which also existed in (see 3.1), operates clause-internally; it cannot be imposed on the relation between adjacent clauses, e.g, between independent tag and report clauses. The fact that “eclipses the sequential verb of saying that introduces it” (Chvany 1990/1996: 289) does not imply that preposed tags cannot be salient as compared to other narrative elements.10 The use of any of the alternative strategies would have had a contrary effect. Tagging with an adjunct — a dependent element without any foregrounded element — would have reduced the prominence of the attribution; the use of a complementizer would have reduced the prominence of the report (see 4.4). If the tag were postposed to or intercalated within the report, it would have been placed in a position where it was not only less salient but also non-asserted, i.e., presupposed from the report (Green 1980: 591–92; Kurihara 1986: 78, 81, 84). Where preposing and postposing are both available, as in (see 1.2), the former is expected at the beginning of dialogues and episodes, since it preemptively identifies the participants and provides other information needed for orientation purposes (cf. Hermon 1979: 7–9, 11, 14; Green
44
R V
1980: 592). However, the majority of preposed tags in trial transcripts are not episode-initial. Hence the relation established by preposing would seem to give the tag undue salience, especially in a “dialogue discourse” (Longacre 1983: 44), in which the most important information is conveyed in . After all, the attribution was generally predictable, since answers were expected to follow questions in the dyadic structure (see 3.1). In fact, preposing the tag was well motivated for several reasons. First, preposed tags served as a form of punctuation, demarcating adjacent reports from one another. Hypothetically this could have been accomplished by a string of postposed tags, but note that the trial record begins with a preposed one; in any case, only preposed tags could serve to focus attention preemptively on the reports. Back-to-back reports, like “colliding speeches” (Brandsma 1996: 257) in general, would have posed processing difficulties not only for the readers, given the graphic conventions of the transcripts, but also for the hearers — a category to which at least some of the intended interpreters belonged. Second, the fact of the utterance, which was asserted by means of the tag, was indeed relevant information for the purpose of the text-kind, not a mere afterthought; the content did not have legal meaning unless it was clear whose position it established. In the verbal duel of the trial, the individual voices were not collaborating to tell a unified story. Third, pre-indexing the reports ensured that acts of testimony would not be interpreted without prior information about their source, which was crucial for appreciating their legal significance. This partially phatic function — much like the identification of the caller at the onset of a telephone conservation — likewise served the needs of hearers, who were unable to scan the transcripts as they were read aloud in the actual trials. The repetition of the same prepatterned tag, with variations only in the identity of the speaker, promoted interpretability in and of itself in that it helped to establish a narrative framework — a plotline that provided orientation clues to the dialogue. Syntactically, the frame is a chain of tag clauses begun by the adverb tak(o) in the first turn at speaking (see (1a), Section 3.1), occasionally interspersed with other kinds of narrative clauses. Each link in the chain is conjoined to its predecessor by the sequencing conjunction i ‘and’ (see (1b), (2), and (4), Section 3.1). It should be emphasized that the combination of and narrative based on the sequentiality of real-world events was intentional rather than automatic,
T
45
since the scribes had other means of representing trials at their disposal. Other principles of organization may be seen, e.g., in verdict records, especially those appended without trial records to other kinds of documents, as in (6), taken from an immunity charter of 1454.11 (6) potomu cˇto [teˇ pustoši izstariny za šest’desjat leˇt because [those fields]-. of.old for sixty- years-. potjagli x Karinskomu selu… a i [sam tot belong-. to [Karinskoe village]- and [self that Vasko skazal, cˇto [teˇ pustoši pašut Vas’ko]- say- that [those fields]-. plow-.3. xrest’jane [Karinskog(o) sela, kak sja dostalo peasants-. [Karinskoe village]- since obtain- [to Karinskoe za manastyr’; a Vasko iskal [that Karinskoe] for monastery- and Vas’ko- seek- ix za [kn(ja)žie zemli za ondreˇevskie. them- for [prince- lands]-. for Andrej-.. … because those fields of old, over sixty years, belonged to the village of Karinskoe… and even Vas’ko himself [the losing litigant] said that the peasants of the village Karinskoe have been farming those fields since that Karinskoe became the monastery’s property; [1] but Vas’ko was seeking them as Prince Andrej’s lands (ASÈI 3: 83, no. 54a).
Here there is minimal attribution and no indication of the real-world ordering of the represented speech events; the report marked [1], in particular, conveys a claim that was probably voiced at the beginning of the trial. Clearly this nonsequential, summarizing form of organization is highly evaluative and so would not have served the institutional function of trial transcripts — providing judges with unprejudiced raw material for their deliberations. Another component in the standard framework was the rek(li) (from recˇi), which, as a perfective preterit verb, served to create a plotline of foregrounded events, in accordance with the narrative norm. The narrative norm in ecclesiastical writing (Church Slavonic) of the same period was the aorist (Chvany 1985a). While this tense was obsolete in northeastern chancery writing, it occurs instead of the preterit in the fifteenth-century
46
R V
transcript from Novgorodian Dvina and in a diplomatically similar settlement charter (rjadnaja gramota) from the same region (GVNP 149, no. 92; 188, no. 132). By contrast, the preterit is used instead of the aorist in a Novgorodian transcript dating from after the 1478 Muscovite annexation (Koreckij 1969: 288, no. 2) — evidently an accommodation to northeastern chancery conventions. The authorial narrative framework, with its high degree of predictability, was an important aid to the task of interpretation. In the text of the transcripts, one finds a continual disruption of cohesive relations, in part because of the constant “gear-shifting” (Page 1988) from to narrative and in part because of the deictic shifts and diverse discourse types (narrative, descriptive, procedural, even reportative) that can appear within the direct reports. This lack of connexity increased the role of verbal devices — including, in particular, recurrent frames — in promoting coherence, especially given the minimal use of graphic clues to discourse structure. While in modern written and especially printed texts tends to be bracketed out by graphic signals such as indentations, quotation marks, or dashes, such devices were little used in writing, which in general had a limited inventory of punctuation marks.12 Colons were used sporadically to signal reports from the sixteenth century, and double commas from the end of the seventeenth; the modern system of punctuation with only became widespread in the eighteenth century, evidently in imitation of Western conventions (Koduxov 1955: 116–17). Thus was not graphically delimited in any special way in trial transcripts, which were written in scriptio continua without paragraph indicators (see the photographs in Cˇerepnin 1956: 289, Kotkov and Filippova (eds.) 1978: 178). Periods were used widely, though inconsistently, to delimit phrases and larger syntagmas — often “measures”, i.e., intonational units reflecting information blocks (see Gvozdanovic´ 1995 on other writings); however, no discrimination was made between and the narrative. There is some evidence that capital letters were sometimes used at the beginning of the tag clauses (see 5.1); unfortunately, modern editions do not permit one to judge how widespread this practice was.13 Even limited graphic signals would be of some benefit to readers; however, one must always keep the primarily oral nature of the judicial process in mind. In medieval Rus’, as in other cultures prior to the spread of printing, reading was primarily “a listening process, simply set in motion by sight” (Ong 1982: 121). The internal evidence of the transcripts (which there
T
47
is no reason to doubt) indicates that it was customary for them to be read aloud before judges of higher instance; in other words, the primary institutional role of at least one part of the intended interpreters was to be hearers rather than readers. I shall discuss the significance of this fact in 3.5; for now, it remains to be considered what other verbal signals the scribes used to assist the interpretation process — that is, to help the judges impose coherence on the texts. As mentioned above, the non-narrative elements within had the potential to disrupt the coherence (connectedness) of the discourse. However, when regularly repeated, they could serve as predictable “quotation markers” to promote cohesion and facilitate interpretation. Noteworthy among the features that demarcated from the authorial narrative is the vocative g(o)s(podi)ne ‘lord’, conventionally employed within the testimony in reference to the judge(s) (cf. Your Honor in American courtroom usage).14 Though singular, this stereotyped form of address is used in the transcripts even when there is more than one represented addressee. (The vocative of the collective gospoda is attested with plural reference in other text-kinds.) Presumably the function of this appellative in the trial context was to show deference to the judges and thus dispose them to accept the testimony as valid.15 This deference was based on relative power in the juridical context rather than social distance; thus two princes are depicted addressing a boyar judge, their social inferior, in this manner in a transcript of 1496 (ASÈI 1: 501, no. 605). The use of g(o)s(podi)ne in writing was typical not only of the testimony in trial transcripts but also of non-narrative (or not primarily narrative) texts ranging over the entire period, including treaties, petitions, and other kinds of documents written from a first-person perspective to specific addressees.16 While the vocative had a straightforward honorific function in texts of this kind, its motivation is less clear-cut in the of trial transcripts — documents framed as third-person narratives with covert addressees. The intended interpreters were not necessarily the represented addressees, and if they were they were on a new footing; in any case, the speech acts marked with the vocative were no longer actual by the time the trial records were read. Moreover, it is inconceivable that the scribes, while taking notes during the hearings, would have bothered with words like g(o)s(podi)ne that had a low informational load; presumably they added the vocatives in the fair copies in accordance with convention and/or their own reconstructions of the
48
R V
events. It is doubtful that they included the vocative out of a desire for verisimilitude, given that they evidently screened out other expressions of interpersonal meaning and expressive elements that were presumably uttered during the trials. (For example, the corpus does not contain a single insult, despite the confrontational nature of the proceedings.) Clearly the vocative had some function other than appellation. It may be observed that g(o)s(podi)ne tends to occur in almost every clause of the testimony in second position — after the the enclitics in Wackernagel’s position or after the first major constituent. This is a slot often occupied by macrostructural connectors and other cohesive devices, and indeed the vocative may be interpreted as a marker — a unique and, in the technical sense, typical feature of that could be exploited to promote coherence by reinforcing other indicators and by disambiguating passages in which such indicators were absent. As Fludernik (1993) cogently argues in her theory of schematic language representation, typicality markers of this kind play an important role in the construction of all modes of reporting. This is true even of , which is not intrinsically mimetic, as often claimed, but often features typification strategies, i.e., relies on adequation to type rather than to token.17 Because the texts were written from a covert thirdperson viewpoint, the appearance of a typical non-narrative feature such as g(o)s(podi)ne was sufficient to identify a passage not only as but also as testimony — indeed, as nonreported portions of testimony; the vocative never occurs in the authorial narrative and tends not to appear in additional layers of within the testimony.18 Table 1 summarizes the distribution of g(o)s(podi)ne by clause type in a sample of 40 northeastern transcripts (nearly one quarter of the corpus). These data show that the scribes tended to insert it in clauses that could be perceived as potentially separate segments, as defined by a change of theme, topic, or subject. The likelihood of its being repeated from one clause to another varies inversely with the cohesiveness of the two clauses. These facts bear out the interpretation that g(o)s(podi)ne served as a signal of continuity with the preceding and, by its very repetition, made the into a cohesive chain.
49
T Table 1. Distribution of the vocative g(o)s(podi)ne by clause type
All clauses Initial clauses Noninitial clauses Second conjunct, same-subject Second conjunct, different-subject19 Preposed subordinate clauses Main clause after preposed subordinate clause Postposed relative clauses Postposed temporal, clausal, adverbial clauses Complement clauses after cognition verbs
3.3
# Vocatives
Total
%
1224 0434 0790 0157 0532 0031 0020 0004 0002 0012
1787 0448 1339 0240 0716 0033 0031 0047 0014 0086
68.5 96.9 59.0 65.4 74.3 93.9 64.5 08.5 14.3 14.0
Direct speech: Verbatimness or relevance?
In the following sections I shall examine the factors that made the preferred strategy for reporting dialogic testimony in the narrative framework of trial transcripts. This discussion will require a foray into debates about the nature of the direct mode. Here my goal will be to rule out two prevalent and related hypotheses — first, that is verbatim reporting; second, that reflects the reporter’s commitment to faithful transmission of the wording and structure of the anterior utterance. At first glance, the fact that was the principal way of representing testimony in trial transcripts seems not to require any explanation. After all, is usually considered the most “simple”, “basic” mode of reporting not only in but in all languages.20 This assumption is, I suspect, founded sometimes on the ontogenetic priority and typological universality of but more often on the notion that is in essence the reproduction of “original utterances.” It is not, in fact, entirely clear that the assumption is justified or, indeed, what it would mean from a pragmatic perspective, in which speech reporting is treated as intentional activity. Whenever a choice is available, there are text-kinds and discourse contexts in which modes other than are not only statistically predominant but also perceptually normal (“basic”) for pragmatically competent speakers (cf. Aaron 1992 on Obolo narrative). For instance, in the earliest Slavic writings (roughly four to six hundred years older than the trial transcripts examined here), there are several well-defined situations in which is far less typical than , though reliance on context-
50
R V
blind overall frequency has misled some scholars into thinking that was nonexistent or incipient (for discussion, see D. Collins 1996; cf. Ebert 1986 on in languages of Nepal). As examples of this kind show, the coexistence of differently distributed reporting strategies imposes a methodological imperative: one must seek out the factors that underlie the context- and textkind-based preferences for each strategy, including the one with overall statistical predominance, before generalizing about the “basic” mode of reporting in the language as a whole. I would argue that the notion of default modes is at best an oversimplification and often an illusion; putatively “basic” modes have pragmatic rationales. In any case, the appearance of in trial records was by no means inevitable, even in representations of dialogue; for example, Muscovite interrogation records (rassprosnye recˇi), which consist mostly of reports, make extensive use of and relatively little of .21 The scribes had other strategies at their disposal and used them extensively in other sections of the documents (see Chapters 6–8), occasionally even in the trial record (see Chapter 4); thus their use of reflects a continual exercise of their speech will. If default is a pragmatically inadequate explanation, what did motivate the choice of as the standard mode for reporting testimony? It cannot be dismissed as “mere” convention, given that conventions have rational bases and serve as solutions to coordination problems (Lewis 1969). One answer that immediately springs to mind is that the scribes of trial transcripts chose as the only feasible means of providing an exact (literal, word-for-word, verbatim) rendition of the trial dialogue for the judges’ information, in the manner of modern court reporting. In such a case, the documents, at least in intention, would be transcripts in a strict sense — reproductions of the actual proceedings. This interpretation would accord with the traditional and still widespread conception of the nature of , which Coulmas (1985a: 41) characterizes as “generally received opinion” — the verbatimness or “reproductionist” (Sternberg 1982b) hypothesis. According to this model, is a mode that features “word for word reproduction (indicating the quoted speaker’s style, in the sense of his lexical choices)” (Banfield 1973: 9). This explanation of is institutionalized in traditional terms like direct speech and oratio recta, which presuppose the possibility of “straight” quotations, without authorial tampering.22 Indeed, the term direct speech (or equivalent) is occasionally used to denote not only a mode of reporting but also what was actually said by the anterior speaker.23
T
51
For all its intuitive appeal, the verbatimness model of is not tenable, as has been demonstrated trenchantly in a number of recent studies. It is, of course, possible to reproduce anterior utterances by means of (Sternberg 1982b: 111), as can be typically seen in written citations from written sources; indeed, as Haberland (1986: 225) points out, “if one wants to commit oneself to the form of the model as well [as the content], then one has to avail oneself of direct speech.” However, verbatim reporting is only one subtype of ; given the existence of others (see below), it would be a fallacy of composition to include the notion of verbatimness in the definition of the category: “From the premise that … can reproduce the original speaker’s words, it neither follows that it must perforce do so nor that it ought to do so nor, of course, that it actually does so…” (Sternberg 1982a: 68). In fact, it has been shown empirically that is often used in a non-reproductive manner — as “constructed dialogue” (Tannen 1989; see below). In spoken language, verbatim reporting is ordinarily precluded by memory constraints, except when the anterior utterance is very brief, formulaic, or specially memorized (a situation that is far from prototypical). Memory limitations come into play almost immediately, particularly when there is intervening talk; according to Chafe (1994: 215), “distal speech events” are stored in long-term memory in “verbally uncommitted” (nonverbatim) form.24 Memory constraints do not, of course, exclude the possibility that reporters, in using , are operating under an illusion or pretense of verbatimness. Accordingly, some scholars have offered definitions of the category that treat verbatimness as a purport, a social rather than empirical issue. Thus Coulmas (1986b: 2) states that “evokes the original speech situation and conveys, or claims to convey, the exact words of the original speaker”; however, he qualifies this by observing that different cultures can have their own criteria for “what counts as the same” (ibid.: 10–13). Another influential view locates verbatimness in norms of conversational cooperation — in a “faithfulness principle” akin to Grice’s Maxim of Quality (Short 1988): by selecting , the reporter has obligated himself, to the extent that it is possible, “to report faithfully (a) what was stated and (b) the exact form of words which were used to utter that statement” (Leech and Short 1981: 320).25 In this reporter-commitment model, the fidelity of depends on felicity conditions — e.g., sincerity and the effort to be accurate. The verbatim that would result from fulfillment of these conditions would be
52
R V
the conceptually basic or prototypical form of the category, a fulfillment of “canonical expectations” (Short 1988: 64; see also 69–70). However, there is considerable evidence that speakers frequently employ without any particular concern for providing, or even appearing to provide, a “faithful reproduction” or exact rendition of some anterior utterance. This is not to say that they are intentionally avoiding verbatimness; rather, the issue of verbatimness forms no part of their intention whatsoever, and it is not required by their audience. To begin with, reporters can make appropriateness repairs in ; if they were really “committed to verbatim reproduction”, they could only make corrections of misquotations (Clark and Gerrig 1990: 797–98). Moreover, they routinely use in manifestly nonverbatim ways (“categories of anti-mimeticism”, Fludernik 1993: 414). For example, can appear in reports intended as condensations of lengthier utterances — Short’s speech summary (“a string which reports in an abbreviated form some longer piece of discourse”, 1988: 74). It is frequently used for reports intended to be perceived as tokens of repeated or habitual statements or as amalgams of discrete statements by several speakers that present a “vox communis image” (Fludernik 1993: 411; “choral speech”, see 4.3). It can occur where verbatimness is excluded a priori, as in translations (Clark and Gerrig 1990: 798–99; Coulmas 1986b: 24, Note 3) and reports with hedged tags. It is frequently employed when the very ideas of verbatimness and commitment to fidelity are irrelevant, as when there is no anterior speech event — for fictitious statements; for non-statements (reports with negated tags) and hypothetical ’s (reports with irrealis tags); for unobservable statements and thoughts ascribed to other people; and for purported communication by animals, body parts, inanimate objects, and signs and other text conveyors (“verbal processes with non-human semiotic source”, Baynham 1996: 73).26 It should be noted that such nonverbatim uses of are attested not only in modern languages but also in medieval Russian writings. There is no evidence that any of these cases should count as abuses or uncooperative usages in the languages with which I am familiar. Unlike, say, prevarications or memory lapses, it would be problematic to argue that they violate felicity conditions. Particularly challenging for the concept of speaker commitment are coconstructed reports, second-hand reports presented in and the category of quotations that Caldas-Coulthard (1994: 299–300) terms factional.27 In the latter instance, the reports are purportedly factual but in reality fabricated by the reporters, often on the probabilistic grounds of what the reportees might
T
53
be expected to say. The issue of actual wording is irrelevant; the reports are constructed to convey what the reporter thinks the represented speaker’s position is or ought to be. (Cf. also Morawski 1970: 700 on fictitious quotes attributed to real authors.) There is no contextual clue that such “factional” reports are fictive, nor is any implicature of nonverbatimness possible. This category of direct reports thus poses problems for the view (expressed, e.g., by Roncador 1988) that the most important aspect of verbatimness is not “factual adequacy” but rather the extent to which the reporters assume responsibility for the wording by signalling their interference. In general, approaches to that appeal to verbatimness, whether as an empirical fact, as a purport, or as a conversational maxim, do not provide an adequate picture of the nature and extent of the category. They ignore its selective character and privilege messages (sentences or words), when in fact speakers can employ to demonstrate any aspect of language use, including code and manner of delivery (Clark and Gerrig 1990: 775–78). Moreover, the message or proposition can be treated as secondary to the other aspects being demonstrated (“supportive, annotative, or incidential”; ibid.: 780). Hence reports with signals can include departicularizers (vague referents) like this and that, which can substitute for part or all of the message (ibid.).28 Examples of this kind are troubling not for only verbatimness-based treatments of but also for the more moderate alternative view that “the use of only implies a commitment to the content of the model speech act, viz. to its intention (illocutionary force) and its propositional content”, not to its form (Haberland 1986: 225). Clearly “what counts as the same” is not always wording or structure; often it is the gist or the basic ideological position. It might be argued that, no matter how common they are in everyday language, the “anti-mimetic” types of cited above stray too far from any conceivable prototype to count as cogent arguments against the fundamental verbatimness of the category. However, countering this objection is the “truly hair-raising sloppiness and disregard for reliable representation” (Fludernik 1993: 17) of actual anterior utterances that is typical in speech and nontechnical transcripts of spoken language, even when a high degree of fidelity is possible. Here again verbatimness is not flouted but simply irrelevant; there is no perception that the report is “faithless”, and precisianist challenges to wording will be viewed as uncooperative unless they bear on the interpretation in some way. As Fludernik shows, the effect of identity or similarity in is achieved not just by imitation but also (and more often) by typification
54
R V
strategies leading to “adequation to type” (ibid.: 18–20, 418–19, 425–26, 435). (See the discussion of the vocative g(o)s(podi)ne in 3.2.) For example, Slembrouck (1992) points out striking changes of syntax, wording, and style in reports of British parliamentary debates in Hansard, as compared with tape recordings of the anterior utterances. As in trial transcripts sent up to a higher court, the reported MPs have the opportunity to review the reports and challenge anything that they regard as inaccurate (ibid.: 104); evidently they are usually willing to accept the Hansard reports despite their sometimes drastic deviations from the form of what was actually said. Baynham (1996) cites cases from classroom discourse in which teachers use virtually to put words in their students’ mouths; that is, they radically reformulate the anterior utterances to represent their views of what the reportees are (or ought to be) “really saying” — the proper reasoning in the appropriate institutional style. Short (1988: 67), though an advocate of the reporter-commitment model, observes that British newswriters “appear to ignore the faithfulness maxims to some extent”, since reports in headlines can differ in a far from trivial manner from reports of the same anterior utterances in the articles themselves. In other words, newspaper writers and editors are not perturbed by discrepancies in syntax, wording, and style even when the readers can collate different versions (and the reportees can potentially issue denials). This latitude suggests that “what counts as the same” is not only culture- but also genre-specific. Even in text-kinds where punctilious factography might be expected, the (putative) commitment to fidelity in is a matter of degree; it can be overridden by other needs (ibid.: 65–70), and the reading conventions of specific text-kinds can license this as a form of cooperative behavior. The weight of evidence favors a different interpretation — that verbatimness (or commitment to form and content) is not a fundamental principle but itself a reading convention for in certain written genres (e.g., quotations in scholarly articles and monographs). It requires a special effort that is only preferred or relevant (in the sense of Sperber and Wilson 1986) in particular text-kinds and discourse contexts. For reporters, it is a pattern of cooperative behavior adopted, in accordance with the behavioral principle of least effort (Zipf 1949), when deemed necessary for specific purposes. Like conversational cooperation in general, it is regulated by institutional, generic, interpersonal, and other contextual considerations (cf. Fludernik 1993: 20, 22; Lakoff 1995: 196).
T
55
Verbatimness, as Fludernik notes (1993: 2), is not the natural, unmarked form but “the most highly artificial, marked and formulaic type of represented discourse.” Its marked status is reflected in the very existence of special strengthening hedges relating to the quality maxim like verbatim, slovo v” slovo (“word for word”).29 There would be no need for such terms if garden-variety truly involved a commitment to verbatimness. Note that the converse use of weakening hedges with — e.g., or words to that effect (Fludernik 1993: 412; see also 3.4) — does not provide support of the reporter-commitment model; there can be contextual motivations not only for emphasizing verbatimness but also for ruling it out as a possible interpretation.30 These include the inability to report verbatim in situations where it is favored; the perception that challenges are likely; or self-subversion, as in the following fictional example, in which the reporter, with tongue in cheek, is cultivating an eccentric identity: “‘I was mucking about very peacefully [in the church], admiring things, when I suddenly thought I wasn’t being very respectful. I thought “God Tonker, you are a stinker, only coming up here when you want something”, so I got down very decently and said something — not aloud, of course, but something simple — like “O God, make me a good Tonker and let Mr. Guggenheim pass that scheme.” Something perfectly normal and ordinary’” (Allingham 1955/1990: 183). In short, fidelity to the wording of the anterior utterance (if any) should not be assumed as the norm — or even as an ethical ideal — for across the board, nor should any appeal to identity of form be included in the definition of the category. There is, indeed, no empirical evidence that there is a “faithfulness maxim” of universal validity, or even that such a principle operates as a general conversational maxim in modern English-speaking cultures. It can certainly not be taken for granted in premodern cultures. The conceptualization of verbatimness as involving syntax, wording, and style seems to be a by-product of writing and especially printing; preliterate and mixed-orality cultures have different understanding of what constitutes sameness.31 In other words, the idea of fidelity to form in definitions of is nothing more than contraband from genre-specific conventions for quoting written sources (Clark and Gerrig 1990: 800; Mayes 1990: 330); it must be avoided if one is to do justice to orality, as in an investigation of medieval texts. A revealing illustration of the different attitudes towards reproduction in largely oral vs. largely chirographic societies can be found in the semantic evolution of naizust ‘by heart’, which now refers to verbatim reproduction
56
R V
of written texts. This adverb derives from the prepositional phrase iz ust” ‘from the mouth’, used of documents composed on the basis of oral instructions (SRJa, s.v.). Thus in a contrast was made between written and unwritten sources, without regard to verbatim transmission. While the concept of verbatimness, as noted above, certainly existed in medieval Russian culture, there is no evidence for any kind of across-theboard commitment to fidelity in form and content. The authoritative religious texts of Eastern Orthodoxy provided models for wording changes and paraphrases in , as in “verified” reports repeated in different forms within a single book of Scripture (Savran 1988: 29–36) or non-identical reports in parallel Gospel passages (Ong 1981: 22–23, 1982: 64–65). On the other hand, medieval churchmen were familiar with the fourth-century bishop Spiridon of Trimithus’ objection to synonymic substitutions in Gospel citations even when they preserved the content (see Arximandrit Panteleimon 1904, 12: 348–49). Uspenskij (1984: 382) asserts that, in ecclesiastical language (Slavonic) of the late Muscovite era, “any deviation from the correct designation may be associated with a change in content.” However, this attitude, which is best documented for the time of the Old Believer Schism (mid- to late seventeenth century), seems to have applied mainly to the copying of scriptural and liturgical texts, i.e., reflect a genre-specific doctrine of verbatim transmission of written sources. Outside of canonical genres, one can see great deviations from fidelity in cases in which it is verifiable. For example, Avvakum, the first leader of the schismatic Old Believers, had no qualms about presenting his opponents’ version of a sacred text (which he regarded as heretical) in that was patently unfaithful (7): (7) Napecˇatano: [duxu lukavomu molimsja print- [spirit evil]- pray-.1. It is printed: “Let us pray to the Evil Spirit” (ibid.: 378).
Also telling is a case in which Avvakum attributes the words of Judas to his persecutor (Uspenskij 1987: 56–57) — that is, uses without regard to either form or content as a way of revealing his understanding of what his reportee was “really saying.” These are not cooperative usages, but they do provide evidence that direct reporting could flout verbatimness in medieval Russian culture. In legal-administrative text-kinds, the concept of verbatimness is likewise restricted to reports of written sources. Documents were generally quoted in
T
57
other documents in a relatively “faithful” manner; we have no knowledge of sanctions against bona-fide departures (i.e., ones that were not forgeries). In some instances, document citations were explicitly marked as verbatim (slovo v” slovo).32 It is important to note that this label was used exclusively with quotations from writing, which indeed is the only case in which extensive verbatimness is feasible in a premodern culture; the scribes never claimed that their reports of speech events were word for word. In fact, medieval Russian law seems to have recognized that its recording methods were imprecise at best and that the scribes could wittingly or unwittingly distort the purport of the testimony. This explains, in part, the institution of a procedure in which the litigants could endorse or dispute trial records in courts of higher instance, whose judges were not at the original hearings (see Chapter 6). As a text-kind, trial transcripts have particular relevance for discussions of reporting in that they provide a clear illustration of the fundamentally creative nature of , even when there are real anterior utterances. There can be no doubt that the scribes were responsible for constructing much of the wording and structure of the dialogue; nevertheless, the reports were subsequently accepted by the principal reportees as a valid representation of their positions. The constructed nature of the trial dialogue follows from the fact that writing is considerably slower than speech, even when technologically assisted (Chafe 1982: 36–38; 1984: 1095). In the production of trial transcripts, the lag between the speech events and its “transcription” (written reconstruction) was naturally increased by premodern technology of writing — coarse paper and quills that had to be continually dipped in ink and periodically resharpened. (Even with technological assistance, some modern court reporters make it a policy to lag somewhat behind the speaker “so that sense [can] easily be made of the utterances” (Walker 1986: 215).) A further factor that slowed the pace of writing was the unavailability of cursive; fifteenth-century chancery texts were written in the semi-uncial hand (poluustav), which had few ligatures and hence did not permit a smooth transition from one letter to another (see Tixomirov and Murav’ev 1982: 19–23). It is inconceivable that the scribes thought they were or should be reporting word for word when the testimony was so consistently outpacing them. Consequently, the scribes who wrote the trial records must have (re)constructed the dialogue from memory, with all its limitations, assisted by
58
R V
whatever sketchy notes they were able to make in the course of the hearings. (There is no reason to think that the testimony was ever held up for the scribe’s benefit.) There is evidence of a slightly later period for the existence of such notes. In a judgment charter of 1531, the plaintiff refused to verify the trial record during a higher-court hearing. Thereupon the judge of higher instance, who happened to be the grand prince, took the routine step of calling on the “men of court” (official witnesses) for confirmation; their statements prompted him to decide in favor of the respondents. However, there was an unusual sequel: six months later, the plaintiff produced a document that he claimed to be the genuine rough (“black”) draft (spisok” cˇernoj prjamoj). Both of the trial judges denounced this as a forgery, and the grand prince refused to withdraw his original verdict (N. P. Lixacˇev (ed.) 1895: 188–92, no. 10). Whether or not it was authentic, this rough draft proves that trial scribes produced some kind of preliminary version before drawing up the fair copy of the transcript. In representing the trial dialogue, the scribes, like modern court recorders (Walker 1986), parliamentary reporters, and newswriters (Slembrouck 1986, 1992) filtered out many features of orality — overlaps, interruptions, hesitations, disfluencies, auto-corrections, vagueness, and, to a large extent, the redundancy that is characteristic of spontaneous dialogue but has little informational value in writing. (Cf. the hemming and hawing in the American trial transcript — mechanically recorded in the first instance — quoted in Philips 1985.) The scribes also edited out most features that had primarily interpersonal meaning (phatic and emotive devices). Moreover, while one cannot know what was actually said, it is a suspicious circumstance that so much of the testimony adheres to the maxims of quality, quantity, relevance, and manner, especially given that many of the trial participants could not have been familiar with the conventions of the speech event. The scribes must have chosen what to report with a strict view to relevance, as defined by the institutional function(s) of the text-kind. (For what it is worth, none of the cases in which the transcripts were challenged mentions any objection to wording discrepancies.) Presumably what the scribes intended to record was a paraphrase accurate enough to satisfy the litigants that their positions were represented fairly — an adequation to type (Fludernik 1993), as defined by the purpose of the text-kind. Any stricter adherence to verbatimness, even if it were possible, would have exceeded the demands of relevance and hence would have constituted uncooperative behavior.
T
3.4
59
A convention of explicit typicality
Trial transcripts have particular significance for the issue of typification in (cf. Fludernik 1993), because they are a text-kind in which the reporters routinely indicated the typicality of their reports. This, at least, is how I interpret the use of the demonstrative adverb tak(o) ‘in this/that/such a manner’ in the standard tag formula (see 3.1).33 The pairing of this adverb with past-tense ’s, which is so characteristic for trial transcripts and related legal text-kinds, was not especially typical in other narrative genres. While undoubtedly a generic convention, it did not conflict with the ordinary functions of the adverb. When used cataphorically, as in the tag, tak(o) indexed the following clause (the beginning of the report) and hence served as a cohesive device, focusing the interpreter’s attention on the testimony as an important element in the discourse (the “attention-getting value” of cataphora, Grimes 1975: 317–18).34 Thus it could serve as an index of reported speech in the authorial narrative. In addition, the continual repetition of tak(o) as part of a set tag clause created a paradigm that linked all of the litigants’ and witnesses’ turns at speaking and marked them as information relevant to the goals of the judge(s) as the intended interpreter(s). Hence, like other anaphoric and cataphoric devices, it could promote the cohesion and enhance the general interpretability of the discourse. In addition, as a deictic manner adverb, tak(o) could do double duty as a comparator/approximative — that is, an explicit signal of typicality.35 In other words, tak(o) could function as a hedge addressed to the Maxim of Quality, ruling out verbatimness — one possible interpretation of — by signalling incompleteness, imprecision, or other kinds of approximation to the type of the utterance being evoked. (Cf. the use of demonstrative manner adverbs in typifying expressions such as so to speak, Modern Russian tak skazat’.) This quasi-evidential strategy (cf. Chafe 1986: 264–66; Mithun 1986: 89–90) would be well motivated in a legal context in that it would preempt litigious quibbling over wording. At this point I must pause to note that the use of tak(o) in tag clauses has been interpreted in the opposite way, as a marker of reproductiveness rather than typicality. Otin (1969: 55) claims that scribes used tak(o) and its near-synonym sice with to indicate “not only what was said… but also how the RS proceeded.” He goes on to argue that, with time, this meaning came to be hedged by the addition of jako ‘as, like’ as a correlative to tak(o)
60
R V
or sice. The resulting construction, which Otin views as nascent , purportedly allowed nonverbatim reporting by introducing an element of comparison (ibid.: 55–56). Otin’s explanation is untenable even apart from the theoretical problems with the verbatim model discussed above and the general unverifiability of any claim about verbatimness in premodern texts. To begin with, the tak(o)/ sice… jako constructions cited by Otin (ibid.: 56) are not correlative. In a correlative construction with adverbs of manner, two predicates would be compared to one another. If jako is indeed a comparator in Otin’s examples, the reports that it complementizes must be correlative not to the main clause but to omitted reports signalled by tak(o) or sice. A second problem is that Otin is not sifting the textual evidence properly. The tak(o)/sice… jako… construction, which by my observations is quite rare, appears only in later texts. By contrast, jako plus (jako recitativum), which Otin treats as a secondary development, is attested in the earliest Slavic writings, as indeed is full-fledged (see D. Collins 1996). Moreover, if the tak(o)/sice… jako… and jako recitativum constructions really did represent inchoate , one would expect to find cases of reports with indirect deictic orientation; in fact, examples of that kind are scantily attested, if at all, and, again, appear after the earliest cases of clear-cut . A final problem with Otin’s hypothesis is that it is based on faulty reasoning. If the scribes felt a need to indicate verbatim by means of tak(o)/sice, they must have perceived it as a special variety; concomitantly, they must have viewed the form of without tak(o)/sice — the unmarked, default variety — as at least potentially non-verbatim. Thus there would have been no need to single out the nonverbatim type as a special category by means of jako. In any case, it is unclear why the relative adverb of manner should signal non-verbatimness when its demonstrative counterpart was signalling the opposite. The fact that the use of tak(o) by itself was sufficient to introduce an element of explicit comparison (typification) is supported by cases of “omitted speech” (Thompson 1996: 518) or “narrator’s report of voice” (Semino, Short, and Culpeper 1997: 24–25) — clausal ellipsis in which the adverb stands in anaphorically for a report (8): (8) I Mamon s tovarišcˇi tako rkli… I and Mamon- with fellows-. thus speak-. and
T
61
[knjazja velikogo znaxori [Ontipa desjatckoj s [prince grand]- witnesses- [Ontipa tenthman]- with tovarišcˇi tak žo skazali. fellows-. thus say-. And Mamon and [his] companions spoke in this way []… And the grand prince’s witnesses Ontipa the Tenthman and [his] companions said likewise (ASÈI 1: 323, no. 432).
A similar example involving tak(o) without the intensifying particle že occurs in a subsequent passage (ibid.) While there is no way to establish whether the grand prince’s witnesses repeated Mamon’s statements word for word, it is unlikely a priori. Clearly the scribe was concerned with content rather than form and utilized omitted speech with tak(o) for the sake of economy. Omitted speech is infrequent in the transcripts; the scribes avoided syntheses, even at the cost of redundancy, in order to leave the main burden of sensemaking to the judges. However, this does not imply that they avoided typification. The use of tak(o) to signal typicality may be verified by examining the only kind of report in which verbatimness was feasible — citations from documents, which litigants put into evidence to establish their claims. Not counting instances embedded in the trial dialogue, the transcripts contain 109 full or partial citations from various kinds of legal documents. The sources are usually mentioned first in the reported testimony; then the citations are introduced in the narrative with an action clause followed by a minimal tag clause. The most common introductory formulae are illustrated in (9a–b); it is noteworthy that they do not include tak(o), even though they tag verbatim citations: (9) a.
I sud’ja gramotu… velel cˇesti; i v gramote and judge- writ- bid- read- and in writ- pišet… write-.3 And the judge ordered the… document to be read; and in the document is written… (ASÈI 1: 403, no. 525).
b. I [Ivan Vasil(’)evicˇ(’) vozreˇl v gramotu, i v gramote and [Ivan Vasil’evicˇ]- look- in writ- and in writ-
62
R V
pišet… write-.3 And Ivan Vasil’evicˇ looked at the document, and in the document is written… (ASÈI 1: 319, no. 430).
The verb of writing appears in the present tense because the tag is a narration of perception (a mode similar to ), with the judge as subject of focalization. This can be seen explicitly in cases in which the tag clause is subordinated to a verbum percipiendi by means of the complementizer ože (ož’) or aže (až’), both typically used for objects of perception.36 It is noteworthy that the tag in (9a–b) does not include an attribution, for all that it features a report index (the tag verb) and a redundant reference to the document as textual conveyor (the adjunct prepositional phrase). The verb pišet(”) in this formula does not have a recoverable subject and is functionally impersonal — a widespread usage in writings (Sreznevskij, s.v. “pisati”), though personal uses are also common.37 The use of subjectless or impersonal verbs to index reports was well motivated, given the general purpose of attributions in the text-kind (see 3.2). While identification of the speakers in the trial was important not only for coherence (tracking) but also on legal grounds, the attribution of the authors (or principals, Goffman 1981: 144–45) of cited documents was superfluous. It was generally reported within or inferable from the citation, and in any case had no evidentiary value; there was no need to (re)construct the position of the author or principal of documents produced outside of the trial context. (While information in the tag was also predictable, the tag itself was functional, since it promoted coherence by indexing the citation. Thus there are only a few examples in which it is omitted — in some cases, due to increased predictability in the context.)38 There is a strong preference for the direct mode of reporting in document citations of this kind. While there are a few cases of paraphrase (“speech summary”),39 verbatim transmission of written sources (not counting errors or minor corrections) was clearly the norm. This is confirmed in several cases in which the document survives in the original or in later copies. Even when there is no means of corroboration, the verbatim nature of the citations may be assumed on the basis of their diplomatic structure, which allows them to be compared with extant documents of the same text-kind. Only rarely do citations depart from the expected formulaic structure sufficiently
T
63
to suggest that they are abridgments or paraphrases rather than close copies of their originals.40 While tak(o) is standard in the clauses that introduce , it is scantily attested in citations of written documents (only three tokens). Significantly, its presence in these cases can be ascribed to typification marking. Two of the examples, which are citations of donation charters, appear back to back; the first of these is quoted in (10): (10) I [kn(ja)z’ Mixailo Andreˇevicˇ(’) v”zreˇl” v” [Olešinu and [prince Mixajlo Andreevicˇ]- look- in [Oleša- danuju gramotu; ož v gramoteˇ pišot tak: donation writ]- in writ- write-.3 thus [Teˇ svoi zemli Oleša dal [živonacˇalnoi [those . lands]-. Oleša- give- [life.giving Troici… [c(e)rk(o)v’ s(vja)tuju Troicju, da [Savrasovskuju Trinity]- [church Holy Trinity]- and [Savrasovskaja derevnju, da [Gridinu derevnju i s leˇsy, i village]- and [Gridina village]- and with forests-. and s požnjami. A poslus(i) v [danoj gramoteˇ: with fields-. and witnesses-. in [donation charter]- [kn(ja)z’ Semen Fedorovicˇ(’) Starodubskoj, da [Timofeˇi [prince Semen Fedorovicˇ Starodubskij]- and [Timofej Dmitreev s(y)n” Korjakina… A [gramotu danuju Dmitrij- son]- Korjakin- and [writ donation]- pisal [Oleša sam svoeju rukoju. write- [Oleša self]- [. hand]- And Prince Mixajlo Andreevicˇ looked at Oleša’s donation charter; lo, in the charter [it] is written thus: “Oleša gave those lands of his to the Life-Giving Trinity [Monastery]… [namely] the Church of the Holy Trinity and the village of Savrasovskaja and the village of Gridina, with both the forests and the fields. And the witnesses in the donation charter [were] Prince Semen Fedorovicˇ Starodubskij and Timofej, son of Dmitrij Korjakin… And Oleša wrote the donation charter himself, with his own hand” (ASÈI 1: 352–53, no. 467).
64
R V
If the citations in (10) and its sequel are compared with extant donation charters, they are revealed to be highly abridged paraphrases rendered in indirect speech rather than verbatim copies. All of the 113 donation charters from the archives of Trinity-Sergius Monastery published in the first volume of ASÈI, which range from 1392 to 1505, are written in the first person and have a different, more complex diplomatic structure than the citation in (10). Significantly, the scribe of the given transcript — an original rather than a copy — did not use tak(o) in tagging a subsequent citation, a letter from an absentee witness (ibid.: 353–54). While this is clearly an excerpt, given the absence of the incipit, closing formula, and other inferable information, it is in and almost certainly a verbatim transmission of the written words, rather than a paraphrase like (10). Therefore, in a series of reports that could be reproduced, the scribe indicated the non-verbatim renditions by tak(o), which he omitted before the verbatim rendition expected for citations. (The citation from the absentee witness’s letter shows that tak(o) was not simply mechanically transposed from the standard tag clause. The motivation for the was probably that the letter served as a written substitute for testimony; indeed, it is followed without any intervening narrative by the statements of witnesses who are present at the trial, which are likewise presented in .) The remaining instance of a documentary citation tagged with tak(o) (ASÈI 3: 55, no. 32) is in ; though selective, it is more or less verbatim, as can be seen by comparing it with its surviving original, a trial transcript of ca. 1416–17 (ibid.: 53–54, no. 31). This need not be interpreted as a counterexample to the use of tak(o) to signal typification, as illustrated in (10), since the citation is somewhat reworked and lacks the incipit and end protocol found in the original transcript. Judging by other examples, the general practice was to cite earlier trial transcripts in their entirety, regardless of their length.41
3.5
From conduit to collaborator
Accounts of that invoke verbatimness reflect the Conduit Metaphor identified by Reddy (1979), in which “the conveyor of information is seen as an inert vessel, … a voice giving form to information for which the quoted party is… responsible” (Tannen 1989: 108). By concentrating on the relation between reports and their “originals” (regardless of whether they exist) and
T
65
by taking exact reproduction as the standard from which to calculate deviations, such approaches ignore the intentionality and creativity inherent in direct reporting and hence present a distorted picture of its nature. The reporter is denied sovereignty — in effect, made subservient to the choices of an anterior speech will, which, if it exists at all, is not necessarily involved in or even relevant to the ongoing act of reporting. In reality, all reports — even apparently verbatim ones — are “communicatively subordinated” (Sternberg 1982b) to their contexts and serve the communicative goals of their reporters. As Cassirer (1955: 183) notes of the Aristotelian conception of imitation, “It no longer implies the mere repetition of something outwardly given, but a free project of the spirit: the apparent ‘reproduction’ (Nachbilden) actually presupposes an inner ‘production’ (Vorbilden).” Even when reports represent actual prior utterances, they are subject to new intentions that are working to coordinate sense-making with new addressees. Thus the crux of is not the relation between the report and the anterior utterance or between the reporter and the reportee; it is the relation between the reporter and the intended audience — the interpreter, to whose speech will the reporter is appealing. This is not to deny that in certain settings reporters can feel constrained by their social roles or by contextually dictated ideas of cooperation to represent the statements of anterior speakers in a relatively exact manner. An extreme case of this sort is the role of mouthpiece, where one speaker actually has the intention of serving as a conduit for the words of another, as when (if one can trust the internal evidence of orders in diplomatic dossiers) medieval Russian ambassadors to foreign courts read aloud, as tagged direct speech, prescribed written statements of the grand prince, for whom they served as official substitutes (e.g., Karpov (ed.) 1887: no. 7, dating from 1537; see also D. S. Lixacˇev 1946/1986). The individual speech will of such “transmitter-senders” (Dirven et al. 1982: 2–3) is not implicated in their official acts of communication, except in the sense that they choose to comply with their orders and, perhaps, select the time, place, and manner in which to do so. Only in special circumstances do reporters abdicate their speech wills in this way; it is rare for reportees to have such control — or indeed any control — over reporters. Thus the of a mouthpiece must not be taken as the core of the category. In more ordinary situations, reporters select the content, mode, and quantity of information to be conveyed on the basis of their communicative
66
R V
goals, which generally appeal not to anterior speakers but to the addressees with whom they wish to collaborate in sense-making. Typically, their choices are guided by several factors, including the purpose of the text-kind or speech in which they are engaged, their assessment of the speech situation, and their individual and genre-based understanding of cooperative behavior and relevance.42 In particular, the reporters must gauge what and how much information they need to give their addressees in order to aid their task of interpretation and steer them toward the intended understanding.43 In doing this, they are influenced by knowledge founded on their own experiences as hearers or readers. This applies not only to reporting but to the selection of any discourse strategy; as Green and Morgan (1981: 168) observe, “Production will be heavily influenced by the speaker or writer’s assessment of how an utterance is likely to be interpreted by a real or hypothetical audience.”44 Accordingly, reporters choose the direct mode of reporting because, by virtue of its formal properties, it allows them to provide their intended audience with a different quality and, potentially, quantity of information (which in the present sense is not limited to content or proposition). As I shall show below, this approach to does not contradict the many functionoriented studies that have correctly emphasized its capacity to create vividness, immediacy, or involvement, nor, indeed, does it rule out the fidelitybased (evidential) hypothesis. However, it does view the mode from a different vantage point in that it focuses on the work of the interpreter. A case can be made that this interpreter-based approach actually accounts for the other effects that have been attributed to . The property that distinguishes from other reporting strategies is not, as shown above, its treatment of the wording or message but rather its perspective. The reporter embeds some aspect of language use from a projected (represented) speech event in the ongoing speech event; the embedded segment has the same deictic orientation as the projected speech event and, optionally, other features not meant to be assigned to the authorial or narratorial voice of the ongoing speech event. (There can be ambiguous cases in which the interpreter has to infer the perspective, e.g., passages without any shifters or other distinguishing features, which are nevertheless most conveniently treated as direct by default because they are marked by paralinguistic signals typically associated with that mode or because they are contiguous to clear-cut stretches of and contain no indications of the authorial or narratorial perspective.) This perspectival property, which makes
T
67
no reference to other syntactic or lexical characteristics, is, together with the concept of a separate speech event, both necessary and sufficient to define the category.45 While some definitions assert that characteristically has main-clause status (see, e.g., Mayes 1990), this is not constitutive (even if prevalent or prototypical), as is proven by phenomena such as complementized , slipping, and “partial quotes” (“incorporated quotations”, Clark and Gerrig 1990: 791). Moreover, reports do not have to be unbound, pace Mayes (ibid.: 338–45, 358) and Romaine and Lange (1991: 267), to include the expressive syntax and lexicon typically, though not invariably, associated with , as is shown by “texture-analyzing ” (Vološinov 1929/ 1986: 130–32; Page 1988: 33, 37). It has long been noted that has a “‘theatrical’, playful, imaginary character” (Wierzbicka 1974: 272).46 By introducing a perspective separate from the ongoing speech event, the reporter is engaging in a form of sanctioned make-believe, an effort to (re)animate the reportee’s viewpoint (“voice”) by assuming — or re-enacting, when there is an actual anterior utterance — his position in the projected speech event (cf. Haberland 1986: 220–21). Acts of direct reporting are not descriptions but selective demonstrations of language behavior, which “work by enabling others to experience what it is like to perceive the things depicted” (Clark and Gerrig 1990: 765). To interpret the signals of perspectival shift, addressees must do more than recognize similarities; they must put themselves in the position of witness in the reported speech events and engage in “make-believe” perception (ibid.: 765, 793; Fónagy 1986: 255; cf. Walton 1973: 301, 303; Rosen 1980: 160). This, along with the technology of written language, may account for the salience of the notion of verbatimness. It is not because verbatimness is the prototype of , as argued by Short (1988) — a claim that would need to be verified culture by culture; nor is it because it is what the reporter is necessarily striving to achieve. Rather, the prominence of the concept is because it is what the addressee, as fictive witness of the reported speech event, is supposed, in effect, to experience. In this sense one can meaningfully talk about the “verbatim qualities of the pretended original” (Chafe 1994: 222) in fictitious , since the observation would then relate not to the actual actions of the reporter but to the expected perceptions of the interpreter (including the reporter as co-interpreter). While all reporting is a game coordinated between the reporter and the interpreter, the reporter gives the interpreter a more demanding, collaborative
68
R V
role in than in modes like that do not involve demonstrations. Here it is useful to recall Lakoff’s (1984) division of discourse strategies — particularly those associated with textual organization — into speaker-based and hearer-based ones, in dependence on which of the two participants in a dialogue bears more of the burden of sense-making (establishing coherence).47 Hearer-based48 strategies defer, or seem to defer, to the interpreter; they can be motivated by convention (as in politeness routines), social roles (as when there is an imbalance of knowledge or power), or other situational factors. In terms of Lakoff’s dichotomy, is undoubtedly a hearer-based strategy. Rather than assimilate the reported information to his own authorial viewpoint, the reporter puts the interpreter on an equal footing with himself, in the position of a witness who must evaluate the represented speech event, analyze characteristically diffuse information, and make the necessary deictic adjustments for himself. As Miller and Johnson-Laird (1976: 630) point out, “A speaker who uses direct quotation leaves more of the interpretive work to his listener” — in particular, the need “to reinterpret deictic elements… in a manner appropriate to the circumstances surrounding [the] original utterance.” The interpreter’s share in the creation of meaning is particularly heavy when the authorial discourse is minimal, as in drama-like narratives of speech events: “Narratives that rely totally or predominantly on records of characters’ speeches… entail more inference than other kinds of narrative, or if not more, at least a special kind of inference. To a greater degree than normal, the reader is required to interpret the illocutionary force of the sentences that are spoken by the characters; that is, he is supposed to infer what they ‘mean’ in the context of the action, even if there are no direct reports of the action, indeed even if the whole action can only be constructed through such inferences” (Chatman 1975: 244–45). The hearer-based nature of actually lies behind some of its most common discourse functions. For example, from Plato onwards it has been noted that creates a sense of “drama” or vividness — in other terminology, that it reflects and stimulates a higher degree of involvement (Chafe 1982; Tannen 1989) than other reporting strategies, largely due to its particularity.49 This is reflected in the typologically widespread, perhaps universal preference for with reports singled out for particular attention because they are narrative peaks, foregrounded, newsworthy, or otherwise prominent in discourse.50
T
69
The involvement-creating capacity of has sometimes been linked with its capacity to express affect or evaluation (e.g., Chafe 1994: 217, 223; Mayes 1990: 338–45; cf. also Fludernik 1993: 426, 428). In fact, as Clark and Gerrig (1990: 793–94) point out, involvement — in their terms, “engrossment” — is an effect of the direct experience characteristic of as a demonstration. I would further suggest that the involvement-creating capacity stems primarily from one specific aspect of this direct experience — the fact that makes greater demands on the interpreter, in comparison to other modes of reporting; in other words, the interpreter of does really have to be more active and hence is more emotionally involved in the sense-making process. According to Romaine and Lange (1991: 267), “Dialogue [ — DEC] creates involvement because it makes it possible for listeners to come closer to imagining the recounted action or speech rather than hearing about it.” This observation could be put in even stronger terms: makes it necessary for the interpreter to assume a perspective in the represented speech event — an associative act that must be performed if the report is to be differentiated from the narrative and assigned its proper place in the discourse; otherwise the discourse would be incoherent, a jumble of undistinguished, competing points of view. In addition to creating involvement, the stronger sense of (or need for) collaborating in a mutual endeavor can lessen the “social distance between participants” (Baynham 1996: 78). This explains the effect that has in promoting a feeling of solidarity: “When speakers demonstrate only a snippet of an event, they tacitly assume that their addressees share the right background to interpret the same way they do” (Clark and Gerrig 1990: 793). Thus Brown and Levinson (1987: 122) treat as part of the positivepoliteness strategy of “stressing common ground.” However, it is problematic to identify direct reporting exclusively with positive politeness or, in general, to set up one-to-one correspondences between the formal varieties of and specific functions (cf. Sternberg’s “Proteus Principle”, 1982b: 148). In specific contexts, depending on the division of labor, the act can focus on deference or self-suppression (see below) rather than on collaboration; in that case, can function as a form of negative politeness à la Brown and Levinson (1987), since the reporter by convention is not infringing on what can be seen as the addressee’s negative face, the right to independent interpretation. A further discourse function of that may be explained by its hearerbased character is the impression that the reporter is to a large extent
70
R V
refraining from “interference” in the report (see Leech and Short 1981: 324). By contrast, nondirect modes of reporting tend to give the impression that the reporter “intervenes as an interpreter between the person he is talking to and the words of the person he is reporting” and hence that the represented speaker is not being allowed to speak for himself (ibid.: 320; Caldas-Coulthard 1994: 304). This concept of narratorial interference or control has been applied not only to reports of actual anterior utterances but also — presumably in a figurative sense — to fictive reports: “The character’s individual manner… appears unmediated in his own ” (Uspenskij 1973: 43; see also Pascal 1977: 2; Short 1988: 67). The general notion of a “cline of interference” in reporting modes (Leech and Short 1981: 324; Slembrouck 1986: 48–49) is undeniably useful; however, it would be more realistic without the Conduit Metaphor and the implicit assumption of reporter commitment. In reality there is no report without some degree of authorial interference. Whatever the desired impression may be, in reality the reporter is not a go-between but the author of the report, regardless of whether, or to what extent, he has appropriated heteroglossic material; except in special cases, the represented speaker is neither a participant nor a consultant in the act of reporting.51 Every report is subordinated to the communicative purposes of the reporter (Sternberg 1982b). In fact, the impression of interference does not relate, except in a metaphorical sense, to how much the reporter lets the “original reportee” (if any) speak. Rather, it relates to the extent and explicitness of the reporter’s own work in imposing meaning — that is, to the degree to which he leaves, or seems to leave, the task of interpretation to his addressee. This “attempt to avoid interpretation and bias (or to seem to avoid them)” (Longacre 1983: 131) is not at all the same as letting the represented speaker have full “say”, i.e., control of the report’s meaning, even if it creates that illusion. (Represented speakers are not the authors, in any immediate sense, of the statements attributed to them. In cases of actual anterior utterances, they may, if reporters are cooperative, be the principals — “someone whose position is established by the words that are spoken, someone whose beliefs have been told, someone who is committed to what the words say” (Goffman 1981: 144).) Accordingly, the reporter can use as a form of apparent, if not necessarily actual, self-suppression; by treating his addressee as a witness, he can cede, or seem to cede, responsibility for imposing meaning on the report. Sometimes this strategy is hearer-based by default. For example, in some
T
71
instances it can be motivated by a difficulty or by an actual inability to interpret, as when is used as “a vehicle by which a message that one person finds incomprehensible reaches another who can understand it” (Miller and Johnson-Laird 1976: 630). In other instances, there can be social or cultural pressure to avoid the self-assertion and analysis involved in explicit interpretation, as in some cases of the verbatim citation of written authorities when “the language that was actually used has some importance because of its authoritativeness” (Chafe 1994: 217). Here the additional burden on the addressee may be a by-product of the reporter’s desire not to analyze what he regards as self-evident (cf. Du Bois 1986) or sacrosanct (“appeal to authorities”, Morawski 1970: 692–93). However, the use of verbatim quotation is not always by default; for example, it can serve to maximize the quality and quantity of information available to the interpreter and so to make the fictive act of witnessing as ample and authentic as possible. The apparent self-suppression of the reporter in deferring to the addressee can have several additional effects. One is a sense of objectivity (cf. Holt 1996: 226, 230), in proportion as the reporter gives fewer signals — within the report, at least — of trying to predetermine or manipulate the evaluation. (This, perhaps, is the “fidelity” on which the reporter-commitment model is grounded.) In a seeming paradox, another of the effects is implicitly evaluative and quasi-evidential — distancing or “disassociation of responsibility” (Clark and Gerrig 1990: 792; Thompson 1996: 513). Here the reporter is, in effect, leaving the report unanalyzed because he would not care to defend an analysis if challenged. To be sure, some degree of distancing is inherent in every form of explicit , since the reporter could simply aver (in the sense of Sinclair 1988: 23) heteroglossic material without identifying it as such. Indeed, distancing (removal of self) is a characteristic of any storytelling (Goffman 1981: 3, 147, 149, 151–52). To promote a stronger degree of distance in , the reporter must provide the addressee with additional signals of reduced responsibility in the context. Thus can create an impression of deferring to the interpreter. This is especially strong and well-motivated in covert, “objective” styles of narration such as that seen in medieval Russian trial transcripts, which offer a clear illustration of the hearer basis of direct reporting. The choice of as the standard reporting mode promoted the main purpose of the text-kind — to provide the judges with information that would promote a fair outcome. Accordingly, it was conventional for the scribes to refrain as much as
72
R V
possible from analysis and to suppress their own evaluations of what they recorded. By presenting reports as , the scribes relegated as much of the interpretive work as possible to those who were the ratified interpreters, legally empowered to dispense justice. Concomitantly, the scribes adopted other hearer-based strategies. For example, in order to avoid evaluating the actions of the participants, they set the reports in a minimal, prepatterned framework, consisting chiefly of tag clauses with semantically general ’s rather than evaluative speech-act verbs. These tags performed the important ancillary function of delimiting turns at speaking and hence of signalling the shift of viewpoints (interests in the litigation), but they provided no information about the represented speaker’s illocutionary intent or about the manner in which the depicted utterance was produced. The use of “a large and finely-differentiated lexicon” is a speaker-based strategy; explicit lexical items encode the speaker’s evaluation and leave little to the imagination of the interpreter (Lakoff 1984: 484). Thus the low type-to-token ratio of the authorial narrative is hearer-based; it contrasts not only with other text-kinds such as ecclesiastical writings but also with other contextualizations within the transcripts. Deference to the hearers may also be seen in the preference for the syntactically inexplicit forms of parataxis and polysyndeton. Hypotaxis is almost never found in the narrative framework, though it appears in the transcripts in other contextualizations and occurs extensively in other kinds of writing. By making syntactic relations explicit, hypotactic structures function as evaluations that the authors use to maintain greater influence over the interpretation (Lakoff 1984: 484, 486; Labov 1972: 387–92; Schiffrin 1981: 48–49). By contrast, paratactic devices are hearer-based: while conveying the same notional relations between clauses, they leave to the interpreters the tasks of inferring the nature of the connections and unravelling other ambiguities (Lakoff 1984: 486–87). Another facet of the choice of was the “theatrical”, re-enactive character of discussed above. It is doubtful that this was motivated by esthetic concerns, as in modern literature; even though the transcripts exhibit purposive, “artful” use of language, aesthesis in and of itself cannot be plausibly regarded as a major goal of the text-kind. If it were, the scribes could be reproached with artless overuse of vividness, given the low “reportability” (Labov 1981: 227–28) of many of the represented messages.
T
73
In fact, the re-enactive nature of promoted even the utilitarian purpose of the text-kind. In cases of judicial referral, the judges who were the intended audience and primary interpreters of the trial records had not been present at the original hearings; this lack of reliable first-hand knowledge could obviously have prejudiced the fairness of the verdicts, since the judges would have had to base their decisions on hearsay. Even the judges of first instance presumably must have often had imperfect recall of the trial proceedings, since the testimony was often complex, and a great deal of time could elapse between the hearings and the judgments. However, by hearing (and sometimes also reading) reports of the testimony presented in the comparatively non-analytic mode of , the judges became, in effect, witnesses of the trial proceedings. They were obliged to do the work of reconstructing the scene of the trial — to construct their own model of the speech event and draw their own inferences on the basis of the diffuse represented statements, which were re-enacted from the viewpoints of the trial participants — principals whose positions the reports established. As noted in Chapter 2, the interested participants had an opportunity during judicial-referral hearings to confirm that their positions were adequately presented. Verbatim reports in even a broad sense were not relevant to the situation; indeed they could never be remembered, given the oral nature of the on-site trial. Relevant, “faithful”, cooperative reporting consisted in providing perspectival information and enough of the content for the interpreters to reconstruct the positions conveyed in the testimony. Thus the use of with semantically general verbs of saying compelled the judges who read or heard the transcripts to acquire their knowledge of the cases piecemeal and to hand down their verdicts on the basis of their own reconstructive observations, in a like manner to the judges who supervised the on-site trials. Direct reporting reanimated the voices of the trial; it was a way of overcoming the evanescence of the oral trial situation, not so much by compiling a permanent written record (though that was a byproduct of the effort) as by providing the means by which an adequate image of the trial could be (re)constructed — one may almost say, reconstituted — in its relevant aspects. One of these aspects, in a time of mixed orality, was in fact the oral nature of the trial as an institution. It is significant, in this regard, that the judges in judicial-referral hearings are said to have heard the trial transcript. Despite the existence of a written record, it was conventional to retranslate back to the spoken medium, to reconstitute
74
R V
the trial primarily by ear — that is, to witness the trial as an oral proceeding, again in the manner of the on-site participants. In closing this chapter, I would like to propose, if only as a thesis for further discussion, that this compulsion to “witness” (recreate) the represented speech event is not a by-product of mimesis but rather the very heart of the direct mode of reporting. While much of the literature on from Plato onwards has focused on its purportedly mimetic nature, many instances of , as shown above, cannot be considered imitative in any strict sense; moreover, as Sternberg (1982a, 1982b) emphasizes, all is to some extent mimetic. What is unique about is the deictic adjustments that it imposes on its audience. Thus, while the concept of mimesis is oriented towards reporters’ fidelity to their “originals”, this may well be less central than the effect that they intend to have on their interpreters. The quintessence of , I would suggest, is not mimesis but methexis (“participation”), to adapt a term from Platonic and early Christian philosophy — an act that, in accordance with culturally determined interpretive conventions, allows the audience to participate both in the new event (the re-creation) and, vicariously, in the prior event, real or imagined, that is being represented.52
C 4 Residual forms in testimony
4.1
Participial tags and foregrounding/backgrounding effects
The standard reporting strategy discussed in the preceding chapter could, judging from the texts, be used to convey any kind of utterance made by one of the interested participants (litigants or witnesses) during the trial hearing. Nevertheless, it should not be regarded as a mere default device, if that is understood as an automatic rather than rational principle of selection, for two reasons. First, as shown above, each of the conventional elements of the standard strategy was aimed at the needs of the interpretation process, inferable from what is known of the institutional functions of trial records as a text-kind. Second, for all its possible breadth of application, the standard reporting strategy was seen to be particularly linked — in a correlation amounting to 100% of the possible examples — with the representation of speech events that can be considered prototypical for the trial hearing, as defined by both statistical prevalence and institutional centrality. While the vast majority of reports are presented in this conventional manner, there is a substantial residue of cases involving other reporting strategies. The methodology used in this study assumes heuristically that such residual forms likewise reflect rationally grounded acts of reporting, but in contextualizations that differ, sometimes subtly, from the prototypical speech event. As will be shown below, the results of the investigation confirm the prediction that the reports that depart from the standard are not random or accidental (unintentional) deviations but rather means of producing special effects when the speech events are perceived as nonprototypical. The use of nonstandard strategies is thus iconic of pragmatic difference. As a general principle, the further a contextualization departs from prototypical testimony, the more hospitable it is to reporting methods other than the standard one associated with such testimony.
76
R V
One of the most frequent of the nonstandard reporting strategies, which was evidently motivated by prominence considerations, is tagged with a formula involving the nonpast gerund of the recˇi ‘speak, say’ (1), instead of the preterit found in the standard tag: (1) a r(’)kucˇ(i) (X-) (pered” Y-) (tak(o)) and speak- (X-) (before Y-) thus And (X), (in the presence of Y) spoke [lit. speaking] (in this way).
The gerund, an uninflected deverbal adverb, derives historically from a form of the present active participle; by the fifteenth century, it no longer conveyed number or adjective-like information about definiteness, gender, or case, though it continued to express the verbal categories of voice, aspect, and Aktionsart.1 Gerund tags like (1) are common in other text-kinds (see Bondar’ 1967; Lopatina 1979: 423–24); indeed, the collocation a r(’)kucˇi is so common that it has been treated as a lexicalized quotative (Barnet 1965: 104–6; SRJa, s.v. “arkucˇi”). There are reasons to doubt this claim if it is meant imply that a r(’)kucˇi was a citation marker divorced from the paradigm of the verb recˇi. Univerbation cannot be proven, given the use of scriptio continua in ; the putative quotative does not occur in modern dialects. Moreover, r(’)kucˇi tends to be preposed, whereas clear-cut quotative particles like deˇ(i) and mol are generally intercalated in second position. Two additional facts suggest that relexicalization had not occurred in the fifteenth century. First, the gerund r’kucˇi could be modified by the adverb tak(o); adverbial modification would be unlikely if it were already a quotative particle or other nonpredicative element. Second, r’kucˇi retained its ability to take subject and indirect-object arguments (although the latter happens not to be attested in the corpus). The general function of nonpast imperfective gerunds like r’kucˇi is to indicate non-sequentiality.2 In context, they typically denote actions that are concursive with and subsidiary to some main event or part of a complex of actions viewed as one event (Barnet 1965: 104–5, 116; Nikiforov 1952: 265; cf. Jakobson 1957/1984: 51–52). Thus it is not surprising to find the gerund in (1) co-occurring with the conjunction a, which links clauses that are not in temporal sequence, in contrast to the conjunction i found in the standard tag and other plotline clauses.
R F T
77
There are three optional elements in the formula given in (1) — the subject noun phrase conveying the represented speaker, a prepositional phrase conveying the addressee, and the cataphoric adverb tak(o). (The corpus does not provide information about the relative order of the last two elements.) The subject of the gerund, which is postposed to the verb when it occurs in the tagging formula, is usually the same as that of the main clause. While the typical pattern in such cases is zero reference, the subject of the gerund can be made explicit under various circumstances, e.g., when it denotes a subset of the main-clause subject or when it is stranded from the main-clause subject. The adverb tak(o) is likewise postposed to the verb, not preposed as in the standard tag formula; it appears immediately before the report if there is an explicit subject. Since clause-final position is ordinarily correlated with newsworthy information in , the postposition of the optional elements suggests that the constatation of the speech act, as conveyed by the verb, is highlighted less here than in the standard tag. This inference is borne out by the use of a nonfinite . Cross-linguistically, participial clauses tend to be used as backgrounding devices.3 While they retain certain verbal categories and can perform the verb-like function of predication, participles convey fewer of the categories needed for full descriptions of events than do finite verb forms (cf. Hopper and Thompson 1984). From a pragmatic perspective, they compel the interpreter to look in adjacent clauses for much of the information that is canonically “verbal” (tense, mood, and person). This inferencing process (which, generally speaking, is more extensive with participial elements than with adjacent finite verbs) creates a subordination-like association between clauses; hence it makes participles and gerunds effective devices for presenting events as being either dependent on other events or less prominent than those denoted by main-clause finite verbs. This is especially the case with forms like the gerund that signal non-sequentiality, since sequenced events tend to be in the foreground (Givón 1984–90, 2: 840; Hopper and Thompson 1980: 281). As noted in 3.2, the institutional purpose of trial transcripts favored foregrounding not only of the testimony but also of the attribution, which comprised an identification of the speaker and a constatation of the speech act. The attribution was important because it facilitated interpretation of the drama-like texts and because every statement made by litigants or their
78
R V
witnesses was binding and potentially probative for the outcome of the trial. Thus prima facie the gerund tag would seem to be less than optimal for testimony (at least of the prototypical variety), since the combined effect of the gerund, the conjunction a, and the placement of the tag verb in nonfinal position (as seen in the postposing of tak(o) and/or the subject) was to background the constatation of the utterance — to present it as relatively less event-like than reports tagged with the standard formula. One explanation for the use of the gerund tag is that not all of the reports so introduced represent prototypical testimony occurring in a dyadic framework. In many cases, the evident purpose of the is to explain the sequenced, foregrounded nonverbal actions of topicalized participants. Commentary of this kind generally belongs to the background of discourse (Hopper and Thompson 1980: 280). However, it is by no means clear that every instance of the gerund tag has a backgrounding function, and some of them present information that is not presupposed and potentially newsworthy. In addition, the peculiar properties of gerunds made them suitable for other pragmatic effects; this fact may pose difficulties for a blanket interpretation of participial clauses as backgrounding devices. It must also be kept in mind that the tag and the report are independently subject to foregrounding or backgrounding. The only reporting method attested in the corpus with the gerund tag is , which implies that the reports themselves are foregrounded — an inference supported by the optional occurrence of the cataphoric adverb in final position in the introductory clause. Thus the use of the gerund to introduce would seem to establish a kind of middle ground in the discourse. The gerund clause signals that the represented utterance is ancillary, the that the content is newsworthy. The combined effect may be to suggest a lesser degree of newsworthiness than in prototypical testimony; in any case, the attribution information is treated as less newsworthy than with the standard tag. The gerund tag formula has three main contextualizations in trial transcripts — in descriptions of boundary-setting; with commentary on unexpected events; and in a special incipit formula. In the first contextualization, boundary-setting descriptions, longtime residents serving as partisan witnesses are depicted conducting the judge(s) around the metes and bounds of the disputed property and indicating where they think the landmarks are (see Kleimola 1972: 361–62 and 1975: 37–40). The witnesses’ actions are expressed in a narrative clause (often very extended) consisting of the
R F T
79
preterit of the verb povesti ‘lead’ (with the judge as understood patient) plus strings of prepositional phrases — e.g., “And [they] led up from the stream Medvežka along a ravine, and, having gone a little ways along a marshy stream, to the left past an ivy bush (it stands on the marshy stream) and past a little birch…”, etc. (ASÈI 1: 553, no. 639). Usually the descriptions end with a report in which the witnesses sum up their actions — e.g., “Having stopped, [they] spoke in this way: ‘On the left, lord, is the grand prince’s land, and on the right is the land of Trinity Sergius Monastery’” (ibid.). There can also be reports in the midst of the descriptions. Diverse reporting strategies are found in this contextualization. While the most common strategy is introduced by the standard tag formula, gerund tags are also attested, as are alternative tag verbs, , , and fused speech. The reason for this latitude is clear: the speech reported from boundarysetting procedures is not the prototypical testimony associated with the standard tag. To begin with, it does not occur in a dyadic framework. The witnesses are on a more active footing during boundary-settings and temporarily assume responsibility for the conduct of the trial. The remarks that they make are offered of themselves, not elicited by the judge (if internal evidence can be trusted; note that elicitation is generally made explicit in other contextualizations). Reported speech is not the sole or even primary conduit of evidence in such passages; the main emphasis is on the witnesses’ movements, and the testimony is, for the most part, ostensive rather than descriptive. Thus the reports that occur are not representations of “canonical talk” whose primary context is a speech event; they depict “coordinated task activity”, in which “nonlinguistic utterances have the floor” (Goffman 1981: 140–43). The reports serve chiefly as commentary on or interpretation of the adjacent narrative of events; thus they can be marked as less important than the prototypical testimony in the dyadic framework of the trial dialogue. Their attributions are also predictable, since boundary-setting canonically involved only a single set of speakers — hence the occasional use of (see 4.3). The gerund tag formula is one of the chief ways of indicating that the given report is ancillary to the surrounding description.4 At times it is used for a more specific effect, as in (2) — to distinguish running commentary [1] from the speech acts that close out the descriptions, which are more eventlike in the plotline and are generally introduced by the standard tag [2]:
80
R V
(2) Da ot mogily poveli tot že Elka and from grave- lead-. that- El’ka- svoimi s tovaryšcˇi napravo Kolycˇevskoju ... with fellow-. to.right [Kolycˇevskaja dorogoju, da z dorogi nalevo [poperecˇnim putem, a road]- and from road- to.left [transverse road]- and rkucˇi: Napravo zemlja [velikogo knjazja Ostrovskaja, speak- to.right land- [grand prince]- Ostrovskaja- a nalevo [zemlja Korobovskaja… Da priveli tem and to.left [land Korobovskaja]- and lead-. that- že vragom k granem k lipe, cˇto ravine- toward boundaries-. toward lime- that s Yvanom z Bitjagovskim roz”ezžali. Da stavši u with Ivan- with Bitjagovskij- ride-. and stand- at [toj lipy, Mantyreevy i Misajlovy [that lime]- Mantyrej-.. and Misail-.. znaxori [El’ka S’janov [svoimi tovarišcˇi witnesses.. [El’ka S’janov]- [. fellows]-. sud’i tako rekli: Cˇto esmja, gospodine, judge. thus speak.. what ..1. lord- veli… ino napravo zemlja i les Andreeva lead-. to.right land- and wood- Andrej-. Okljacˇeeva, [Isxodckogo sela, a nalevo zemlja, i Okljacˇeev- [Isxodskoe village]- and to.left land- and les, i pašnja, i požni, i selišcˇa wood- and field- and fields-. and hamlets-. Precˇistye [Simonovskogo monastyrja [Korobovskogo Virgin- [Simonov Monastery]- [Korobovskoe selca. village]- And from the grave that same El’ka with his fellows led to the right along the Kolycˇevskaja Road, and from the road to the left by a transverse road, and [1] they said, “To the right is the grand prince’s land Ostrovskaja, and to the left is the Korobovskaja land” … And they led
R F T
81
along that same ravine to the boundaries, to the lime tree, that they had demarcated with Ivan Bitjagovskij. And, having taken their stand at that lime tree, [2] Mantyrej and Misail’s witnesses El’ka S’janov [and] his fellows spoke in this way to the judge: “Where we, lord, have led, … to the right [is] the land and wood of Andrej Okljacˇeev’s village Isxodskoe, and to the left is the land and wood and plowed field and stubble fields and hamlets of the Theotokos Simonov Monastery’s village Korobovskoe (ASÈI 2: 435–36, no. 411).
The report with the gerund tag formula [1] is followed by two similar cases, which are omitted here for the sake of brevity. All three of these reports are commentaries on the preceding narrative of actions. Judging from (2) and two later boundary-setting descriptions in the same transcript (ibid.: 437–38, 438–39), the scribe used the standard tag formula for major stopping-points and the gerund formula for minor pauses and commentary. The same distributional strategy occurs in in a lengthy boundary-setting description in another trial transcript, in which the running commentary of the respondent’s witnesses is tagged eight times with a r’kucˇi and once, at the very end, with tak(o) rek(li) (AFZX 225–26, no. 258). An earlier boundary-setting description in the same transcript follows this pattern, though less consistently: the first four remarks during the boundary-setting are tagged with tak(o) rek(li), the next five (which occur after a long stretch of description) with a r’kucˇi, and the concluding remark with tak(o) rek(li) (ibid.: 224–25). The second main contextualization for the gerund tag with is with commenting on narrative events that contradict institutional expectations. As in boundary-setting descriptions, the reports function not as events per se but as amplifications of events; hence they are good candidates for backgrounding, like other collateral material. In one of the examples, the reported utterance is an attempt to explain why a witness is not competent to testify; in four others, including (3), it conveys litigants’ excuses for not producing witnesses or evidence that they had promised prior to an adjournment.5 (3) I sud’ja dal srok Ondrejku [dve and judge- give- term- Ondrejko- [two nedeli samomu stati i [gosudarja svoego knjazja weeks]-. self- stand- and [liege . prince Borisa postaviti… I na [tot srok… Ondrejko Boris]- put- and on [that term]- Ondrejko-
82
R V
stal že, a [gosudarja svoego knjazja Borisa ne stand- but [liege . prince Boris]- postavil, a rkucˇi Ondrejko: [Gosudarja moego, put- and speak- Ondrejko- [liege my]- gospodine, [knjazja Borisa ne otpustili [sic] s lord- [prince Boris]- release-. from Moskvy [knjaz’ velikij… Moscow- [prince grand]- And the judge gave Ondrejko an adjournment of two weeks to appear himself and produce his liege lord Prince Boris… And at the reconvenement … Ondrejko appeared but did not produce his liege lord Prince Boris, and Ondrejko said: “My lord, the grand prince did not grant my liege lord Prince Boris leave from Moscow…” (ASÈI 2: 536, no. 493).
This is not testimony in a legal sense, since it does not provide evidence about the contested property or further the gathering of such evidence. Nevertheless, the absence of promised witnesses or documents was potentially a crucial factor in the judge’s decision, since it could cast doubt on the litigant’s bona fides. This may explain why all of the gerund tags in this particular contextualization feature explicit subjects, which are seemingly pleonastic since the gerunds are not stranded from the main clauses. The identity of the speakers, though predictable, bears some emphasis in this situation because the judges would have had to pay special attention to how liable the represented speakers were for the no-shows. Since some excuse for no-shows was expected in the real-world context, the fact that the utterance occurred was predictable in (3) and similar cases and did not need to be asserted in a way that could distract from the central information about the litigant’s nonfeasance. By contrast, when, after a second excuse (also tagged with the gerund), the represented speaker in (3) finally fulfills his promise, his statement is introduced more prominently, with the standard narrative tag discussed in Chapter 3 (ASÈI 2: 537, no. 493). Unlike the first two, the third contextualization for the gerund tag does not appear to be a backgrounding device in any straightforward sense. It occurs in a variant form of the incipit that reflects a special procedure for initiating trials. The point of departure in all five tokens is a centering on the phrasal verb biti cˇel”m’ ‘petition; state in a petition’ (literally, ‘strike
R F T
83
one’s head’, a kowtowing deference ritual), which is followed by the gerund tag (4) [1].6 Superficially this use of the gerund resembles the class of collateral explanations exemplified by (3); however, given that the gerund is introducing the opening charges, it seems problematic to treat the reports as backgrounded material. (4) Se bil cˇelom [kn(ja)zju Mixailu Andreˇevicˇ(ju) [igumen lo hit- forehead- [prince Mixailo Andreevicˇ]- [Abbot Paseˇja [živonacˇalnye Troici Sergeˇeva Paseja]- [life-giving Trinity]- [Sergius- monastyrja, da igumen [Kirilova monastyrja Monastery]- and Abbot- [Cyril- Monastery]- Nifont”, a rkucˇ(i) [igumen Paseˇja tak: Dal Nifont- and speak- [Abbot Paseja]- thus: give- nam, g(o)s(podi)ne, … [Oleša Ofonas’ev s(y)n” Vnukova… us- lord- [Oleša Ofonasij- son]- Vnukov- [svoi zemli… A [igumen Nifont” bil [. lands]-. and [Abbot Nifont]- hit- cˇelom: A u nas v [Kirilloveˇ monastyreˇ [togo forehead- but at us- in [Cyril- monastery]- [that Oleši [ot(e)c’ Ofonasei postrig”sja v cˇernci… Oleša]- [father Ofonasij]- be.tonsured- to monks-. [1] Abbot Paseja of the Life-giving Trinity-Sergius Monastery petitioned Prince Mixail Andreevicˇ, as did the abbot of Cyril Monastery, Nifont; and Abbot Paseja spoke in this way: “My lord, Oleša son of Ofonasij Vnukov gave us… his properties…” [2] And Abbot Nifont petitioned: “But that Oleša’s father Ofanasij was tonsured as a monk in our Cyril’s Monastery…” (ASÈI 1: 352, no. 467).
As [2] shows, the verb biti cˇel”m’ could introduce by itself, without any supplementary tag. What, then, was the function of the seemingly pleonastic gerund? Evidently the same iconic principle of individuation that motivates hendiadys is in operation here: “The linguistic separateness of an expression corresponds to the conceptual independence of the object or event which it represents” (Haiman 1983: 783). The attribution at the beginning of the
84
R V
incipit conveys several bits of information, including the participation framework, the special character of the trial, and the content of the plaintiff’s opening salvo. When presented as separate blocks, these bits of information are treated as if they belonged to separate events or, at least, to separate phases of a single event. The overall effect of this analysis is to prolong and hence to highlight the attribution. While the first reported utterance in (4) is treated in this manner, the second (the respondent’s counterpetition) need not be singled out because the participants and the special character of the trial are already givens in the discourse. Within this analytic strategy, the gerund in (4) [1] serves as a sort of id est, singling out relevant portions of the preceding represented utterance without suggesting a new event in the plotline as a finite-verb tag would. This possibility also made gerunds suitable for “repairing” (disambiguating) untagged . For example, where an original fifteenth-century transcript has in its boundary-setting descriptions (see 4.3), a seventeenth-century version (reflecting a lost fifteenth-century original) has the introduced by gerunds (ASÈI 2: 391–82, no. 388, 394–95, no. 388a). In sum, the gerund tag formula presented in (1) is used for rather different effects in the corpus. In accordance with a cross-linguistic tendency to use participial clauses for backgrounding, it can introduce reports viewed as elaborations of the surrounding narrative rather than as separate speech events. Like other participial forms, gerunds convey less grammatical information than finite verbs and hence are iconically appropriate for presenting speech acts that are not perceived as central or full-fledged events. With , the gerund tag de-emphasizes the constatation of the utterance without backgrounding the information in the report, which is kept separate from the surrounding narrative. The overall effect is to set the report in a middle ground of discourse. In addition, the formula can be used to de-emphasize the attribution by allowing omission of the explicit subject and fronting of the to a position typically occupied by less newsworthy information. However, examples like (4) [1] suggest that the syntactic dependence of gerunds, the very property that makes them suitable for backgrounding, also allows them to be used to create what is arguably a foregrounding effect — the fission of a single speech act into quasi-independent chunks, each of which requires special attention. The possibility of the same form being used for two strikingly different effects illustrates the more general point that discourse strategies, including the varieties of reported speech, need not
R F T
85
exemplify a straightforward form-to-function relation. Sternberg (1982b: 148) has termed this the Proteus Principle: “In different contexts… the same form may fulfill different functions and different forms the same function [author’s emphasis]”. This lack of isomorphism not only justifies but necessitates a context-sensitive approach to the function of discourse units.
4.2
Alternative tag verbs and different participant status
Another departure from the standard tagging procedure is the cross-linguistically familiar strategy of varying the introductory verb in the preposed clause. While had an extensive repertory of possible tags, the verbs that are attested with in the authorial narrative of trial records are generally not lexically specific speech-act verbs (’s) or “graphic introducers” (Tannen 1986: 322) but rather ’s and hence near synonyms of recˇi. As these ’s are, for the most part, perfective and occur in the preterit tense, they differ only subtly from the conventional strategy. While the data are limited, it appears that the use of these alternative ’s was motivated in part by an ongoing lexical shift and in part by the possibility of indexing the participant status of the represented speakers. The verb recˇi, which appears in the standard tag clause ((1b) in Section 3.1), was obsolescent in later (see Lopatina 1979: 425); it was eventually displaced by other ’s such as s(”)kazati ‘say, tell-” (cf. skazat’). For example, in the heresy-trial transcript of Maksim Grek (1525–31), a text that was reworked because of its religious significance, forms of recˇi that appear in the earliest version are regularly replaced by other ’s in the later redactions (Pokrovskij (ed.) 1971: 27). If the assessment of the first foreign grammar of Russian (Ludolf 1696/1959: 5) is accurate, recˇi (ecclesiastical-register rešcˇi) had become restricted to the archaizing ecclesiastical register (Church Slavonic) by the end of the seventeenth century. The progress of this shift can be traced in trial transcripts of the fifteenth through seventeenth centuries. The first transcripts in which the preterit of s(”)kazati is the sole or chief means of tagging dyadic testimony date from the early 1540s, a period of political and social instability and hence perhaps of relaxed conventions in legal language.7 Although recˇi continued to appear in conservative documents, s(”)kazati came to predominate in transcripts from the second half of the sixteenth century onwards.8
86
R V
It is tempting to dismiss the sporadic appearance of ’s other than recˇi in the trial record as a mere “stylistic mistake” reflecting scribal dialect — an accidental violation of an archaizing convention rather than a purposive use of acceptable alternatives. This interpretation may perhaps be borne out by the scarcity of the other ’s as compared with the more than a thousand tokens of recˇi. However, two facts suggest that this explanation, while not necessarily false in effect, is inadequate and should only be used as a final resort. First, the alternative ’s are not mechanically plugged into the slot of recˇi in the standard tag formula. The adverb tak(o) tends not to be used; in the one case in which it occurs, it is postposed to the . If the appearance of the other verbs were due to an arbitrary violation of convention, one would expect contamination from the standard tag. Second, the distribution of the other verbs is not arbitrary; they are used exclusively before the of secondary participants. The canonical tag verb recˇi is the only verb used in tagging the testimony of the litigants, the leading figures in the trial hearing. While this verb is also by far the most common way of tagging the testimony of witnesses and other minor players, the usage in this context is less uniform; not only do alternative ’s appear, but the reports themselves can be treated in uncanonical ways (see 4.3–4.7). This latitude was functionally motivated, in accordance with the principle that the further a contextualization departs from prototypical testimony, the more hospitable it is to nonstandard reporting methods. Evidently iconicity is at work, with the scribes marking less usual contexts in less usual ways, or, perhaps, feeling less bound by convention in reporting the utterances of those with lower stakes in the trial. The fact that the convention being flouted involves an obsolescent verb provides intriguing evidence of pragmatic effects influencing the implementation of a lexical shift. Witnesses were on a peculiar footing in the trial, and their has a different status in the transcripts than that of the litigants (see also 4.3). This distinction is reflected, inter alia, in the address-forms that the judges are depicted using: while the litigants are addressed by name, if at all, the witnesses are addressed as brate ‘brother-’ or brat’e ‘brethren-.9 In the actual trial hearing, these terms, if used, were doubtless positive-politeness devices that softened commands addressed to participants without any direct stake in the trial (cf. Brown and Levinson 1987). In the transcripts, they had the function of signaling a shift to a new turn at speaking and perhaps also to a new type of verbal encounter (see also 5.1). Whereas the
R F T
87
litigants’ speech, undoubtedly the central part of the testimony, takes place in a dialogue with the judge that usually runs for the entire length of the trial, the witnesses typically make one individual or collective statement and then cease to be participants; they have their say and then retire “backstage”. Moreover, the witnesses’ testimony has a different illocutionary intent: the litigants speak to establish their own claims, the witnesses to corroborate those claims. It is significant here that the witnesses are the only participants who are asked to testify on oath — i.e., to indicate the reliability of their evidence. This may be seen from the formulae in (5a–c), which routinely accompany questions addresssed to witnesses (173 tokens in the corpus).10 (5) a.
S”kaži(te) (brat(’)e) (v”) (bož’ju) prav’du tell-./ (brother(s)-) (in) [(God-) truth]- Tell [us], brother(s), (in) God’s truth…
b. S”kaži(te) (brat(’)e) kak” pravo (pered” bogom’) tell-./ (brother(s)-) as truly (before God-) Tell [us], brother(s), as is true before God… c.
S”kaži(te) (brat(’)e) po velikogo knjazja kr’st’nomu tell-./ (brother(s)-) by [grand prince]- [of.cross ceˇlovan’ju kiss]- Tell [us], brother(s), in accordance with [your] fealty oath to the grand prince…
One may therefore posit that the witnesses’ peculiar footing and the discrete character of their testimony may have influenced the occasional use of tag verbs other than the standard recˇi. Note that the the latter never appears in formulae like (5a–c). There are six instances of witnesses’ testimony presented as tagged by the preterit of s(”)kazati (usually written sk-), the most widespread of the alternative verbs of saying (see also Chapters 6–8). Four of the cases tag what may be termed multiplex testimony, in which a single group attribution covers a series of separate reports, each of which is a response to the same question (see 4.3). Hence the witnesses’ statements do not appear in a straightforward dyadic context (6):
88
R V
(6) I sud’i vsprosili starcev i xrest’jan: and judges-. ask-. elders-. and peasants-. A vam kak vedomo o [tex luzex? I and you-. how known about [those meadows]-. and starcy i xrest’jane tak rkli: Nam, elders-. and peasants-. thus speak-. us- gospodine, vedomo po tomu ž, kak skazali [deti lord- known by that- as tell-. [children bojarskie. A [gorodnye xrest’jane [Vasko of.boyars]-. and [of.town peasants]-. [Vasko Overin, da Negodja, da Oleksejko skazali: Overin]- and Negodja- and Oleksejko- say-. My, gospodine, iz Mencˇakova vyšli… we- lord- from Mencˇakovo- come.out-. And the judges asked the elders and peasants, “How do you know about those meadows?” [1] And the elders and peasants spoke in this way: “We, lord, know in the same way as the boyars’ children said”. [2] And the town peasants Vasko Overin, and Negodja, and Oleksejko said, “We, lord, came from Mencˇakov…”, etc. (ASÈI 2: 534, no. 492).
The first of the responses, in (6) [1], illustrates the fact that the standard recˇi can also appear in this context. In choosing s(”)kazati to tag the second response [2], the scribe may have been influenced by its use as a metapragmatic label for the witnesses’ institutionalized speech act, as in (5a–c).11 Of the remaining two cases of s(”)kazati with , one involves testimony that differs from the prototype only in its being attributed to a witness (ASÈI 1: 547, no. 635); however, it is worth noting that the scribe (or the mid-sixteenth-century copyist) had a proclivity for this verb, since he also made use of it with (see Section 4.4). The other case of s(”)kazati with occurs in a recognizable non-dyadic contextualization — a boundary-setting description. As discussed in 4.1, the diversity of reporting strategies found in this context is probably due to the fact that the reports are commentary on visual evidence rather than prototypical testimony. In the given instance, s(”)kazati appears in the middle of a description, while recˇi introduces the witnesses’ closing remarks (ASÈI 1: 492–93, no. 595). This distribution can
R F T
89
be compared to that of the gerund instead of the preterit in some of the cases discussed in Section 4.1. The tendency to present nonprototypical speech events by nonstandard strategies is illustrated clearly by a case of the verb ot”kazati ‘answer’, which, given its meaning, could conceivably have been used for any statement elicited by a question. In fact, there is only one example (7), which conveys a speech event that is unusual for the text-kind: (7) I [knjaz(’) Ivan Jur’evicˇ(’) poslal k” mitropolitu and [prince Ivan Jur’evicˇ]- send- to metropolitan- [s(y)na svoeg(o) kn(ja)zja Ivana, veleˇl ego sprositi, [son . prince Ivan]- bid- him- ask- on li posadil [teˇx mužikov, Okulika da he- settle- [those peasants]-. Okulik- and Olferka, na tom na [Šiškinskom selceˇ, i skol’ Olferko- on that- on [Šiškinskoe village]- and how.much davno oni tut živut. I [mitropolit long.ago they- there live-.3 and [metropolitan Zosima otkazal [kn(ja)zju Ivanu Jur’evicˇ(u) tak: [To Zosima]- answer- [prince Ivan Jur’evicˇ]- thus [that selišcˇo Šiškinskoe — zemlja [Simanovskog(o) manastyrja… village Šiškinskoe]- land- [of.Simonov monastery]- And Prince Ivan Jur’evicˇ sent his son Prince Ivan to the metropolitan [and] bid [him] to ask him whether he had settled those peasants, Okulik and Olferko, in that village Šiškinskoe, and how long they had been living there. And Metropolitan Zosima answered Prince Ivan Jur’evicˇ as follows: “That village Šiškinskoe is land of Simonov Monastery…”, etc. (ASÈI 2: 411, no. 402).
Here the judge, verifying an assertion made during the hearing, is communicating with an witness through an intermediary, outside the context of the trial hearing. As the mediated question is itself unusual for the text-kind, it is not surprising that it is presented as , i.e., differently than prototypical questions (see Chapter 5). The metropolitan’s relayed answer is presented as evidence and given as , as if part of the hearing. The tag clause contains no reference to mediation, and it is followed without seam by the continuation of the trial narrative. Nevertheless, the heterogeneous nature of the
90
R V
metropolitan’s reply is clearly signalled in three ways: by the atypical tag verb, by the postposition of tak(o), and by the reference to the addressee. Judging from its use in other text-kinds, ot”kazati ‘answer’ often takes a dative of addressee (Sreznevskij, s.v.; SRJa, s.v. “otkazati”; cf. AI 1: 144, no. 101) — naturally enough, since it includes a semantic component of reaction to an interlocutor — whereas recˇi rarely does so in the transcripts investigated (see 3.1). A final case that illustrates the correlation between unusual speech events and unconventional representational strategies involves the m”lviti ‘say’ (usually written molv-). This verb is attested several times in reports within (see 7.2). The sole instance in the narrative portion of the trial record (8) occurs with what is essentially non-testimony — a profession of ignorance by the witness of an equally unknowing respondent. (8) I Mixal’ tako rek: Jaz, gospodine, zemli ne and Mixal’- thus speak- I- lord- land- znaju, mne sja zemlja dostala novo, a know-.1 me- land- obtain- newly and znaet, gospodine, zemlju [Fedor Kokoš, poselskoj know-.3 lord- land- [Fedor Kokoš steward]- [velikogo knjazja. I sud’ja vsprosil Kokoša, i [grand prince]- and judge- ask- Kokoš- and Kokoš molvit: Jaz, gospodine, mež ne Kokoš- say-.3 I- lord- boundaries-. znaju. know-.1 And Mixal’ spoke in this way: “I don’t know the land, lord, I just obtained it, but Fedor Kokoš, lord, the grand prince’s steward, knows the land”. And the judge asked Kokoš, and Kokoš says, “I don’t know the boundaries, lord” (ASÈI 1: 321, no. 431).
Unusual features tend to cluster in the corpus. This probably reflects purposive use of language rather than scribal ineptitude; there is a strong correlation between departures from the genre-specific narrative norm and contextualizations that represent nonprototypical speech events. The given transcript contains at least four atypical features in addition to the lexeme m”lviti: the use of the historical present; omission of the judge’s question to Kokoš (see
R F T
91
5.3); pronominal anaphora in the text immediately following (8) (see (4) in 3.1); and the preterit of recˇi as a tag for the judge’s speech in the same passage (see 5.4). Cross-linguistically, the historical present tends to be used to increase vividness by substituting the perspective of the ongoing speech situation for that of the narrative. In other words, it is an icon of foregrounding, like other elements replicating “the ego-hic-nunc context of the speech event” (Chvany 1990/1996: 288). Recent studies treat the historical present as an involvement strategy that allows listeners to experience the narrated event themselves and draw their own conclusions about its significance.12 As shown by trial transcripts and other text-kinds, the historical present in legal language tends to co-occur with — in particular, reports ascribed to adversaries. It is most common in emotionally charged contexts such as petitions, where the writers or principals express personal viewpoints and appeal to the sympathy of their addressees. This explains why it often occurs within the in trial transcripts while being rare in the narrative, which is written from the relatively impersonal perspective of an effaced scribe, often unnamed in the text, who presumably had no stake in the outcome of the trial. This general distribution supports the idea that the historical present had an evaluative, involvement-creating function in , as in other languages. It is significant that one of the few cases in which the historical present occurs in the authorial narrative involves the verb m”lviti, which was typical of emotive contexts in northeastern documents. According to Mixajlovskaja (1980: 47–50, 181–83), this verb was colloquial and had negative connotations in . However, this nuance may not have been present in every dialect; for example, in western and northwestern documents, m”lviti appears to have been neutral, as it is in West Slavic (where it denotes unfocused linguistic action). Also problematic is the notion that the verb was colloquial, given that it can appear in trial transcripts and other kinds of chancery writing, which are not colloquial in character any more than they are literary. There is no evidence that its occurrence in (8) and elsewhere in the transcripts is due to stylistic mistakes. In fact, most of the examples of m”lviti found within the dialogue seem to be used purposively, to convey a sense of futility or even exasperation (e.g., the admission of defeat in Kaštanov 1970: 372, no. 16) — a nuance furthered in some cases by the use of the historical present.13 The same connotation of ineffective speech is found in (8), since the uncooperative
92
R V
represented speaker, Kokoš, frustrates the expectations created by his wouldbe principal. (Uninformative or otherwise uncooperative speech can, of course, also be tagged by the neutral recˇi, e.g., ASÈI 1: 236, no. 326). While evaluations of this kind are relatively common within the testimony, they are not hearer-based discourse strategies (see 3.5) and thus exceed the usual interpretive role of the scribes; this explains why m”lviti is otherwise unattested in the authorial narrative of trial records.
4.3
Free direct speech and cohesion effects
As noted above, the standard reporting method in the trial narrative foregrounds both the report and its attribution (the constatation of the speech act); the latter is prominent because any statement made by one of the interested parties in the lawsuit is potentially probative. While tagged with a preposed narrative tag is accordingly the most common method by far for reporting testimony, there are also a number of cases of free () (“zero quotatives”, Mathis and Yule 1994; “null quotation formulae”, Longacre 1994). This mode of reporting is typologically widespread. To a large extent, the literature has emphasized its capacity to create “drama” — vividness or involvement above and beyond the effects of tagged , particularly at discourse peaks. The use of this strategy in trial records shows that there can be other effects than involvement; indeed, some of the cases could only be infelicitious if involvement were a factor, given the low degree of reportability. It is generally agreed that felicitous use of depends typically on the identity of the reportee being recoverable or inferable from the context, e.g., through “regular alternation of speakers” and/or “content appropriate to a given speaker” (Longacre 1994: 130). (It can also be felicitous in the less common cases in which the attribution is immaterial, as when “talking heads” are depicted collaborating in the development of an argument or story — for example, Eco’s three fantasizing editors in Foucault’s Pendulum.) In 3.1, I discussed two cases that illustrate this principle in that the was used to present the opening statements of the trial hearings, where the attribution was easily recoverable. Interpreters familiar with the text-kind would be able to infer that the plaintiff, whose identity was established in the preamble, would be the first represented speaker — a deduction rein-
R F T
93
forced by the itself, which included trial-initiating formulae and forms dependent on the plaintiff’s viewpoint. If predictability of attribution alone was sufficient to motivate , the strategy should have been possible with most of the testimony presented in a dyadic framework, since the represented speaker’s identity can generally be inferred from the tag to the preceding question. The fact that is actually rare in this context suggests that such predictability was not the primary factor. Scribal carelessness — i.e., nonpurposive usage — is not a likely or viable explanation, because the three cases attested in dyadic all appear in representations of the same kind of speech act — a litigant naming and presenting his witnesses — and involve the same formula (9).14 (9) I sud(’)ja vsprosil Stepanka i v” vseˇx” and judge- ask- Stepanko- in [all kr(e)st’jan meˇsto: Komu ž to u vas peasants]-. place- who- that- at you-. veˇdomo, [dobrym ljudem, starožilcem? Veˇdomo be.known- [good people old.residents]- be.known- to, g(ospodi)ne, u nas” [Ostašu Paninu da Fetku… that- lord- at us- [Ostaš Panin]- and Fedko- And the judge asked Stepanko on behalf of all the peasants, “To what good people, longtime residents, among you is this known?” “Among us it is known to Ostaš Panin, and to Fedko…” [etc.] (ASÈI 1: 478, no. 589).
Though found in a dyadic context, the underlined report is not prototypical testimony whose purpose is to provide evidence about the disputed property; rather, it is a preliminary to the gathering of such testimony. There are several additional cases in the corpus in which the act of naming witnesses is conveyed by an uncanonical method (see Section 4.6).15 This follows the principle that deviations from the conventional reporting methods tend to occur when the represented speech event is atypical in some respect — a factor also present with the used to present the opening charge. In (9), the atypical strategy may have been promoted not only by the predictability and lesser importance of the attribution but also by the high degree of cohesion within the question-and-answer dyad (cf. the repetition of to and veˇdomo and the echo of u vas by u nas), which may have worked against the
94
R V
use of a tag clause (a delimiting device). Any doubt about the attribution would have been dispelled by the use of the vocative g(o)s(podi)ne, a testimony marker (see Section 3.2). Another contextualization in which is found is in boundary-setting descriptions. Three such cases, all from a single document, were discussed briefly in 4.1 (ASÈI 2: 391–82, no. 388); it was noted that a seventeenthcentury copy of the same trial record, thought to reflect a fifteenth-century version made shortly after the lawsuit, features the same reports presented as with a gerund tag (ibid.: 394–95, no. 388a). While it is impossible to determine when the gerunds were inserted, the fact of the substitution provides independent evidence for the general preference for tagged , which I identified on the basis of statistical predominance (see Section 3.1). Boundary-setting descriptions are, indeed, a favorable contextualization for , as for other uncanonical reporting methods; ten cases are attested, counting the three mentioned above. In each instance, the presence of is signalled unambiguously by forms dependent on the reported speech event, as in (10a). (10) a.
I Uvar”, i Gavša, i Ignat tak rkli: and Uvar- and Gavša- and Ignat- thus speak-. Znaem, g(o)s(podi)ne; poidite [sic], g(o)s(podi)ne, know-.1. lord- go-.2. lord- za nami; a my, g(o)s(podi)ne, tebja po after us- and we- lord- you-. along meže vedem. Iz poleˇs’ja poveli… boundary- lead-.1. out.of wood- lead-. na bereg po vetlu po vilovatuju, onto bank- as.far.as willow- as.far.as winding- po [samye rozsoxy. Po [ta meˇsta, as.far.as [very forks]-. as.far.as [those places]-. g(o)s(podi)ne, znaem: to, g(o)s(podi)ne, lord- know-.1. lord- meži [mitropolicˇe požne s [Sysoevoju boundaries-. [metropolitan- field]- with [Sysoj- požneju. field]-
R F T
95
And Uvar and Gavša and Ignat spoke in this way: “We know, lord; come, lord, after us, and we, lord, will lead you along the boundary”. From the wood [they] led… [etc.] to the bank as far as a winding willow, as far as the very forks [of the willow]. “This far, lord, we know. There, lord, are the boundaries of the metropolitan’s field with Sysoj’s field” (ASÈI 3: 462, no. 477; also in AFZX 253, no. 308).
Direct deictic elements and the vocative g(o)s(podi)ne ‘lord’ serve as boundary signals, reinforcing the interpreters’ expectations of what should come next, based on their background knowledge of the text-kind and trial procedure. As noted in 4.1, the statements made during boundary-setting depart from the dialogic framework characteristic of prototypical testimony in the trial hearing. The witnesses are depicted speaking without prompting, and the primary form of testimony is gestural rather than verbal; much of the emphasis is on what the witnesses show the judges rather than on what they say. Reported speech tends to be concentrated at the end of the descriptions, as a summarizing evaluation of the ostensive testimony; often, though by no means always, it appears in a set form like (10b) (where A and B are individual or institutional owners, and X and Y the names of properties): (10) b. napraveˇ — zemlja (NP-) NP-/, a naleveˇ — to.right land- (X-) A-g/, and to.left zemlja (NP-) NP-/ land- (Y-) B-/ To the right is A’s land (X), and to the left is B’s land (Y).
The occurrence of a report in some form was conventional in boundarysetting descriptions; the fact that the speaker’s identity was easy to recover permitted, though it did not motivate, the felicitous use of tagless strategies. One factor that apparently promoted was the possibility of viewing the witnesses’ statements and the nonverbal actions that they performed during boundary-settings as parts of a single complex event — a single instance of multimodal testimony. As the and demonstrations were perceived as more closely related than two adjacent turns in ordinary testimony, they could be treated as a cohesive unit by omission of the intrusive tag that would otherwise delimit them — an icon of distance (cf. Haiman 1983: 67; Mathis and Yule 1994: 67).
96
R V
The existence of such a perception can be deduced from the seamless transition from to narration and back that is characteristic of boundarysetting descriptions. While uninterrupted themes are typically, though not invariably, indicated by zero in , a theme that is interrupted by (including ) does not have be resignalled explicitly after the report; this is true even when the contains intervening themes or boundaries, as in (10a): “And Uvar and Gavša and Ignat spoke in this way: [] From the wood [they] led…”16 In these examples, the abruptly gives way to narrative, without any transition. (By contrast, in , the theme tends to be reindicated after any instance of direct speech, despite the fact that uninterrupted discourse themes are generally assigned a zero marking; see Nichols 1985: 173, 177.) Given that thematic continuity indicates a natural subdivision of the discourse, the successive actions of the participant(s) in these cases are evidently treated as a unit, even when one of the actions is an utterance reported by the interruptive strategy of . It would seem that is the converse of zero-theme marking: the fact that the narrative fades into without any overt signalling presents the witnesses’ statements and movements as manifestations of a single continuous instantiation of testimony. The same kind of grouping effect may be seen in a further contextualization of , which also involves what can be viewed as continuous multimodal testimony. In (11), a litigant is depicted placing a document into evidence before the judge [1]. In accordance with custom, the document is cited verbatim [2] (as far as can be determined from the internal evidence, see 3.3); then the report of the same litigant’s testimony is resumed without any new tag clause, despite the intervening topics [3]: (11) I Fegnast” tako rek: Na tot, g(ospodi)ne, leˇs and Fegnast- thus speak- for that- lord- wood- u nas [gramota danaja [kn(ja)zja Mixaila Ondreˇevicˇa… at us- [charter donation]- [prince Mixail Andreevicˇ]- a s toe, g(ospodi)ne, gramoty spisok pered and from that- lord- charter- copy- before toboju. I sud(’)ja vozril v” spisok; i v” you-. and judge- look- into copy- and in spiskeˇ pišet: [text] I tot, g(ospodi)ne, leˇs… copy- write-.3 and that- lord- wood-
R F T
97
[1] And Fegnast spoke in this way: “For that wood, lord, we have a donation charter of Prince Mixail Andreevicˇ… [etc.] and, lord, a copy of that charter is before you.” [2] And the judge looked at the copy; and in the copy was written: [text]. [3] “And that wood, lord…” [etc.] (ASÈI 2: 329, no. 337).17
Experienced interpreters — especially those familiar with the cited text-kind — would have had no trouble determining where the documents left off and the testimony resumed, given the clues provided by vocatives and other features dependent on the represented speech event. The attribution of the was predictable because it served as commentary on the documents; it did not need to be demarcated in a way that would have falsely implied the onset of an entirely new turn at speaking (not to be equated with a new utterance). The absence of a tag has the effect of clustering the report with the document that it elucidates. A comparable case of creating multimodal testimony occurs with a report placed before a document citation to explain why a copy is being offered instead of the original (ASÈI 2: 416, no. 404). As background to the documentary evidence, the report can be treated as part of the same testimony. Multimodal testimony is analogous to another common contextualization for — multiplex testimony, in which a single question is followed by a series of separate reports with a collective attribution (see 4.2). Multiplex testimony permits a number of tagging techniques, all of which differ, to various degrees, from the standard way of representing testimony in the investigated transcripts, in accordance with the principle noted in 4.1 that the further a contextualization departs from prototypical testimony, the more hospitable it is to noncanonical reporting methods. In multiplex testimony, the parallel responses to a single question are presented as equipollent parts of a unified speech event with a single purpose — to support the claim of one of the litigants. In other words, they are treated as a single instantiation of testimony delivered over more than one utterance, just as reports and gestures in boundary-setting descriptions or with document citations are treated as a single instantiation of testimony delivered in more than one medium. This recalls the use of when reporters wish to indicate the “convergent behavior” or “similarity and shared knowledge” of two reportees (“two speakers, one voice”; Mathis and Yule 1994: 64, 72). The distinct nature of multiplex answers is manifest even when there is
98
R V
no departure from the standard reporting strategy (see 3.1). In a prototypical question-and-answer dyad, the narrative clauses that tag each represented utterance generally begin with the sequencing conjunction i; as a result, the entire trial dialogue is presented as a single cohesive chain of identically conjoined sequential events. In multiplex testimony, by contrast, only the first tag clause begins with i; those that frame the subsequent parallel answers almost invariably begin with the nonsequencing conjunction a.18 When the multiplex testimony is completed, the tag for the judge’s next question — the following turn in the real-world speech event — begins again with i, as a new link in the narrative chain, as in (12).19 (12) I sud(’)ja v”prosil [Vaska Bužoniny, da Manaka, da and judge- ask- [Vasko Bužonina]- and Manak- and Jurki, — da Maksima: Skažite, brate, v” [b(o)ž’ju Jurka- and Maksim- tell-. brethren- in [God- pravdu, cˇ’e to selišcˇo, na koem” truth]- whose- village- on . stoite? I Maksim” tak” rek”: Jaz”, stand-.2. and Maksim- thus speak- I- g(ospodi)ne, pomnju za šest’desjat leˇt… A lord- recall-.1 for sixty- years-. and Manak” tak rek”: A jaz, g(ospodi)ne, pomnju Manak- thus speak- and I- lord- recall-.1 za pjatdesjat leˇt… A [Vas(’)ko Bužonina tak for fifty- years-. and [Vas’ko Bužonina]- thus rek: A jaz, g(ospodi)ne, pomnju za sorok” speak- and I- lord- recall-.1 for forty- leˇt… A [Jurka Melexov tako rek: A jaz, years-. and [Jurka Melexov]- thus speak- and I- g(ospodi)ne, pomnju za polpjatadesjat(’) leˇt… I lord- recall-.1 for 45- years-. and sud(’)ja v”sprosil Olekseˇika… judge- ask- Olekseik- And the judge asked Vas’ko Bužonina and Manak and Jurka and Maksim: “Say, brothers, in God’s truth, whose village is it in which
R F T
99
you stand?” And Maksim spoke in this way: “I, lord, remember for 60 years…” [etc.] And Manak spoke in this way: “And I, lord, remember for 50 years…” [etc.] And Vas’ko Bužonina spoke in this way: “And I, lord, remember for 40 years…” [etc.] And Jurka Melexov spoke in this way: “And I, lord, remember for 45 years…” [etc.] And the judge asked Olekseik… [etc.] (ASÈI 2: 519, no. 481).
The same pattern is found in a combination of multiplex and multimodal testimony, when an absentee’s letter is treated in parallel with the of cowitnesses (ASÈI 1: 353–54, no. 467), and in multiplex written evidence, when two documents produced by the same party are cited back to back (e.g., ASÈI 1: 538–39, no. 628). Rather than being linear like the reports whose tags begin with i, the parallel responses in multiplex testimony fan out from the plotline; their order can make no difference to the meaning of the document. Indeed, it possible that the statements in (12) have been reordered, since the witnesses are quoted in a different sequence than they are named in the tag; the editors found the displacement of da Maksima odd enough to warrant distinct punctuation. Ordinarily the of longtime residents goes from oldest to youngest (Kleimola 1975: 35). The use of a instead of i in multiplex testimony signals the scribes’ perception that only one speech event is taking place — a complex response to a single question. Significantly, the standard pattern of coordination, with i instead of a, is employed when the parallel answers to a single question cannot be viewed as a unified event — e.g., when the speakers are opposed parties (ASÈI 2: 523, no. 483), when one of the witnesses changes the topic (ibid. 1: 573, no. 651), or when one of the witnesses’ testimony clashes with that of the others (AFZX 98, no. 103; ASÈI 1: 466, no. 584; 468, no. 485).20 The perceived unity of multiplex testimony motivates the use of methods like that minimize the interruption of going from one witness’s response to another. In (13), the multiplex reply begins as tagged choral [1], then fans out into separate untagged reports [2–4] before merging back, again without a tag, into choral speech [5].21 (13) I [knjaz’ Vasilej sprosil Rodivonika s and [prince Vasilij]- ask- Rodionik- with tovarišcˇi: Skažite v [bož’ju pravdu, cˇ’ja to fellows-. tell-. in [god- truth]- whose-
100
R V
zemlja, na kotoroj stoim? I Rodivonik s land- on . stand-.1. and Rodionik- with tovaryšcˇi tako rekli: Skazati, gospodine, v [bož’ju fellows-. thus speak-. tell- lord- in [god- pravdu, to, gospodine, zemlja, na kotoroj stoim, truth]- lord- land- on . stand-.1. istari arxangel’skaja [Plotnicˇa sela [Bortnikovy of.old of.archangel- [Plotnicˇe village]- [Bortnikova derevni. Jaz, gospodine, Rodivonik, pomnju za 40 hamlet]- I- lord- Rodionik- recall-.1 for 40 let; a jaz, gospodine, Petruška da Ivaško years-. and I- lord- Petruška- and Ivaško- pomnim za 50 let; a jaz, gospodine, recall-.1. for 50 years-. and I- lord- Mixalko i Palka pomnim za 60 let; Mixalko- and Palka- recall-.1. for 60 years-. vse to my, gospodine, pomnim i all-. we- lord- recall-.1. and znaem za tol’ko let, cˇto [ta know-.1. for that.many- years-. that [this zemlja arxangel’skaja [Plotnicˇa sela [Bortnikovy land]- of.archangel- [Plotnicˇe village]- [Bortnikova derevni… hamlet]- And Prince Vasilij asked Rodionik and his fellow-witnesses, “Say, in God’s truth, whose land is it on which we are standing?” [1] And Rodionik and his fellow-witnesses spoke in this way: “To say, lord, in God’s truth, the land, lord, on which we are standing belongs of old to Archangel Cathedral, to hamlet Bortnikova of the village Plotnicˇe. [2] I, lord, Rodionik remember for 40 years; [3] and I, lord, Petruška and also Ivaško remember for 50 years; [4] and I, lord, Mixalko and Palka remember for 60 years; [5] we all, lord, remember and have known for that many years that this land belongs to Archangel Cathedral, to hamlet Bortnikova of the village Plotnicˇe…” [etc.] (ASÈI 3: 76, no. 50).
R F T
101
Here has the effect of coalescing the individual acts of testimony into a single continuous whole — one statement, as it were — whose parts are less demarcated and more cohesive than are ordinary statements tagged by intrusive narrative clauses. Accordingly, can be used for the testimony of agreeing witnesses in contexts where the statements of uncooperative witnesses appear as tagged (AJu 16, no. 8; ASÈI 2: 422, no. 406). It is typical for the first response in multiplex testimony, which can be choral, as in (13), or individual, to be tagged, and for some or all of the rest to be .22 (The untagged reports are thus alternatives to tagged reports with the non-sequencing conjunction a; indeed, they themselves often begin with a.) Even when the opening tag refers to more than one speaker, the first report can represent a single individual response. There is never any ambiguity in the attribution of the untagged reports; first-person singular forms and proper names signal shifts into individual or the transition from one speaker to another, while first-person plural forms pinpoint the onset of choral speech. As (13) illustrates, tends be used in the context of choral speech — reports presented as if uttered by several speakers simultaneously (cf. Tannen 1989: 113). In (13), the individual statements of how long the witnesses remember are separable from, though contiguous with, the choral testimony at the beginning and end, which identifies the owner of the property.23 In other examples, the transition from individual to choral can occur within the bounds of a single sentence; the joint information appears in a complement clause after the final speaker’s individual statement.24 Such constructions, in which one witness caps off the testimony of an entire group, plainly reveal the constructed nature of choral speech. They also serve as a strong proof of why must be approached as a discourse phenomenon rather than a syntactic construction. To view the joint complement of the parallel instances of ‘remember’ as mere clausal deletion would overlook the basic function of the construction — to present the statements as equipollent parts of a whole, as more similar to one another than to other statements in the hearing. The chief function of choral speech is to reduce the diffuseness and redundancy of multiple utterances with the same propositional content. It may sometimes also be a way of framing statements by group spokesmen; this can be inferred from cases in which the tag is collective but the itself reflects a single perspective (e.g., ASÈI 1: 538, no. 628). Both of these functions link choral speech with the pragmatic effects of discussed
102
R V
above. If unites separate acts of speaking into a continuous whole by removing narrative intrusions, choral speech goes further toward total cohesion: by eliminating linear segmentation, it coalesces the acts and eliminates their identity as discrete utterances. The combination of and choral speech allows the witnesses to be portrayed both as individuals and as members of an internally coherent group united for a single purpose in the real-world event. Given that the central purpose of multiplex testimony is to communicate shared knowledge, functions in this contextualization as a centrifugal, individualizing device and choral speech as a centripetal, communalizing one. In sum, it is clear that the use of , though unconventional, was not random; like other uncanonical reporting methods, it tends to occur when a represented speech event departs from the prototype for the text-kind. While it may be promoted by predictability of attribution, the number and distribution of the examples suggest that this is not its primary motivation. Indeed, the attribution is not externally predictable in some cases of multiplex testimony, where otherwise unusual strategies like naming the reported speaker after first-person pronouns are employed to compensate for the ambiguity. In verbal art, is often used to create vividness because it mimics direct experience (Clark and Gerrig 1990: 793; cf. Leech and Short 1981: 324) — hence the alternative term drama. In addition, like other forms of ellipsis, compels interpreters “to fill in and hence to become more involved in the storytelling” (Tannen 1983: 365–66). As an involvement strategy, it tends cross-linguistically to occur at narrative peaks (Longacre 1985: 86, 94–97; 1994: 130, 136). However, such peaks are not evident in the in trial records, nor does involvement seem to be an factor in multiplex or multimodal testimony. In the trial transcripts investigated, functions primarily as a cohesion strategy. This is most evident in the case of multiplex testimony and multimodal testimony involving document citations, in which discrete utterances are presented as parts of a single complex speech event. The omission of intervening tags creates cohesion in part because narrative clauses of this kind iconically suggest separate events. Without the intrusive tags, the juxtaposed passages of can be seen as closely related, in accordance with the “effort after meaning” that prompts interpreters to try to co-interpret adjacent segments of discourse (Brown and Yule 1983: 65–66). The use of minimizes the distances between the parallel reports and the interruption of going from one to another (“gear-shifting”, Page 1988: 33); it may thus
R F T
103
increase the pace of interpretation, which produces the iconic suggestion of fast-moving events (Leech and Short 1981: 322, 323). The same pragmatic effect appears in the use of in the multimodal testimony in boundarysetting descriptions. While in general there is little coherence between narrative and , in this context the reports are commentary on the narrated gestures and hence, in a sense, part of the same events in a different medium. This event-integrating use of recalls Longacre’s idea (1994: 132) that “variations in Q[uotation] F[ormulas] in regard to mention/non-mention of Speaker… are indexical of the intensity of participant interaction in reported dialogue” [Longacre’s emphasis]. For Longacre, continual explicit tagging that focuses on the represented speaker can be an index of verbal confrontation (130–32, 138–41). Most of the cases of in trial transcripts may be viewed conversely as non-confrontational, especially given that the represented speakers themselves tend to have little stake in the outcome of the trial. This lack of special focus on alternations between speakers may promote omission of the tag as a marker of such give-and-take.
4.4
Complementized reports: presupposition and grounding effects
Another alternative to the conventional reporting strategy in the trial dialogue is complementized indirect speech with a preposed tag. While this is the predominant mode of reporting in the verdict and certain other contextualizations (see Chapters 6–8), it is rarely used with testimony, and then only with statements of minor players; six tokens are attested, in three documents. One of the cases (16a) is tagged with the standard formula, like most testimony presented as ; the other five are introduced by the preterit of the s(”)kazati (see also 4.2), which often co-occurs with in other contextualizations. All of the examples feature the complementizer cˇ’to ‘that’ (usually spelled cˇto or što). From the onset it must be stressed that the only reliable diagnostic for in is whether person deixis is oriented to the reporter’s viewpoint rather than that of the reportee (see D. Collins 1996). In the absence of such clues, there is no secure way to distinguish it from , given that there is no backshifting of tense or automatic (matrix-dependent) change of modality. Explicit subordination, though common, is neither a necessary nor a sufficient criterion for ; both uncomplementized and complementized are
104
R V
attested. Thus one must account separately for the complementizer and the deictic orientation. In indeterminate cases, it is best to speak neither of direct nor of indirect but rather of nondirect speech, though I would hesitate to identify this as a separate reporting strategy. Nondirect speech is not a true halfway point between the two poles of the formal continuum of reporting strategies. Because it lacks the deictic features that make salient in discourse, it more closely resembles indirect speech in its pragmatic effects. From a functional standpoint, would seem to be polarized into direct and lessthan-direct types. Another point that needs to be stressed from the first is that was not used as a means of avoiding verbatim reproduction — a traditional interpretation, corollary to the reproductionist approach to (see 3.3). The statements presented as were by and large as accessible to the scribes’ observation as those presented as . Some of them are actually adjacent to direct reports; they can even represent the answers to questions in , as in (14), below. Thus it seems likely that the contextually unexpected use of where is ordinarily preferred was a way of achieving certain pragmatic effects. In fact, we see in the cases of the same correlation between nonprototypical speech events and unconventional (or less conventional) strategies as in other residual forms. For example, one of the tokens represents a statement made in the interlude between the on-site trial and a higher-court hearing (14) — a “low-keyed” transition between major episodes (cf. Diver 1969: 49, 57). During the trial, the plaintiff had used a copy of a charter to prove that his monastery owned the disputed property; he stated that the original was in the keeping of a chancery secretary, Ivan Kobjak Naumov (ASÈI 1: 539, no. 628). Copies were not considered firm evidence in trials; thus the respondents challenged the authenticity of the charter (ibid. 1: 540, no. 628). This prompted the judge, Prince Vasilij Ivanovicˇ Golenin, to visit Kobjak before submitting the trial record to the grand prince. (14) I [knjaz’ Vasilej, priexav k Moskve, vprosil [dijaka and [prince Vasilij]- come- to Moscow- ask- [clerk Kobjaka: Prislal esi ko mne spisok z Kobjak]- send- ..2 to me- copy- from [danye gramoty [Vasil’ja Borisovicˇa, a gramotu mi [donation writ]- [Vasilij Borisovicˇ]- and charter- me-
R F T
105
danuju starcy skazali u tebja, — i ty donation- elders- say-. at you-. and you-. [toe gramotu položi pered [velikim knjazem. I [that charter]- place- before [grand prince]- and Kobjak [knjazju Vasil’ju skazal, cˇto spisok z Kobjak- [prince Vasilij]- say- that copy- from danye slovo v slovo, a u spiska donation- word- into word- and at copy- podpis’ ego ruka; a gramota sja, signature- his hand- and charter- . skazyvaet, u nego uterjala. say-.3 at him- lose- And Prince Vasilij, having come to Moscow, asked the chancery secretary Kobjak, “You sent me a copy from a donation charter of Vasilij Borisovicˇ, and the elders told me that you have that donation charter. Place that charter before the grand prince.” [1] And Kobjak told Prince Vasilij [1a] that the copy from the donation [charter] was word for word, and the signature on the copy was his; [1b] but, he says, he has lost the charter (ASÈI 1: 541, no. 628).
The indirect (narrative) deictic orientation of the report in (14) [1] can be seen from the use of third-person pronouns and possessives referring to the represented speaker. Also suggestive (necessary but not sufficient) is the absence of the vocative g(o)s(podi)ne, which conventionally appears in testimony (see 3.2). The first two clauses of the report [1a] are presupposed, representing Kobjak’s confirmation of claims made in the trial hearing; the third [1b] provides new information and, in addition, features a topic change. Significantly, at this point of discontinuity the attribution is reinforced by an intercalated present-tense — a cohesion-promoting intrusion from the authorial framework, which is typical in extended passages of (e.g., (16a), below). (In all probability, the present tense was chosen here because the intercalated tag conveys a continuation rather than a separate event, so that the sequencing denoted by the preterit would be irrelevant.) The depiction of interludes between hearings, as in (14), is a contextualization that favors reporting strategies other than (see also 4.5). As “low-keyed” transitions, interludes tend to feature strategies that present
106
R V
reports in a less salient manner than is normal in the chief episode, the trial narrative. The statements that are reported in interludes are never testimony proper and often have a predetermined character. In some cases, they relate to adjournment procedures and have no bearing on the evidence; in others, they corroborate testimony made during the trial hearing, so that their content is largely predictable. Corroboration is the key factor in another case of complementized in this contextualization (15), which cannot be identified as either direct or indirect due to the absence of deixis referring to the represented speech situation. (15) A k Elizaru sud’ja poslal [Ivana Turab’eva da and to Elizar- judge- send- [Ivan Turab’ev]- and Timošku sprašivat’ o Ikonnicˇe zemle. I Elizar Timoška- ask- about [Ikonnicˇa land]- and Elizar- skazal, cˇto [ta zemlja Ikonnicˇe… po [Spasskuju dorogu say- that [that land Ikonnicˇa]- up.to [Spasskaja Road]- protivu Gorbova [zemlja mitropolicˇja Kulikovskaja, a opposite Gorbovo- [land of.metropolitan Kulikovskaja]- and kupil [Ivan Fedorovicˇ u Mixaila do moru, buy- [Ivan Fedorovicˇ]- at Mixajlo- before plague- a v knigax pisana v dannyx [Danila and in books-. write- in of.donations-. [Danilo Ivanovicˇja bojarskaja zemlja, a ne [velikogo knjazja. Ivanovicˇ]- [boyar land]- and [grand prince]- And the judge sent Ivan Turab’ev and Timoška to Elizar to ask about the Ikonnicˇa property. And Elizar said that that property Ikonnicˇa… as far as the Spasskaja Road opposite Gorbovo belongs to the metropolitan’s property of Kulikovskoe, and Ivan Fedorovicˇ bought [it] from Mixajlo before the Plague, and in the donation books of Danilo Ivanovicˇ it is recorded as boyar property and not [property] of the grand prince (AFZX 228–29, no. 259; trial transcript quoted in another transcript).
The represented speaker Elizar in (15) is a witness for the plaintiff who was unable to testify in situ due to illness (ibid.: 228). As in (14), the message of the (including the boundary-setting description, whose beginning is omitted here) is presupposed from the prior text; it repeats and corroborates
R F T
107
evidence given by the plaintiff’s other witnesses, as well as the plaintiff’s own claims about what his witnesses will say (ibid.: 227–28). Since Elizar’s remarks are expected — to some extent, already actuated — in the context, it is chiefly his willingness to support the plaintiff’s case that is being asserted in this passage. Corroboratory statements like those in (14)–(15) tend strongly to be presented as complementized in trial transcripts and other contemporary text-kinds (see Chapter 6). This accords with a cross-linguistic tendency for presupposed information to be conveyed in complementized clauses (Townsend and Bever 1977: 5). In corroborations, as opposed to ordinary testimony, more attention is given to the assertion and attribution of the utterance than to the message, which is largely predictable. Thus corroboratory reports can be backgrounded even when their tags are foregrounded, as in (14)–(15). One way of accomplishing this is the use of a complementizer, which explicitly indicates that the is secondary to the authorial narrative, since subordination is iconic of backgrounding (Chvany 1985b: 1). The predictability of corroborations obviates one of the main functions of in the transcripts — to present new evidence as prominently as possible and with the fullest possible contextualization in order to provide the intended interpreters (the judges) with complete, unprejudiced information (see 3.5). Though not inherently backgrounding, is well suited for conveying backgrounded information because it avoids the deixis switch that animates and hence actualizes passages of — the ego-hic-nunc context of the represented speech situation that is iconic of foregrounding (ibid.: 8, 15; Chvany 1990/ 1996: 288). (Traditionally has been interpreted as a shift away from the deictic pivot of the “original” utterance. This is an epiphenomenon of the reproductionist fallacy; actually it is that represents a shift of orientation.) The unobtrusiveness of in narrative allows it to be interpreted, ceteris paribus, more quickly than — an effect that accelerates the pace of the discourse (Page 1988: 33). This makes it an effective strategy for presenting statements perceived as less newsworthy or salient than the narrative norm. Two of the other cases of complementized , found in a single transcript, likewise convey statements which are at least partially corroboratory and which deviate sharply in other ways from the prototypical speech event — a factor already shown to motivate unconventional strategies. In the given trial, the respondent takes the extraordinary step of naming two of the judges as witnesses. The first indirect report (16a) represents the statement of the
108
R V
judge Zaxarij Mikulicˇ Gavinskij, who previously participated in a survey of the disputed property. The appears in a dyadic frame and eventually fades into . (16) a.
I Zaxar’ja tako rek: cˇto grani po and Zaxarij- thus speak- that boundaries-. by primetam — ix kladen’ja, a roz”ezda, landmarks-. their placement- and survey- skazyvaet, ne pomnit, po tomu li say-.3 recall-.3 up.to that- mestu ili ne po tomu li, be[z] spiska; da place- or up.to that- without copy- and tex, skazyvaet, krest’jan ne those-. say-.3 peasants-. pomnit že, [kotorye krest’jani s recall-.3 [which peasants]-. with nimi na [tom roz”ezde byli. «A [rubežnoj them- on [that survey]- be-. «and [of.boundary spisok», skazal Zaxar’ja, «u [velikogo knjazja v copy]- say- Zaxarij- «at [grand prince]- in kazne, po [kotoromu spisku roz”ezžali i treasury- by [which copy]- survey-. and grani klali. A koli [te zemli boundaries-. put- and if [those lands]-. monastyrem otdavali, ezdjacˇi, [Mixajlo Šapkin monasteries-. give-. go- [Mixajlo Šapkin]- da Golova i [gramoty svoi davali, and Golova- and [writs .]-. give-. a menja, gospodine, toldy s nimi ne bylo, — v and me- lord- then with them- be- in [tu poru menja [knjaz’ velikij poslal v [that time]- me- [prince grand]- send- to Murom pisati». Murom- write-
R F T
109
And Zaxarij spoke in this way: [1] that they had [indeed] set the boundaries, in accordance with the landmarks, [2] but, he says, he does not recall whether the survey went up to this place or not without [seeing] the document; and, he says, he does not recall which peasants were with them on that survey. [3] And, said Zaxarij, the boundary writ by which they surveyed and set the boundaries is in the grand prince’s treasury. [4] And if Mixajlo Šapkin and Golova, when departing, awarded those lands to the monasteries and gave [them] their own writs, I, lord, was not with them then; at that time the grand prince had sent me to Murom to do scribal work (ASÈI 2: 270, no. 310).
While this report features the standard tag, it begins as clear-cut , indicated by third-person verbs and pronouns coreferential with the subject of the tag clause. The first sentence of the report [1], which conveys Zaxarij’s corroboration of previous testimony (cited in the previous question, which is in ), features the highly presuppositional devices of nominalization and anaphora (ix kladen’ja, literally “of their placing”). Nominalizations are low in discourse saliency as compared with finite predicates (see Chvany 1990/ 1996: 293). The second and third sentences [2] convey new information — Zaxarij’s admission that he is unable to testify on other issues. Here the tag is reinforced, as in (14), by intercalated present-tense ’s, which reiterate the fact of reportedness without demarcating the clauses as separate utterances. The non-initial tag skazal Zaxar’ja in the fourth sentence [3] may have the same continuity-promoting function; however, it is more likely that the change in tagging strategies, from present to preterit, marks, as it were, a new paragraph — a transition from non-probative, negative remarks to positive statements that could be of use for the judgment. The switch to in the final sentences [4] marks a further change in the character of the speech event and is iconic of increasing relevance; the report now conveys a statement of personal experience bearing on one of the respondent’s claims. (The fourth sentence in (16a) can bear a different interpretation; see 4.5.) The second indirect report in the same transcript occurs in an ostensibly non-dyadic framework (16b). Here the judge Mixajlo Gnevaš Stoginin is represented testifying in response to the preceding statement; there is no report of any intervening question. When the respondent turns the tables by stating that Gnevaš had previously been appointed to judge a suit over the property, Gnevaš has to step out of his judicial role to explain himself.
110
R V
(16) b. I Galaseja tako rek: … [knjaz’ velikij, gospodine, and Galasij- thus speak- [prince grand]- lord- dal nam sud na nix [Gnevaša give- us- court- against them- [Gnevaš Stoginina, i Gnevaš, gospodine, na zemle ne Stoginin]- and Gnevaš- lord- on land- byval za [velikogo knjazja dely. I be- due.to [grand prince]- cases-. and Gnevaš pered piscy skazal, cˇto emu Gnevaš- before clerks-. say- that him- bylo na [tu zemlju exati nedosug za [gosudar’skimi . to [that land]- go- no.time due.to [of.sovereign dely [velikogo knjazja. cases]-. [grand prince]- And Galasij spoke in this way: “… The grand prince, lord, granted us a trial against them with Gnevaš Stoginin as judge; and Gnevaš, lord, has not been on the land due to the grand prince’s business.” And Gnevaš stated before the clerks that he had had no time to go to that land due to the sovereign grand prince’s business (ASÈI 2: 270, no. 310).
The underlined report is straightforward , with a third-person pronoun referring to the represented speaker. Gnevaš’s change of footing is also indexed by the use of the prepositional phrase pered piscy ‘before the clerks [the judges]’ in the tag; this suggests an unexpected participant configuration, especially since Gnevaš is otherwise one of the “clerks.” His statement, which has the illocutionary force of an excuse, resembles the first sentence of the report in (16a), since it corroborates — indeed, largely repeats — the preceding claim. It is noteworthy that Gnevaš’s entirely presupposed statement is tagged by the unconventional s(”)kazati, while Zaxarij’s in (16a), which includes new information with evidentiary value, is introduced by the standard tag. This difference seems to index their relative degrees of distance from the prototype associated with the standard strategy. The three remaining cases of complementized indirect or nondirect speech, which occur in a single transcript (17a), appear in the nondyadic contextualization of multiplex testimony, in which is more typical (see
R F T
111
4.2–4.3). Here the information that is individual — the longevity of each witness’s memory — is introduced by separate tags [1a], [2a–b]; the overlapping information — what the witnesses remember — is presented collectively as complementized indirect speech IS, once for the priests [1b] and once for the laymen [2c]. (17) a.
Ivan vsprosil [popa Semena i [popa Ivana, Ivan- ask- [priest Semen]- and [priest Ivan]- Isaka, Rodivona Cˇej to ogorod i [ta Isak- Rodion- whose- garden- and [that požnja istari, na kotoroj stoim? I [pop field]- of.old on . stand-.1. and [priest Semen i [pop Ivan skazali, cˇto Semen]- and [priest Ivan]- say-. that pomnjat to do Suzdalskogo boju za recall-.3. that- before [of.Suzdal battle]- for desjat’ let, cˇto [tot ogorod i [ta požnja ten- years-. that [that garden]- and [that field]- [troeckago dvora. A Isacˇko skazal, cˇto [of.Trinity farmyard]- and Isacˇko- say- that pomnit za 50 let, a Rodivon recall-.3 for 50- years-. and Rodion- skazal, cˇto pomnit za 60 let, cˇto [tot say- that recall-.3 for 60- years-. that [that ogorod i [ta požnja [troetckago dvora. garden]- and [that field]- [of.Trinity farmyard]- Ivan asked the priest Semen and the priest Ivan, Isak, [and] Rodion, “Who of old was the owner of the garden and that field on which we stand?” [1a] And the priest Semen and the priest Ivan said that they have known since ten years prior to the Suzdal battle [1b] that that garden and that field belong to Trinity Monastery’s farm. [2a] And Isacˇko said that he has known for 50 years [2b] and Rodion said that he has known for 60 years [2c] that that garden and that field belong to Trinity Monastery’s farm (ASÈI 1: 548, no. 635).
112
R V
The priests’ statements [1a–b] are coalesced because they are identical in content. Their testimony may have been cited separately because it was made under a peculiar kind of oath: laymen swore “before God” or “by a crosskissing oath” (see 4.2), clerics “by the priesthood” (po svjašcˇen’stvu ‘by priesthood-’). The underlined passages in (17a) are in unambiguous , with the subjects of the tag clauses coreferential with the implicit subjects of thirdperson verbs within the reports. The omission of the subject pronouns indicates relatively cohesive relations between the report and the tag clause; it may imply subordination, as does the complementizer (see Haiman and Thompson 1984: 511–12). By contrast, multiplex testimony conveyed as usually includes first-person subject pronouns. Several motivations can be posited for the use of in (17a). First, it can be seen as another means of minimizing redundancy in multiplex testimony, where the reports tend to overlap in content. Indirect speech is a relatively ˇ umakov 1975: 19, 26) because it eliminates economical reporting mode (cf. C certain elements referring to the represented speech situation that are commonly found in testimony. The use of a complementizer in (17a) obviates the disambiguating function ordinarily performed by these elements. A second possible motivation for the use of complementized in (17a) is the corroboratory and hence partially presupposed character of the witnesses’ — a feature shared by the indirect reports in (14)–(16). The statements of witnesses tend to have predictable content because of their circumscribed, normatively partisan role in the trial; indeed, the litigants are often represented anticipating what their witnesses will say, as in the given transcript. (Usually the witnesses are not mentioned at all until the judge asks the litigants for corroboration; see Kleimola 1972: 360.) In (17a), the fact that the witnesses support the respondent is contextually new information, but the actual message conveyed by their statements is presupposed (or partially so, as when they assert their longevity). The backgrounding effect of the in (17a) becomes obvious when it is compared with a passage from later in the transcript (17b) — the statement of the plaintiff’s witness, which is tagged in the conventional way and presented in a salient manner, as : (17) b. Ivan velel Stepanka postaviti, i Stepanko Ivan- bid- Stepanko- put- and Stepanko-
R F T
113
stal… I Stepanko tako rek: Jaz, stand.up- and Stepanko- thus speak- I- gospodine, [otca svoego v [tom dvore ne lord- [father .]- in [that farm]- pomnju, otec moj v tom dvore ne recall-.1 [father my]- in [that farm]- žival. live- Ivan commanded [the plaintiffs] to produce Stepanko, and Stepanko appeared… And Stepanko spoke in this way: “I, lord, do not remember my father [being] on that farm; my father did not live on that farm” (ASÈI 1: 548, no. 635).
The difference in strategies in (17a) and (17b) conveys a distinction in content and illocutionary force. In (17b), Stepanko blatantly contradicts the claims of the litigants who call him; thus his testimony carries far more weight than the corroboration offered by the other side’s witnesses in (17a), which fulfills institutional expectations. This may also explain the unusual attention to the “stage-directions” leading up to Stepanko’s testimony. The norms of medieval Russian law demanded that witnesses uphold their principals’ claims — even to defend them by force in the event of a judicial duel (Kaiser 1980: 136–37). The Muscovite Law Code (Sudebnik) of 1497, which to a large extent institutionalized already existing practices, instructed judges to decide against any litigants who were not supported by their witnesses (Grekov (ed.) 1952: 26, §51, and 86–87). The fifteenth-century Pskovian Law Code made a similar provision (Zimin (ed.) 1953: 289, 344; Martysevicˇ 1951: 126). The verdicts in several trial transcripts, including the one in question (ASÈI 1: 548, no. 635), state that uncooperative witnesses were the chief factor in a negative judgment (see also Kleimola 1975: 88). Consequently, Stepanko is acting, in effect, as a witness for the opposing side — a striking change in his socially predetermined footing. The crucial significance of the message in (17b) is indicated by a foregrounding strategy, uncomplementized ; by contrast, the presupposed status of the message in (17a) favors backgrounding of the report, though not of its attribution. In sum, all of the examples of in the trial dialogue convey information that is wholly or partially presupposed and corroboratory in character. This
114
R V
factor favors the choice of explicit subordination in the form of the complementizer cˇ’to, which gives the attribution in the narrative framework greater salience than the report — a striking effect in a “dialogue discourse” (Longacre 1983: 44), where the most important information tends to be within . Presupposition of the message may also favor , which avoids the discourse-slowing deixis shifts of . Slowing the pace of interpretation, as in , makes the report contextually salient — an effect that can be likened to the modern Russian convention of putting spaces between each letter of a word for emphasis. By contrast, does not place the report in special prominence; it offers a quick account, much like a summary, of a message that has, in effect, already been processed. In Sections 4.1–4.3, it was noted that, as predicted by the method of residual forms, unconventional strategies tend to index statements that depart from the prototype for the represented speech event. This tendency also applies to the cases of discussed above; it may be a factor conspiring with presupposition to promote complementized . The differences between ordinary testimony and the multiplex statements exemplified in (17a) were discussed in 4.2–4.3. In (16a, b), the indirect reports reflect a drastic change in footing; it was not a normal part of the judges’ role to bear witness to a litigant’s claims or to account for their own conduct. The use of complementized in (14)–(15) may also be influenced by the unusual nature of the represented speech events as compared with testimony during the trial hearing. The given speech events have a different participation framework than the trial hearing, and not only because of the change in venues (cf. Goffman 1981: 128). The interested participants — litigants and witnesses — are absent or inactive and thus unable to respond to the new evidence. In (14), the secretary Kob’jak does not fall into one of the usual categories of witnesses — longtime residents (znaxari, starožil’ci), or reputable spokesmen of local opinion (dobrye muži). If deviation from the prototypical speech event does play a role in the choice of complementized , it raises two potential problems. First, why should the statements of witnesses who fulfill social expectations in (17a) be presented as when that of the uncooperative witness in (17b), whose footing in the trial changes drastically, is presented in the conventional manner? Second, why should , a strategy typically used in backgrounding, be employed to index deviant speech events, which one one think would be a foregrounding function? Recall Silverstein’s remark (1985: 142) that “there
R F T
115
are normal, ‘appropriate’, presupposed contexts for the use of particular forms — unmarked usage… — and highly marked contexts, unexpected for the occurrence of such forms, although occur they do with a ‘foregrounded’… effect” (see 1.5). The solution to these problems lies in the fact that the conventional reporting method, , is privileged as a foregrounding device by virtue of its formal properties. In accordance with institutional needs, the writers preferred to convey the prototypical speech event, probative testimony, in this highly salient manner. When other strategies are employed, no matter how unexpected they may be, they cannot exceed the foregrounding power of or, in general, assume functions that are incompatible with their properties. Rather, they establish contrasts with the other reports in their larger context. The trial dialogue exhibits a high degree of parallelism (see Chapter 2); thus, in terms of global coherence, the statements of the respondent’s witnesses in (17a) are naturally paired with those of the plaintiff’s witness in (17b). Accordingly, the use of the unobtrusive form of for the run-of-the-mill testimony in (17a) serves to give greater prominence to the unusual testimony presented as in (17b), which could be characterized as “hyper-foregrounded” in the macrostructure. (Cf. the contrast between free and tagged in some multiplex testimony, 4.3.) In general, it is a mistake to assume that unexpected usages (“marked” ones, to recall Silverstein’s somewhat problematic term) are inherently foregrounded. In explaining unconventional usages, one must never lose sight of the properties of the deviant strategies or their relation to other reports in the context. Conceivably, the appearance of in the trial dialogue could be interpreted differently — as the beginning of a long-term stylistic change, like the cases of the s(”)kazati where the standard tag verb recˇi is expected (see 4.2). There is, in fact, a tendency for to replace as the preferred method for reporting testimony in transcripts from the 1590s onwards. However, it is uncertain that the indirect reports in (14)–(17a) are connected with this development; transcripts from the intervening years do not provide clear evidence of expanding at the expense of . In any case, it would be a mistake to assume that the new preference for was an unmotivated stylistic “drift”, since developments of that kind are strongly influenced by pragmatic factors. The trial transcripts of the late sixteenth and early seventeenth centuries in which predominates reflect marked changes in lawsuits as a social
116
R V
institution. They are compilative dossiers (sudnye dela or “trial cases”) in which abridged reports, often conveying corroborations, are appended to copies or summaries of documents. Trial records of this kind reflect the diminishing status of spoken testimony in the legal process, as lawsuits came to be fought primarily through documents such as žalobnicy (“complaints”, petitions that replace the plaintiffs’ opening statements) and ssylocˇnye pamjati (“referral memoranda”, briefs that provide judges with the names of witnesses). The use of in these dossiers was evidently motivated, in large part, by the some of the factors that prompted it in the trial transcripts examined here — backgrounding of presupposed corroborations and other dialogue that played a minor role in the judges’ decisions.25 There was less call for such backgrounding in trial transcripts of the fifteenth and early sixteenth centuries, in which testimony in spoken confrontations was still the principal form of evidence (the prototypical speech event). In these early documents, in the trial dialogue occurs, just as in the later dossiers, when the writers categorized reports as less probative for the verdicts than statements presented as .
4.5
Fused reported speech and streamlining effects
In the cases of and examined above, the report clauses retain the syntactic features of independent clauses, regardless of whether they are explicitly complementized, since the grammatical relations within the reports are not controlled by elements in the tag clauses. Indeed, the tags themselves are omissible, as seen in . The deictic orientation in , while congruent with that of the tag clause (narrative), is not grammatically dependent on it, while the optional use of a complementizer is “a coding acknowledgment that the two clauses are semantically still independent of each other, at least to some extent” (Givón 1980: 371; see also 1989: 110; Thompson 1986: 520). By contrast, in the strategy that I will term fused , the report clause lacks some or all of the properties of independent clauses (see Givón 1980, 1989: 108–12; Haiman and Thompson 1984: 511; Li 1986: 36–37). In , fused may be divided into fused topic statements, or syntheses of content (cf. Potebnja 1899/1958–65, 1–2: 296), and fused reports of messages. In the latter type, which is the only one attested in the narrative framework, the notional subject of the is treated as the direct object of the and
R F T
117
appears in the accusative case, or genitive if the is negated (as is expected in ). The notional verb — either a null copula or a third-person nonpast existential verb — has no surface realization; its mood is predictably indicative, and its tense coreferential with that of the . The predicate of the report (if any) can be a noun phrase or adjective, whose case agrees with that of the direct object of the (the notional subject); or else it can be a nonagreeing constituent — a prepositional phrase, possessive noun phrase, or locative adverb. If the fused report contains pronouns or possessives referring to the represented speaker (the subject of the main clause), they are reflexive even when they belong to the noun phrase that is the underlying subject. This form of small-clause structure is attested not only with ’s but also with verbs of perception and cognition. In the trial transcripts investigated, fused is most common in contextualizations other than testimony — in additional layers of within and in reports of the judges’ rationales in verdicts (see 7.5, 8.3.2). There are only eight certain cases in the authorial narrative of the trial record; in all these examples, the is s(”)kazati/s(”)kazyvati ‘say, tell-/’ (see also 4.2). There is also a possible case involving the present tense of kazati ‘say, tell’ (ASÈI 1: 492–93, no. 595). This example, which occurs in a boundarymarking description, can be interpreted either as fused or as non-integrated with an intercalated . It is indeterminate because the notional subject is an inanimate masculine noun, for which nominative and accusative are homophonous, while the notional predicate is a non-agreeing possessive phrase. In the clear-cut examples, the perfective s(”)kazati is attested in the preterit tense, while the imperfective s(”)kazyvati is found in both the preterit and the historical present. The variation in tense and aspect is connected here, as elsewhere, with grounding effects. The perfective preterit is used in narrative to encode events (including speech events) that are plot-advancing and salient (“figure”, Wallace 1982: 208–9, 212, 215), whereas events encoded by the imperfective tend to be backgrounded. As predicted by the method of residual forms, none of the cases of fused occurs in prototypical dyadic testimony; that is, none represents an informative (evidentiary) response to a question in the immediately preceding context. One example appears in the incipit of a transcript (18); here the report is mere background orientation explaining why one of the previously mentioned litigants is absent. (As there is no equivalent of fused after bleached ’s in English, I will translate the examples as .)
118
R V
(18) tjagalisja [mitropolicˇi posel’skie [Jur’evskix sue-. [metropolitan- stewards]-. [of.Jur’evo sel… s Yvanom, da s Fedorom, da s villages]-. with Ivan- and with Fedor- and with Kostjantinom s [Yvanovymi det’mi Kucˇeckogo, Konstantin- with [Ivan- children]-. Kucˇeckij- a Fedor da Kostjantin otvecˇivali i v [brata and Fedor- and Konstantin- respond-. in [brother svoego [mesto Ivanovo, a [brata svoego .]- [place Ivan-]- and [brother . Ivana skazali v Novegorode v Velikom na pomest’i. Ivan]- say-. in Novgorod- in Great- on estate- The metropolitan’s stewards of the villages of Jur’evo… sued Ivan and Fedor and Konstantin, the sons of Ivan Kucˇeckij; and Fedor and Konstantin responded in lieu of their brother Ivan, and they said their brother Ivan was in Novgorod the Great on [his/their] estate (AFZX 138, no. 157).
There are two clear signals that the report in (18) is grammatically integrated. First, the notional subject (“brat… Ivana”) is accusative; second, the possessive adjective that modifies it (“svoego”) is reflexive because it is coreferential with the subject of the main verb (the ). In , as in , reflexive elements are usually restricted to the predicate; the report in (18) must represent an utterance with a non-reflexive possessive in its subject phrase. The word order in the fused report in (18) reflects the ordinary information structure in , just as in independent sentences. Whereas the right-hand constituent (the locational prepositional phrase) conveys new information, the left-hand one (the notional subject brata svoego Ivana) is presupposed from the immediately preceding context; it is thematized in a particularly salient manner by being preposed to the , so that it becomes the topic of the main (and only real) clause. This possibility is not available with less integrated forms of , with the exception of reports with intercalated tags. If the report in (18) forms a prelude, two other cases of fused (both in a single transcript) appear in the interlude between two acts — the on-site trial and a higher-court referral hearing (19). As mentioned in Section 4.4, such “low-keyed” episodes favor reporting strategies that are less intrusive
R F T
119
in the narrative than . The statements conveyed as fused in interludes are not prototypical testimony, since they provide no information about the ownership of the disputed property; they are background material for the subsequent actions of the judge, which are part of a procedure to ensure that neither litigant will profit from the land until the case is resolved after the adjournment.26 (19) A Fed’ko, da Andrejko, da Jakuš skazali and Fed’ko- and Andrejko- and Jakuš- say-. [svoego košen’ja sena dva stogu da [. mowing]- hay- two- stacks-. and vzorod, a [mitropolicˇi xristiane [Karinskogo rick- and [metropolitan- peasants]-. [Karinskoe sela [Senka dvornik da [Gracˇko Paxomov village]- [Senka tenant]- and [Gracˇko Paxomov]- skazali [svoego košen’ja sena [tri stogi. say-. [. mowing]- hay- [three stacks]-. And Fedko and Andrejko and Jakuš said the hay from their mowing [came to] two stacks and a rick, and the metropolitan’s peasants of the village of Karinskoe, the tenant Senka and Gracˇko Paxomov, said the hay from their mowing [came to] three stacks (AFZX 110, no. 117).
The only elements in (19) that point unambiguously to fused are the possessives, which, as in (18), are reflexive because they corefer with the subjects of the ’s. The notional subject is postposed, in accordance with the usual word order in existential-quantifying sentences; the left-most elements are presupposed from the dialogue of the earlier hearing. Another contextualization for fused is in boundary-setting descriptions. In addition to the ambiguous example with kazati mentioned above, there are four tokens, two of them in variant transcripts from a single trial (20). As discussed in Sections 4.1–4.3, much of the emphasis in such descriptions is on actions rather than on statements, which can be treated as background commentary.27 (20) I povel Fedosko poperek [dorogi bolšie Ozereckie and lead- Fedosko- along [road big Ozereckaja]-
120
R V
k ozeru, cˇto ezdjat xrest’ja[nja] k Spasu k to lake- ride-.3. peasants-. to Savior- to cerkvi, i privel k ozeru da doro[go]ju podle church- and lead.to- to lake- and road- along ozera novo prosecˇeno. A skazyval na [toj doroge lake- newly cut- and say- on [that road]- osinu, a na nej gran’, i on osiny ne aspen- and on her- marker- and he- aspen- ukazal, a grani net. I privel ko show- and marker- be-. and lead- to pnju… stump- And Fedosko [the respondent] led along the big Ozereckaja road to the lake, which the peasants use to go to the Church of the Savior, and he led to the lake and by the road alongside the lake, [where there was] a new cutting. And he said there was an aspen on that road and a marker on [that aspen], and he did not point out the aspen, and there is no marker. And he led to a stump… (ASÈI 2: 455, no. 419; similarly 453, no. 418).
In (20), the use of an imperfective instead of the perfective, the narrative norm, may be explained by the collateral nature of the represented statement. The imperfective preterit, especially of derivationally iterative verbs like s(”)kazyvati, could express noniterative pluperfect meaning in .28 This nuance was probably contextual rather than inherent. The imperfective preterit generally conveyed actions without reference to sequentiality; thus it was appropriate for pluperfect-like “flashbacks” that provided the setting for subsequent events or non-events. In (20), the fused report that is backgrounded in this way is preliminary to the mention of what the respondent does not show, which was evidently an important factor in the verdict and is conveyed by a negated perfective. The final example of fused in the authorial narrative of a trial record represents a litigant identifying witnesses to speak on his behalf (21) — a contextualization in which is also attested (see Section 4.3). This is not testimony proper but a preliminary to such testimony; it introduces new participants in a prepatterned manner.29
R F T
121
(21) I sud’i vsprosili Dyma: A u tebja and judges-. ask-. Dym- and at you-. komu vestno, cˇto [te lugi who- be.known that [those meadows]-. tjanut k selu k Mencˇakovu? I belong-.3. to village- to Mencˇakovo- and Dym skazal [svoix starožilcev, mencˇakovskix Dym- say- [. old.residents of.Mencˇakovo xrest’jan [Trofimka Fomina… na tex sja, peasants]-. [Trofimko Fomin]- to those-. gospodine, xrest’jan šljusja, cˇto [te lord- peasants-. refer-.1 that [those lugi tjanut k [selu Mencˇakovu meadows]-. belong-.3. to [village Mencˇakovo]- And the judges asked Dym, “Whom do you have that knows that those meadows belong to the village of Mencˇakovo? And Dym said his witnesses were the Mencˇakovo peasants Trofimko Fomin [et al.]: “I call those peasants, lord, [as witnesses] that those meadows belong to the village of Mencˇakovo” (ASÈI 2: 533, no. 492).
This report differs from (18)–(20) in that an illocutionary meaning other than declaration can be ascribed to the . By identifying individuals who he believes are willing to support his claims, the speaker is attempting to make them legally accountable witnesses, who could even be called on to fight a duel on his behalf. This speech act is the equivalent of English name in the sense of “I name X (to be) the executor of my will” — name2 in Wierzbicka’s analysis (1987: 365): “the speaker utters a name to cause people to associate a particular predication with that particular person (at a certain time).” Significantly, there is a switch from fused to at the end of this speech act, i.e., at a seam in the represented utterance: the litigant begins a new speech act of assertion that could provide probative information. Structurally this continuation does not lend itself to fused , since its predicate consists of a verb other than ‘be’. The motivation for the use of untagged is that a second could have falsely implied a separate turn at speaking. In sum, none of the fused reports represent speech events that can be interpreted as salient or probative testimony. One conveys running commentary
122
R V
(20), and the others setting information ((18)–(19), (21)). Two ((18)–(19)) appear in “low-keyed” transitional episodes (cf. Diver 1969) and reflect unsequenced, non-testimonial phases of the trial. The messages themselves carry a relatively low information load; the boundary-setting reports are particularly terse as compared with similar passages in other transcripts. In (21), the message is prepatterned and thus partially presupposed. While fused preserves the central propositional content, it renders statements less diffusely than either and by omitting inferable information about verbal categories. This terseness makes it suitable for presenting brief reports, especially ones that carry a low information load and need not be presented in a prominent, elaborated manner, with everything made explicit for the interpreter. Fused is not suitable for presenting structurally more elaborated statements with multiple clauses, as this would require sustained inference due to the reduced cohesion of stranded constituents. Fused can thus have a streamlining effect, which makes it suitable for conveying presupposed or otherwise backgrounded reports. The small-clause structure has the textual effect of eliminating the demarcation between report and authorial context (in the present instance, narrative), though fused does not go as far in this respect as (see 4.7), since it still marks reportedness in an explicit manner. The lack of “gear-shifting” (Page 1988: 33) between and narrative increases the pace of interpretation. This accelerando or glossing-over effect tends, in general, to counteract discourse salience. At this point a seeming contradiction emerges: if fused reports are incorporated as constituents of the main clause, they must reflect the salience characteristic of main clauses. This, in fact, is true not of the report as such but of the separate constituents of the report, as is illustrated clearly in (18) by the topicalization of the notional subject and the assertion of the notional predicate. While the separate parts of the are thus amenable to salience effects, the report as a whole presents neither a figure nor a ground that can be treated separately from the narrative tag clause. Indeed, the capacity to single out particular items within the reported information is clearly a special property of the strategy. It is noteworthy in this connection that fused constructions have been treated specifically as salience devices. Nichols and Schallert (1983) view the accusative with null ‘be’ as a subtype, derived by deletion of an underlying participle, of the accusativus cum participio (an ecclesiastical-register construction unattested in trial transcripts). They argue that writers chose these
R F T
123
constructions because they allowed elements of notionally subordinate clauses to come under the domain of some pragmatic factor that operated only in main clauses (ibid.: 240–43). While Nichols and Schallert do not identify this factor, they propose that raising-derived participial constructions had a discourse function that they call plot salience — conveying “information important to the development of the plot” (ibid.: 239–40). Participial clauses can have plot salience where less integrated complement clauses cannot because the “pared-down character” of the participials allows each clause constituent a measure of prominence that would not be available in a more elaborate clause” (ibid.: 240; see ibid.: 242–43, Nichols 1980: 428). However, it is somewhat problematic, in my opinion, to view the text function of fused reports and other small-clause constructions as salience, at least as that term is usually understood — as a grounding effect. Causal importance cannot always be equated with actual prominence in the discourse or in the plot at the given moment of narrative time (cf. Chvany 1985b: 5, 11; 1990/1996: 288). The constructions to which Nichols and Schallert ascribe plot salience are of secondary importance in their context; they comment on or set the stage for other, genuinely highlighted, plotadvancing events.30 While the matrix verb itself may denote a highlighted event in the narrative structure, this is an effect of factors other than the grammatical integration of the embedded clause (“raising”, in Nichols and Schallert’s approach). The assumption that integrated constructions have plot salience by virtue of their main-clause status prompts Nichols and Schallert to divorce them from participial relative clauses, which do not have that function (1983: 240). On the whole, it would be preferable to assign the accusativus cum participio the typical function of participles in (Barnet 1965: 41–42) — indicating the secondary status of represented actions (see 4.1). In general, cases of “deep embedding”, including participles and embedded predicate nominals, are icons of backgrounding, relatively low in discourse saliency (Chvany 1985b: 14–15, 1990/1996: 292–93). The cases of fused in the authorial narrative of trial records are likewise used to convey contextually secondary events. The examples all occur in contexts in which the fact of testifying is not emphasized, and the reports serve primarily as explanations for other events. This is also true of the fused reports in the verdict (see 8.3.2). Other forms of — in particular, — are used to present information viewed as potentially probative. Thus, considering the contextualization of the evidence, salience (however it
124
R V
is understood) does not seem to be a plausible explanation for the accusative with null ‘be’ seen in fused . In fact, the ancillary character of the statements presented as fused lies at the very heart of the strategy’s function. Dyadic testimony with minimal tag clauses is not the primary form of narrative in the contexts in which fused reports are found. The is not the center of attention here, as it is elsewhere in the trial record; the emphasis is on the participants themselves or on their nonverbal actions. There is no reason to dwell on the per se or present it in a way that occasions delay. Fused is well motivated in such contexts, since it impedes the narrative less than do the more diffuse nonintegrated strategies such as and , which are used to encode more crucial information. As Nichols (1980: 435) observes, the effect of the null ‘be’ construction is “to minimize surface subordination and evidence of complementation.” There is no independent non-narrative clause or shift of perspective to slow down the transition between the clauses that are foregrounded in the narrative. This accelerando effect is furthered by the reduced nature of fused reports — the “pared-down character” to which Nichols and Schallert (1983) allude. As a rule, the fewer the words, the faster the tempo. Thus the pragmatic effect of fusing the with the main clause is to enhance the continuity, pacing, and economy of narrative passages that would be impeded by grammatically more independent and hence more interruptive strategies. The gain in pacing that fused speech promotes makes it particularly suitable with reports in transitional episodes and obiter dicta in the middle of coherent episodes of other types of action.
4.6
Narrative reports of speech acts
While fused reports (see Section 4.5) are incorporated grammatically into the narrative to a greater degree than and , they nevertheless retain, if not their identity as independent clauses, at least their propositional structure and perhaps also some elements of their original surface structure. A greater degree of integration can be found in narrative reports of speech acts (’s), to use Leech and Short’s term (1981: 323–24). In this strategy, is, as it were, “submerged” (Page 1988: 34) into the narrative by means of speech-act verbs (’s); verbal acts are recounted in much the same way as
R F T
125
nonverbal ones. The primary function of ’s is not to convey messages but rather to assert that the given types of illocutionary acts or perlocutionary effects occurred.31 However, some ’s can, in fact, convey message information; these “summaries” must be distinguished from “omissions”, the non-reportive use of linguistic-action verbs or nouns to denote “speaking or writing treated much like any other physical or mental event” (Thompson 1996: 517–18). Within the of trial transcripts, the participants are depicted using an extensive inventory of lexically specific ’s to denote claims, accusations, denials, oaths, promises, requests, refusals, lies, and other kinds of speech acts (see 7.7). Various ’s are also found in verdicts (see 8.3.5). Thus it is curious, at first glance, that ’s are scantily attested in the authorial narrative of the trial record (twelve tokens in all). For example, there are only three cases of the volitional speech-act verb () veleˇti ‘order’, even though this strategy is often used for the judges’ procedural commands (see 5.7). Its relative scarcity in the trial narrative is certainly due in part to the limitations of the litigants’ role in the real-world context; however, a more important disincentive would seem to be the covert style of narration, which tends to avoid evaluations (see below). As with the reductive form of fused and, to some extent, the other nondirect strategies, the chief motivations for choosing ’s with veleˇti rather than the standard strategy were pacing and relevance. Both these factors are at work in one narrative report of a litigant’s speech act, which occurs in a string of narrative clauses recounting adjournment procedures outside the give-and-take of the trial dialogue (22): (22) I sud’i dali srok knjazja postavit’. A and judges-. give-. term- prince- put.up- and [istec Okul, sena pudernja [sic] velel sud’jam [plaintiff Okul]- hay- pull.out- bid- judges-. zapecˇatat’… I [sami sud’i na [tretej srok… seal- and [self judges]-. on [third term]- stali… a [drugo[go] stog[a] na požne net; a stand.up-. and [second stack]- on field- be- and Nor[a] že stal že… Nora- stand.up-
126
R V
And the judges gave [Nora] an adjournment to produce the prince. And the plaintiff Okul, having pulled out some hay, bid the judges put a seal [on it]… And, after the third adjournment, the judges themselves appeared…, but the second haystack was not on the stubblefield; and Nora also appeared… (ASÈI 2: 496–97, no. 458).
In (22) the plaintiff is concerned with ensuring that some haystacks on the disputed property are not used during an adjournment (which stretches out to two further adjournments, as recounted in the omitted portions of the above passage). His speech act, conveyed in the underlined report, cannot be construed as testimony, and its wording is irrelevant to the judge’s eventual decision. However, it is reportable as a fact (a narrated action) because it provides a background for a subsequent event — the disappearance of one of the haystacks, presumably at the hands of the respondents (who are ultimately fined for an equivalent amount of money). If the standard strategy of had been used instead of the , it would have provide superfluous detail. Moreover, given that the report appears in the midst of a narrative of nonverbal actions, the use of would have had an unduly interruptive, highlighting effect without a commensurate information load. A clear illustration of how ’s contrast functionally with may be seen in (23), which depicts a steward ordering his witnesses to repeat the boundary-setting procedure reported in a document that he has just given into evidence. This is a non-dyadic context, in that the expected response to such a command is nonverbal action: (23) I Ivan… [xupanskomu poselskomu starcu Savel’ju tak and Ivan- [of.Xupanskaja steward elder Savelij]- thus rek: Veli ž peredo mnoju po [toj roz”ezšcˇoj speak- bid- before me- by [that survey gramote [Xupanskuju zemlju… ot [velikogo knjazja zemli… writ]- [Xupanskaja land]- from [grand prince]- land- otvesti; a ty, [Okul sotnik i delimit- and you-. [Okul hundredman]- and [Vasko Blekloj i s [svoimi znaxori… [Vasko Blekloj]- and with [. witnesses]-. pojdite za [manastyrskimi rozvodšcˇiki… I [starec go-. behind [of.monastery surveyors]-. and [elder
R F T
127
Savelej velel… otvesti. I [Okul sotnik i Savelij]- bid- delimit- and [Okul hundredman]- and [Vasko Blekloj… tako rkli: Uže esmja [Vasko Blekloj]- thus speak-. already -.1. pered toboju… [velikogo knjazja before you-. [grand prince]- zemlju… ot [monastyrskie zemli… otveli… a land- from [of.monastery land]- delimit-. and za [manastyrskimi rozvodšcˇiki po ix [roz”ezžoj behind [of.monastery surveyors]-. by their [survey gramote na rozvod ne idem… writ]- to survey- go-.1.] And Ivan [the judge]… spoke in this manner to the monk Savelij, the Xupanskaja steward: “Bid [your witnesses] delimit the Xupanskaja property… from the grand prince’s property… in accordance with that survey document; and you, Okul the hundredman and Vasko Blekloj, follow the monastery surveyors with your witnesses…” [1] And the monk Savelij bid [his witnesses]… delimit [the properties]. [2] And Okul the hundredman and Vasko Blekloj… spoke in this manner: “We have already delimited… the grand prince’s property… from the monastery’s property… before you, and we are not going on the survey following the monastery’s surveyors in accordance with their survey document…” (ASÈI 1: 559, no. 642).
Under ordinary circumstances, procedural commands by litigants tended to be omitted as irrelevant; such commands did not contribute directly to the pool of evidence, and compliance on all sides was expected. The steward Savelij’s command in (23) [1] is highly reportable in one particular respect: it contrasts with the stonewalling of his opponents. However, it is the fact of the speech act that matters for the verdict; the precise content and form are irrelevant. This set of circumstances promotes reporting in a condensed manner, with omission of superfluous nonnarrative details. By contrast, Okul and Vas’ko’s refusal [2], cited only in part, merits the maximally detailed form of because it contains unpredictable content and is probative for the verdict. (In medieval Russian law, uncooperative behavior typically raised suspicions of guilt. The judge in the case from which (23) is drawn ultimately
128
R V
decided against the stonewallers.) Rendering the refusal as a would have entailed either drastic abridgment, with loss of potentially revealing detail, or else the creation of a highly complex narrative passage with more embeddings than is characteristic for the speech genre. A further illustration of the functional contrast between and ’s is found in (24). This passage depicts a nonprototypical speech event — a speech act performed in a non-dyadic, non-testimonial context after a break in the hearing. (24) I igumen tak rek: Verju, gospodine, and abbot- thus speak- believe-.1 lord- Dimitriju, veli emu krest pocelovat’, cˇto [ta Dmitrij- bid- him- cross- kiss- that [that zemlja kuplja otca evo, a ne bogojavlenskaja land]- purchase- father- his and of.Theophany- ot [Tjuševki debr’e po [Koleminskuju dorogu. I from [Tjuševka thicket]- up.to [Koleminskaja road]- and Dmitrej tak rek: Velju, gospodine, [svoemu Dmitrij- thus speak- bid-.1 lord- [. cˇeloveku krest pocelovat’. I kak na sroce k celovan’ju, man]- cross- kiss- and when on term- to kissing- i Dmitrej [cˇeloveku svoemu celovat’ ne velel. and Dmitrij- [man .]- kiss- bid- [1] And the abbot spoke in this way: “I believe Dmitrij, lord, bid him kiss the cross [and swear] that that land is a purchase of his father’s, and not [property of] Theophany [Monastery] from the Tjuševka thicket up to the Koleminskaja Road.” [2] And Dmitrij spoke in this way: “I do, lord, bid my servant to kiss the cross.” [3] And when [they] were back for the cross-kissing, Dmitrij did not bid his man to kiss [the cross] (Morozov 1988: 301, no. 2)
In negated commands typically denote acts of forbidding (“order not to”); thus the in (24) [3] probably represents a refusal rather than a genuine non-event. Dmitrij’s unexpected refusal to let his witnesses kiss the cross (i.e., to swear an oath) has considerable significance for the verdict, since it is tantamount to surrender. However, a summary of his speech act is sufficient, since it is his nonfeasance — the fact of his refusal rather than its
R F T
129
wording — that is important for the outcome of the trial. The use of would introduce elements of the speech situation with no relevance for the verdict, give the reported information a degree of prominence incommensurate with its functional load, and add diffuseness without a clear function in the discourse. By contrast, the choice of for the preceding exchange [1–2] is well motivated for two reasons. First, the two litigants’ statements occur in a string of dyadic exchanges; second, they serve as testimony (or testimony-like information) in that they convey the litigants’ joint agreement to settle the suit by oath. The juxtaposition of with has the effect of contrasting Dmitrij’s words (testimony) [2] with his deeds (a speech act viewed less as a message than as a form of narratable action) [3]. Two further cases of ’s conveying nonfeasance where verbal action is expected appear in what is, to my knowledge, a unique ecclesiastical lawsuit over the jurisdiction of a monastery (25).32 The judge is the metropolitan; the litigants are proxies for the prince of Belozersk and the archbishop of Rostov. (25) I [gospodin mitropolit vsprosil Feodora: Starym and [lord metropolitan]- ask- Feodor- old-. pak [knjazem rostovskim i belozerskim i [princes of.Rostov]-. and of.Belozero-. and [bojarom starym est’ li komu vedomo, [kotorye [boyars old]-. ..3 who- be.known [which prežnie arxiepiskopy v [Kirilov monastyr’ previous archbishops]-. to [Cyril- Monastery]- [pristavov svoix slali, a igumena i [bailiffs .]-. send-. and abbot- and brat’ju sudili, i desjatilnicy ix v”ezžali brethren- judge-. and tithemen-. their come-. i [pošliny svoi imali? I [Feodor and [taxes .]-. take-. and [Feodor Poluxanov i na to pered gospodinom [dovoda Poluxanov]- also to that- before lord- [argument nikakova ne ucˇinil že not.any]- make-
130
R V
And the lord metropolitan asked Feodor [the archbishop’s proxy], “Again, are there any old princes of Rostov and Belozero and old boyars to whom it is known whether any previous archbishops sent their bailiffs to Cyril Monastery and judged the abbot and brothers, and [whether] their tithemen came and [whether] they took their taxes?” And Feodor Poluxanov did not make any argument against that either before the lord (ASÈI 2: 281, no. 315).
A similar instance occurs earlier in the same transcript (ibid.). The in these examples is a pure classification of function with no trace of the message; indeed, it leaves it unclear whether Feodor made any reply at all or simply remained silent. As in (24), the content of Feodor’s statement (if any) has no relevance for the judge’s decision; the mere fact of his verbal nonfeasance demolishes his case. The phrasal verb ucˇiniti dovod” (approximately, ‘advance an argument’) points to what Feodor should have done; like any negated expression in narrative, it is evaluative, because it creates the expectation that the opposite might have been expected (Labov 1972: 380–81; Grimes 1975: 64–65). As evaluative devices, negated expressions are rare in the trial record — predictably, given the covert style of narration; they occur mostly in accounts of no-shows.33 Given its ecclesiastical origin, the transcript from which (25) is drawn could conceivably reflect an independent norm that was more tolerant of evaluations than the usual chancery practice seen elsewhere in the corpus. In some other cases, ’s seem to be used merely as reductive devices. This evidently applies to three cases involving v”imenovati ‘name’, which represent litigants’ naming their witnesses — a contextualization in which other nonstandard strategies are also found (see 4.3 and 4.5).34 Three further examples involve the pos”latisja ‘refer to’, which, in its judicial usage, denotes a litigant’s reference to a document or else his authorization to send for an absent witness (26a).35 (26) a.
I Vanja tak rek: na išcˇeiny, and Vanja- thus speak- on plaintiffs-.. gospodine, starožil’ci šljus’… I sud’ja lord- witnesses-. refer-.1 and judge- vsprosil Frola i Koposa: vy na ask- Frol- and Kopos- you-. to
R F T
131
[otvetšikovy starožil’ci i na knigi [respondent- witnesses]-. and to books-. [Mixajlova pisma Volynskogo šlete [Mixajlo- writing]- Volynskij- refer-.2. li sja? I Frol i Kopos tak rkli: and Frol- and Kopos- thus speak-. šlemsja, gospodine. I poslalis’ [oboi refer-.1. lord- and refer-. [both istci na [oboix starožil’cev. litigants]-. to [both witnesses]-. And Vanja spoke in this way: “I refer, lord, to the plaintiffs’ witnesses…” And the judge asked Frol and Kopos [the plaintiffs], “Do you refer to the respondent’s witnesses and to the books written by Mixajlo Volynskij?” And Frol and Kopos spoke in this way: “We do [so] refer, lord.” [1] And both parties referred to each other’s witnesses (AFZX 223, no. 258).
Here the [1] is used to summarize two speech acts that coincide in force and are non-confrontational. The diffuse detail offered by would be superfluous in this context, given that the underlined information is presupposed from the preceding testimony. The in (26a) reflects a well-attested performative judicial formula (26b), which is represented elliptically in the plaintiff’s represented utterance: (26) b. šljusja/šlem”sja na X (cˇto…) refer-..1./ to X- (that I/we refer to X (as evidence that…)
Given their structural similarity, it is tempting to treat ’s with pos”latisja as free indirect reports of the formula in (26b); however, this proves not to be feasible. The preterit of pos”latisja in (26a) conveys an event that is temporally ordered in the narrative; an indirect report would need a nonsequential present-tense šljut’sja, with time-coreference, or, conceivably, an imperfective preterit s”lalisja (a treatment I have not seen in ). This distinguishes ’s in from comparable reports in backshifting languages like English, where sentences like John named Mary as a witness that… can be interpreted equally as ’s and as free indirect reports of sentences like I name Mary as a witness that…
132
R V
Even in their diversity, the narrative reports of the litigants’ speech acts share certain pragmatic functions. Like the other nondirect reporting strategies, ’s occur when the form of the message or the message itself has a low informational value for the judge’s decision — that is, when it has little weight for the primary function of the text-kind. The statements presented as ’s are not testimony; they relate either to trial procedure or to non-events (anti-testimony, as it were). Like fused , ’s can be viewed in part as reductive devices that free non-evidentiary segments of narrative from being encumbered by the superfluous details, perspective changes, and diffuseness characteristic of the standard strategy, . In other words, ’s are “useful for summarizing relatively unimportant stretches of conversation” (Leech and Short 1981: 324). At the same time, because of their descriptive, summarizing character, ’s function as a means of embedding evaluation into the narrative — a way of commenting on events without breaking the flow of narrative clauses (see Labov 1972: 371). This probably explains why ’s are so rare in the authorial narrative of trial records. If the scribes had strayed from the (admittedly partial) objectivity of their reporting methods by including their own evaluations of the events and statements, they would have been acting as judges themselves; this would have undermined the central purposes of the text-kind — to refresh the memory of the trial judges and to provide information to judges of higher instance, as well as judges in future trials to whom the transcript could be submitted as evidence of the victorious litigant’s ownership (see Chapter 2). The contextualizations in which ’s freely occur — within the testimony and in the verdict — are precisely those in which participants other than the scribes express evaluations. Thus ’s are a reporting method in which the writer or speaker assumes most of the interpretive responsibility — a speaker-based discourse strategy, in which “explicitness and clarity are primary desiderata” (Lakoff 1984: 481). Such reports are rare in the trial record because that contextualization, by reason of its function, favors hearer-based discourse strategies (ibid.: 482) that give the intended interpreter the primary responsibility for imposing meaning on the text.
R F T
4.7
133
Free indirect speech and the boundary with narrative
Free indirect speech () in the European languages has frequently been treated — and continues to be so, even in the face of abundant counterexamples — as a primarily or exclusively literary phenomenon of postmedieval origin. In Russistic scholarship, the emergence of is generally ascribed to the late eighteenth or early nineteenth century, the period of the rise of modern Russian literature. In the following discussion and in subsequent chapters, I will provide evidence that existed in texts of purely utilitarian purpose, where it was motivated by factors other than those typically cited for its equivalent in modern languages (e.g., distancing, polyphonic complication, interior monologue).36 Many studies of have discussed the profusion of terms that have been used to designate this method. I prefer the term free indirect speech because it is parallel with direct and indirect speech and because it describes ostensible features of the strategy. (The most frequent English alternative, represented speech, is problematic, in my opinion, because every form of is a representation.) Because is a heterogeneous category, I use a definition based on central examples; some kinds of reports that have traditionally been viewed as do not fit the two essential criteria of this definition. In establishing the very existence of the reporting mode in or any other premodern language, it is necessary to use a stringent definition to exclude possible members of other categories. The first criterion is that, as implied by the label free, central examples of appear in the discourse without explicit signals of reportedness — tag clauses, adjuncts, or citation markers — in the authorial context (“absence of an indication of the speaker and of reported speech by means of a reporting clause”, Slembrouck 1986: 60). This includes unbound reports that follow ’s or occur as fadeouts from tagged or .37 Reports signalled only by particles are a grey area; I treat them as tagged . The criterion definitely excludes any report with a contiguous tag, even when the latter is “parenthetical” (intercalated or postposed) or an adverbial adjunct, and even when the report itself contains syntactic constructions or lexical and stylistic elements putatively forbidden in tagged (treated as by, e.g., Banfield 1973: 25–26, 1982; Leech and Short 1981: 329–31). These fall into my category of tagged , even if they are not prototypical examples.38 The second criterion is that central examples of , like tagged , share
134
R V
the person deixis of the enframing (authorial or narratorial) context; this means that, in the covert narrative characteristic of trial records, forms referring to the represented speaker are in the third person. Many definitions of stipulate that verbal tense must also be dependent on the enframing context, as in tagged , so that the reference points of the narrative are generally maintained. However, this is language-specific, and it is not even consistent in languages with backshifting (see Pascal 1977: 10, 31; Slembrouck 1986: 53). Where there is no backshifting, tense in can be oriented either to the narrative or, perhaps more typically, to the represented speaker, in accordance with the pattern in tagged (Paducˇeva 1996: 344). The use of the “personal present” (oriented on the here-and-now of the represented speaker, ibid.: 379–81) instead of the narrative preterit is thus a diagnostic of in languages without backshifting.39 However, the use of the narrative preterit instead of the present is not in itself a clear counterindication to . Other deictic elements such as time and space adverbials are typically oriented, as in , to the perspective of the represented speaker/perceiver rather than the narrator. However, it is not clear than this is a necessary criterion, and there appear to be inconsistent usages in various languages. Appellative, emotive, and evaluative lexical elements in can also reflect the perspective or “ideology” of the represented speaker, in the manner of .40 This possibility, though suggestive, is not criterial and only arguably sufficient. Some authors allow it for tagged as well — “texture-analyzing” (Vološinov 1929/1986: 131–32) or “coloured” (Page 1988: 33, 35–36; see also Slembrouck 1986: 61–62) — while others, as noted above, classify such reports ipso facto as despite the presence of the tagging elements. In accordance with these two essential criteria, is really a subvariety of the same mode of reporting as tagged ; it has the same essential relation to tagged that (with which it forms a continuum) has to tagged . Given the peculiarities of the central forms of deixis (person and tense), free indirect reports can be hard to demarcate from authorial or narratorial discourse, especially when they contain narrative verbs.41 Ultimately the chief diagnostic is not any formal feature but the very fact of a heteroglossic source: “the recognition of echoes [sc. ] depends on shared knowledge. They use features which are not in themselves identifiable as reporting signals but which are intended to be recognized as another voice” (Thompson 1996: 514). This likens to proverbs and allusive (“hidden”) quotations; the difference is that the interpreter of cannot look outside the text
R F T
135
for sources of the heteroglossia (Fónagy 1986). Therefore, is the most hearer-oriented (to use Lakoff’s [1984] term) of the varieties of . Its reported status is identified primarily or solely by the interpreter’s ability to distinguish between information sources and to reconstruct the configuration and viewpoints of participants in the represented speech event. This knowledge can be promoted by various contextualization cues, including “instantial norms” (Thompson 1996: 513) established within texts or genres, as well as “assessments, interpretations, judgments, ‘intended meanings’ more plausibly attributable to a character than to the narrator” (Prince 1987: 35). The ambiguity between and narrative explains, to some extent, why the strategy had been used for hundreds of years before it came to the attention of linguists. Many of the early investigators believed, indeed, that was a modern, purely literary development; for example, C. Bally claimed that it only occurred “spasmodically” in French before the nineteenth century, while E. Lerch connected it with the rise of the European novel (cited in Pascal 1977: 11, 15–16) — a view revived by Banfield (1993: 354–57).42 These claims are contradicted by the occurrence of in various premodern languages: French, as seen in the earliest specimen of Old French literature, the Sequence of St. Eulalia (Lips 1926: 118–29; Ullmann 1957: 99); English, e.g., in a 1612 account of witchcraft trials (Leech and Short 1981: 332); Danish (Haberland 1986: 232–33); and other languages. Despite this body of evidence, Banfield (1993: 340) calls (“represented speech and thought”) a “relatively recent literary form” and asserts that “the general consensus of scholarship first finds the style in a fully developed form in French in La Fontaine and in English in Jane Austen and Sir Walter Scott” (ibid.: 354; cf. Lee 1993: 382 on French). The validity of this dismissal depends not so much on empirical evidence as on how the “fully developed form” is defined — by universal criteria or by optional characteristics. Thus A. Neubert distinguishes full-fledged from a “not sufficiently marked” inchoate form found in premodern English fiction, which he terms “auctorial” (cited by Doležel 1973: 50, Note 42). In fact, many perfectly good cases of in modern literature could equally be dismissed as “insufficiently marked” if the required marking is lexicon rather than deixis. The optional characteristics may define the prototype (though that has yet to be established), but they do not exhaust the category. In Russistic scholarship it has traditionally been asserted, quite erroneously, that is an exclusively literary construction that emerged in Russian
136
R V
in the late eighteenth or early nineteenth century, either organically or (more often) under the influence of Western European models.43 Only Uspenskij (1973: 35) posits an earlier origin, but his argument is based on a false premise; his putative examples of “quasi-direct speech” (the usual Russian term for ) are actually tagged, complementized — reports that are neither indirect nor free. The use of a complementizer is not an intrusion of the authorial or narratorial point of view in . Paducˇeva (1996: 360), who argues for a nineteenth-century origin, suggests that this is a terminological muddle rather than a claim about the history of , with Uspenskij using quasi-direct speech for “citations” and interior monologue for (e.g., 1973: 41–44). In fact, Uspenskij (1973: 33) explicitly identifies “quasi-direct speech” with uneigentliche direkte Rede and style indirect libre, and makes his case while criticizing Bulaxovskij’s (1954: 442–44) discussion of uncomplementized, untagged . The trial transcripts examined, as well as other specimens of chancery writing, prove empirically that existed in medieval writings. To identify examples of in the transcripts, one must follow the original interpreters and take into account not only formal features but also information sources and reading conventions. The style of narration in the documents is covert and objective, dealing exclusively with observables. Thus the scribes are unlikely to have been the ultimate source of evaluations and cannot have been the source of information about events that occurred away from the scene of the trial unless there are explicit indications of their involvement. However, even with these rules of thumb, it is not always easy to determine whether certain stretches of discourse are as opposed to narrative or some other reporting strategy.44 These problems are illustrated by some reports in an idiosyncratic transcript of a Rjazanian criminal trial, which I will argue are . I will discuss these in detail because, as I will show, the use of and other departures from the norm are motivated here by staging considerations. One of the reports in question appears in the opening complaint (27a). Here subscript i is used in the translation for terms coreferential with the represented speaker, Kuz’ma, and subscript j for terms referring to the accused, Fedorec. (27) a.
Tako rek Kuzma: “Tot, gosudar’, Fedorec thus speak- Kuz’ma- that- liege- Fedorec-
R F T
137
ukral pcˇely, i on de s [temi pcˇely [sic] steal- bees-. and he- with [those bees svoimi jal, i muži, gospodine, u .]-. take- and men-. lord- at [togo dela byli [oboi, moi i [that affair]- be-. [both my]-. and Fedorcovy; i Fedorec, gospodine, togo pered Fedorec-.. and Fedorec- lord- that- before muži ne zapersja, cˇto u menja pcˇely men-. deny- that at me bees-. ukral; a [te pcˇely evo. A muži, stole-. and [those bees]-. his and men-. gospodine, moi i Fedorcevy, Fedorca v lord- my-. and Fedorec-.. Fedorec- in [moix pcˇelax obmuževali… [my bees]-. denounce-. Kuz’mai spoke in this way: [1] “That Fedorecj, lord[s], stole [some] bees, [2] and hei caught [Fedorecj] with those bees of hisi. [3] And present, lord[s], were men of both sides, minei and Fedorec’sj; and Fedorecj, lord[s], did not deny before those men that hej stole the bees from mei; [4] and those bees are hisi. [5] And the men, lord[s], minei and Fedorec’sj, have denounced Fedorecj for [stealing] myi bees.” (Krotov and Smetanina 1987: 68, no. 5).
This passage is unique in the corpus in that it features no fewer than four changes of strategy in quick succession. It begins conventionally, with the standard tag plus [1], as shown by the appellative gosudar’ in reference to the judges (a mistake for g(o)s(podi)ne, which occurs elsewhere in the document and itself is a conventionalized mistake, given the plurality of judges). It then slips into [2], with no vocative; here the third-person subject pronoun refers to the speaker, Kuz’ma, as shown by the subjectcontrolled reflexive possessive in the predicate and by subsequent statements of Fedorec and the witnesses (ibid.). Reportedness is indicated by the quotative particle deˇ(i), which is attested in the verdict section as well as in other types of chancery writing (see 8.3.3). The following two sentences [3]
138
R V
are in , with first-person elements referring to the speaker and vocatives referring to the hearers. Then comes another clause of [4], this time without any marker of reportedness; the referent of the third-person predicative possessive pronoun is Kuz’ma, the owner of the stolen bees, as shown by Fedorec’s subsequent testimony (ibid.). The non-narrative present tense in this segment depends on Kuz’ma’s perspective. The report concludes with a further slip into , as defined by first-person elements and vocatives [5]. It is perhaps arguable whether both indirect reports in (27a) should be treated as and the direct reports that follow them as , since by convention the entire turn at speaking falls under the scope of the tag. The writer evidently felt that the first shift to [2] threatened some loss of cohesion, since he signalled the continuation of the report by means of a quotative particle. (Intercalated quotation markers are well-suited to maintain cohesion because their syntactic position does not disrupt continuity with the preceding context; see 7.6). By contrast, the second indirect report [4] matches the formal criteria for , even in juxtaposition with attributed to the same speaker in the same turn at speaking. Slipping into is a common phenomenon: “So dependent is on context that it appears more readily in the vicinity of verbs of speech or thought, or next to instances of direct or indirect discourse…” (Prince 1987: 35). At first glance, it scarcely seems possible that the repeated fadeins and fadeouts in (27a) could reflect a purposive use of reporting strategies. One is tempted to explain the switches away as later corruptions, since the document only survives in an eighteenth-century copy. This explanation is moot in the absence of any textual history; however, it should be kept in mind that copyists generally avoided changing the substance of the texts as much as possible in order to maintain their legal utility. (Exceptions to this convention can be found in formulary redactions and some ecclesiastical writs with an edifying purpose.) Alternatively, the slipping might be ascribed to the scribe’s carelessness or difficulty in maintaining Kuz’ma’s first-person perspective. This explanation is unconvincing, since is used consistently throughout Fedorec’s subsequent testimony and in the questions of the judges. Morever, it seems odd that the writer should have slipped out of a direct perspective only in passages representing Kuz’ma’s claims about his own actions and belongings while maintaining it when he discusses the actions of others. There is, in fact, a regularity in the switches that is based on distinctions
R F T
139
in the kind of evidence being discussed. Kuz’ma’s testimony contains three pieces of evidence: he caught Fedorec with the bees in his possession; Fedorec confessed his crime in front of witnesses; and those witnesses went through a legal form of denouncing the culprit. This evidence falls into two categories: verbal, comprising Fedorec’s confession and the witnesses’ denunciation; and material, comprising the stolen property (the policˇ’noe; see Grekov (ed.) 1952: 62; Kaiser 1980: 152–53). These were distinct categories in medieval Russian law, as may be seen from the fact that they were treated in separate articles in the section of the Muscovite Lawcode of 1497 dealing with theft (Grekov (ed.) 1952: 20, §§12–13). (Even though the trial took place in Rjazan’, it was judged by Muscovite boyars.) The difference is reflected in the of the judges that follows (27a). The judges ask the defendant two questions — the first about his confession and the witnesses’ denunciation, i.e., about the verbal evidence against him; the second about the material evidence. This divides the trial dialogue into two episodes, one for each kind of evidence, with the incipit and Kuz’ma’s report as preface and the verdict as epilogue. While Fedorec’s response, a confession presented as , covers both testimonial and material evidence, the judges proceed to question him separately about the material evidence, in a further dyad of (Krotov and Smetanina 1987: 68, no. 5). At first glance it seems as though they cannot take yes for an answer; in fact, they are following an agenda based on the two kinds of evidence. In light of this legal distinction, it is significant that the portions of the complaint in (27a) that concern testimonial evidence appear as , in accordance with the convention of the text-kind, while those that deal with the material evidence appear as . Far from changing strategies haphazardly, the writer used slipping from into () and vice versa as a way of partitioning the discourse. He presented the parts of Kuz’ma’s statement dealing with the testimonial evidence to be examined in the immediately following section as , but conveyed the identification of the stolen property, which was to come later, as . This made the portions to be tabled less salient — less topical, as it were — by putting them in contrast with the adjacent, “livelier” segments of . Due to its narrative-like deixis, () is generally less prominent than when the two modes are in juxtaposition. Thus the use of () where is usual can perform a backgrounding function, with linguistic form iconic of relevance. The division of the report in (27a) into foregrounded and backgrounded parts may also reflect how the
140
R V
writer estimated the legal weight of different portions of the message. Unlike real-estate lawsuits, criminal trials were most concerned with establishing guilt; the restoration of the stolen property was a secondary issue. Backgrounding and summarizing are apparently at work in a case of in the witnesses’ statement (27b), which follows the defendant’s confession and belongs to the testimonial evidence. (It is noteworthy that the judges’ question to the witnesses, which would disrupt the continuity of the testimonial portion, is omitted.) (27) b. I [te muži, prišed pered bojare, da [vse te and [those men]- come- before boyars- and [all those recˇi po tomu ž bojaram raskazali, statements]- as.per that- boyars-. tell-. kak Kuz’ma govoril. I [tot Fedorec pered as Kuz’ma- say- and [that Fedorec]- before nimi ne zapersja togo, cˇto on u Kuzmy them- deny- that- that he- at Kuz’ma- pcˇel ukral, i s [temi pcˇelami bees-. steal- and with [those bees]-. Kuzma ego jal, po tomu ego Kuz’ma- him- take-. for that- him- obmuževali i vinu položili denounce-. and guilt- place-. And those men, having come before the boyars, told the boyars all the same things that Kuz’ma had said. And that Fedorec did not deny before them that he had stolen bees from Kuz’ma, and Kuz’ma took him with those bees [in his possession]; for that reason [they] denounced and indicted him (Krotov and Smetanina 1987: 68, no. 5).
Initially the witnesses’ statement is presented as a involving the preterit of raskazati ‘tell, relate’ with the noun reˇcˇi ‘statements’ as its object. As discussed in 4.6, ’s are used as reductive devices to present relatively minor information in an economical manner and thus to prevent the narrative from being bogged down in more details than would be necessary for present purposes. These motivations are clearly at work in (27b), since the witnesses do nothing more than confirm Kuz’ma’s statement and Fedorec’s confession.
R F T
141
The status of the underlined passage following the is not immediately evident. If interpreted as new narrative rather than an elaboration of the preceding, it could reflect an “eye to eye” confrontation (face to face with the accuser) like those found in sixteenth-century interrogation records (e.g., AAÈ 1: 143–44, no. 172). The verbs zapersja ‘denied’, obmuževali ‘denounced’, and vinu položili ‘indicted’ would then be part of the authorial narrative. Three facts militate against this view in favor of the passage being . First, it is unclear why such a confrontation would be necessary when the culprit himself had admitted his crime both in and out of court. Second, as shown by the opening complaint (27a), the question that follows it, and Fedorec’s confession, the witnesses performed the speech acts denoted by obmuževali and vinu položili when they caught Fedorec red-handed, prior to the current speech event. This suggests that the verbs form part of a report rather than the authorial narrative; they are “personal preterit” (the recollection of the represented speakers) rather than “narrative preterit” (cf. Paducˇeva 1996: 381). Third, the causal po tomu ‘for that reason’ is an evaluation of motives rather than an objective account of observable actions; as such, its most probable source is the witnesses rather than the writer, given the covert style of narration. Spitzer (1928/1988: 74–75) shows that causal clauses that are formally part of a third-person narrative often do not belong to the viewpoint of the narrator but rather express the rationales of dramatis personae, as filtered through the narrator’s perspective (“pseudo-objective discourse”).45 In other words, causal clauses often provide clues to some internal focalization. As , the underlined statement in (27b) can be understood as a fade-in, a selective amplification of the preceding dealing only with the evidence that the witnesses can uniquely offer — their corroboration of Fedorec’s confession and their own subsequent actions. When occurs alongside ’s, it represents a relative “movement towards directness” (Leech and Short 1981: 325, 336). Like other nondirect strategies, the in (27b) serves to report information of lower contextual importance without foregrounding it. This ranking function is similar to the backgrounding, deferring use of in the opening statement. Since much of the witnesses’ testimony is presupposed from Kuz’ma’s statement and Fedorec’s confession, there is no need to report it in the same actualizing manner as the preceding dialogue; that would retard the pace of the narrative more than necessary for perception of the content. The lack of a tag and subordinator in promotes
142
R V
a maximally nondisruptive (nonactualizing) transition between the witnesses’ statement and the surrounding narrative of action (cf. Doležel 1973: 47–48; Page 1988: 33). Indeed, the motivation may well be that there is no ostensible transition at all. Free indirect or nondirect speech can also be used to integrate heteroglossic information into narrative when the fact that the statement occurred per se is contextually unimportant. A case of this kind occurs in a report of essoin within a narrative of action (28). The transcript recounts that the judge has sent the constable to summon witnesses, one of whom, it turns out, cannot appear due to illness: (28) I sud’ja velel [Kuzemke Brjuxatomu [znaxorej and judge- bid- [Kuzemka Brjuxatyj]- [witnesses Malginyx postaviti i Korovaevyx. A Malga-]-. place- and Korovaj-.. and [Korovaevy znaxori na sude byli. I [Korovaj- witnesses]-. at trial- be-. and Kuzemka postavil [Savu cˇern’ca, da Mikljaja, da Kuzemka- place- [Sava monk]- and Mikljaj- and [Andrejka mel’nika, da Pozdejka, a Elizar bolen [Andrejko miller]- and Pozdejko- and Elizar- ill- And the judge ordered Kuzemka Brjuxatyj to place Malga’s and Korovaj’s witnesses [on the stand]. And Korovaj’s witnesses were at the trial. And Kuzemka placed Sava the monk and Mikljaj and Andrejko the miller and Pozdejka [on the stand], but Elizar [was] ill (AFZX 228, no. 259).
In this passage, the narrative events and states denoted by the preterit verbs all fell within the purview of the writer, as well as other observers at the trial hearing. However, the information that Elizar was ill was not thus observable; it must have come from someone who had left the scene of the trial to be present at Elizar’s home. It is, in fact, a case of or free nondirect speech, which recalls A. Thibaudet’s classic example of a sergeant asking his officer to grant a soldier leave: “‘Sa soeur fait sa première communion’” (“‘His sister is having her first communion’”; cited in Ullmann 1957: 98). Paducˇeva’s (1996: 344) illustration of is virtually identical: “He was ill [‘he said’].” The ultimate source of the report, given the nature of the
R F T
143
information, was probably Elizar himself or a member of his household. The report is filtered through the perspective of the constable Kuzemka, since it functions as an excuse for why he cannot produce the witness as ordered; note that it is conjoined with a clause that recounts his disposition of the other witnesses. (There is no indication that the scribe left the trial to follow Kuzemka’s movements.) The reported status of the underlined clause is confirmed by the fact that it appears in the “personal” present tense, which reflects the perspective of a trial participant rather than the retrospection of the scribe writing the final version of the transcript ex post facto. The motivation for the free nondirect speech in (28) is that the attribution is a matter of indifference — an unusual circumstance for the text-kind. The fact that Elizar is sick has no bearing on the issue for which the trial transcript is written; it is background for the following narrative rather than testimony, and as such there is no reason to single it out from the narrative by an explicit attribution or shift of deixis. The source of the information could be inferred from the context if there were any need to recover it, and the information itself is verified at the end of the transcript; thus tagging the report would raise a complication incommensurate with its functional load. In effect, the scribe has taken on the responsibility, ordinarily reserved for the judge, of accepting this one instance of hearsay as truth and has put it on a level with observable narrated events. This recalls Jespersen’s remark (1924/1992: 291) that — in his terminology, “represented speech” — “is chiefly used in long connected narratives where the relation of happenings in the exterior world is interrupted… by a report of what the person mentioned was saying or thinking at the time, as if these sayings or thought were the immediate continuation of the outward happenings.” The use of and free nondirect speech can also be prompted by predictability of attribution when there is no threat of ambiguity. This is evidently part of the motivation for the cases that appear within boundarysetting descriptions — the culmination of the tendency to integrate into the narrative in this contextualization.46 (29) I poveli [xr(e)st’jane cˇernye [Nazarik Senkin and lead-. [peasants black]-. [Nazarik Senkin]- s tovarišcˇi… A ot [Mininskoi d(e)r(e)vni… na [dubovoi with fellows-. and from [of.Minin village]- to [oak
144
R V
pen(’); naprave — zemlja lug [selišcˇa stump]- on.right land- meadow- [village Kermedinovskog(o) i [selišcˇa Cˇovyrevskogo [cˇernaja Kermjadinovskoe]- and [village Cˇevyrevskoe]- [black zemlja, cˇto eˇe [simonovskye xr(e)st’jane land]- her- [of.Simonov peasants]- pašut; a naleve [zemlja cˇernaja [Mininskoi plow-.3. and on.left [land black]- [of.Minin Korjakinskoi d(e)r(e)vni. A ot [dubovog(o) pni… Korjakinskaja village]- and from [oak stump]- And the state [lit. ‘black’] peasants Nazarik Senkin and his companions led… And from Minin’s hamlet… to an oak stump: on the right the land [is] a meadow belonging to the village of Kermjadinovskoe and state land belonging to the village of Cˇevyrevskoe, which the Simonov [Monastery] peasants cultivate; and on the left is state land belonging to Minin’s hamlet Korjakinskaja. And from the oak stump… (ASÈI 2: 443–44, no. 414).
The peasants who are showing the judge what they claim to be the landmarks of the disputed property are the plaintiffs in the trial; this is the only such procedure in the transcript. The first report in the description (omitted here) is separated from the initial narrative clause by several prepositional phrases; it is given as , distinguished by the vocative g(o)s(podi)ne, which is impossible in the narrative (see 3.2). This is followed by a second string of narrative prepositional and gerund phrases initiated by A ot Mininskoi d(e)r(e)vni, which precede the underlined report. The rest of the description consists of strings of prepositional phrases alternating with two additional reports in free indirect or nondirect speech and then with two relatively elaborate passages of tagged in the standard manner (ibid.: 444). This recalls the use of gerund tags medially and preterit tags finally in some boundary-setting descriptions (see 4.1). In the underlined report in (29) and in the two that follow it, there are no elements such as vocatives, deictic elements, or citation tags that could index reportedness. (The adjective Mininskoi in Mininskoi Korjakinskoi d(e)r(e)vni may be an index of , if it is a possessive referring to the plaintiff Kostja Minin rather than a fixed toponym.) The reports in (3) are, as it were, concealed in the narrative of
R F T
145
events — an effect recalled in terms such as verschleierte/verkleidete Rede and the punning oratio tecta (Coulmas 1986b: 7; Fónagy 1986: 301; Haberland 1986: 233). Nevertheless, the evaluative nature of the underlined sentence, which makes partisan claims that are ultimately for the judge to decide, precludes them from being part of a purely descriptive narrative based on observables. The intended interpreters of the transcripts, the judges, could infer the presence of heteroglossia through their knowledge of customary trial procedures and the conventions of the text-kind; they could identify the source of the evaluations from the context, which concerns the movements of specific trial participants. The extreme integration seen in (29), which makes heavy demands on the interpreter, is too sporadic to be anything other than a feature of individual style. (Cf. the occurrence of and another atypical strategy, , in the transcript from which (29) is drawn.) In this connection, it is interesting to compare the usage in a different text-kind, which resembles trial transcripts in having a dialogic format but reflects a different, nonconfrontational legal procedure. Otvodnye gramoty are records of amicable settlements in which landowners and their old-resident witnesses, under the supervision of specially appointed survey officials, agree on the boundaries of adjoining properties and set up corresponding markers, sometimes in connection with land transfers (see Pokrovskij 1973: 84). The principal judicial procedure recounted in such documents is boundary-setting of the kind seen in trial transcripts, with the distinction that the two parties’ witnesses cooperate (as a rule) and give the officials a single version of the boundary’s location. A series of otvodnye gramoty was written in 1492 as part of a survey of the lands of the Cyril Belozersk Monastery, which had just come under the jurisdiction of the Muscovite crown (or, depending on the date, would soon do so). These documents generally consist of lists of villages owned by the monastery; brief dialogues between the survey officials and the monk Martem’jan that establish the legal grounds for the monastery’s claims, on a village-by-village basis; copies of the monastery’s documentation; and boundary-setting descriptions. The dialogues are framed just like dyadic testimony in trial transcripts, with the standard strategy identified in 3.1. However, the numerous reports in the boundary-setting descriptions are usually presented as free nondirect speech, without any explicit signals of reportedness (see, e.g., ASÈI 2: 231–32, no. 290, with four tokens). Evidently the norms of the two text-kinds differ in this respect because the reports
146
R V
in trial transcripts, in general, were meant to inform the judge’s decision, whereas in otvodnye gramoty they do not, as a rule, have an evidentiary function because the documents themselves are records of a bilateral agreement. Thus it is significant that the few cases in the Beloozero collection of otvodnye gramoty in which reports in boundary-setting descriptions are presented as tagged (often ) involve disagreements among the participants (e.g., ibid.: 238, with one amicable statement given as free nondirect speech and six disagreeing statements given as tagged direct or nondirect speech; similarly ibid.: 226–27). Disagreeing statements made at major breaks in the action are represented, like dyadic testimony, by means of the standard tag formula; disagreeing statements within the episodes of boundary-setting tend to be given as nondirect (or not explicitly direct) speech tagged by forms of s”kazati. In other words, the presence of controversy, in contrast to the prevailing unanimity recounted in this text-kind, motivated the use of explicitly signalled ; cf. the formula “And about that there was a trial” at the end of boundary-setting descriptions that involve disagreements (ibid.: 238). In general, explicit marking is preferred in trial transcripts, as in otvodnye gramoty, in contexts where the fact that someone has made a certain claim can have potentially evidentiary value. Even though it is the logical culmination of the observable tendency toward reduction, the use of free nondirect speech in boundary-setting descriptions remains exceptional because it violates the need for clarity and explicitness that is crucial to the purpose of trial transcripts as a text-kind. The scribes were naturally aware that attributions could play an important role in the decisions of the judges. A final case of free indirect or nondirect speech is used to present an opening complaint (30), a nonprototypical contextualization in which another tagless stategy, , is also attested (see 3.1). The attribution can be omitted because it is trivial; by custom, the first word in the hearing belongs to the plaintiff, who is introduced as the subject of the incipit formula. The occurs as a fade-in (cf. Tannen 1989: 117–18) from a topic-setting prepositional phrase in a unique extended variant of the incipit: (30) Tjagalsja [bogojavlenskij igumen Iona s Dmitriem s sue- [of.Theophany abbot Iona]- with Dmitrij- with [Feodorovym synom Jur’evicˇem o [cerkovnoju zemlju, cˇto [Feodor- son Jur’evicˇ]- about [of.church land]-
R F T
147
pridal [Bogdan Kurunbeev na recˇki na Tjuševki give- [Bogdan Kurunbeev]- on stream- on Tjuševka- [Velikomu Bogojavlen’ju. I [Feodor Jur’evicˇ [tu zemlju [great Theophany]- and [Feodor Jur’evicˇ]- [that land]- poxal, a doval s [toe zemli v [dom plow- and give- from [that land]- to [house Bogojavlenski[j] na god po desjati altyn. I of.Theophany]- for year- at ten- altyn-. and otxodja [sego sveta [Feodor Jur’evicˇ [tu zemlju leave- [this world]- [Feodor Jur’evicˇ]- [that land bogojavlenskuju otdal [Velikomu Bogojavlen’ju, i of.Theophany]- return- [great Theophany]- and [Dmitrej Feodorov syn [tu zemlju ne vsju [Dmitrij Feodor- son]- [that land]- all- ot”exal bogojavlenskuju, kak prikazal emu otec ego cede- of.Theophany- as bid- him- father- his Feodor. I [Feodor Grigor’evicˇ sprosil Dmitrija: Feodor- and [Feodor Grigor’evicˇ]- ask- Dmitrij- Otvecˇaj! respond- The Theophany [Monastery] abbot Iona sued Dmitrij, Feodor Jur’evicˇ’s son, over church land that Bogdan Kurunbeev had donated to Great Theophany on the Tjuševka River. And Feodor Jur’evicˇ farmed that land [on rental] and gave the Monastery of the Theophany ten altyns a year from [the produce of] that land. And, while dying, Feodor Jur’evicˇ gave that Theophany land back to Great Theophany, and Dmitrij, Feodor’s son, did not cede back all of that Theophany land as his father Feodor had commanded him. And Feodor Grigor’evicˇ asked Dmitrij, “Respond!” (Morozov 1988: 300–301).
The prepositional phrase o cerkovnoju zemlju in the incipit initiates a free report of the opening complaint. Even though its object is the overarching topic, the form of reference (“church land”) reflects the monastery’s partisan claim that the disputed property belonged to the church and had only been rented by the respondent’s father. The respondent himself could not have called the disputed property “church land” or, as elsewhere in the passage,
148
R V
“Theophany land” without undermining his case; indeed, in his rejoinder (conveyed as ), he is depicted referring to the land as “the village of Otabevo” (ibid.: 301). Abbot Iona is also the only likely source of the information in the following relative clause, which conveys the monastery’s version of the property’s history. The scribe could not have supplied this information without compromising the objectivity of his recording methods, though he relayed it from his own perspective. The sequence of preterits that follow are a personal narrative within the narrative; they relate, again, to Abbot Iona’s point of view. In general, the use of and other uncanonical reporting strategies in the plaintiff’s charge may be explained, like other departures from the norms of the text-kind, as a reflection of the non-prototypical character of the speech event. The opening statement of the trial does not occur in dyadic testimony, i.e., as a response to one of the judge’s questions. Moreover, it functions as a rationale for the dialogue proper; that is, it sets the scene for the following narrative of action. Hence it can be incorporated into the preamble of the transcript along with other background information. Indeed, when trial transcripts are cited in a condensed form — e.g., within other documents — the opening charge can be omitted or replaced by a synopsis that continues the narrative of the incipit formula: (31) Tjagalsja Pozdyx s Lavrentiem…, a iskal sue- Pozdyx- with Lavrentij- and seek- Pozdyx zemli na Lavrentie Gorlovskie za Pozdyx- land- on Lavrentij- Gorlovskaja- for knjaže [sic]. I sudia sprosil Lavrentija… prince-. [and judge- ask- Lavrentij- Pozdyx sued Lavrentij…, and Pozdyx sought from Lavrentij the property of Gorlovskaja [which he claimed] as belonging to the grand prince. And the judge asked Lavrentij… (ASÈI 2: 504–5, no. 465).
Free indirect or nondirect speech was particularly well suited for condensations of this kind, since it could convey the gist of the plaintiff’s claims while eliminating the attribution and other contextually superfluous information. The same properties that make effective when the source is unimportant lie at the basis of another of its functions, which is not at work in trial transcripts — promoting the perspectival ambiguity and polyphony that is
R F T
149
valorized in much modern literature (“complication of the issue of who is speaking”, McHale 1978: 278; cf. Paducˇeva 1996: 352). This “potential for ambiguity concerning the source” (Thompson 1996: 514) can also be exploited for obfuscation or manipulation in non-literary text-kinds. In sum, the examples of embedded in the trial narrative provide further empirical proof that the strategy existed and was used purposively well before the rise of the modern novel. They also bear witness to the fact that is not an exclusively literary phenomenon, as sometimes claimed.47 One may suspect that the esthetic effects of are, in fact, by-products of more quotidian discourse functions. In any case, it is evident that the function of in trial transcripts is not esthetic at all. There is certainly no trace of the emotive coloring thought to be typical for the strategy in modern literary usage; the style of the discourse remains objective and dispassionate. This is not to deny that is used artfully in the transcripts. On the contrary, it appears in contexts where the effects that result from its properties are highly motivated. Several of the examples appear as continuations of narratives of nondyadic, nonverbal (or not primarily verbal) action outside the main narrative of dyadic verbal exchanges — stage-managing (28), boundary-setting descriptions (29), and orientation at the onset of the trial ((30)–(31)). Here the deictic set of allows it to function more or less as heteroglossic narrative, with the reports presented as if they were “the immediate continuation of the outward happenings” (Jespersen 1924/ 1992: 291; see above). The fusion is particularly thorough when, as in trial records, the does not depart from the style of the surrounding narrative. It is pertinent here that co-occurs with ’s in (27b), (30)–(31) — a reflection of a cross-linguistic tendency grounded in large part on the deictic overlap of and narrative. The infiltration of the narrative by heteroglossic material through occurs primarily in cases where the constatation of has a small functional load, either because it is fully predictable from the context — as in boundary-setting descriptions (29) or incipits (30)–(31) — or because it is unimportant per se — as probably in (28), where semi-predictability may also be a factor. The omission of attribution and ongoing-event perspective information optimizes the pacing of the discourse and hence the rate of processing and ensures that the reported material does not obtrude in the surrounding narrative clauses in a more salient fashion than warranted by its functional weight. This motivates the use of for backgrounding effects, as when it is
150
R V
juxtaposed, as in (27a), with the first-person-oriented and ipso facto relatively salient form of , with its pace-slowing, attention-getting deictic shifts. The same unobtrusiveness makes , at least when it shares the style of the surrounding narrative, suitable for “low-keyed” transitional passages, as in (28), and for summaries, as in (30)–(31) and perhaps also (29). (Cf. the use of in modern academic “reports of previous research”, Thompson 1996: 515.) Some of the effects of that I have mentioned are also associated, to varying degrees, with the other nondirect strategies. These, as noted above, tend to co-occur both with and with each other. It appears, therefore, that the nondirect strategies conspire together as a “compact” mode of reporting opposed to the diffuseness of .
C 5 The question framework
5.1
The general function of the judges’ reported speech
The of the judges in the trial narrative, which is illustrated in some of the excerpts in the last chapter, has a different character from the reports of the litigants and witnesses. The content of their statements was evidently limited both by their social role and by the conventions of the text-kind. The transcripts themselves depict the judges acting primarily as moderators during the on-site trial; their principal task (other than listening) was to elicit as much evidence as possible: “judges tended to limit themselves to an essentially supervisory role, ensuring each participant an opportunity to counter evidence produced by the opposition with proof of his own” (Kleimola 1975: 74–75; see also 80–81). The judges assumed a leading role only at the moment of the verdict. Indeed, they were not always legally empowered to render verdicts; in many cases they were restricted to gathering evidence, which they had to refer to a court of higher instance (see Chapter 2) — one of the motivations for the very existence of trial transcripts as a text-kind. The trial judges’ peculiar role is reflected in the way that they are identified in the transcripts. As was discussed in 3.1, the litigants and witnesses are usually indicated by their names or nicknames in tag clauses preposed to their reports; less frequently, titles or other lexical substitutes are used, but in general potentially ambiguous forms of reference are avoided. By contrast, it is rather common (ninety transcripts) for the judges to be identified in tag clauses not by their names but by the generic terms sud’ja ‘judge’ or, in a few cases, pis’c’ ‘chancery secretary.’ More explicit forms of reference (names and/or unique titles) are found in fifty-seven of the transcripts, and a mixture of unique and generic terms occur in eight others; in
152
R V
many of these examples, the judges were prominent boyars or princes, whom there was presumably some social pressure to identity. However, in several cases the judges are left anonymous not only in the tags but also in the incipit and end protocols, so that there is no clue whatsoever to their identity. In addition, when there are several judges hearing a trial, named or unnamed, they are always presented as a collective, never individuated; that is, choral speech is the norm.1 In one transcript involving teams of judges, the collective attribution sud’i ‘judges’ is used when the teams act together; when they act separately, they are labelled not by their names but by their liege lords — sud’i knjaži Borisovy Vasil’evicˇa ‘Prince Boris Vasil’evicˇ’s judges’ and velikogo knjazja sud’i ‘the grand prince’s judges’ (Antonov and Baranov 1997: 287–88, no. 296). Even here, the judges are presented as a chorus rather than as individual voices. These data suggest that, in many cases, the scribes considered the judges’ identity irrelevant for the purpose of the documents. If the trial judges referred the trials upwards, they would be at hand to supply the higher judges with all the necessary information. In any event, there was no need for precision in attribution, because what the judges said was not testimony and had no direct bearing on the verdict. What, then, was the function of the judges’ , if it did not supply the evidentiary information central to the text-kind? It was not included for the sake of verisimilitude, even though the question-and-answer format undoubtedly reflected actual trial procedures. One may note that the questions are omitted altogether in certain other text-kinds that report speech from dialogic exchanges — for example, inquest and interrogation records (doprosnye recˇi, rassprosnye recˇi) and depositions (skazki) (e.g., AI 1: 174, no. 120; 192, no. 130). If these documents are compared to trial transcripts, one of the differences that emerges is that each episode tends to focus on the testimony of a single speaker or a group of associated speakers; thus that there is little danger of misattributing partisan information. This contrast reveals the essential conventional function of the judges’ in trial transcripts — to enhance interpretability by promoting coherence in the discourse. (It is noteworthy in this connection that the judge’s questions, along with the tags of the interested participants’ , are omitted altogether in a trial précis used for other purposes than ordinary trial transcripts [ASÈI 3: 348, no. 319].) One of the ways in which the judges’ increased coherence was by separating the interested participants’ reports from one another (thus supporting
T Q F
153
the function of the preposed tags) and by creating dyads, the divisions that are the fundamental organizational units in the transcripts. The sentences that tag the questions — distinguished, as a rule, from those that introduce the responses by the use of a verb of asking rather than a — partition the discourse by indicating the beginning of a new dyad; the messages conveyed by the questions act like titles in establishing the overall theme of the dyad. The perceptual reality of this structure is borne out graphically in fourteen out of the fifty-eight autographs and in one of the six contemporary copies published in diplomatic editions, in which the questions as the beginnings of the dyads are singled out by capital letters (indicated by boldface), while the responses as continuations are not marked in any special way: (1) a.
I sud(’)i sprosili Onikeja da Semena… I and judges-. ask-. Onikej- and Semen- and Onikei i Senka tak rkli… I Onikej- and Senka- thus speak-. and sud(’)i sprosili Savki… I Savka tak judges-. ask-. Savka- and Savka- thus rek… I sud(’)i sprosili Makuty da speak- and judges-. ask-. Makuta- and Filiska… I Filisko i Makuta tak rkli… Filisko- and Filisko- and Makuta- thus speak-. And the judges asked Onikej and Semen… And Onikei and Senka spoke in this way… And the judges asked Savka…. And Savka spoke in this way… And the judges asked Makuta and Filisko… And Filisko and Makuta spoke in this way… (ASÈI 2: 193–94, no. 285).
This system of capitalization divides the trial dialogue into subsections initiated by the questions rather than into individual turns at speaking.2 The same question-dominant partitioning method operates in the nondyadic context of multiplex testimony (1b) — a further proof that the answers in this contextualization tend to be treated as a single instantiation of testimony spread over more than one (see 4.2). (1) b. I [knjaz(’) Ivan Jur’evicˇ(’) sprosil” [manastyr’skix and [prince Ivan Jur’evicˇ]- ask- [of.monastery
154
R V
znaxarei… I [Timoxa Denisov tako rek… witnesses]-. and [Timoxa Denisov]- thus speak- A Levon tako rek… A [Ivaško Medveˇd’ da and Levon- thus speak- and [Ivaško Medved’]- and Foka tako rkli… I [knjaz(’) Ivan Jur’evicˇ(’) Foka thus speak-. and [prince Ivan Jur’evicˇ]- sprosil tex žo [manastyr’skix znaxarej… ask- those-. [of.monastery witnesses]-. I Timoxa i eg(o) tovaryšcˇi tak rkli… and Timoxa- and his fellows-. thus speak-. And Prince Ivan Jur’evicˇ asked the monastery’s witnesses… And Timoxa Denisov spoke in this way… And Levon spoke in this way… And Ivaško Medved and Foka spoke in this way… And Prince Ivan Jur’evicˇ asked those same monastery witnesses… And Timoxa and his companions spoke in this way… (ASÈI 2: 411, no. 402).
In most of the other autographs and contemporary copies, capitals are reserved for the beginnings of major sections (incipit, record of the highercourt hearing, and verdict); in a few documents, they are used without any evident system. There are no transcripts in which the beginnings of both questions and answers are consistently signalled. The paragraphing function of the judges’ is supported by its peculiar linguistic profile — i.e., by characteristic features that reinforce the tag in demarcating the adjacent portions of separate dyads. There is almost always some formal signal of interrogativity. Even apart from the use of verbs of asking as tags, the reports tend to contain interrogative words — pronouns, adjectives, adverbs and the particle li — and they tend to be in moods other than the indicative. The prevailing locutions are questions and commands; in testimony, by contrast, questions are rare and imperatives occur in only a few, mostly formulaic expressions — usually invitations to the judge (e.g., ASÈI 1: 450, no. 571). Finally, the vocative g(o)s(podi)ne, which is ubiquitous in reports of testimony (see 3.2), as a rule does not appear in the judges’ reports. The only appellatives that occur with any frequency are brate ‘brother-’ or brat’e ‘brethren-’, used solely in reference to witnesses. In the transcripts, the primary function of these vocatives is to
T Q F
155
signal a shift to a new turn at speaking and perhaps even a new kind of verbal encounter (see 4.2). (This function is also seen in the sole case of g(o)s(podi)ne, which occurs in a transcript of a trial being judged by two teams of judges, one appointed by the local prince, the other by the grand prince. At a certain point, the local prince’s judges, who otherwise speak in chorus with the grand prince’s judges, are depicted turning from a litigant to consult their colleagues on a procedural matter. The vocative is used to mark this transition from the dominant encounter to byplay.)3 Another way in which the judges’ promotes coherence is by providing a context (frame) for the testimony. The continual shifts of perspective caused by the alternation of represented speakers had the potential to disrupt coherence — an effect that could only have been aggravated by the possibility of layering within the report. This was mitigated by the redundancy generated by the judges’ questions, which provided an ample context for each statement in the line of testimony. The wording of the questions in the written transcript created expectations in the minds of the intended interpreters, just as it did in the actual trials. The echoes of the judges’ inquiries in the testimony, like other repeated elements (cf. Halliday and Hasan 1976: 13; Brown and Yule 1983: 194; Tannen 1989: 50–51), promoted connexity and, in particular, enhanced the internal cohesion of the dyads. Elements topicalized in the questions became quickly processed givens in the answers; this focused attention on the less expected parts of the replies, the new information (a typical function of frames, cf. Tannen 1989: 49–50). The very fact that a question was asked highlighted the following response, since it is ordinarily assumed that questions will be answered; cf. the cataphoric use of the adverb tak(o) in the standard tagging formula. Highlighting is a general function of questions, which, like other collateral information or comparators, raise the possibility of alternative nonevents to heighten the significance of the actual events (see Grimes 1975: 64; Labov 1972: 384–85, 387). (There are, to be sure, isolated cases in which the judge’s report has been omitted, if analogous contexts in other transcripts are any indication — e.g., when witnesses just named by a litigant are depicted testifying without any intervening question [ASÈI 1: 247–48, no. 340 (2×)]. This is a context with high predictability and little risk of mistaking one position for another.) The framing function of the judges’ questions became less important in transcripts of the late sixteenth and seventeenth centuries. During this period trials came to be conducted primarily through petitions and documentary
156
R V
evidence, and verbal confrontations became less central to the process. The format of these lawsuit records resembles that of the earlier interrogation transcripts mentioned above, to some extent. Testimony became increasingly compartmentalized, and separate hearings were held to corroborate the written evidence of a single party or speaker in the suit. With these limitations on the possible speakers in a given hearing, it would have been uneconomical to maintain the diffuse dyadic structure characteristic of transcripts of the earlier period, and the framing function of the judges’ questions, in particular, became superfluous. Consequently, in later transcripts (e.g, Jakovlev 1943/1970: 331, no. 4.1; 342, no. 5.7), the questions were often concentrated at the beginning of the transcript or, in many cases, omitted altogether as the least informative portion of the trial dialogue. Features such as anaphora and the citation marker deˇ(i) now functioned as the chief signals of partitioning in the discourse.
5.2
Dyadic questions
The conventional strategy for reporting the judges’ utterances is uncomplementized with a preposed narrative tag. As in the testimony, the use of is motivated in part by the foregrounding of dialogic exchanges in the narrative, and in part by the explicit style of the transcripts, whose goal was to provide ample information about the speech situation as an aid to interpretation. It does not appear to be motivated by the occurrence of in the second, testimonial portion of the dyad, to judge from the cases in which responses to questions in are presented in .4 This possibility of a switch of reporting strategies within the dyads reflects the functional difference between the judges’ speech as a framing device and the other participants’ speech as a conduit for information that could be weighed against other reported information in the context. The conventional tag for the judges’ is a narrative formula centered on the preterit of a verb of inquiring — Wierzbicka’s ask2 (1987: 66): (2) i sud’ja/sud’i v”sprosil”/v”sprosili X-: “[].” and judge(s)-./ ask-./ X-: “[].” And the judge(s) asked X, “[].”
T Q F
157
As noted above, the subject of this formula can involve more specific terms (e.g., titles and/or names). The genitive object5 is a name or noun phrase referring to the addressee; this identification is necessary because the ratified addressee can change from dyad to dyad. The tag verb v”sprositi can appear in the variant spellings v(”)s(”)prositi or vosprositi.6 A related verb of inquiring, s”prositi (with spelling variant sprositi), also frequently appears. As the following passages suggest, the two verbs are evidently in free variation, without any difference in meaning, dialect, style, or function: (3) a.
I sud’ja vs”prosil Kazaka da i eg(o) tovarišcˇov: and judge- ask- Kazak- and his fellows- Otvecˇaite!… I sud’ja vs”prosil [stroitelja Ieva: answer-. and judge- ask- [abbot Iev]- Pocˇemu žo ty zoveš(’) [teˇ požni why you-. call-.2 [those fields]-. sp(a)sk(i)mi izstariny?… I sud’ja vs”prosil of.Savior-. of.old and judge- ask- Kazaka da i eg(o) tovarišcˇov: A vy Kazak- and his fellows- and you-. zovete [teˇ požni [velikog(o) kn(ja)zja, call-.2. [those fields]-. [grand prince]- komu to u vas veˇdomo… whom- that- at you-. be.known [1] And the judge asked Kazak and his companions, “Respond!”… And the judge asked Abbot Iev, “Why do you claim those fields are Savior [Monastery’s] of old?… [2] And the judge asked Kazak and his companions, “And you claim those fields are the grand prince’s, whom do you have as witnesses to that?” (ASÈI 2: 500, no. 463).
b. I sudia sprosil Vaska: Otveˇcˇai!… I sudia and judge- ask- Vasko- answer- and judge- sprosil [starca Semiona: Pocˇemu že ty [tu ask- [monk Semion]- why you-. [that požnju zoveš(’) monastyr’skoju?… I sudia field]- call-.2 of.monastery- and judge-
158
R V
sprosil Vaska: A ty [tu požnju pocˇemu ask- Vasko- and you-. [that field]- why zoveš” [Gorodišcˇ”skoju požneju? call-.2 [of.Gorodišcˇe field]- [1] And the judge asked Vasko, “Respond!”… [2] And the judge asked the monk Semion, “Why do you claim that field is the monastery’s?… [3] And the judge asked Vasko, “And you, why do you claim that field is Gorodišcˇe’s?” (ASÈI 2:383, no. 383).
This is not to say that the two verbs were used indiscriminantly; the scribes tended to keep to one or the other within the confines of a single document (thus in 104 transcripts), though the verbs do co-occur, generally to an unequal extent, in 32 transcripts.7 V”sprositi is the sole verb in 59 transcripts and the principal one in 20,8 while s”prositi is the sole verb in 45 and the principal one in 20.9 (The discrepancies are due to co-occurrence with other verbs of inquiring or with ’s.) Twenty-three of the transcripts with v”sprositi alone and 18 of the ones with s”prositi alone are autographs or contemporary copies; this suggests that the preferrence for one verb over the other was not necessarily the result of later recopying. In 14 transcripts, the verb that tags the judge’s question in the record of the higher-court hearing (see Chapter 6) differs from the verb preferred in the record of the lowercourt hearing; this may imply that the two sections were written by different scribes.10 Indeed, the interference of a second scribe is a potential factor in most of the 32 transcripts in which the verbs co-occur. Seventeen are documents that underwent referral to higher courts,11 while ten others are later copies.12 Thus only three of the transcripts in which both verbs occur can safely be considered the work of a single scribe.13 In general, the choice of one verb over the other was probably dictated by the scribe’s preference, based on the patterns in model documents or on the rules of the scriptorium in which he was trained. This is borne out by the differences between the copies of two trial transcripts of 1495–99 that appear in duplicate in the mid-sixteenth-century copy book of Trinity-Sergius Monastery. The versions from the monastery archives have v”sprositi, while those from the monastery’s representatives in Galicˇ (near Kostroma) have s”prositi (ASÈI 1: 476–77, no. 588; 488–90, no. 594). (Both verbs are attested in transcripts from the Kostroma region; thus the Galicˇ copies do not bear witness to a dialectal difference.) In a similar fashion, in one 1498/99
T Q F
159
lawsuit recorded in two slightly different documents, v”sprositi is found in one variant and s”prositi (as well as the derived imperfective v”prašati) in the other (ibid. 2: 451–53, no. 418; 453–56, no. 419). A third tag verb, v”prositi/v”prašati ‘inquire-/’ (with spelling variants vprositi/vprašati, voprositi/voprašati), is attested with in isolated examples in seven transcripts. Like v”sprositi and s”prositi, this lexeme denoted both inquiring and requesting (see Sreznevskij, s.v. “v”prositi”; SRJa, s.v. “voprositi”). In four of the documents, v”sprositi is the most common tag verb,14 while s”prositi predominates in two others.15 Although the evidence is small, v”prositi does not appear to differ from v”sprositi or s”prositi in meaning or function. This can be inferred from the fact that its imperfective v”prašati can be used to present the repetition of a command tagged by s”prositi in the preceding context (4): (4) I Volodja sprosil Fedoska: Povedi ž ty and Volodja ask- Fedosko- lead- you-. [svoeju dorogoju I povel Fedosko poperek [dorogi [. road]- and lead- Fedosko- along [road bolšie Ozereckie k ozeru… I privel ko pnju k big Ozereckaja]- to lake- and lead.to- to stump- to gnilomu da k ystoku… I Volodja vprošal [Fedoska rotten- and to spring- and Volodja- ask- [Fedosko Bašlova: Povedi ž ty ešcˇe! Bašlov]- lead- you-. more And Volodja asked Fedosko, “Lead [me] by your road.” And Fedosko led along the big Ozereckaja road to the lake… and led up to a rotten stump and a spring… And Volodja asked Fedosko Bašlov, “Lead on further!” (ASÈI 2: 455, no. 419).
It is also suggestive that v”sprositi occurs instead of v”prašati in the corresponding passage in a variant copy of the same transcript (ibid. 2: 453, no. 418). The use of the imperfective preterit in (4), suggesting a departure from the strict sequence of plotline events, has at least two possible explanations. It may represent an action overlapping with the achievement conveyed by privel ‘he led to’, which implies a stopping point; or it may be motivated by the repetition of the command thus tagged and may convey some nuance of
160
R V
insistence. The latter would resemble the “constative general factual” or “simple denotative” use of the imperfective preterit in when “the main logical emphasis is put not on the verb itself, but on some other element in the sentence”, that is, when the verb has a copula-like function and “is not required to introduce ‘kinetically’ any new concept of action” (Forsyth 1970: 84, see also 82–91; Comrie 1976: 112–14). This function of the imperfective may account for an instance of s”prošati, one of the imperfectives of s”prositi, in an unusual case involving presupposed information (5): (5) I [išcˇeja Nefetko tako rek: A sverx togo, and [plaintiff Nefedko]- thus speak- and besides that- [gospodine [sic] sud’i šljusja na knigi [lord judges]-. refer-.1 to books-. [velikogo knjazja… I sud’i sprošali [grand prince]- and judges-. ask-. otvetcˇikov… Vy šletes’ li na [te respondents-. you-. refer-.2. to [those knigi…? books]-. And the respondent Nefedko spoke in this way: “And besides that, lord judges, I refer to the books of the grand prince…” And the judges asked the respondents…, “Do you refer to those books…?” (ASÈI 2: 254, no. 296).
The situation in (5) departs from the norm in that the plaintiff is depicted introducing unsolicited documentation at the end of the trial. In general, imperfective preterit tags are uncommon in trial transcripts and, like other uncanonical reporting strategies, tend to appear precisely in unusual, marked contexts. The use of verbs of inquiring does not imply that questions are the only kind of messages in the judges’ . Although the most characteristic modality is the interrogative, commands expressed in the imperative, as in (3a) [1], (3b) [1], and (4), are frequent — a reflection of the judges’ role as moderators in the trial hearing. Because the judges’ directives are typically aimed at eliciting testimony, there is no great distance between the prototypical conveyed by the verbs of inquiring and the illocution expressed in the reports. (Actually, all of the tag verbs attested are polysemous and can also
T Q F
161
denote the of requesting — Wierzbicka’s ask1 [1987: 49–50]. However, verbs of requesting — or, at least, those that are unambiguously so — do not occur with complements in texts.) Verbs of inquiring can also tag declarative sentences, which serve chiefly as preambles to subsequent questions and commands, as in (3a) [2].16 Declarations of this kind often consist of a second layer of (see Chapter 7), because the questions at certain stages of the trial are typically aimed at clarifying the preceding testimony. While preambles of this sort are common, declarative statements in the judge’s that do not lead up to questions or directives are comparatively rare: (6) I sud’ja vsprosil Korovaa: zoveš’ Ikonnicˇe and judge- ask- Korovaj- call-.2 Ikonnicˇe- [velikogo knjazja, a kosil ego xto [to dvorišcˇo [grand prince]- and mow- it- who- [that farm]- da i [te luzi pod dvorišcˇom? I Korovaj and [those meadows]- below farm- and Korovaj- tak rek: kosili ego, gospodine, [ljudi thus speak- mow- it- lord- [people mitropolicˇi. I sud’ja vsprosil: davno li ego of.metropolitan]-. and judge- ask- long it- kosjat? I Korovaj tak: izstariny, gospodine, mow-.3. and Korovaj- thus of.old lord- kosjat [to dvorišcˇe da i [te luzi mow-.3. [that farm]- and [those meadows]- mitropolicˇi, a zemlja, gospodine, [velikogo of.metropolitan-. and land- lord- [grand knjazja. I sud’ja vsprosil Korovaa: zoveš’ prince]- and judge ask- Korovaj- call-.2 zemlju [velikogo knjazja, a oni kosjat land- [grand prince]- and they- mow-.3. izstariny, a ty dvorskoj, a živeš’ tut, a of.old and you-. steward- and live-.2 here and o tom molcˇiš’, to vedaet bog about that- be.silent-.2 that know-.3 God-
162
R V
da [knjaz’ velikij. I sud’ja vsprosil Malgi… and [prince grand]- and judge- ask- Malga- [1] And the judge asked Korovaj, “You claim Ikonnicˇe as the grand prince’s; but who has been using that farm and those meadows near the farm?” And Korovaj spoke in this way: “The metropolitan’s people, lord, have been using it.” [2] And the judge asked, “Have they been using it long?” And Korovaj [spoke] as follows: “The metropolitan’s [people], lord, have been using that farm and those meadows of old, but the land, lord, is the grand prince’s.” [3] And the judge asked Korovaj, “You claim the land as the grand prince’s, but they have been using it of old; and you are the steward and live here but you have kept silent about it — that is a matter for God and the grand prince.” [4] And the judge asked Malga… (AFZX 227, no. 259).
Only two other examples are attested (ASÈI 2: 254, no. 296; 389–90, no. 387). In all three passages, the judges’ declarations are nondyadic. In (6) [3–4], the judge turns abruptly to the other litigant without awaiting Korovaj’s response; in the other cases, the judge’s closes out the trial record. Thus, despite the use of a verb of inquiring, the judges’ represented statements are not intended to elicit further testimony; they are reproaches to the litigants for not fulfilling official duties (6) [3], for bringing in useless evidence, or for other violations of legal procedure. In two of the transcripts, the judges find against the errant litigants — in the case of (6), explicitly because of the nonfeasance in question (ibid.: 229); in the third there is no verdict. Such pre-verdict evaluations are a departure from the norm; the usual function of the judge’s in the trial section is to provide a context for the testimony, not to evaluate — the task of an active participant. The use of verbs of inquiring to tag the declarative sentences follows convention rather than semantic appropriateness.
5.3
Reduction of the tag clause and rapid-fire exchanges
There are a few cases in which the object of the verb is omitted, so that there is no indication of the addressee — e.g., (6) [2] (see Section 5.2). These examples can be regarded either as random deviations or as idiosyncratic but purposive usages. Although the evidence is slight, it suggests that
T Q F
163
the departure from the convention may have been promoted by certain features of the context. (7) I sud(’)i sprosili Kur’ana: Komu to and judges-. ask-. Kur’jan- whom- that- veˇdomo… I Kur’an” tak rek: Veˇdomo, g(ospodi)ne, be.known and Kur’jan- thus speak- be.known lord- [ljudem dobrym starožilcem [Ivašku Braginu, da [Ivašku [people good old.residents]-. [Ivaško Bragin]- and [Ivaško Šestakovu… I sud(’)i v”sprosili: Skažite v pravdu, Šestakov] and judges-. ask-. tell-. in truth- cˇ’i to požni, na koix stoim? whose-. fields-. on -. stand-.1. [Ivaško Šestakov tak rek… [Ivaško Šestakov]- thus speak- And the judges asked Kur’jan, “To whom is that known…” And Kur’jan spoke in this way: “It is known, lord, to good people, the longtime residents Ivaško Bragin and Ivaško Šestakov…” And the judges asked, “State in truth, whose fields are those on which we are standing?” [And] Ivaško Šestakov spoke in this way… (ASÈI 2: 321, no. 334).
This passage exemplifies the rhetorical use of to introduce new participants (cf. Larson 1978: 60–64), which is typical in trial transcripts. The fact that the respondent has just named his witnesses creates a presupposition that one of the named parties will be the next addressee, in accordance with ordinary procedure. This presupposition is supported by the content of the question, since the formula S”kaži(te) v” prav’du ‘state in truth’ is reserved for interrogating witnesses. Thus the context obviates the need to identify the addressee, especially as the witnesses’ testimony follows the question without interruption. (This is reminiscent of the use of names alone as tags in dramatic texts, where the reader expects a change of speakers with every paragraph. Cf. also the use of with statements in boundary-marking descriptions, where the identity of the speakers is predictable and there is a quick alternation between narrative of actions and .) In (6) [2] (see Section 5.2), the omission of the object also seems to be conditioned by context-generated expectations, although here the addressee
164
R V
is given in the narrative rather than in the preceding . The ellipsis occurs in an exchange in which the judge has just been interrogating the same addressee, whose identity is thus well established in the context. This is also true of a further example from the same transcript: (8) I sud’ja vsprosil Korovaja: a ty pak and judge- ask- Korovaj- and you-. zoveš’ Ikonnicˇe zemlja [velikogo knjazja call-.2 Ikonnicˇe- land- [grand prince]- pocˇemu? I Korovaj tak rek: vedomo, gospodine, why and Korovaj- thus speak- be.known lord- [ljudem dobrym, cˇto [to Ikonnicˇe zemlja [velikogo [people good]-. that [that Ikonnicˇe]- land- [grand knjazja. I sud’ja vsprosil: komu vedomo, cˇto to prince]- and judge- ask- whom- be.known that zemlja [velikogo knjazja? Vedomo, gospodine, to [Ofone land- [grand prince]- be.known lord- that- [Ofona Stafurovu… Stafurov]- And the judge asked Korovaj, “And you, why do you claim Ikonnicˇe is the grand prince’s land?” And Korovaj spoke in this way: “It is known, lord, to good people that that Ikonnicˇe is the grand prince’s land.” And the judge asked, “To whom is it known that it is the grand prince’s land?” “It is known, lord, to Ofona Stafurov…” (AFZX 227–28, no. 259).
Judging from internal evidence, most litigants named their witnesses right after they asserted that they could produce them; Korovaj’s hesitation thus gives the appearance of hedging, as, indeed, it does in the other example from this transcript. The question with the abridged tag in (8) is a rejoinder, with predetermined content, to Korovaj’s tantalizing statement. Clearly the reduction of the tag is motivated by and, as it were, iconic of the decrease in new information and the increase in the pace of the dialogue that presumably results from it. The same cumulative effect probably accounts for why is used for Korovaj’s reply at the end of (8), and also why the is omitted in the otherwise standard tag clause following the abridged question in (6) [2] (Section 5.2). In general, ellipsis of parts of the tag is characteristic of
T Q F
165
passages that may be called “rapid-fire” exchanges, in which much of the information is presupposed; that is, it occurs in contexts in which abridgment is motivated as a narrative technique. The fact that four cases of such ellipsis occur in a single transcript suggests that the scribe employed it as a purposive stylistic device.
5.4
Verba-dicendi tags: turning the tables on the judge
In addition to the formula cited in (2) (see Section 5.2), there are thirteen instances (in eight transcripts) in which of the judges is tagged with tak(o) rek(li) ‘spoke in this way.’ Most of these tag clauses include datives referring to the judge’s addressees, parallel to the genitive object in the conventional formula — a clarifying device motivated by the judge’s customary role as the initiator of exchanges. The use of this strategy with the judges’ is not surprising, given it is the standard formula for tagging testimony, the basic speech event in the text-kind. Moreover, none of the thirteen examples involves an interrogative .17 Ten are directives, some with declarative preambles, while the other three involve refusals, in two cases accompanied by promises. Thus the violation of convention goes in the direction of semantic appropriateness. Although the recˇi is semantically broad enough to tag any kind of reported sentence, it is associated prototypically with noninterrogative ’s, unlike verbs of inquiring, although, as mentioned above, directives that elicit information are relatively close to the ’s that verbs of inquiring convey. In one transcript, indeed, the choice of verbs clearly reflects the division between interrogative and stage-managing ’s: questions and commands that elicit verbal testimony are tagged by the preterit of v”sprositi, while the four statements about trial procedure and directives relating to the participants’ movements are tagged by tak(o) rek(li) plus the dative of addressee (ASÈI 1: 556–60, no. 642). In trial transcripts, the recˇi, as the core of the standard tag, is prototypically associated with responses in dyadic exchanges. This dialogic function accounts for the use of tak(o) rek(li) in refusals, where the judge is put in the non-dominant position — unusual for the text-kind — of having to respond to interested participants (9):
166
R V
(9) I Trofim, i Ostaš… tako rkli:… A ne and Trofim- and Ostaš- thus speak-. and not odny, gospodine, [te… selišcˇa [monastyrskie only-. lord- [these villages]-. [of.monastery xrest’jane pašut…, pašut [tex peasants]-. plow-.3. plow-.3. [these selišcˇ mnogo… i my tebe [te vse villages]-. many- and we- you-. [those all selišcˇa skažem… I sud’ja [mišutinskim villages]-. tell-.1. and judge- [of.Mišutinskaja znaxorem tako rek: Jaz [inye zemli ne witnesses]-. thus speak-. I- [other lands]-. poslan sudit’ opricˇ’ [tex dvu selišcˇ… I send- judge- except [these two villages]-. and [mišutinskie znaxori… tako rkli: K nam, [of.Mišutinskaja witnesses]-. thus speak-. to us- gospodine, [gosudarja velikogo knjazja prišla gramota, a lord- [sovereign grand prince]- come- writ- and velel nam [tebe, sud’e, [vse svoi zemli bid- us- [you- judge]- [all . lands]- ukazati ot [monastyrskix zemel’. point.out- from [of.monastery lands]-. And Trofim and Ostaš spoke in this way: “… And [it’s] not only these villages, lord, [that] the monastery peasants are farming; they are farming many of the [local] villages… And we will tell you all those villages…” And the judge spoke in this way to the Mišutinskaja witnesses: “I have not been sent to judge lands other than these two villages…” And the Mišutinskaja witnesses spoke in this way: “To us, lord, came a writ of the grand prince, and he ordered us to demarcate for you, the judge, all our lands from the monastery lands” (ASÈI 1: 450–51, no. 571).
The underlined report represents the judge’s ultimately unsuccessful attempt to suppress a line of investigation being forced on him by some of the interested participants. Muscovite courts, as Kleimola (1975: 79) observes, were concerned “primarily with settling the issue immediately at hand.” In
T Q F
167
the other cases of refusal, the judges are likewise depicted reacting to situations initiated by interested participants that they perceive as outside their legal competence (ASÈI 1: 235, no. 326; ASÈI 1: 559–60, no. 642). Thus the reports tagged by tak(o) rek(li) reflect sharp changes in footing; the judges cease to be dominant moderators, and other speakers assume, for a time, the leading role in driving on the dialogue. (The reversal is particularly striking in the second case, in which stonewalling litigants threaten to go over the judge’s head if he does not decide the suit in their favor.) The verb recˇi, associated with responses, is more motivated in these contexts than verbs of inquiring would be, since the intent of the judge’s statement is not to elicit information. It may be possible to treat some of the other reports of the judges tagged by tak(o) rek(li) as reactive and discourse-continuing rather than dominant and discourse-initiating — e.g., when the judge is answering a request for an adjournment (ASÈI 1: 558, no. 642). This may account for the unusual instance in which the plaintiff produces a witness preemptively in his opening complaint; in violation of the customary procedure, the judge reacts to this not by asking the respondent to counter but by ordering the plaintiff and his witness to show him the boundary — a tagged by tak(o) rek(li) (ASÈI 2: 364, no. 370). When the judge does eventually turn to the respondent, he asks him not for a rebuttal but for a demonstration of his version of the boundary — another procedural anomaly tagged by tak(o) rek(li) (ibid.). Another of the examples depicts the judge reacting to a bipartisan agreement, initiated by the respondents, to accept whatever version of the boundary a specific group of peasants will offer (AFZX 139, no. 157; cf. also ASÈI 2: 351, no. 358). These cases are further illustrations of the tendency, seen frequently in the transcripts, for unusual speech events to be depicted by means of uncanonical reporting strategies.
5.5
Free direct speech and crossplay within a dominant encounter
Free direct speech is not a typical strategy with the judges’ reports; this is hardly surprising, since the omission of a tag clause would tend to counteract the framing function discussed in 5.1. However, is employed in a strikingly purposive manner in one idiosyncratic trial transcript, where it occurs five times over a series of ten questions. While the judge is depicted interrogating five allied participants in this sequence, the tag clauses are only
168
R V
omitted before certain questions to the plaintiff, Ondron. This is clearly not the result of scribal error; on the contrary, the writer of this extraordinary transcript utilized selectively to create a tightly structured discourse. From the scribe’s viewpoint, there are only two real participants in the series of questions, the judge Mikita and the plaintiff Ondron; and there are only two issues being discussed in this dominant encounter — which properties Ondron intends to claim and what basis he has for those claims. Each issue forms the subject of an episode or paragraph consisting of several dyads. Within each paragraph, only the first of Mikita’s questions is presented by means of the conventional tag clause; the following questions are consistently given in , while Ondron’s responses are all presented as introduced in the standard way. This full ellipsis of the tag clause may, to some extent, be a device for conveying rapid-fire exchanges, as with the cases of partial ellipsis discussed in 5.3. This structure is complicated by the fact that, in three cases, Ondron asserts that someone else owns the properties in question. Each time this occurs, the judge’s question to the putative owner is tagged, even though his following question to Ondron is presented as . In effect, the scribe is treating Mikita’s brief exchanges with Ondron’s allies Torop, Uvarko, and Frolec as crossplay — “communication between ratified participants and bystanders across the boundaries of the dominant encounter” (Goffman 1981: 133). The omission of the tag in questions to Ondron is iconic of continuation — a return to the main speech event in the sequence. The sequence of questions begins with two dyads dedicated to clarifying the extent of Ondron’s claims (10a). Mikita’s first question is tagged, with Ondron identified as the addressee [1]; his second is presented as , with an appellative (a nominative instead of the conventional vocative) at the end to show that Ondron is still the addressee [2]. (10) a.
I Mikita vsprosil Ondrona: Tvoja li to and Mikita- ask- Ondron- your-. zemlja Merinovo? I Ondron tak rek: land- Merinovo- and Ondron- thus speak- To, gospodine, Merinovo — zemlja [šurina that- lord- Merinovo- land- [brother-in-law moego Gridkina… A Podymskaja [zemlja my]- Gridka-. and Podymskaja- [land
T Q F
169
tvoja li, Ondron? I Ondron tak your.]- Ondron- and Ondron- thus rek: To, gospodine, — zemlja [moego brata speak- lord- land- [my brother]- Toropova; a se, gospodine, Torop pered Torop-. and lo lord- Torop- before toboju. you-. [1] And Mikita asked Ondron, “Is Merinovo your property?” And Ondron spoke in this way: “That Merinovo, lord, is the property of my brother-in-law Gridka.” [2] “And is Podymskaja your property, Ondron?” And Ondron spoke in this way: “That, lord, is the property of my brother Torop; and here, lord, is Torop before you” (ASÈI 1: 235–36, no. 326).
As noted in 5.1, appellatives in the judges’ tend to be used as signals of new turns at speaking. This tracking function, which is not restricted to , follows from “everyday understanding of conversational exchanges and how they are carried out” (Klewitz and Couper-Kuhlen 1999: 482). The two dyads in (10a) are followed by a single dyad between Mikita and Ondron’s ally Torop, in which both reports are tagged conventionally (10b) [1]. Then the interrogation of Ondron resumes with a report in and Ondron identified as addressee by an intercalated appellative [2]: (10) b. I Mikita vsprosil Toropa: Pocˇemu žo, Torop, and Mikita- ask- Torop- why Torop- [tu zemlju zoveš’ svoeju… Est’ [that land]- call-.2 .- be-.3 li u tebja na [tu zemlju gramota… I Torop at you-. for [that land]- writ- and Torop- tak rek: Zemlja, gospodine, moja izstariny, a thus speak- land- lord- my- of.old and gramoty, gospodine, u menja net. A writ- lord- at me- be-. and Karpecovo, Ondron, zemlja tvoja li? I Karpecovo- Ondron- land- your-. and
170
R V
Ondron tak rek: To, gospodine, [zemlja Ondron- thus speak- lord- [land Uvarkova; a se, gospodine, Uvarko pered Uvarko-]- and lo lord- Uvarko- before toboju. you-. [1] And Mikita asked Torop, “Why, Torop, do you claim that property as your own?… Do you have a writ for that land?” And Torop spoke in this way, “The property, lord, is mine of old, but, lord, I do not have a writ.” [2] “And Karpecovo, Ondron, is it your property?” And Ondron spoke in this way: “That, lord, is Uvarko’s property; and here, lord, is Uvarko before you (ibid.: 236).
The pattern in (10b) is repeated in two more cycles, with Mikita’s tagged questions to Uvarko and Frolec each followed by an untagged question to Ondron (ibid.). Ondron’s reply to Mikita’s second untagged question completes the paragraph dealing with the extent of Ondron’s claims. The subsequent two dyads (10c) are concerned with the basis for Ondron’s claims; here there are only two participants, Mikita and Ondron. The onset of the new paragraph is signalled by a conventionally tagged question [1], while is used in the second dyad, the concluding exchange in the sequence [2]. (10) c.
I Mikita vsprosil Ondrona: Pocˇemu že ty, and Mikita- ask- Ondron- why you-. Ondron, zoveš’ [te zemli Ondron- call-.2 [those lands]-. svoimi… I Ondron tak rek: .-. and Ondron- thus speak- Est’ u menja, gospodine, na [te zemli na be-.3 at me- lord- for [those lands]-. for vse gramoty. Skazyvaeš’ gramoty, položi all-. writs-. say-.2 writs-. put- že peredo mnoju. I Ondron tak rek: Pered before me- and Ondron- thus speak- before
T Q F
171
toboju, gospodine, ne položu, položu you-. lord- put-.1 put-.1 ix pered [velikim knjazem. them- before [grand prince]- [1] And Mikita asked Ondron, “Why do you, Ondron, claim those properties as your own?…” And Ondron spoke in this way: “I have, lord, writs for all those properties.” [2] “You say [you have] writs; place them before me.” And Ondron spoke in this way: “I will not place them before you, lord; I will place them before the grand prince (ibid.).
It is important to remember that the scribe of this transcript did not have at his disposal the elaborate spacing, paragraphing, and punctuation of modern printed texts, in which is a familiar device. While he did supply some address-forms as aids for tracking the represented speakers, his choice of made strong demands on the interpretive abilities of his hearers or readers — in particular, their capacity to distinguish the discrete parts of a continuous text and their knowledge of the typical linguistic profile of the judge’s . The only other example of in a report of a judge’s utterance may also be interpreted as marking a return to a dominant encounter. In the given instance, the judge and the plaintiff are depicted having a standard dyadic exchange, presented by means of the conventional strategies (AFZX 226–27, no. 259). This is followed by six manuscript pages of crossplay. A trial transcript mentioned by the plaintiff is quoted in its entirety, and the judge calls on the respondent in the cited transcript to verify it; his corroboration involves a second document, also quoted in its entirety and verified by two further witnesses. All of these secondary encounters are represented in tagged . The dialogue then shifts back to the dominant encounter, with the judge eliciting testimony from the plaintiff of the ongoing trial, who was last mentioned six pages earlier (ibid.: 230). This question is presented as ; while there is no vocative to identify the addressee, there is a reference to the previously cited trial transcripts. While it is tempting to dismiss the omission of the tag clause as an error, the fact that it occurs in a context similar to (10a–c) suggests purposiveness, as does the fact that the scribe shows a tendency to reduce tag clauses in other contexts (cf. (6) in Section 5.2 and (8) in Section 5.3).
172 5.6
R V
Stage-managing: The dative with infinitive after speech-act verbs
While the judges’ directives are mostly reported as tagged , there are certain classes of exceptions that tend to be presented in a semantically reduced, grammatically dependent form. As the judges’ chief role in the trial hearing was to gather evidence, most of the commands reported in the transcripts serve as requests for information and require verbal responses. The framing function that the commands peform in adjacency pairs motivates the use of , with its diffuseness and deviations from the narrative orientation. However, the judges also had a secondary role as “directors” or “stage managers” responsible for organizing the movements and judicial actions of the interested participants. Directives stemming from this role are typically conveyed in a way that reduces the intrusiveness of the in the narrative — as infinitival complements after the veleˇti ‘bid, command’ (see also 4.6). This kind of speech event is not a prototypical dyadic interchange, as Longacre (1983: 57–58) points out: “Such structures contain initiating utterances but probably should not be considered to contain resolving utterances since the resolution is non-verbal.” Most of the directives reported as complements of veleˇti reflect three kinds of situations: the judges are setting a time for reconvenement,18 having documents read before the court,19 or ordering the litigants to produce previously named witnesses.20 Such commands are concerned with the participants’ movements or other behavior, and only indirectly with testimony; they expect no response besides compliance, and, as a rule, their reports are followed not by but by narrative. Thus in one case the judge is depicted sending an intermediary for information (ASÈI 2: 411, no. 402; see (7) in 4.2); this is stage-managing, not eliciting testimony by direct communication. The only recurrent situation in which the infinitive after veleˇti overlaps with is in orders initiating boundary-setting procedures. Here the illocutionary intent may involve both eliciting testimony and directing the participants’ movements; however, cooperative responses are not primarily verbal in nature. The prevalent strategy is ; ’s are only attested five times.21 It is suggestive that, in a transcript where both methods are used, a command followed by narrative of action is given as a , while one that unexpectedly occasions dialogue is presented as tagged (11):
T Q F
173
(11) I sud’ja velel pojti [znaxorem mišutinskim and judge- bid- go- [witnesses of.Mišutinskaja]-. Trofimku s tovarišcˇy, a [starcju Markelu tak Trofimko- with fellows-. and [monk Markel]- thus rek sud’ja: Poidi ty so mnoju, a speak- judge- come- you-. with me- and veli ty [svoim znaxorem s [velikogo bid- you-. [. witnesses]-. with [grand knjazja zemleju [tem selišcˇem meži prince]- land- [those villages]-. borders-. ukazati. I Markel tako rek: Jaz na [te show- and Markel- thus speak- I- to [those selišcˇa ne poslan… a s toboju na [te villages]-. send- and with you-. to [those selišcˇa ne edu… villages]-. go-.1 [1] And the judge bid the Mišutinskaja witnesses Trofimko et al. to go [to show him the disputed properties]; [2] and to the monk Markel the judge spoke in this way: “You come with me [too] and bid your witnesses show [me] the boundary of those villages with the grand prince’s land.” And Markel spoke in this way: “I have not been sent to those villages, … and I will not go with you to those villages” (ASÈI 1: 451, no. 571).
Here, conforming to the usual pattern, the successful order — a mere stage direction — is presented in a narrative form [1], while the command with an uncooperative response [2] (a highly unusual situation for the text-kind) is reported in the form preferred in dyadic contexts. The few other examples of commands met with resistance are likewise tagged as (see ASÈI 1: 509, no. 607 and its variant, 514, no. 607a; 559, no. 642). The real-world situation reflected here is complex dialogue, “when the second speaker does not like to accept the dialogue on the terms suggested by the first speaker” (Longacre 1983: 51). At the close of the trial hearing and in interludes between hearings, the judges’ reports, like those of the litigants and witnesses, tend to be presented in a grammatically integrated form. For example, the testimonial portion of
174
R V
transcripts sent to courts of higher instance typically concludes with the judges’ formulaic promise to refer the case upwards.22 (12) a.
I o tom sud’i reklis’ doložiti and about that- judges-. promise-. ask- [gosudarja velikogo knjazja, kak ukažet. [sovereign grand prince]- how order-.3 And the judges promised to ask the sovereign grand prince about that [to see] how he would command [the trial to be resolved] (AFZX 98, no. 103).
In (12a), the presence of submerged explains the future tense (nonpast perfective) of the verb in the indirect question kak ukažet, which reflects the perspective of the narrator rather than the speaker in the hypothetical speech situation.23 The preterit forms of the verbs recˇisja and jatisja ‘promise’ (or ‘agree’, Kleimola 1972: 364) serve as narrative reports of a formulaic commissive speech act, which is attested in in two cases: (12) b. I Mikita tako rek: A koli ty, and Mikita- thus speak- and if you-. dvorskoj, zoveš’ zemleju [knjazja velikogo, i steward- call-.2 land- [prince grand]- and jaz vam zemli delit’ ne smeju, a I- you-. land- divide- dare-.1 and o sem doložu [gosudarja velikogo knjazja, about this- ask-.1 [sovereign grand prince]- kak ukažet. how order-.3 And Mikita spoke in this way: “If you, the [grand prince’s] steward, claim [it] as the grand prince’s land, I am not empowered to divide the property for you, but I will ask the grand prince about this, [to see] how he will order [it to be resolved]” (ASÈI 1: 235, no. 326; see also 560, no. 542).
If, in addition to the promise exemplified in (12a), the judges issue directives at the close of the trial prior to the higher-court hearing, they are conveyed, like other nondyadic commands, as narrative — as infinitive complements
T Q F
175
after veleˇti ‘command’ or, when appropriate, after the phrasal verb dati/ ucˇiniti s“rok” ‘set a date (for reconvening)’ (13): (13) I o sem Suxoj reksja doložiti [gosudarja and about this- Suxoj- promise- ask- [sovereign velikogo knjazja a [oboim istcem i grand prince]- and [both litigants]-. and znaxorem ucˇinil srok stati pered [velikim witnesses-. make- term- appear- before [grand knjazem o [tem dele… a cˇto na [tom selišcˇe prince]- about [that matter]- and in [that village]- seno v stozex, i Suxoj [togo sena ne hay- in stacks-. and Suxoj- [that hay]- velel svoziti s [togo selišcˇa s Pavlecova bid- take.away- from [that village]- from Pavlecovo- do doklada before referral.hearing- And Suxoj promised to ask the sovereign grand prince about this, and he set a term for the litigants and witnesses of both sides to appear before the grand prince in that matter… And, as regards the hay in the haystacks in that village, Suxoj commanded that it not be taken away from that village of Pavlecovo before the higher-court hearing (AFZX 234, no. 261).
These reported commands recall the way that the judges’ directives are typically conveyed in another non-dialogic context, the verdict (see Chapter 8).24
C 6 Reporting from judicial-referral hearings
6.1
The nature of judicial-referral hearings
As previously noted, the judges who heard the testimony in the trials were not always able or empowered to hand down verdicts. There are many transcripts that illustrate a further stage in the legal process, in which the trial judges referred the suits to judges of higher instance, who could be grand princes or princesses, metropolitans, or boyars with special judicial authority (see Kleimola 1975: 68–74). The general term for hearings before higher judges (which were not restricted to lawsuits) is doklad ‘judicial referral.’ Judicial referral was one of the motivations for trial transcripts as a text-kind; the higher judges used the records of the on-site trial hearings as the principal source of information for their verdicts. Written records of the higher-court hearings and verdicts were then appended to the recopied trial records as proof of the judgment in the event of future litigation. The resulting documents are known as “referred-judgment trial records” (dokladnye sudnye spiski) or, if certified by court secretaries, “signed referred-judgment trial records” (podpisnye (dokladnye) sudnye spiski). However, many medieval copies are labelled more loosely as “judgment charters” (pravye gramoty), the term for any trial transcript that includes a verdict. The scribes who composed the appendices are often anonymous; however, given the nature of the judicial system, there is good reason to presume that they were not the scribes who transcribed the trial proceedings in situ. Thus referred trial copies can potentially reflect the interference of a second writer. Most judicial-referral records are brief and staged according to the set pattern illustrated in (1), which evidentally reflects a ritualized procedure.1
178
R V
(1) Pered [kn(ja)z(e)m” Danilom Aleksandrovicˇem sud(’)i before [prince Danilo Aleksandrovicˇ]- judges-. [Mixailo Šapkin da [Golova Semenov [ses(’) spisok [Mixajlo Šapkin]- and [Golova Semenov]- [this copy]- položili i [oboix istcov… postavili. I [knjaz(’) put-. and [both litigants]-. put.up-. and [prince Danilo Aleksandrovicˇ(’), vyslušav spisok, vsprosil [oboix Danilo Aleksandrovicˇ]- hear- copy- ask-. [both istcov: Byl li vam [takov sud, kak v litigants]-. be- you-. [such trial]- as in [sem spisku pisano? I [oba istca skazali, [this copy]- write- and [both litigants]-. say-. cˇto im [sud takov byl, kak v [sem spisku. that them- [trial such]- be- as in [this copy]- [1] Before Prince Danilo Aleksandrovicˇ, the judges Mixajlo Šapkin and Golova Semenov placed this [trial] copy and produced both litigants… [2] And Prince Danilo Aleksandrovicˇ, having heard the copy, asked both litigants, “Did you have a trial such as is written in this copy?” [3] And both litigants said that they had had a trial such as in this copy (ASÈI 2: 199–200, no. 287).
As in many, though by no means all, judicial-referral records, this example is demarcated from the trial record by a capital letter (indicated by boldface) and by the lack of an initial conjunction. Thus the beginning of the section ends the chain of conjoined clauses that constitute the trial record and initiates a new polysyndetic chain, which is continued in the verdict (ibid.: 200). As can be seen in (1), the typical ways of presenting the litigants’ in judicial-referral records differ from those preferred in trial records. A typical judicial-referral record usually begins, as in (1), with a brief narrative relating how the on-site judge or his proxy (cf. ASÈI 1: 548, no. 635) presented the trial record and the litigants or their proxies to the higher court [1]. The trial judge’s presentation can be selectively reported when it includes information that is not conveyed in the trial record (see 6.4). Next the higher judge is depicted ordering the transcript to be read aloud and asking the litigants whether it is an accurate representation of the proceedings [2]; the litigants are then depicted confirming or, in a few cases, trying
R J-R H
179
to falsify the trial transcript [3] (see 6.2). This verification usually completes the judicial-referral record; however, in some cases it is followed by reports of additional testimony — documents the litigants were unable or unwilling to submit during the on-site trial; statements elicited by the judge from government officials; or corroborations of trial record by the “men of the court” (sudnye muži), the observers of the on-site trial (see 6.2–6.3). The verification process recounted in (1) was one of the principal institutions in which trial records were utilized in medieval Russian law. It was motivated, at least in part, by a recognition that the reporting methods were fallible (see 3.3). This also accounts for the use of typicality markers like takov sud ‘such a trial’ in reference to the reported dialogue, in contrast to the token marker used in reference to the written document in ses(’) spisok ‘this [trial] record’ (see 3.4). As noted in Chapter 2, the document did not “fully constitute the legal act” (Kittay 1988: 211) until it was ratified as the official reading with the verdict; prior to that, it was a memorandum ancillary to the recollections of the litigants and men of court. Indeed, the primarily oral, reconstitutive nature of the judicial-referral hearing is striking. The higher judges, though presumably literate, are conventionally said to have “heard” the trial, with its voices reanimated by the spoken performance of the trial record and, as suggested in some transcripts, by the spoken report of the trial judge (2). (2) I [knjagini velikaja, vyslušav sud, po spisku i and [princess grand]- hear- trial- through copy- and po [sud’inu slovu po Vasil’evu, velela… through [judge- word]- through Vasilij-. bid- And the grand princess, having heard the trial, according to the [trial] record and the [trial] judge Vasilij’s statement, ordered… (AFZX 259, no. 308).
Judicial-referral hearings led to significant shifts in the trial participants’ footing; these are reflected in the composition of the text. As a rule, once the trial judges have made their reports, presented their trial records, and produced the litigants, they retire into the background; they reappear in the verdict, first as addressees and then as executors of the higher judge’s directives. The litigants mostly play a limited corroboratory role similar to that of the witnesses in the on-site trial. This reduction in their prominence
180
R V
is reflected linguistically in their lack of individuation in the clauses that introduce the judge’s question and in their own reply (if they agree with one another). In other words, despite their opposed stances, the litigants are usually presented here as an undifferentiated chorus, just as allied witnesses can be in the trial record. If nomination strategies are any guide, the central characters in judicialreferral records are the judges of higher instance. They are typically thematized and, unlike the litigants, are consistently identified in a unique, individuated way, by name and title. (As noted in Section 5.1, the lower judges in trial copies are often referred to by generic terms. However, they are identified by name at the beginning of records of higher-court hearings, where they essentially have the role of witnesses rather than moderators.) The discourse salience of the higher judges accords with their participation standing. While the purpose of on-site hearings is to allow litigants and witnesses to present their evidence, judicial-referral hearings are intended to give the higher judges an opportunity to check their facts before rendering judgment; in other words, they are preludes to the verdicts. Unlike the on-site judges, the higher judges are not primarily moderators of a two-sided verbal exchange; they play the more active roles of decision-makers and executors of judgment. Another probable factor in their discourse salience is their social prominence. Reports in judicial-referral records fall into four categories: the verification procedure itself, a ritual that is certainly presented in a prepatterned way in the written record; dialogue that ensues if one of the litigants denies the accuracy of the trial transcript; testimony in which participants adduce previously unavailable documents or other evidence; and preliminary reports made by the trial judges. As will be shown, each of these situations favors a particular reporting strategy (or complex of strategies). These context-based preferences are conditioned, at least in part, by the participants’ divers roles and changing alignment in the speech event; therefore, they provide evidence that can be used in determining some of the pragmatic properties of the reporting devices.
6.2
Verifications
The verification procedure is presented as a question-and-answer dyad. As in the trial transcript, the judge’s question, which is prepatterned as in (1),
R J-R H
181
above, serves to frame (contextualize) the following response. This explains the consistent use of tagged by the same formula as in the trial record; the only variable is the tag verb, which can be either v”sprositi or s”prositi (see Section 5.2). The litigants’ report echoes the question and is typically one of the following formulae (or minor variants thereof): (3) a.
X- [takov” sud” byl”, kak v” [sem’ s”pis”keˇ X- [such trial]- be- as in [this copy]- (pisano) (write-) X had a trial such as is written in this copy.
byl”, kakov” s”pis”k” b. X- [takov” sud” X- [such trial]- be- what.sort- copy- X had a trial like the copy.
The pronoun X refers to the represented speaker(s); its case is the dative of either relation or experiencer. Since the subject of the formula is an inanimate noun, the person of the pronoun is the only reliable indicator of whether the report is in or . In one transcript it is reported that both parties were represented by proxies (ASÈI 2: 328, no. 336); here it is impossible to differentiate between and , since the pronoun is third-person from the perspective of both speakers and narrator. Ambiguity is also present in the three cases in which there is no personal pronoun (ASÈI 1: 400, no. 523 (2×); 463, no. 582). The proximal demonstrative sei (sii, ses’) ‘this’ in (3a) may also point to if it is interpreted as an exophoric reference, i.e., if it is dependent on the external viewpoint of the scribe. This possibility emerges clearly in a judicial-referral record where one of the verification reports is in and the other in (3c): (3) c.
I [išcˇeja Toropko skazal, cˇto im [sud and [plaintiff Toropko]- say- that them- [such takov byl, kak v [sem spisku pisano; a trial]- be- as in [this copy]- write- and [otvetcˇik Mitja Mel’nev tako rek: Mne, [respondent Mitja Mel’nev]- thus speak- me-
182
R V
gospodine, sud byl, da ne takov, kak v [tom lord- trial- be- but such- as in [that spisku pisano copy]- write- And the plaintiff Toropko said that they [the litigants] had had such a trial as [is] written in this copy; but the respondent Mitja Mel’nev spoke in this way: “I had, lord, a trial, but not one such as [is] written in that copy” (ASÈI 3: 269, no. 250).
Cf. also the use of toi (tyi, tot”) instead of sei within in (8), below, and in ASÈI 2: 461, no. 422. In this endophoric usage, the two demonstratives correspond to the different deictic orientations of the speakers (distal) and the judges before whom the transcript is placed (proximal). In exophoric usage, whenever a reference to the transcript appears in the narrative or in reports presented as clear-cut , the demonstrative is invariably sei (e.g., (1) in the preceding section). However, this issue is muddied by the occasional appearance of sei within the litigants’ in reference to the trial transcript (e.g., ASÈI 1: 509, no. 607); this may be a de re intrusion of the scribe’s viewpoint. (On exophoric vs. endophoric deixis in , see Brecht 1974.) The most frequent way of reporting the verification is complementized tagged by the preterit of s”kazati ‘say, tell’ (4a). This pattern is attested 56 times in 52 transcripts, the earliest of which dates from ca. 1484–90.2 There are also four cases of uncomplementized , including (4b). The litigants (provided they agree) are presented as a chorus — an economizing device; they are usually identified generically, as in (4b), although more specific forms of reference are found when there are absentees or proxies, as in (4a): (4) a.
I [išcˇeja Karpik, i v [brata svoeg(o) and [plaintiff Karpik]- in [brother .]- meˇsto v Fedkovo, i, v [starcevo meˇsto place- in Fedko-. and in [elder- place Aleksandrovo, [Mitja Galaseˇev skazali, cˇto Aleksandr-]- [Mitja Galaseˇev]- say-. that im [sud takov byl, kak v [sem spiskeˇ them- [trial such]- be- as in [this copy]- pisano. write-
R J-R H
183
And the plaintiff Karpik, also representing his brother Fedko, and Mitja Galaseev, representing the monk Aleksandr, said that they had had a trial such as [is/was] written in this copy (ASÈI 3: 222, no. 209). b. I [oba istca skazali: Sud im and [both litigants]-. say-. trial- them- takov byl, kak v spisku pisano. such- be- as in copy- write- And both litigants said they had a trial such as [is] written in the copy (ASÈI 2: 419, no. 405; see also ibid.: 416–17, no. 404).
The presence of can be seen from the third-person pronouns referring to the represented speakers; thus there is a shift from to in the transition from the judge’s question to the litigants’ choral response. Given the absolute identity of the contextualizations, the absence of the complementizer in (4b) should perhaps be ascribed to scribal idiosyncrasy. It is likely, indeed, that both of the transcripts with this strategy are the work of a single scribe; they both record lawsuits of 1490–98 involving the Simonov Monastery and have the same trial judge, the same judge of higher instance, and the same chancery secretary as signatory (ibid. 2: 417, 419). Moreover, their verification formulae share another peculiarity — the omission of the demonstrative sem’ in the second part of the correlative construction. In a further case (ASÈI 1: 401, no. 523), uncomplementized tagged by s(”)kazati appears in the second of two conjoined reports, the first of which is complementized . Here the first report conveys the verification of the trial record, while the second, uncomplementized one recounts the corroboration of an earlier judicial-referral record. (The transcript contains two such records, one from the court of a boyar judge, the other from the court of the son of the grand prince, who had been given the privilege of deciding all real-estate litigation.) Judging from other contextualizations, there appears to be a tendency to omit the complementizer in the second report of a tautosubjective sequence (see 7.4, 8.3.1). In any case, the scribe of the second judicial-referral record shows a proclivity for reduced structures; he also omits the judge’s and does not individuate the litigants in the tag clause, despite the fact that one of them is a proxy. These abridgments are evidently motivated by the fact that the information in the second judicial-referral record is repetitive.
184
R V
It is surprising, at first glance, that complementized after s”kazati should be the chief method for reporting the responses in the verification procedure. After all, this strategy is hardly ever found in the narrative framework of the trial record (see Section 4.4), to which the judicial-referral report is appended. Moreover, as noted above, the question that elicits the response is always in , so that the use of involves a shift in orientation. It has already been noted that the litigants play a less prominent role in the judicial-referral hearing than in the trial. This change of footing reflects the mission of the higher court — to check the testimony and settle other issues in preparation for the verdict. The litigants testify primarily as witnesses of the on-site trial; their statement is meant to corroborate a written artifact, not to add to the body of evidence. Moreover, the litigants’ statement is partially presupposed; it was not expected to be more, in essence, than yes or no. Presupposition often favors complementization and indirectness (see 4.4). When the litigants agree, as they do in most cases, there is no need to present the hearing as a confrontation; thus their statements, which were undoubtedly separate in the actual hearing, can be merged, just as the of allied witnesses can be in the trial record. The only information that needs to be foregrounded is the constatation of their , i.e., the attribution; the use of a complementizer furthers this by rendering the tag more prominent than the report. These factors enable the statements to be viewed as a preliminary to the more important business ahead and, accordingly, to be presented in a less salient manner, in a way that demands less attention than the typically found in full-fledged testimony. This explanation is borne out by the way additional is treated in the judicial-referral record. When participants offer truly new, potentially probative information, it is given in , for the most part tagged by tak(o) rek(li), like testimony in the trial record. This includes twelve examples in which litigants adduce new documents (5a), writs that they mentioned but did not produce during the on-site trial, or local land-cadastre books (5b);3 one in which the opposing litigant responds to such evidence under the judge’s prompting (ASÈI 3: 269, no. 250); and one that involves discussion of a no-show litigant (Antonov and Baranov 1997: 288, no. 296).4 (5) a.
I [oba isca skazali, cˇto im [sud and [both litigants]-. say-. that them- [trial
R J-R H
185
takov byl, kak v [sem spisku pisano. A such]- be- as in [this copy]- write- and [troeckoi zakašcˇik Ofonasei tako rek”: Ta, [of.Trinity rent-collector Ofonasij]- thus speak- that- g(ospodi)ne, [zemlja Noskovo tro(e)ckaja [Sergieva lord- [land Noskovo]- of.Trinity- [Sergij- mona[styrja]; a es(t’), g(ospodi)ne, u nas na monastery]- and be-.3 lord- at us- for [tu zemlju [gramota pravaja [Kuzmy Klimentieva [that land]- [charter of.law]- [Kuz’ma Klimentiev]- a vo se g(ospodi)ne [gramota pravaja pered and lo lord- [charter of.law]- before toboju. I [knjaz(’) Vasilei Ivanovicˇ(’) veleˇl pered you-. and [prince Vasilij Ivanovicˇ]- bid- before sobou [gramotu pravuju cˇesti. I v gramoteˇ v . [charter of.law]- read- and in charter- in pravoi napisano, cˇto [ta zemlja Noskovo — of.law- write- that [that land Noskovo]- troeckaja [Sergieva monastyrja, a dal of.Trinity- [Sergij- monastery]- and give- eeˇ k monastyrju [Iev Obeˇrucˇev. her- to monastery- [Iev Oberucˇev]- [1] And both litigants said that they had had a trial such as [is] written in this copy. [2] And the Trinity rent-collector Ofonasij [the respondent] spoke in this way: That property of Noskovo, lord, belongs to Trinity-Sergius Monastery, and we, lord, have a judgment charter for that land [given by] Kuz’ma Klimentiev; and there, lord, is the judgment charter before you. [3] And Prince Vasilij Ivanovicˇ commanded the judgment charter to be read before him. And in the judgment charter [is] written that that property of Noskovo belongs to Trinity-Sergius Monastery, and Iev Oberucˇev gave it to the monastery (ASÈI 1: 471, no. 586).
Here the complementized [1] is used to foreground the attribution of a presupposed verification, while uncomplementized [2] is used to present new evidence in a salient way. This shift in discourse prominence reflects a
186
R V
change not only in the character of the reported information but also in footing of the represented speaker in the hearing. The respondent ceases to be a passive corroborator and takes the initiative; it is his reported statement [2] that determines the further direction of the speech event. It is noteworthy that the additional testimony in (5a) [2] is conjoined to the preceding context by a instead of the sequencing conjunction i. As noted in Sections 4.2–4.3, the non-sequencing a is used contexts such as multiplex testimony to imply that the reports thus conjoined are viewed as distinct parts of a single statement. If this is true in (5a), the portion of the respondent’s statement that overlaps with that of the plaintiff is presented as choral [1], while the rest is presented individually, of necessity, but still not as a fully separate turn at speaking [2]. This bracketing is reinforced graphically in (5a) and three other autographs (ASÈI 1: no. 583, 587, and 589) where the initial a is not capitalized, unlike the i of the following turn at speaking (boldfaced). As (5a) shows, any new documents that the litigants adduce are summarized and backgrounded through the use of a complementizer and without any features of “direct” (non-narrative) orientation. This usage shares the motivation for — more precisely, the disincentive for — that is under discussion. The judgment charter mentioned in (5a) [3] is only cited as a confirmation of the respondent’s previous assertion [2], which it echoes; in other words, the summary of its pertinent information is largely presupposed. In general, tends not to be an optimal strategy for summarizing because of its diffuseness and the perspectival information that it conveys, which is superfluous in the given function. By contrast, extracts from land-cadastre books, which were housed in state chanceries, are cited in (5b). In such cases, the judge sends a secretary — an official of some dignity — to obtain the information from the chancery secretary outside the hearing; (5) b. I [knjaz(’) Vasilei Ivanovicˇ(’) veleˇl [d’jaku svoemu and [prince Vasilij Ivanovicˇ]- bid- [secretary . Oleške Bezobrazovu u d’jaka u Mikifora is Oleška Bezobrazov]- at secretary- at Mikifor- from [kostromskix knig [manastyr’skie zemli [of.Kostroma books]-. [of.monastery lands
R J-R H
187
tr(oe)cskie vypisati, da sobeˇ skazati. I [Oleška of.Trinity]-. extract- and . tell- and [Oleška d’jak u Mikifora u d’jaka iz secretary]- at Mikifor- at secretary- from knig vypisal, da skazal [kn(ja)zju Vasil(’)ju books-. extract- and say- [prince Vasilij Ivanovicˇ(ju): Napisano, g(o)s(podi)ne, v [kostromskix Ivanovicˇ]- write- lord- in [of.Kostroma knigax [leˇt(a) semdesjat tret(’)jag(o) pisma books]-. [year- seventy- third]- writing- [Mixaila Volynskog(o): [zemli tr(oe)cskie [Sergeeva [Mixajlo Volynskij]- [lands of.Trinity]-. [Sergij- manastyrja v [Zaleˇzskom Berezuice v Verxnem, na monastery]- in [Zalesskij Berezovec]- in upper- in Ogloblineˇ Ondronko. Oglobino- Ondronko- And Prince Vasilij Ivanovicˇ commanded his secretary Oleška Bezobrazov to extract [records about] the Trinity Monastery lands from the Kostroma books kept by the secretary Mikifor and to tell [them] to him. And the secretary Oleška made the extract from the books kept by the secretary Mikifor and told Prince Vasilij Ivanovicˇ, “[It is] written, lord, in the Kostroma books of the year ’73 [1464/1465] written by Mixajlo Volynskij: the lands of Trinity-Sergius Monatery in Zalesskij Verznij Berezovec — in [the village of] Oglobino [is] Ondronko” (ASÈI 1: 475, no. 587).
As can be seen, the court takes the accuracy of the extract on trust; in other words, the secretary is called on to make a statement — in fact, to testify, though not as an interested party. This explains why is used and also why the secretary is named, whereas the officials who read the documents that the interested participants provide remain implicit and anonymous (5a). (The dignity of the secretary’s rank may also be a factor here.) The use of s”kazati to tag the secretary’s report is perhaps to be explained by the nondyadic nature of the report; the same verb is used to introduce the nondyadic reports made by the trial judges in the higher-court hearings (see Section 6.4, below).
188
R V
While is used, as in (5a), for new testimony in the judicial-referral hearing, complementized is used in two cases where the higher judges elicit additional statements to check on specific points raised but not settled in the on-site trials. One of these examples is an outright repetition. During the trial the plaintiff had cited a copy of a donation charter, the original of which he said was in the keeping of the chancery secretary Kobjak; this was challenged by the respondents. In the interlude between the trial and the higher-court hearing, the lower judge took Kobjak’s testimony, which is reported as complementized after s”kazati (see 4.4). Since his word was the only outside proof of a crucial piece of evidence, Kobjak was called as a witness in the judicial-referral hearing. Although slightly more explicit in content, Kobjak’s is presented in the same way as in the narrative of the interlude, from which it is separated by only one sentence (ASÈI 1: 541, no. 628). It is not new evidence but rather preliminary corroboration of already given information, much like the litigants’ responses in the verification procedure. The same is true of the second example, in which the higher judge calls on a nobleman absent at the trial hearing to confirm that he had in fact donated the disputed land to the respondents as claimed: (6) I [knjaz(’) Vasilei Ivanovicˇ(’) v”sprosil [kn(ja)zja Semena and [prince Vasilij Ivanovicˇ]- ask- [prince Semen Romanovicˇ(ja): Dal li esi v” manastyr’ Romanovicˇ]- give- ..2 to monastery- k” [s(vja)t(o)mu Sp(a)su [požnju Kr(e)stcy? I [knjaz(’) to [holy Savior]- [field Krestcy]- and [prince Semen Romanovicˇ(’) pered [kn(ja)z(e)m” Vasil(’)em Semen Romanovicˇ]- before [prince Vasilij Ivanovicˇem skazal, cˇto [požnju Kr(e)stci… on Ivanovicˇ]- say- that [field Krestcy]- he- dal… v manastyr(’)… give- in monastery- And Prince Vasilij Ivanovicˇ asked Prince Semen Romanovicˇ, “Did you give the Holy Savior Monastery the field of Krestcy?” And Prince Semen Romanovicˇ said before Prince Vasilij Ivanovicˇ that he had given the field of Krestcy… to the monastery (ASÈI 3: 219, no. 208).
R J-R H
189
Here again the report is treated like the verification procedure immediately preceding it — as complementized tagged by s”kazati; note the emphatic third-person pronoun referring to the represented speaker. (Where there is no emphasis, subject pronouns are often omitted in tautosubjective complement clauses.) The fact that statements are corroboratory and partially presupposed does not imply that they are trivial. Thus the indirect report in (6) is quoted in the verdict (ibid.) as being of pivotal significance, since the plaintiff’s claim to the land had revolved around its not being a different property. There is considerable evidence that complementized was generally preferred for reporting simple, presupposed, corroboratory messages in chancery language of the fifteenth and early sixteenth centuries. This can be seen in other text-kinds that record official verifications. For example, many of the surviving fifteenth-century wills were submitted to high churchmen for probate; during these hearings, the scribes and witnesses were called on to corroborate the testaments. The records of these proceedings, illustrated by an example from 1433 (7), were appended to the proven wills. (7) A [sija duxovnaja gramota [Ione vladyce javlena, a [pop and [this spiritual writ]- [Iona bishop]- show- and [priest Genadej i vo [vsex posluxov mesto tuto ž Genadij]- in [all witnesses]-. place- there stojal pered vladykoju i skazal, cˇto [sja duxovnaja stand- before bishop- and say- that [this spiritual gramota pered nimi pisana. A [Ivan Vasil’evicˇ’ writ]- before them- write- and [Ivan Vasil’evicˇ]- skazal, cˇto [sju gramotu on pisal. say- that [this writ]- he- write- And this will [was] shown to Bishop Iona, and the priest Genadij, representing all the witnesses, stood there before the bishop and said that this will [had been] written before them. And Ivan Vasil’evicˇ said that he had written this document (ASÈI 1: 87, no. 108).
Complementized after s”kazati, as in the underlined segments, is the sole or main way of presenting in 14 out of the 16 cases of this text-kind examined (including all of the cases in ASÈI, AJu, and AJuB).5 In one example, an indirect report is continued by an even more integrated strategy, fused (ASÈI 1: 100, no. 6); in another, the first report is complementized and the
190
R V
second uncomplementized (ibid: 473–74, no. 494). One of the exceptional cases has complementized after jatisja ‘promise’ (ibid.: 180, no. 251). The other has the standard formula for testimony, after tak(o) rek(li); however, this reflects the usage of a different scriptorium, since the text is the only one from outside the metropolitan’s chancery (ASÈI 2: 513, no. 474). Complementized indirect (or nondirect) speech tagged by s”kazati is also typical in similar higher-court verification records appended to other contemporary text-kinds, e.g., surety bonds (porucˇnye kabaly), purchase deeds (dokladnye kupcˇie gramoty), and records of amicable land settlements (dokladnye delovye gramoty).6 While tagged may also be found in some records (e.g., AJuB 1: col. 493–94, no. 69.2; col. 496–97, no. 70), the statements thus presented are presented as dyads and have more content, including new information, than do mere verifications. Further evidence for a correlation between and new, non-corroboratory information will be discussed in Section 6.3. This is not to say that is never used in reporting the litigants’ verifications. There are, in fact, eight cases in which the litigants’ statements in the verification procedure are presented in the conventional strategy for reporting testimony in trial records, tak(o) rek(li) plus .7 The earliest example dates from 1442–62, and at least two of the others from the 1460s; this makes them the earliest cases of in judicial-referral records, whereas the more common pattern involving first appears, as noted above, in documents of the 1480s. The verification formula can have a different, looser structure in these early cases (as does the judge’s question, not cited here): (8) I [istci oboi da i ix [znaxori muži and [litigants both]-. and their [witnesses men vse pered [velikim knjazem tako rkli: all]-. before [grand prince]- thus speak-. sud gospodine, takov nam o [tex zemljax trial- lord- such- us- about [those lands]-. byl pered Fedorom, kakov [toj spisok ego. be- before Fedor- what.sort- [that copy]- his And both the litigants and also their witnesses, all their men, spoke in this way before the grand prince: “We did, lord, have a trial about those lands before Fedor like that copy of his” (AFXiX 1: 117, no. 125).
R J-R H
191
The use of here is indicated by the first-person pronoun and the vocative referring to the participants of the represented speech event. This example is particularly unusual in that it depicts the witnesses as well as the litigants participating in the verification routine. Even though is not optimal for representing presupposed corroborations, there are at least two reasons why it is not surprising to find it in this contextualization. First, is associated with testimony, the prototypical speech event for the text-kind. It would require little stretch of the imagination to view the litigants’ statements and behavior during the judicial-referral hearing as additional evidence (cf. Kleimola 1975: 85–86). Analogous verifications of documents, including earlier trial transcripts, can be recounted within the record of the on-site trial on a par with ordinary testimony — as they are, in a way, since the history of the cited documents is not an observable for the judge or scribe (see AFZX 229, no. 259; ASÈI 2: 419, no. 405; 461, no. 422). In such cases, the answers not only verify but also amplify the information contained in the old trial transcript; even negative responses add to the body of evidence by casting doubt on the claims of the opposing parties or, ultimately, on the veracity of the speakers. Second, is associated with dyadic reports in narrative, and the verification procedure is dyadic (or quasi-dyadic, given the choral speech) and set in a minimal narrative. Indeed, in some transcripts, the trial narrative flows continuously into the judicial-referral record, without the closure provided by the final , as if the statements in the higher court continue those of the lower without any temporal break (see, for example, ASÈI 1: 434, no. 557; and 541, no. 628). While the use of in the judicial-referral verification can be interpreted as an extension or continuation of the pattern ordinarily employed for testimony in the trial record, there is another possible explanation. The chronological distribution of the examples suggests a drift from direct to indirect style. Especially noteworthy in this regard is the preponderance of complementized in the early-sixteenth-century transcripts (13 out of 15 cases). However, the earlier transcripts investigated do not provide fully secure evidence, given that the time gap between the earliest attested examples of and is some forty years at the outside. Preliminary examination of later trial transcripts shows that complementized after the preterit of s”kazati likewise predominates in judicial-referral records from the reigns of Vasilij III Ivanovicˇ (1505–33) and Ivan IV Vasil’evicˇ (1533–84).8
192
R V
Only a single case of tagged by tak(o) rek(li) was found (AJu 31, no. 16). It is possible that the two instances of uncomplementized tagged by s”kazati, which occur in separate transcripts of the same lawsuit of 1498/99 (ASÈI 2: 453, no. 418; 455–56, no. 419), represent an intermediate stage in the proposed drift or a compromise between the two strategies. Perhaps a further stage can be seen in an early sixteenth-century transcript from Muscovite-annexed Novgorod, where the verification report is tagged by s”kazati and complementized by cˇ’to (Koreckij 1969: 290, no. 2). In any event, there is no evidence for any unusual contextual factors to distinguish these examples from the better-represented cases presented as .
6.3
Falsifications
The functional differences between and other reporting strategies stand out with particular clarity in cases in which one of the litigants is depicted denying the accuracy of the trial transcript. This contextualization appears in eight judicial-referral records, two of them in variant transcripts of a single trial. Internal evidence suggests that there was a set procedure for checking disputed transcripts: the higher judge would ask the litigants whether they were willing to call on the testimony of the “men of court” who had been present at the original trial (cf. Kleimola 1975: 18, 69–71; Kaiser 1980: 136). The threat of being unmasked by the men at court sometimes led the recalcitrant litigants to admit that the record was accurate or to withdraw their protest without such an admission. These recantations tend to be reported differently from the verifications.9 In cases of falsification, the litigants’ statements are introduced separately, with tag clauses conjoined by the non-sequencing conjunction a. In six of the examples, the verification reports are in complementized tagged by the preterit of s”kazati, the usual pattern with corroborations. In the other two (one of which is the earliest attestation of the procedure), they are presented like standard testimony. In one of these cases, significantly, the plaintiff’s statement, which is restricted to the formula and thus presupposed, is given as complementized , while the respondent’s falsification, which is amplified by extensive new information (including a reference to a landcadastre book), is presented as uncomplementized (ASÈI 3: 269, no. 250). Likewise, if the uncooperative litigants make any subsequent statements that
R J-R H
193
go beyond the ritualized verification and offer information pertinent to the verdict, they are conveyed by the strategy that is standard for testimony, even if the falsification itself is in . The use of reflects the fact that such statements count as testimony, even if they are retractions10 or maneuvering, as in (9), because they cast doubt on the bona fides of the litigant who challenged the transcript. (9) I Šalaba skazal, cˇto emu [sud takov byl, kak and Šalaba- say- that him- [trial such]- be- as v [sem spiske pisano; a Mixejko skazal, cˇto in [this copy]- write- and Mixejko- say- that emu sud byl ne takov, kak v [sem spiske him- trial- be- such- as in [this copy]- pisano; i poslalsja s sud’jami po spisku na write- and refer- with judges-. by copy- to [sudnye muži. Na [tretej den’, stav pered [of.court men]-. on [third day]- stand- before Dmitreem, Mixejko tako rek: Cˇto esmi, Dmitrij- Mixejko- thus speak- that ..1 gospodine, poslalsja s sud’jami po spisku na [sudnye lord- refer- with judges-. by copy- to [of.court muži, i jaz, gospodine, na [sudnye muži ne men]-. and I- lord- to [of.court men]-. šljusja; a sud ne takov byl, kak v [sem refer-.1 and trial- such- be- as in [this spisku pisano. copy]- write- And Šalaba said that he had a trial such as [is] written in this copy; but Mixejko said that he had not had a trial such as [is] written in this copy, and, with the judges, he referred to the men of court, in accordance with the copy. On the third day, having appeared before Dmitrij [the judge of higher instance], Mixejko spoke in this way: “As for the fact that I referred, lord, with the judges to the men of court, in accordance with the copy, [now] I do not, lord, refer to the men of court; but the trial was not such as [is] written in this copy” (ASÈI 3: 271, no. 251).
194
R V
The remarks of the higher judges are generally treated like those of the onsite judges in the trial record. When they elicit testimony, they are presented as after a verb of asking; when they relate to trial procedure (“stage managing”), they are presented as grammatically integrated speech after veleˇti, that is, as part of the third-person narrative (cf. the nondialogic commands to clerks and other court personnel in ASÈI 1: 493, no. 595; 509, no. 607; 514, no. 607a; 3: 186, no. 172; 269, no. 250). Procedural statements by other participants likewise tend to be integrated into the narrative; thus, when the litigants agree to call on the men of court for corroboration, their statements are given as ’s, as in (9).11 The same distinction between narrated procedure and reported testimony appears in two examples in which bailiffs presenting the men of court make preliminary reports before the judges. In both these cases, the litigants’ separate responses in the verification procedure, which are presented as , are followed by narratives of action, including ’s (ASÈI 1: 493, no. 595; 3: 186, no. 172). In the first instance (10), the bailiff’s report begins as a continuation of the narrative: (10) I pered [knjazem Vasil’em Ivanovicˇem [nedelšcˇik Vaska and before [prince Vasilij Ivanovicˇ]- [bailiff Vaska Karacˇev [sudnyx mužej… postavil; a [sud’ja Ivan Karacˇev]- [of.court men]-. place- and [judge Ivan Kuzmin da [starec Efrem stali že. A Kuz’min]- and [elder Efrem]- stand-. and [otvetcˇika Stepanka Dorogu, — skazal [nedelšcˇik Vaska [respondent Stepanko Doroga]- say- [bailiff Vaska Karacˇov, — dal na poruku, a vyrucˇil ego, Karacˇev]- give- to bail- and bail- him- gospodine, u menja [Ofremko Bulgak Jakušov syn, da lord- at me- [Ofremko Bulgak Jakuš- son]- and [Oleško Dmitrov syn Bobrovnikov; i tot, gospodine, [Oleško Dmitr- son Bobrovnikov]- and that- lord- Stepanko zbežal [s] [svoimi porucˇniki. Stepanko- run- [with [. bondsmen]- [1] And the bailiff Vaska Karacˇev placed the men of court… before Prince Vasilij Ivanovicˇ; and the [trial] judge Ivan Kuz’min and the
R J-R H
195
monk Efrem [the plaintiff] also appeared. [2] But, said the bailiff Vaska Karacˇev, [he] set bail for the respondent Stepanko Doroga, “and, lord, Ofremko Bulgak, Jakuš’s son, promised me to be his surety, as did Oleško Dmitr’s son Bobrovnikov; and that Stepanko, lord, ran away with his sureties” (ASÈI 1: 493, no. 595).
In accordance with normal expectations, the passage marked [2] should be a continuation of the narrative [1] affirming that the respondent — the party who rejected the trial transcript — is present; note the fronting (topicalization) of the corresponding noun phrase otvetcˇika Stepanka Dorogu. Instead, [2] initiates the bailiff’s report of how the respondent has absconded. As the events described are not observables in the ongoing speech event, the scribe has to disclaim responsibility for this information; however, he demarcates it in a way that minimizes the transition from the narrative, by intercalating the preterit of the s”kazati — the usual tag for the of officials other than the judges — right before the point at which he identifies the subject, who is also the information source. As an interruption, intercalation is a strategy that delays the complete reception of heteroglossic information for various pragmatic ends (see Shapiro 1984: 73–76; Leech and Short 1981: 333, 351 n. 11; Page 1988: 27). In (10), it allows the reported information to be perceived as a continuation of the narrative by postponing the attribution. This effect of seamless transition is furthered by the absence of elements exclusively dependent on the reporter’s perspective in the underlined clause. While the strategy is ambiguous here, the subsequent clauses are clear-cut , with a first-person pronoun and vocatives oriented on the bailiff’s viewpoint. This information is pertinent to the verdict, since flight constituted an admission of guilt (see Kleimola 1975: 89–90), and the sureties were financially culpable for their actions. Hence serves as a foregrounding device, and the report itself counts as testimony. By contrast, the underlined quasi-narrative report is to some extent presupposed, since it describes a standard procedure; as background for the more immediately relevant information that follows, it need not be conveyed in a salient manner. A second bailiff’s report (ASÈI 3: 186, no. 172) follows much the same pattern.12 Presented as a continuation of the narrative, it is attributed by an intercalated , just as in (10). There is no fade-in to . Presumably was not optimal here, since the report is limited to background information
196
R V
without any bearing on the verdict (accounting for the absence of one of the men of court, who has died in the interval between hearings). The of the men of court can be presented like standard testimony (ASÈI 1: 494, no. 595) or like typical corroborations, with s”kazali plus (ibid. 3: 186, no. 172). In the case with (in the passage following (10)), the statement goes beyond the restricted content of the verification formula — a reference to the absconding litigant. While this information has already been established in the context, it zeroes in on the truly crucial point in the judgment and thus merits foregrounding. By contrast, the statement in is restricted to the verification formula. The same pattern may be observed in a case in which a litigant returns to recant his falsification (11): (11) Po [semu spisku, stav pered [Mikitoju Vasil(’)evicˇem by [this copy]- stand- before [Mikita Vasil’evicˇ Beklemišovym, … Andreˇiko skazal, cˇto sud byl Beklemišov]- Andrejko- say- that trial- be- ne takov, kak pisano v [sem spisku. Da sšed s such- as write- in [this copy]- and leave- from suda, opjat(’) stav pered Mikitoju, skazal, cˇto court- again stand- before Mikita- say- that sud byl takov, kak pisano v [sem spisku trial- be- such- as write- in [this copy]- In accordance with this copy, having appeared before Mikita Vasil’evicˇ Beklemišov, Andrejko said that he did not have a trial such as [is] written in this copy. And, after leaving the court, he appeared again before Mikita [and] said that the trial was such as [is] written in this copy (ASÈI 1: 400, no. 523).
The underlined report is simply a positive restatement of the verification formula. It differs from the other recantations (ibid.: 509, no. 607; 514, no. 607a; Kaštanov 1970: 359, no. 9) and from the unrepentant retraction in (9), which are in clear-cut , in that it does not include any new information such as expressions of guilt or resignation. The largely presupposed nature of the report may explain why it does not contains any of the features of that clash with the narrative perspective — that is, any of the elements that could actualize and foreground information perceived as probative and prominent. In the absence of personal pronouns, there is no way to determine
R J-R H
197
which reporting strategy is used, though the proximal demonstrative in v sem spisku can perhaps be interpreted as an exophoric reference to the narrator’s viewpoint. The use of a complementizer, though not a reliable indication of , is typical with presupposed reports like the one in (11) (see 4.4). Unlike the other cases of falsification, the passage in (11) is patently reduced, with all the non-probative reports such as the judge’s question and the plaintiff’s verification omitted. These abridgments make the respondent’s falsification and its aftermath the sole focus of the narrative. Evidently the scribe was striving for the bare minimum and so restricted the judicialreferral record to information that could not be inferred from background knowledge — the very information that was pertinent to the verdict. The probable motivation for this abridged style was the fact that the hearing depicted was a mere formality; even the judge of higher instance was not authorized to decide the case, since Grand Prince Ivan III Vasil’evicˇ had reserved all real-estate lawsuits for his heir, Ivan Ivanovicˇ. The given trial transcript thus includes a second judicial-referral record. As a preliminary ritual, the first hearing could be summarized; there was no need to recount it in any detail. (The record of the second hearing is also relatively brief, but the corroboratory statements of both the litigants are reported in the usual manner.) Thus the use of complementized in (11) and the omission of elements dependent on the represented speech situation were well motivated; they minimized the amount of information transmitted and avoided two potentially retarding factors — shifts in perspective and the syntactic discontinuity characteristic of uncomplementized reports.
6.4
Preliminary reports
In several transcripts, the trial judges are depicted making a statement at the beginning of the higher-court hearing, before the verification procedure. Here the choice of strategies is influenced by various factors, including the givenness and diffuseness of the reported information. In some cases, the report is said to be an account of the trial; it is then not depicted but merely referred to by means of a (12).13 Presumably the message did not have to be recounted because it was more or less equivalent to the written transcript.
198
R V
(12) I postavja sud(’)i [oboix ist’cov pered and place- judges-. [both litigants]-. before ospodarem peredo [knjazem Ondreˇem Vasil(’)evicˇ[em], [sud sovereign- before [prince Andrej Vasil’evicˇ]- [trial svoi skazali i spisok položili. I [knjaz(’) .]- tell-. and copy- put-. and [prince Ondreˇi Vasil(’)evicˇ(’), vozreˇv” v spisok, vyslušav sud, Andrej Vasil’evicˇ]- look- into copy- hear- trial- vsprosil [oboix” istcev: Takov li vam ask-. [both litigants]-. such- you-. sud byl, kakov spisok? trial- be- what.sort- copy- And the [trial] judges, having placed both litigants before the sovereign prince Andrej Vasil’evicˇ, related their trial and put the [trial] copy [into evidence]. And Prince Andrej Vasil’evicˇ, having looked at the [trial] copy [and] having heard [their account of] the trial, asked both witnesses: “Did you have a trial like the copy?” (ASÈI 3: 88, no. 56).
This procedure may have originated before the extensive use of writing in medieval Russian courts; its orality is sometimes highlighted, as in (12), by a reference to the higher judge’s having listened to it. The consists of the preterit of s”kazati with a textual conveyor or speech-event label as direct object (either sud” ‘trial’ or doklad” ‘judicial referral’). In several other cases, the occurs without any report of the verification.14 The use of a in this contextualization conforms to the general preference for that strategy with procedural statements, which did not have to be spelled out for the intended interpreters. By contrast, the message of the trial judge’s statement is specified, at least in part, when it has a direct bearing on the higher judge’s agenda. All of the examples involve some nonfeasance on the part of one of the litigants — in particular, not showing up at the higher-court hearing, in defiance of the trial judge’s instructions. In medieval Russian law, this was grounds for a default judgment in favor of the other litigant (see Kleimola 1975: 69, 89–90). Such reports are attested in both and . The earliest example (13), in a trial transcript of the early 1460s, has a introducing a complementized multiclausal stretch of , i.e., it seems to represent a compromise with the strategy seen in (12):
R J-R H
199
(13) I [Dmitrei Davydovicˇ(’), postavja [pristava Petelju and [Dmitrij Davydovicˇ]- place- [constable Petelja nedeˇlšcˇika i [Sidora istca pered [knjazem Andreˇem bailiff]- and [Sidor plaintiff]- before [prince Andrej Vasil(’)evicˇem, [sud svoj skazal, i [srok tot Vasil’evicˇ]- [trial .]- say- and [term that]- skazal, cˇto [poselskoi krutickoi starec’ Davyd vzjal, say- that [steward of.Krutica elder Davyd]- take- g(o)s(podi)ne, srok sobeˇ takov… a poselskoi, lord- term- . such- and steward- g(o)s(podi)ne, krutickoi na [tot srok sam ne lord- of.Krutica- on [that term]- self- stal peredo mnoju otveˇcˇjat(’) i [igumena stand- before me- respond- and [abbot svoeg(o) ne postavil i gramot ne položil. .]- put- and writ-. place- And Dmitrij Davydovicˇ, having placed the constable Petelja the bailiff and Sidor the plaintiff before Prince Andrej Vasil’evicˇ, related his trial and related that term, that “the Krutica steward, the monk Davyd, accepted, lord, such a term… and the Krutica steward, lord, did not appear at [the end of] that term before me to respond himself, and did not produce his abbot or put any documents [into evidence]” (ASÈI 3: 85, no. 55).
Here the s”kazati has two direct objects — sud” ‘(on-site) trial’ and s”rok” ‘term (to reappear after an adjournment).’ The first is certainly a speech-event label; this may be true, at least metonymically, of the second, since the “term” is conveyed as a directive of the judge in the trial hearing (ibid.). The noun s”rok” is followed by the complementized ; the complementizer cˇ’to can be interpreted either as a relativum generale dependent on s”rok” or an explicative conjunction introducing an epexegetical construction (explanatio dicti superioris; see Potebnja 1899/1958–85, 3: 30–13). If the first, the is actually a fade-in or a “partial quote” (Thompson 1996: 513) inset in and overlapping with the narrative; cf. Clark and Gerrig’s (1990: 791) “incorporated quotations”, which “depict, but what they depict is simultaneously appropriated for use in the containing utterance. They both
200
R V
demonstrate and describe…” While complementized is not unknown in trial transcripts, in the present case it seems likely that the scribe began the report nondirectly, in accordance with the integrated narrative character of the preceding context, and then slipped into because of the length, diffuseness, and syntactic complexity of the information he was recounting. This would explain why the vocative g(o)s(podi)ne, which almost always appears after the first phonological word (after the enclitics in Wackernagel’s position) or, in later , after the first (thematized) intermediate constituent, is displaced several positions to the right in the first clause of the report, precisely before the repetition of s”rok”, the first element incompatible with a tight relative-clause structure. In , as in other languages, there is a tendency for to fade into in convoluted passages, under the influence of production constraints; the grammatical changes involved in can lead to confusion of multiple third-person referents and an overly complex syntactic structure involving multiple embeddings and other potentially confusing constructions. Moreover, the scribe of (13) has a proclivity for presenting reports in the judicial-referral report as , as seen in the verification statement of the remaining litigant and the corroborative testimony of the constable. (The fact that the transcript dates from the 1460s may also be significant, since is found in the earliest examples of the verification procedure.) Of the remaining instances of preliminary reports in the judicial-referral record, one is conveyed in a more ordinary fashion as uncomplementized after tak(o) rek”, like testimony in the trial section; like (13), it involves a lengthy exposition of how one of the litigants’ failed to return after an adjournment (ASÈI 1: 503, no. 604). The other two reports are complementized and tagged by s”kazati. There is no way to determine whether they are , since they contain no elements marked for person that refer to the represented speakers; however, ipso facto they are more homogeneous with the narrative than , and they lack features that could define them specifically as that strategy. In both examples, the preliminary statements are conveyed in considerably less detail than in (13), and verification procedures are not recounted; the use of grammatically dependent, nondirect speech may in fact be a concomitant of the abridged style of exposition. While the first such report (ASÈI 3: 462, no. 477) consists of only a single clause, the second (AFZX: 234, no. 261) is multiclausal, like (13); however, the reported message conveys information that has already been given in the immediately
R J-R H
201
preceding narrative (ibid.). Givenness militates against , since it obviates the need to go into diffuse detail; is most functional when the interpreter is confronted with the reported information for the first time.15 Two other factors are pertinent in both cases: the trial judges’ reports serve as orientation (background), a function that does not favor ; and the avoidance of preserves the continuity of the perspective, so that the narrative of the verdict follows the trial judges’ reports without any gear-shifting (cf. Page 1988: 33). All of these motivations conspire to favor a grammatically dependent, semantically compact strategy that presents reports homogeneously with the narrative. In sum, reports of the lower judges’ reports are subject to sometimes conflicting tendencies. On the one hand, the reports were intended as background information; on the other, they could be perceived as testimony, since the erring litigants’ nonappearance or noncompliance was tantamount to a confession. Thus there was a certain tension between the need to condense background information, which is characteristic of trial transcripts as a text-kind, and the need to expound new information at a level of detail commensurate with its potential importance for the verdict. When the trial judge’s statement adds nothing to the written account, it is simply alluded to without any detail (12). When it repeats what has already been stated in the portion of the text that describes the interlude between the on-site trial and the higher-court hearing, it is grammatically integrated into the narrative; this recalls the way statements made in the interlude itself are treated (see 4.4–4.5). When it provides new information, the form of the seems to depend on the complexity of the message; the longer and more convoluted the report is, the greater the preference is for (13). These tendencies follow the general patterns for in the judicial-referral record: statements that are perceived to be new and/or potentially probative are presented more directly, as a rule, than statements that merely corroborate what is already known from the context or that provide information viewed as mere background for what follows.
C 7 Layered reports
7.1
Preliminary remarks
The possibility of layering — forming reports of reports — may be a universal of represented speech, inherent in its ability to refer beyond the given situation (Clark 1987: 15–18; Goffman 1981: 3, 151). The recursiveness of has special ramifications for judicial procedures grounded on verbal evidence; witness the restrictions on hearsay (which take the function of the reports into account) institutionalized in various modern legal systems (see, for example, Philips 1985: 157–59). Indeed, Morawski (1970: 690) suggests that the use of as a way of establishing truth in law provided the first impetus in Western society for the now prevalent perception that citations should, at least ideally, be verbatim. However, this ideal of verbatim speech is inseparably linked not only with the law but also with the spread of written record-keeping, for only with writing can the form of reports become subject to the intense scrutiny that necessitates careful fidelity to the original (Goody 1977: 118; Coulmas 1986b: 10–11) — and even this is only possible, or at least verifiable, when the original itself is also written (Sternberg 1982a). Medieval Russian legal writing still preserved a high degree of orality (see Birnbaum 1985: 167). The trial transcripts studied include numerous reports, conveyed in various ways, within the represented dialogue of the participants in the trial hearing. It must be stressed that this kind of layering provides direct evidence only for the usage of the scribes, given the imprecision inherent in the recording methods of the pre-technological era. Little can be said, apart from pure speculation, about the status of the in the oral context of the trial, although presumably some of its functions were the same as in written record. Therefore, even though there is no particular reason to doubt that the
204
R V
given reports took place, it would be risky, at best, to attribute their form, including the choice of reporting strategy, to the represented reporter, whose statements, after all, are only known through the filter of the writer’s language. In this sense, despite their distinct epistemological status, the additional layers of in trial transcripts present methodological problems similar to the ones posed by the in fiction, where “there is no direct ‘original’ prior to or behind an instance of [] or []” and where “the supposedly ‘derived’ utterances are not versions of anything, but themselves the ‘originals’ in that they give as much as the reader will ever learn of ‘what was really said’” (McHale 1978: 256; see also Chatman 1975: 221–22). The layered reports found within the trial dialogue are attested in many of the same strategies as the questions and answers attributed in the authorial narrative. However, the distribution of the strategies is quite different — indeed, almost in an inverse proportion. This undoubtedly reflects the contextual diversity and distinct function of the additional layers of speech, as compared with the quotations in the narrative. Because speech within speech can convey a spectrum of verbal interactions limited only by the inventory of speech genres in medieval Russian society, it represents a much wider range of types than the testimony that contains it. Even though the scribes provided minimal background and avoided evaluations in their exposition of the trial hearing, they depicted the participants speaking “for themselves”, without any constraint on what they could say — presumably as a reflection of reality. Since representations of the participants’ statements can contain not only complicating action clauses but also extensive orientation, reports of current states of affairs, and evaluations of events (Labov 1972: 363–65; Grimes 1975: 51–64), it follows that third-person narrative is not the sole or even principal context in which additional layers of can be found. First-person narrative, projections in the future, commands, and other frames are also attested. The functional differences between the record of the testimony and additional layers of are manifest. As already observed, the trial dialogue is intended to provide the judges with ample and unbiased information, in order to facilitate their decisions; thus it is generally presented in as neutral and comprehensive a form as possible. By contrast, the chief function of the within is rhetorical; it is often intended to persuade — to advance or discredit one side of an argument. Unlike the authorial narrative, it always takes a definite, personal stance towards the events that it describes. This
L R
205
subjective perspective, along with the discrepancies of context outlined above, explains why the reporting strategies are distributed differently in the within and in the testimony.
7.2
Additional layers of direct speech: narratives and dyads
Although is by far the most common strategy for reporting the participants’ statements in the trial dialogue, it is rarely used with additional layers of within those statements. Only twelve clear examples are attested (two in variant transcripts from the same trial), as against more than sixty cases of indirect or nondirect speech. Whatever other factors were present, it can be assumed that the need for clarity militated against additional layers of , which in scriptio continua could be misinterpreted as the onsets of new turns at speaking in the trial dialogue. This was especially true when the layered reports were tagged with tak(o) rek(li) and contained the vocative g(o)s(podi)ne. Another disincentive was the difficulty in keeping track of attributions and deictic pivots when multiple perspectives are present; this undoubtedly accounts for the total absence of additional layers of , a strategy largely dependent on predictability of attribution. The layered direct reports share several traits with the direct reports in the trial dialogue. For example, is privileged with reports in narrative sequences, especially when the represented is “hyperforegrounded” as a narrative peak or a crucial point in a line of argumentation. Thus is preferred for the responses in dyadic or quasi-dyadic exchanges, and is sometimes also used for the initiating ’s in such dyads. Dyadic responses function, of course, as the peaks of minimal narratives. All these factors can be seen in the two cases in which layered reports are presented in accordance with the conventions for testimony. Both examples occur within narrative accounts of legal or quasi-legal proceedings; they are presented as responses in dyadic exchanges with third-person speakers. In general, the configuration of the participants in the represented speech events resembles that of the participants in a trial hearing. In the first case, the layered report is a rebuff to a warning that the represented reporters officially witness (1). The onset of the dyad is presented as tagged by the izveˇcˇati ‘warn, notify (of misconduct)’ (on this , see Zaliznjak 1986a: 177):
206
R V
(1) I tot, gospodine, Oksenko [tomu Dmitroku i ego and that- lord- Oksenko- [that Dmitrok]- and his tovarišcˇom izvecˇjal ot poselskogo: To, brat[’]e, fellows-. warn- from steward- brethren- les secˇete [monastyrskoj Dubrovinskoj, ne forest- cut-.2. [of.monastery of.Dubrovino]- dokladyvaja [poselskogo troickogo Nikona… I tot, inform- [steward of.Trinity Nikon]- and that- gospodine, Dmitrok, i Timoška, i Merinec, i lord- Dmitrok- and Timoška- and Merinec- and Xaritonko tako rkli: To les Xaritonko- thus speak-. forest- secˇem [knjazja velikogo, a ne manastyrskoj, a cut-.1. [prince grand]- and of.monastery- and secˇi nam ego [vse gody i voziti iz nego cut- us- it- [all times]- and cart- from it- drova wood-. “And that Oksenko, lord, gave that Dmitrok and his companions warning from the steward: ‘It’s the monastery’s forest of Dubrovino you’re cutting, brothers, without notifying the Trinity steward Nikon…’ And that Dmitrok, lord, and Timoška and Merinec and Xaritonko spoke in this way: ‘It’s the grand prince’s forest we’re cutting, not the monastery’s, and we can cut it at any time and cart wood from it’” (Kaštanov 1970: 358, no. 9).
In the second case (ASÈI 1: 400, no. 523), the represented exchange is part of a murder investigation. Here the narrative framework virtually duplicates the authorial narrative: the dyad begins with direct question after the preterit of s”prašivati, an imperfective of one of the usual verbs of inquiring, s”prositi (on the choice of aspect, see 5.2). This question has no relevance for the lawsuit except as a frame for contextualizing the response, in which the represented speakers are disclaiming any legal responsibility to identify the criminal because their land (the property disputed in the lawsuit) belongs to a different district than the one in which the murder took place. (In Muscovite law, a bloodwite was exacted from districts in which murders had
L R
207
been committed when “the community either failed to identify the murderer or refused to yield him once discovered”; Kaiser 1980: 68.) This report, which is tagged like testimony, has evidentiary value, because it shows that potential witnesses for the respondents once made an official deposition that the contested land was the property of the plaintiff. In both examples of layered tagged by tak(o) rek(li), the reporters are bystanders rather than active participants in the exchanges that they depict. In (1), the witnesses were only overhearers, not the addressees of the retort addressed to their leader Oksenko, while the reporters in the second case make no mention of their own role in the interrogation. Thus the reporters have a low degree of involvement; their status in the represented exchanges resembles the recording scribe’s detached, observational role in the trial hearing. The difference between such detached, disinterested reports and more involved accounts emerges with clarity when the testimony of Oksenko’s witnesses in (1) is compared with Oksenko’s own version of the same exchange (2), with a mixture of first- and third-person reports. As depicted in the trial transcript, Oksenko is the transmitter of the warning (and hence, from his own viewpoint, the representative of the law); thus he is highly involved both in the events recounted in his testimony and, as plaintiff, in the trial itself. Unlike his witnesses, Oksenko was not in a position to observe impartially his own actions or those of his opponents; as his goal is to persuade the judge, he has no reason to report statements with any degree of objectivity. (2) I jaz im, gospodine, izvecˇjal ot poselskogo, and I- them- lord- warn- from steward- cˇtob iz [manastyrskogo lesu iz Dubrovinskogo that- from [of.monastery forest]- from of.Dubrovino- [secˇenyx drov ne vozili, a bole togo by ne [cut- wood]-. cart-. and more that- sekli; ini, gospodine, skazyvajut to cut-. others-. lord- say-.3. les ne troickoj ne monastyrskoj, a forest- of.Trinity- of.monastery- and zovut [tot les monastyrskoj Dubrovinskoj call-.3. [that forest of.monastery of.Dubrovino]-
208
R V
[knjazja velikogo lesom, a molvjat: secˇ’ [prince grand]- forest- and say-.3. cut- nam ego ežoden’ i vozit’ iz nego drova. us- it- every.day and cart- from it- wood-. “[1] And I, lord, gave them warning from the steward that they not cart cut word from the monastery’s forest of Dubrovino and [that] they no longer cut [wood]. [2] They, lord, say that forest [is] not Trinity Monastery’s, [3] and call that Dubrovino forest of the monastery’s the grand prince’s forest, [4] and say, ‘we can cut it every day and cart wood from it’” (Kaštanov 1970: 355, no. 9).
The involved, evaluative treatment of in (2) is signalled, inter alia, by the switch from the narrative tense [1] (which is adhered to strictly in (1)) to the historical present [2–4]. Cross-linguistically, the historical present functions as an evaluative, involvement-creating device; it tends to co-occur with another actualizing device, (see Schiffrin 1981: 58, 60; Chafe 1994: 218). In (2) it serves to single out the remarks ascribed to Oksenko’s opponent as the central portion of the charge — the part pertinent to the ongoing trial. Another evaluative device is the subdivision of the woodcutters’ retort. Whereas in (1) it is conveyed as a single report with throughout, in (2) it is analyzed into three separate ’s [2–4] (“they say… and call… and say…”). Oksenko is depicted picking apart his adversaries’ statement rather than giving them an uninterrupted say. The claims in the first two clauses are glossed over by the use of nonintrusive reporting strategies — nondirect or fused after the s”kazyvati (the imperfective of s”kazati) [2] and a after z”vati ‘call, claim’ [3]. Within the [3], the illegality of the woodcutters’ position is highlighted by the semantic anomaly of juxtaposing the highly presuppositional phrase “that Dubrovino forest of the monastery’s” with their contradictory claim (“the grand prince’s forest”). The woodcutters are allowed to speak for themselves, as it were, only in the last of the three reports [4], given in after the m”lviti (underlined). The transition from nondirect and integrated strategies [2–3] to diffuse [4] acts as a build-up to the climax — the core of Oksenko’s accusation, in which the woodcutters assert not only their right to the land but also their intention to despoil it regularly for their own gain. Thus the shift of perspectives highlights what from Oksenko’s viewpoint are aggravating circumstances. The in [4] does
L R
209
double duty, not only actualizing crucial evidence but also disassociating the reporter from responsibility (Clark and Gerrig 1990: 792; Sternberg 1982b). A third evaluative device in (2) is the use of the m”lviti (molviti) [4], which also tags seven other cases of layered in the transcripts. Like recˇi, this verb tends to appear in dyadic or at least reactive contexts, but it seems to have a special emotive force. Often, as in (2) [4], statements tagged by m”lviti do not resolve the dyad in a way expected by or acceptable to the reporters (cf. Longacre 1983: 73); they are uninformative, otherwise uncooperative, or unsuccessful. This connotation emerges clearly in some of the other examples of layered . In one case, the represented reporters had petitioned the steward and received his promise of a deed for the property under contention — a promise that he did not live to fulfill. While the petition is given as tagged by a , the steward’s promise is given as introduced by m”lviti (ASÈI 3: 187, no. 173). In three other cases, litigants are depicted defending themselves against the charge of having “kept silent” — being negligent in prosecuting their claim — by telling how the grand prince responded to repeated petitions with promises that have come to nothing. The grand prince’s statements, conveyed as tagged by m”lviti, have failed to have their expected perlocutionary effect (3).1 (3) Bivali esmja, gospodine, cˇelom [velikomu hit-. ..1. lord- forehead- [grand knjazju ne odinova, i [knjaz’ velikij tak molvil: prince]- once and [prince grand]- thus say- Kak poedet [moj pisec Pereslavlja pisati, i when come-.3 [my clerk]- Pereslavl’- write- and vy emu ukažite [moi zemli, i you-. him- show-.2. [my lands]- and on mne skažet, — i pisec v Pereslavle he- me- tell-.3 and clerk- in Pereslavl’- davno ne byval. long be- [1] “We petitioned, lord, the grand prince more than once, [2] and the grand prince spoke in this way: ‘When my clerk comes to write up Pereslavl’, you can show him my lands, and he will tell me’; and the clerk hasn’t been in Pereslavl’ in a long time” (ASÈI 1: 452, no. 571).
210
R V
The frustration of reasonable expectations is very much at issue here. Litigants could lose their cases if they failed to protest in what the judges felt to be a timely manner or if they otherwise delayed initiating proper legal action; for example, the Sudebnik (Lawcode) of 1497 set three- to six-year statutes of limitation for bringing suit in property disputes (see Kleimola 1975: 30–33, 87–88). Thus the grand prince’s statement [2] is reported as evidence of the reporters’ good faith. Tagged is well motivated, given that the report is dyadic, foregrounded, and evidentiary; the switch of perspectives and modality holds the attention and hence makes the information salient, like the cataphoric adverb tak(o), a focusing device. The use of a or nondirect form in (3) [1], as also in (2) [1], for the first half of what may be called an “unequal dyad” is typical of self-quotations (see 7.7). The reporters downplay their own remarks, which are treated as mere background for the that is the focus of the discourse, by using reporting strategies that interrupt the flow of the discourse less than . As proprietors of their own remarks, they are free to rework them in accordance with their knowledge of their original illocutionary intent without compromising their truthfulness (which, in the context of the trial, can be equated with their verifiability). The emotive nuances conveyed by m”lviti can also appear in nondyadic narrative contexts, as in the continuation of (3), where the represented reporters are the receiving end of a warning like the one in (2) [1] (ibid.).2 Here m”lviti, which introduces the warning, conveys a negative, “frustrative” evaluation of the plaintiff’s , which from the reporters’ viewpoint is unexpected and unjust. A further instance (4), found in the testimony of warrant officers in a criminal trial, seems to connote unexpectedness rather than a negative evaluation, but it nonetheless reflects a high degree of involvement in the discourse. (4) K nam, gospodine, pritekl [tot Fetko Poluev rano na to us- lord- run- [that Fedko Poluev]- early on [Ustreten’ev den’, a molvit tako: Ukradeno, [of.Presentation day]- and speak-.3 thus steal- gospodine, u nas nocˇes’ s lugu stog sena, lord- at us- tonight from meadow- stack- hay- tricat’ kopen, a sled, gospodine, pošol v thirty- shocks-. and trail- lord- go- to
L R
211
[Pavlovskoe prisel’e, i vy, gospodine, poedte [of.Pavlovskoe settlement]- and you-. lord- go-. s nami na sled. with us- to trail- “That Fedko Poluev, lord, rushed to us early on the Feast of the Presentation, and he speaks in this way: ‘Last night, lords, a haystack was stolen from our meadow, thirty shocks, and the trail, lord, goes to the settlement of Pavlovskoe; and you, lords, come with us on that trail’” (Kaštanov 1970: 411, no. 40).
This passage recreates a vivid scene, in which one of the plaintiffs runs in clamoring for the officers to come while the trail is still hot. There are several other signals of high involvement, including the historical present at the narrative peak (cf. Schiffrin 1981; Longacre 1985) and the verb pritecˇi ‘rush in’, which is unusually specific in the characteristically restrained style of trial transcripts. In sum, the basic contextualization for additional layers of is narrative recounting pretrial confrontations; dyads with third-person speakers are an especially privileged environment. All of the examples are foregrounded in their narrative sequences either as important, sometimes unexpected evidence or as frames for such evidence. Thus the use of in additional layers of reporting recalls its function in the authorial narrative. Indeed, the two of the examples that are representations of prior legal procedures are treated precisely like standard testimony. Though foregrounding is not an exclusive prerogative of , it is a typical motivation, since the diffuse form and perspectival shifts of that strategy tend to single it out from its context and actualize the discourse. By injecting into the current speech situation the reference points or other elements of a different speech situation, causes a change in footing (Goffman 1981: 126, 128) that interrupts and slows the pace of the narrative and thereby creates salience. All other things being equal, reports in tend to be more intrusive and hence more prominent in the discourse than reports in or other strategies. (On the salience of the first and second persons as compared with the third, see also Chvany 1990/1996: 288; Wallace 1982: 212–13, 215.) In general, additional layers of tend to occur in evaluative contexts, with the represented reporters controlling the interpretation, as it were. This speaker-orientation (Lakoff 1984) may have been something of a disincentive
212
R V
for , which puts the interpreter in control as a witness able to draw personal conclusions about the statement thus depicted. While this “directexperience” effect (Clark and Gerrig 1990: 793) could potentially create empathy, this risk could be offset by the use of evaluative introducers like m”lviti and by a careful selection of what to report from direct perspective. The zeroing-in, foregrounding effect of could conspire with presupposed negative evaluations to create distancing rather than empathy, holding, as it were, the given under microscopic scrutiny. The distribution and foregrounding effect of additional layers of can be readily compared with the usage in the modern American trial transcript studied by Philips (1985). According to Philips (ibid.: 154), (“quoting”) signals a high degree of relevance — “information being presented as evidence directly related to proof of the elements of a criminal charge” (see also 162, 168). By contrast, background information and other parts of the testimony that are perceived as less relevant (“central”) to the outcome are presented transcript as ’s (“topic naming”) or (“reports of substance”; ibid.: 160–62).
7.3
Complementized indirect speech: rejoinders, hearsay, and hypothetical reports
The most common way of conveying reports within the trial participants’ statements is by the use of complementized indirect or nondirect speech tagged by forms of the s”kazati or its imperfective partner, s”kazyvati or, in a few instances, by other verbs. There are 92 examples of this strategy in the trial transcripts investigated (or 93, given an ambiguous example). In most cases, the complementizer is the explicative conjunction cˇ’to; the synonymous conjunction koe is also attested in a single transcript, where it is the only complementizer used (ASÈI 1: 419, no. 540). The contextualizations in which complementized is preferred fall into several speech genres. Twelve examples occur as preambles to judges’ questions and function as requests for amplification of the immediately preceding remarks or previous testimony. In such cases, the queries are intended to clarify the content of the , to resolve apparent contradictions, or to pursue lines of inquiry stemming from the reports. Six of the examples are tagged with s”kazyvati in the present tense (5).
L R
213
(5) Sami skazyvaete, cˇto [tret(’)ego leˇta selves-. say-.2. that [third year]- pokosili este [to selišcˇo, a dotudy mow-. ..2. [that village]- and until.then ix xto kosil? them- who- mow- “You yourselves say that you have mowed that village for three years, but who mowed it before that?” (ASÈI 2: 520, no. 481).
There are also six cases involving other tag verbs, also in the present tense — the ’s kazati and govoriti and the ’s z”vati ‘call; claim (as/that)’ and nazyvati ‘name; call; claim (as/that)’.3 While kazati seems to be synonymous with s”kazyvati, it tends to be used more in intercalated (parenthetical) rather than in preposed tags (see 7.6). The case of govoriti (cf. govorit’, the basic imperfective ) is one of only two examples in the corpus of this verb tagging a report; this may reflect a dialectal difference or an idiosyncratic linguistic preference.4 Its use here seems comparable with that of s”kazyvati (cf. the aspectual pairing of govorit’ with perfective skazat’ in ); however, govoriti is more typical in contexts that focus on the occurrence of the speech as a linguistic action. Thus it seems to be the only possible after phase verbs, which denote how the speech act is implemented (cf. examples in legislative texts, e.g., AAÈ 1: 164, no. 187); it also seems to be preferred for denoting linguistic action in which the message is omitted or presented only as a pronoun, textual-conveyor noun, or prepositional phrase (e.g., Kaštanov 1970: 359, no. 9). The related ’s z”vati and nazyvati occur much more frequently as ’s in double-accusative or accusative-instrumental constructions (see, e.g., Antonov and Baranov 1997: 117, no. 146 (3×)). They are openly evaluative, in that they rephrase the litigants’ assertions as quasi-performative speech acts and thereby draw attention to their potentially counterfactual value. This connotation explains why, both as ’s and as tag verbs, z”vati and nazyvati are virtually restricted to the judges’ questions and to litigants’ disparaging references to the claims of their opponents. Another contextualization that privileges complementized or nondirect speech after the present of s”kazyvati is when the judges ask litigants to respond to the statements of other participants (6). This speech genre, which
214
R V
is attested in five examples, resembles requests for amplification like (5) and had the same basic motivations in the actual trials.5 (6) [Vaši starožilcy skazyvajut, cˇto [te [your- old.residents]-. say-.3. that [those zemli paxali na monastyr’ let s lands]-. plow-. for monastery- years-. about šestdesjat’, i vy o cˇem monastyrju za sixty- and you-. for what- monastery- for tolko let molcˇali, i [velikomu so.many- years-. be.silent-. and [grand knjazju este o [tex zemljax prince]- ..2. about [those lands]-. bivali li cˇelom? hit-. forehead- “Your witnesses say that they farmed those properties for the monastery for some sixty years. Why did you go so many years without complaining to the monastery? And have you petitioned the grand prince about those properties?” (ASÈI 1: 540, no. 628).
If reports of this kind are compared with their antecedents in the trial record, they turn out to be far from mechanical transmissions. The messages are greatly condensed, with omission of perspectival information like the vocative g(o)s(podi)ne; in addition, they show the inevitable changes of meaning that result from recontextualization. This is epitomized by (6), in which the judge, as depicted in the transcript, subverts the illocutionary intent of the testimony; the witnesses had evidently wanted to inculpate the opposing side’s supporters but unwittingly unmasked the nonfeasance of their own principals. Subversion, indeed, is the function of fourteen other cases that challenge the statements of the opposing party. In these examples, the reports are adduced exclusively as material to be refuted. The verbs in the tag clauses are mostly in the present tense, as with the requests for amplification and for reaction discussed above.6 (7) A cˇto, g(o)s(podi)ne, skazyvaet [cˇernec’ Kirilo, cˇto and that lord- say-.3 [monk Kirilo]- that
L R
215
[teˇ pustoši dal [Danilo Blin, a skazyvaet [those fields]- give- [Danilo Blin]- and say-.3 sem(’) leˇt pustoši kosili, to, g(o)s(podi)ne, seven- years-. fields- mow-. lord- lžet Kirilo; a ne kašivali, g(o)s(podi)ne, lie-.3 Kirilo- and mow-. lord- [teˇx pustošei do [seˇx meˇst”. [those fields]-. to [these places]-. And, lord, as for the fact that the monk Kirilo says that Danilo Blin gave those fallow fields, and says [they] have used those fields for seven years, Kirilo is lying, lord; they haven’t yet, lord, used those fields” (ASÈI 2: 315–16, no. 333).
Besides the sixteen cases after s”kazyvati, there is one case tagged by the present of the govoriti (ASÈI 2: 373, no. 375) and one by the present of the negative-presuppositional l”gati ‘lie’ (AFZX 127, no. 140), which is more common without a report. One can conclude, in general, that presenttense tags are typical with rejoinders, i.e., with variously motivated reports of statements made or defended during the trial hearing. Nevertheless, there are a number of cases tagged by preterit verbs of both imperfective and perfective aspect. These seeming exceptions actually index differences in the character of the layered reports. For example, the reports tagged by the preterit perfective are accompanied by information about the participation framework, so that the situation of represented speech acts is brought into focus rather than ignored as in (5)–(7). This can be seen in two cases, which occur in a single passage in which the plaintiff casts doubt on the honesty of his opponents (8). One of the reports is presented as complementized [1], the other as an epexegetical construction involving complementized nondirect speech after the textual-conveyer noun znaxar’stvo ‘testimony (of witnesses)’ [2]. (8) Sam, g(ospo)dine, [knjaz(’) Ivan skazal v otveˇte self- lord- [prince Ivan]- say- in response- pered toboju, cˇto [to selišcˇo Zelenevo pašet before you-. that [that village Zelenevo]- plow-.3 poltret(’)jatcat(’) leˇt, a znaxori, g(ospo)dine, twenty-five- years-. and witnesses-. lord-
216
R V
[znaxar’stvo svoe skazali pered toboju, cˇto [testimony .]- say-. before you-. that [tu zemlju Zelenevo selišcˇe pašet [knjaz(’) [that land Zelenevo village]- plow-.3 [prince Ivan tritcat(’) leˇt… ino, g(ospo)dine, eg(o) [znaxorju Ivan]- thirty- years-. lord- his [witness Gridkeˇ Alekseˇeveu tritcati leˇt ot rodu Gridka Alekseev]- thirty- years-. from birth- neˇt… obyšcˇy leˇt eg(o), da postavi eg(o) be-. seek- years-. his and put- him- pered [velikim kn(ja)zem. before [grand prince]- “[1] Prince Ivan himself, lord, said in [his] response before you that that he has farmed that village of Zelenevo for twenty-five years; [2] and [his] witnesses, lord, gave their testimony before you, that Prince Ivan has farmed that land of Zelenevo village for thirty years… [3] But look, lord, his witness Gridka Alekseev is not [even] thirty years old… [4] Find out his age and place him before the grand prince” (ASÈI 1: 507–8, no. 607; similarly in the variant transcript from the same trial, 513, no. 607a).
The underlined reports are subtly different from the challenges exemplified by (7), in that the represented reporter, the monk Isaja, is trying not so much to refute the statements as to discredit the speakers by proving that they have broken the rules of the trial. Thus he sets up an inferential relation between [1] and [2] to expose the five-year discrepancy in the opposing side’s claims; he contradicts the credibility of [2] directly by adducing counterevidence in his own voice [3]. It is noteworthy that both tag clauses contain seemingly pleonastic references to the legal roles of the attributed speakers (“in [his] response” [1]; “their testimony” [2]) and to the judge’s status as official witness of the testimony (“before you”).7 Thus the reporter is purposefully situating the ’s in the trial context to emphasize the specifically legal nature of their misconduct. The use of perfective preterit tag verbs creates a mini-narrative that foregrounds the fact that inconsistent and false testimony have been offered in the trial — a focus that culminates in Isaja’s request that the judge prove the perjury and take the culprit before the grand prince
L R
217
for judgment [4]. The content of the false testimony, while relevant to the charge, is presupposed and presented as a matter of lesser concern through the use of devices that integrate it into the narrative — complementizers and . Isaja has shown his enemies to be liars and has no need to dispute their words as such. In other cases, the layered reports tagged by the perfective preterit, which convey statements made in earlier legal procedures or confrontations, are situated in narratives (fourteen instances in four transcripts).8 These indirect reports differ from direct ones tagged by tak(o) rek(li) or m”lviti (see Section 7.2) in that they are nondyadic, and there is no special focus on the reported information per se. The narrative in which the ’s occur is (unlike the authorial framework) not primarily a sequence of ’s. The layered reports serve as background; with the surrounding narrative, they set the stage for that part of the testimony directly relevant to the case at hand. For example, in one of the cases, a witness for the respondents is recounting the results of an obysk, an investigation conducted by collecting depositions from reputable members of the community (9); however, the underlined report is not the ultimate point of his statement but rather an explanatory preamble whose purpose is to bring out the full significance of the more recent conflicts that prompted the lawsuit.9 (9) vylgal sebeˇ u Timofeˇja u Mixailovicˇja get.falsely- . at Timofej- at Mixajlovicˇ- [gramotu lgotnuju na [tu rozseˇcˇ(’), a skazal, [writ of.privilege]- for [that clearing]- and say- g(ospodi)ne, to leˇs [velikog(o) kn(ja)zja… I my, lord- forest- [grand prince]- and we- g(ospodi)ne, bili cˇelom [Timofeˇju Mixailovicˇju; lord- hit-. forehead- [Timofej Mixajlovicˇ]- i Timofeˇi… obyskal [ljudmi dobrymi [vseju and Timofej- seek.out- [people good]-. [all volostiju, cˇto, g(ospodi)ne, Onikei da Senka skazali region]- that lord- Onikej- and Senka- say- Timofeˇju, cˇto [tot Savka popaxal [našu zemlju Timofej- that [that Savka]- plow- [our land monastyrskuju kocˇevinskuju; i Timofeˇi, g(ospodi)ne, of.monastery Kocˇevinskaja]- and Timofej- lord-
218
R V
[gramotu svoju lgotnuju u neg(o) vzjal [see [writ . of.privilege]- at him- took- [this oseni, a zemlju… nam otdal… A see, g(ospodi)ne, fall]- and land- us- give- and this- lord- oseni… on, g(ospodi)ne, [tu izbu pocˇal staviti fall- he- lord- [that house]- begin- build- cˇerez [Timofeˇevo slovo… across [Timofej- word]- “… [Savka] obtained, under false pretenses, a writ of privilege for that clearing from Timofej Mixajlovicˇ, and said, lord, that [was] a forest of the grand prince… And we, lord, petitioned Timofej Mixajlovicˇ, and Timofej… found out, by questioning reputable witnesses representing the entire region, that, lord, Onikej and Senka [the witnesses] told Timofej that that Savka had farmed our land of Kocˇevinskaja, belonging to the monastery. And Timofej, lord, took his writ of privilege from him this fall and gave us the land… And this fall, lord, he [Savka], lord, began to build that house [previously mentioned] in defiance of Timofej’s command” (ASÈI 2: 193, no. 285).
Besides the collateral nature of the reported information, there are three additional factors that may have motivated the use of complementized . First, the reported is not on the plotline of main events (“obtained/ claimed… petitioned… found out… took/gave… began to build”) but embedded within one of those events (“found out”). Second, the tag is itself subordinated; the complexity of the frame may have militated against , with its diffuse structure and perspectival shifts. Third, as shown by the choral tag, the report is a synthesis of separate depositions, which were undoubtedly not limited to the content of the underlined message (cf. a case in which multiplex testimony is summarized as a single complementized selfquotation; ASÈI 2: 434–35, no. 411). In the other twelve reports in narrative contexts (ten of them in a single transcript), the perfective tag verb s”kazati conveys plotline events in past legal precedings.10 The choice of was evidently motivated, as in (9), by pacing. When tagged with the perfective, can have the effect of presenting a as a link in a chain of equipollent foregrounded events (fast-paced action, cf. Diver 1969) without making it more salient than the other complicating actions in the sequence. Because of its formal properties,
L R
219
would retard the pace of the narration by singling out the statement per se for special attention. Such a ritardando effect would only be well motivated if the reported information itself was a special focus of the discourse.11 The preterit perfective is also used to tag a further case of layered complementized conveyed with situational specifications (10). This report, which appears in an early-sixteenth-century larceny trial, forms the preamble to a question following up on testimony that the defendant is a known criminal. In Muscovite law, the fact of a prior conviction created a presupposition of guilt (see Kaiser 1980: 91, 139). Like the testimony in (8), the judge’s inquiry, as recorded in the transcript, has the illocutionary intent of exposing a contradiction and does not expect an exonerating reply: (10) Pered [knjazem Ondreem Vasil’evicˇem v [tom dele before [prince Andrej Vasil’evicˇ]- in [that matter]- sud tebe byl li i bit li trial- you-. be- and beat- esi knutom v [toj tat’be so [knjaž..2 whip- in [that theft]- from [princeOndreeva suda? I peredo mnoju esi nynecˇa Andrej- trial]- and before me- ..2 now skazal, cˇto [tem senom tebja podkinuli, a koli say- that [that hay]- you-. frame-. and when u tobja [to seno vynjali, i ty ego at you- [that hay]- take.away-. and you-. it- skazyval svoim, a skazyval esi, cˇto ego say- .- and say- ..2 that it- kupil, i u kogo že ty paki [to seno buy- and at whom- you-. [that hay]- kupil? buy- “[1] Were you tried in that matter before Prince Andrej Vasil’evicˇ? And were you flogged for that theft after Prince Andrej’s judgment? [2] And you have now said before me that you were unjustly convicted in [the theft of] that hay; [3] but when they took that hay away from you, [3a] you called it your own and [3b] said that [you] bought it. [4] From whom was it that you bought that hay?” (Kaštanov 1970: 413, no. 40).
220
R V
Here, as in the layered report in (8), the underlined tag clause includes a topicalized reference to the judge’s role in the hearing (“before me”). This implicitly calls attention to the solemn, binding nature of the act of testifying, which the defendant Gridja has violated. The juxtaposition of Gridja’s past offense [1] with the present charge against him [2] (cf. nynecˇa ‘now’) reflects a presupposition of his guilt. This is made explicit in the statements following the underlined segment [3], in which hearsay ascribed to Gridja by hostile witnesses and denied by Gridja himself is taken as given, as background for a question. The question itself [4] contains the emphatic particles že and paky, which, by signalling repetition, call attention to the fact that a satisfactory answer has yet to be given. The two hearsay reports in (10) — the first conveyed as a [3a] and the second as [3b] — actually represent a third layer of set within , since they are themselves heteroglossic. It is noteworthy than the hearsay is tagged with yet another form — the preterit of the imperfective s”kazyvati, which indexes a different kind of speech event than the present or preterit perfective. This way of indicating hearsay is found in another instance in which a judge is asking his addressees to confirm that they did say what preceding speakers had alleged of them (AFZX 116, no. 125). The statements tagged by the imperfective preterit are repeated from the testimony; as presupposed reports, it is not surprising to find them in complementized . The use of the preterit also has a straightforward explanation. The putative original statements are supposed to have been made before the trial hearing and hence outside of the judge’s (and scribe’s) observation; the messages that they convey have not yet been acknowledged by the speakers to which they have been ascribed, so that the attribution has not become a given, activated as “present” testimony like the statements repeated from the trial hearing in (5)–(7). As for the choice of aspect, it is typical in the transcripts for imperfective preterits to be employed in the judges’ questions with untagged heteroglossic information that is being offered for verification (11). The confirmations themselves can be perfective if that is contextually appropriate. (11) I sud(’)i sprosili Danila: Davyval li and judges-. ask-. Danilo- give- ty… pustoš(’)… I Danilo tak rek: Dal you- field]- and Danilo- thus speak- give-
L R
221
esmi… pustoš(’)… ..1 field- And the judges asked Danilo, “Did you donate []… the field…?” And Danilo spoke in this way: “I gave []… the field…” (ASÈI 2: 316, no. 333).
This quasi-evidential use of aspect can be explained by the fact that the actions denoted by the perfective are situated in a definite network of contextual relations, either explicitly, as in (8)–(10), or in the implicit view of the reporter. By contrast, those denoted by the imperfective are treated in a decontextualized manner, and the primary issue under discussion is the occurrence of the actions. (Cf. the case in which the judge adduces an invalidated claim of the plaintiffs as the chief reason for the forthcoming negative judgment; ASÈI 2: 389–90, no. 387.) Another category of hearsay involves statements that litigants or, more frequently, witnesses ascribe to their fathers or grandfathers. These are most often presented (14 out of 16 examples) as complementized indirect or nondirect speech tagged, like other hearsay, by the preterit of s”kazyvati or, in two cases in a single transcript, by slyxati ‘hear’ (AFZX 214, no. 249 (2×)). (Verbs of hearing can act as the converses or relational opposites (see Palmer 1981: 97–100) of ’s, with their focus on the addressees rather than the speakers.) The choice of the imperfective for ancestral hearsay is determined by the fact that the statements are said to have occurred in the unsituated past; the imperfective thus functions as constatation of the , without specifying whether the action was iterative or punctiliar. The focus, indeed, is on the source of the witnesses’ knowledge — i.e., on the represented speaker rather than on the .12 The tag clauses involving s”kazyvati all include first-person forms — either dative pronouns or possessive adjectives — referring to the reporters, who depict themselves as the addressees of the layered report.13 (12) a.
My, g(o)s(podi)ne, slyxali u [otcev” we- lord- hear-. at [fathers svoix; [otci naši skazyvali, cˇto to .]-. [fathers our]-. say-. that zemlja [velikogo kn(ja)zja… land- [grand prince]-
222
R V
“We, lord, heard from our fathers; our fathers said that it [was] the property of the grand prince” (ASÈI 2: 425, no. 407).
The witnesses in medieval Russian trials generally testified on what they “remembered” or “knew” (cf. the term znaxar’ ‘one who knows’); the authority of their statements was thought to increase in proportion with their longevity (cf. the term starožil’c’ ‘longtime resident’; see also Kaiser 1980: 128). This was undoubtedly the social motivation for the ancestral hearsay exemplified in (12a), which added, in effect, several decades to the witnesses’ memory. By reporting a speech event in which they were participants (addressees), the witnesses were actually asserting something about their own knowledge (cf. Townsend and Bever 1977: 5), as can be seen from the dual attribution in (12b): (12) b. a skazyvali nam, gospodine, [otci naši da and say-. us- lord- [fathers our]-. and i my zapomnim, cˇto izstariny [to we- recall-.1. that of.old [that Popkovo tjanet k Biserovu… Popkovo]- belong-.3 to Biserovo- “And our fathers, lord, told us, and also we remember that of old that Popkovo belongs to Biserovo” (AFZX 107, no. 114).
The primary purpose of the witnesses’ testimony, whether it invoked only their own recollections or ancestral hearsay, was to state their own ability to corroborate the claims of the litigants who called them. In keeping with this, the chief element foregrounded in their testimony is not the content of their memory nor (in ancestral hearsay) the actual message of the layered speech, which tends to be presupposed; it is the attribution — their assertion that their knowledge, including what their forbearers told them, bears out their principal. The use of complementized in such contexts is well motivated: the report is presented as ground rather than a more salient (event-like) figure that could vie with the more pertinent information, the attribution, for the attention of the interpreters (see Talmy 1978: 640; Wallace 1979: 214–15). The emphasis on the attribution also explains the inverted word order of the tag clauses in (12b) and most of the other examples of ancestral hearsay: the represented source of the message is asserted in the
L R
223
normal position for new information. (In (12a) the represented source has already been topicalized.) This analysis is confirmed if the cases of ancestral hearsay in complementized after the preterit of s”kazyvati are compared with the sole instance of in this context, which is tagged by the historical present of m”lviti (12c). (12) c.
a [otec moj, gospodine, togdy žil za and [father my]- lord- then live- under mitropolitom v sele v Kulikovskom, i jaz, metropolitan- in village- in Kulikovskoe- and I- gospodine, s [otcem svoim i kosil na lord- with [father .]- mow- on [tex luzex, na kotoryx stoim, i [those meadows]-. on -. stand-.1. and [otec moj… mne molvit tak — nynecˇja my [father my]- me- say-.3 thus now we- [te luzi kosim za [those meadows]- mow-.1. as mitropolicˇi, potomu cˇto ešcˇe ne sel nixto of.metropolitan-. because yet settle- no one- na Parašine, a kak na Parašine sjadut on Parašino-, and when on Parašino- settle-.3. ljudi žiti, i oni to u nas u people- live- and they- at us- at mitropolicˇix [te luzi of.metropolitan-. [those meadows]- ot”imut, zanež to zemlja [velikogo take.away-.3. because land- [grand knjazja Parašinskaja [te luzi… prince]- of.Parašino- [those meadows]-. “… And my father, lord, then lived under the metropolitan in the village of Kulikovskoe, and I, lord, with my father mowed [hay] on those meadows on which we stand, and my father… spoke to me in this way: ‘Now we mow these meadows as the metropolitan’s,
224
R V
because no one has yet settled in Parašino, but when people come to live in Parašino, they will take those meadows away from us, from the metropolitan’s [people], because those meadows are the property of the grand prince, of Parašino’” (AFZX 230–31, no. 259).
There are at least three factors that may have influenced the unusual reporting strategy in this example. First, like the other cases of layered and unlike the other cases of ancestral hearsay, the report is embedded in a narrative. Second, the message of the layered report is unexpected. For one thing, it provides far more information than is usual in ancestral hearsay; for another, it is not a straightforward, presupposed corroboration of the principal’s position. The represented reporter is testifying on behalf of a servant of the grand prince in a lawsuit against the metropolitan’s steward, yet he is depicted testifying that he and his father were servants of the metropolitan at the time of the narrated speech event. Therefore, the underlined report, which ascribes the disputed property to the grand prince, is in effect the admission of a party opponent, unique in the investigated transcripts. This fact, which takes on a devastating significance in the verdict, distinguishes the report from the other instances of ancestral hearsay. The unexpected nature of the additional layer of makes it contextually salient and promotes foregrounding, a function usually performed by . This foregrounding effect, which is a form of evaluation, may also have motivated the use of the historical present in the tag and the choice of the emotively charged m”lviti (see Section 7.2), which may, indeed, have its “frustrative” nuance here. The third probable factor in the choice of in (12c) is the diffuseness and complexity of the layered report. The cases of ancestral hearsay in complementized (12a–b) involve relatively simple clauses with a noncoreferential third-person subject. By contrast, the layered report in (12c) is relatively lengthy and more hypotactic than is typical for the language of trial transcripts. The use of a complementizer would lead to multiple embeddings, while a switch to the narrator’s indirect orientation would have caused referential complications; for example, the phrase i oni to u nas u mitropolicˇix te luzi ot”imut, which includes a coreferential inclusive first-person plural pronoun, would have two competing third-person plural referents. The final category of complementized report that should be noted is hypothetical , which falls into two subtypes. The first is projected statements;
L R
225
these include first-person utterances with the ’s ceˇlovati kr’st” ‘swear [lit. kiss the cross]’ and s”latisja ‘cite, refer to’ (13). The latter denotes ’s in which litigants simultaneously identify witnesses and predict how they will testify or in which they make reference to documentary evidence to which they have no immediate access.14 The present tense is used because the ’s have performative force; they are claims that are pertinent to the ongoing trial, like rejoinders. (13) Vedomo to, gospodine, našie že volosti be.known- that- lord- our- region- xristianom [Jakušu Tokarevu… da [Bornjaku Ontuševu, peasants-. [Jakuš Tokarev]- and [Bornjak Ontušev]- na nix sja, gospodine, i šlem, cˇto to on them- lord- cite-.1. that zemlja [velikogo knjazja. land- [grand prince]- “That is known, lord, to peasants of our region, Jakuš Tokarev… and Bornjak Ontušev; I cite them, lord, that that is the grand prince’s land” (AFZX 223, no. 258).
The imperative of the verb s”prositi ‘ask’ introduces a hypothetical report complementized by cˇ’to in a similar situation, in which the represented speaker asks the judge to ask witnesses about his assertion (ASÈI 1: 492, no. 595). This is essentially the same illocution as in (13). The verb of asking is used factively, as it were: “Ask the hundredman and peasants, (they will say that) that village Kozlovo is the monastery’s property of old”. The second subcategory of hypothetical occurs only in questions and is tagged with preterit verbs. Here the judges are trying to obtain further information; however, unlike in requests for amplification, they are referring not to statements made in the course of the trial but to ’s that the addressees could be expected to have performed before the trial (hence the preterit) in accordance with normative procedure (14a). (14) a.
Bivali li este cˇelom o [toi beat-. ..2. forehead- about [that zemleˇ [g(o)s(u)d(a)rju velikomu kn(ja)zju, cˇto u land]- [sovereign grand prince]- that at
226
R V
vas [tu zemlju pašet [simanovskoi you-. [that land]- plow-.3 [of.Simonov poselskoi [s”] [svoimi xr(e)st’jany silno steward]- [with [. peasants]-. by.force naeˇzdom? from.outside “Did you petition the sovereign grand prince about that land, that the Simonov [Monastery] steward and his peasants were farming that land of yours by force from outside?” (ASÈI 2: 445, no. 414).
The denoted by the imperfective (indeed, iterative) verb bivati cˇelom ‘petition’ is not presupposed; this use of the imperfective aspect conforms to the pattern seen in (11), above. By contrast, the perfective is used with a hypothetical report that is presupposed (perhaps only for rhetorical purposes), which is tagged with the javiti ‘declare to’ (14b): (14) b. i komu esi javil li, koe u and whom- ..2 notify- that at tebja [gramota kupcˇaja zgorela? you-. [writ of.purchase]- burn- “… And to whom did you declare that your purchase deed had burned up?” (ASÈI 3: 269, no. 250).
However, the imperfective preterit is used, as expected, with the presupposed stative verb m”lcˇati ‘keep silent’ (see ASÈI 1: 512, no. 607a and in a variant transcript from the same trial, 507, no. 607a). This verb (which is much more common without a report) bears the negative presupposition that the addressee has failed to initiate requisite redress procedures — typically a strongly prejudicial factor in a judgment. In sum, complementized indirect and nondirect reports are preferred for rejoinders (statements repeated from early in the trial), hearsay from outside the trial context, and hypothetical speech. These contextualizations involve messages that serve as background to the ongoing concern rather than the center of attention; would distract from the focus by interjecting extraneous information such as the reference points of the original speech event — a violation of the Maxim of Quantity (Grice 1975/1989). Most of the messages convey presupposed information, which is processed most efficiently without the explicitness and salience characteristic of .
L R
227
There are several reasons why presupposition and backgrounding favor strategies such as subordination and indirectness. Because of their explicit dependence on the embedding context, complement clauses are less intrusive in the discourse than independent clauses that are structurally separate, paratactically connected layers. For this reason, they are typically used with information that is perceived as less prominent than the center of attention (Townsend and Bever 1977: 1, 4 and passim; Wallace 1982: 212, 215). Likewise, , which often co-occurs with complementizers, is on the whole more suitable than for presenting backgrounded information because, by maintaining the reference points of the current speech situation, it avoids abrupt changes of footing that can impede the flow of the discourse (cf. Goffman 1981: 150–52). This is also true of “nondirect” reports that lack shifters referring to the current speech situation. As Wallace observes (1982: 215), “the characteristic properties of figure and ground suggest that speaker and addressee will be figures against a ground of other entities not directly involved in basic face-to-face conversation”. All other things being equal, reporting strategies with nondirect orientation promote smoother transitions out of and into the embedding context than clear-cut ; they are less intrusive in the discourse and more quickly processed. Another factor that militates against in additional layers is the potential for clashes between the deictic orientation of the reporter and the represented speaker, which is especially acute in rejoinders and other speech genres in which the tag verb is in the first or second person. While perspectival alternations are commonplace in third-person narrative, especially when the narrator is effaced, they tend to be avoided in contexts with a dominant first-person orientation, especially when they affect the participants of the current transaction (cf. Ebert 1986: 150–51) — in self-quotations and secondperson rejoinders within actual dialogue and hence in additional layers as representations of reports within dialogue. Such drastic changes of footing would tend to retard the pace of the discourse — i.e., to have a highlighting effect, which would go contrary to the program in most cases of layered . In general, viewpoint ambiguities would counteract the purpose and explicit style of trial transcripts; both communicative efficacy and verisimilitude (since it probably reflected the practice of the actual speakers) made it desirable to maintain the reference points of the outer layer of within the inner layers. As Ebert (1986: 157) points out, from the vantage point of actual conversation it is that has a shifted perspective, not .15 (The clash
228
R V
of reference points is less problematic when the additional layer of is itself set in a narrative involving referents other than the reporter or addressee; here the choice between and is largely a question of foregrounding or backgrounding.) The tag verbs with additional layers of complementized are present tense when the reports represent claims made in the trial that are relevant for immediate purposes (ones the addressees have a current interest in upholding), and preterit when they introduce pretrial statements. Imperfective preterit tags tend to convey unsituated past statements or ones whose purpose is to provide a context for foregrounded ’s (e.g., questions and other dyad openers). Perfective preterit tags tend to be used with situated statements, including claims made in the trial context that are being exposed as illegal and indefensible. They are particularly associated with inset in narrative sequences when the ’s are plotline events but when there is no need to highlight the messages. The fact that an utterance has occurred can be foregrounded even when the actual message remains in the background if the message is presented in a way that minimizes the transitions between authorial framework to reported discourse (“gear-shifting”, Page 1988: 33).
7.4
Additional layers of uncomplementized indirect speech
While it is far more common for the additional layers of indirect or nondirect speech to be explicitly subordinated, there are 17 cases in the corpus in which such reports follow their introductory clauses without any intervening complementizer. The tag verbs represented are the ’s s”kazati/s”kazyvati and kazati, the z”vati ‘call, claim (as/that)’, and the verb slyxati ‘hear’, which is the semantic converse of s”kazyvati. The uncomplementized reports occur in some of the same contextualizations as complementized ones. Nine of the examples appear in the judges’ questions as requests for amplification or reaction.16 Five others come in testimony as subversive requotations of statements made (or allegedly made) by the represented reporters’ opponents.17 Like their complementized counterparts, the uncomplementized reports in these contextualizations are tagged with the present imperfective, which indicates that the content of the requotation is topical — a claim that is still being made by one of the parties in the litigation. This also accounts for the present imperfective tag in an example in which a litigant ascribes his
L R
229
own position to the witnesses that he has just named (AFZX 227, no. 259). There is also a case of ancestral hearsay tagged by the hearer-oriented verb slyxati (ibid.: 106, no. 114) in the preterit imperfective, as is expected in this contextualization, since it denotes a speech event in the unsituated past. The remaining case is tagged by the preterit of the perfective s”kazati; like other additional layers of tagged in that way, the report appears in a narrative context and depicts a statement situated in an earlier legal procedure (ASÈI 2: 194, no. 285). These overlaps in contextualization suggest that the absence of the complementizer does not index any distinctions in meaning (at least, as far as can be determined). Presumably one of the discourse factors that promoted it was pacing, since its general effect is to smooth transitions by eliminating a barrier between the narrative and that shares the narrative reference points. It is noteworthy that four of the examples follow other reports attributed to the same speakers — complementized (15a), or ’s (15b); AFZX 222, no. 258; 260, no. 308). In all of these cases, there are null subjects in the clauses introducing the uncomplementized reports, since their referents have already been thematized; in addition, all of the tags are linked to the preceding context by the non-sequencing conjunction a, which is typical in multiplex testimony and other situations with subdivided speech events (see 4.1, 4.3). (15) a.
A cˇto, g(o)s(podi)ne, skazyvaet [cˇernec’ Kirilo, cˇto and that lord- say-.3 [monk Kirilo]- that [teˇ pustoši dal [Danilo Blin, a [those fields]- give- [Danilo Blin]- and skazyvaet sem(’) leˇt pustoši kosili, say-.3 seven- years-. fields- mow-. to, g(o)s(podi)ne, lžet Kirilo… lord- lie-.3 Kirilo- “… And, lord, as for the fact that the monk Kirilo says that Danilo Blin gave those fallow fields and says [they] have used those fields for seven years, Kirilo, lord, is lying…” (ASÈI 2: 315–16, no. 333; see also (7) in 7.3).
b. vylgal sebeˇ u Timofeˇja u Mixailovicˇja get.falsely- . at Timofej- at Mixajlovicˇ-
230
R V
[gramotu lgotnuju… a skazal, g(ospodi)ne, to [writ of.privilege]- and say- lord- leˇs [velikog(o) kn(ja)zja forest- [grand prince]- “… [Savka] obtained, under false pretenses, a writ of privilege… and said, lord, it [was] a forest of the grand prince” (ASÈI 1: 193, no. 285; for the full context, see (9) in 7.3).
Since the prior reports thematize the fact of a speech event, the subsequent ’s have a reduced load; they may function merely as quasi-evidential disclaimers or citation markers, signalling the continuation of heteroglossic information. This diminution in the role of the tag clause works in a direction contrary to one of the functions of complementizers — to render attributions dominant over reports (cf. Thompson 1996: 518–19). The presence of the complementizer would demarcate the continuation more than was requisite, given the low functional load of the attribution. Indeed, given that most of the uncomplementized reports are presupposed from earlier testimony and serve as background for questions or rebuttals, some of the other attributions may have a low functional load as well. Two of the tags can be interpreted as intercalated rather than preposed (16): (16) Pocˇemu ty zoveš(’) [tu zemlju [velikie why you-. call-.2 [that land]- [grand knjaini, a si skazyvajut, — za tolko princess]- and these-. say-.3. for so.many- leˇt [tu zemlju… pašut? years-. [that land]- plow-.3. “Why do you claim that property as [land of the grand princess], while these [men] say they’ve been farming that land… for so many years?” (ASÈI 1: 247, no. 340; for the other example, see ASÈI 2: 194, no. 285).
The underlined segment can also be read as a si (skazyvajut) za tolko leˇt tu zemlju… pašut, with the intercalated in the usual position for parentheticals. If this interpretation is correct, the reported information is treated, in effect, as the reporter’s discourse, with the functioning as an unobtrusive quasi-evidential disclaimer.
L R
231
In addition to prior reports (15a–b), certain other syntactic conditions favor the absence of (or omission of) the complementizer. For example, the subjects of several of the report clauses are thematized in the immediately preceding discourse within the same turns at speaking. (The reported information is almost always activated in the larger discourse context.) Five of the uncomplementized reports (including (16) under the first interpretation) have the same subjects as their tag clauses (ASÈI 2: 194, no. 285; 315–16, 317, no. 333 (2×); 3: 268, no. 250), so that there is null reference. Another has a subject (again, with null reference) that is thematized in the previous report (AFZX 222, no. 258); a looser form of thematic continuity may obtain in some of the other examples (e.g., ASÈI 1: 193, no. 285). Shared thematic elements promote cohesion between adjacent segments of discourse; this tends to obviate the connective function of the complementizer. A further condition that was evidently amenable to uncomplementized is the presence in the tag clause of a deictic reference to the report (17): (17) To, g(o)s(podi)ne, skazyvaet [Danilo Blin, — podaval this- lord- say-.3 [Danilo Blin]- give- [pustoši Sar’skie [v”] [Farafontov monastyr’; a [fields of.Sara]- [to [Ferapont- monastery]- and teˇ, g(o)s(podi)ne, pustoši votcˇina [g(o)s(u)d(a)rja those- lord- fields- patrimony- [liege moeg(o) Gneˇvaševa… my]- Gnevaš-. “This, lord, says Danilo Blin — [he] gave the fallow fields on the Sara [River] to Ferapont Monastery; but those fields, lord, are the patrimony of my liege Gnevaš” (ASÈI 2: 317, no. 333; for the other example, see AFZX 106, no. 114).
The deictic particle or demonstrative pronoun to at the beginning of this excerpt refers to the following uncomplementized report. The presence of this cohesion-promoting element lessens the need for the explicit connexity provided by complementizers. (By the same token, complementized reports are rare after the cataphoric adverb tak(o).) Many of the reports exhibit marked word orders, which tend to characterize main rather than subordinate clauses. In four cases, including (15b), the reports begin with to, a particle used in equational (cleft-like) sentences.18
232
R V
In three other cases a prepositional or adverbial phrase is topicalized, as in (16),19 while in five the uncomplementized report has inverted word order, with particular elements fronted for emphasis.20 For example, in (18a) the noun phrase sem’ godov ‘seven years’ is the subject of an existential construction, which in neutral order would follow the presupposed temporal clause or an element correlative to it, as indeed can be seen in the report of the original testimony (18b) (tomu… kak…). In (18c), from a judicial-referral record, the usual order of the verification formula (18d) is partially inverted: (18) a.
Skazyvaeš’ ty, sem’ godov kak say-.2 you-. seven- years-. since [gramota ta u tebja zgorela; a Jurka [writ that]- at you-. burn- and Jur’ka- togdy živ li byl? then alive- . “You say it’s been seven years since that writ in your possession burned up. Was Jur’ka alive at that time?” (ASÈI 3: 269, no. 250).
b. a tomu, gospodine, let sem’, kak and that- lord- years-. seven- since posad gorel suburb- burn- “And from then, lord, it’s been about seven years since the suburb burned” (ibid.). c.
Skazyvaete sud vam takov ne byval… say-.2. trial- you-. such- be- “You say you did not have a trial like this…” (Kaštanov 1970: 359, no. 9).
d. Takov nam, gospodine, sud byl… such- us- lord- trial- be- “We had, lord, such a trial” (ASÈI 2: 453, no. 418).
In some of these examples, the report-initial elements may be sufficient to signal the onset of a new clause; in addition, the special focus in these examples may be a disincentive to complementization because of the
L R
233
increased syntactic complexity. In this light, it is suggestive that one of the instances of an uncomplementized report occurs within a preposed subordinate (topicalizing) clause and itself contains a postposed temporal clause (ASÈI 2: 438, no. 411). The use of a complementizer would thus have created a heavy structure with multiple embeddings in a position where light structures tend to be preferred. In sum, there are several contextual conditions that seem to favor the absence (or omission) of the complementizer — in particular, thematicity continuity, with the attendant increase in cohesion, and the avoidance of syntactic complexity. It does not appear that the distinction between uncomplementized and complementized reflects pragmatic or semantic differences, given the use of both strategies in much the same contextualizations. However, the use of uncomplementized reports may have been motivated by emotive nuances that cannot be reconstructed from the examples; conceivably, for instance, it may have conveyed iconically the rapid-fire pace of rejoinders in heated exchanges.
7.5
Additional layers of fused reported speech
As was discussed in 4.5, fused , a strategy unattested in , is a form of small-clause structure, in which reports of existential or copulative sentences are grammatically integrated into the tag clauses. The notional subjects of the reports are treated as direct objects of ’s or ’s; accordingly, they appear in the accusative or, when the tag verb is negated, the genitive. Pronouns and possessive adjectives that co-refer with the subject of the tag clause are reflexivized, even when they belong to the noun phrase that is the notional subject of the report. Forms of the verb ‘be’ have a null realization. The notional predicates of the reports can be nonagreeing genitives of possession or prepositional phrases, several of which also express possession; there are also two cases of predicative adjectives in the instrumental case, a construction that was expanding during this period.21 While fused reports are scantily attested in the narrative framework, they are better represented in additional layers of . Most of the examples are tagged by s”kazati/s”kazyvati, though kazati ‘say’ and its relational opposites slyxati and zaslyšeti ‘hear-/’ occur in isolated instances. The tense of the tag verb depends on the same factors as with additional layers of
234
R V
complementized (see 7.3): the present tense introduces requotations from the trial dialogue and other claims of continuing relevance; the preterit, statements purportedly made outside of the trial hearing. Within the preterit, the perfective aspect conveys relatively situated speech events, though under negation it may convey the modal nuance of inability. The contexts that favor fused reports are similar or identical to those that favor complementized and uncomplementized , which likewise show a relatively high degree of grammatical integration and reduction in comparison to . Fused reports do not occur in dyadic contexts or as plot-advancing events in narrative sequences; they tend to appear as presupposed background information — e.g., as preambles to requests for amplification (19), or as requotations of previous testimony that is about to be subverted (20).22 (19) Pocˇemu ty skazyvaeš’ [Ontonovu gramotu i ego why you-. say-.2 [Onton- writ]- and his brat’i lživoj? brethren- false- “Why do you say the writ of Onton and his brothers is false?” (Koreckij 1969: 289, no. 2). (20) To, g(o)s(podi)ne, naleveˇ [zemlja manastyr’skaja Vasilovskoe selišcˇo lord- on.left [land of.monastery of.Vasilovo village Kuznecovo, i ta, g(o)s(podi)ne, [zemlja Malecˇkinskaja Kuznecovo]- and that- lord- [land Malecˇinskaja manastyr’skaja žo, cˇto otnimaet Karp”, a of.monastery]- take.away-.3 Karp- and skazyvaet [Par”fenovskoi zemli… say-.3 [of.Parfenovo land]- “That, lord, on the left is the monastery’s property, the village of Vasilovo Kuznecovo, and that property, lord, of Malecˇinskaja also belonging to the monastery, which Karp is appropriating and [he] says [belongs to] the Parfenovo property” (ASÈI 2: 392, no. 388).
The notional predicate in (20), Par”fenovskoi zemli, seems to have confused a later copyist, probably because it was stranded from its notional subject and because the given type of small-clause construction itself was increasingly archaic; in a second variant, extant in a seventeenth century copy, the
L R
235
passage is altered to cˇto otnimaet Karp k Parfenovskoj zemle (“that Karp is appropriating for the Parfenovo property”; ibid. 2: 395, no. 388a). There is also a fused report of ancestral hearsay (21). As noted in 7.3, this contextualization favors the integration of heteroglossic information into the discourse of the reporter because the attribution is the focus of attention and the reported information tends to be presupposed. (21) I [otec moj, gospodine, mne togdy i meži and [father my]- lord- me- then boundaries- skazyval [toj zemle Parašinskoj da i požnjam s say- [that land Parašinskaja]- and fields- with [Ykonnicˇskoju zemleju ot recˇki ot Sendegi… po [of.Ikonnicˇe land]- from river- from Sendega- up.to kamen’, da po grjazcu… da ottole popereg stone- and up.to swamp- and from.there along lužku da na bor… meadow- and to forest- “… And my father, lord, then also told me the boundaries of that Parašinskaja land and [its] fields with the Ikonnice land were from the Sendega River… as far as the stone and as far as the swamp, and from there along the meadow and to the coniferous forest” (AFZX 231, no. 259).
In (21) the hearsay report is equivalent to a boundary-setting description, a speech genre that would ordinarily occur as plotline in the authorial narrative rather than as collateral information within the testimony. The backgrounding here has two explanations. First, the report is embedded in a narrative, where it is not a subject of special focus; indeed, it follows a segment of foregrounded that proves critical in the trial. Second, it serves as a preamble to the reporter’s invitation to demonstrate the boundary himself (ibid.). A fourth contextualization for fused reports is in hearsay cited in excuses — e.g., a defense against a charge of nonfeasance (22) or an explanation for a no-show witness (23). (22) Skazali nam, g(o)s(podi)ne, u sebja na [teˇ say-. us- lord- at . for [those d(e)r(e)vni [gramotu žalovalnuju, i my im, hamlets]- [charter of.immunity]- and we- them-
236
R V
g(o)s(podi)ne, potomu že molcˇali lord- therefore be.silent-. “[They] told us, lord, they had an immunity charter for those villages, and therefore, lord, we didn’t say anything to them” (ASÈI 2: 449, no. 416). (23) est’, g(ospodi)ne, u nas na to… starožilcy, be-.3 lord- at us- to that- witnesses-. tot že [Jakov Fedorov syn Gubin, da tot that- [Jakov Fedor- son Gubin]- and that- že Ivan… da [Suxoi Vasil’ev syn” Nikonov; a se, Ivan- and [Suxoj Vasilij- son Nikonov]- and lo g(ospodi)ne, [teˇ naši starožilcy… a Suxovo, lord- [those our witnesses]-. and Suxoj- g(ospodi)ne, n(y)necˇa doma neˇt, a lord-, now at.home be-. and skazyvajut ego v zadeˇleˇ. say-.3. him- in corvée- “… We have, lord, witnesses to that — the aforementioned Jakov Fedor’s son Gubin, and the aforementioned Ivan… and Suxoj Vasilij’s son Nikonov; and here, lord, are those our witnesses… but Suxoj, lord, is not at home now, and [they] say he is on corvée” (ASÈI 3: 69, no. 48).
In excuses of this kind, the focus is not only the per se but on the nonaction that it is thought to mitigate; cf. the causal relation in (22). In other words, excuses are explanations, collateral information rather than events; thus they lend themselves to backgrounding. It is noteworthy that an orientation excuse similar to the one in (23) is presented as a fused report in the authorial framework in the incipit of one trial transcript — a contextualization that favors reductive strategies (see (18) in 4.5).23 (It should be noted that the underlined passage in (23) is ambiguous; in the absence of clear indices of , it could be interpreted as a fused report in the authorial narrative (“‘… Suxoj, lord, is not at home now’, and they say he is on corvée”). In that case, the subject of the skazyvajut would be the preceding represented speakers, with the same pattern of zero anaphora as when
L R
237
reports are followed by narratives of action with the same subjects (see 4.3). By contrast, if the underlined passage in (23) is interpreted as speech within speech, the subject of skazyvajut must be indefinite.) As the examples show, fused was not a homogeneous phenomenon in . While the notional subject of the report is consistently treated as direct object, there are a number of different ways of realizing the underlying predicate, up to and including its omission in syntheses of content, where the topic is reported without any message (24).24 (24) Dvorskoj, gospodine, na [tom lese byl da ukazal, steward- lord- in [that forest]- be- and show- gospodine, nam mežu ot [Kijasovoj pustoši prjamo lord- us- boundary- from [Kijas- field]- straight lesom, a ne tudy, kudy Nikon skazyvaet. forest- and there where Nikon- say-.3 “The steward, lord, was in that forest and showed us, lord, the boundary — from Kijas’ fallow field straight through the forest, but not where Nikon says” (ASÈI 1: 459, no. 581).
While the reduction of the report is not always as extreme as in (24) and similar cases, fused is characteristic of reductive contextualizations where there is no particular focus on the report as a whole (although individual portions of the report can be put into emphatic positions). Like the use of omission of the complementizer, the choice of fused is due, in part, to pacing; such integration eliminates the transition out of and into the surrounding non-reported discourse. In addition, certain common types of sentences are evidently amenable to fused reporting when in isolation; for example, there are five cases, including (22), indicating that certain parties possess (or do not possess) charters giving them rights to the disputed properties.25 The use of fused in this contextualization is a general preference in the text-kind; similar cases may be found in verdicts (see Chapter 8).
7.6
Two functions of intercalated tags: Cohesion and disclaimers
Although the use of ’s as citation markers in “medial position” (Fónagy 1986: 262), within rather than before reports, is exceptional in the authorial
238
R V
narrative (see 4.4), it is relatively well attested in additional layers of . Of the sixteen tokens in the corpus with finite verbs, s”kazyvati occurs ten times, kazati five times, and s”kazati one time in a patently imperfective usage. The intercalated verbs are all in the present tense; their second- or third-person subjects, the represented speakers, are left implicit in most cases. There is also a single example of rek”ši, a nonagreeing (perhaps grammaticalized) past gerund of the verb recˇi ‘speak, say’ (ASÈI 2: 268–69, no. 310) — a usage attested in other legal text-kinds (cf. a case in a default judgment charter, ibid. 1: 362, no. 479). The data are insufficient to reveal pragmatic distinctions among the individual ’s. However, the frequency of kazati is conspicuous, given its scarcity elsewhere; intercalation may be its preferred function. Intercalated tags in additional layers, as in other contexts, have two main functions — to disambiguate and to promote cohesion. Like any tags, they demarcate reported information, but their position makes them comparatively inobtrusive because they do not block the reports off fully from the preceding context. In they tend to occur after the first phonological word (Wackernagel’s position) or, in later texts, after the first major constituent — positions associated with macrostructural connectors like the emphatic/ topicalizing particle že and parenthetical phrases indicating the relation of the clause to the prior text. Because intercalated tags, unlike preposed and postponed ones, do not intervene between adjacent clauses, they lend themselves to tagging continuations and can be used to promote cohesion by indexing cointerpretation in extended passages of (25): (25) Jaz, gospodine, velel im na [tex selišcˇax I- lord- bid- them- in [those villages]-. dvory staviti, a mne, gospodine, skazali farms-. build- and me- lord- say-. ivašovskie ž xristiane… cˇto [te selišcˇa of.Ivašov-. peasants-. that [those villages]-. Altynovo i Dubrovka — zemli [velikogo Altynovo- and Dubrovka- lands-. [grand knjazja, a tjanuli, skazyvajut, k [Ivašovu Uglu… prince]- and belong-. say-.3. to [Ivašov Ugol]-
L R
239
“I, lord, bid them to build farmhouses in those villages, and the Ivašov Ugol peasants, lord, told me… that those villages Altynovo and Dubrovka are lands of the grand prince, and belong, they say, to Ivašov Ugol” (AFZX 116, no. 125).
(For a similar example, see ibid.: 181, no. 204.) In such cases, the intercalated tags, without implying the occurrence of a new , specify the information source of an otherwise ambiguous segment of polysyndetic discourse that is stranded from the tag in the outer context. This resembles the typical use of the grammaticalized citation particle deˇ(i) in other text types.26 Such explicitness is, of course, particularly necessary when reporters (actual or represented) are citing claims they do not wish to endorse. The inobtrusiveness of intercalated tags makes them suitable for conveying heteroglossic information from previous testimony that is assumed for the sake of argument, as is typical in requests for amplification and other rejoinders (26). This function is particularly common (fourteen cases) when the use of a preposed tag clause would lead to unusual syntactic complexity, e.g., when the report is embedded in subordinated relative (26a), topicalizing, or complement clauses (26b). (26) a.
Komu to veˇdomo, cˇto [teˇ derevni… whom- that- be.known that [those villages]-. [vaši zemli monastyr’skie [Vasil(’)evskie [your- lands of.monastery]-. [Vasil’evskaja slobodki, cˇto vam, skazyvaete, dal slobodka]- you-. say-.2. give- [Roman Ivanovicˇ(’)? [Roman Ivanovicˇ]- “To whom is it known that those villages… are your (the monastery’s) properties pertaining to Vasil’evskaja Slobodka, which you, you say, were given by Roman Ivanovicˇ?” (ASÈI 2: 202, no. 288).
b. A vy celuete li krest na tom, and you-. kiss-.2. cross- on that- cˇto kažete [tot lug [velikogo knjazja that say-.2. [that meadow]- [grand prince]- [Dorofeevskoj zemli? [Dorofeevskaja land]-
240
R V
“And do you swear that (you say) that meadow belongs to the grand prince’s Dorofeevskaja property?” (AFZX 260, no. 308).
Example (26b) illustrates particularly well the difference between intercalated (parenthetical) and preposed (embedding) tags. Its placement is actually ambiguous, since the complementizing conjunction cˇ’to serves as the first phonological word.27 However, the addressees are not being asked to swear to the observable fact that they say the given message (“And do you swear that you say that meadow belongs…”); rather, they are being asked to uphold the content of the message, which is taken as a given for present purposes but marked as reported lest it be mistaken for the position of the represented speaker. Rephrasing examples like (26a–b) in a way that would allow preposed tags would lead to further structural complexity or a degree of diffuseness that would be problematic with presupposed information.28 It is noteworthy that, in two cases, intercalated tags appear in noninitial clauses in multiclausal complements (second clause, ibid. 212, no. 248; third clause, ibid: 212–13, no. 248), where they also serve to promote cohesion under conditions of syntactic complexity, like the example in (25). When reports are presented in this way, the intercalated tags can have a distancing effect similar to that produced by hedges. The reporters are using the heteroglossic information for their own purposes without dwelling on it per se or endorsing it as the truth; in some cases, they are even preparing to undermine it in what follows. The noncommittal character of the strategy explains its prevalence in the judge’s requests for amplification, where eight cases are found, including those in (26). However, this quasi-evidential function is contextual rather than inherent. Every report marker is a form of disclaimer, given that reporters have the option of accepting heteroglossic information as their own truth. Greater degrees of distancing can be expressed by using report markers “pleonastically”, where the fact of heteroglossia is evident, e.g., after the negative-presuppositional l”gati ‘lie’: (27) tot, gospodine, [Ivaško Vnuk lžet, cˇto ot that- lord- [Ivaško Vnuk]- lie-.3 that from nego, skazyvaet davyval najmu pjat’ altyn… him- say-.3 give- rent- five altyns-. “That Ivaško Vnuk, lord, says falsely that for [that property], he says, he paid five altyns in rent…” (AFZX 127, no. 140).
L R
241
Here the intercalated is used like a dubitative evidential to disassociate the message from the represented reporter’s viewpoint. The meaning of l”gati creates an amenable condition for the reinforced distancing caused by multiple attributions. Syntactic complexity may also be a factor here, since the given rarely takes a complement clause.29
7.7
Narrative reports of speech acts
As was noted in 4.6, ’s are infrequent in the authorial framework because the essentially evaluative character of such reports clashes with the objective style of narration that is conventional for the text-kind. By contrast, ’s, sometimes in combination with an infinitival complement, are widespread in the represented discourse of the trial participants. This is not surprising, since that is a context devoted to the expression of overt individual viewpoints and evaluations of meaning. The rich inventory of lexemes includes verbs that denote acts of naming, inquiries, answers, requests, commands, promises, claims, accusations, denunciations, falsifications, lies, etc. Their tense generally matches that of tag verbs in additional layers of (see 7.3), with the present used for ’s with current pertinence for the ongoing trial, and the preterit in narrative as expected. Typically ’s are found in contexts where the focus is not on the content of messages. For example, they can appear in sequence in narratives like (28), where the primary kinds of action depicted are not linguistic and where clausal reports would tend to provide an undue amount of detail (“fast-paced action”). (28) My, gospodine, im ne molcˇali. Kak we- lord- them- be.silent-. when pocˇali oni paxati siloju, i my, begin-. they- plow- force- and we- gospodine, bili cˇelom na nix [velikomu lord- hit-. forehead- against them- [grand knjazju… I [knjaz’ velikij, gospodine, poslal na prince]- and [prince grand]- lord- send- to zemlju [Mixaila Karpova. I Mixajlo… na [tu zemlju land- [Mixajlo Karpov]- and Mixajlo- to [that land]-
242
R V
v”ezžal, da jalsja… skazati [velikomu knjazju, da come- and promise- tell- [grand prince]- and [velikomu knjazju ne skazal… [grand prince]- tell- “We, lord, did not keep silent with them. When they began to plow by force, we, lord, petitioned the grand prince against them… And the grand prince, lord, sent Mixajlo Karpov to the land; and Mixajlo… came to that land and promised… to tell the grand prince, but he did not tell the grand prince…” (ASÈI 2: 369, no. 374).
While m”lcˇati ‘be/keep silent’ in (28) does not, prima facie, look like a , in a legal context it denotes ‘not report’, an act that the courts were inclined to interpret as nonfeasance.30 Thus it functions as an antonym of biti cˇel”m’ ‘petition’, also illustrated in (28). The latter is frequent within the testimony because of the court’s concern that litigants follow due procedures — in particular, petitioning promptly for redress — before initiating lawsuits.31 Frequently, ’s are chosen when the fact of the is crucial but the details of the message are unimportant for the argumentation (29a–b). As in judicial-referral records, there is a pronounced tendency to convey directives as infinitival complements of veleˇti ‘bid’ (or near-synonyms) rather than as nonintegrated report clauses after a (29a).32 (29) a.
Ot”jal”, g(o)s(podi)ne, [arximandrit Jakim u take.away- lord- [abbot Jakim]- at nas derevni i [lugi Gorodec’skii us- hamlets-. and [meadows Gorodeckij]-. po [staruju dorogu k selu k Filipovskomu… a up.to [old road]- to village- to Filipovskoe- and nyneˇ nas v[ysyla]et” von”, a ne velit now us- send-.3 away and bid-.3 nam [zeml’ svoix prodavati nikomu. us- [lands .]-. sell- no one- “Abbot Jakim, lord, took away from us the Gorodeckij villages and meadows up to the old road to the village of Filipovskoe… and now he is sending us away and commands us not to sell our [or his] lands to anyone” (ASÈI 3: 54, no. 31).
L R
243
b. Tot, gospodine, [Sergeec Vasilov syn that- lord- [Sergeec Vasil- son]- podbael [moix xolopov obelmyx trex… i lure.away- [my slaves full three]-. and uvel ix… za rubež… take.away- them- beyond border- “That Sergeec, Vasil’s son, lord, incited my three full bond-slaves to run away and took them… abroad” (ASÈI 3: 380, no. 357; cf. ibid.: 381).
The veleˇti in (29a), like biti cˇel”m’ and jatisja ‘promise’ in (28), is emotively neutral. Nevertheless, the use of these and other metapragmatic labels is the result of an interpretive (categorizing) activity absent in reports after lexically bleached ’s and, indeed, in the ’s (“omitted messages”, Thompson 1996: 518) involving the s”kazati in (28).33 More overt evaluations can also be found, as in (29b) and in ’s involving verbs such as l”gati/s”l”gati ‘lie (/)’ (many examples, e.g., ASÈI 2: 438–39, no. 411), vyl”gati ‘obtain by lying’ (ASÈI 1: 193, no. 285), oposluš’stvovati l”živo ‘bear false witness’ (ASÈI 2: 322, no. 334; Kaštanov 1970: 358, no. 9), poklepati ‘slander’ (ibid.), etc. The use of ’s is the culmination of the tendency for additional layers of to appear in integrated forms. Unlike reports in the authorial framework, layered reports tend to show referential strategies identical to those of the surrounding narrative (cf. Ebert 1986: 157); in addition, there is a preference for complementized and fused speech, which render the relations between tag and the report explicit. Similarly, the use of lexically specific ’s rather than faded ’s is a “speaker-based” strategy (Lakoff 1984: 481–84); it reflects the effort of the represented reporters — more precisely, the scribes, in conveying their statements or positions — to control the interpretation of the messages. Indeed, all of the reductive strategies characteristic of additional layers reflect a high degree of authorial control over the interpretation, with the “author” understood as either the represented reporter or the actual reporter (scribe), whereas in the narrative framework most of the responsibility for the interpretation is typically given to the intended hearer or reader. This control is in accord with the overtly subjective character of within , which represents the viewpoints of individuals trying to accomplish their own agendas in the trial situation.
C 8 Reporting the verdict
8.1
Preliminary remarks
The norms of reporting in the final major section, the verdict (found in the judgment charter and referred trial transcript subtypes) differs strongly from the patterns found elsewhere in the documents. In all of the verdicts examined, the dyadic question-and-answer structure found in the trial record is abandoned for a concise third-person, past-tense narrative devoted exclusively to the actions — more precisely, the ’s — of the judges, who are the topic of the discourse and the subject of most or all of the main clauses. The verbs conveying the judges’ actions appear in the tense that is the narrative norm — the preterit in Northeast Russian texts, the aorist in Pskovian and Novgorodian ones. The third-person style of the verdict precludes the use of appellative elements, even though the judges’ ’s had addressees in the actual context that is represented. As a rule, the verdict section begins with the conjunction i, which is used elsewhere to link consecutive events in the polysyndetic chain of the narrative framework. Thus the verdict is presented as the final event — the last word, as it were — in the trial hearing. In cases of judicial referral, the verdict serves as the conclusion of the judicial-referral record, which is usually joined to the preceding context asyndetically, so that it forms a separate narrative from the trial dialogue. In most of the trial transcripts that survive as autographs, where there is sparing use of capital letters, the initial letter of the verdict section is capitalized (boldfaced in (1)).1 This usage suggests that writers viewed the verdict section as a discrete paragraph in the discourse: (1) I [Kostantin Grigorievicˇ(’) v [teˇx zemljax and [Konstantin Grigor’evicˇ]- in [those lands]-.
246
R V
[Simanovskog(o) ma[na]styrja [poselskog(o) starca Ignat(’)ja [Simonov monastery]- [manager elder Ignatij Travina opravil, a išcˇei… obvinil, Travin]- find.right- and plaintiffs-. find.wrong- potomu cˇto starožilcov u nix ne bylo i za because old.residents-. at them-. be- and for pole ne poimali(s’), a [sami oni field- undertake-. and [selves they]-. išcˇei i starožilci; i prisudil [teˇ plaintiffs-. and old.residents-. and award- [those zemli [Simanovskomu manastyrju. lands]-. [Simonov monastery]- [1] And Konstantin Grigor’evicˇ decided in favor of the Simonov Monastery estate manager, the monk Ignatij Travin, in those lands, [2] and decided against the plaintiffs [3] because they had no longtime residents and did not agree to a judicial duel, and they themselves are plaintiffs and longtime residents; [4] and [Konstantin] awarded those lands to Simonov Monastery (ASÈI 2: 445–46, no. 414).
This passage exemplifies the typical internal structure of the verdict. The two halves of the primary adjudication — the judge’s decision in favor of one party [1] and against the other [2] — are grouped together as a single subsection divided by the conjunction a, which indicates a nonsequential (in the given instance, complementary) connection (cf. its use in multiplex answers). Any additional reports of ’s, like [4] in (1), are conjoined by means of i; this sets them up as separate subsections, correlated to discrete events in the trial context. Nevertheless, the reports of the secondary ’s cohere with the subsection containing the primary adjudication by sharing the same subject/topic. As as a rule, this subject/topic is neither repeated nor anaphorized after its first occurrence at the beginning of the verdict, even when, as in (1), there are several intervening embedded clauses with different subjects. This principle of cohesion can hold true over large expanses of text in trial transcripts of the sixteenth and seventeenth centuries, where the verdicts often include résumés that can cover several folia (see, e.g., Lixacˇev (ed.) 1895: 216–19). Thus the clauses that report the judge’s decisions are treated as if adjacent in the macrostructure.
R V
247
The judges’ prominence in the verdict, as implied by their strong thematization, and the other shifts in the representational technique mentioned above reflect changes in the configuration of the participants in the actual scene of the trial. As was discussed in Chapter 5, the judges play the basically ancillary roles of moderator and ratified listener in the trial hearing, which is set up primarily as a confrontation between the litigants and between their witnesses. In the judicial-referral hearing (see Chapter 6), all the participants are on a new footing; typically, the litigants serve merely as witnesses to the trial transcript, while the judges come to the forefront by assuming the role of decision-makers. However, the participants return to their previous configuration when new testimony is offered. In the verdict, by contrast, the judges are the only active participants and the focus of attention. The litigants now play a purely passive role as patients of the actions projected by the judges’ decision; indeed, they are not even identified as ratified auditors of the verdict, although this status can be inferred from the context (see below). When judges of higher instance issue directed verdicts for their subordinates in the chain of command to implement, the litigants and witnesses cease to be participants altogether; the trial judges are the only ratified listeners of the higher judges’ orders, while the interested parties are merely the patients of the trial judges’ projected actions in execution of the judgment. In view of these changes in footing, the dyadic structure of the preceding sections would be inappropriate in the verdict. There is no dialogue, in the strict sense; there is no exchange of views, no dissent, no polyphony. The judges’ viewpoint, refracted through the third-person narrative, is the only perspective presented. Consequently, the scribes felt no need to signal the reported status of the judges’ ’s; the topicalized judges are the only socially ratified speakers. Furthermore, the listeners in the verdict context are not true participants in the sense-making; they are receptors rather than interactors (in the sense of Dirven et al. 1982: 3). While the receptors do not reply verbally to the judges’ commands, they may respond in their subsequent behavior — in the normative case, by compliance. The litigants’ and witnesses’ statements are reported, if at all, only in requotations from the record of earlier hearings that appear, filtered through the judges’ viewpoint, in the ratio decidendi — their rationale for their decision ([3] in (1)). In this context, the may receive various subversive interpretations that the original speakers could scarcely have intended.
248
R V
In the trial transcripts investigated, the verdict section includes the adjudication proper — the ruling on the contested property — and, in many documents, the ratio decidendi. Two reporting strategies can be found in the adjudication: ’s, including volitional ’s with infinitival complements; and free nondirect speech. Like the additional layers of speech within the trial dialogue (Chapter 7), with which they share the property of being overtly shaped by a specific subjective viewpoint, the requotations in the ratio decidendi run the gamut of reporting techniques, from overt to submerged forms of representation; indeed, as will be discussed below, the entire ratio decidendi should be regarded as the covert, unattributed speech of the judges.
8.2
Reported speech in the adjudication
The judges’ adjudications are reported by means of strategies that integrate into the third-person, past-tense narrative of the verdict. The actual judgment, giving one of the litigants ownership rights and penalizing the other, is presented in a maximally incorporated form — as a sequence of ’s involving opraviti ‘find/say to be right; decide in favor of’, ob(v)initi ‘find/say to be wrong; decide against’, and occasionally other ’s such as prisuditi ‘award (in judgment); give the right to’ and otsuditi ‘take away (in judgment)’ (2). (2) I potomu [kn(ja)z’ Jur’i Vasil(’)evicˇ(’)… [Nestera, starca and therefore [prince Jurij Vasil’evicˇ]- [Nester elder]- [Simanovskog(o) monastyrja… opravil, a Ondreˇika [of.Simonov monastery]- find.right- and Ondrejko- da [Ivaška Strelu [Lopotovyx detei obvinil and [Ivaško Strela]- [Lopot- children]- find.wrong- i prisudil [zemlju Malecˇkinskuju k Verznev’skomu and award- [land Malecˇkinskaja]- to Verznevskoe selu [Simanovskogo manastyrja. village]- [of.Simonov Monastery]- And therefore Prince Jurij Vasil’evicˇ… decided in favor of Nester, the monk of Simonov Monastery, and decided against Ondrejko and
R V
249
Ivaško Strela, Lopot’s sons, and awarded the Malecˇinskaja property to Simonov Monastery’s village of Verznevskoe (ASÈI 2: 390, no. 387).
In adjudication passages like (2), the verbs opraviti, ob(v)initi, prisuditi, otsuditi convey communicative acts — rulings — rather than an implementation of the judgment; they are “verdictives”, in Austin’s (1975) classification, ’s like acquit, convict, rule, which “consist in the delivering of a finding…” (153; see also 88, 141, 151, 153–55). Wierzbicka (1987: 239) likewise posits a spoken component in English convict and acquit, which are similar in meaning to opraviti and ob(v)initi (though they relate to criminal rather than civil trials). Without a spoken component, the actions denoted could not have been observed by the scribes, unlike the actions conveyed by verbs of nonlinguistic action. (The oral nature of the adjudication tends to be made explicit in seventeenth-century transcripts — texts that reflect a trial process in which writing was the basic form of communication. Seventeenthcentury judges reached their decisions in camera after examining the dossiers that had largely replaced face-to-face confrontations.)2 The presence of the spoken component in the adjudication — “submerged speech” (Page 1988: 34) — explains the sporadic appearance of changes in perspective in fifteenth- and early sixteenth-century trial transcripts (3): (3) I [knjaz’ velikij Ivan Vasil’evicˇ’… [Danila Stromilova i and [prince grand Ivan Vasil’evicˇ]- [Danilo Stromilov]- and [Osifa cˇern’ca opravil, a… xrestijan [Pexorskogo [Osif monk]- find.right- and peasants-. [of.Pexorka stanu obvinil, i prisudil… derevni i district]- find.wrong- and award- villages-. and pustoši k [Simanovskomu manastyrju po unsettled.lands-. to [of.Simonov monastery]- by starine, opricˇe [bortnicˇ’ix dereven’, cˇto old.usage- except [beekeeper- villages]-. otnjal u nix [otec moj knjaz’ velikij Vasilej take.away- at them- [father my prince grand Vasilij Vasil’evicˇ’ u manastyrja u Simonova. Vasil’evicˇ]- at monastery- at Simonov- And Grand Prince Ivan Vasil’evicˇ… found Danilo Stromilov and the monk Osif right and found the peasants of the Pexorka district wrong,
250
R V
and awarded… the villages and unsettled lands to Simonov Monastery in accordance with the previous state of affairs, “except for the beekeepers’ villages that my father Grand Prince Vasilij Vasil’evicˇ took away from Simonov Monastery” (ASÈI 2: 380, no. 381).
In (3), amidst third-person narrative, there is a slip to the first person in a relative clause, though up to now there has been no explicit indication that anyone is speaking — a sign of latent . (Similarly, in the Pskovian transcript, the judges are presented in the third person in the trial record but in the first in the adjudication, without other explicit indication of ; GVNP: 327, no. 340.) The first person refers to the judge, Grand Prince Ivan Vasil’evicˇ’, as the reference to his father shows; thus it is coreferential with the third person at the start of the passage and in the sequel (not cited here). The slip occurs in the one clause where there would be multiple third-person referents if the narrator’s perspective were maintained. Fade-ins to diffuse free indirect or nondirect speech, with preservation of the narrative perspective, are also attested; these tend to occur when the reported adjudications are multiclausal and hence do not lend themselves to compact presentation — typically, legislative directives with continuing relevance (4): (4) i prisudil [gospodin Gerontij, mitropolit [vsea and give.right- [lord Gerontij metropolitan]- [all Rusii, [knjazju Mixailu Andreevicˇju [Kirilova Rus’]- [prince Mixajlo Andreevicˇ]- [Kiril- monastyrja igumena suditi po starine… opricˇ’ monastery]- abbot- judge- by old.custom- except [duxovnyx del; a igumen sudit [svoju [spiritual cases]-. and abbot- judge-.3 [. brat’ju starcov sam… i [pristavov brethren]- elders-. self- and [bailiffs svoix arxiepiskopu v [Kirilov monastyr’ .]-. archbishop- to [Kiril- monastery]- ne vsylati… send.out- And the lord Gerontij, metropolitan of all Rus’, gave Prince Mixajlo Andreevicˇ the right to judge the abbot of Cyril Monastery, in accordance
R V
251
with the old custom… except in spiritual cases, but the abbot judges his brethren, the monks, himself… And the archbishop is not to send his bailiffs to Cyril Monastery… (ASÈI 2: 281, no. 315).
In (4), taken from a lawsuit over ecclesiastical jurisdiction, the adjudication of the verdict begins in an integrated, attributed form as the complement of the prisuditi. It then slips into a sequence of unattributed clauses (only part of which are cited here), with nonpast indicative verbs or unbound infinitives conveying imperative modality (prescribed future behavior). The meaning of the passage would be essentially unchanged if the first directive were expressed as nondirect speech or if the later commands were presented, like the first, as infinitive complements. However, the latter would create multiple embeddings, which could be hard to process, especially given the graphic format. Moreover, integration with past-tense ’s indicating temporally limited events would be less effective than unbound present-tense discourse for conveying directives meant to have ongoing validity.3 Directives of limited validity are, in fact, generally conveyed by ’s involving the volitional verb veleˇti ‘bid’, which takes an infinitival complement with a dative for the recipient of the command (if explicit). (In prohibitions, the is negated, even though the scope of the negation is the action conveyed in the complement clause.) Thus the commands are presented in the same way as procedural directives in the trial narrative (5). (5) I [knjaz(’) velikii Vasilei Dmitreevicˇ(’)… [teˇx zeml’ and [prince grand Vasilij Dmitrievicˇ]- [those lands c(e)rk(o)vnyx… ne veleˇl prodavat[i] ni kupiti nikomu, of.church]-. bid- sell- nor buy- noone- a kto imet prodavati ili kupiti ili arximandrita and who- will-.3 sell- or buy- or abbot- ne slušati, teˇ[x] veleˇl [veliki [knjaz’] [arximandritu obey- those- bid- [grand prince]- [abbot mixailov’skomu vo[n] metati. of.Michael]- thence drive- And Grand Prince Vasilij Dmitrievicˇ… commanded that no one sell or buy those church lands; and if anyone buys or sells or does not obey the abbot, the grand prince commanded the abbot of St. Michael to drive them out (ASÈI 3: 54, no. 31).
252
R V
The strategy of veleˇti with the dative plus infinitive maintains the narrator’s perspective and allows the reported command to be condensed into its barest essentials, while contextually trivial deictic information such as person, tense, and mood is eliminated. However, in elaborated commands such as (5), in which the complements of veleˇti have dependent clauses of their own, the verbs in the subordinate clauses remain finite and retain the grammatical marking of independent verbs. Examples of this kind demonstrate how bound can slip into less integrated forms — in the given instance, nondirect speech — in cases in which compact presentation can lead to syntactic complexity and processing difficulties. (Slipping into is not the only possible interpretation here. Semino, Short, and Culpeper (1997: 30) view English constructions involving elaborated ’s as a special subcategory — “narrator’s representation of speech act with topic” — that has “a greater summarising effect than is normally associated with .”)
8.3
Reported speech in the judges’ ratio decidendi
In more than 40 trial transcripts (approximately one quarter of the documents in the corpus), the narrative in the verdict section includes the judges’ ratio decidendi.4 In most cases, this rationale is expressed in the form of a clause or string of clauses complementized by a causal conjunction (po tomu cˇ’to or, in a few cases, togo deˇlja cˇ’to ‘because’) or by the all-purpose complementizer cˇ’to in a causal usage. This causal construction is usually embedded in the clause predicated by the verb ob(v)initi ‘say/find to be wrong; decide against’, i.e., in the that announces the losing party. Less often, it is embedded in the clause predicated by opraviti ‘say/find to be right; decide in favor of’. As mentioned above, the discourse theme of the main clause (the judge) is generally not reindicated when the verdict narrative resumes after the causal clause, even when the ratio decidendi is lengthy and includes intervening, potentially competing themes.5 The causal clauses of the ratio decidendi furnish another example of “submerged ”. The evaluative information that they convey could only have been furnished by the judges, even though this is never indicated explicitly in the texts. The scribes, who otherwise conceal their evaluations, would scarcely have undertaken to provide rationales for decisions that they themselves did not make, especially given the exalted status of many of the
R V
253
judges. Indeed, the majority of fifteenth-century trial transcripts offer no justification whatsoever for the judges’ verdict, which was to be accepted without demur. (When both a directed and an executed verdict are recorded, the ratio decidendi is often omitted in the latter, since the decision belonged to the higher judge alone.) There was no need to tag the information in the causal clauses as because it fell within the scope of the preceding . The causal clauses themselves were part — more precisely, were constructed by the scribes as part — of the statements made by the judges in announcing their verdicts, which were transmitted in a converted form as past events rather than quotations in the narrative. As a rule, the information in the ratio decidendi tends to be condensed, especially when it relays the of the trial participants. The principal reason for this reduction is that the information is almost always repeated from earlier in the transcript.6 Consequently, although it is naturally asserted as a factor in the judges’ decision, it is not put forth as new evidence but rather presupposed as part of the previous discourse that has been temporarily deactivated (displaced from the center of attention; Brown and Yule 1983: 41, 75, 79, 173, 179–83). As a known entity, the requoted information does not have to be represented in the same detail as when it first appeared as testimony; indeed, such detail would make it more prominent than would be functional, given its ancillary role in the verdict narrative, where the judges’ decisions are the principal type of event. The givenness of the information transmitted in the ratio decidendi is encoded syntactically, in part, by its being conveyed in a subordinate clause. It is typical for the causing events in complex causal constructions to be conveyed in the subordinated rather than in the main clauses (see Talmy 1978: 632, 639–41). This hypotactic structure is another factor that favored syntactic and semantic reduction of the repeated information as a way of avoiding undue complexity in the embedded sentence(s). 8.3.1
Indirect and nondirect speech
If the ratio decidendi as a whole consists of messages implicitly reported by the judge and re-reported by the scribe, overtly attributed statements from the trial dialogue represent an additional layer of within , like similar second-hand reports found within the testimony (see Chapter 7). Indeed, the strategies employed for explicit reports in the ratio decidendi are largely
254
R V
congruous with those used for within in the trial record. There was a strong tendency to reduce such requotations; this was probably promoted by the givenness (recoverability) of the message, which favored compact presentation, and by the transition from a dyadic framework to the continuous narrative format of the verdict. Much of the detail — in particular, deictic elements — needed to convey the give-and-take of conversational exchanges would be superfluous outside the dyadic framework, in speech reported at second hand — indeed, at third hand, since the judges’ statements themselves are converted and reduced from their original first-person orientation. In accordance with this reductive tendency, there is a strong preference for nondirect and grammatically integrated forms of in the ratio decidendi. The best-attested strategy is s”kazati ‘say ()’ followed by complementized indirect or nondirect speech (6). The only complementizer attested is the all-purpose conjunction cˇ’to ‘that’. This strategy is attested, sometimes alongside other methods, in 34 examples in 25 documents.7 (6) I [knjaz(’) velikii Ivan Vasil(’)evicˇ(’)… [Vaska Danilova and [prince grand Ivan Vasil’evicˇ]- [Vasko Danilov]- veleˇl obiniti… potomu cˇto skazal Vasko, cˇto [sud bid- find.wrong- because say- Vasko- that [trial takov byl, da i to sam že skazal, cˇto such]- be- and that- self- say- that za sorok leˇt o [toi požneˇ molcˇal, da for forty- years-. about [that field]- be.silent- and n(y)neˇcˇa [tu požnju pokosil. now [that field]- mow- And Grand Prince Ivan Vasil’evicˇ… bid [the trial judge] to decide against Vasko Danilov… because [1] Vasko said that there had been such a trial [2] and also said himself that he had been silent about that field for forty years and had now mowed that field (ASÈI 2: 384, no. 383).
While complementized is rarely used for testimony, it is the most common strategy with within the testimony, where statements are filtered through dominant subjective viewpoints.8 In addition, it occurs frequently with verificational statements in judicial-referral reports — that is, with given,
R V
255
prepatterned reports. The in the ratio decidendi is likewise given (albeit in need of reactivation), and it is also filtered through the judge’s explicit viewpoint. The choice of the tag verb s”kazati is perhaps by default; the use of perfective aspect is motivated by the punctiliar character of the reported statement (situated within the previous dialogue) and by the fact that the message is no longer being mooted. The constructed, reduced nature of within the ratio decidendi can be seen by comparing the requotations with their anterior reports. The written trial record was undoubtedly the immediate source of the reports in (6), since the verdict is issued by a judge of higher instance. While the first report in (6) conveys, in abbreviated form, the prepatterned verification procedure from the judicial-referral hearing [1] (see 6.2), the second [2] is a selective, highly interpretive amalgam of two statements (7a–b), which subverts the intention of the represented speaker (the respondent): (7) a.
I Vasko tak rek: Tu, g(ospodi)ne, požnju and Vasko- thus speak- that- lord- field- nynecˇa pokosil jaz; a požnja to [velikog(o) now mow- I- and field- [grand knjaz(ja) [Gorodišcˇ”skogo sela. prince]- [of.Gorodišcˇe village]- And Vasko spoke in this way: “That field, lord, I have now mowed; but the field belongs to the grand prince’s village of Gorodišcˇe” (ASÈI 2: 383, no. 383).
b. I Vasko tak rek: … [ta požnja byla and Vasko- thus speak- [that field]- . Pašin pokos; a posleˇ Paši [tot pokos Paša- meadow]- and after Paša- [that meadow]- v mor” kosil s(y)n” ego Filja; a tomu in plague- mow- son- his Filja- and that- uže, g(ospodi)ne, boleˇ soroka leˇt… a posleˇ already lord- more forty- years-. and after ix [tog(o) pokosa ot [teˇx meˇst”… them- [that meadow]- from [those places]-. ne kašival nixto… mow- no one-
256
R V
And Vasko spoke in this way: “… That field was Paša’s meadow, and after Paša his son Filja mowed that meadow during the plague; and that, lord, was more than 40 years ago… And after them from that time… no one mowed that meadow (ibid.).
The statement that Vas’ko kept silent about his claim to the property for more than forty years does not correspond directly to any of his remarks; the author has read between the lines of the testimony in (7b). Nothing in the record suggests that Vas’ko admitted such nonfeasance, though a statement to that effect is tagged as his . The requotation in (6) features omission of both the coreferential subject pronoun and the vocative g(o)s(podi)ne, which is characteristic of the switch from direct to nondirect speech. Since the vocative relates to the original speech situation, it offers no information pertinent for interpreting the judge’s reported decision; consequently, its omission — attested in all the cases of complementized speech in the verdict — reduces the semantic complexity caused by the shift in viewpoint, which does not carry any functional (informative or disambiguating) load in this context. The same is true of the deletion of coreferential subject pronouns, which is common though not obligatory in requotations in the ratio decidendi. Like other redundant elements, such pronouns would facilitate interpretation of the testimony in the trial record (as in the real-world dialogue that it reflects); there was no need for this function in the verdict, inasmuch as the critical evaluation was a fait accompli. The greater brevity and simplicity that resulted from these deletions probably added to interpretability of the causal clause. In addition to complementized reports, there are fourteen tokens of uncomplementized indirect or nondirect speech, in twelve transcripts. There is one case of the kazati, which, as usual, appears in the present tense, as does the verb of knowing in the second clause of the rationale (ASÈI 3: 88, no. 56). The remaining examples are all tagged by the preterit of s”kazati, like the complementized reports discussed above. In most of the examples, the report is the first or only item in the rationale; hence the tag clause comes immediately after the causal conjunction.9 This suggests that distance between the causal conjunction and the reported-speech clause plays some role in the pragmatic choice between inclusion or omission of the complementizer; in other words, one of the motivations may have been to avoid explicit multiple embeddings in close proximity.
R V
257
There may be additional motivations in some cases. One possible factor is a high degree of cohesion between the tag and the first clause of the report. For example, in four cases (two of which occur in a single trial transcript), the subject/theme of the report is coreferential with that of the tag clause (the represented speaker) and hence omitted (8).10 (8) Po [semu spisku [knjaz’ velikij… [selcˇan by [this copy]- [grand prince]- [villagers voislavskix Larivonka s tovarišcˇi of.Voislavskoe]-. Larionko- with companions-. velel obiniti, potomu cˇto [dvoi znaxori bid- find.wrong- because [two witnesses Serapionovy skazali, odni — najmovali u Serapion-]-. say-. some-. rent-. at nego [tu zemlju po pjat’ let’, a him- [that land]- each five- years-. and drugie ego ž znaxori skazali — po others-. his witnesses-. say-. each tri gody najmovali … a [poselskoj voislavskoj three- years-. rent-. and [manager of.Voislavskoe Leša i s xristiany skazali, cˇto za Leša]- with peasants-. say-. that for dvatcat’ let [ta zemlja [knjažšcˇina twenty- years-. [that land]- [prince’s.land Voislavskaa, a paxali ee [te of.Voislavskoe]- and plow-. it- [those selcˇane… villagers]-. In accordance with this copy, Grand Prince Ivan Vasil’evicˇ… bid [the trial judge] to decide against the Voislavskoe villagers Larionko et al. because Serapion’s two sets of witnesses said, some rented that land from him for five years each, and others of his witnesses said for three years each… And the Voislavskoe estate manager Leša and the peasants said that for twenty years that land was [part of] the prince’s property of Voislavskoe, and those peasants farmed it… (AFZX 98–99, no. 103).
258
R V
Cohesiveness would tend to counteract the complementizer’s function of demarcating the from the authorial context; hence it could influence omission of the complementizer. In three of the remaining cases, the subjects are non-coreferential and appear in postposition to the predicate (9).11 This is a marked order as compared to the narrative norm (subject-predicate); it has the effect of making the postponed subject noun phrases salient. (9) I [pisec Ivan Poluextov… Ozarka, da Gavrilka, and [scribe Ivan Poluekt-]- Ozarko- and Gavrilko- da Xarkju… obvinil, potomu cˇto sami and Xar’ka- find.wrong- because selves-. skazali: sekli les [monastyrskie xrestijane say-. cut-. forest- [of.monastery peasants]-. okolo [monastyrskix dereven’ around [of.monastery hamlets]-. And the scribe Ivan Poluekt’s son… decided against Ozarko and Gavrilko and Xar’ka… because [they] themselves said the monastery peasants cut the forest around the monastery hamlets (ASÈI 1: 460, no. 581).
The crucial point of this rationale is the plaintiffs’ admission that the disputed forest is surrounded by their opponent’s property (cf. emphatic repetition of monastyr’skij). The fact that the reports begin with verbs prevents their being interpreted as part of the tag clauses; this makes the complementizer redundant and promotes its omission. Judging from the evidence, two opposing tendencies are at work in the choice between zero and an explicit complementizer with in the ratio decidendi. On the one hand, the givenness or, at least, the low degree of assertedness of most of the transmitted information and the summarizing (and hence fast-paced) character of the section as a whole tends to promote as much reduction as is feasible without detriment to interpretability. On the other hand, the fact that the ratio decidendi establishes a specific, official viewpoint tends to favor explicitness in order to clarify the accepted interpretation and to prevent misinterpretations. The first tendency conditions deletion of the complementizer, subject pronoun, and other redundant information; the second leads both to the inclusion of such redundant
R V
259
elements (to varying degrees) and to the use of grammatically bound forms of or lexical evaluations of illocutions in place of clauses. Syntactic factors and discourse cohesion also seem to exert some influence. The occurrence of complementizing conjunctions is evidently linked with the need to avoid ambiguous structures that could potentially be misinterpreted in the continuous linear organization of the text. For example, when there is some interrupting element such as a prepositional phrase in the matrix clause right before the subordinate clause (e.g., po kr’st’nomu ceˇlovaniju ‘in accordance with a cross-kissing oath’; e.g., ASÈI 3: 291, no. 276), explicit complementizers seem to be preferred in order to promote the proper segmentation of the text. Bolinger (1972: 38) notes a typologically parallel use of that in English when the complement clause “is separated from the main verb by intervening complements or interpolations… to preserve the identity of the clause.” Conversely, complementizers can be omitted when the predicate comes first in the speech clause rather than the subject, because there is less possibility of misassigning reported information — in particular, noun phrases and prepositional phrases — to the matrix clause when the word order is marked/salient than when it is neutral. Likewise, it would appear that the complementizing conjunction is omissible when the itself begins with a complementizer such as kak” that clearly demarcates it from the tag clause (e.g., ASÈI 2: 198, no. 286; 203, no. 288); this possibility may also be promoted by the production and interpretation constraints on multiple embeddings. Syntactic parallelism also has some disambiguating power in this respect, as is shown by the first two reports in (8). 8.3.2
Fused reported speech
The tendency to reduce in the ratio decidendi is carried further when the report loses its clausal identity and fuses with the tag. In seven examples (in six documents), the subject of the is encoded as the object of s”kazati ‘say ()’ in the accusative case, or genitive when the verb is negated by raising (cf. Noonan 1985: 90) (10). (10) I… [Vasilei Grigor’evicˇ(’) Naumov… [Okiša Oljunova da and [Vasilij Grigor’evicˇ Naumov]- [Okiš Oljunov]- and
260
R V
Fomku… ob’vinil potomu cˇto skazali u sobja Fomka- find.wrong- because say-. at . [starožilca mertvovo, apricˇ(’) tog(o) ne skazali u [old.resident dead]- except that- say-. at sobja starožilca… . old.resident- And… Vasilij Grigor’evicˇ Naumov… decided against Okiš Oljunov and Fomka… because they said their witness was dead; they said they had no witness besides him (ASÈI 1: 583, no. 658).
As in most of the examples of fused , the statements after which the reports in (10) are modelled (11) have the verb of existence predicating a nominative subject to a prepositional phrase with locational or, as in (10), possessive meaning.12 Nonreflexive pronouns in the anterior reports that refer to the represented speakers are reflexivized in the fused constructions (“logophoric reflexives”, see Kuno 1987: 105, 136–45). (11) Veˇdomo, g(o)s(podi)ne, u nas bylo [igumenu Germanu, be.known lord- at us- . [abbot German]- a uže ego, g(o)s(podi)ne, v živote ne stalo; a and already him- lord- in life- become- and tomu, g(o)s(podi)ne, desjat(’) leˇt, kak ego ne that- lord- ten- years-. as him- stalo; apricˇ(’) togo, g(o)s(podi)ne, u nas ne become- besides that- lord- at us- vedomo nikomu. be.known no one- “Our witness, lord, was Abbot German, but, lord, he is no longer alive; it’s been ten years, lord, since he died. Besides him, lord, we have no witness” (ibid.: 582).
As (11) shows, the reports in (10) are compact syntheses rather than a straightforward transmission of the diffuse model found in the testimony. In four additional cases, which come from the directed and executed verdicts of a single trial recorded in two variants, the subject of the , which is coreferential with the represented speaker, is deleted, and the itself is reflexive (12):
R V
261
(12) I… [knjaz(’) Ivan Jur’evicˇ(’)… [kn(ja)zja Ivana Kostjantinovicˇ(a) and [prince Ivan Jur’evicˇ]- [prince Ivan Konstantinovicˇ Obolenskog(o) veleˇl obiniti potomu cˇto na [sudnye Obolenskij]- bid- find.wrong- because to [of.court muži is [sc. sja] ne poslal, a u doklada men]- refer- and at referral.hearing- stav, skazalsja vinovat… stand- say-. guilty- And… Prince Ivan Jur’evicˇ… bid [the trial judge] to decide against Prince Ivan Konstantinovicˇ Obolenskij because he didn’t refer to the men of court and, having appeared at the judicial-referral hearing, said he was in the wrong… (ASÈI 1: 515, no. 607a; similarly in the trial judge’s verdict, ibid., and in both verdicts in ibid.: 509, no. 607).
While the reflexive verb s”kazati sja does not appear elsewhere in the corpus, it is common in some varieties of chancery writing, especially in text-kinds like indenture documents that report brief personal statements (see Sreznevskij, s.v. “s”kazyvatisja”). It should be noted that the use of s”kazati sja does not necessarily entail fused (loss of clausal identity); there are later examples from other text-kinds in which finite forms of the copula or other verbs are maintained (e.g., NZKK Part 2: 207, 286). 8.3.3
Direct speech
Given the compact and partially given character of reports in the ratio decidendi, the diffuse form of prototypically associated with new messages (testimony) is dispreferred. There is one clear instance, which is tagged with the standard formula tak(o) rek(li) (13): (13) I [knjaz’ Ivan Jur’evicˇ’ velel… [Andreja Okljacˇeeva and [prince Ivan Jur’evicˇ]- bid- [Andrej Okljacˇeev]- obviniti… potomu cˇto Ondrej i ego znaxori find.wrong- because Andrej- and his witnesses-. zabludilisja… vodili po [ugrešskoj zemle go.astray-. lead-. through [Ugrešskij land]- [Kostjantinovskogo sela; i [ugrešskoj krylošanin [Ofonasej [of.Konstantinovo village]- and [Ugrešskij cleric [Ofonasij
262
R V
Morž da [starec Poleuxt [sic], naexav ix na Morž]- and [elder Poluekt]- come.upon- them- in lese, tako rekli [sud’i Cˇjubaru da Ondreju forest- thus speak-. [judge Cˇubar]- and Andrej- i ego znaxorem: To gospodine, Ondrej i ego and his witnesses-. lord- Andrej- and his znaxori zabludilisja, vodjat ne witnesses-. go.astray-. lead-.3. gorazdo… a vedut, gospodine, Ondrej i ego accurately and lead-.3. lord- Andrej- and his znaxori po [ugrešskoj zemle po našej witnesses-. through [Ugrešskij land]- through our- [Kostjantinovskogo sela; i s [tex mest [of.Konstantinovo village]- and from [those places]-. Ondrej i ego znaxori vorotilisja, a… Andrej- and his witnesses-. turn-. and Ofonasiju da [starcu Poleuxtu [sic] ne otvecˇav nicˇego… Ofonasij- and [elder Poluekt]- answer- nothing- And Prince Ivan Jur’evicˇ bid [the trial judge]… decide against Andrej Okljacˇeev… because Andrej and his witnesses went astray… [they] led through the land of Ugrešskij Monastery’s village of Konstantinovo; and the Ugrešskij cleric Ofonasij Morž and the monk Poluekt, coming upon them in the forest, spoke in this way to the judge Cˇubar and to Andrej and his witnesses: “Andrej, lord, and his witnesses have gone astray, they are leading [you] inaccurately… Andrej and his witnesses, lord, are leading, through our (Ugrešskij’s) land belonging to the village of Konstantinovo.” And thereupon Andrej and his witnesses turned, without replying to Ofonasij and the monk Poluekt (ASÈI 2: 440, no. 411).
The deictic information provided by the direct perspective is functional here, since it helps to distinguish between two sets of referents that would otherwise both be in the third person — Ofonasij Morž and Poluekt, on the one hand, and Andrej’s witnesses, on the other. In comparison with other instances of in the verdict, the report in (13) is long and syntactically complex — factors that may have favored the diffuseness of for the sake of clarity.
R V
263
In addition, it appears in a sequence of narrative events (“went astray… spoke… turned…”), a situation that favors in other contextualizations. The inclusion of such a complex passage — especially one that involves a shift of deictic pivot — within a causal clause poses the risk of loss of cohesion, with the potentially being cointerpreted with the authorial discourse. In this light, it is significant that the in (13) includes a vocative that is absent from the anterior report in the trial record (“a vedut, gospodine, Ondrej i ego znaxori…”; see ASÈI 2: 439, no. 411). In their appellative function, vocatives have no place in the ratio decidendi. However, as mentioned in 3.2, within narratives vocatives or other phatic forms can serve as ancillary markers of reportedness. While the first clause of the report in (13) is demarcated by means of the tag clause, the status of the following clauses is less clear because of the polysyndetic structure of the passage. Consequently, the additional vocative acts as a kind of verbal quotation mark similar to an intercalated or citation marker. The fact that g(o)s(podi)ne ceases to occur in the following narrative is the only overt signal that the has ended (cf. English unquote); without the vocative, the correct interpretation would have been uncertain up to the point at which the ’s of the judge’s verdict resume at the end of the excerpt. Another device used to maintain cohesion in extended is the quotation marker deˇ(i), which generally appears in intercalation after the first phonological word or, in less conservative texts, after the first major constituent of the sentence. This particle (a grammaticalized form of an obsolete , which can appear as deˇi, dei, deˇ, de, or di) is widespread both by itself and in combination with ’s in text-kinds such as immunity charters and edicts that are written from a first-person perspective, where is backgrounded and filtered through a single authoritative viewpoint (e.g., ASÈI 1: 409, no. 532; 3: 274, no. 253). In trial transcripts, by contrast, it is scantily attested; apart from the one case in testimony (Section 4.7), the only three examples occur in the ratio decidendi section. Two appear in the directed and executed verdicts of a single transcript within what is probably a case of slipping into . The particle is used, like the vocative discussed above, to signal that the is a continuation of the preceding, explicitly tagged report (14): (14) I… [knjaz(’) Danilo Aleksandrovicˇ(’)… [Martynka Kosjaka and [prince Danilo Aleksandrovicˇ]- [Martynko Kosjak]-
264
R V
veleˇl obviniti potomu cˇto ego že znaxori, na bid- find.wrong- because his witnesses-. to kotyryx [sic] sja on poslal, pered sud(’)jami .. he- refer- before judges-. skazali, cˇto rozvod mež [teˇmi d(e)r(e)vnjami say-. that demarcation- between [those villages]-. ne byval, a my dei rozvoda ne be- and we- demarcation- veˇdaem. know-.1. And… Prince Danilo Aleksandrovicˇ… bid [the trial judges] to decide against Martynko Kosjak because his witnesses, to which he had referred, said before the judges that there had not been a demarcation between those villages, “and we (dei) don’t know the demarcation” (ASÈI 2: 200, no. 287; similarly in the executed verdict, ibid.).
Since the first part of the report does not contain any elements unambiguously coreferential with the original speech situation, there is no way to determine whether the entire report is complementized or whether there is a slip from nondirect to in the underlined clause. Militating favor of the latter possibility is the absence of the vocative g(o)s(podi)ne. In , slipping typically occurs in extended passages of that, like (14), present several competing deictic orientations and/or changes of the discourse theme. Such midstream changes in the reporting mode probably posed a certain risk of loss of cohesion. In general, self-contained subdivisions of discourse tend to be organized around a single participant’s perspective as a point of departure (“unity of participant orientation”, Grimes 1975: 41, 104–5, 109, 323; see also Brown and Yule 1983: 96). Consequently, changes in perspective within a segment may affect the coherence of a text and deceive the hearer or reader in thinking that a new section has begun. (In some cases, slipping facilitates the writer’s task by eliminating the need for long passages with shifted deixis. However, this puts a greater burden on the interpreter.) In (14), the potentially confusing character of the shift was counterbalanced by the additional reportedness marking provided by deˇ(i), which made the status of the second clause explict and reasserted the reporter’s perspective. In this sense, the seemingly pleonastic use of deˇ(i) and
R V
265
other, functionally similar markers could play an important supplementary role in maintaining textual cohesion in extended . The motivation for the in (14) is not entirely clear; it may be a way of highlighting decisive evidence, since the direct portion of the report is the crux of the judgment. Vološinov (1929/1986: 132) compares the effect of slipping to “those sculptures of Rodin’s, in which the figure is left only partially emerged from stone”: “the subjectivity of speech acquires a heightened definition…” This insight recognizes that , far from being limited to reproduction, can actually be a way of manipulating the meaning of the message to accomplish some context-specific pragmatic goal never intended by the original speaker (if indeed there was any original speaker). As will be shown below, this effect of increased prominence is also a potential result of intercalated citation markers. The co-occurrence of slipping and deˇ(i) in (14) may be a complex strategy for highlighting the aggravating circumstance in the judgment. 8.3.4
Intercalated verba dicendi and other quotation markers
Cohesion in extended reports can be promoted not only by grammaticalized quotation markers like the particle deˇ(i) but also by non-particularized (or semi-particularized) ’s in intercalated position. Intercalation of ’s is widespread cross-linguistically; nevertheless, there have been few studies of its linguistic function. One important observation made in earlier studies is that intercalated elements, by virtue of their intrusive syntactive position, delay the full reception of the surrounding ; in verbal art, this manipulation of narrative time causes various pragmatic effects, up to and including subverting “the primary meaning of the utterance” (Shapiro 1984: 76; see also 73–77; Kieckers 1920: 202; Leech and Short 1981: 333, 351 n. 11; Page 1988: 27). In extended , intercalation has another important property besides intrusiveness. Though intercalated citation markers cause breaks within the individual clauses of a report, they do not disrupt the continuity between the consecutive clauses; they interrupt without obstructing. This property, which distinguishes intercalated tags from preposed and postposed ones, makes them eminently suitable for strengthening cohesive relations within strings of that represent a single act of speaking. In , this effect was furthered by the fact that the intercalated elements were usually placed next to or in the
266
R V
midst of the discourse theme, which was sentence-initial, as in Modern Russian. (Indeed, Lopatina [1985: 59–60; 1987: 240, 243] suggests that one of the functions of deˇ(i) was to single out the discourse theme in the separate segments of a string of . This claim merits further investigation.) By serving as ancillary elements of the theme, intercalated citation markers could reinforce preposed tags by outright replication or paraphrase; hence, like other repetitions, they could promote continuity when the organization of the became diffuse.13 This effect can be seen in a verdict in which the present tense of the kazati is intercalated in the second clause of a string of nondirect speech (15). The first clause of the report has a preposed tag involving the preterit of s”kazati and is complementized with cˇ’to, so that it conforms to the most common pattern of reporting (see 8.3.1): (15) I [knjaz’ velikij Dmitrej Ivanovicˇ… [Ivaška Izbina and [prince grand Dmitrij Ivanovicˇ]- [Ivaško Izbin]- velel sud’e obviniti potomu cˇto Ivaškov bid- judge- find.wrong- because Ivaško-. že [starožilec Vas’ko Roskorjaka pered sud’eju [old.resident Vas’ko Roskorjaka]- before judge- skazal, cˇto s [svoim otcem žil za say- that with [. father]- live- under mitropolitom v sele v Kulikovskom, a [te metropolitan- in village- in Kulikovskoe- and [those požni, kažet, kosili [mitropolicˇi fields]-. say-.3 mow-. [metropolitan- xristiane za pusto… peasants]-. as empty And Grand Prince Dmitrij Ivanovicˇ… bid the [trial] judge decide against Ivaško Izbin because Ivaško’s witness Vas’ko Roskorjaka said before the judges that he lived with his father under the metropolitan’s [jurisdiction] in the village of Kulikovskoe, and those fields, he says, the metropolitan’s peasants mowed, [regarding them] as empty (AFZX 232, no. 259).
While both clauses of this report are drawn from a single turn at speaking in the testimony, they are not adjacent to one another in the anterior report in
R V
267
the trial transcript; moreover, the second clause is actually a compilation of three separate statements (ibid.: 230–31). Thus the report in (15) is at once selective and synthetic (and subversive, since it goes against the reported speaker’s intention); its parts do not necessarily constitute a natural whole. The cohesion of the is further imperiled by the change of theme — from the represented speaker Vas’ko, the (null) subject of the first clause, to the contested properties in the second clause, which contains no further mention of Vas’ko. The fact that the second clause belongs to the testimony of Vas’ko, a witness for the respondent, is crucial; his statement is an admission that the land had been unoccupied before the plaintiff’s peasants took possession — a circumstance that strongly prejudices the respondent’s claims. The first clause of the report, by contrast, is background, which prepares the way for the aggravating circumstance by establishing the speaker’s authority. Thus, by unambiguously indicating the relation between the two parts of the report, the intercalated reinforces the attribution and underlines its legal significance — the fact that the respondent’s sole witness had willy-nilly bolstered the case for the plaintiff. It is not surprising that a seemingly redundant tag should occur in (15) amidst information that is pivotal for the verdict; the repetition of ’s has a foregrounding effect in addition to (or perhaps concomitant with) its cohesion-promoting function. Tannen (1989: 50–51) points out that repetition of similar structures can foreground not only the repeated elements but also the distinct elements that surround them. Similarly, Labov (1972: 379) observes that repetition can have an evaluative function in narrative by emphasizing one particular event and by suspending the action as a whole. Suspension of the action is, as mentioned above, a general property of intercalated tags, which can thereby cause foregrounding because of their retarding effect on the assimilation of the surrounding report. According to Page (1988: 27), attributions can interrupt “in a way that creates suspense or provides emphasis by isolating a word or phrase (cf. the effect in speech of what is customarily referred to as ‘a significant pause’).” Likewise, Shapiro (1984: 67) notes that the delay caused by the intrusive attribution “gives way to a finality that is completed by the reader’s inclination to understand an utterance as tending toward its final stress point” (see also ibid.: 68, 74–77). These effects of intercalation can be likened (as was in Chapter 5) to the Russian graphic convention of putting spaces between the letters of words for emphasis. This foregrounding effect can be seen in the following verdict (16), in
268
R V
which the intercalated — the present of s”kazyvati ‘say ()’ — is the only explicit marker of .14 (16) I [knjaz(’) Jur’i Ivanovicˇ(’)… Stepka da Oksenka and [Prince Jurij Ivanovicˇ]- Stepko- and Oksenko- veleˇl obviniti… potomu cˇto vošli sami bid-. find.wrong- because go.in-. selves-. na [teˇ pocˇinki i v xoromy v to [those clearings]-. and into houses-. into gotovye v monastyr’skye, a dvor’skoi i ready-. into of.monastery-. and steward- and [volostnye xr(e)stijane v [teˇ pocˇinky ix, [local peasants]-. in [those clearings]-. them- skazyvajut, ne sažyvali, ni gramot(y) im say-.3. settle-. nor writ- them- lgotnye na [teˇ pocˇinki ne davyvali… of.privilege- to [those clearings]-. give-. And Prince Jurij Ivanovicˇ… bid [the trial judge] decide against Stepko and Oksenko… because they went in on their own to those clearings and into the monastery’s ready-made houses, and the steward and local peasants did not settle (they say) them in those clearings, nor did [they] give them writs of privilege to those clearings (ASÈI 3: 189, no. 173).
As in (15), the in (16) is intercalated at a crucial point in the report — amidst a damaging admission. The given example differs somewhat from its counterpart in (15) in that it does not occur in either of the standard positions for interruptive elements, i.e., after the first phonological word or after the first major constituent; rather, it splits the predicate of the report in two and thus causes a delay before the most detrimental portion of the message. To be sure, if the tag had been placed in one of the standard positions, it would have counteracted its own disambiguating function; the report would probably be associated with the first constituent of the sentence (“the steward and local peasants say”) rather than with the actual represent speakers. Indeed, this allusion to potentially rival speakers was probably one of the factors that motivated the inclusion of a citation marker right at the point in which the narrated viewpoint became ambiguous due to a change in subject/topic.
R V
269
There is also a case in which an intercalated quotation particle deˇ(i) is the only marker of (17). This instance appears in the verdict of a trial with a default judgment: the judge of higher instance ruled against the respondents because they failed to appear in court after an adjournment, as they had agreed. (17) I [knjaz(’) Andreˇi Vasil(’)evicˇ(’)… [igumena krutickog(o) and [prince Andrej Vasil’evicˇ]- [abbot of.Krutica Galaseˇja i eg(o) poselskog(o) veleˇl obvinit(’) Galasij]- and his manager- bid- find.wrong- togo deˇlja, cˇto poselskoi… vzjal sobeˇ [srok dobrovolnoi because steward- take- . [term voluntary]- na [Nikolin d(e)n’, stat’ bylo emu u ozera to [Nicholas- day]- stand- . him- at lake- u Poleckog(o), igumena bylo emu postavit(’)… u at Poleckoe- abbot- . him- put.up- at otveˇta [starcem storožev’skim; i on” dei sam reply- [elders Storoževskij]-. and he- self- ne stal i [igumena svoeg(o) ne postavil… stand- and [abbot .]- put.up- And Prince Andrej Vasil’evicˇ… bid [the trial judge] decide against the Krutica abbot Galasij and his estate manager because the manager… took for himself a voluntary term until St. Nicholas’ Day; he was to appear himself at Lake Poleckoe and produce the abbot… to reply to the Storoževskij monks. And he (deˇ(i)) did not appear himself and did not produce his abbot… (ASÈI 3: 85–86, no. 55).
Since the reasoning in verdicts can only be ascribed to the deciding judges, the first three clauses of this rationale must be regarded as relayed by the higher judge. Their source is the trial judge’s statement reported at the beginning of the judicial-referral record (see 6.4). If the passage in (17) is compared with this statement (ASÈI 3: 85, no. 55), the first untagged clause turns out to be a reflecting some statement of acceptance, while the second and third are free nondirect speech syntheses. This raises the question of why only the fourth clause was tagged, when all were the voice of the court. The explicit citation marking was evidently motivated, in part, by the fact that the marked portion of the message is the crux of the rationale,
270
R V
while the preceding, unattributed clauses serve only to set the stage for it. Thus the use of deˇ(i) in (17), like that of the ’s in (15)–(16), may be a highlighting technique. Another factor may have been the need for cohesion. The three clauses that open the ratio decidendi cohere strongly because they share a single theme; the second and third elaborate the content of the in the first. By contrast, the fourth clause is less tightly connected to the preceding text because of the shift of focus from the steward’s obligations to his actual behavior. This shift is indicated by the retopicalization of the given referent, as signalled by the pronominal rather than zero subject in the clause tagged by deˇ(i). 8.3.5
Narrative reports of speech acts
The reports in verdicts discussed up to now have mostly been condensations of statements or assertions, the basic ’s found in testimony. However, as seen in the preceding section, other kinds of ’s performed during the trial hearing can also take on significance for the judgment and so be recounted in the ratio decidendi. In accordance with the tendency to reduce information, such illocutions, as a rule, are not relayed as clausal with a ; rather, they are categorized by their lexical label — the corresponding — and converted into past-tense ’s. This evaluative reporting strategy is rare in the objective style of the trial narrative; however, it is common in verdicts, which represent the interpretation of the judge, just as it is in additional layers of within the testimony. In a few cases, the ’s reflect explicit performative formulae, which may or may not appear in the transcript of the testimony; here there is a change of both tense and aspect. For example, the formula in (18a), predicated on the imperfective verb s”latisja ‘cite, refer to’, is frequent in contexts in which the litigants are depicted calling on witnesses or citing documents for corroboration. A preterit renarration involving pos”latisja, the perfective of s”latisja, appears in the ratio decidendi from the same trial transcript (18b).15 (18) a.
I sud’ja vsprosil Stepanka: A ty and judge- ask- Stepanko- and you-. šleš’ li sja na [te muži? I refer-.2 to [those men]-. and
R V
271
Stepanko tako rek: šljusja, gospodine. Stepanko- thus speak- refer-.1 lord- And the judge asked Stepanko, “Do you refer to those men?” And Stepanko spoke in this way: “I refer, lord” (ASÈI 1: 491, no. 595). b. I [knjaz’ Vasilej Ivanovicˇ’… [Stepanka Dorogu and [prince Vasilij Ivanovicˇ]- [Stepanko Doroga]- velel obviniti, potomu cˇto [spisok sudnoj bid- find.wrong- because [copy of.trial]- obolživil i poslalsja na [sudnye muži, da ot falsify- and refer- to [of.court men]- and from sroka s Moskvy zbežal, i s term- from Moscow- run.away- and with porucˇniki… bondsmen-. And Prince Vasilij Ivanovicˇ… bid [the trial judge] to decide against Stepanko Doroga because he had declared the trial record false and had referred to the men of court, and [then] during the adjournment had run away from Moscow with his bondsmen (ASÈI 1: 494, no. 595).
The conversion of the from to narrative can be viewed as part of the more general effort to reduce information in the nondyadic contextualization of the ratio decidendi. A related method of renarration appears in several cases in which one of the sides loses the suit because they refuse to fight a judicial duel. In juridical practice, one party’s witnesses could challenge their counterparts from the other side to a duel; the outcome was considered evidence of God’s justice (Kleimola 1972: 363; 1975: 60–66, 72, 89; Kaiser 1980: 150–51). Such challenges seem often to have been nothing more than a ritual; records of actual judicial duels are rare in trial transcripts (see, e.g., AFZX 110–11, no. 117). One formula invoked in such cases can be seen in (19a): (19) a.
Znaxori, g(ospodi)ne, monastyrskie… witnesses-. lord- of.monastery-. poslušestvovali lživo; daite nam s bear.witness-. falsely give-. us- with
272
R V
nimi [b(o)žiju pravdu:… celovav kr(e)st”, them- [God- justice]- kiss- cross- da leˇzem s nimi na pole go-.1. with them- to field- bitisja. fight- “The monastery’s witnesses, lord, … have borne false witness. Give us God’s justice with them:… after kissing the cross, let us go to the field to fight with them (ASÈI 2: 322, no. 334).
Unlike the formula renarrated in (18b), the prepatterned performative utterances for inviting and accepting judicial duels cannot be converted to narrative by simply substituting perfective preterit verbs for present-tense ones, since the resulting statement would not imply volitionality. Given that the duels were not necessarily fought, the participant’s willingness to fight was the decisive factor in the such cases. Hence duelling formulae are conveyed in the ratio decidendi by means of the preterit of the verb poimatisja, which in its speech-act meaning denotes ‘promise, undertake’ (Sreznevskij, SRJa, s.v.). This conventional substitution can be seen in the verdict of the same document (19b).16 (19) b. I… [knjaz(’) Danilo Aleksandrovicˇ(’)… veleˇl Efreˇma and [prince Danilo Aleksandrovicˇ]- bid- Efrem- obviniti potomu cˇto znaxori ego find.wrong- because witnesses-. his monastyrskie s [Kur(’)janovymi znaxori of.monastery-. with [Kur’jan- witnesses]-. za pole sja ne poimali… for field- undertake-. And… Prince Danilo Aleksandrovicˇ… bid [the trial judge] to decide against Efrem because his monastery witnesses did not agree to fight with Kur’jan’s witnesses (ASÈI 2: 322, no. 334).
In one verdict, the refusal is renarrated by the negated prositi ‘request’ (ASÈI 2: 262, no. 306). In other cases, the preterit verbs used in the renarrations are interpretive labels of indirect ’s. An instance of this type occurs in (18b), where it is stated that the losing litigant declared the trial record
R V
273
false (spisok sudnoj obolživil) in the higher court (ASÈI 1: 494, no. 595). Here the verb obol”živiti (obl”živiti) ‘declare false’ is not a straightforward equivalent of the litigant’s statement from the verification procedure (ibid.: 493) but an analysis of its illocutionary intent. The original report, which is in , could have been transplanted to the verdict without violating clarity or stylistic conventions; the further stage of reduction is used — as often in the verdict — to indicate the significance of the statement as filtered through the judge’s interpretation. This same principle of labelling is found in a verdict that emphasizes the respondent’s act of naming witnesses as background for the crucial information that the witnesses did not bear him out (AFZX 98–99, no. 103). The v”imenovati ‘name’ functions as a narrative substitute for the formulaic expression used to call witnesses (a list of names without any ; ibid.: 98). Since it follows a string of three independently tagged clauses, the represents a slip into a more integrated, evaluative form in a contextualization that favors compactness. Highly evaluative perlocutionary-based ’s are found in several verdicts. Typically, these feature negations of ’s that refer to the illocutionary intent desired by the principal. In some cases, the ’s refer to the testimony of uncooperative witnesses and involve ’s such as posluš’stvovati ‘bear witness’ (ASÈI 1: 548, no. 635); muževati ‘bear witness’ (AFZX 98–99, no. 103); and m”lviti v” ist’cevy reˇcˇi ‘say what litigant says’ (Gorcˇakov 1871, appendix [separate pagination]: 65, no. 4.5). Another perlocutionary-based centers on ocˇistiti ‘prove ownership of’ (literally, ‘clear’) in (20): (20) I [kn(ja)z’ veliki… Kirila veleˇl obviniti, and [prince grand]- Kirilo- bid- find.wrong- potomu cˇto [Danilo Blin [teˇx zemel’… po gramotam because [Danilo Blin]- [those lands]-. by writs-. ne ocˇistil… clear- And the grand prince… bid [the trial judge] to decide against Kirilo, because Danilo Blin did not manage to establish his ownership of those lands… with [his] documents (ASÈI 2: 318–19, no. 333).
The plaintiff Kirilo had based his case on the fact that Danilo Blin had donated the disputed property to the monastery. In the higher court, Danilo
274
R V
produced documents to prove that the property was originally boyar land (ibid.: 318); however, this new evidence was not accepted for technical reasons. The use of the ocˇistiti in (20) probably reflects the judge’s evaluation of the perlocutionary success of Danilo’s statement. Note that the negated perfective in (20) implies not inaction but an action essayed without the desired achievement. Inaction — failure to perform a legally normative — can also be reported as grounds for an adverse judgment. The typical verb used is m”lcˇati ‘be silent’ (e.g., AFZX 107, no. 114); however, more specific negated ’s like bivati cˇel”m’ ‘petition’ can also be found (ibid.). In such cases, silence is viewed as intentional withholding of judicially relevant information. (On verbs of reticence and silence, see Verschueren 1985: 73–121.) 8.3.6
Free reported speech in the ratio decidendi
As mentioned in 8.3, the deciding judges must be the immediate source of the information in the ratio decidendi, although the scribes were responsible for its precise form. Consequently, the entire contents of the ratio decidendi are a form of free , with the attribution (“the judge said”) omitted. At least two factors influenced this omission. First, the causal clause that contains the rationale falls within the scope of the ’s in the adjudication, i.e., the judge has already been established as the sole speaker. Second, since the judge had the last word, the attribution was superfluous; the case would never be subject to further judgment, given the lack of an appeals process. Besides forming the rationale as a whole, free can be used to represent a second layer of reporting within the rationale, i.e., to convey requotations of testimony as outright narrative, without any marking. This is always the case in the directed verdicts issued after judicial-referral hearings, where the deciding judge has received his information at second hand, from the trial judge and/or the written record. In such examples, the conceivable attributions (e.g., the bracketed clauses in “the higher judge… ordered the trial judge to decide against X because {the higher judge said [the trial judge/transcript said]}…”) were unnecessary because the sources could be taken as the voice of the court once the transcripts had been verified in the judicial-referral hearing, as, for example, in (21).
R V
275
(21) I [knjaz’ velikij… Nekrasa velel obviniti and [prince grand]- Nekras- bid- find.wrong- potomu cˇto na srok u doklada ne stal, a because on term- at judicial.referral- stand- and seno s zemli svezl… hay- from land- take.away- And the grand prince… bid [the trial judge] to decide against Nekras because he did not appear at the judicial-referral hearing, and he took hay from the [disputed] land (AFZX 234, no. 261).
The underlined information has two sources in the preceding trial transcript: the authorial narrative of what the scribe himself witnessed; and the trial judge’s report at the beginning of the judicial-referral record.17 Likewise, in (22), a document citation is presented in the ratio decidendi without any attribution. (22) I [kn(ja)z’ veliki… [starca Kirila veleˇl obviniti, and [prince grand]- [elder Kirilo]- bid- find.wrong- potomu cˇto… u kn(ja)ži u [Mixailovy gramoty u because at prince-. at [Mixajlo- writ]- at žalovalnye pecˇjat’ ne [knjaža Mixailova, — of.immunity- seal- [prince- Mixajlo-]- potpisal i zapecˇjatal [gramotu knjažemixailovu sign- and seal- [writ prince.Mixajlo-]- bojarin eg(o) [Grigorei Fedorovicˇ(’)… boyar- his [Grigorij Fedorovicˇ]- And the grand prince… bid [the trial judge] to decide against the monk Kirilo because… the seal on Prince Mixajlo’s immunity charter was not Prince Mixajlo’s; his boyar Grigorij Fedorovicˇ had signed and sealed Prince Mixajlo’s charter (ASÈI 2: 318, no. 333).
The information that the boyar signed and sealed the immunity charter is taken from the verbatim quotation of that document in the judicial-referral record (ibid.); the only changes are a switch from the first-person possessive to the third-person, and the addition of the label knjažemixailovu to prevent confusion. (By contrast, in the later executed verdict, the same documentary evidence appears as tagged by pisati ‘write’; ibid.: 319.)
276
R V
The voice of the court is also heard in verdicts in which the rationale is that a litigant had lost an earlier suit over the same property (ASÈI 2: 462, no. 422); had neglected to petition for redress in a timely manner (AFZX 107, no. 114; ASÈI 1: 541, no. 628); had failed to name new witnesses when his first were discredited (ibid.: 535, no. 492); or was unable to name witnesses altogether, as in (23a): (23) a.
I [Kostantin Grigorievicˇ(’)… išcˇei… and [Konstantin Grigor’evicˇ]- plaintiffs-. obvinil, potomu cˇto… [sami oni find.wrong- because [selves they]- išcˇei i starožilci… plaintiffs-. and old.residents-. And Konstantin Grigor’evicˇ… decided against the plaintiffs… because… they themselves [were] plaintiffs and longtime residents (ASÈI 2: 445–46, no. 414).
b. Kak im, g(o)s(podi)ne, u nas starožilcom na how them- lord- at us- old.residents-. for [tu zemlju byti? starožilci, g(o)s(podi)ne, na [that land]- be- old.residents-. lord- for [tu zemlju my… [that land]- we- “How, lord, could we have them, longtime residents, for that land? The longtime residents, lord, for that land are us” (ibid.: 443).
The juxtaposition of (23a) with its anterior report in the testimony (23b) reveals it to be a clear case of , with the first person replaced by the third. The crux of the verdict was evidently the interpretation of the term “longtime resident”, which in legal usage denoted a special kind of witness (Kleimola 1975: 35–37; Kaiser 1980: 137–38). The judge evidently found an anomaly in the peasants’ claim to be at once litigants and “witnesses”; this crossing of purposes motivated his decision to reject their testimony. Unattributed reports of testimony — more precisely, — occur in some cases in which the crucial evidence against the losing litigants was their witnesses’ claim not to know or remember information about the disputed property (24)–(25).18 Though the reported status of cognitive
R V
277
predicates is explicitly indicated in a few verdicts (e.g., ASÈI 3: 219, no. 208), such marking is redundant when the subject is not in the first person. Since it is impossible to introspect another person’s knowledge, the information conveyed by third-person cognitive verbs must be reported or inferred (cf. Austin 1946/1989: 113–14). The possibility of inferences about knowledge can be ruled out in this context; thus, the reportedness of knowledge predicates is presupposed and can be omitted. (24) I… [Dmitrij Volodimerovicˇ’… Efimka velel and [Dmitrij Volodimerovicˇ]- Efimko- bid- obvinit’, potomu cˇto ego ž starožylci pered find.wrong- because his old.residents-. before sud’eju skazali, paxali [tu zemlju [Ivanovy judge- say-. plow-. [that land]- [Ivan- krest’jane Saraeva iz [Sergeevskie derevni peasants]-. Saraev- from [Sergeevskaja village]- desjat’ let, a Efim molcˇal; a napered ten- years-. and Efim- be.silent- and before Ivana xto [tu zemlju paxal, i one Ivan- who- [that land]- plow- and they- togo ne vedajut… that- know-.3. And… Dmitrij Volodimerovicˇ… bid [the trial judge] decide against Efimko because his witnesses said before the judge Ivan Saraev’s peasants from the village of Sergeevskaja had farmed that land for ten years, and Efim had not complained; and they don’t know who farmed that land before Ivan (ASÈI 3: 292, no. 276).
If the renarration of the witnesses’ statements in (24) is juxtaposed with the anterior report in the testimony (ASÈI 3: 291, no. 276), it is seen that the message has undergone reduction in form and content. The cognitive verb exhibits the typical shift from first person to third, while the temporal reference is, predictably, unchanged. Comparing (24) with its anterior report also reveals that free can be employed in the ratio decidendi in the same way as in boundary-marking descriptions in the trial record — as a way of continuing tagged interrupted by narrative commentary (see 4.7). The clause “and Efim had not complained” (literally, “was silent”) in (24) is an
278
R V
evaluative interpolation that must be ascribed to the judge; it does not appear in the testimony. Its function is to highlight the damaging significance of the information in the preceding, explicitly attributed clause of the rationale. The interpolation occurs at a natural seam in the , since the following untagged admission of ignorance comes from a separate turn at speaking in the trial dialogue. A similar case of involving a cognitive predicate occurs in (25) [2], where the ratio decidendi also contains two further instances of free : (25) I [Grigorej Romanovicˇ’… Martynka… da Ermolku… and [Grigorij Romanovicˇ]- Martynko- and Ermolka- obvinil”, potomu cˇto iskali [teˇx” zemel’ za find.wrong- because seek-. [those lands]-. after sorok” leˇt”; da i starožilci ix”… forty- years-. and old.residents-. their togo ne upomnjat”, [koj mitropolit” [teˇ that- recall-.3. [which metropolitan]- [those zemli poimal” za sebja; a v” [Mixajlovyx” lands]-. take- under . and in [Mixajlo- knigax” Volynskogo [teˇx” zeml’ [starye books]-. Volynskij- [those lands]-. [old derevni pisany mitropolicˇi, a [novye hamlets]-. write- metropolitan-.. and [new derevni stavleny promež” teˇx” že zeml’… hamlets]-. place- amidst those-. lands-. And Grigorij Romanovicˇ… decided against Martynko… and Ermolka… because [1] they had sought those lands after forty years; and [2] their witnesses… can’t remember which metropolitan took those lands under his jurisdiction. [3] And in Mixajlo Volynskij’s books the old villages of those lands are recorded [as] the metropolitan’s [property]; [4] and the new villages have been built amidst those lands (AJu: 17, no. 8).
One of the other cases of free in (25) appears in the third point in the rationale (“And in Mixajlo Volynskij’s [cadastre] books the old villages of those lands are recorded [as] the metropolitan’s [property]; and the new
R V
279
villages have been built amidst those lands”). This is an untagged, condensed renarration of the respondents’ testimony (cf. AJu: 15, no. 8). While the ultimate source of the first part [3], Mixajlo Volynskij’s cadastre book, is attributed by an adjunct prepositional phrase, the second part [4] (about the “new villages”) is an unattributed synthesis of information from two other cadastre books, which are explicitly named in the respondents’ testimony. Indeed, since it is ascribed neither to the ultimate sources nor to the speakers who introduced it into evidence, it is doubly unattributed (or triply, given that the judge is the immediate source). The other case of untagged in (25) is at the onset of the rationale [1], in the statement “they had sought those lands after forty years” — i.e., longer than the three- to six-year statute of limitations legislated by the Lawcode of 1497 (Grekov (ed.) 1952: 28; see also Kleimola 1975: 30–31, 87–88). The plaintiffs did not, in fact, admit that their opponents had been in de facto possession of the land for forty years; this information was furnished, probably unwittingly, by their witnesses (AJu: 15, no. 8). In the filter of the judge’s interpretation — i.e., “taken out of context” — the witnesses’ statement assumed a different significance, which subverted the intention of the original testimony (cf. (23a)). This evaluative use of was neither an accident or an abuse; interpretation is the essence of the requotations in the verdict. Indeed, changes in meaning — the addition of a new perspective, if nothing else — are an inevitable effect of any recontextualization of because of the inseparability of meaning and context (Bakhtin 1975/1981: 340; Sternberg 1982a; 1982b: 108–9; Tannen 1989: 100–101, 105–110). This is true not only of the ratio decidendi, where evaluation is institutionalized, but also of the other sections, including even the testimonial portion with its objective style of narration. As Sternberg remarks (1982b: 108), no matter how objective a reporter tries to be, “the difference… is ultimately one of degree: to quote is to mediate and to mediate is to interfere.” Within this difference of degree, it is evident that the typical manner of reporting in the testimonial portion differs from the one generally employed in the verdict in the demands that it makes on the interpreter. Whereas the standard strategy in the trial record gave the intended interpreters most of the responsibility for sense-making, those used in the verdict often gave them hardly any at all. This was appropriate, given the function of the completed document and of the verdict in particular: it was now essential to leave the
280
R V
potential interpreters — rival claimants or future judges — with no doubt of the represented speakers’ (the deciding judges’) intentions. Moreover, the messages reported in the verdict were already accessible to the interpreters in their fullest form; hence the verdict could be compact and evaluative where the testimonial portion was diffuse and ostensibly objective. The distribution of speech-reporting techniques in the ratio decidendi, as compared with the trial record, is commensurate with its differences of function; in fact, the two contexts exhibit an almost perfectly inverse relationship. The conventional method in the trial record is the diffuse mode of ; and ’s are rare. In the verdict, by contrast, the compact mode of is usual, and ’s are common, while is virtually unattested. The use of untagged in both of these contexts reflects a rapprochement of function: it is frequent in the ratio decidendi, where isolated requotations are used to explain, in brief, the judges’ decisions; it is used in the testimonial portion to explain the actions of the participants with a minimum of delay in the narrative. The compact reporting strategies that tend to occur in verdicts are comparable to those in additional layers of . Both contextualizations are similar in that the is filtered through a single perspective that excludes or dominates all of the other possible viewpoints in its context. Significantly, the same strategies tend to be preferred in other text-kinds such as immunity charters, state orders, and other chancery documents of local legislative application, which were written from the first-person but depersonalized (contextually autonomous) perspective of rulers or other administrative authorities. In such text-kinds one may occasionally find accounts of entire lawsuits that, in sharp contrast to the usual diffuse presentation in trial records, are treated in a compact, writer-based way because a single authoritative interpretation is imposed on the testimony from the onset. This may be seen in an apparently unique example of an immunity judgment charter (žalovannaja pravaja gramota) of the 1470s, a document granting fiscal immunities on the basis of a lawsuit. Like ordinary immunity charters, the immunity judgment charter begins with a of the act that it performs, followed by an account of a petition reported as : (26) a.
Se jaz, [knjaz(’) veliki Ivan Vasilievicˇ(’), požaloval lo I- [prince grand Ivan Vasil’evicˇ]- grant-
R V
281
esm’ [kn(ja)zja Vasilia Ivanovicˇ(a). Cˇto mi -.1 [prince Vasilij Ivanovicˇ]- that me- bil cˇelom o tom: moi de [selcˇane hit- forehead- about that- my-. [villagers dobromer’skie eg(o) [zemlju glumovskuju of.Dobromer’e]-. his [land of.Glumovo]- poorali… k [moe]mu selu Dobromerju… i jaz… plow-. to [my village Dobromer’e]- and I- dal emu… [svoeg(o) s”zoršcˇika Semena Nakvasu, give- him- [. surveyor Semen Nakvasa]- a velel es[mi em]u [teˇ zemli and bid- -.1 him- [those lands]-. obyskati [tamošnimi st[arožil’cy n]a [o]beˇ investigate- [local witnesses]-. for [both storony bez polja. sides]-. without field- I, Grand Prince Ivan Vasil’evicˇ, hereby grant a boon to Prince Vasilij Ivanovicˇ, [regarding the fact] that [he] petitioned me about the following: my villagers from Dobromer’e [] have cultivated his property of Glumovo… [claiming it] for my village of Dobromer’e… And I… gave him… my surveyor Semen Nakvasa, and ordered him [Semen] to investigate those properties by means of local witnesses for both parties, without [resorting to] duelling (ASÈI 2: 506–7, no. 467a).
The judicial process narrated here is also recounted in trial records with the variant incipit examined in Section 4.1, in which a third-person involving biti cˇel”m’ ‘petition’, as in (26a), is followed by a gerund tag plus . In (26a), the opening complaint is presented by means of a strategy that is not characteristic for testimony in trial records — marked with the quotative particle deˇ(i) (see Section 8.3.4). While the complaint, as a nondyadic statement, is not prototypical testimony, it can still have evidentiary value in the trial setting; thus is strongly preferred in that contextualization. However, the plaintiff’s complaint is background information in the context of the immunity judgment charter, which is, in essence, a verdict; hence the diffuse, undigested strategy of is not well-motivated.
282
R V
The opening in (26a) is followed by a narrative of a trial hearing that covers the same ground as a judgment charter, ending with a verdict and a list of the men of court (omitted here): (26) b. I [Semen Nakvasa privezl” ko mneˇ… spisok], a and [Semen Nakvasa]- bring- to me- copy- and v nem pisano, cˇto v”sprašival [moix starožilcev… in it- write- that ask- [my witnesses]-. o [glumovskoi zemli s [moeju zemleju… o about [of.Glumovo land]- with [my land]- about mežeˇ; i oni pered nim” skazali cˇto border- and they-. before him- say-. that mežy [teˇm zemljam ne veˇdajut. A border- [those lands]-. know-.3. and [glumovskix starožilcev v”sprašival… i oni [of.Glumovo witnesses]-. ask- and they- skazali mežu [dobromer’skoi zemleˇ z say-. border- [of.Dobromer’e land]- with [Glumovskoju zemleju… dorogu; da i veli [of.Glumovo land]- road- and lead-. [dorog]oju, da z dorogi […] i [moi sta[rožilci [road- and from road- and [my witnesses]-. i s [moimi] selicˇany za nimi po and with [my villagers]-. after them-. along [toi mežeˇ ne poeˇxali. I Nakvasa pered [that border]- go-. and Nakvasa- before [moim sudom] [oboix starožilcev postavil; i [my court]- [of.both witnesses]-. put- and oni peredo [mnoju to ž] skazali, [kak they-. before [me- that- say-. [as v” [Na]kvasineˇ spiskeˇ pisano. I jaz… potomu in [Nakvasa- copy]- write- and I- therefore Vasilja opravil, a dobromer’cov… Vasilij- find.right- and Dobromer’e.men-. [obvinil]. [find.wrong-
R V
283
And Semen Nakvasa brought the [trial] record to me…, [1] and in it was written that [2] [he] asked my witnesses… about the border of the Glumovo property with my property… [3] And they stated before him that they did not know the border between those properties. [4] But he asked the Glumovo witnesses, [5] and they said that the border of the Dobromer’e property with the Glumovo property [was] the … [lacuna — DEC] road; and they also led [him] along the road, and from the road… [boundary-setting description with lacunae — DEC]. And my witnesses and villagers did not follow them along that border. And Nakvasa placed the witnesses of both parties before my court; [6] and before me they said the same thing as is written in Nakvasa’s [trial record]. And I… for that reason found in favor of Vasilij and against the Dobromer’e men (ibid.: 507).
While the narrative in (26b) resembles trial records in content, it resembles the ratio decidendi (apart from the use of the first person) in its preference for less-than-direct speech — complementized indirect or nondirect speech ([1], [3]); fused speech ([5]); and ’s or syntheses of content ([2], [4], [6]). While the account of the hearing includes representations of both the questions and the answers, it cannot be considered dyadic in format, since it relays only the theme of the questions, not the message ([2], [4]); moreover, the represented answers appear to be conflations of several statements ([3], [5]). All of these reports are filtered through the perspective of the grand prince; in addition, they are all backgrounded, since they are recorded not to supply information for a verdict but as an ex post facto record of a favor granted to a petitioner. In many ways, the style of reporting in (26a–b) and in the ratio decidendi anticipates the patterns seen in trial records of more than a century later. In the late sixteenth and seventeenth centuries, the judges’ interpretation of documents in camera came to assume greater importance than testimony offered in face-to-face confrontations; hence more and more of the space in trial records was given over to the judges’ reasoning. Accordingly, oral testimony was relegated to a primarily corroborative role in the judicial process and was increasingly presented not in the diffuse, undigested form of but in the less-than-direct reporting modes characteristic of compact, writer-based contextualizations.
C 9 Conclusions
9.1
The purposiveness of reporting in trial transcripts
At the beginning of this work, I invoked one of the central methodological problems of historical-pragmatic research. With only written texts as a guide, can one really reconstruct aspects of pragmatic competence and so reanimate, as it were, the silenced voices of premodern writers? Can one actually come to understand the patterns of variation in the writers’ texts as their intended interpreters would, in a way that adequately approximates the “emic” framework of the original communicative context? To demonstrate that the answer to these questions can be in the affirmative, I examined the distribution of speech-reporting strategies in a large corpus of medieval Russian trial transcripts, documents in which was the principal means of conveying relevant information. The availability of multiple formal alternatives for conveying the single “semantic motif” of heteroglossia raised the pragmatic issue of intention.1 Accordingly, my goal was to recover the factor(s) that motivated the writers to choose one strategy over the others in a particular context. These factors, I hypothesized, would reflect purposive language behavior and hence pragmatic competence, since it was evident that the choice generally was not determined syntactically. It was assumed that the writers were rational agents striving to accommodate the needs of the interpreters served by the texts, in accordance with culturally and institutionally specific notions of relevance and cooperation. The primary method of analysis, matching forms to functions, was facilitated by the nature of the trial transcripts that comprised the corpus. These utilitarian documents permit a fine-grained analysis because they incorporate several recurrent contextualizations, which had functions that could be inferred from the known institutional purposes of the text-kind.
286
R V
These traits minimized the element of conjecture, which is always a potential danger when implicit intentions are invoked to explain observable behavior. The intentions involved in my explanations of function are recurrent and institutional (shared) rather than purely individual — a feature that is generally not present in belletristic writings, the focus of many previous studies of . The posited functions themselves are likewise not conjectural because they derive from the ostensible, formal properties of the reporting strategies; in many cases, they receive additional typological support in that they accord with functions ascribed to comparable strategies in other languages. The results of form-to-function matching, which will be detailed in the following section, showed that the writers strongly preferred a single strategy in each of the major recurrent contextualizations. While conventional, these preferences were neither random nor arbitrary; rather, they were contextsensitive and purposive, in that they coordinated the interpretation in a way that met the needs of the intended audience. The scribes chose them based on their experiential knowledge of the interpretive process — not because the strategies were mechanized clichés but because, by virtue of their formal properties, they produced effects that were appropriate to the communicative function(s) of the given portions of the transcripts. This is not to deny that syntactic factors were sometimes present. For example, fused was only possible with specific types of sentences in legal-administrative writing, and the use or omission of a complementizer with seems to have been influenced in part by clausal structure. However, even in these cases the scribes had alternative strategies at their disposal; their decisions were motivated chiefly by the pragmatic effect that they desired. Reporting is fundamentally a coordinative process, which cannot be understood property without reference to the interpreter. While every major contextualization in the transcripts had a preferred reporting strategy, there was always a residue of unconventional cases, which could be matched to text functions by attention to the context and application of the method of residual forms. It was determined that, like the norms, the exceptions generally had a functional motivation; they were most often used to index represented utterances that departed in various ways from what could be identified as the prototypical (perceptually basic and central) speech event for the given stage of the trial process and hence for the given section of the written transcript. An iconic (isomorphic) principle proved to be at work: the standard reporting strategies were used consistently with prototypical speech
C
287
events and also, by convention, with many nonprototypical ones; departures from the normative patterns were used exclusively with nonprototypical speech events and became more frequent in proportion as the represented utterance deviated from the prototype. It was also suggested that, for all their contextual unexpectedness, the departures were not necessarily associated with foregrounding in any of the usual senses of the term; rather, their grounding effect was determined by the contrast between their formal properties and those of the contextually expected strategy. For example, in the idiosyncratic criminal-trial transcript examined in 4.7, and were used unexpectedly in the midst of contextually expected to signal elements of the report that were of less immediate interest — a backgrounding function. The results of my investigation have several immediate ramifications for historical-pragmatic methodology. First, they validate the use of the method of residual forms for analyzing variation in discourse units in a large corpus and, in general, point to the need for fine-grained attention to sometimes minute features of context. Second, they demonstrate the use of the concept of prototypes in a new sphere — the relations between representational conventions and perceptions of basic speech acts in represented institutionalized speech events. Third, they provide additional evidence that conventions tend to be functionally motivated (cf. Lewis 1969; Brown and Levinson 1987). Thus an adequate approach to discourse variations cannot afford to dismiss conventions or operate with automated, unmotivated defaults; rather, it is necessary to look for purposiveness in statistically predominant or “unmarked” patterns as well as in “marked” ones. The context-sensitivity of reporting strategies in medieval Russian trial transcripts provides further evidence for the view that literary techniques have their basis in ordinary language use.2 It shows that even the most utilitarian premodern texts can reflect the kind of purposive usage that has often been treated as the prerogative of modern verbal art. In addition, the purposiveness of the reporting invalidates the widespread notion that medieval Russian chancery texts are unpolished, artless, and too bound by cliché and mundane reality to permit the purposive use of linguistic devices — a claim that can be found even in authoritative studies such as the following: “In contradistinction to the sophisticated structure of the high language [sc. Church Slavonic — DEC], legal texts display simplicity, a frequently poor syntactic organization and, sometimes, an awkwardness in the wording of comparatively simple ideas” (Issatschenko 1980: 122). The
288
R V
impersonal, effaced style of narration that was conventional — and, what is not always acknowledged, functional — in most kinds of chancery writing may have contributed to this dismissive attitude. The work of the scribewriter has often been confused with that of the scribe-copyist. Chancery writers have been regarded not as authors with a say in the organization and wording of the texts but as a mere amanuenses whose speech will is not involved in the production, except perhaps in the most trivial respects.3 As my study shows, this view stems from the deficiencies of context-blind interpretations rather than from any impoverishment in the medieval texts themselves.4 The same can be said for the anti-uniformitarian notion found in some treatments of the history of Russian (e.g., Molotkov 1958; Rinberg 1985) that evolved from a primitive state in to greater complexity and “flexibility” in .
9.2
The main functions of the reporting strategies in the corpus
The scribes who wrote the investigated trial transcripts drew from a rich inventory of reporting strategies. After deciding initially to report rather than aver, the scribes could choose from among four modes — , , ’s, and fused , a small-clause construction not found in — and from a variety of tag formulae, which could be preposed to or intercalated in the reports. In addition, with and they could choose between complementized and uncomplementized and between tagged (attributed) and untagged (free) varieties. Though has been treated in previous studies as a postmedieval innovation, it is, in fact, well attested in medieval trial transcripts, and it can even be the dominant reporting strategy in other premodern text-kinds, e.g., third-person memoranda summarizing orders (nakazy) to the tsar’s officers.5 The distribution of the various kinds of suggests that they are functionally polarized into two broad types, in accordance with their effect in creating prominence in the discourse. Reporting strategies may be classified as relatively compact in proportion as they blend in with the surrounding nonreported discourse, and as relatively diffuse in proportion as they stand out from it. This salience-related distinction is not equivalent to the traditional dichotomy of and , since it is not defined by deixis alone; nondeictic factors also can have the effect of integrating reports into the authorial discourse or, conversely, singling them out.6 Another reason why the
C
289
distinction should not be equated with (in)directness is that it is also relevant to discourse strategies other than . Compact reporting draws mostly on strategies in which inset reports are deictically homogeneous with their frames — not just in its different varieties but also ’s, fused , reports that are deictically indeterminate (“nondirect speech”), and, in other text-kinds, nominalizations and adjunct prepositional phrases. These “less-than-direct” varieties tend to cluster in the transcripts in contextualizations where is not favored. The quality of compactness (homogeneity with the authorial discourse) is scalar; it increases when the reports are not demarcated from the narrative, as in ’s and , or when they are grammatically integrated in narrative predicates, as in fused . While direct deixis is typically associated with relative diffuseness (heterogeneity with the authorial discourse), complementized is apparently more compact than other kinds of . The level of diffuseness increases when additional means are used to single the reports out — e.g., in attributed , especially when the tag is prominent rather than parenthetical; involving supplementary nondeictic markers of the represented speech event; and, presumably, cases in which the is in a different code or style than the authorial discourse (a situation that is not attested in trial transcripts). In addition, the scope of diffuseness effects can span more than one report; this is evidently the case in dyads, since the enframing questions serve as “build-ups” for the responses, which are thereby singled out from the authorial discourse. This explains, in part, why the more integrated strategies hardly ever occur in dyadic frames. The distinction between compact and diffuse types is more than just a convenient way to label cooccurrence patterns; it reflects two different orientations in the process of coordinating the interpetation. Reports that blend with the authorial discourse are generally interpreted with less effort than those that stand out. Thus the connection between salience and diffuseness may be explained as an iconic principle: in cooperative, relevant communication, elements that take more effort to interpret are meant to be perceived as more important (i.e., worth the additional effort). While this principle can be utilized or flouted for literary stylization, it is grounded ultimately (and immediately, in utilitarian text-kinds such as trial transcripts) in a practical concern — the generically or institutionally defined needs/ responsibilities of the interpreter. In Lakoff’s (1984) terminology, diffuse
290
R V
reporting is more hearer-based — that is, it requires more effort on the part of the interpreter, who then has a greater role in sense-making; by contrast, compact discourse is more speaker-based. The institutional purpose of the trial record — the “dialogue discourse” that recounted the on-site trial hearing — was to provide judges with detailed, impartial accounts of the on-site hearings as raw material for their verdicts. Accordingly, in this contextualization the conventions of reporting were directed at maximizing the interpretive contribution of the judges while minimizing that of the scribes, who generally avoided evaluations and effaced themselves to the point of anonymity. This hearer-based orientation is seen first of all in the choice of a diffuse strategy — signalled by a preposed narrative tag clause — as the standard way of reporting testimony. By virtue of its formal properties, typically functions not as an analysis but as a quasi-dramatic reenactment of events, even though under ordinary circumstances it cannot be considered a true facsimile (see the discussion of verbatimness, above). Consequently, the occurrence of in the trial record compelled the interpreters to adopt the perspective of the participants in the represented speech event in order to impose coherence on the . In other words, the interpreters were obliged to hear the trial hearing as it were from inside, without any apparent interference from the reporters; they became, in effect, witnesses of the hearings, like the judges who observed the actual trials, with no other vade mecum than their own reconstructive abilities. Thus the standard reporting strategy matched the organizational principles of the actual trial hearings, in that it allowed “every man… to tell his own tale and plead for himself so well as he can”, as Dr. Giles Fletcher, an early English ambassador to Muscovy, wrote in Of the Rus Commonwealth (1591/1966: 71). The hearer-based orientation that motivated insets is further reflected in the extended frame. The use of standardized preposed tags, though often redundant, ensured that the interpreters could keep track of the positions of the principals and established a predictable structure that facilitated quick recognition of represented utterance boundaries. (It is noteworthy that the recˇi, which appears in the standard tag clause, is infrequent in additional layers of within the testimony; presumably this is due to the fact that it could have been misinterpreted as a signal of a new turn at speaking in the authorial narrative.) Likewise, the diffuse dyadic format, with the judges’ questions in , was not chosen for the sake of verisimilitude, though it
C
291
mirrored the structure of the actual hearing; rather, it permitted the scribes to supply an ample frame for testimony independent of the narrative. This, in turn, compelled the intepreters to situate the testimony in a context consisting primarily of represented speech events and thus to perceive it much as the participants in the hearing did, though in a different medium. In addition, the questions topicalized elements that the interpreters could process quickly while focusing on the new information in the responses. Moreover, since statements were conventionally followed by questions and vice versa, the dyadic format provided many additional clues such as modality or the occurrence of specific vocatives to supplement the delimitive function of the tag clauses. The hearer-based orientation also motivated other features of trial records — in particular, the spare, covert style, reflected in the preference for minimal lexical elaboration and paratactic organization (cf. Lakoff 1984). Paradoxically, the existing treatments of syntax typically treat parataxis as if it were “simpler” and less “sophisticated” than hypotaxis, even though the latter actually makes fewer demands on the interpreter. Likewise, the use of “dialogue discourse” and minimal narrative limited the possibility of authorial intrusions. The scribes concentrated important information within the and conveyed events without indicating how they unfolded and without analyzing their relative weight by departures from real-world sequencing. This gave the interpreters evaluative responsibility for making connections between the represented events (cf. Rader 1982: 192) and so fulfilled the social function and, in particular, the cooperative norms of the text-kind. The use of a third-person narrative style likewise assisted the interpreters by ensuring that any first- and second-person elements could be taken as signals of . As mentioned in 9.1, violations of the standard pattern in a given contextualization occur chiefly with reports that depart from the prototypical speech event in that contextualization. The prototypical speech event in the trial record was testimony, which in this context must be understood in a narrow sense: an uninterrupted statement by one of the principals, which provided information intended to establish the speaker’s rights to the disputed property and was elicited by the judge(s) in a face-to-face exchange, presumably in a standard configuration determined by judicial custom. There can be little doubt that this speech event was perceptually central in the trial institution; no other kind of utterance was as likely to supply information
292
R V
that could be probative for the verdict — proofs of ownership offered by the parties who were interested and legally liable in the lawsuits. (There was, of course, a different prototyical speech event for the judges — requests for information about the disputed property, addressed to and cooperatively answered by individual participants, again in a face-to-face exchange.) The violations of the standard in the trial record include cases in which the litigants’ remarks were not directly concerned with providing evidence, or when they were presupposed, limited in content, or transparent in attribution. Other examples suggest that unconventional reporting strategies could indicate that the represented speaker was on an unexpected footing in the trial, as when two judges were asked to give evidence, or when witnesses testified against their party’s interests or made statements outside the hearing. Another type of nonprototypical speech event is the opening complaint, which ordinarily deals with the respondents’ malfeasances rather than with the plaintiffs’ own proofs of ownership. Since this nondyadic contextualization is relatively close to the prototype, the standard strategy predominates. By contrast, in certain other nondyadic situations, there is a distinct preference for less-than-direct strategies — and fused . This is the case in the incipit and in narrative accounts of reconvenements, where reports generally convey backgrounded orientation information. Similarly, ’s are common in reports of nondyadic stage-managing utterances by both the litigants and the judges. Since reports of this kind usually had no evidentiary value for the verdicts, it would have been a violation of relevance to present them in the detailed form of , which would have brought the narrative to a standstill. The judges’ procedural commands are generally followed not by dialogic responses but by citations of documents or narrative passages; this obviated the framing function ordinarily performed by the judges’ . When they do serve, in accordance with the prototype, as frames for verbal action such as testimony or refusals, they tend to be presented in the conventional way, as . The of the witnesses is conveyed more diversely than that of the litigants; this is explained by the fact that the statements of these secondary participants are not prototypical testimony in the sense defined above. The witnesses were on a different footing in the trial than the litigants; they spoke on oath, often acted as a group, and were only responsible for evidence contained in their own memories. The witnesses’ evidence generally falls into two categories: boundary-setting descriptions, which differ from
C
293
prototypical testimony in being nondyadic and largely nonverbal; and corroborations, which, unlike prototypical testimony, involve assertions that are largely or wholly presupposed. In boundary-setting descriptions, there were two factors that were disincentives to the conventional strategy — rapid gear-shifting between actions and words, and emphasis on the narrated sequence of landmarks. Foregrounding the reports or otherwise retarding the narrative would have been effects incommensurate with the evidentiary value of the ; hence the scribes had recourse to strategies that worked against salience and/or promoted seamless transitions — tagged by gerunds, nondirect (deictically indeterminate) speech, , and fused speech. This backgrounding was appropriate for reports perceived as commentary to movements and gestures. In the sporadic cases of , the evaluative character of the reports prevented them from being interpreted as the averrals of the covert scribe. However, generally speaking, was too inexplicit to be favored in trial records; it appears more often in boundary-setting descriptions in amicable settlements — documents where the attribution and representation of discrete viewpoints carried little weight. Corroboratory reports tend to be presupposed — a factor that promoted economical, compact strategies rather than the standard one used for testimony. In the situation termed multiplex testimony, allied witnesses made statements that overlapped in content, to a large extent. In order to minimize redundancy, the scribes often combined the shared messages into a single, as it were choral report in or , or else indicated the unity of the statements by tagging the first report and using for the others. The use of reflects a general tendency to present presupposed messages as complementized , a strategy that renders the main-clause attribution, which is asserted, more salient than the subordinated report. The same tendency to convey predetermined, overlapping messages as complementized can be seen in judicial-referral records, where the prototypical speech event involving the litigants was a choral verification of the trial record rather than testimony in the strict sense. Such corroborations, whose content was circumscribed, in effect, to yes or no, functioned as background information for the main focus of the end protocols, the verdict section. Nevertheless, the standard strategy associated with testimony is also attested, particularly in earlier documents; while this may reflect stylistic drift, it was probably also influenced by the dyadic structure of the verification and by the evidentiary value that could be ascribed to the litigants’
294
R V
comportment in the hearing. Indeed, even in later documents, there is a tendency to use , often introduced by the standard tag formula used with testimony, whenever the statements are unpredictable in any way. These include falsifications of the trial record, retractions by formerly refractory litigants, verifications by the men of court, and statements of new evidence. The same correlations of with unpredictable messages and less-thandirect (compact) with presupposed ones are reflected in cases where the trial judges are represented making preliminary reports before the higher courts. These nondyadic reports were perceptually distinct from prototypical testimony; the fact that they conveyed preliminary information favored their being presented in a condensed manner. When the reports contributed no new information, they were given in the maximally compact, non-descriptive form of ’s; however, when they concerned no-shows, whose bona fides was suspect, they were presented in a more diffuse way — as or (in one case slipping into ). In general, the scribes had greater latitude for evaluative reporting in the juridical-referral section than in the trial record, because they no longer needed to orient their discourse to interpreters who would have the responsibility of handing down verdicts solely on the basis of the reported information. An even greater tendency to condense and evaluate reports can be seen in verdict records, which did not inform pending decisions so much as regulate future behavior. With the addition of a verdict, trial transcripts became quasi-legislative texts of local application. The judges’ version was the sole authoritative interpretation of the testimony; there was no institutional need to present already processed reports diffusely, in a hearer-based manner that would allow potential interpreters to arrive at independent readings. Accordingly, verdict records feature language that is maximally autonomous — as independent of context and background knowledge as possible — to reduce the possibility of misunderstandings in discontinuous transmission (cf. Kay 1977; Rader 1982: 186). The writer-based style in verdict records favored reporting strategies that ceded little of the interpretive work to the reader. Reports of adjudications were invariably conveyed by ’s, a strategy that reduced messages, integrated them into the narrative, and classified them, leaving little material for independent evaluation. However, avoidance of syntactic complexity could prompt slipping from the ’s into less integrated modes when the adjudications contained additional, unpredictable provisions.
C
295
One such case of slipping is the ratio decidendi optionally appended to adjudications. Typically framed as a causal clause, it is in its entirety a free indirect (or nondirect) report of the judges’ remarks, with the attribution omitted because it was predictable. Reports within these causal clauses are never presented dyadically, since the framing function of the trial judges’ becomes superfluous once an authoritative interpretation has been reached. The use of conflicts, in general, with the style of the ratio decidendi, since it involves diffuseness and viewpoint shifts that could have no informational value in the context of the verdict. Instead, reports tend to appear in less-than-direct, bound, or integrated forms — tagged indirect or nondirect speech, often in the complementized form associated with presupposed messages; and fused speech, which eliminate gear-shifting; and ’s that classify illocutionary intent or perlocutionary effect. This pattern of compact reporting was motivated first of all by the fact that the reports were not asserted as new evidence but repeated as background information to shed light on the judges’ reasoning. (Indeed, the entire ratio decidendi is backgrounded as a subordinated causal clause.) The fact that the reports appeared in a subordinate clause may also have promoted reductive strategies such as fused and uncomplementized as a way of minimizing syntactic complexity and, in particular, multiple embeddings. However, this was offset, to some degree, by the need to impose a single interpretation on the in the verdict, which favored the lexical explicitness and syntactic clarity of autonomous language. In this connection, it is significant that the less explicit form of uncomplementized tends to occur precisely when additional factors are present to obviate the demarcational function of the complementizer — e.g., marked word order, strong cohesive relations, or disambiguating elements within the reports. The reporting strategies favored in the ratio decidendi were also preferred, in general, in the additional layers of speech that occurred within the reported dialogue in the trial and juridical-referral records. This is not surprising, since both of these contextualizations reflect non-effaced, controlling viewpoints. To be sure, the reporting patterns in the two contextualizations do not fully coincide; there are divergences linked with differences in the characteristic style — the objective, third-person voice of the law in the ratio decidendi, and the subjective first-person voice of participants in the additional layers of . This explains, for example, why was little used in the additional layers, where lack of attribution could lead to undesirable
296
R V
confusion between two overt viewpoints. (Paradoxically, the subjective viewpoints were communicated only through the mediation of self-effacing narrators. However, the principals were able to protect their own interpretations during the juridical-referral process, in the case of additional layers of , or during the signing of the completed trial transcript, in the case of the ratio decidendi.) The most frequent reporting mode in the additional layers is complementized indirect or nondirect speech, which can be tagged by ’s or by evaluative ’s. This strategy was preferred when reports were presupposed and served as background — preambles to questions or material about to be refuted. It was also used in cases of hearsay, when the litigants or witnesses cited others’ statements non-dyadically to support their testimony. In “ancestral hearsay” (statements that evidenced community memory), the messages were partially presupposed — a factor favoring , as in corroborations in general. In all these cases, the preference for less-than-direct speech was motivated in part by the need to avoid ambiguity between the deictic orientation of the inset report and that of the surrounding discourse, which was also reported. In some cases, the additional layers of involve further integration and reduction. Fused occurs in the same kinds of non-dyadic, backgrounded contexts as ; it tends to be used with certain types of statements that fulfill the syntactic precondition of having copular predicates. By virtue of its single-clause structure, fused speech allows for rapid interpretation by promoting quick transitions to and from the surrounding non-reported discourse. Also widespread are ’s, whose evaluative character matches the general nature of the contextualization. Both these strategies maximize the interpretive role of the speakers and minimize that of the audience. Additional layers of occur only in sequenced, foregrounded narrative. There are a few examples that are tagged with the standard formula associated with testimony; these appear in dyadic frameworks and reflect legal or quasi-legal exchanges. In general, however, is atypical in additional layers. Not only was it susceptible to misinterpretation as a new turn at speaking, it posed the risk of conflicting first- and second-person references. Moreover, it provided information superfluous in a speaker-based style and drew unnecessary attention to the statements per se. In both additional layers of and the ratio decidendi, the preference for a speaker-based style promoted the use of intercalated citation markers to
C
297
reinforce the function of preposed tags in multiclausal reports, where each new clause could potentially be mistaken for the beginning of a new report or a resumption of the narrative. The intercalated verbs or particles served to disambiguate the discourse by explicitly indicating the connections between successive clauses of . The redundancy of such tags promoted cohesion within the reports; their intercalated — intrusive but not obstructive — position made them iconically suitable for strings of conveying single acts of speaking. In sum, the preferred patterns of reporting in the trial transcripts investigated are neither random nor “merely” conventional; they are purposive, motivated primarily by the socially dictated communicative functions of the individual sections or recurring contextualizations and, ultimately, by institutional (generic) notions of cooperative behavior. As a rule, the strategies are correlated with the degree to which cooperative norms permitted the scribes to assume control of the interpretive process — a factor that also influences the overall structure of the discourse. One may see a general division between relatively diffuse, hearer-based styles, which give the intended interpreters the greatest share in the sense-making process, and relatively compact, author/writer-based styles, where there is no institutional requirement of deferring to an authoritative interpreter — that is, where the scribes can digest the reported information themselves without violating norms of relevance or cooperation. The subsequent development of the text-kind suggests that new norms of cooperation arose in the later sixteenth and seventeenth centuries, in accordance with changes in the trial institution. During this period, lawsuits were conducted mainly through memoranda and petitions; verbal confrontations ceased to be the chief conduit of evidence, and hearings became largely corroboratory. In transcripts of this period, compact, writer-based styles involving less-than-direct speech become more prevalent, since the judges no longer needed transcripts of undigested testimony for their decisions. I would suggest heuristically that the pragmatic factor of control or deference to the interpreter lies behind many of the effects of reporting strategies that have been identified in previous studies — for example, the vividness and immediacy characteristically associated with , or the effect of intervention attributed to other modes of reporting. Indeed, it may be the overarching phenomenon of which the effects are contextually variable (“Protean”, cf. Sternberg 1982b) epiphenomena.
298
R V
Thus the use of form-to-function linking within a genre-based approach permits one to recover purposive behavior in premodern texts and to reconstruct, without speculation, collective intentions to which individual writers had to subscribe as part of cooperative communication. Subsequent application of the method of residual forms shows that even the exceptions to the conventional patterns of reporting are purposive. Far from being random or merely reflecting the influence of decontextualized systemic pressures (e.g., diachronic developments), they are often motivated by unexpected or nonprototypical features of the represented speech event that the writers could perceive as meriting special treatment. In other cases, the exceptions may reflect the scribes’ attempts to avoid undue syntactic complexity. Both of these motivations are ultimately oriented to the needs of the intended interpreter, so that the exceptions, like the conventions, for the most part reflect efforts to maintain relevance and cooperation.
9.3
Some methodological implications
This case study of reporting in medieval Russian trial transcripts has at least two larger implications for historical-linguistic methodology. First, it demonstrates a rigorous, non-arbitrary method of recovering the crucial factors of intention and purpose in premodern writings. The method approaches the data from two directions — from the observable formal properties of discourse strategies, which are seen to be conducive to specific textual functions; and from the needs of speech genres and text-kinds, which motivate one or a subset of those functions in a given context. These institutional needs or aims provide a social locus for intentionality that lies not in the unrecoverable whims of individual writers but in the norms of cooperative behavior in institutional situations. The second ramification follows from the generally purposive nature of reporting as one kind of behavior involving large-scale textual units. It was shown that the choice of reporting strategies could, in virtually all cases, be linked with textual functions that matched the needs of the given contextualization. The method of residual forms proved that this was true even in cases when the strategies in question were infrequent — perhaps even statistically insignificant — in the corpus. If this is any indication of what may be expected in other cases of alternative discourse strategies, it reveals
C
299
a possible pitfall for approaches that pay insufficient attention to often finegrained details of context and genre. To put it briefly, it is premature to interpret the nonoccurrence or scarcity of a unit of syntax, discourse, or lexicon as an evolutionary indicator before one has determined whether there are likely contextualizations for the unit in the attested texts. The crux is, of course, how to define what the likely contextualizations are; typological data cannot be more than suggestive here, since the functions of higher-level units can vary from one language to another. In fact, one is compelled to examine an exhaustive or, where that is not feasible, at least a representative range of contextualizations; yet in order to decide what is representative one must have a good working knowledge of the repertory and reading conventions of the text-kinds produced in the given culture, and hence of the culture itself. It is not difficult to find object lessons to justify this cautionary note; indeed, some may be found in the history of the study of . The existence of in premodern Russian was overlooked — or, perhaps, not looked for — not only because of false a priori assumptions about its modern, Western European origin but also because of the decontextualized nature of most previous approaches to reporting (as to discourse in general) in medieval texts. In order to recognize , one must be familiar with interpretive frames and with the reading conventions of specific text-kinds; however, most of the previous investigations drew their conclusions about from isolated sentences — a dubious unit in medieval discourse — culled from miscellaneous sources. While studies that cull examples from indiscriminately chosen texts can accidentally cover a large number of contextualizations, they are unlikely to give any weight to the functional, generic, and contextual factors that can motivate variation. In any case, assembling a corpus that represents a sufficiently broad spectrum of functions — when it is possible at all — requires extensive background research, which is not exclusively linguistic and does not always lead to the immediate location of the sought-for tokens. This preliminary work is often omitted; consequently, conclusions have sometimes been built on the shifting sand of unrepresentative data. It has been said with some justification that the study of discourse “is of necessity the study of particularity” (Becker 1984: 435) — an observation that, without rejecting the validity of generalizations per se, warns against overgeneralizing neglect of contextual detail.7 The results of my case study
300
R V
indicate that historical-linguistic explanations involving higher-level units — especially ones based on statistical distributions — need to be grounded on nuanced synchronic investigation of contextual factors, including in particular the institutional needs and functions of genres. Without this pragmatic spadework, one risks not only overgeneralizing but also misunderstanding or misusing the evidence. This possibility is illustrated by the pragmatic complications involved in the apparent lexical shift discussed in 4.2. It is generally agreed that there is a tendency in the history of Russian to replace the recˇi, thought to be obsolescent in spoken language of the Muscovite period, with s(”)kazati, the standard perfective in later centuries (cf. skazat’). While one may view the sporadic occurrence of s(”)kazati instead of recˇi in trial records as part of this development, my study demonstrates that both the putatively conservative and the putatively innovative tokens have pragmatic functions that are independent of the lexical level; that is, they are motivated by more than just random synchronic dynamism as a stage in lexical change. Presumably this observation could be extended heuristically to the mutual distribution of these verbs in other text-kinds. This is not to deny that s(”)kazati eventually displaced recˇi; however, in any synchronic slice there may be more to the patterning of these verbs than meets the eye, and what might be taken as evidence for the shift may bear an entirely different interpretation. Clearly, then, one must proceed with caution when offering explanations for the distribution of alternative units of syntax, discourse, or lexicon that are based on purely systemic factors or on drift alone. Many previous works on the history of and other discourse phenomena in Slavic languages, including Russian, offer explanations based not on the functions of the attested strategies but on the assumption that the distribution reflects diachronic “competition” among the alternatives — in other words, that it is not purposive. This is simply fallacious, a form of begging the question; cooccurrence is to be expected, given the observable multiplicity of discourse functions. Moreover, disparities in the distribution of strategies cannot be taken as sure signs of competition or change. If one were to view the reporting patterns in trial records on an exclusively statistical basis, without sensitivity to context, one might conclude erroneously that strategies such as the s”kazati or were marginal and/or at some extreme of development (incipience or obsolescence). In fact, many scholars have denied the existence of in the earliest Slavic writings precisely on the basis of its relatively scanty
C
301
attestation as compared with , without pausing to consider the functional and cultural reasons for its scarcity (see D. Collins 1996 for criticism). The same error lies behind the argumenta e silentio of scholars such as Bondar’ (1967) and Rinberg (1985: 50–51), who claim that the inventory of verbs that could introduce was initially restricted virtually to ’s and expanded only slowly. In fact, the rarity and often late attestation of tags in texts do not necessarily imply that such devices were unavailable; one must pay careful attention to the relationship between function and context here. Even in , such devices are virtually restricted to highly emotive literary genres. The covert, documentary style of narration of much writing and, in particular, the utilitarianism of trial transcripts and other legal-administrative text-kinds would not have favored the use of graphic introducers. However, a first-person letter of admonition written by Metropolitan Filipp in 1471 (AI 512–14, no. 280) includes tagged by not only by ’s but also by prikazyvati ‘charge’, v”zveˇstiti ‘proclaim’, napisati ‘write’, and several nominal introducers, and tagged by biti cˇelom’ ‘petition’ and the phrase slyšanie moe takovo ‘my hearing [is] this.’ Similarly, some investigators have asserted that was “inchoate” in (more precisely, Old East Slavic) and that it “expanded” at the expense of tagged in the late Muscovite period; as evidence they adduce the fact that the older the text is, the fewer the tokens of are (Molotkov 1958: 46–49; Rinberg 1986: 51). It is doubtful that such claims are actually grounded on a reading of the entirety or even the majority of the oldest texts. In any case, this kind of evolutionary argument ignores the generic character of the earliest texts and the functions of the reports that occur in them, not to mention the general nature of medieval textual attestation and transmission. The oldest surviving texts form a random corpus, which is certainly not representative of all the contemporary genres; the number of manuscripts and attested text-kinds increases over the centuries, and with them the range of observable contextualizations for the different varieties of . Similar examples of false reasoning are easy to find in the literature on higher-level units in premodern languages; in some cases, they have been carried over from standard sources to historical-linguistic studies that are not based on independent work with the primary sources. This, of course, is not to deny that drift can be a factor in the synchronic distribution of functionally related units of syntax, discourse, or lexicon. However, there is considerable validity to Bakhtin’s observation (1952–53/
302
R V
1986: 65) that “there is not a single new phenomenon… that can enter the system of language without having traversed the long and complicated path of generic-stylistic testing and modification.” Before positing sweeping longterm tendencies or independent developments in the linguistic code, one should first take into account “the teleology of the authorial context” (to use Bakhtin/Vološinov’s term), as reflected in such factors as the individual characteristics of primary sources, the function and evolution of text-kinds and the speech genres that they contain, and the changing needs of the interpreters and social institutions that the text-kinds serve. This highly contextualized level of analysis has as yet been little practiced — a fact that gives particular urgency to the further development of the field of historical pragmatics and to pragmalinguistic studies of text-kinds and speech genres.
Notes
Preface 1. See, for example, Arnovick 1999; Brown and Gilman 1989; Shippey 1993, van der Walle 1993, and the articles in Jucker, Fritz, and Lebsanft (eds.) 1999 and Jucker (ed.) 1995 (in particular, the discussion in Jacobs and Jucker’s introduction). 2. See, inter alia, Baynham 1996; Fludernik 1993; Thompson 1996, and the articles in Coulmas (ed.) 1986; Dirven et al. 1982; Janssen and van der Wurff (eds.) 1996; Lucy (ed.) 1993; Verschueren (ed.) 1987. For further references, see van der Wurff 1997. 3. See D. Collins 1996; J. Collins 1987; Johnstone 1987; Mayes 1990; Philips 1985; Romaine and Lange 1991; Silverstein 1985, 1993; Tannen 1986, 1989: 98–133; Yule 1993. 4. Bakhtin acknowledged his authorship on various occasions; Vološinov’s widow supported his claim (Clark and Holquist 1984: 146–48, 166; Ivanov 1973: 44; V. L. Maxlin in Vološinov 1929/1993: 177).
Chapter 1 1. On the universality of (in particular, direct speech), see Coulmas 1986b: 2; Cram 1978: 42; Li 1986: 39; Lucy 1993: 9; Silverstein 1976: 50; Wierzbicka 1974: 271. 2. See Jakobson 1957/1984: 41; Longacre 1983: 44; Vološinov 1929/1986: 85–87. 3. On prior texts, see also Bakhtin 1975/1981: 278–80; Becker 1984, 1988: 24–26; Chafe 1994: 212; Coulmas 1986a: 1–2; Fónagy 1986: 283; Hopper 1988: 117–23; Prince 1987: 46; Tannen 1989: 99–100. 4. On metapragmatics, see also Lucy 1993: 17; Silverstein 1976: 50–52, 1977: 146–47, 1985, 1993: 36–45; Verschueren 1987: 125–26, and other works in Verschueren (ed.) 1987. 5. On “thick” descriptions, see Ryle 1966–67/1971, 1968/1971; on “emic” understanding, see Becker 1984. 6. On the automatization of intentional strategies, see Enkvist 1987: 24; Keller 1985: 213. 7. Note that intentional acts can subsume acts whose intentionality is irrelevant (Blackburn 1994: 196). Uttering a relative clause is intentional; the accompanying movements of the speech organs are usually not (Keller 1985: 213).
304
R V
8. The sequence in which I discuss these decisions is not intended as a claim about the order in which they are actually made in the reporting process. 9. Among the many alternative terms are verbs of saying/utterance/communication, linguistic action verb(ial)s (Goosens 1987; Verschueren 1987), and metapragmatic descriptors (Silverstein 1985). 10. For the citation conventions, see “Conventions for Citing Cyrillic Sources” in the front matter. 11. See, e.g., Munro 1982; Coulmas 1985a, 1985b; Roncador 1988; Frajzingier 1991; the studies in Coulmas (ed.) 1986 and Janssen and van der Wurff (eds.) 1995. For bibliography, see van der Wurff 1997. 12. On ’s with infinitives and participles, see D. Collins 1994: 202–44, and the references cited there. 13. On the nonconvertibility of and , see Banfield 1973; McHale 1978: 250–57, Partee 1973. 14. Banfield’s attempts (1973, 1978, 1982, 1993) to work English and French into a generative syntax proscribe many of the features that are typical of the mode. For discussion, see McHale 1978: 252–57. 15. Even as object clauses, reports are syntactically problematic. See Munro 1982 for discussion. 16. See Bulaxovskij 1958: 409; Ivanov 1965: 80–81; Machek 1968, s.v. “dít”; Šaxmatov 1941: 268; Sreznevskij s.vv. “deˇti, deˇju” and “deˇjati”; Vasmer 1953–58, s.v. “de.” 17. See Hickmann 1993: 66–67; Leech and Short 1981: 324; McHale 1978: 258–60; Page 1988: 35; Semino, Short, and Culpeper 1997; Tannen 1986: 323. 18. On slipping in , see Bulaxovskij 1958: 416; Lopatina 1979: 445–46; Sundberg 1982: 178. On , see Borkovskij 1981: 222–30; Koduxov 1955: 136–40; Lopatina 1979: 446–47; Peškovskij 1956: 485; Vinogradov and Istrina (eds.) 1960: 413–14. On other languages, see Fónagy 1986: 276–77; Haberland 1986: 230–31; Holt 1996: 243, Note 1; Schuelke 1958; Tannen 1986: 314, 1989: 117–18; Yule 1993. 19. See, for example, Givón 1980; Wierzbicka 1987, 1988: 23–168. 20. Cf. Keller 1985: 236; see also Lewis (1969) on the emergence of conventions in language. 21. Cf. Ehlich (1981: 160–63) on the parallel activities of “philologists” and “linguist native speakers.” 22. The possibility that writers may have had private agendas poses similar difficulties. However, unless there is internal evidence for such agendas, there is no reason to suspect them in the kinds of documents used in this study, which had disinterested writers and clear-cut institutional functions. 23. Inability to make a match of this kind does not, of course, exclude the possibility of a rational motivation that cannot be reconstructed due to inadequate information.
Chapter 2 1. For descriptions of typical lawsuits, see Dewey and Kleimola 1973: 42–45; Kleimola 1972, 1975. On general characteristics of the text kind, see also Cˇerepnin 1948–51, 2: 230, 1952: 649; Dewey and Kleimola 1973: 41–48; Kaštanov 1998: 160; Kleimola 1972, 1975; Leont’ev 1969: 45–46; Pokrovskij 1973: 20–22. 2. Not included in this count are a few documents that are cited within other transcripts but are not independently attested.
N
305
3. On the terminology, see Dewey and Kleimola 1973: 47–48; Kaštanov 1998: 160; Kleimola 1972: 356, 1975: 10; Pokrovskij 1973: 21. 4. On the referral process, see Kleimola 1972: 367, 370; 1975: 68–74; Leont’ev 1969: 40–41. 5. Cf. the reliance on communal memory, as evidenced by the testimony of “longtime residents” (starožil’ci). 6. The Muscovite Law Code (Sudebnik) of 1497 legislates different processing fees (see Grekov (ed.) 1952: 21–22, §§15, 16, 22, 24; see also the commentary, 64–65, 68). On these subtypes, see also Kaštanov 1988: 151, 1998: 160; Kleimola 1975: 6; Leont’ev 1969: 45. 7. One later summary of a trial transcript, of uncertain function, has the date at the beginning, much as in a chronicle entry (ASÈI 3: no. 319); however, this paraphrase evidently had a nonjudicial purpose. 8. See ASÈI 2: no. 229, 388a, 464 (no year), GVNP no. 340 (1463–90); AFZX no. 117, 157, 204, 248, 259, ASÈI 2: no. 416, 418, 419, 422, 3: no. 105, 276 (1491–1500); AFZX no. 306, ASÈI 1: no. 651, 658, 2: no. 309, 310, 336, 338, 428, 495, 3: no. 48, 172, 173, 221, 223, 224, 250, 251, 478, Kaštanov 1970, no. 40 (1500–1505). In tallying up the total number of transcripts for a specific period, I have assigned the documents without internal dates to the earliest possible year given by the editors.
Chapter 3 1. On perspective and the typology of , see, inter alia, D. Collins 1996; Doležel 1973: 26, 28–29; Ebert 1986: 157; Goffman 1981: 147–52; Koduxov 1955: 118–19; Li 1986: 29–34; Palmer 1986: 163–65; Romaine and Lange 1991: 228–29; Roncador 1988: 3–4. Alternative definitions of are discussed in 3.3. 2. There was a different convention for introducing the judges’ ; see Chapter 4. 3. For settlement records presented as dialogue discourse, see ASÈI 1: no. 420, 421, 422, 428, 432; 2: no. 210, 289, 290, 335, 409, 429; 3: no. 213, 214; Kaštanov 1970: no. 70. For a default-judgment charter, see ASÈI 2: no. 167. 4. The following forms occur: masculine rek(”) or rekl(”), rekli or r(’)kli. The letters transcribed ” and ’ were essentially diacritic and often omitted in writings of the Muscovite period. 5. Such buck-passing may be seen in other transcripts, e.g., ASÈI 1: 324, no. 336; AFZX 97, no. 103. 6. See AFZX: 116, no. 125; ASÈI 2: 251, no. 296; Kaštanov 1970: 357, no. 9. 7. This is attested in two related transcripts — ASÈI 2: 415–16, no. 404 (5×); 417, no. 405. 8. See ASÈI 2: 435 (2×), 436–37 (5×), 438, 439 (3×), 440, no. 411. 9. I am not claiming that the decision-making process followed the order presented here. 10. Given that preposed tags are less salient than (“eclipsed by”) the following direct reports, it seems somewhat problematic to characterize this relation as “separate: equal” (Thompson 1996: 520). 11. For a similar example, see ASÈI. 1: 315, no. 427. 12. On punctuation in , particularly with reference to , see also Cˇerepnin 1956: 159, 374–76; Osipov 1992; Preobraženskaja 1983: 102–3; Rinberg 1985: 49; Starovojtova 1988: 70. The same
306
R V
basic observation could be made for premodern writings in other European languages, e.g., Hungarian (Fónagy 1986: 257) and English (Lennard 1995: 67–68). On the diverse uses of quotation marks in premodern English texts, see also Page 1988: 30–31). 13. On the usual editorial practices, see ASÈI 1: 8; Cˇerepnin 1956: 564–69; Kotkov and Popova 1986: 3. ˇ erepnin 1956: 246; Zaliznjak 14. On this word, customarily abbreviated as (>n or (c>n, see C 1990: 9–24. 15. Cf. Brown and Levinson 1987; Dewey and Kleimola 1973: 41–42; Kleimola 1972: 364. 16. See Volkov 1974: 115–16; Zaliznjak 1986b: 272, 289, 1990: 12–14. 17. On typicality markers in , see also Bauman 1986: 67; Chafe 1994: 217; Dubois 1989. 18. On vocatives as a marker, see Roncador 1988: 4; Wierzbicka 1970: 644. 19. In heterosubjective conjuncts, many of the cases without g(o)s(podi)ne show topic continuity involving elements other than the subject. 20. For representative expressions of this view, see Otin 1969: 54–55; Wierzbicka 1974: 267, 272; ˇ umakov 1975: 16; Longacre 1983: 132; Li 1986: 39–40. Xaburgaev 1974: 425–26; C 21. On the use of in this text kind, see also Larin 1961: 30. For examples, see AI 1: 340–41, no. 179.2; ibid. 2: 62–63, no. 53; ibid.: 123, no. 92; ibid.: 249–50, no. 212; ibid.: 314–15, no. 262; ibid.: 341–42, no. 282; ibid.: 351–52, no. 290; ibid.: 354, no. 295; ibid.: 357–58, no. 301; ibid.: 358–59, no. 303; ibid.: 364–65, no. 307; AJuB 3: col. 271–76, no. 330; Kotkov, ed. 1984: 161, no. 126; Kotkov, ed. 1990: 81–82, no. 67; ibid.: 108, no. 89; Mordovina and Stanislavskij 1979: 101; Stanislavskij 1981: 288, no. 2; ibid.: 293–94 no. 8; ibid.: 298–99, no. 18; ibid.: 300, no. 21; ibid.: 303–4, no. 28; ibid.: 305, no. 32; and Kotkov et al., eds. 1993: 119–47. 22. The reproductionist model is common in Russistics — e.g., Cˇumakov 1975: 16–17; Gvozdev 1958: 274; Koduxov 1955: 114–19; Peškovskij 1956: 484; Rozental’ 1988: 247; Vinogradov and Istrina (eds.) 1960: 402. For reproductionist treatments of in , see Molotkov 1958: 28; Sprincˇak 1960–64, 2: 119; Otin 1969; Lopatina 1979: 442; Rinberg 1985: 53; SchmückerBreloer 1987: 248–49. For other languages, see, e.g., Xaburgaev 1974: 425–26; Li 1986: 40; Massamba 1986: 99–100. 23. has also been treated as an “opaque” context (cf. Quine 1960: 151) in which referential terms are necessarily de dicto rather than de re, i.e., expressions of the “original speaker” (see Coulmas 1986b: 3–4; Kiparsky and Kiparsky 1970: 157–58, Note 7; Li 1986: 29–30; Partee 1972; Zwicky 1974: 198–99). There are several reasons to doubt this approach, apart from its reproductionist character. First, it only pertains to reports of actual utterances. Second, it is logical-philosophical contraband to associate verbatimness in natural language with truth value. The question “Is this a valid argument?” is not equivalent to “Is this an exact rendition of what was said?” Third, actual observation proves that reporters allow substitutions in unless they have some special reason to avoid them. (For further criticism, see Cram 1978: 42–44.) 24. On and memory constraints, see Chafe 1994: 215–17; Clark and Gerrig 1990: 796–97; Coulmas 1986b: 25, Note 8; Mayes 1990: 332; Tannen 1986: 313–14. 25. See also Comrie 1986: 266; Coulmas 1986b: 2, 6, 11–13, 25; Leech 1974: 353; Li 1986: 40; Romaine and Lange 1991: 231–32, 243, 263; Short 1988: 66–67, 70. 26. On non-verbatim uses of , see also Baynham 1996: 66–68; Chafe 1994: 216; Clark and Gerrig 1990; Dubois 1989; Fludernik 1993: 408–14; Fónagy 1986: 255, 278–82; Haberland 1986: 224–25; Hanks 1990: 206; Mayes 1990; Page 1988: 6–7, 25–26, 29–30; Romaine and Lange 1991: 229–31; Sternberg 1982a, 1982b; Tannen 1986: 313–14, 1989: 98–119.
N
307
27. On co-construction of (“chiming in”, Couper-Kuhlen 1999: 18, 21), see also Mathis and Yule 1994; Yule 1995: 189. 28. On departicularizers in , see also Chafe 1994: 216; Haberland 1986: 223–24; Fludernik 1993: 410; Mayes 1990: 334; Sternberg 1982a: 96, 99–104. 29. On strengthening hedges, see Brown and Levinson 1987: 145. 30. Cf. Romaine and Lange’s observation (1991: 231–32, 243, 263) that different tags are used with to indicate differing degrees of commitment to fidelity. 31. See Coulmas 1986: 10–13; Goody 1977: 118; Lyons 1977: 17; Ong 1981: 21–22, 1982: 57–68. 32. For example, ASÈI 3: 270, no. 251; Anpilogov 1977: 79, 208, 212. 33. This word was usually written without the o, with a superscript k over the a. Thus there is usually no way to distinguish between tako, which in the later medieval period occurred mostly in ecclesiastical texts, and the innovative form tak, first attested in a text of 1356 (Vasmer 1953–58, s.v.). 34. On cataphora in discourse, see also Halliday and Hasan 1976: 17–19, 69, 75. 35. Cf. Romaine and Lange (1991: 244–51, 262) on the strategy be like in recent American English. 36. See AFZX 97, no. 103; ASÈI 1: 335, no. 447; 352–54, no. 467 (3×); 2: 119, no. 188; 150, no. 229 (2×); 389, no. 387 (2×); 416, no. 404; 418, no. 405; 3: 55, no. 32. 37. The true impersonals pisano ‘write-.’ and napisano ‘write-.’ are also found in seven examples. See AFZX 97, no. 103; ASÈI 1: 491, no. 595; 2: 335, no. 338; 451, no. 418; 544, no. 496; 3: 55, no. 32; Kaštanov 1970: 371, no. 16. 38. Two cases directly follow perception reports similar to (9b) (ASÈI 2: 320–21, no. 334; 378, no. 381). Others occur when the citation is the second in a series (ibid.: 265, no. 307; 3: 182–83, no. 172 (2×)). 39. See ASÈI 2: 422, no. 406; 3: 463–64, no. 478. 40. For possible examples, see ASÈI: 201, no. 288; 437, no. 411; 523, no. 483. 41. Cf. AFZX 227–29, no. 259; 258–59, no. 308; ASÈI 1: 472–73, no. 587; 2: 324–27, no. 336; 332–33, no. 338; 418–19, no. 405; 459–61, no. 422. For exceptions, see ibid.: 333–34, no. 338; 422, no. 406; 436–37, no. 411 (said by the editors to be an abridgment). 42. See Bakhtin 1952–53/1986: 76–81, 86; Fludernik 1993: 20, 22; Grice 1975/1989; Lakoff 1995: 196; Slembrouck 1992: 109–10; Sperber and Wilson 1986; Walker 1986: 214, 220. 43. In a sense this is true even of mouthpieces or, at least, of the ambassadors mentioned above, since the rulers they addressed would require as unmediated a representation of the grand prince’s position as possible. 44. On assessments of audience, see also Bakhtin 1952–53/1986: 97–99, 1975/1981: 279–81; Lakoff 1984: 481; Sinclair 1988: 15; Walker 1986: 216–17, 220. 45. Clark and Gerrig (1990: 787) treat both and as demonstrations that differ in perspective: “Free indirect quotations, like direct quotations, are demonstrations that are components of language use. It is just that free indirect quotations take the vantage point of the current instead of the source speaker.” 46. See also Fónagy 1986: 255; Goffman 1981: 149–52; Mayes 1990: 346; Tannen 1983: 365, 1989: 125.
308
R V
47. Zipf (1949: 20–21), approaching speech behavior from a different standpoint, makes a similar distinction between speaker’s economy, where the hearer must exert more effort, and auditor’s economy. 48. It would be more precise to say addressee-based, since such strategies are oriented to the knowledge and sense-making ability of ratified addressees rather than to hearers of other kinds (cf. Clark 1992: 205–6, 248–74; Goffman 1981: 260). Moreover, they are not restricted to spoken language but also appear in writing. 49. See, e.g., Baynham 1996: 78; Chafe 1982: 48, 1994: 217, 223; Cˇumakov 1975: 19; Jespersen 1924/1992: 290; Leech and Short 1981: 319–20; Li 1986: 41; Mayes 1990: 326; Page 1988: 32–33; Schiffrin 1981: 58–60; Schuelke 1958: 97; Tannen 1983: 365, 1986: 311, 1989: 25–26; Thompson 1996: 512. 50. See, for example, Fludernik 1993: 435; Glock 1986: 46, 48; Larson 1978: 60–68, 183; Li 1986: 40; Philips 1985: 154–55, 169; Tannen 1983: 364. 51. This is not to be confused with the moral or ethical question of the extent to which one bears responsibility for the use or misuse of one’s statements by others — an issue that need not be discussed here. 52. Cf. Huizinga (1944/1955: 14–15) on “the mystic repetition or re-presentation” involved in sacred rites.
Chapter 4 1. See Barnet 1965: 62–65; Borkovskij and Kuznecov 1965: 290, 318; Kuz’mina and Nemcˇenko 1980: 152–54; Stola 1958. 2. One complication is the aspect of the gerund. Since the verb recˇi was perfective by the fifteenth century, r’kucˇi might be expected to be perfective as well and hence to convey sequentiality; however, its use in discourse suggests that it was imperfective (cf. Barnet 1965: 104–5; Larin 1975: 217). 3. See Chvany 1985b: 15; 1990/1996: 292–93; Fox 1983; Givón 1984–90, 2: 839–40; Hopper and Thompson 1984: 740–41. 4. For examples, see ASÈI 1: 556, 557, no. 642 (2×); ibid. 2: 394, 395, no. 388a (3×); 436, no. 411 (1×); AFZX 231, no. 259 (2×); 259–60, no. 308 (2×). 5. For the other cases, see ASÈI 2: 431–32, no. 410 (3×); 536, no. 493. It was not, of course, obligatory to use a gerund tag in such contexts; the standard formula could be used, as in ASÈI 2: 539, no. 493. 6. See also ASÈI 2: 119, no. 188; 2: 377, no. 381; AFZX 181, no. 204; 213, no. 249; and in transcripts from Pskov and Dvina (GVNP 148, no. 92; 188, no. 132; 326, no. 340). The same incipit is found in a 1391 monastic statutory charter (ASÈI 3: 16, no. 5) and in default judgment charters (ASÈI 1: 238–39, no. 329; 362, no. 479; 2: 102, no. 167; AFZX 221, no. 257; 229, no. 259). 7. See N. P. Lixacˇev (ed.) 1895, pt. 2: 198, no. 10 (1543); AJuB 1: no. 52.5 (1547). 8. See, for example, AJu no. 24 (1584), 52.7 (1561), 52.8 (1567); DAI no. 157 (1609); Ivina (ed.) 1983: no. 154 (1566); Kotkov and Filippova (eds.) 1978: no. 26 (1555), 83 (1576–77); Liberzon
N
309
(ed.) 1988–90, pt. 2: 656 (1578); N. P. Lixacˇev (ed.) 1894: no. 2 (1583), 3 (1589); N. P. Lixacˇev (ed.) 1895, 2: no. 11 (1551), no. 15.4 (1584). For conservative texts that largely maintain tak(o) rek(li), see AJu no. 22 (1547), 23 (1571); AJuB 1: no. 52.6 (1555); N. P. Lixacˇev (ed.) 1895, pt. 2: no. 12 (1552). 9. Brat’e is the vocative of the collective (functionally plural) noun brat’ja. Typically the two vocative forms were spelled identically in scribal usage — as brate, with a superscript t. 10. E.g. Antonov and Baranov 1997: 117, no. 146. For the other cases, see D. Collins 1994: 138, Note 192. 11. For the remaining cases, see ASÈI 1: 477, no. 588; 2: 452, no. 418; 455, no. 419. Another instance of this appears in nonmultiplex testimony (ibid. 1: 547, no. 635). 12. On the historical present in , see also Nikiforov 1952: 174. For other languages, see Chafe 1994: 207–11; Johnstone 1987; Page 1988: 48; Schiffrin 1981; Schuelke 1958: 97; Tannen 1983: 365, 1989: 214, n. 13; Uspenskij 1973: 71; Vinogradov and Istrina (eds.) 1960: 482. 13. Cf. the strong association of the historical present and emotive ’s such as go, be like, be all in American English (Schiffrin 1981: 58; Romaine and Lange 1991: 243). 14. For the other cases, see ASÈI 3: 221, no. 209; AFZX 228, no. 259. 15. The distinct status of this speech act is borne out in seventeenth-century trial dossiers, where it is typically performed by means of a special text-kind, ssylocˇnye pamjati (“referral memoranda”; see 4.4). 16. For other examples in boundary-setting descriptions, see AFZX 109–10, no. 117 (3×); 260, no. 308; AJu 16, no. 8; ASÈI 1: 321, no. 431 (integrated ); 492, no. 595; 2: 197, no. 286; 202, no. 288; 435, no. 411 (2×); 438; 444, no. 414; 452, no. 418; 455, no. 419 ( and integrated ); 540, no. 493; 3: 76, no. 50 (2×); 220, no. 209. The same usage is found after citations of documents (AFZX 97, no. 103; ASÈI 1: 235, no. 326). 17. For an additional example, see ASÈI 1: 538–39, no. 628. 18. This pattern is violated in a few cases where some of the parallel tags are conjoined with i (AFZX 212, no. 248; ASÈI 1: 460, no. 581; 2: 415, no. 404). Other factors may be at work in these examples. 19. For similar examples, see AFZX 106, no. 114; 116–17, no. 125 (2×); 214, no. 249; ASÈI 1: 247–248, no. 340 (with a choral tag); 288, no. 397; 400, no. 523; 464, no. 583 (2×); 466–67, no. 584; 469, no. 585; 470–71, no. 586 (2×); 474, no. 587 (2×); 476–77, no. 588 (2×); 478–79, no. 589 (2×); 481, no. 590 (3×); 483, no. 591 (2×); 485, no. 592 (2×); 487, no. 593 (2×); 488–89, no. 594 (2×); 573–74, no. 651 (3×); 2: 151, no. 229; 321, no. 334; 411, no. 402; 439, no. 411; 519–20, no. 481; 3: 188, no. 173; 218–19, no. 208; 271, no. 276 (2×); Kaštanov 1970: 411–13 (2×). For cases of multiplex testimony by co-respondents, see ASÈI 1: 398, no. 522; 501, no. 604; 556–57, no. 642. 20. For an exception in which conflicting responses are presented as part of the same speech event, see ASÈI 2: 407, no. 400. Clearly this is a case of nonevaluation rather than a true counterexample. 21. For similar cases, see AFZX 110, no. 117; AJu 15, no. 8; ASÈI 1: 248, no. 340; 450, no. 571; 506–7, no. 607; 512, no. 607a; 552, no. 639; 2: 407, no. 400; 452, no. 418; 455, no. 419; 3: 182, no. 172; Gorcˇakov 1871, appendix: 63–64, no. 4.5; Koreckij 1969: 289, no. 2. For amid tagged , see ASÈI 1: 248, no. 340; 3: 69, no. 48.
310
R V
22. The standard tag is used in all but two cases, which feature the s(”)kazati (ASÈI 2: 452, no. 418; 455, no. 419; see 4.2). 23. For similar cases, see AFZX 223, no. 258; AJu 16, no. 8; Antonov and Baranov 1997: 118, no. 146; ASÈI 1: 540, no. 628; 2: 263–64, no. 307; 434–35, no. 411; 3: 69, no. 48; 219, no. 208; 235, no. 218; Gorcˇakov 1871, appendix: 63–64, no. 4.5. 24. See AFZX 109, no. 117; 127, no. 140; 230, no. 259; 259–60, no. 308; ASÈI 1: 528, no. 615; 557, no. 642; 540, no. 628; 2: 313–14, no. 332; 423, no. 406; 469, no. 428; 540, no. 493; 3: 143, no. 105; 221–22, no. 209; 321, no. 334. 25. See Kleimola 1975: 9–10, especially note 46. For examples of trial dossiers, see AJu no. 25, 26, 29; AJuB 1: no. 52.9, 104; D’jakonov (ed.) 1897: no. 57; Kotkov (ed.) 1990: no. 98; N. P. Lixacˇev (ed.) 1895: no. 5; Zabelin 1848; and Jakovlev 1943/1970, appendix. 26. For other instances of this procedure, see AFZX 234, no. 261; ASÈI 2: 496–97, no. 458. 27. For the other example, see ASÈI 1: 321, no. 431. Excerpt (20) is is unusual in that the respondent conducts the boundary-setting rather than his witnesses. 28. See Borkovskij and Kuznecov 1965: 307–8; Nikiforov 1952: 115, 157, 165; Uspenskij 1987: 152–53. 29. For treatments of this kind of formula, see earlier in the same transcript or ASÈI 1: 248, no. 340. 30. Note that the constructions can also occur in non-narrative contexts, where there is no plot. 31. On the definition of speech-act verbs, see Verschueren 1980: 4–5, 1985: 5; Wierzbicka 1987. 32. The heresy trials known from the first half of the sixteenth century have a different character (Kazakova 1960: 285–318; Pokrovskij (ed.) 1971). 33. Cf. the use of dovoda… ne dovel ‘did not produce an argument’ as a within the judges’ in Antonov and Baranov 1997: 287, no. 296. 34. See GVNP 326, no. 340 (2×); ASÈI 2: 434, no. 411 (in a relative clause). 35. For the other examples, see ASÈI 3: 234, no. 218; GVNP 326, no. 340. 36. For a survey of the vast bibliography on , see Fludernik 1993. 37. Cf. Vološinov 1929/1986: 134; Leech and Short 1981: 325–32; Clark and Gerrig 1990: 788; Thompson 1996: 514–15. 38. See also Slembrouck 1986: 47–48, 59–60, Page 1988: 33, 35, 37. 39. See Vinogradov and Istrina (eds.) 1960: 428 on ; Doležel 1973: 27 on Czech; Fónagy 1986: 284 on Hungarian; and Bamgbos® e 1986: 78–80, 95 on Yoruba. On in , see also Vološinov 1929/1986: 138–40, 155–59; Koduxov 1955: 136–38, 140–44; Uspenskij ˇ umakov 1975: 38–48. 1973: 33–36; C 40. See Banfield 1973: 21–22; Doležel 1973: 32–39; McHale 1978: 265; Spitzer 1928/1988. 41. This is not the place to discuss the arguments for a “narratorless” , advocated by Banfield (1973, 1978, 1982, 1993). See the counterarguments in Fludernik 1993; McHale 1978: 253–57, 1983; McKay 1978: 14–20; Paducˇeva 1996: 348–49; Toolan 1992: 129–37; Wierzbicka 1974: 294–97. 42. Among the scholars claiming that is exclusively literary are C. Bally and E. Lorck (cited by Pascal 1977: 13, 18–19), Banfield (1973, 1978, 1982, 1993), and Paducˇeva (1996: 337). Vološinov (1929/1986: 156) sees it as primarily a written phenomenon dependent on the “’silencing’ of prose.” Other investigators have pointed to examples in spoken narratives (Pascal
N
311
1977: 18–19; Ullmann 1957: 98; Haberland 1986: 232–33; Clark and Gerrig 1990: 787; Hickmann 1993: 67) and in non-literary text kinds such as parliamentary reporting and journalism (McHale 1978: 282–83; Slembrouck 1986; Short 1988: 72). 43. See, e.g., Kozlovskij 1890: 4 (the first Russian study of ); Vološinov 1929/1986: 139; Švedova 1952: 113; Bulaxovskij 1954: 442–44; Koduxov 1955: 138, 145, 168; Molotkov 1958: 27. 44. Some of the reports examined here for convenience’ sake are technically not but free nondirect speech, since they do not contain any forms with person marking. Given that reporting strategies form a continuum, free nondirect speech does not represent a distinct category and has the same non-actualizing discourse effects as central . 45. See also Bakhtin 1975/1981: 305–6, 318; Doležel 1973: 35–39, 50; Thompson 1996: 513. 46. For other cases in boundary-setting procedures, see ASÈI 2: 534–35, no. 492; 540, no. 493. 47. For additional counterarguments to the literariness claim, see Fludernik 1993: 4, Polanyi 1982.
Chapter 5 1. On choral speech, see also Section 4.3. 2. For texts with a similar use of capitals, see ASÈI 1: no. 587 (a contemporary copy, including an earlier transcript quoted in its entirety), 607a, 651, and 658; 2: no. 286, 307, 309, 414, 428, and 495; and 3: no. 48 and, with two exceptions, in 3: no. 223. The same system is used, as far as can be determined, in an autograph that survives in three fragments (ASÈI 3: 224). 3. See Antonov and Baranov 1997: 287–88, no. 296. On byplay, see Goffman 1981: 134. 4. The reverse situation, with the question in and the answer in , is not attested in the corpus. The one apparent instance involves a question that the judge asks through an intermediary (see 4.2). Given that the ratified addressee was absent, the original question as spoken during the trial was probably also in . 5. The genitive eventually gave way to the accusative after verbs of asking, as part of a general tendency to reanalyze genitive objects (Sprincˇak 1960, 1: 132; Timberlake 1975: 135; Klenin 1983: 103–4; 1987: 413–18). In the present corpus, the object is clearly genitive with a-stem nouns, where there is no genitive-accusative syncretism (e.g., ASÈI 2: 392, no. 388; 420, no. 406; 448, no. 416). An accusative, Il’ju ‘Il’ja’ is found in a transcript from the northwestern city of Pskov (GVNP 327, no. 340). 6. The verb form is in the aorist (v”sprosiša) in the sole extant transcript from Novgorodian Dvina (GVNP 149, no. 92; cf. s”prosiša in a similar document, 188 no. 132). 7. See AJuB 1: no. 103.1; ASÈI 1: no. 447, 522, 524, 538, 571, 581, 587, 592, 595, 607, 607a, 615, 628; 2: no. 229, 287, 288, 296, 309 (equal numbers), 334, 375, 405, 410, 416, 418, 422, 483, 492; 3: no. 208, 288; Kaštanov 1970: no. 9, 27. 8. It occurs to the exclusion of other verbs in AFZX no. 125, 140, 157, 258, 259 (including a quoted transcript), 261; AJu no. 8, 12; AJuB 1: no. 103.3; ASÈI 1: no. 397, 430, 431, 467, 485, 523, 525, 537, 538, 540, 583, 586, 590, 593, 635, 639, 640, 642, 658; 2: no. 90, 315, 368, 387, 388, 388a, 400, 411, 428, 463, 464, 481, 495, 496; 3: no. 31, 32, 35, 48, 55, 172, 173, 218, 250, 357, 477; Gorcˇakov 1871: no. 4.1, 4.3, 4.5; Morozov 1988: no. 2; and two northwestern
312
R V transcripts (GVNP no. 92, 340). It is the predominant verb in AJuB 1: no. 103.1; ASÈI 1: no. 326, 587, 589, 592, 595, 607, 607a, 615, 628; 2: no. 229, 336, 338, 416, 418, 492; 3: no. 288, 364; Kaštanov 1970: no. 9, 27.
9. It occurs to the exclusion of other verbs in AJuB 1: no. 103.2; ASÈI 1: no. 340, 521, 539, 557, 582, 584, 585, 588, 591, 594, 604; 2: no. 285, 286, 306, 307, 332, 333, 337, 374, 383, 401, 402, 404, 406, 407, 414, 458, 465, 493; 3: no. 50, 105, 209, 221, 223, 224, 251, 276, 478; Kaštanov 1970: no. 6, 16, 40; Koreckij 1969: no. 2; Krotov and Smetanina 1987: no. 5; GVNP no. 132. It is the predominant verb in ASÈI 1: no. 447, 522, 524, 538, 571, 581; 651; 2: no. 287, 288, 296, 310, 334, 375, 405, 410, 419, 421, 422, 483; 3: no. 208. 10. See AJu no. 12; ASÈI 1: no. 538, 587, 607, 607a; 2: no. 287, 288, 336, 405, 422, 483, 492; 3: 172; Kaštanov 1970: no. 9. 11. Eleven are originals or contemporary copies (ASÈI 1: no. 524, 538, 587, 592, 607, 607a; 2: no. 287, 288, 334, 422; 3: no. 208), and six are later copies (AJuB 1: no. 103.1; ASÈI. 1: no. 595; 2: no. 229, 375, 418, 483). 12. ASÈI 1: no. 447, 571, 581, 615, 628; 2: no. 405, 410, 492; Kaštanov 1970: no. 9, 27. 13. ASÈI 1: no. 522, 2: no. 416, 3: no. 288. 14. ASÈI 1: 235, no. 326; 478, no. 589; 541, no. 628; 3: 386, no. 364. 15. ASÈI 2: 252, no. 296; 455, no. 419. V”prositi is the sole verb of inquiring in the remaining example (Gorcˇakov 1871: 51, no. 4.2). 16. For examples, see AFZX 228, no. 259; ASÈI 1: 459, no. 581; 509, no. 607; 514, no. 607a (a variant of the preceding document); 540, 541, no. 628; 2: 196, no. 286; 202, no. 288; 313, no. 332; 369, no. 374; 497, no. 458; 523, no. 483; 539, no. 493; 3: 291, no. 276; Gorcˇakov 1871, appendix (separate pagination): 50, no. 4.1; Kaštanov 1970: 359, no. 9. 17. See AFZX 139, no. 157; 229, no. 259; ASÈI 1: 235, no. 326; 321, no. 431; 450–51, no. 571 (2×); 557, 558, 559–60 (2×), no. 642; 2: 351, no. 358; 2: 364, no. 370 (2×). 18. See ASÈI 2: 431–32, no. 410 (5×); 352 no. 467 (2×); 536, 539–40 no. 493 (3×); 3: 84–85, no. 55; 302, no. 288. 19. See AFZX no. 129 (2×); 229, no. 259; 257, no. 308; Antonov and Baranov 1997: 117, no. 146; ASÈI 1: 539, no. 628; 2: 420–21, no. 406 (3×); 459, no. 422; 537–38, no. 493 (2×); 3: 141, no. 105. 20. See AFZX 106, no. 114; 109, no. 117; 228, no. 259; ASÈI 1: 481, no. 590; 502, no. 604; 2: 252, no. 296; 372, no. 375; 411–12, no. 402 (2×); 425, no. 407; 434, no. 411; 448, no. 416; 533–34, no. 492 (2×); 3: 240, no. 221; 290, no. 276; and, in a northwestern transcript, GVNP 327, no. 340. 21. For the other cases, see AFZX 231–32, no. 259 (2×); 260, no. 308; ASÈI 1: 434, no. 557. 22. The occurrence of the nonreflexive rek in one case of this formula (ASÈI 1: 474, no. 587) is probably a mistake for rek”sja, and it is so treated by the editors (ibid.: 475, Note 5). For further examples of recˇisja, see AFZX 117, no. 125; 228, no. 259; ASÈI 1: 248, no. 340; 290, no. 397; 1: 403–4, no. 525; 414, no. 537; 416, no. 538; 417, no. 539; 465, no. 583; 467, no. 584; 471, no. 586; 477, no. 588; 478, no. 589; 481, no. 590; 483, no. 591; 485, no. 592; 487, no. 593; 489, no. 594; 493, no. 595; 509, no. 607; 514, no. 607a; 548, no. 635; 2: 197, no. 286; 202, no. 288; 254, no. 296; 314, no. 332; 317, no. 333; 322, no. 334; 392, no. 388 (with jatisja in its variant, 395, no. 388a); 416, no. 404; 419, no. 405; 455, no. 419 (with jatisja in its variant, 453, no. 418); 461, no. 422; 501, no. 463; 520, no. 481; 523, no. 483; 535, no. 492;
N
313
541, no. 493; 3: 88, no. 56; 219, no. 208; 222, no. 209; 292, no. 276; 462, no. 477. For further examples of jatisja, see ibid. 1: 236, no. 326; 400, no. 523; 402, no. 524; 452, no. 571; 469, no. 585 (with recˇisja in a variant copy); 502, no. 605; 2: 365, no. 370; 370, no. 374; 373, no. 375. 23. For similar examples, see ASÈI 2: 535, no. 492; 3: 88, no. 56. 24. For further examples with veleˇti, see AFZX 110, no. 117; 228–29, no. 259 (cited within another transcript); 232, no. 259; 261, no. 308; ASÈI 2: 194, no. 285. For further examples with ucˇiniti or dati s“rok”, see ASÈI 1: 560, no. 642; 2: 263, no. 307.
Chapter 6 1. In a few documents, there is no record of the hearing; the verdict of the higher court comes immediately after the trial copy (AFZX: 98, no. 103; ASÈI 1: 501, no. 463; 2: 392 no. 388; 395, no. 388a. In others, the hearing is recounted without (ibid. 1: 322, no. 431; 2: 370, no. 374; 373, no. 375; 390, no. 450). 2. For the other cases, see AFZX: 232, no. 259; 213, no. 248; ASÈI 1: 401, no. 523; 420, no. 540; 465, no. 583; 467, no. 584; 469, no. 585; 471, no. 586; 475, no. 587; 477, no. 588; 479, no. 589; 481–82, no. 590; 483, no. 591; 486, no. 592; ibid: 489, no. 594; 493, no. 595 (2×); 503, no. 604; 509, no. 607 (2×); 513–14, no. 607a (2×); 560, no. 642; 575, no. 651; 2: 194, no. 285; 198, no. 286; 202–3, no. 288; 330, no. 337; 314, no. 332; 317, no. 333; 322, no. 334; 440, no. 411; 462, no. 422; 470, no. 428; 520, no. 481; 535, no. 492; 541, no. 492; 3: 70, no. 48; 186, no. 172 (2×); 189, no. 173; 219, no. 208; 235, no. 218; 269, no. 250; 272, no. 251 (2×); 292, no. 276; Gorcˇakov 1871, appendix (separate pagination): 56–57, no. 4.3. 3. For the other document summaries, see ASÈI 1: 465, no. 583; 467, no. 584; 471, no. 586; 477, no. 588; 479, no. 589; 486, no. 592; 2: 317–18, no. 333; 3: 70–71, no. 48 (tagged with skazal). For the references to cadastre books, see ibid.: 469, no. 585; 475, no. 587; 481, no. 590; and 483, no. 591. The transcripts from ASÈI 1 are all records of lawsuits of 1495–99 between Kostroma peasants and the Trinity-Sergius Monastery’s rent collector Ofonasij; they have the same trial judge and higher judge. Thus they may reflect the practice of a single scribe or (more probably, given the orthographic differences) a group of scribes working under the same supervision. 4. See ASÈI 2: 317–18, no. 333 (tagged with tak(o) rek(li)); 3: 70–71, no. 48 (tagged with skazal). 5. See ASÈI 1: 182, no. 253; 287, no. 394; 338, no. 450; 343–44, no. 456; 345, no. 457; 378, no. 499; 442, no. 562; 524, no. 612; 3: 102, no. 67a; 3: 103, no. 68; 138, no. 100. 6. See ASÈI 1: 297, no. 406; 334, no. 446; 2: 404–5, no. 399; 524, no. 483a; 3: 36, no. 19. In two other referred purchase deeds (Kotkov and Filippova (eds.) 1978: 9, no. 1; GVNP: 243, no. 220), reports in the higher-court record are presented as tagged by the preterit of v”sprositi and recˇi. However, the appendices contain no mention of the deeds to which they are attached, so that they are not really verification records. 7. For the other examples, see ASÈI 1: 402, no. 524; 404, no. 525; 435, no. 557; 3: 85, no. 55; 3: 88, no. 56; 269, no. 250; Kaštanov 1970: 358, no. 9. The first two cases may belong to the pen of a single scribe; both are copies of originals from 1485–90, with the same respondent, trial judge, and higher judge.
314
R V
8. See, for example, AJu 28, no. 14; 45, no. 20; 52, no. 22; Ivina (ed.) 1983: 11, no. 3; 51, no. 46; 74, no. 63; N. P. Lixacˇev (ed.) 1895: 170, no. 7; 231, no. 12; Novosel’skij et al. (eds.) 1975: 45, no. 40; 48, no. 41; 83, no. 77; 197, no. 194; 234, no. 230; and 261, no. 255. 9. For the falsifications, see ASÈI 1: 400, no. 523; 493–94, no. 595; 509, no. 607 and the variant from the same trial, 514, no. 607a; 3: 186, no. 172; 269, no. 250; 271, no. 251; Kaštanov 1970: 359, no. 9. 10. See ASÈI 1: 509, no. 607; 514, no. 607a; Kaštanov 1970: 359, no. 9. 11. See also ASÈI 1: 493–94, no. 595; 3: 186, no. 172. In a further instance, the echoes a performative statement given in (Kaštanov 1970: 359, no. 9). 12. This passage is misinterpreted by the editors of ASÈI. See D. Collins 1994: 360–61 for discussion. 13. For similar examples, see AFZX 117, no. 125; ASÈI 1: 434–35, no. 557. 14. See AFZX 214, no. 249; ASÈI 1: 320, no. 430; 541, no. 628; 2: 152, no. 229; 318, no. 333; 365, no. 370; 384, no. 383; 408, no. 400; 3: 77, no. 50. 15. Cf. a fully presupposed report of a litigant’s nonappearance made by a surety, which is presented as complementized (ASÈI 1: 236, no. 326).
Chapter 7 1. For the other cases, see ASÈI 1: 507, no. 607, and the variant transcript from the same trial, 513, no. 607a. 2. A further case of after m”lviti (AFZX 230–31, no. 259) is discussed in 7.3. 3. For s”kazyvati, see AFZX 214, no. 249; ASÈI 2: 316, no. 333; 523, no. 483; 534, no. 492; 537, no. 493. For the other verbs, see AFZX 222, no. 258 (kazati); ASÈI 1: 397, no. 521 (govoriti); AFZX 234, no. 261; ASÈI 2: 150, no. 229 (z”vati); ASÈI 1: 459, no. 581; 2: 468–69, no. 428 (nazyvati). 4. The same transcript includes another unique case in which govoriti appears as a phatic device in the judge’s speech (ASÈI 1: 396, no. 521) — a context in which s”kazati is abundantly attested. 5. For the other examples, see ASÈI 1: 514, no. 607a and the variant transcript from the same trial, 509, no. 607; 2: 269–70, no. 310 (2×). 6. For subversive requotations treated in a similar way, see AFZX 231, no. 259 (2×); ASÈI 1: 416, no. 538; 2: 268–69, no. 310; 316, no. 333; 452, no. 418 and in a variant transcript from the same trial, 455, no. 419; 503, no. 464; 3: 269, no. 250; Gorcˇakov 1871, appendix (separate pagination): 51, no. 4.2. One fragmentary example may involve a statement made prior to the trial (ASÈI 3: 245, no. 224). 7. The prepositional phrase peredo mnoju co-occurs in a tag clause with the perfective preterit in another example (ASÈI 1: 557, no. 642), where its precise motivation is unclear. 8. Imperfective preterit tags can be found in narrative contexts when the given report serves primarily as a frame or background, e.g., when it initates dyads or explains subsequent nonevents (ASÈI 1: 538, no. 628; 2: 411, no. 402; 3: 173, no. 187; 3: 222, no. 209; Kaštanov 1970: 355, no. 9).
N
315
9. Cf. the layered report (in complementized after s”kazati) that is set in a mini-narrative in AFZX 116, no. 125. Here the reporter’s purpose is evidently to give background on the witnesses whom he is about to introduce; that is, the emphasis on the source of the information rather than the content. 10. See the other example in ASÈI 2: 193, no. 285; ibid. 1: 460, no. 581; Kaštanov 1970: 411–13, no. 40 (10×). There are also two or possibly three cases of the negated perfective preterit of the zaperetisja ‘deny’, functioning as collateral information rather than plotline (Krotov and Smetanina 1987: 68, no. 5). 11. The gerund of the verb rosprositi ‘find out by means of interrogation’ also introduces a complementized report in a narrative context (ASÈI 2: 369, no. 374). The backgrounding in this example is to be expected, since it conveys the reason for a subsequent nonevent (a nonjudgment). 12. Cf. the use of the preterit of the imperfective javljati ‘declare’ to tag hearsay of a different kind (ASÈI 1: 480, no. 590). 13. For the other cases after s”kazyvati, see AFZX 106–7, no. 114; Antonov and Baranov 1997: 117, no. 146; ASÈI 1: 485, no. 592; 2: 197, no. 286 (2×); 199, no. 287; 414–15, no. 404; 3: 141, no. 105 (2×); and, with a left-shifted topic clause intervening between tag and complementized report, 1: 528, no. 615. 14. For other cases of s”latisja with a report, see the other instance in AFZX 223, no. 258; also ASÈI 2: 533, no. 492; 3: 141, 143, no. 105 (2×); Koreckij 1969: 289, no. 2. For cases of ceˇlovati kr’st” with a report, see AFZX 260, no. 308; ASÈI 3: 386, no. 364; Koreckij 1969: 290, no. 2. 15. Likewise, Hanks (1990: 206) treats as “decentered” (indexically displaced from the reporter’s frame). 16. Six are tagged by s”kazyvati (ASÈI 1: 247, no. 340; 2: 194, no. 285; 3: 182, no. 172; 268–69, no. 250 (2×); Kaštanov 1970: 359, no. 9), two by z”vati (AFZX 227, 230 no. 259 (2×)), and one by kazati (AFZX 260, no. 308). 17. Four are tagged by s”kazyvati (AFZX 222, no. 258; ASÈI 2: 315–16, 317, no. 333 (2×); Kaštanov 1970: 355, no. 9), and one by kazati (ASÈI 2: 438, no. 411). 18. For the other examples, see AFZX 227, 230, no. 259; Kaštanov 1970: 355, no. 9. 19. For the other examples, see ASÈI 2: 194, no. 285; 315–16, no. 333. 20. For the other examples, see AFZX 222, no. 258; 260, no. 308; ASÈI 3: 182, no. 172. 21. The prepositional phrases can be left implicit (ASÈI 1: 235–36, no. 326; 3: 54, no. 31). 22. For other requests for amplification, see ASÈI 1: 235–36, no. 326; 461, no. 582; 541, no. 628; 2: 539, no. 493; 3: 54, no. 31; Kaštanov 1970: 413. For other subversive requotations, see ASÈI 1: 541, no. 628 (with the nazyvati ‘call, claim’); 2: 407, no. 400; 414, no. 404; 3: 54, no. 31. 23. For another excuse for nonfeasance similar to (22), see ASÈI 2: 264, no. 307. For another excuse for a no-show witness (as well for absence of promised documents), see ibid. 2: 432, no. 410. 24. For other syntheses of content, see ASÈI 1: 193, no. 285 (after zaslyšeti ‘hear-’); 396, no. 521 (after slyxati ‘hear-’); 450, no. 571; 512, no. 607a and in a variant transcript from the same trial, 507, no. 607; 2: 539, no. 493; 3: 85, no. 55. For examples with interrogatives or relatives, see ASÈI 1: 502, no. 604; 2: 193, no. 285; Kaštanov 1970: 359, no. 9; 372, no. 16; Krotov and Smetanina 1987: 68, no. 5 (2×)). 25. See also ASÈI 1: 541, no. 628; 2: 264, no. 307; 432, no. 410; 539, no. 493.
316
R V
26. For sporadic instances of deˇ(i) in the transcripts investigated, see 4.7, 8.3.3. The particle occurs once in an additional layer of within the testimony of a poorly preserved trial transcript of the early sixteenth century (ASÈI 3: 245, no. 224), where it performs a disassociating or even dubitative function. 27. There can also be ambiguity in main clauses when the tag occurs after a noun phrase that can be interpreted either as its subject or as the subject of the report clause (e.g., ASÈI 2: 532, no. 492). 28. For further examples in complement clauses, see AFZX 127, no. 140; 212–13, no. 248 (2×); ASÈI 1: 501, no. 604; 507, no. 607 and in the variant transcript from the same trial, 512, no. 607a; 2: 268–69, no. 310; 3: 291, no. 276. For other cases in relative clauses, see AFZX 230, no. 259; 260, no. 308; ASÈI 2: 194, no. 285. For an instance in a topicalizing clause, see ASÈI 2: 438–39, no. 411. For examples in main clauses, see ibid.: 369, no. 374; 532, no. 492. 29. Cf. the intercalated tags in complements after iskati ‘seek; sue’ (ASÈI 2: 268–69, no. 310) and m”lcˇati ‘be silent; not report’ (ASÈI 1: 507, no. 607; 512, no. 607a). 30. For further examples, see AFZX 120, no. 129; 227, no. 259 (2×); ASÈI 1: 320, no. 430; 398, no. 522 (2×); 507, no. 607; 513, no. 607a; 2: 151, no. 229 (2×); 369, no. 374 (2×); 371, no. 375; 415, no. 404 (2×); 445, no. 414; 449, no. 416 (2×). The perfective zam”lcˇati is attested in ibid.: 1: 247, no. 340. 31. For further examples of biti/bivati cˇel”m’, see AFZX 227–28, no. 259 (4×); ASÈI 1: 289, no. 397; 396–97, no. 521; 452, no. 571; 460, no. 581; 501, no. 604 (3×); 507, no. 607; 513, no. 607a; 540–41, no. 628 (2×); 2: 193, no. 285; 369, no. 374 (4×); 432, no. 410; 445, no. 414 (2×); 490, no. 450; 500, no. 463; 503, no. 464; 490, no. 450; 537, no. 493. On the use of this phrasal verb, see also Tarabasova 1963. 32. For some examples with veleˇti, see AFZX 97, no. 103; 116, no. 125 (2×); 229, no. 259; ASÈI 1: 460, no. 581; 501, no. 604; 507, no. 607 (2×); 512, no. 607a (2×); 539, 541, no. 628 (2×); 2: 280, no. 315 (3×); 439, no. 411; 445, no. 414; 490, no. 450 (2×); 496, no. 458; 500–501, no. 463 (4×); 503, no. 464; 536–37, no. 493 (2×); 3: 54, no. 32; 58, no. 35; 84, no. 55. For ukazati, see ibid. 3: 461–62, no. 477 (2×). For prikazati ‘command, entrust’, see AFZX 228, no. 259; Morozov 1988: 304, no. 2. 33. For other cases, see AFZX 97, no. 103; 227–28, no. 259 (2×); ASÈI 1: 193, no. 285; 2: 534, no. 492 (s”kazati); 1: 193, no. 285; 459, no. 581; 3: 141, no. 105 (s”kazyvati); 1: 418, no. 539; 502, no. 604; 2: 415, no. 404; 541, no. 495 (govoriti/govarivati); Krotov and Smetanina 1987: 68, no. 5 (izgovoriti).
Chapter 8 1. For similar cases, see ASÈI 1: 465, no. 583; 471, no. 586; 475, no. 587; 509, no. 607; 515, no. 607a; 2: 194, no. 285; 198, no. 286; 203, no. 288; and 412, no. 402. No capital letter is indicated at the beginning of verdicts in three other autographs (ibid. 1: 479, no. 589; 2: 426, no. 407; 3: 462, no. 477). 2. See, for example, Jakovlev 1943/1970: 321–22, no. 1.5–1.6. 3. Similar cases can be found in northwestern trial transcripts (GVNP 149, no. 92; 327–28, no. 340). 4. For a legal-historical examination of these passages, see Kleimola 1975: 82–83, 86–91.
N
317
5. The ratio decidendi can also be given in a prepositional phrase when judgments are “in accordance with [po]” documents presented during the trial (e.g., ASÈI 2: 203, no. 288; 314, no. 332; 3: 143, no. 105). Such prepositional phrases can be treated in parallel with causal clauses (ibid. 1: 541, no. 628). 6. In a few early transcripts, the rationales incorporate reports of higher-court verifications. Since none of the attested cases involve denials, the reports were inferrable, if not given, since verification was a routine. In two other cases, the reports evidently comprise information presented during the higher hearing but not recorded in the transcript itself (ASÈI 2: 98, no. 286; 541, no. 493). 7. For other examples, see AFZX 98–99, no. 103; 229, no. 259 (quoted in another transcript); 232, no. 259; ASÈI 1: 248–49, no. 340 (2×); 489–90, no. 594; 503, no. 604; 528–29, no. 615; 554, no. 640; 2: 194–95, no. 285; 200, no. 287 (2×); 2: 271, no. 310; 370, no. 374; 423, no. 406; 426, no. 407; 458, no. 421; 470, no. 428; 541, no. 493 (2×); 3: 186, no. 172 (2×); 219, no. 208 (4×); 243, no. 223; 245, no. 224; 272, no. 251; 3: 292, no. 276; and 464, no. 478. 8. The report in (6) may not be prototypical , if the deictic adverb nyneˇcˇa ‘now’ is oriented to the original speaker’s viewpoint, as opposed to being a generalized present including the time of the verdict. 9. In addition to (8)–(9), see AFZX 98–99, no. 103; 140, no. 157; Antonov and Baranov 1997: 118, no. 146; ASÈI 1: 460, no. 581; 2: 198, no. 286; 203, no. 288; 3: 77, no. 50; 88, no. 56; 186, no. 172; 240, no. 221; 245, no. 224; 292, no. 276. 10. For the other cases, see AFZX 140, no. 157; Antonov and Baranov 1997: 118, no. 146; ASÈI 3: 71, no. 48. On null anaphora and cointerpretation, see Halliday and Hasan 1976: 13, 142–44; Brown and Yule 1983: 185, 193; Nichols 1985: 173–74. 11. For the other cases, see ASÈI 3: 77, no. 50; 292, no. 276. Postposition of the subject is also a factor in one of the coreferential examples — Antonov and Baranov 1997: 118, no. 146. 12. For the other examples, see Antonov and Baranov 1997: 118, no. 146; ASÈI 2: 501, no. 463; 541, no. 493; 3: 71, no. 48; 272, no. 251. 13. On repetitions as cohesive devices, see Halliday and Hasan 1976; Brown and Yule 1983: 194; Tannen 1989: 50–51; and (with special reference to citations markers) Kieckers 1920: 202. 14. For a similar example involving the preterit of s”kazati ‘say ()’, see ASÈI 3: 71, no. 48. 15. Similarly, in the formula kladusja/kladem”sja na X (lay-..1./ on X- ‘I/we agree to accept what X says’), the imperfective is replaced by perfective položitisja in the renarration (e.g., AFZX 140, no. 157). 16. For similar cases, see AFZX 107, no. 114; ASÈI 2: 445, no. 414; 3: 243, no. 223; Gorcˇakov 1871, appendix (separate pagination): 65, no. 4.5. 17. Other untagged reports of no-shows are found in AFZX 111, no. 117; ASÈI 1: 354, no. 467; 494, no. 595; 2: 318–19, no. 333 (directed and executed verdicts); 3: 462, no. 477. 18. For similar examples, see ASÈI 2: 410, no. 401; 3: 88, no. 56.
318
R V
Chapter 9 1. On as a semantic motif, see Thompson 1996: 502. 2. Cf. Jakobson 1956/1980: 85; Lucy 1993: 16; Polanyi 1982; Tannen 1989. 3. Cf. Larin 1961: 40; 1975: 256; Lopatina 1985; Tarabasova 1964: 159–60. 4. On the undervaluation of legal-administrative writing, see also Derjagin 1980: 97. 5. For examples, see AI 2: 31–33, no. 34; 146, no. 118; AJu: 357, no. 334.1; AJuB 2: col. 673, no. 230.5; col. 678–83, no. 230.8; Anpilogov 1967: 475–76, ll. 52–53; Juškov 1898: 288, no. 269. 6. Since the compactness/diffuseness distinction is a matter of discourse prominence, it can be graded by means of Chvany’s Saliency Hierarchy (1990/1996), which has axes involving both Subordination and Dialogue (i.e., ) vs. Narrative. However, in treating the salience of , it may be useful to view the salience of “Dialogue” as scalar in itself and to add axes referring to other factors such as the type and placement of the attribution (the “build-up” in the authorial context). 7. On particularity, see also Becker 1988; Goffman 1981: 130–31; Johnstone 1987: 34–35; Tannen 1989.
References
Aaron, Uche E. 1992 “Reported speech in Obolo narrative discourse”. In Language in Context: Essays for Robert E. Longacre [Summer Institute of Linguistics and The University of Texas at Arlington Publications in Linguistics 107], Shin Ja J. Hwang and William R. Merrifield (eds.), 227–40. Arlington, Texas: Summer Institute of Linguistics. Allingham, Margery 1955/1990 The Estate of the Beckoning Lady. New York: Avon Books. Andersen, Henning 1989 “Understanding linguistic innovations”. In Language Change: Contributions to the Study of its Causes, Leiv Egil Breivik and Ernst Håkon Jahr (eds.), 5–27. Berlin: Mouton de Gruyter. Anpilogov, G. I. 1967 Novye dokumenty o Rossii konca XVI-nacˇala XVII v. Moscow: Izdatel’stvo Moskovskogo Universiteta. 1977 Nižegorodskie dokumenty XVI veka. Moscow: Izdatel’stvo Moskovskogo Universiteta. Antonov, A. V., and K. V. Baranov (eds.) 1997 Akty služilyx zemlevladel’cev XV — nacˇala XVII veka: Sbornik dokumentov. Vol. 1. Moscow: Arxeograficˇeskij Centr. Arnovick, Leslie K. 1999 Diachronic Pragmatics: Seven Case Studies in English Illocutionary Development [Pragmatics and Beyond, New Series 68]. Amsterdam: John Benjamins. Arxeograficˇeskaja Kommissija 1836 Akty, sobrannye v bibliotekax i arxivax Rossijskoj Imperii Arxeograficˇeskoju Èkspedicieju Imperatorskoj Akademii Nauk. 1836. Vol. 1. St. Petersburg: Arxeograficˇeskaja Kommissija. 1838 Akty juridicˇeskie, ili sobranie form starinnogo deloproizvodstva. St. Petersburg: Arxeograficˇeskaja Kommissija. 1841 Akty istoricˇeskie, sobrannye i izdannye Arxeograficˇeskoju Kommissieju. Vol. 1. St. Petersburg: Arxeograficˇeskaja Kommissija.
320
R V 1846
Dopolnenie k Aktam istoricˇeskim. Vol. 1. St. Petersburg: Arxeograficˇeskaja Kommissija.
Austin, J. L. 1946/1989
“Other minds”. In J. L. Austin, Philosophical Papers, J. O. Urmson and G. J. Warnock (eds.), 76–116. 3d ed. Oxford: Clarendon. 1966/1989 “Three ways of spilling ink”. In J. L. Austin, Philosophical Papers, J. O. Urmson and G. J. Warnock (eds.), 272–87. 3d ed. Oxford: Clarendon. 1975 How to Do Things with Words. J. O. Urmson and Marina Sbisà (eds.). 2d ed. Cambridge, Massachusetts: Harvard University. Avanesov, R. I. (ed.) 1963 Smolenskie gramoty XIII-XIV vekov. Moscow: Akademija Nauk SSSR. Avanesov, R. I., et al. (eds.) 1975– Slovar’ russkogo jazyka XI — XVII vv. 23 vv. to date. Moscow: Nauka. Bakhtin, M. M. 1929/1971 “Discourse typology in prose”. In Readings in Russian Poetics: Formalist and Structuralist Views, Ladislav Matejka and Krystyna Pomorska (eds.), 176–96. Cambridge, Massachusetts: M. I. T. 1952–53/1986 “The problem of speech genres”. In Speech Genres and Other Late Essays, Vern W. McGee (trans.), 60–102. Austin, Texas: University of Texas. 1975/1981 “Discourse in the novel”. In The Dialogic Imagination, Caryl Emerson and Michael Holquist (trans.), 259–422. Austin, Texas: University of Texas. Bamgbos® e, Ayo 1986 “Reported speech in Yoruba”. In Direct and Indirect Speech [Trends in Linguistics 31], Florian Coulmas (ed.), 77–97. Berlin: Mouton de Gruyter. Banfield, Ann 1973 “Narrative style and the grammar of direct and indirect speech”. Foundations of Language 10: 1–39. 1978 “The formal coherence of represented speech and thought”. PTL: A Journal for Descriptive Poetics and Theory of Literature 3: 289–314. 1982 Unspeakable Sentences: Narration and Representation in the Language of Fiction. Boston: Routledge. 1993 “Where epistemology, style, and grammar meet literary history: The development of represented speech and thought”. In Reflexive Language: Reported Speech and Metapragmatics, John A. Lucy (ed.), 339–64. Cambridge: Cambridge University. Barnet, Vladimír 1965 Vývoj systému participií aktivních v ruštineˇ [Acta Universitatis Carolinae, Philologica 9]. Prague: Karlová Universita.
R
321
Battistella, Edwin L. 1990 Markedness: The Evaluative Superstructure of Language. Albany, New York: State University of New York. Bauman, Richard 1986 Story, Performance, and Event: Contextual Studies of Oral Narrative. Cambridge: Cambridge University. Baynham, Mike 1996 “Direct speech: What’s it doing in non-narrative discourse?” Journal of Pragmatics 25: 61–81. Becker, A. P. 1984 “The linguistics of particularity: Interpreting superordination in a Javanese text”. In Proceedings of the Tenth Annual Meeting of the Berkeley Linguistics Society, Claudia Brugman and Monica Macaulay (eds.), 425–36. Berkeley: Berkeley Linguistics Society. 1988 “Language in particular: A lecture”. In Linguistics in Context: Connecting Observation and Understanding [Advances in Discourse Processes 29], Deborah Tannen (ed.), 17–35. Norwood, New Jersey: Ablex. Birnbaum, Henrik 1985 “Orality, literacy, and literature in Old Rus’”. Die Welt der Slaven 30: 161–96. Blackburn, Simon 1994 The Oxford Dictionary of Philosophy. Oxford: Oxford University. Bloomfield, Leonard 1933/1984 Language. Chicago: University of Chicago. Bolinger, Dwight 1972 That’s That [Janua Linguarum, Series Minor 155]. The Hague: Mouton. Bondar’, I. P. 1967 “Obstojatel’stvennaja svjaz’ nekotoryx glagolov recˇi v staroslavjanskom i v drevnerusskom jazykax”. In Ocˇerki po russkomu jazyku i stilistike, 192–201. Saratov: Saratovskij Gosudarstvennyj Pedagogicˇeskij Institut. Borkovskij, V. I. 1981 Sintaksis skazok: Russko-belorusskie paralleli. Moscow: Nauka. Borkovskij, V. I., and P. S. Kuznecov 1965 Istoricˇeskaja grammatika russkogo jazyka. Moscow: Nauka. Brandsma, Frank 1996 “The presentation of direct discourse in Arthurian romance: Changing modes of performance and reception?” In The Medieval “Opus”: Imitation, Rewriting, and Transmission in the French Tradition, Douglas Kelly (ed.), 245–60. Amsterdam: Rodopi. Brecht, Richard D. 1974 “Deixis in embedded structures”. Foundations of Language 11: 489–518. Brown, Gillian, and George Yule 1983 Discourse Analysis. Cambridge: Cambridge University.
322
R V
Brown, Penelope, and Stephen C. Levinson 1987 Politeness: Some Universals in Language Usage. 2d ed. Cambridge: Cambridge University. Brown, Roger, and Albert Gilman 1989 “Politeness theory and Shakespeare’s four major tragedies”. Language in Society 18: 159–212. Bulaxovskij, L. A. 1954 Russkij literaturnyj jazyk pervoj poloviny XIX veka: Fonetika, morfologija, udarenie, sintaksis. Moscow: Gosudarstvennoe Ucˇebno-pedagogicˇeskoe Izdatel’stvo. 1958 Istoricˇeskij kommentarij k russkomu literaturnomu jazyku. 5th ed. Kiev: Radjanska Škola. Caldas-Coulthard, Carmen Rosa 1994 “On reporting reporting: The representation of speech in factual and factional narratives”. In Advances in Written Text Analysis, Malcolm Coulthard (ed.), 295–308. London: Routledge. Cassirer, Ernst 1955 The Philosophy of Symbolic Forms, Ralph Manheim (trans.). Vol. 1: Language. New Haven: Yale University. ˇ erepnin, L. V. C 1948–51 Russkie feodal’nye arxivy XIV — XV vekov. Moscow: Akademija Nauk SSSR. 1952 “Akty konca XIV — XV vv., ix raznovidnosti i znacˇenie v kacˇestve istoricˇeskogo istocˇnika”. In Akty social’no-èkonomicˇeskoj istorii severovostocˇnoj Rusi konca XIV-nacˇala XVI v., B. D. Grekov and L. V. Cˇerepnin (eds.), 1: 638–50. Moscow: Akademija Nauk SSSR. 1956 Russkaja paleografija. Moscow: Nauka. Cˇerepnin, L. V. (ed.) 1951 Akty feodal’nogo zemlevladenija i xozjajstva XIV — XVI vekov. Vol. 1. Moscow: Akademija Nauk SSSR. Chafe, Wallace 1982 “Integration and involvement in speaking, writing, and oral literature”. In Spoken and Written Language: Exploring Orality and Literacy [Advances in Discourse Processes 9], Deborah Tannen (ed.), 35–53. Norwood, New Jersey: Ablex. 1986 “Evidentiality in English conversation and academic writing”. In Evidentiality: The Linguistic Coding of Epistemology [Advances in Discourse Processes 22], Wallace Chafe and Johanna Nichols (eds.), 261–72. Norwood, New Jersey: Ablex. 1994 Discourse, Consciousness, and Time: The Flow and Displacement of Conscious Experience in Speaking and Writing. Chicago: University of Chicago.
R
323
Chatman, Seymour 1975 “The structure of narrative transmission”. In Style and Structure in Literature: Essays in the New Stylistics, Roger Fowler (ed.), 213–57. Ithaca, N. Y.: Cornell University. Chvany, Catherine V. 1973/1996 “On ‘root’ and ‘structure-preserving’ — disposable blades for Occam’s Razor”. In Selected Essays of Catherine V. Chvany, Olga T. Yokoyama and Emily Klenin (eds.), 104–27. 1985a “Backgrounded perfectives and plot line imperfectives: Toward a theory of grounding in text”. In The Scope of Slavic Aspect [UCLA Slavic Studies 12], Michael S. Flier and Alan Timberlake (eds.), 247–73. Columbus, Ohio: Slavica. 1985b “Foregrounding, ‘transitivity’, saliency (in sequential and non-sequential prose)”. Essays in Poetics 10: 1–27. 1990/1996 “Verbal aspect, discourse saliency, and the so-called ‘perfect of result’ in Modern Russian”. In Selected Essays of Catherine V. Chvany, Olga T. Yokoyama and Emily Klenin (eds.), 286–99. Clark, Herbert H. 1987 “Four dimensions of language use”. In The Pragmatic Perspective: Selected Papers from the 1985 International Pragmatics Conference [Pragmatics and Beyond Companion Series 5], Jef Verschueren and Marcella BertuccelliPapi (eds.), 9–25. Amsterdam: John Benjamins. 1992 Arenas of Language Use. Chicago: University of Chicago. Clark, Herbert H., and Richard J. Gerrig 1990 “Quotations as demonstrations”. Language 66: 764–805. Clark, Katerina, and Michael Holquist 1984 Mikhail Bakhtin. Cambridge, Massachusetts: Belknap. Collins, Daniel E. 1994 The Pragmatics of Reported Speech in Medieval Russian Trial Transcripts: The Contextualization and Interpretation of Metapragmatic Framing Strategies in Old Russian Juridical Language, 1400–1505. Unpublished Ph. D. Dissertation, UCLA. 1996 “The pragmatics of indirect speech in Old Church Slavonic and other early Slavic writings”. Studies in Slavic and General Linguistics 23: 21–86. 1997 “The function and social context of prošcˇal’nye gramoty (Toward a thicker description of a medieval Russian text kind)”. Zeitschrift für Slawistik 42: 192–231. Collins, James 1987 “Reported speech in Navajo myth-narratives”. In Linguistic Action: Some Empirical-Conceptual Studies [Advances in Discourse Processes 23], Jef Verschueren (ed.), 69–84. Norwood, New Jersey: Ablex.
324
R V
Comrie, Bernard 1976 Aspect: An Introduction to the Study of Verbal Aspect and Related Problems. Cambridge: Cambridge University. 1986 “Tense in indirect speech”. Folia Linguistica 20: 265–96. Coulmas, Florian 1985a “Direct and indirect speech: General problems and problems of Japanese”. Journal of Pragmatics 9: 41–63. 1985b “Nobody dies in Shangri-La: Direct and indirect speech across languages”. Georgetown University Round Table on Languages and Linguistics 1985, 140–151. 1986a “Introduction: Conversational routine”. In Conversational Routine: Explorations in Standardized Communication Situations and Prepatterned Speech, Florian Coulmas (ed.), 1–17. The Hague: Mouton. 1986b “Reported speech: Some general issues”. In Direct and Indirect Speech [Trends in Linguistics 31], Florian Coulmas (ed.), 1–28. Berlin: Mouton de Gruyter. Coulmas, Florian (ed.) 1986 Direct and Indirect Speech [Trends in Linguistics 31]. Berlin: Mouton de Gruyter. Couper-Kuhlen, Elizabeth 1999 “Coherent voicing: On prosody in conversational reported speech”. In Coherence in Spoken and Written Discourse: How to Create it and How to Describe it [Pragmatics and Beyond, New Series 63], Wolf Bublitz et al. (eds.), 11–33. Amsterdam: John Benjamins. Cram, D. F. 1978 “The syntax of direct quotation”. Cahiers de lexicologie 33: 41–52. Cˇumakov, G. M. 1975 Sintaksis konstrukcij s cˇužoj recˇ’ju. Kiev: Vyšcˇa Škola. Derjagin, V. Ja. 1971 “K voprosu ob individual’nom i tradicionnom v delovoj pis’mennosti (na materiale važskix porjadnyx XVII v.)”. In Russkij jazyk: Istocˇniki dlja ego izucˇenija, S. I. Kotkov (ed.), 151–88. Moscow: Nauka. 1980 “Ob istoriko-stilicˇ eskom issledovanii aktovyx tekstov”. Voprosy jazykoznanija, 1980, no. 4: 97–107. Dewey, H. W., and A. M. Kleimola (eds.) 1973 Russian Private Law in the XIV — XVII Centuries: An Anthology of Documents [Michigan Slavic Materials 9]. Ann Arbor, Michigan: University of Michigan Department of Slavic Languages and Literatures. Dirven, Rene, Louis Goosens, Yvan Putseys, and Emma Vorlat 1982 The Scene of Linguistic Action and its Perspectivization by Speak, Talk, Say and Tell [Pragmatics and Beyond 3 (6)]. Amsterdam: John Benjamins.
R
325
Diver, William 1969 “The system of relevance of the Homeric verb”. Acta Linguistica Hafniensia 12: 45–68. D’jakonov, M. A. 1897 Akty, otnosjašcˇiesja k istorii tjaglogo naselenija v Moskovskom gosudarstve. Part 2: Gramoty i zapisi. Jur’ev: Tip. K. Matisena. Dmitriev, L. A., and D. S. Lixacˇev (eds.) 1982 Pamjatniki literatury Drevnej Rusi: Vtoraja polovina XV veka. Moscow: Xudožestvennaja Literatura. Doležel, Lubomír 1973 Narrative Modes in Czech Literature. Toronto: University of Toronto. Dubois, Betty Lou 1989 “Pseudoquotation in current English communication: ‘Hey, she didn’t really say it’”. Language in Society 18: 343–59. Du Bois, John W. 1986 “Self-evidence and ritual speech”. In Evidentiality: The Linguistic Coding of Epistemology [Advances in Discourse Processes 22], Wallace Chafe and Johanna Nichols (eds.), 313–36. Norwood, New Jersey: Ablex. Ebert, Karen 1986 “Reported speech in some languages of Nepal”. In Direct and Indirect Speech [Trends in Linguistics 31], Florian Coulmas (ed.), 145–59. Berlin: Mouton de Gruyter. Ehlich, Konrad 1981 “Native Speaker’s heritage: On philology of ‘dead’ languages”. In A Festschrift for Native Speaker [Janua Linguarum, Series Maior 97], Florian Coulmas (ed.), 153–65. The Hague: Mouton. Enkvist, Nils Erik 1987 “A note towards the definition of text strategy”. Zeitschrift für Phonetik, Sprachwissenschaft und Kommunikationsforschung 40: 19–27. Fleischman, Suzanne 1990 “Philology, linguistics, and the discourse of the medieval text”. Speculum 65: 19–37. Fletcher, Giles 1966 Of the Rus Commonwealth. Albert J. Schmidt (ed.). Ithaca, N. Y.: Cornell University. Fludernik, Monika 1993 The Fictions of Language and the Languages of Fiction. London: Routledge. Fónagy, Ivan 1986 “Reported speech in French and Hungarian”. In Direct and Indirect Speech [Trends in Linguistics 31], Florian Coulmas (ed.), 255–309. Berlin: Mouton de Gruyter.
326 Forsyth, J. 1970
R V
A Grammar of Aspect: Usage and Meaning in the Russian Verb. Cambridge: Cambridge University.
Fox, Barbara 1983 “The discourse function of the participle in Ancient Greek”. In Discourse Perspectives on Syntax, Flora Klein-Andreu (ed.), 23–41. New York: Academic Press. Frajzingier, Zygmunt 1991 “The de dicto domain in language”. In Approaches to Grammaticalization [Typological Studies in Language 19], Elizabeth Closs Traugott and Bernd Heine (eds.), 1: 219–51. Amsterdam: John Benjamins. Geertz, Clifford 1973 “Thick description: Toward an interpretive theory of culture”. In Clifford Geertz, The Interpretation of Cultures, 3–30. New York: Basic Books. Gennadij 1499/1992 Biblija. Vol. 8. Reprint. Moscow: Moskovskij Patriarxat. Givón, Talmy 1980 “The binding hierarchy and the typology of complements”. Studies in Language 4: 333–377. 1984–90 Syntax: A Functional-Typological Introduction. 2 vv. Amsterdam: John Benjamins. 1989 Mind, Code and Context: Essays in Pragmatics. Hillsdale, New Jersey: Laurence Erlbaum. Glock, Naomi 1986 “The use of reported speech in Saramaccan discourse”. In Pragmatics in Non-Western Perspective [Summer Institute of Linguistics Publications in Linguistics 73], George Huttar and Kenneth Gregerson (eds.), 35–61. Arlington, Texas: Summer Institute of Linguistics. Goffman, Erving 1981 Forms of Talk. Philadelphia: University of Pennsylvania. Goody, Jack 1977 The Domestication of the Savage Mind. Cambridge: Cambridge University. Goosens, Louis 1987 “Dealing with linguistic action verbs in depth”. In Linguistic Action: Some Empirical-Conceptual Studies [Advances in Discourse Processes 23], Jef Verschueren (ed.), 95–116. Norwood, New Jersey: Ablex. Gorcˇakov, M. 1871 O zemel’nyx vladenijax vserossijskix mitropolitov, patriarxov i sv. Sinoda (988–1738): Iz opytov izsledovanija v istorii russkogo prava. St. Petersburg: Tip. A. Transhelia. Green, Georgia M. 1980 “Some wherefores of English inversions”. Language 56: 582–601.
R
327
Green, Georgia M., and Jerry L. Morgan 1981 “Pragmatics, grammar, and discourse”. In Radical Pragmatics, Peter Cole (ed.), 167–81. New York: Academic Press. Grekov, B. D. (ed.) 1952 Sudebniki XV — XVI vekov. Moscow: Akademija Nauk SSSR. Grekov, B. D., and L. V. Cˇerepnin (eds.) 1952–64 Akty social’no-èkonomicˇeskoj istorii severo-vostocˇnoj Rusi konca XIV — nacˇala XVI v. 3 vv. Moscow: Nauka. Grice, Paul 1975/1989 “Logic and conversation”. In Studies in the Way of Words, 22–40. Cambridge, Massachusetts: Harvard University. Grimes, Joseph E. 1975 The Thread of Discourse. The Hague: Mouton. Gvozdanovic´, Jadranka 1995 “Parameters underlying the organization of medieval Russian texts”. In Historical Linguistics 1993 [Current Issues in Linguistic Theory 124], Henning Andersen (ed.), 177–90. Amsterdam: John Benjamins. Gvozdev, A. N. 1958 Sovremennyj russkij jazyk. Vol. 2: Sintaksis. Moscow: Gosudarstvennoe Ucˇebno-pedagogicˇeskoe Izdatel’stvo. Haberland, Harmut 1986 “Reported speech in Danish”. In Direct and Indirect Speech [Trends in Linguistics 31], Florian Coulmas (ed.), 219–53. Berlin: Mouton de Gruyter. Hagenaar, Elly 1996 “Free indirect speech in Chinese”. In Reported Speech: Forms and Functions of the Verb [Pragmatics and Beyond, New Series 43], Theo A. J. M. Janssen and Wim van der Wurff (eds.), 289–98. Amsterdam: John Benjamins. Haiman, John 1983 “Iconic and economic motivation”. Language 59: 781–819. Haiman, John, and Sandra A. Thompson 1984 “’Subordination’ in universal grammar”. In Proceedings of the Tenth Annual Meeting of the Berkeley Linguistics Society, Claudia Brugman and Monica Macaulay (eds.), 510–23. Berkeley: Berkeley Linguistics Society. Halliday, M. A. K., and Ruqaiya Hasan 1976 Cohesion in English [English Language Series 9]. London: Longman. Hammer, A. E. 1983 German Grammar and Usage. 2d ed. Baltimore: E. Arnold. Hanks, William F. 1990 Referential Practice: Language and Lived Space among the Maya. Chicago: University of Chicago.
328
R V
Havránek, Bohuslav 1932 “Úkoly spisovného jazyka a jeho kultura”. In Spisovná cˇestina a jazyková kultura, Bohuslav Havránek and Miloš Weingart (eds.), 32–84. Prague: Melantrich. 1964 “The functional differentiation of the standard language”. In A Prague School Reader on Esthetics, Literary Structure, and Style, Paul L. Garvin (trans.), 3–16. Washington, D. C.: Georgetown University. Hermon, Gabriella 1979 On the Discourse Structure of Direct Quotation [Center for the Study of Reading, Technical Report 143]. Cambridge, Massachusetts: Bolt Beranek and Newman, Inc. Hickmann, Maya 1993 “The boundaries of reported speech in narrative discourse: Some developmental aspects”. In Reflexive Language: Reported Speech and Metapragmatics, John A. Lucy (ed.), 63–90. Cambridge: Cambridge University. Holt, Elizabeth 1996 “Reporting on talk: The use of direct reported speech in conversation”. Research on Language and Social Interaction 29: 219–45. Hopper, Paul J. 1979 “Aspect and foregrounding in discourse”. Syntax and Semantics 12: 213–41. 1988 “Emergent grammar and the a priori grammar postulate”. In Linguistics in Context: Connecting Observation and Understanding [Advances in Discourse Processes 29], Deborah Tannen (ed.), 117–34. Norwood, New Jersey: Ablex. Hopper, Paul J., and Sandra A. Thompson 1980 “Transitivity in grammar and discourse”. Language 56: 251–99. 1984 “The Discourse Basis for Lexical Categories in Universal Grammar”. Language 60: 703–52. Huizinga, J. 1944/1955 Homo Ludens: A Study of the Play-Element in Culture. Boston: Beacon Press. Ickler, Nancy 1977 “Topicalization and relativization in Old Russian”. In Proceedings of the 3rd Annual Meeting of the Berkeley Linguistics Society, Kenneth Whistler et al. (eds.), 656–69. Berkeley: Berkeley Linguistics Society. Issatschenko, Alexander V. 1980 “Russian”. In The Slavic Literary Languages: Formation and Development [Yale Russian and East European Publications 1], Alexander M. Schenker and Edward Stankiewicz (eds.), 119–42. New Haven, Connecticut: Yale Concilium on International and Area Studies. Ivanov, Vjacˇ. Vs. 1965 Obšcˇeindoevropejskaja, praslavjanskaja i anatolijskaja jazykovye sistemy: Sravnitel’no-tipologicˇeskie ocˇerki. Moscow: Nauka.
R 1973
329
“Znacˇenie idej M. M. Baxtina o znake, vyskazyvanija i dialoge dlja sovremennoj semiotiki”. Ucˇenye Zapiski Tartusskogo Gosudarstvennogo Universiteta 6: 5–45. Ivina, L. I. (ed.) 1983 Akty feodal’nogo zemlevladenija i xozjajstva: Akty moskovskogo Simonova monastyrja (1506–1613 gg.). Leningrad: Nauka. Jacobs, Andreas, and Andreas H. Jucker 1995 “The historical perspective in pragmatics”. In Historical Pragmatics: Pragmatic Developments in the History of English [Pragmatics and Beyond, New Series 35], Andreas H. Jucker (ed.), 3–33. Amsterdam: John Benjamins. Jakobson, Roman 1956/1980 “Metalanguage as a linguistic problem”. In The Framework of Language, 81–92. Ann Arbor, Michigan: University of Michigan. 1957/1984 “Shifters, verbal categories, and the Russian verb”. In Russian and Slavic Grammar: Studies 1931–1981, Linda R. Waugh and Morris Halle (eds.), 41–58. Berlin: Mouton de Gruyter. Jakovlev, A. 1943/1970 Xolopstvo i xolopi v Moskovskom gosudarstve XVII v.: Po arxivnym dokumentam Xolop’ego i Posol’skogo prikazov, Oružejnoj palaty i Razrjada [Slavistic Printings and Reprintings 217]. Vol. 1. Reprint. The Hague: Mouton. Janssen, Theo A. J. M., and Wim van der Wurff (eds.) 1996 Reported Speech: Forms and Functions of the Verb [Pragmatics and Beyond, New Series 43]. Amsterdam: John Benjamins. Jespersen, Otto 1924/1992 The Philosophy of Grammar. Reprint. Chicago: University of Chicago. Johnstone, Barbara 1987 “’He says… so I said’: Verb tense alternation and narrative depictions of authority in American English”. Linguistics 25: 33–52. Jucker, Andreas H. (ed.) 1995 Historical Pragmatics: Pragmatic Developments in the History of English [Pragmatics and Beyond, New Series 35]. Amsterdam: John Benjamins. Jucker, Andreas H., Gerd Fritz, and Franz Lebsanft (eds.) 1999 Historical Dialogue Analysis [Pragmatics and Beyond, New Series 66]. Amsterdam: John Benjamins. Juškov, A. I. 1898 Akty XIII — XVII vv., predstavlennye v Razrjadnyj prikaz predstaviteljami služilyx familij posle otmeny mestnicˇestva. Vol. 1: 1257–1613. Moscow: Universitetskaja Tip. Kaiser, Daniel H. 1980 The Growth of the Law in Medieval Russia. Princeton, New Jersey: Princeton University.
330
R V
Kalacˇov, Nikolaj (ed.) 1857–84/1970 Akty, otnosjašcˇiesja do juridicˇeskogo byta Drevnej Rossii, izdany Arxeograficˇeskoju Kommissieju [Slavistic Printings and Reprintings 252]. 3 vv. The Hague: Mouton. Karpov, G. O. (ed.) 1887 Pamjatniki diplomaticˇeskix snošenij drevnej Rusi s deržavami inostrannymi [Sbornik Imperatorskogo Russkogo Istoricˇeskogo Obšcˇestva 59]. Vol. 2. St. Petersburg: Imperatorskoe Russkoe Istoricˇeskoe Obšcˇestvo. Kaštanov, S. M. 1970 Ocˇerki russkoj diplomatiki. Moscow: Nauka. 1988 Russkaja diplomatika. Moscow: Nauka. 1998 Aktovaja arxeografija. Moscow: Nauka. Kay, Paul 1977 “Language evolution and speech style”. In Sociocultural Dimensions of Language Change, Ben G. Blount and Mary Sanches (eds.), 21–33. New York: Academic Press. Kazakova, N. A. 1960 Vassian Patrikeev i ego socˇinenija. Leningrad: Akademija Nauk SSSR. Keller, Rudi 1985 “Towards a theory of linguistic change”. In Linguistic Dynamics: Discourses, Procedures and Evolution [Research in Text Theory 9], Thomas T. Ballmer (ed.), 211–37. New York: W. de Gruyter. Kieckers, E. 1920 “Zum ‘pleonastischen’ inquit”. Glotta 10: 200–209. Kiparsky, Paul, and Carol Kiparsky 1970 “Fact”. In Progress in Linguistics, Manfred Bierwisch and Karl Erich Heidolph (eds.), 143–73. The Hague: Mouton. Kittay, Jeffrey 1988 “Utterance unmoored: The changing interpretation of the act of writing in the European Middle Ages”. Language in Society 17: 209–30. Kleimola, Ann M. 1972 “Formulae and formalism in Muscovite judgment charters”. CanadianAmerican Slavic Studies 6: 355–73. 1975 Justice in Medieval Russia: Muscovite Judgment Charters (Pravye Gramoty) of the Fifteenth and Sixteenth Centuries [Transactions of the American Philosophical Society, New Series, 65, Part 6]. Philadelphia: American Philosophical Society. Klenin, Emily 1983 Animacy in Russian: A New Interpretation [UCLA Slavic Studies 6]. Columbus, Ohio: Slavica. Klewitz, Gabriele, and Elizabeth Couper-Kuhlen 1999 “Quote — unquote? The role of prosody in the contextualization of reported speech sequences”. Pragmatics 9: 459–85.
R
331
Koduxov, V. I. 1955 “Sposoby peredacˇi cˇužoj recˇi v russkom jazyke”. Ucˇenye Zapiski Leningradskogo Gosudarstvennogo Pedagogicˇeskogo Instituta 104: 107–72. Koreckij, V. I. 1969 “Vnov’ otkrytye novogorodskie i pskovskie gramoty XIV-XV vv.”. Arxeograficˇeskij ežegodnik za 1967 god, 275–90. Kotkov, S. I. (ed.) 1990 Pamjatniki južnovelikorusskogo narecˇija: Konec XVI-nacˇalo XVII v. Moscow: Nauka. Kotkov, S. I., and I. S. Filippova (eds.) 1978 Pamjatniki russkoj pis’mennosti XV — XVI vv.: Rjazanskij kraj. Moscow: Nauka. Kotkov, S. I., and Z. D. Popova 1986 Ocˇerki po sintaksisu južnovelikorusskoj pis’mennosti XVII veka. Moscow: Nauka. Kozlovskij, Pavel 1890 “O socˇetanii predloženij prjamoj i kosvennoj recˇi v russkom jazyke”. Filologicˇeskie zapiski 4–5: 1–7. Krotov, M. G. and S. I. Smetanina (eds.) 1987 “Istocˇniki po istorii feodal’nogo zemlevladenija Velikogo knjažestva Rjazanskogo v XV v.”. Sovetskie arxivy, 1987, no. 1: 66–69. Kuno, Susumu 1987 Functional Syntax: Anaphora, Discourse and Empathy. Chicago: University of Chicago. Kurihara, Takehiko 1985 “The syntax and semantics of quotation reconsidered”. Hokkaid˙ao Daigaku 7: 67–89. Kuz’mina, I. B., and E. V. Nemcˇenko 1980 “K istorii kratkix form dejstvitel’nyx pricˇastij v russkom jazyke”. In Obšcˇeslavjanskij lingvisticˇeskij atlas: Materialy i issledovanija 1978, R. I. Avanesov (ed.), 151–200. Moscow: Nauka. Labov, William 1972 “The transformation of experience in narrative syntax”. In William Labov, Language in the Inner City: Studies in the Black English Vernacular, 354–96. Philadelphia: University of Pennsylvania. 1981 “Speech actions and reactions in personal narrative”. Georgetown University Round Table on Languages and Linguistics 1981: 219–47. Lakoff, Robin Tolmach 1984 “The pragmatics of subordination”. In Proceedings of the Tenth Annual Meeting of the Berkeley Linguistics Society, Claudia Brugman and Monica Macaulay (ed.), 481–92. Berkeley.
332
R V 1995
“Conversational logic”. In Handbook of Pragmatics Manual, Jef Verschueren, Jan-Ola Östman, and Jan Blommaert (eds.), 190–98. Amsterdam: John Benjamins.
Larin, B. A. 1961 “Razgovornyj jazyk Moskovskoj Rusi”. In Nacˇal’nyj ètap formirovanija russkogo nacional’nogo jazyk, B. A. Larin, I. P. Eremin, and M. A. Sokolova (eds.), 22–34. Leningrad: Izdatel’stvo Leningradskogo Universiteta. 1975 Lekcii po istorii russkogo literaturnogo jazyka (X — seredina XVIII v.). Moscow: Vysšaja Škola. Larson, Mildred Lucille 1978 The Functions of Reported Speech in Discourse [Summer Institute of Linguistics, Publications in Linguistics 59]. Dallas, Texas: Summer Institute of Linguistics. Lee, Benjamin 1993 “Metalanguages and subjectivities”. In Reflexive Language: Reported Speech and Metapragmatics, John A. Lucy (ed.), 365–91. Cambridge: Cambridge University. Leech, Geoffrey N. 1974 Semantics. New York: Penguin Books. Leech, Geoffrey N., and Michael H. Short 1981 Style in Fiction: A Linguistic Introduction to English Fictional Prose [English Language Series 13]. London: Longman. Lennard, John 1995 “Punctuation: And — ‘Pragmatics.’” In Historical Pragmatics: Pragmatic Developments in the History of English [Pragmatics and Beyond, New Series 35], Andreas H. Jucker (ed.), 65–98. Amsterdam: John Benjamins. Leont’ev, A. K. 1969 “Pravo i sud”. In Ocˇerki russkoj kul’tury XIII — XV vekov, A. V. Arcixovskij et al. (eds.), 1: 5–48. Moscow: Izdatel’stvo Moskovskogo Universiteta. Lewis, David 1969 Convention: A Philosophical Study. Cambridge, Massachusetts: Harvard University. Li, Charles 1986 “Direct speech and indirect speech: A functional study”. In Direct and Indirect Speech [Trends in Linguistics 31], Florian Coulmas (ed.), 29–45. Berlin: Mouton de Gruyter. Liberzon, I. Z. (ed.) 1988–90 Akty social’no-èkonomicˇeskoj istorii severa Rossii konca XV — XVI v.: Akty Soloveckogo monastyrja. 2 vv. Leningrad: Nauka. Lips, Marguerite 1926 Le style indirect libre. Paris: Payot.
R Lixacˇev, D. S. 1946/1986
333
“Russkij posol’skij obycˇaj XI — XIII vv.”. In Issledovanija po drevnerusskoj literature, 140–53. Leningrad: Nauka. 1952 Vozniknovenie russkoj literatury. Moscow: Akademija Nauk SSSR. Lixacˇev, N. P. (ed.) 1894 Mestnicˇeskie dela 1563–1605 gg. St. Petersburg: Sinodal’naja Tipografija. 1895 Sbornik aktov sobrannyx v arxivax i bibliotekax. St. Petersburg: V. S. Balašev. Longacre, Robert E. 1983 The Grammar of Discourse. New York: Plenum. 1985 “Discourse peak as zone of turbulence”. In Beyond the Sentence: Discourse and Sentential Form, Jessica R. Wirth (ed.), 79–98. Ann Arbor, Michigan: Karoma. 1994 “The dynamics of reported dialogue in narrative”. Word 45: 125–43. Lopatina, L. E. 1979 “Sposoby peredacˇi cˇužoj recˇi v ramkax složnogo predloženija”. In Istoricˇeskaja grammatika russkogo jazyka: Sintaksis: Složnoe predloženie, V. I. Borkovskij (ed.), 422–47. Moscow: Nauka. 1985 “Sredstva organizacii teksta pri peredacˇe cˇužoj recˇi: Na materiale sudebnyx dokumentov XVII v.”. In Vostocˇnye slavjane: Jazyki, istorija, kul’tura, J. N. Karaulov (ed.), 57–64. Moscow: Nauka. 1987 “Zakonomernosti sintaksicˇeskoj struktury teksta v pamjatnikax narodnorazgovornoj raznovidnosti russkogo literaturnogo jazyka XVI — XVII vv.”. In Drevnerusskij literaturnyj jazyk v ego otnošenii k staroslavjanskomu, L. P. Žukovskaja (ed.), 237–45. Moscow: Nauka. Lucy, John A. 1993 “Reflexive language and the human disciplines”. In Reflexive Language: Reported Speech and Metapragmatics, John A. Lucy (ed.), 9–32. Cambridge: Cambridge University. Lucy, John A. (ed.) 1993 Reflexive Language: Reported Speech and Metapragmatics. Cambridge: Cambridge University. Ludolf, Heinrich Wilhelm 1696/1959 Grammatica Russica Oxonii A. D. MDCXCVI. Reprint. B. O. Unbegaun (ed.). Oxford: Clarendon. Lyons, John 1968 Introduction to Theoretical Linguistics. Cambridge: Cambridge University. 1977 Semantics. 2 vv. Cambridge: Cambridge University. Machek, Václav 1968 Etymologický slovník jazyka cˇeského. 2d ed. Prague: Academia. Martysevicˇ, I. D. 1951 Pskovskaja sudnaja gramota: Istoriko-juridicˇeskoe issledovanie. Moscow: Izdatel’stvo Moskovskogo Universiteta.
334
R V
Massamba, David P. B. 1986 “Reported speech in Swahili”. In Direct and Indirect Speech [Trends in Linguistics 31], Florian Coulmas (ed.), 99–119. Berlin: Mouton de Gruyter. Mathis, Terrie, and George Yule 1994 Zero quotatives. Discourse Processes 18: 63–76. Mayes, Patricia 1990 “Quotation in spoken English”. Studies in Language 14: 325–63. McHale, Brian 1978 “Free indirect discourse: A survey of recent accounts”. PTL: A Journal for Descriptive Poetics and Theory of Literature 3: 249–87. McKay, Janet Holmgren 1978 “Some problems in the analysis of point of view in reported discourse”. Centrum 6: 5–26. Miller, George A., and Philip N. Johnson-Laird 1976 Language and perception. Cambridge, Massachusetts: Belknap. Mithun, Marianne 1986 “Evidential diachrony in Northern Iroquoian”. In Evidentiality: The Linguistic Coding of Epistemology [Advances in Discourse Processes 22], Wallace Chafe and Johanna Nichols (eds.), 89–112. Norwood, New Jersey: Ablex. Molotkov, A. I. 1958 “K istorii sintaksicˇeskix konstrukcij dlja peredacˇi cˇužoj recˇi v russkom jazyke: Konstrukcija prjamoj recˇi”. Ucˇenye Zapiski Leningradskogo Gosudarstvennogo Universiteta 235: 23–49. Morawski, Stefan 1970 “The basic functions of quotation”. In Sign, Language, Culture [Janua Linguarum, Series Maior 1], A. J. Greimas et al. (eds.), 690–705. The Hague: Mouton. Morozov, B. N. 1988 “Gramoty XIV — XVI vv. iz kopijnoj knigi Rjazanskogo arxierejskogo doma”. Arxeograficˇeskij ežegodnik za 1987 god, 298–309. Munro, Pamela 1982 “On the transitivity of ‘say’ verbs”. Syntax and Semantics 15: 301–17. Nichols, Johanna 1980 “The syntax of Old Russian mıæneˇti (sja)”. In American Indian and Indoeuropean Studies: Papers in Honor of Madison S. Beeler [Trends in Linguistics 16], Kathryn Klar, Margaret Langdon, and Shirley Silver (eds.), 421–39. The Hague: Mouton. 1985 “The grammatical marking of theme in literary Russian”. In Issues in Russian Morphosyntax, Michael S. Flier and Richard D. Brecht (eds.), 170–86. Columbus, Ohio: Slavica.
R
335
Nichols, Johanna, and Joe Schallert 1983 “The pragmatics of raising in Old Russian and Common Slavic”. In American Contributions to the Ninth International Congress of Slavists, vol. 1: Linguistics, Michael S. Flier (ed.), 221–46. Columbus, Ohio: Slavica. Nikiforov, S. D. 1952 Glagol: Ego kategorii i formy v russkoj pis’mennosti vtoroj poloviny XVI veka. Moscow: Akademija Nauk SSSR. Noonan, Michael 1985 “Complementation”. In Language Typology and Syntactic Description, Timothy Shopen (ed.), vol. 2: Complex Constructions, 42–140. Cambridge: Cambridge University. Novosel’skij, A. A. et al. (eds.) 1975 Akty russkogo gosudarstva 1505–1526. Moscow: Nauka. Ong, Walter J. 1981 Oral remembering and narrative structures. Georgetown University Round Table on Languages and Linguistics 1981, 12–24. 1982 Orality and Literacy: The Technologizing of the Word. New York: Methuen. Osipov, B. I. 1992 Istorija russkoj orfografii i punktuacii. Novosibirsk: Izdatel’stvo Novosibirskogo Universiteta. Otin, E. S. 1969 “K istorii razvitija formy kosvennoj recˇi v drevnerusskom jazyke”. Naucˇnye Doklady Vysšej Školy. Filologicˇeskie Nauki, 1969, no. 3: 54–64. Paducˇeva, E. V. 1996 Semanticˇeskie issledovanija (Semantika vremeni i vida v russkom jazyke; Semantika narrativa). Moscow: Jazyki Russkoj Kultury. Page, Norman 1988 Speech in the English Novel. 2d ed. London. Palmer, F. R. 1981 Semantics. 2d ed. Cambridge: Cambridge University. 1986 Mood and modality. Cambridge: Cambridge University. Panteleimon, Arximandrit 1904–11 Žitija svjatyx na russkom jazyke, izložennye po rukovodstvu Cˇet’ix-Minej sv. Dimitrija Rostovskogo. 12 vv. Moscow: Sinodal’naja Tipografija. Partee, Barbara Hall 1973 “The syntax and semantics of quotation”. In A Festschrift for Morris Halle, Stephen R. Anderson and Paul Kiparsky (eds.), 410–18. New York: Holt, Rinehart, and Winston. Pascal, Roy 1977 The Dual Voice: Free Indirect Speech and its Functioning in the NineteenthCentury European Novel. Manchester: Manchester University. Pearce-Higgins, Lucinda 1989 Goldilocks and the Three Bears. Austin, Texas: Steck-Vaughn.
336
R V
P’ecux, Vjacˇeslav 1990 Novaja moskovskaja filosofija: Roman. In Vjacˇeslav P’ecux, Ja i procˇee, 218–333. Moscow: Xudožestvennaja Literatura. Peškovskij, A. M. 1956 Russkij sintaksis v naucˇnom osvešcˇenii. 7th ed. Moscow: Gosudarstvennoe Ucˇebno-pedagogicˇeskoe Izdatel’stvo. Philips, Susan U. 1985 “Reported speech as evidence in an American trial”. In Georgetown University Round Table on Languages and Linguistics 1985, 154–70. Pokrovskij, N. N. (ed.) 1971 Sudnye spiski Maksima Greka i Isaka Sobaki. Moscow: Glavnoe Arxivnoe Upravlenie. 1973 Aktovye istocˇniki po istorii cˇernosošnogo zemlevladenija v Rossii XIV — nacˇala XVI v. Novosibirsk: Nauka. Polanyi, Livia 1982 “Literary complexity in everyday storytelling”. In Spoken and Written Language: Exploring Orality and Literacy [Advances in Discourse Processes 9], Deborah Tannen (ed.), 155–69. Norwood, New Jersey: Ablex. Potebnja, A. A. 1899/1958–85 Iz zapisok po russkoj grammatike. 4 vv. in 3 pts. Reprint. V. I. Borkovskij (ed.). Moscow: Ucˇpedgiz. Preobraženskaja, M. N. 1983 “Specifika složnyx sintaksicˇeskix konstrukcij v uslovijax drevnerusskogo teksta”. In Struktura predloženija v istorii vostocˇnoslavjanskix jazykov, V. I. Borkovskij (ed.), 102–8. Moscow: Nauka. Prince, Gerald 1987 A Dictionary of Narratology. Lincoln, Nebraska: University of Nebraska. Quine, Willard Van Orman 1960 Word and Object. Cambridge, Massachusetts: M. I. T. Rader, Margaret 1982 “Context in written language: The case of imaginative fiction”. In Spoken and Written Language: Exploring Orality and Literacy, Deborah Tannen (ed.), 185–98. Advances in Discourse Processes, 9. Norwood, New Jersey. Reddy, Michael J. 1979 “The Conduit Metaphor — a case of frame conflict in our language about language”. In Metaphor and Thought, Anthony Ortony (ed.), 284–324. Cambridge: Cambridge University. Remnick, David 1994 “The exile returns”. The New Yorker, February 14, 64–83. Rimmon-Kenan, Shlomith 1983 Narrative Fiction: Contemporary Poetics. London: Methuen.
R
337
Rinberg, V. L. 1985 “K probleme konstrukcii prjamoj recˇi i svjaznogo teksta v istoricˇeskom sintaksise”. In Vostocˇnye slavjane: Jazyki, istorija, kul’tura, J. N. Karaulov (ed.), 49–57. Moscow: Nauka. Romaine, Suzanne, and Deborah Lange 1991 The use of like as a marker of reported speech and thought: A case of grammaticalization in progress. American Speech 66: 227–79. Roncador, Manfred von 1988 Zwischen direkter und indirekter Rede: Nichtwörtliche direkte Rede, erlebte Rede, logophorische Konstruktionen und Verwandtes [Linguistiche Arbeiten 192]. Tübingen: M. Niemeyer. Rosen, Harold 1980 “The dramatic mode”. In Coming to Know, Phillida Salmon (ed.), 152–69. London: Routledge and Kegan Paul. Rozental’, D. È. 1988 Russkij jazyk. 2d ed. Moscow: Vysšaja Škola. Ryle, Gilbert 1966–67/1971 “Thinking and reflecting”. In Gilbert Ryle, Collected Papers, 2: 465–79. New York: Barnes and Noble. 1968/1971 “The thinking of thoughts: What is ‘Le Penseur’ doing?” In Gilbert Ryle, Collected Papers, 2: 480–96. New York: Barnes and Noble. Savran, George W. 1988 Telling and Retelling: Quotation in Biblical Narrative. Bloomington: Indiana University. Šaxmatov, A. A. 1941 Sintaksis russkogo jazyka. E. S. Istrina (ed.). Leningrad: Ucˇpedgiz. Sayers, Dorothy L. 1927/1987 Unnatural Death. New York: Harper and Row. Schiffrin, Deborah 1981 “Tense variation in narrative”. Language 57: 45–62. Schmücker-Breloer, M. 1987 Peredacˇa prjamoj recˇi v delovyx tekstax XVII v.: Po materialam arxiva Požarskix. Istorija russkogo jazyka i lingvisticˇeskoe istocˇnikovedenie, V. V. Ivanov and A. I. Sumkina (ed.). Moscow: Nauka. Schuelke, Gertrude L. 1958 “’Slipping’ in indirect discourse”. American Speech 33: 90–98. Semino, Elena, Mick Short, and Jonathan Culpeper 1997 “Using a corpus to test a model of speech and thought presentation”. Poetics 25: 17–43. Shapiro, Marianne 1984 “How narrators report speech”. Language and Style 17: 67–78.
338
R V
Shippey, T. A. 1993 “Principles of conversation in Beowulfian speech”. In Techniques of Description: Spoken and Written Discourse, John M. Sinclair, Michael Hoey, and Gwyneth Fox (eds.), 109–26. London: Routledge. Short, Michael 1988 “Speech presentation, the novel and the press”. In The Taming of the Text: Explorations in Language, Literature and Culture, Willie Van Peer (ed.), 61–81. London: Routledge. Silverstein, Michael 1976 “Shifters, linguistic categories, and cultural description”. In Meaning in Anthropology, Keith H. Basso and Henry A. Selby (eds.), 11–55. Albuquerque, New Mexico: University of New Mexico. 1977 “Cultural prerequisites to grammatical analysis”. Georgetown University Round Table on Languages and Linguistics 1977, 139–51. 1985 “The culture of language in Chinookan narrative texts: or, on saying that… in Chinook”. In Grammar Inside and Outside the Clause: Some Approaches to Theory from the Field, Johanna Nichols and Anthony C. Woodbury (eds.), 132–71. Cambridge: Cambridge University. 1993 “Metapragmatic discourse and metapragmatic function”. In Reflexive Language: Reported Speech and Metapragmatics, John A. Lucy (ed.), 33–58. Cambridge: Cambridge University. Sinclair, John McH. 1988 “Mirror for a text”. Journal of English and Foreign Languages 1: 15–44. Slembrouck, Stef 1986 “Towards a description of speech presentation and speech reportage in newspaper language”. Studia Germanica Gandensia 9: 44–115. 1992 “The parliamentary Hansard ‘verbatim’ report: The written construction of spoken discourse”. Language and Literature 1: 101–19. Sperber, Dan, and Deirdre Wilson 1986 Relevance: Communication and Cognition. Oxford: Blackwell. Spitzer, Leo 1928/1988 “Pseudo-objective motivation in Charles-Louis Philippe”. In Leo Spitzer, Representative Essays, Alban K. Forcione, Herbert Lindenberger, and Madeline Sutherland (eds.), 41–81. Stanford, California: Stanford University. Sprincˇak, Ja. A. 1960–64 Ocˇerki russkogo istoricˇeskogo sintaksisa. 2 vv. Kiev: Radjanska Škola. Sreznevskij, I. I. 1893–1912/1989 Slovar’ drevnerusskogo jazyka. Reprint. 6 vv. Moscow: Kniga. Starovojtova, O. A. 1988 “O vydelenii granic složnogo sintakticˇeskogo celogo v drevnerusskom tekste: Na primere konstrukcii s prjamoj recˇ’ju”. Vestnik Leningradskogo Gosudarstvennogo Universiteta, 2d ser., no. 3: 70–74.
R
339
Sternberg, Meir 1982a “Point of view and the indirections of direct speech. Language and Style 15: 67–117. 1982b “Proteus in quotation-land: Mimesis and the forms of reported discourse”. Poetics Today 3: 107–56. Stola, R. 1958 “Zur Frage der Entwicklung des deepricˇastie im Altrussischen”. Slavjanskaja filologija: IV Meždunarodnyj s”ezd slavistov, 189–94. Moscow: Akademija Nauk SSSR. Sundberg, Hagar 1982 The Novgorod Kabala Books of 1614–1616: Text and Commentary [Stockholm Slavic Studies 14]. Stockholm: Almqvist and Wiksell International. Švedova, N. Ju 1952 “K voprosu ob obšcˇenarodnom i individual’nom v jazyke pisatelja”. Voprosy jazykoznanija 1952, no. 2: 104–25. Talmy, Leonard 1978 “Figure and ground in complex sentences”. In Universals of Human Language, Joseph H. Greenberg (ed.), vol. 4: Syntax, 627–49. Stanford, California: Stanford University. Tannen, Deborah 1983 “’I take out the rock — dok!’: How Greek women tell about being molested (and create involvement)”. Anthropological Linguistics 25: 359–74. 1986 “Introducing constructed dialogue in Greek and American conversational and literary narrative”. In Direct and Indirect Speech [Trends in Linguistics 31], Florian Coulmas (ed.), 311–32. Berlin: Mouton de Gruyter. 1989 Talking Voices: Repetition, Dialogue, and Imagery in Conversational Discourse [Studies in Interactional Sociolinguistics 6]. Cambridge: Cambridge University. Tarabasova, N. I. 1963 “Ob odnom frazeologizme v cˇastnoj perepiske XVII veka”. In Issledovanija po lingvisticˇeskomu istocˇnikovedeniju, S. I. Kotkov and O. A. Knjazevskaja (eds.), 144–55. Moscow: Nauka. 1964 “O nekotoryx osobennostjax jazyka delovoj pis’mennosti”. In Istocˇnikovedenie i istorija russkogo jazyka, S. I. Kotkov and V. F. Dubrovina (ed.), 157–72. Moscow: Nauka. Thompson, Geoff 1996 “Voices in the text: Discourse perspectives on language reports”. Applied Linguistics 17: 501–30. Timberlake, Alan 1975 “Hierarchies in the genitive of negation”. Slavic and East European Journal 19: 123–38. Tixomirov, M. N., and A. V. Murav’ev 1982 Russkaja paleografija. 2d ed. Moscow: Vysšaja Škola.
340
R V
Tokarev, S. A. (ed.) 1970 Materialy po istorii Jakutii XVII veka (Dokumenty jasacˇnogo sbora). Vol. 3. Moscow: Nauka. Tolstoj, L. N. 1978 Vojna i mir. Vol. 1–2. Moscow: Xudožestvennaja Literatura. Toolan, Michael J. 1990 The Stylistics of Fiction: A Literary-Linguistic Approach. London: Routledge. Townsend, David J., and Thomas G. Bever 1977 Main and subordinate clauses: A study in figure and ground. Bloomington: Indiana University Linguistics Club. Ullmann, Stephen 1957 Style in the French Novel. Cambridge: Cambridge University. Uspenskij [Uspensky], B. A. 1973 A Poetics of Composition: The Structure of the Artistic Text and Typology of a Compositional Form. Valentina Zavarin and Susan Wittig (trans.). Berkeley: University of California. 1984 “The language situation and linguistic consciousness in Muscovite Rus’: The perception of Church Slavic and Russian” [California Slavic Studies 12]. In Medieval Russian Culture, Henrik Birnbaum and Michael Flier (eds.), 365–85. Los Angeles: University of California. 1987 Istorija russkogo literaturnogo jazyka (XI — XVII vv.) [Sagners Slavistische Sammlung 12]. Munich: Otto Sagner. Valk, S. N. (ed.) 1949 Gramoty Velikogo Novgoroda i Pskova. Moscow: Akademija Nauk SSSR. van der Walle, Lieve 1993 Pragmatics and Classical Sanskrit: A Pilot Study in Linguistic Politeness [Pragmatics and Beyond, New Series 28]. Amsterdam: John Benjamins. van der Wurff, Wim 1997 “Reported discourse: A bibliography”. Unpublished. Vasmer, Max 1953–58 Russisches etymologisches Wörterbuch. 3 vv. Heidelberg: Carl Winter. Verschueren, Jef 1980 On Speech Act Verbs [Pragmatics and Beyond 4]. Amsterdam: John Benjamins. 1985 What People Say They Do with Words: Prolegomena to an Empirical Conceptual Approach to Linguistic Action [Advances in Discourse Processes 14]. Norwood, New Jersey: Ablex. 1987 “Metapragmatics and universals of linguistic action”. In Linguistic Action: Some Empirical-Conceptual Studies [Advances in Discourse Processes 23], Jef Verschueren (ed.), 125–40. Norwood, New Jersey: Ablex. Verschueren, Jef (ed.) 1987 Linguistic Action: Some Empirical-Conceptual Studies [Advances in Discourse Processes 23]. Norwood, New Jersey: Ablex.
R
341
Vinogradov, V. V., and E. S. Istrina (eds.) 1960 Grammatika russkogo jazyka. Vol. 2, pt. 2: Sintaksis. Moscow: Akademija Nauk SSSR. Volkov, S. S. 1974 Leksika russkix cˇelobitnyx XVII veka: Formuljar, tradicionnye ètiketnye i stilevye sredstva. Leningrad: Izdatel’stvo Leningradskogo Universiteta. Vološinov, V. N. 1929/1986 Marxism and the Philosophy of Language. Ladislav Matejka and I. R. Titunik (trans.). 2d ed. Cambridge, Massachusetts. 1929/1993 Marksizm i filosofija jazyka [Baxtin pod maskoj 3]. Moscow: Labirint. Walker, Anne Graffam 1986 “The verbatim record: The myth and the reality”. In Discourse and Institutional Authority: Medicine, Education, and Law [Advances in Discourse Processes 19], Sue Fisher and Alexandra Dundas Todd (eds.), 205–22. Norwood, New Jersey: Ablex. Wallace, Stephen 1982 “Figure and ground: The interrelationships of linguistic categories”. In Tense-Aspect: Between Semantics and Pragmatics [Typological Studies in Language 1], Paul J. Hopper (ed.), 201–23. Philadelphia: John Benjamins. Walton, Kendall L. 1973 “Pictures and make-believe”. The Philosophical Review 82: 283–319. Wierzbicka, Anna 1970 “Descriptions or quotations?” In Sign, Language, Culture [Janua Linguarum, Series Maior 1], A. J. Greimas et al. (eds.), 627–44. The Hague: Mouton. 1974 “The semantics of direct and indirect discourse”. Papers in Linguistics 7: 267–307. 1987 English Speech Act Verbs: A Semantic Dictionary. Sydney: Academic Press. 1988 The Semantics of Grammar. Amsterdam: John Benjamins. Worth, Dean S. 1985 “Vernacular and Slavonic in Kievan Rus’”. In The Formation of the Slavonic Literary Languages, Gerald Stone and Dean S. Worth (ed.), 233–41. Columbus, Ohio: Slavica. Xaburgaev, G. A. 1974 Staroslavjanskij jazyk. Moscow: Prosvešcˇenie. Yule, George 1993 “Vera Hayden’s dilemma, or the indirection in direct speech”. In Principles and Prediction: The Analysis of Natural Language [Current Issues in Linguistic Theory 98], Mushira Eid and Gregory Iverson (eds.), 233–42. Amsterdam: John Benjamins. 1995 “The paralinguistics of reference: Representation in reported discourse”. In Principles & Practice in Applied Linguistics: Studies in Honour of H. G. Widdowson, Guy Cook and Barbara Seidlhofer (eds.), 185–96. Oxford: Oxford University.
342
R V
Zabelin, Ivan 1848 “Sysknoe delo o ssore meževyx sudej, stol’nika Knjazja Vasil’ja Bol’šogo Romodanovskogo i dvorjanina Lariona Sumina”. Cˇtenija v Imperatorskom Obšcˇestve Istorii i Drevnostej Rossijskix pri Moskovskom Universitete 3 (7): 85–122. Zaliznjak, A. A. 1986a “Novgorodskie berestjanye gramoty s lingvisticˇeskoj tocˇki zrenija”. In Novgorodskie gramoty na bereste: Iz raskopok 1977–1983 gg., V. L. Janin and A. A. Zaliznjak (eds.), 89–219. Moscow: Nauka. 1986b “Slovoukazatel’ k berestjanym gramotam”. In Novgorodskie gramoty na bereste: Iz raskopok 1977–1983 gg., V. L. Janin and A. A. Zaliznjak (eds.), 260–306. Moscow: Nauka. 1990 “*Ogospodin””. In Voprosy kibernetiki: Jazyk logiki i logika jazyka, Vjacˇ. Vs. Ivanov (ed.), 6–25. Moscow: Akademija Nauk SSSR. 1995 Drevnenovgorodskij dialekt. Moscow: Jazyki Russkoj Kul’tury. Zimin, A. A. (ed.) 1953 Pamjatniki prava feodal’no-razdroblennoj Rusi: XII — XV vv. [Pamjatniki Russkogo Prava 2]. Moscow: Gosudarstvennoe Izdatel’stvo Juridicˇeskoj Literatury. Zipf, George Kingsley 1949 Human Behavior and the Principle of Least Effort. Cambridge, Massachusetts: Addison-Wesley. Zwicky, Arnold M. 1974 “Bibliography IV: Direct and indirect discourse”. Ohio State University Working Papers in Linguistics 17: 198–205.
Appendix: Text-kind and date of the investigated trial transcripts
In the list below, the trial transcripts used in the study are listed by source (see the References). The classifications of the text-kind and the dates are taken from the editions; I have indicated “with referral” or “with verdict” in certain cases whose classification does not correspond to the text-kind descriptions given in Chapter 2. Where the documents are not originals, the dates of the extant copies, when given in the editions, are provided in square brackets after the posited dates of the protographs. AFZX 1 103 114 117 125 129 140 157 204 248 249 254 258 259 261 308
Pravaja gramota (with referral, 1462–64 [1525–50] Pravaja gramota, end of the 15th century [1525–50] Pravaja gramota (with referral, 1498 [1525–50] Pravaja gramota (with referral, 1462–1505 [1525–50] Pravaja gramota, 1495–99 [1525–50] Pravaja gramota, end of the 15th century [1525–50] Pravaja gramota, 1499 [1525–50] Pravaja gramota, 1492 [1525–50] Pravaja gramota (with referral, 1493–94 Pravaja gramota (with referral, 1473–89 [1525–50] Pravaja gramota, ca. 1501–2 [1525–50] Pravaja gramota, ca. 1501–2 [1525–50] Pravaja gramota (with referral, 1498 [1525–50] Pravaja gramota (with referral, 1473–89 [1525–50] Sudnyj spisok, 1498–99 [1525–50]
Antonov and Baranov 1997 145 Pravaja gramota, 1500–1501 [1628] 296 Pravaja gramota, 1490–94 [1665] ASÈI 1 326 340
Sudnyj spisok, 1462–73 [mid-16th century] Dokladnoj sudnyj spisok, ca. 1464–78
344
R V 397 430 431 447 467 485 521 522 523 524 525 537 538 539 540 557 571 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 604 607 607a 615 628 635 639 640 642
Sudnyj dokladnoj spisok, ca. 1470–78 [mid-16th century] Pravaja gramota, ca. 1474–75 Sudnyj spisok (with referral, ca. 1474–75 [mid-16th century] Pravaja gramota, 1474–91 [mid-16th century] Pravaja gramota, 1478–82 Pravaja gramota, ca. 1470–1490s [1641] Sudnyj spisok (pravaja gramota?), ca. 1485–90 [mid-16th century] Pravaja gramota, 1485–90 Pravaja gramot (with referral, 1485–90 Pravaja gramot (with referral, 1485–90 Pravaja gramot (with referral, 1485–90 [mid-16th century] Sudnyj spisok (with referral, 1488–90 [mid-16th century] Sudnyj spisok (with referral, 1488–90 [mid-16th century] Sudnyj spisok (with referral, 1488–90 [mid-16th century] Sudnyj spisok (with referral, ca. 1488–90 [mid-16th century] Pravaja gramota (with referral, ca. 1490–95 Sudnyj spisok, ca. 1492–94 [mid-16th century] Pravaja gramota, ca. 1495–97 [16th century] Pravaja gramota, 1495–99 [mid-16th century] Dokladnoj sudnyj spisok, 1495–99 Sudnyj spisok (2 variants, one with referral, 1495–99 [mid-16th century] Dokladnoj sudnyj spisok, 1495–99 [mid-16th century] Dokladnoj sudnyj spisok, 1495–99 Dokladnoj sudnyj spisok, ca. 1495–99 [contemporary copy?] Dokladnoj sudnyj spisok, 1495–99 [mid-16th century] Dokladnoj sudnyj spisok, 1495–99 Dokladnoj sudnyj spisok, 1495–99 [mid-16th century] Dokladnoj sudnyj spisok, 1495–99 [mid-16th century] Dokladnoj sudnyj spisok, 1495–99 Sudnyj spisok, 1495–99 [mid-16th century] Dokladnoj sudnyj spisok (2 variants), 1495–99 [mid-16th century] Pravaja gramota (with referral, 1495–99 [mid-16th century] Dokladnoj sudnyj spisok, 1496 [mid-16th century] Dokladnoj sudnyj spisok, 1496–98 [contemporary copy] Pravaja gramota (with referral, 1496–98 Sudnyj spisok (with verdict), 1497–98 [mid-16th century] Sudnyj spisok (with referral, ca. 1499–1502 [mid-16th century] Dokladnoj sudnyj spisok, ca. 1500–1501 [mid-16th century] Sudnyj spisok (with verdict), ca. 1501–2 [mid-16th century] Sudnyj spisok (with verdict), ca. 1501–2 [mid-16th century] Dokladnoj sudnyj spisok, ca. 1502–4 [mid-16th century]
A 651 658
345
Dokladnoj sudnyj spisok, 1504 Pravaja gramota, 1505
ASÈI 2 90 Pravaja gramota, 1435–47 188 Pravaja gramota (with referral, ca. 1460s — 1470s 229 Pravaja gramota (with referral, 1475–76 [17th century] 285 Pravaja gramota (with referral, ca. 1492 286 Pravaja gramota (with referral, ca. 1492 287 Pravaja gramota (with referral, ca. 1492 288 Pravaja gramota (with referral, ca. 1492 296 Sudnyj spisok, ca. 1496–1505 [17th century] 306 Sudnyj spisok (with verdict), 1503 307 Pravaja gramota, 1503 309 Pravaja gramota, 1505 310 Pravaja gramota, 1505 [17th century] 315 Sudnyj spisok (with verdict), 1478–79 [17th century] 332 Pravaja gramota (with referral, ca. 1490–92 [end of the 16th century] 333 Sudnyj spisok (with referral, 1493 334 Pravaja gramota (with referral, ca. 1499–1500 [1502] 336 Pravaja gramota (with referral, 1502 337 Pravaja gramota (with referral, 1502 [1504] 338 Sudnyj spisok (with verdict), 1504 358 Pravaja gramota (abridged), 1450s — 1460s [17th century] 368 Pravaja gramota, ca. 1462–70 370 Pravaja gramota (with referral, 1462–78 [17th century] 374 Pravaja gramota (with referral, ca. 1463 [17th century] 375 Pravaja gramota (with referral, ca. 1463 [17th century] 381 Pravaja gramota, ca. 1465–69 [17th century] 383 Dokladnoj sudnyj spisok, ca. 1465–70 387 Pravaja gramota, ca. 1471 388 Podpisnoj dokladnoj sudnyj spisok, 1472 388a Pravaja gramota (with referral, 1472 [17th century] 400 Pravaja gramota (with referral, 1485–90 [17th century] 401 Pravaja gramota, 1485–1490s [1500] 402 Pravaja gramota, 1490 404 Pravaja gramota (with referral, 1490–98 [17th century] 405 Pravaja gramota (with referral, 1490–98 [17th century] 406 Pravaja gramota, 1490–1501 407 Pravaja gramota, 1492–1503 410 Pravaja gramota, 1494–99 [17th century]
346
R V 411 414 416 418 419 421 422 428 450 458 463 464 465 481 483 492 493 495 496
ASÈI 3 31 32 35 48 50 55 56 105 172 173 208 209 218 221 223 224 250 251 276 288
Pravaja gramota (with referral, 1494–99 [17th century] Pravaja gramota, 1497–98 Pravaja gramota, 1498 Podpisnoj sudnyj spisok (with referral, 1498–99 [17th century] Pravaja gramota (with referral, 1498–99 [17th century] Pravaja gramota, ca. 1498–99 [17th century] Pravaja gramota (with referral, 1500 Podpisnoj sudnyj spisok (with referral, 1504 Pravaja gramota (with referral, ca. 1449–50 Pravaja gramota, ca. 1456–64 [18th century] Dokladnoj sudnyj spisok, 1465–66 (tentative dating) Pravaja gramota, ca. 1465–71 Pravaja gramota, ca. 1470–85 (probably abridged) [1529] Podpisnoj dokladnoj sudnyj spisok, ca. 1484–90 Pravaja gramota (with referral, 1485–1490 [1529] Pravaja gramota (with referral, ca. 1497–98 [18th century] Pravaja gramota (with referral, ca. 1497–98 [1556] Sudnyj spisok, 1503 Pravaja gramota, 1453–62 Pravaja gramota, ca. 1416–17 Pravaja “sudnaja” gramota, 1425 Pravaja gramota, ca. 1448–52 [15th century] Pravaja gramota (with referral, 1505–6 Pravaja gramota (with referral, ca. 1499–1502 [17th century] “Pravaja i bessudnaja” gramota (with referral, ca. 1461–62 Pravaja gramota (with referral, ca. 1462–69 Pravaja gramota, 1498–99 [17th century] Podpisnoj dokladnoj sudnyj spisok, 1504 Pravaja gramota (with referral, 1504 Pravaja gramota (with referral, ca. 1495–97 Dokladnoj sudnyj spisok, ca. 1495–97 Dokladnoj sudnyj spisok, 1505 Pravaja gramota, 1501 [18th century] Pravaja gramota, 1504–5 Pravaja gramota, 1504–5 Podpisnoj dokladnoj sudnyj spisok, 1501 [1576] Podpisnoj dokladnoj sudnyj spisok, 1501 [1576] Pravaja gramota (with referral, 1499–1500 Pravaja gramota, ca. 1490–1509
A 319 357 364 477 478
Pravaja gramota (Rjazanian; paraphrased), 1464–71 [unspecified date] Pravaja gramota (Rjazanian), 1483–1500 [17th century] Pravaja gramota (Rjazanian), 1464–82 [18th century] Pravaja gramota (with referral, 1478–85 Sudnyj spisok, 1503–4 [late 17th century]
Gorcˇakov 1871 5 [Pravaja gramota], 1504 [16th century] GVNP 92 340
347
Pravaja gramota (Novgorodian), 1st quarter of the 15th century Pravaja gramota (Pskovian), 1483 [16th century]
Kaštanov 1970 6 Sudnyj spisok (with verdict), 1438–40 or 1440–41 [1530s] 9 Pravaja gramota (with referral, 1442–62 16 Sudnyj spisok (with verdict), 1454–56 [1530s] 27 Sudnyj spisok (with verdict), ca. 1470 [1530s] 40 Pravaja gramota, 1503 [1530s] Koreckij 1969 2 Pravaja gramota (Novgorodian), 1501–5 (or 1510) [17th century] Krotov and Smetanina 1987 5 1457–62 [18th century] Morozov 1988 2 Pravaja gramota (Rjazanian), 1483–1500 [18th century]
Name index
A Aaron, Uche E. 14, 49 Allingham, Margery 55 Andersen, Henning 16 Anpilogov, G. I. 307, 318 Antonov, A. V. 152, 184, 213, 309, 310, 311, 312, 315, 317 Aristotle 65 Arnovick, Leslie K. 303 Austen, Jane 135 Austin, J. L. 3, 249, 277 Avanesov, R. I. 28 Avvakum, Protopope 56 B Bakhtin, M. M. xiv, xv, 1, 2, 4, 10, 11, 19, 20, 279, 301, 302, 303, 307, 311; see also Vološinov, V. V. Bally, C. 135, 310 Bamgbose, » Ayo 310 Banfield, Ann 50, 133, 135, 304, 310 Baranov, K. V. 152, 184, 213, 309, 310, 311, 312, 315, 317 Barnet, Vladimír 76, 123, 308 Battistella, Edwin L. 24 Bauman, Richard 306 Baynham, Mike 52, 54, 69, 303, 306, 308 Becker, A. P. 299, 303, 318 Bever, Thomas G. 107, 222, 227 Birnbaum, Henrik 18, 203 Blackburn, Simon 303
Bloomfield, Leonard 23 Bolinger, Dwight 259 Bondar’, I. P. 76, 301 Borkovskij, V. I. 304, 308, 310 Brandsma, Frank 44 Brecht, Richard D. 182 Brown, Gillian 102, 155, 253, 264, 317 Brown, Penelope 3, 20, 25, 38, 69, 86, 287, 306, 307 Brown, Roger 303 Bulaxovskij, L. A. 136, 304, 311 C Caldas-Coulthard, Carmen Rosa 52, 70 Cassirer, Ernst 65 Cˇerepnin, L. V. 46, 304, 305, 306 Chafe, Wallace 51, 57, 59, 67, 68, 69, 71, 208, 303, 306, 307, 308, 309 Chatman, Seymour 68, 204 Chvany, Catherine V. 8, 43, 45, 91, 107, 109, 123, 211, 308, 318 Clark, Herbert H. 13, 52, 53, 55, 67, 69, 71, 102, 199, 203, 209, 212, 306, 307, 308, 310, 311 Clark, Katerina 303 Collins, Daniel E. xv, 25, 50, 60, 103, 301, 303, 304, 305, 309, 314 Collins, James 1, 10, 23, 303 Comrie, Bernard 160, 306 Coulmas, Florian 20, 50, 51, 52, 145, 203, 303, 304, 306, 307 Couper-Kuhlen, Elizabeth 12, 169, 307
350
R V
Cram, D. F. 303, 306 Culpeper, Jonathan 13, 60, 252, 304 Cˇumakov, G. M. 112, 306, 308, 310 D Derjagin, V. Ja. 25, 318 Dewey, H. W. 304, 305, 306 Dirven, Rene 10, 65, 247, 303 Diver, William 15, 104, 122, 218 D’jakonov, M. A. 310 Dmitriev, L. A. 6 Doležel, Lubomír 135, 142, 305, 310, 311 Dömötör, A. 12 Du Bois, John W. 71 Dubois, Betty Lou 306 E Ebert, Karen 25, 50, 227, 243, 305 Eco, Umberto 92 Ehlich, Konrad 304 Enkvist, Nils Erik 303 F Filippova, I. S. 46, 308, 313 Fleischman, Suzanne xvi, 29 Fletcher, Giles 290 Fludernik, Monika 2, 5, 35, 48, 52, 53, 54, 55, 58, 59, 69, 303, 306, 307, 308, 310, 311 Fónagy, Ivan 5, 12, 67, 135, 145, 237, 303, 304, 306 , 307, 310 Forsyth, J. 160 Fox, Barbara 308 Frajzingier, Zygmunt 304 Fritz, Gerd 303 G Geertz, Clifford 3 Gennadij, Archbishop of Novgorod 7 Gerrig, Richard 13, 52, 53, 55, 67, 69, 71, 102, 199, 209, 212, 306, 307, 310, 311 Gilman, Albert 303
Givón, Talmy 19, 77, 116, 304, 308 Glock, Naomi 308 Goffman, Erving 5, 22, 42, 62, 70, 71, 79, 114, 168, 203, 211, 227, 305, 307, 308, 311, 318 Goody, Jack 203, 307 Goosens, Louis 10, 304 Gorcˇakov, M. 273, 309, 310, 311, 312, 313, 314, 317 Green, Georgia M. 7, 43, 66 Grekov, B. D. 113, 139, 279, 305 Grice, Paul 51, 226, 307 Grimes, Joseph E. 59, 130, 155, 204, 264 Gvozdanovic´å, Jadranka 46 Gvozdev, A. N. 306 H Haberland, Harmut 51, 53, 67, 135, 145, 304, 306, 307, 311 Hagenaar, Elly 11 Haiman, John 83, 95, 112, 116 Halliday, M. A. K. 155, 307, 317 Hammer, A. E. 16 Hanks, William F. 25, 306, 315 Hasan, Ruqaiya 155, 307, 317 Havránek, Bohuslav 25 Hermon, Gabriella 7, 43 Hickmann, Maya 5, 304, 311 Holquist, Michael 303 Holt, Elizabeth 71, 304 Hopper, Paul J. 22, 77, 78, 303, 308 Huizinga, J. 308 I Ickler, Nancy 37 Issatschenko, A. V. 287 Istrina, E. S. 15, 304, 306, 309, 310 Ivan Ivanovicˇ, Grand Prince 197 Ivan III Vasil’evicˇ, Grand Prince 197 Ivan IV Vasil’evicˇ, Tsar’ 191 Ivanov, Vjacˇ. Vs. 303, 304 Ivina, L. I. 308, 314
N I J Jacobs, Andreas 17, 18, 303 Jakobson, Roman 76, 303, 318 Jakovlev, A. 29, 156, 310, 316 Janssen, Theo A. J. M. 303, 304 Jespersen, Otto 16, 143, 149, 308 Johnson-Laird, Philip N. 68, 71 Johnstone, Barbara 303, 309, 318 Jucker, Andreas H. 17, 18, 303 Juškov, A. I. 318 K Kaiser, Daniel H. 113, 139, 192, 207, 219, 222, 271, 276 Karpov, G. O. 65 Kaštanov, S. M. 30, 38, 91, 196, 206, 208, 211, 213, 219, 232, 243, 304, 305, 307, 309, 311, 312, 313, 314, 315 Kay, Paul 294 Kazakova, N. A. 310 Keller, Rudi 16, 303, 304 Kieckers, E. 265, 317 Kiparsky, Paul and Carol 306 Kittay, Jeffrey 30, 179 Kleimola, Ann M. 29, 30, 78, 99, 112, 113, 151, 166, 174, 177, 191, 192, 195, 198, 210, 271, 276, 279, 304, 305, 306, 310, 316 Klenin, Emily 311 Klewitz, Gabriele 12, 169 Koduxov,V. I. 46, 304, 305, 306, 310, 311 Koreckij, V. I. 46, 192, 234, 30, 312, 315 Kotkov, S. I. 46, 306, 308, 310, 313 Kozlovskij, Pavel 311 Krotov, M. G. 137, 139, 140, 312, 315, 316 Kuno, Susumu 260 Kurihara, Takehiko 43 Kuz’mina, I. B. 308 Kuznecov, P. S. 308, 310
351
L La Fontaine, Jean de 135 Labov, William 72, 130, 132, 155, 204, 267 Lakoff, Robin Tolmach 6, 54, 68, 72, 132, 135, 211, 243, 289, 291, 307 Lange, Deborah 10, 67, 69, 303, 305, 306, 307, 309 Larin, B. A. 306, 308, 318 Larson, Mildred L. 163, 308 Lebsanft, Franz 303 Lee, Benjamin 135 Leech, Geoffrey N. 4, 51, 70, 102, 103, 124, 132, 133, 135, 141, 195, 265, 304, 306, 308, 310 Lennard, John 306 Leont’ev, A. K. 27, 304, 305 Lerch, E. 135 Levinson, Stephen C. 3, 20, 25, 38, 69, 86, 287, 306, 307 Lewis, David 20, 21, 42, 50, 287, 304 Li, Charles 116, 303, 305, 306, 308 Liberzon, I. Z. 308 Lips, Marguerite 135 Lixacˇev, D. S. 6, 65 Lixacˇev, N. P. 58, 246, 308, 309, 310, 314 Longacre, Robert E. 27, 42, 44, 70, 92, 102, 103, 114, 172, 173, 209, 211, 303, 306 Lopatina, L. E. 76, 85, 266, 304, 306, 318 Lorck, E. 310 Lucy, John A. 303, 318 Ludolf, Heinrich Wilhelm 85 Lyons, John 17, 307 M Machek, Václav 304 Maksim Grek 85 Martysevicˇ, I. D. 113 Massamba, David P. B. 306 Mathis, Terrie 92, 95, 97, 307
352
R V
Maxlin, V. L. 303 Mayes, Patricia 55, 67, 69, 303, 306, 307, 308 McHale, Brian 149, 204, 304, 310, 311 McKay, Janet H. 310 Miller, George A. 68, 71 Mithun, Marianne 59 Molotkov, A. I. 8, 288, 301, 306, 311 Morawski, Stefan 53, 71, 203 Morgan, Jerry L. 66 Morozov, B. N. 128, 147, 311, 316 Munro, Pamela 304 Murav’ev, A. V. 57 N Nemcˇenko, E. V. 308 Neubert, A. 135 Nichols, Johanna 96, 122–23, 124, 317 Nikiforov, S. D. 76, 309, 310 Noonan, Michael 259 Novosel’skij, A. A. 314 O Ong, Walter J. 46, 56, 307 Osipov, B. I. 305 Otin, E. S. 59–60, 306 P Paducˇeva, E. V. 11, 134, 136, 141, 142, 149, 310 Page, Norman 6, 35, 46, 67, 102, 107, 122, 124, 134, 142, 195, 201, 228, 249, 265, 267, 304, 306, 308, 309, 310 Palmer, F. R. 221, 305 Panteleimon, Arximandrit 56 Partee, Barbara Hall 7, 14, 304, 306 Pascal, Roy 70, 134, 135, 310 Pearce-Higgins, Lucinda 14 P’ecux, Vjacˇeslav 5–6 Peškovskij, A. M. 304, 306 Philips, Susan 10, 58, 203, 212, 303, 308
Photios, Metropolitan 30 Plato 68, 74 Pokrovskij, N. N. 85, 145, 304, 305, 310 Polanyi, Livia 311, 318 Popova, Z. D. 306 Potebnja, A. A. 116, 199 Preobraženskaja, M. N. 305 Prince, Gerald 135, 138, 303 Q Quine, Willard Van Orman 306 R Rader, Margaret 291, 294 Reddy, Michael J. 64 Remnick, David 13 Rimmon-Kenan, Shlomith 43 Rinberg, V. L. 288, 301, 305, 306 Romaine, Suzanne 10, 67, 69, 303, 305, 306, 307, 309 Roncador, Manfred von 53, 304, 305, 306 Rosen, Harold 67 Rozental’, D. È. 306 Ryle, Gilbert 303 S Savran, George W. 56 Šaxmatov, A. A. 304 Sayers, Dorothy L. 2 Schallert, Joe 122–23, 124 Schiffrin, Deborah 72, 208, 211, 308, 309 Schmücker-Breloer, M. 306 Schuelke, Gertrude L. 13, 304, 308, 309 Scott, Sir Walter 135 Semino, Elena 13, 60, 252, 304 Shapiro, Marianne 195, 265, 267 Shippey, T. A. 303 Short, Michael H. 4, 13, 51, 52, 54, 60, 67, 70, 102, 103, 124, 132, 133,
N I 135, 141, 195, 252, 265, 304, 306, 308, 310, 311 Silverstein, Michael 1, 5, 11, 15, 22, 23–25, 114–15, 303, 309 Sinclair, John 5, 42, 71, 307 Slembrouck, Stef 54, 58, 70, 133, 134, 307, 310, 311 Smetanina, S. I. 137, 139, 140, 312, 315, 316 Sperber, Dan 4, 54, 307 Spiridon of Trimithus, Bishop 56 Spitzer, Leo 141, 310 Sprincˇak, Ja. A. 306, 311 Sreznevskij, I. I. 62, 90, 159, 261, 272, 304 Starovojtova, O. A. 305 Sternberg, Meir 2, 50, 51, 65, 69, 70, 74, 85, 203, 209, 279, 297, 306, 307 Stola, R. 308 Sundberg, Hagar 304 Švedova, N. Ju 311 T Talmy, Leonard 222, 253 Tannen, Deborah 5, 13, 51, 64, 68, 85, 101, 102, 146, 155, 267, 279, 303, 304, 306, 307, 308, 309, 317, 318 Tarabasova, N. I. 316, 318 Thibaudet, A. 142 Thompson, Geoff 4, 60, 71, 116, 125, 134, 135, 149, 150, 199, 230, 243, 303, 305, 308, 310, 311, 318 Thompson, Sandra A. 77, 78, 112, 116, 308 Timberlake, Alan 311 Tixomirov, M. N. 57 Tokarev, S. A. 6 Tolstoj, L. N. 14 Toolan, Michael J. 310 Townsend, David J. 107, 222, 227 Turgenev, I. S. 13
353
U Ullmann, Stephen 135, 142, 311 Uspenskij, B. A. 56, 70, 136, 309, 310 V van der Walle, Lieve 303 van der Wurff, Wim 303, 304 Vasilij III Ivanovicˇ, Grand Prince 28, 191 Vasmer, Max 304, 307 Verschueren, Jef 274, 303, 304, 310 Vinogradov, V. V. 15, 304, 306, 309, 310 Volkov, S. S. 306 Vološinov, V. N. xiv, xv, 3, 10, 11, 20, 35, 67, 134, 265, 302, 303, 310, 311; see also Bakhtin, M. M. W Walker, Anne G. 57, 58, 307 Wallace, Stephen 117, 211, 222, 227 Walton, Kendall L. 67 Wierzbicka, Anna 10, 15, 67, 121, 156, 161, 249, 303, 304, 306, 310 Wilson, Deirdre 4, 54, 307 Worth, Dean S. 35 X Xaburgaev, G. A. 306 Y Yule, George 14, 92, 95, 97, 102, 155, 253, 264, 303, 304, 307, 317 Z Zabelin, Ivan 310 Zaliznjak, A. A. 14, 205, 306 Zimin, A. A. 113 Zipf, George K. 54, 308 Zwicky, Arnold M. 306
Subject index
A a 76, 78, 98–99, 101, 186, 192, 229, 246 a r(’)kucˇi 76, 81 Abridgments 63, 64, 116, 183, 197, 200; see also Trial transcripts, abridged Accelerando see Pacing Accusations 125, 141, 208, 241 Accusative after verbs of asking 311 double 213 in fused reported speech 117, 118, 122, 124, 233, 259 syncretism with genitive 311 syncretism with nominative 117 with infinitives 9 with instrumental 213 with participles 9, 122–23 acquit 249 Actualization see Foregrounding Address-forms see Appellatives; Vocatives Addressees appellatives for 168, 169, 171 genitive of 157; omitted, 162–64 dative of 41–42, 90, 165, 221 in prepositional phrase 77 thematization of 37 Adequation to type see Typicality Adjacency pairs 27, 172
Adjournment procedures 33, 81–82, 106, 119, 125–26, 167, 198–99, 200, 269 Adjudications 246, 247–52, 274, 294–95 Adjuncts as elaborations 40 as tags 6, 15, 37, 43, 62, 133, 279, 289 Adverbials 134 Allusive quotations 4, 134 Amicable-settlement records 28, 37, 145–46, 190, 293, 305 Amplifications 3, 81, 141, 191 requests for 212–13, 214, 225, 228, 234, 239, 240, 315 Anaphora 39–40, 59, 60, 91, 109, 156 zero see Subjects, null Anterior utterances mediated by mental images 2, 3 and verbatimness 49, 51–58, 64–65, 67, 68, 70, 203 changed in requotations 203–4, 220, 232, 254, 255, 260, 263, 266–67, 273, 276, 277, 279 Aorist tense 45–46, 245, 311 Appellatives 47, 86, 134, 137, 154, 168, 169, 171, 245, 263; see also Vocatives Approximatives 59 ask 156, 161
356
R V
Asking, verbs of see Inquiring, verbs of Aspect, choice of 6, 117, 206, 215, 220–21, 226; see also Imperfective; Perfective Asyndeton 178 Attitudinal verbs 6 Attribution see Tag formula Author-based styles see Speaker-based strategies Authorization formula 30, 31 Automatized uses 3, 5, 21, 24, 75, 303 Autonomous language 294, 295 Averrals 4, 5, 42, 43, 71, 247, 288, 293 aže (až’) 62 B Background information see Commentary; Orientations Backgrounding 263, 283 amplifications 81 of attribution 77 commentary 78, 97, 119, 236, 293 corroborations 107, 112, 116, 293 free direct speech 293 free indirect speech 139–40, 141, 143, 148, 149–50, 287, 311 fused reported speech 119, 120, 122, 123, 234, 235, 236, 293, 296 gerunds 78, 82–83, 84, 293, 315 imperfectives 117, 228, 314 indirect/nondirect speech 107, 112, 113, 114, 212, 227, 228, 287, 296, 315 narrative reports of speech acts 126, 212 nominalizations 109 null subjects 8 participial clauses 77–78, 84, 123 preposed elements 43, 77–78 presupposed elements 43, 113, 116, 122, 195, 226, 234, 296 reports of less immediate interest 138–40, 141, 287
subordination 186, 227, 295, 315 summaries 140, 186 Backshifting 11, 103, 131, 134 Bailiffs, reported speech of 194–96 Bessudnye pravye gramoty see Defaultjudgment charters Binary oppositions 24 Birchbark letters 13–14 biti/bivati cˇel”m’ 82–83, 226, 242, 243, 274, 281, 301, 316 Bloodwite 206–7 Boundary-setting descriptions 22, 97, 235 with addressee indicated 42 compact strategies in 292–93 commands initiating 167, 172–73 direct speech in 79, 81, 144, 146, 309 free direct speech in 79, 84, 94–96, 103, 144, 163, 293 free indirect/nondirect speech in 143–46, 149, 277, 293, 311 fused reported speech in 79, 117, 119–20, 122, 293, 310 gerund tags in 78–81, 84 nondirect speech in 106 s(”)kazati in 88 brate 86–87, 154–55, 309 brat’e 86–87, 154–55, 309 Byplay 155, 311 C Cadastre books 184, 186, 192, 278–79, 313 Capital letters 46, 153–54, 178, 186, 245, 311, 316 Cataphora 37, 59, 77, 78, 155, 210, 231, 307 Causal constructions 12, 141, 252–53, 256, 263, 274, 295, 317 cˇelobitnye gramoty see Petitions ceˇlovati kr’st” 225, 315; see also Crosskissing oaths
S I Challenges 42, 58, 104, 188, 193, 214–15, 216 Chancery writing see Legaladministrative writing Choral speech 22, 52, 99, 101–102, 152, 155, 180, 182, 183, 186, 191, 218, 293, 309, 311 Chronicles xv, 37, 305 Church Slavonic see Ecclesiastical register Citation particles see Particles, quotative Citations of documents see Quotations of written records Classroom discourse 54 Closing formula 173–74, 191; see also End protocols Cognition, verbs of 49, 117, 276–77 Coherence deictic shifts 46, 107, 114, 150, 155, 205, 227, 263, 264, 296 direct speech 47, 48, 69, 263–65, 290, 296 free direct speech 103 promoted by frames 44, 46, 115, 152–55 promoted by judges’ reports 152–55 promoted by repetitions 44, 48, 59, 256 role of cataphora 59 role of intercalated elements 47–48, 264–65 and slipping 264–65 role of tag formula 15, 44, 62, 68, 290 role of vocatives 47–48 Cohesion and gear-shifting 46, 138, 263, 264 free direct speech 93–94, 95, 98, 101, 102 promoted by intercalated elements 105, 138, 238–40, 263, 264–66, 267, 270, 297
357
promoted by repetition 47, 155, 267, 297, 317 and second position 48 and tak(o) 59 in dyadic structure 93–94, 98, 155 and thematic continuity 112, 231, 233, 246, 257–58, 295 and complementizers 257–59, 295 Collateral information, reported speech as 81, 83, 120, 155, 218, 235–36, 315 Colons, marking reported speech 13, 46 Commands framing function of 172 in layered reported speech 204, 241, 242–43 of judges 33, 154, 159, 160–61, 165, 172, 179, 199, 247, 250–52, 316 of litigants 40, 125 procedural 125, 127, 149, 165, 172–75, 194, 198, 251, 292 Commas, double 46 Commentary, reported speech as 22, 43, 78, 79, 81, 88, 97, 103, 119, 121, 277, 293 Commissive speech acts 174; see also Promises Compact strategies 288–90, 297, 318 in boundary-setting descriptions 292–93 in corroborations 293–294 counteracting foregrounding 288, 293, 318 free indirect speech as 150, 289, 295 fused reported speech as 288–89, 293 and givenness 131, 197, 240, 253, 254, 258, 293–94, 295 in judicial-referral records 201, 293–94 in layered reports 243 Narrative reports of speech acts as 132, 280, 283, 289, 294 and pacing 201, 293
358
R V
in trial dossiers 283, 297 in verdicts 254, 261, 273, 295 Comparators 59–60, 155 Competence generic 4, 66 pragmatic 3, 4, 18, 66, 21, 285 Complaints (žalobnicy) see Petitions Complement clauses 8, 11, 49, 62, 101, 123, 189, 227, 239, 240, 241, 259, 304, 316; see also cˇ’to; koe Complement type 15, 21 Complementized direct speech see Direct speech, complementized Complementized indirect speech see Indirect speech Complementizers see also aže (až’); cˇ’to; koe; ože (ož’); that and backgrounding of report 43, 107, 184–86, 217, 227, 230 with direct speech 11, 67, 103–4, 136, 199–200, 264, 289 explicative function of 11, 199, 258–59 multiple embeddings as disincentive for 224, 233, 256, 259, 286 not criterial for indirect speech 103–4 omission of see Indirect speech, uncomplementized and presupposition 107, 184, 197, 217, 227, 295 Complexity semantic 201, 224, 256 syntactic 298; see also Embeddings, multiple favoring direct speech 201, 262 and complementizers 218, 232–33 and intercalated elements 240, 241 and slipping 252, 294 Complicating action 204, 218 Condensations 3, 52, 127, 148, 201, 214, 252, 253, 270, 279, 294; see
also Abridgments; Summaries; Syntheses Conduit metaphor 64–65, 70 Conjunctions coordinating see a; i explicative see Complementizers Constructed dialogue 51, 53, 57, 101; see also Direct speech Control of interpretation 65, 70, 211–12, 243, 295, 297; see also Coordination of interpretation; Deference; Interference, authorial/narratorial Conventions automatization of 20–21 and coordination of interpretation 20, 50, 75, 286, 290, 298 and institutionalized intentions 21, 75, 297 intentional character of 26, 42, 287 and pragmatic competence 21 as proven efficient means 20–21 purposive nature of 20, 24, 42, 286, 287, 297, 298 rational basis for 20, 21, 26, 50 reading 4–5, 18, 54, 74, 135, 136, 299 convict 249 Cooperation, conversational 2, 38 65–66, 70, 73 in trial procedure 36, 145, 172–73, 292 for trial scribes 42–43, 73, 285, 289, 291, 297–98 and verbatim reporting 51–54, 58, violations of 91–92, 101, 113, 114, 127–28, 173, 192–93, 209, 273 Coordinated task activity 79 Coordination of interpretation 6, 10, 20, 42, 65, 67, 286, 289; see also Control of interpretation Copula 117, 233, 261, 296 Correlatives 59–60, 183, 232
S I Corroborations; see also Verification procedure in direct speech 191, 200 favoring indirect/nondirect speech 106–7, 109, 110, 112–13, 116, 188–89, 192, 196, 201, 293, 296 in judicial-referral records 179, 183–84, 188–89, 293–94 as narrative reports of speech acts 141 as prototypical speech act of witnesses 87, 112–13, 222 in trial dossiers 156, 283, 297 in wills 189–90 Court reporting 50, 57 Criminal trials 27, 136–42, 210, 219, 287 Cross-kissing oaths 33, 112, 128, 225, 259, 315 Crossplay 168, 171 cˇ’to 12, 103, 114, 192, 199, 212, 225, 240, 252, 254, 266 Cursive 57 Cyril Belozersk Monastery 145 Czech 310 D Danish 135 Dannye gramoty see Donation charters Dashes, marking reported speech 46 Data problem in Historical Pragmatics xvi, 16–17, 285 Dati s”rok” 175, 313 Dative of addressee 41–42, 90, 165, 221 of experiencer 181 of recipient of command 172–75, 251–52 of relation 181 De dicto/de re 182, 306 Decentering 25, 315 Declarations 121, 161–62, 165, 226, 315
359
Default-judgment charters 37, 238, 305, 308 Default judgments 198, 269 Default strategies 24, 50, 60, 75, 287 Deference in linguistic politeness 47, 69, 83 to interpreters 68, 69, 71, 72, 297 deˇ(i) 12, 76, 137, 156, 239, 263–66, 269–70, 281, 316 Deictic adverbs 59, 35, 134, 317; see also tak(o) Deictic pivot 11, 21 Deixis endophoric 182 exophoric 181–82, 197 shifts of 46, 67, 107, 114, 150, 155, 205, 211, 218, 227, 263, 264, 296 spatio-temporal 35, 134, 317 deˇ(ja)ti 12, 263, 304 Demonstrations 67, 68, 69, 95, 167, 307 Demonstratives 59–64, 181–82, 183, 197, 231 Denials 125, 141, 180, 220, 317 Denunciations 139, 141, 241 Departicularizers 53, 307 Depositions (skazki) 152, 207, 217, 218 Dialogue, complex 173 Dialogue discourse 27, 44, 114, 290, 291, 305 Diffuse strategies 73, 124, 200, 224, 250, 260, 280, 283, 288–90, 294, 295, 297, 318 Diplomatic dossiers 65 Diplomatics 25 Direct experience 69, 102, 212 Direct speech ₍₎ 35–103, 305 appropriateness repairs in 52 “basic” status of 49 in boundary-setting descriptions 79, 81, 144, 146, 309 in choral speech 52, 99, 293 with co-constructed reports 52, 307
360
R V
and coherence 47, 48, 69, 263–65, 290, 296 as collaboration 69–70 in commentary 81 complementized 11, 67, 103–4, 136, 199–200, 264, 289 in Contemporary Standard Russian 5, 14–15 and continuity of theme 96 creative character of 2, 57, 65 and direct experience 69, 102, 212 and disassociation of responsibility 71, 195, 208–9 decentered nature of 25, 107, 227, 315 deixis in 14–15, 35, 66–67, 68, 74, 95, 107, 114, 132, 134, 150, 196, 211, 218, 227, 262, 263, 288, 295, 305 as demonstration 53, 67, 69, 307 diffuse character of 122, 124, 129, 131, 132, 150, 172, 186, 201, 208, 211, 218, 224, 261, 262, 280, 281, 283, 288–90, 294, 295 and distancing 71, 212 in dyads 191, 205–7, 290, 296 emotive elements in 35, 67, 134 and empathy 212 and evaluation 68, 69, 71, 212, 224 in falsifications 192, 294 with fictitious statements 52–53, 67, 70 and foregrounding 25, 43, 68, 76, 78, 84, 112, 113, 129, 150, 156, 195, 196, 205, 208, 210, 212, 224, 227, 228, 265, 296; compared with free indirect speech 138–41, 149–50; compared with fused speech 121, 122, 123, 234, 235; compared with indirect speech 107, 114, 115, 184, 185–86, 211, 218–19, 222, 226, 293; compared with narrative reports of speech acts 126;
compared with nondirect speech 104 and gear-shifting 96, 107, 119, 124, 201, 227 and givenness 186, 191, 200–201, 226, 261 with habitual statements 52 as hearer-based strategy 67–72, 212, 290 and historical present 91, 208 with hypothetical speech 52 illocutionary force 53, 68, 72 and immediacy 66, 297 contrasted with indirect/nondirect speech 11, 103–4, 107, 112, 115, 208, 226, 227, 299–300, 304 introduced by alternative verba dicendi 85–92, 208–11, 314 introduced by gerunds 76–85, 94, 281, 293 introduced by speech-act verbs 5, 83, 85, 86, 161, 205, 301 and involvement 66, 68–69 in judges’ questions 156, 165, 172, 180–81, 183, 184, 194 in judge’s report 198–201, 294 in layered reports 205–12, 223–24, 296 and main-clause syntax 67 in narrative frames 205, 211, 224, 263, 296 at narrative peaks 68, 92, 102, 205, 211 contrasted with narrative reports of speech acts 126–29, 131, 172–73, 194, 208, 210, 212, 280, 292, 294 nonverbatim uses of 2, 51, 52–58, 306, 307 and objectivity 71–72 in Old Russian 9, 11, 35, 49, 161, 306 ontogenetic priority of 49 opacity of 306
S I open-endedness of 15 and pacing 107, 114, 124, 126, 150, 211, 218–19, 227 in paraphrases 56, 58, 62–64 and politeness 69 pronouns in 102, 112, 154, 181, 191, 195, 196–97, 221, 224 prosodic features in 12 with prototypical testimony 36, 191, 292 punctuation with 46, 305 as re-enactment 67, 72–74, 290 and relevance 109, 116, 121, 123, 124, 129, 138–39, 184, 201, 212, 226, 281, 292 with second-hand reports 52 selectiveness of xv, 53, 67 and self-evident reports 71 “simplicity” of 25, 49 and slipping 13–14, 29, 67, 108, 109, 121, 137–39, 195, 199–200, 250, 263–65, 294 in speech summaries 52, 62 in standard reporting strategy 35, 36, 42, 44, 79, 86, 89, 92, 103, 129, 137, 165, 190, 192, 205, 261–62, 290, 296 as suppression of evaluation 43, 69, 70–72 theatricality of 67, 72 in translations 52 in trial dossiers 115–16, 283 as universal 9, 49, 303 untagged see Free direct speech and verbatimness 2, 48, 49–58, 60, 64–65, 67, 107, 265, 306, 307; see also Reporter-commitment model of direct speech; Reproductionist approach to direct speech in verdicts 248, 261–65, 280, 295, 313 in verifications 181–82, 190–92
361
and vividness 66, 68, 72, 91, 92, 102, 297 vocatives in 47–49, 105, 154, 205, 306 in wills 190 Directives see Commands Disassociation of responsibility 71, 195, 209 Disclaimers 195, 206, 230, 240 Distance, social 47, 69 Distancing in reported speech 71, 133, 212, 240, 241 in storytelling 71 Document citations see Quotations of written records doklad” 177, 198 Dokladnye delovye gramoty 190 Dokladnye sudnye spiski see Referredjudgment charters Donation charters 63–64, 188 Drafts 58 Drama 68, 92, 102; see also Involvement; Vividness Dramatic texts 163 Drift stylistic 86, 115–16, 191–92, 293 syntactic 16, 300–301 Duels, judicial 33, 113, 121, 271–72 Dvina 46, 308, 311 Dyadic structure 19, 27, 33, 42, 44, 93, 98, 129, 153–65 departures from 283, 294, 296; and indication of addressee 41–42; and gerund tags 78, 79; and alternative verba dicendi 87–88, 187, 217; and indirect speech 109–10, 292; and fused reports 117, 124, 234, 292; and narrative reports of speech acts 126, 128, 172–74; and free indirect speech 148–49; self-quotations as 210
362
R V
framing function of 155–56, 289, 290, 291 indirect speech in 108 in judicial-referral record 180, 191, 293 in layered reported speech 205–6, 209, 210, 211, 228, 296, 314 in other text-kinds 145, 190 contrasted with verdict 245, 247, 254, 271, 295 E Eastern Orthodoxy 56 Ecclesiastical register (Church Slavonic) 8, 9, 37, 41, 45, 56, 72, 85, 122, 138, 287, 307 Economy and choral speech 101, 182 and fused reported speech 122, 124 and indirect speech 112 and narrative reports of speech acts 127, 132, 140 and omitted speech 61 and presupposition 293 speaker’s vs. auditor’s 308 Edicts (ukaznye gramoty, ukazy) 263, 280 Elaborations see Amplifications Ellipsis 8, 38, 60, 102, 112, 162–65, 168, 181, 189, 236–37, 246, 256, 258, 317 Embedded quotations see Quotations, incorporated Embeddings, multiple 128, 200, 218, 224, 233, 239, 251, 253, 256, 259, 295 Emic interpretations 3, 18, 233, 285, 303 Emotive elements in direct speech 35, 67 in free indirect speech 134, 149 in indirect speech 67, 233 m”lviti 91, 209–11, 224
screened out by scribes 48, 58 Empathy 212 Enclitics 48, 200 End protocols 19, 30, 34, 64, 152, 282, 293; see also Closing formula Endophoric deixis 182 English 5, 7–8, 10, 14, 43, 55, 117, 121, 131, 135, 249, 252, 259, 263, 304, 306, 307, 309 Epexegetical constructions 199, 215 Episode boundaries 41, 43–44, 81, 88–89, 128, 146, 159, 191 Episodes, subdivision of 96, 139, 168, 208, 229, 264 Essoin, report of 142 Evaluation 3, 162, 204, 208–11, 252–53, 256, 279–80, 294 causal information as 141, 252–53 and choice of tag 6, 43, 72, 91–92, 209–10, 212, 213, 224, 259, 296 and direct speech 68, 69, 71, 212, 224 and free indirect speech 134, 136, 141, 145, 277–78, 279, 293 and historical present 91, 208, 224 hypotaxis as 72 narrative reports of speech acts as 132, 213, 241, 243, 270, 273, 274, 278, 294 negation as 130 repetitions as 267 and summaries 45, 95 suppression of 43, 71–72, 125, 130, 132, 136, 141, 145, 162, 204, 252, 290, 291 Explanations 83, 123, 236; see also Commentary; Excuses Excuses 81, 82, 110, 143, 235–36, 315 Existential constructions 119, 232, 233 Existential verb 117 Exophoric deixis 181–82, 197 Expressive elements see Emotive elements
S I F Fade-ins and fade-outs see Quotations, incorporated; Slipping Faithfulness see Verbatimness purported maxim of 51, 53, 54, 55 Falsifications 28–29, 58, 178–79, 192–97, 241, 272–73, 294, 314; see also l”gati/s”l”gati Felicity conditions 51, 52 Fiction, reported speech in 67, 135, 204, 286, 287 Fictive reports 1, 52–53, 67, 70, 204 Fidelity of transmission see Verbatimness First-person narrative 23, 204, 283 Flashbacks 120 Focalization 35, 62, 141 Footing, changes in 22, 41–42, 47, 79, 85, 110, 113, 114, 167, 179, 184, 186, 211, 247, 292 Foregrounding 7, 23, 24–25, 105–6, 180, 253, 288 and adjuncts 43 of attributions 43, 77, 92, 107, 114, 180, 184–85, 222, 228, 253, 288, 289, 293 and cataphora 210 and compact strategies 288, 293, 318 and comparators 155 and complement clauses 114, 184, 227, 293 and diffuse strategies 288, 289, 318 and foregrounding 25, 43, 68, 76, 78, 84, 112, 113, 129, 150, 156, 195, 196, 205, 208, 210, 212, 224, 227, 228, 265, 296; compared with free indirect speech 138–41, 149–50; compared with fused speech 121, 122, 123, 234, 235; compared with indirect speech 107, 114, 115, 184, 185–86, 211, 218–19, 222, 226, 293; compared with narrative reports of speech acts 126;
363
compared with nondirect speech 104 and first person 211 in fused reported speech 118, 122–24 and gerunds 77, 84 and historical present 91 and intercalated elements 43, 265, 267–68, 270 and marked word order 118, 258, 259 and pacing 107, 114, 122, 211, 218–19, 227 and participial clauses 77, 123 and perfective 25, 45, 117, 216, 218, 228, 229, 308 and postposed tags 43 Prague School concept of 23, 25 and preposed tags 43, 44, 305 role of questions 155 repetition 267 and second person 211 and sequentiality 77, 216 and slipping 265 and standard reporting strategy 82, 92, 115 and suspension of action 267–68 of testimony 43, 77, 92, 115 of unbound reports 43, 92 and unexpected forms 24–25, 114–15, 287 Forgeries 57, 58 Formulary redactions 138 Free direct speech (₎ 2, 4, 92–103, 138, 145, 167–71, 288 and backgrounding 293 in boundary-setting descriptions 79, 84, 94–96, 103, 144, 163, 293 in choral speech 99–102 and cohesion 93–94, 95, 98, 101, 102 and coherence 103 with commentary 97, 103 in Contemporary Standard Russian 14–15
364
R V
with convergent statements 97, 101 in crossplay 168, 171 and direct experience 102 and gear-shifting 96, 102–103, 163, 293 and involvement 92, 102 with judges’ reports 167–71 in multimodal testimony 95, 96–97, 99, 102, 103 in multiplex testimony 97–102, 115 in naming witnesses 93–94, 120–21, 130 in Old Russian 301 in opening complaints 38, 92–93, 146 and pacing 99, 102–3, 163, 164–65, 168, 293 at narrative peaks 92, 102 reading conventions for 4 and redundancy 101 and tagged direct speech 14–15, 115, 134 with trivial attributions 38, 62, 79, 92–93, 95, 97, 102, 163, 205 universality of 9 and vividness 92, 102 Free indirect speech (₎ 2, 11, 122, 133–50, 204, 250, 274–79, 288, 304, 307, 310, 311 and backgrounding 139–40, 141, 143, 148, 149–50, 287, 311 in boundary-setting descriptions 143–46, 149, 293, 311 characteristics of 133–35 with cognition predicates 276–78 as compact strategy 150, 289, 295 in Contemporary Standard Russian 133, 135–36, 310 in continuations 144, 277 deixis in 133–34, 135, 139, 143, 144, 149–50 as demonstration 307 emotive elements in 134, 149
evaluations in 134, 136, 141, 145, 277–78, 279, 293 and foregrounding, compared with direct speech 139–40, 141, 149–50; compared with narrative reports of speech acts 141 as hearer-based strategy 134–35 in incipits 146–48, 149 and indirect speech 134 in layered reported speech 220, 274, 295–96 “literariness” of 133, 135–36, 148–49, 310–11 as “modern” development 133, 135–36, 149, 288, 299 and narrative discourse 134–35 with narrative reports of speech acts 131, 133, 140–41, 149 and narratives of perception 62 “narratorless” 310 in orientations 143, 148, 149 and pacing 149–50 personal preterit in 134, 141, 143 and perspectival ambiguity 148–49 and polyphony 148–49 pronouns in 137–38 reading conventions for 135, 136, 145, 293, 299 and slipping 138, 139, 141, 146, 250 and staging 136–40 in summaries 140, 148, 150 in syntheses 279 terminology for 133, 136, 145 tense in 131, 134, 141–42, 143, 148 in testimony 133–50 in transitions 150 with trivial attributions 141, 142–48, 149, 274, 276–77, 293, 295 in verdicts 250, 274–79, 295 French 135, 142, 304 Old xv, 135 Fronting 84, 195, 232; see also Topicalization
S I Frustration 92, 210, 224 Fused reported speech 116–24, 189, 208, 233–37, 259–61, 283, 286, 292; see also Reported speech, grammatically integrated; Small clauses accusative in 117, 118, 122, 124, 233, 259 aspect in 117, 120 and backgrounding 119, 120, 122, 123, 234, 235, 236, 293, 296 in boundary-setting descriptions 79, 119–20, 122, 293, 310 characteristics of 116–17, 233, 237 in commentary 119, 121–22 as compact strategy 288–89, 293 and economy 124 and foregrounding 118, 122–24; compared with direct speech 121, 122, 123, 234, 235 and gear-shifting 122, 124, 237, 293, 295 and givenness 122, 234, 235 in incipits 117–18, 236, 292 in layered reports 117, 208, 233–37, 243, 296 in Old Russian 116–17, 122–23, 237 in orientations 117, 119, 122, 234, 236, 292 and pacing 118–19, 122, 124, 125, 237 pronouns in 117, 233, 260 as reductive strategy 124, 132 reflexivization in 117, 118, 119, 233, 260–61 and slipping 121 tense in 117, 120 word order in 118, 119 G Gear-shifting 46, 96, 102, 107, 119, 122, 124, 201, 227, 228, 293, 295; see also Pacing
365
Genitive genitive of 157; omitted, 162–64 under negation 117, 233, 259 objects 157, 165, 311 of possession 233 German 16 Gerunds 76–85, 94, 281, 293, 308 in amplifications 31, 40–41, 81, 84, 308 and backgrounding 78, 82–83, 84, 293, 315 in boundary-setting descriptions 78–81, 84 in commentary 78, 79, 81–82 and foregrounding 84 in incipit formulae 31, 78, 82–84, 281 and individuation 83–84 as intercalated quotation markers 238 with minor pauses 81, 89, 144 with nonsequenced actions 76, 77 in orientations 41 and repairs 84, 94 with secondary actions 76, 77, 79, 84 Givenness and compact strategies 131, 197, 240, 253, 254, 258, 293–94, 295 as disincentive for direct speech 186, 191, 200–201, 226, 261 as factor in ellipsis 38, 164–65 and frames 44, 46, 47, 290 and fused reported speech 122, 234, 235 and indirect speech 107, 112, 184, 185, 188, 189, 192, 217, 220, 222, 226–27, 255–56, 293, 295, 296, 314 and nominalizations 109 and subordination 106, 107, 117, 112, 184, 197, 217, 227, 253 g(o)s(podi)ne 47–49; see also Vocatives in judges’ questions 154–55 in judicial-referral records 191, 195
366
R V
in layered reports 205; omitted, 214 position of 48, 200 as typification strategy 48, 54 in testimony 47–48, 94, 95, 105, 264 in verdict 263; omitted, 256 gosudar’ 137 govorit’ 213 govoriti/govarivati 213, 215, 314, 316 Graphic cues to reported speech 4, 12, 46 Graphic introducers 5, 85, 301 H Hansard see Parliamentary reporting Hearer-based strategies 68–72, 92, 132, 135, 290–91, 294, 297, 308 Hearing, verbs of 221, 228, 229, 233, 315 Hearsay 143, 203 “ancestral” 221–24, 229, 235, 296 in complementized indirect speech 220–24, 226, 296, 315 in fused reported speech 235 in uncomplementized indirect speech 229 Hedges 52, 55, 59–60, 240, 307 Hendiadys 83 Heresy trials 85, 310 Heteroglossia 1–2, 285 and free indirect speech 134–35, 142, 145, “subjective” 4 Historical Pragmatics 4, 16–17, 302 Honorifics 47 Hungarian 12, 306, 310 Hypotaxis 72, 224, 253, 291 I i 44, 76, 98–99, 186, 245, 246 Icons of backgrounding 84, 107, 123 of continuation 168, 297 of distance 95
of foregrounding 91 of rapid pacing 233 of relevance 109, 139, 164, 289 of separation 83, 102 of unusual events 75, 86, 286–87 Illocutionary force, verbs denoting 125, 259, 270, 273, 295; see also Narrative reports of speech acts Immediacy 66, 297 Immunity charters 45, 263, 275, 280 Immunity judgment charter (žalovannaja pravaja gramota) 280–83 Imperatives 12, 154, 160, 225, 251 Imperfective 131, 159–60, 206, 208, 212, 226, 270, 308 and backgrounding 117, 120 and framing 206, 220–21, 228, 314 and intercalated elements 238 and negative presupposition 226 with nonpresupposed actions 226 with nonsequenced actions 76, 120, 159 and repetition 159–60 with topical actions 228–29 with unsituated actions 220–21, 223, 228, 229, 234, 314, 315 Impersonal verbs 62, 307 Inaction see Nonfeasance Incipits 19, 30–31, 34, 152, 154 and predictability of attribution 38, 146, 149 alternative 78, 82–84, 281, 308 free indirect speech in 146–48, 149 fused reported speech in 117–18, 236, 292 Indentations 46, 171 Indenture documents 261 Indiction, dating by 30 Indirect speech (₎ and backgrounding 107, 112, 113, 114, 212, 227, 228, 287, 296, 315 characteristics of 103–4 “complexity” of 25
S I in Contemporary Standard Russian 5, 11 deixis in 60, 103–4, 105, 107, 114, 116, 133–34, 288, 289 in dyadic structure 108 in earliest Slavic writings 49–50, 300–301 and economy 112 emotive elements in 67, 233 and foregrounding, compared with direct speech 107, 114, 115, 222, 139, 184, 185–86, 211, 218–19, 226, 293 and givenness 107, 112, 184, 185, 188, 189, 192, 217, 220, 222, 226–27, 255–56, 293, 295, 296, 314 in layered reports 209, 212–33, 254, 296, 315 in Old Russian 9, 11, 103–4 and orientations 210, 212, 217, 226, 228, 230, 234, 236, 281, 292, 296, 315 pronouns in 105, 109, 110, 137–38, 181, 183, 189, 196–97 and slipping 13–14, 108, 109, 195, 199–200, 263–65, 273 in summaries 114, 186, 218 “texture-analyzing” or “colored” 35, 67, 134 uncomplementized 9, 103–4, 182–83, 190, 197, 228–33, 234, 237, 256–59, 286, 288, 295 and verbatimness 104 Individuation 83, 180 Infinitival constructions 9, 172–75, 241, 242, 248, 251–52, 304 Ineffective statements see Frustration Information blocks 46, 84 Information structuring 7, 37, 118; see also Theme; Word order Inquest records (doprosnye recˇi) 152 Inquiring, verbs of
367
in judges’ tag formulae 153, 154, 156–62, 165, 167, 311, 312 in judicial-referral record 181, 194 in layered reported speech 206, 225, 315 Insistence 160 Instrumental, predicative 213, 233 Intentions 1–4, 285–86 and automatization 3, 5, 20–21, 24, 303 and conventions 20–21, 26, 42, 287 institutional(ized) 21, 286, 298 slipping and 14, 138–39 and reproductionist approach 49, 52, 53, 65 and standard strategy 37, 42, 44–45 Intercalated tags 7–9, 37, 117, 118, 237–40, 265–70, 288 aspect in 238 and cohesion 47–48, 105, 138, 238–40, 263, 264–66, 267, 270, 296–97 in Contemporary Standard Russian 8 and distancing 240 and foregrounding 43, 265, 267–68, 270 and gear-shifting 195 gerund as 238 interruptive nature of 8, 195, 297 kazati as 117, 213, 265 in layered reports 237–41, 296–97 and pacing 265, 267, 268 quotative particles as 76, 263, 269; see also deˇ(i) reinforcing preposed tags 9, 105, 109, 238–41, 263, 265–67, 296–97, 316 as speaker-based strategy 296–97 tense in 105, 109, 117, 195, 238, 266, 268 with trivial attributions 43, 230
368
R V
Interference, authorial/narratorial 53, 69–70, 279, 290; see also Control of interpretation Interior monologue 133, 136 Interludes, accounts of 104–7, 118–19, 122, 124, 142–43, 150, 173–74, 188, 201 Interpersonal meaning 10, 48, 54, 58 Interposed tags see Intercalated tags Interpretability see Coherence Interrogation records (rassprosnye recˇi) 50, 141, 152, 156 Interrogatives 154, 160, 165, 315 Intervention, effect of 70, 297 Intonational cues to reported speech 4, 12, 66 Intonational units 46 Inversions see Word order, inverted Invitations 154, 235, 272 Involvement and direct speech 66, 68–69, 210–11 and free direct speech 92, 102 and historical present 91, 208, 210–11 and m”lviti 210–11 iskati 317 Iteratives 120, 226 iz ust” 56 izgovoriti 316 izveˇcˇati 205 J jako 59–60 recitativum 60 jatisja 174, 190, 243, 312, 313 javiti/javljati 226, 315 Judgment charters (pravye gramoty) 28, 29–30, 31, 33, 58, 177, 186, 245, 282 K kazati (kažet”) 8, 117, 119, 213, 228, 233, 238, 256, 266, 314, 315 Knowledge
background 38, 40, 69, 95, 197, 294 experiential 4, 20, 66, 97, 145, 171, 286; see also Responsive understanding koe 212 L Law Codes Pskovian 113 Muscovite 18, 113, 139, 210, 279, 305 Layered reports 203–43, 253 and coherence 155 compact strategies in 243 in direct speech 205–12, 223–24, 296 as evaluation 241, 243, 270, 280, 295 function of 204–5 in free indirect speech 220, 274, 295–96 in fused reported speech 117, 208, 233–37, 243, 296 in indirect speech 209, 212–33, 254, 296, 315 intercalated tags in 237–41, 296–97 in judges’ reported speech 161–62 as narrative reports of speech acts 208, 210, 220, 229, 241–43, 270, 295 orientations in 204, 210, 212, 217, 226, 228, 230, 296, 315 quotative particles in 296–97, 316 situated 217, 221, 228, 229, 234, 255 speaker-based strategies in 243, 296 as universal 203 unsituated 221, 228, 229 with standard tag 207, 290, 295 and verdict 248, 253–54, 280, 295 vocatives in 48 Least effort, principle of 54 Legal-administrative writing 8, 25, 28, 30, 36–37, 39, 45–46, 56, 57, 59, 85, 91, 130, 136, 137, 189, 203,
S I 238, 261, 280, 286, 287–88, 301, 318 Lexical approaches to reported speech 15 Lexical change 85, 300–301 l”gati/s”l”gati 215, 240–41, 243 li154 Ligatures 57 Linguistic action nouns 125 Linguistic action verbs 91, 125, 213, 304 Longtime residents (starožil’ci) 78, 99, 114, 222, 276, 305 M Macrostructural connectors 48, 238 Manner adverbs 43, 59, 60; see also tak(o) Manner, Maxim of 58 Manner-of-speaking verbs 5, 6 Markedness 24, 25 Maxims, conversational 51, 53, 55, 58, 59, 226 Measures 46 Medial tags see Intercalated tags Memory, communal 111, 222, 296, 305 Memory constraints 51, 57, 306 Men of court (sudnye muži) lists of 34, 282 reported speech of 58, 179, 196, 294 role of 29, 30, 58, 179, 192, 194, 196 Mental activity, verbs of 6 Metapragmatic function of language 1, 303 Methexis 74 Method of residual forms 17, 22, 23–26, 75, 114, 117, 286, 287, 298 Mimesis 48, 52, 53, 74 m”lcˇati/zam”lcˇati 226, 242, 274, 316 m”lviti 90–92, 208, 209–11, 212, 217, 223–24, 273, 314 Mode of reporting, defined 5
369
mol 76 Monastic-rule charter 28, 308 Mood changes 11, 103, 210, 291 Mouthpieces 65, 307 Multiclausal reports supplementary tagging in 8–9 intercalated tags in 105, 109, 238–40, 265–67, 296–97 and compact strategies 122, 262–63 slipping in 198–200, 250 Multimodal testimony 95–97, 99, 102, 103 Multiplex testimony alternative verba dicendi in 87–88 capitalization in 153–54 choral speech in 101–102, 293 conjunctions in 98–99, 153–54, 186, 229, 246, 309 free direct speech in 97–102, 293 indirect speech in 110–12, 114, 115, 293 summaries of 218 Muscovy, legal procedures in 113, 139, 166, 206–7, 219, 290, 305 muževati 273 N naizust 55–56 Nakazy see Orders, memoranda of Naming witnesses and free direct speech 93 and fused reported speech 120–21 in layered reported speech 241 and narrative reports 130–31, 270, 273 as responsibility of litigants 40, 164, 276 napisati see pisati (pišet”) Narration, covert style of 43, 48, 71–72, 125, 130, 134, 136, 141, 291, 293, 301 Narrative, first-person 23, 204, 283
370
R V
Narrative norm 27, 45, 90, 107, 120, 245, 258 Narrative reports of speech acts (’s) 9, 124–32, 172–75 in adjudication 33–34, 248–53 and backgrounding 126, 129, 210, 212 as compact strategy 280, 283, 288–89, 294 deixis in 252, 254 and direct speech 126–29, 131, 172–73, 194, 208, 210, 212, 280, 292, 294 and economy 127, 132, 140 elaborated 251–52, 294 and evaluations 132, 213, 241, 243, 248, 270, 273, 294, 295, 296 and foregrounding, compared with direct speech 126; compared with free indirect speech 141 and free indirect speech 131, 133, 140–41, 149 in incipits 82, 280–81 in judicial-referral records 197–98, 314 in layered reported speech 220, 229, 241–43, 310 and narration of nonverbal actions 124–25, 126, 194, 241 and orientations 126, 210, 212, 273 and pacing 125–26, 241 and reduction 129, 130–32, 140, 172, 248, 270, 271, 273, 294 and relevance 125–26, 129, 242, 292, 294 in ratio decidendi 259, 263, 269, 270–74 and small-clause constructions 213 and slipping 141, 146, 251–52, 273, 294–95 as speaker-based strategy 132, 243 in summaries 125, 128–29, 131, 132, 252
tense in 131 Narrative representation of speech act with topic 252 Narrator’s report of voice 60 Navajo 10 nazyvati 213, 314, 315 Negation 52, 117, 120, 128, 130, 233, 234, 251, 259, 272, 273, 274, 315 Nepal, languages of 50 New Philology xvi Newspaper reporting 54, 311 Newsworthiness 68, 77, 78, 84, 107 No-shows 82, 130, 184, 235, 294, 315, 317 Nominal introducers 6, 198, 199, 289, 301 Nominalizations 109, 289 Nomination strategies 38–39, 101, 151–52, 156–57, 180, 182 Nondirect speech in adjudications 251–52 as compact strategy 289, 293, 295 and foregrounding, compared with direct speech 104, 208 relation to direct speech 12–13, 35, 104 relation to indirect speech 9, 104 in layered reported speech 205, 208, 210, 212, 213, 215, 221, 226–28, 296 in ratio decidendi 252, 254, 256, 264, 266, 283, 295 in testimony 106, 110–11, 117, 293 in verifications 190 Nonfeasance 82, 128–30, 162, 198, 209, 214, 235–36, 242, 256, 274, 276, 315 Nonverbal action 27, 78, 79, 95, 124–25, 126, 149, 249, 293 Notes, of scribes 28, 47–48, 57–58 Novgorod 13, 46, 192, 245, 311
S I O Oaths 33, 87, 112, 125, 128, 129, 225, 240, 292, 315 Object clauses see Complement clauses Objects, ellipsis of 162–65 obl”živiti 273 ob”muževati 141 Obolo 49 ob(v)initi 34, 248–49, 252 Obysk” 217 ocˇistiti 273–74 Old Believer Schism 56 Omitted speech 60–61, 213, 243 Opacity 306 Opening complaint 31–32, 167, 292 free direct speech in 38, 92–93, 146 (free) indirect speech in 136, 141, 146–48 and gerund tags 83–84, 281–82 orientations in 41 and word order 37 oposluš’stvovati l”živo 243 opraviti 34, 248–49, 252 Orality 29, 46–47, 55–56, 58, 73–74, 179, 198, 203, 249, 283 Oratio recta see Direct speech Orders, memoranda of (nakazy) 288 Orientations in covert style 43, 44 dates as 30 free indirect speech in 143, 148, 149 fused reported speech in 117, 119, 122, 234, 236, 292 gerunds in 41 in incipits 148 in indirect speech 210, 212, 217, 226, 228, 230, 234, 236, 281, 292, 296, 315 in interludes 118–19, 126, 292 in judicial-referral records 195–96, 201 in layered reports 204, 210, 212, 217, 226, 228, 230, 296, 315
371
as narrative reports 126, 210, 212, 273 in opening complaint 41 Originals see Anterior utterances ot”kazati 89–90 ot”suditi 248, 249 Otvodnye gramoty 37, 145–46 ože (ož’) 62 P Pacing see also Gear-shifting and compact strategies 293 and complementized reports 197, 229 and direct speech 107, 114, 124, 126, 150, 211, 218–19, 227 and foregrounding 107, 114, 122, 211, 218–19, 227 and free direct speech 99, 102–3, 163, 164–65, 168, 293 and free indirect speech 149–50 and fused speech 118–19, 122, 124, 125, 237 and intercalated elements 265, 267, 268 and narrative reports of speech acts 125–26, 241 and omission of complementizer 233 and presupposition 141, 258 paky 220 Paralinguistic signals 66 Parallelism 115, 259 Parataxis 72, 227, 291 Parentheticals 8, 133, 213, 230, 238, 240, 289; see also Intercalated tags; Macrostructural connectors Parliamentary reporting 54, 58, 311 Partial quotes see Quotations, incorporated Participant configuration see Footing Participant status see Footing Participial constructions 9, 31, 75–85, 122–23, 304; see also Gerund clauses
372
R V
Particles deictic 231 emphatic/intensifying 61, 220, 238 interrogative 154 quotative 6, 11–12, 21, 76, 133, 137, 138, 156, 239, 263–66, 269–70, 281, 297, 316; see also deˇ(i) topicalizing 238 Pauses in action 81 Peak, narrative 68, 92, 102, 205, 211 Perception objects of 62 reports of 62, 117, 307 Perfective 85, 174, 213, 215, 220, 300, 314, 316 in foregrounded narrative 25, 45, 117, 218, 228, 229, 308 and inability 234, 274 as narrative norm 45, 120 negated 120, 234, 274, 315 and performative formulae 270, 272, 317 and presupposed past events 226 punctiliar 255 and situated past events 215–20, 221, 228, 229, 234, 255 Performatives 131, 213, 225, 270, 272, 314, 317 Periods, marking reported speech 46 Perjury, accusations of 216–17, 240–41, 243; see also l”gati/s”l”gati; oposluš’stvovati l”živo; poklepati Perlocutionary force, verbs denoting 125, 209, 273–74, 295; see also Narrative reports of speech acts Perspectival shifts see Deixis, shifts of Petitions 47, 82, 91, 116, 155, 209, 276, 280–81, 297 Phase verbs 5–6, 213 Phatic elements 44, 58, 263, 314 Phonetic reconstruction 23 pisati (pišet”)/ napisati 62, 275, 301, 307
Pivot, deictic 11, 21 Plaintiff’s charge see Opening complaint Plotline conjunctions in 76, 99 and imperfective 159–60 and indirect speech 218 as orientation for testimony 44 and perfective 45–46, 228, 315 Pluperfect meaning 120 po tomu 141 po tomu cˇ’to 252 Podpisnye dokladnye sudnye spiski see Referred trial records poimatisja 272 poklepati 243 Politeness 68 negative 69 positive 69, 86 Poluustav see Semi-uncial Polyphony 133, 148–49, 247 Polysyndeton 36, 72, 178, 239, 245, 263 Porucˇnye kabaly see Surety bonds pos”latisja see s”latisja posluš’stvovati 273 Postposition of tags 7–8, 37, 43–44, 133, 265 of subjects see Word order, inverted Power, relative 38, 47, 68 Pragmaphilology 18 Pravye gramoty see Judgment charters Preambles to questions 161–62, 165, 212–13, 219, 234, 296 Predictability see Givenness Preposed tags 7, 37, 42–43, 85, 92, 103, 288 and coherence 44, 152–53 as diffuse strategy 290 delimiting function of 44, 153 and foregrounding 43–44, 305 reinforced by intercalated tags 9, 105, 109, 238–41, 263, 265–67, 296–97, 316
S I with the judges’ reports 151–53, 156 phatic function of 44 Present tense in adjudications 251 with continuations 105 and focalization 62, 138, 143, 256 in free indirect speech 131, 134 historical 90–91, 117, 208, 211, 223, 224, 309 with intercalated verbs 105, 109, 117, 238, 266, 268 with nonsequenced reports 131 personal 131, 143 with rejoinders 215 with reports of current relevance 212–13, 214, 215, 220, 225, 228, 234, 241 Presupposition see Givenness Preterit tense in free indirect speech 131, 134, 141–42, 148 imperfective, and backgrounding 117; in flashbacks 120; in frames 206, 228, 314; and insistence 159–60; nonsequenced past events 120, 131, 159–60; and repetition 159–60; and unsituated past events 220–221, 223, 225–26, 228, 229, 234, 314, 315 with intercalated verbs 195 as narrative norm 45, 120, 245 with narrative reports of speech acts 174, 198, 270, 272, 317 perfective, and foregrounding 25, 45, 117, 218, 228, 229, 308; in narrative 79, 105, 117, 131, 134, 141, 216, 241, 245; negated 120, 234, 274, 315; and situated past events 215–20, 221, 228, 229, 234, 255; personal 141, 148 prikazati/prikazyvati 301, 316 Printing 46, 55, 171
373
Prior texts 1, 303 prisuditi 34, 248–49, 251 Probate 189 Projected statements 3, 204, 224–25 Prominence see Foregrounding Promises 33, 82, 125, 165, 174, 190, 209, 241, 243, 272 Pronouns 40, 102, 181, 213, 231 in direct speech 102, 112, 154, 181, 191, 195, 196–97, 221, 224 ellipsis of see Subjects, null in free indirect speech 137–38 in fused reported speech 117, 233, 260 in indirect speech 105, 109, 110, 137–38, 181, 183, 189, 196–97 prositi 272 Prosodic features 12 Proteus Principle 69, 85, 297 Prototypes, salience of 19 Prototypical speech events 19–20, 22, 34, 36, 89, 160 departures from 22, 42, 75, 286–87, 291, 292–94, 298; and gerund tags 78, 79; and alternative verba dicendi 86, 88, 89, 90; and free direct speech 93, 95, 97–98, 102; and indirect/nondirect speech 104, 107, 110, 114–15; and fused reports 117, 119; and narrative reports 128, 172; and free indirect speech 146, 148; and ellipsis 162–63; within hearsay 224 in judicial-referral report 178–79, 293 in trial narrative see Testimony, prototypical for witnesses 87, 112–13, 222 Proverbs 134 Proxies 129, 178, 181, 182, 183 Pseudo-objective discourse 141 Pskov 29, 113, 245, 250, 308, 311 Punctuation, in Old Russian 46, 171, 305
374
R V
Purchase deeds, referred (dokladnye kupcˇie gramoty) 190, 313 Purposive, definition of 3 Q Quality, Maxim of 51, 55, 58, 59 Quantity, Maxim of 58, 226 Quasi-direct speech see Free indirect speech Question-and-answer structure see Dyads Questions and foregrounding 155 omission of 90 Quotation marks 46, 306 Quotations allusive 4, 134 incorporated 13, 67, 199–200 of self 10, 210, 218, 227 in written works 54, 203 of written records 51, 56, 57, 61–64, 71, 97, 102, 203, 275, 292, 307, 309 Quotative markers see Particles, quotative R Raising 122–23, 259 Rapid-fire exchanges 165, 168, 233 raskazati 140 Rassprosnye recˇi see Interrogation records Raz”ezžie gramoty 37 Razvodnye gramoty 37 Reaction, requests for 213–14, 228 Reading, in medieval Rus’ 46–47 Realignments, verbs denoting 5–6 Recantations 192, 193, 196, 294 recˇi in standard strategy 37, 41, 45, 88, 90, 92, 290, 313 gerunds of 76, 238, 308 obsolescence of 85–86, 115, 300 with judges’ reports 91, 165–67
with layered reports 205–9, 290 recˇisja 174, 312, 313 Reconvenements 172, 292 Recoverability see Givenness Reductionist approaches 11–16 Referral formula 33, 174, 190, 243, 312, 313 Referral memoranda (ssylocˇnye pamjati) 116, 309 Referred trial records (dokladnye sudnye spiski) 28, 29, 33, 34, 177, 245 Reflexives, logophoric 260 Reflexivization, in fused speech 117, 118, 119, 233, 260–61 Refusals 58, 125, 127–29, 165–67, 271–72, 292 Refutations see Rejoinders Rejoinders 31, 148, 164, 214, 215, 225, 226, 227, 233, 239, 296 rek”ši 238 Relation, Maxim of 58; see also Relevance Relevance 3, 4; see also Cooperation, conversational icons of 109, 139, 212, 289 and institutional purpose 27, 30–31, 38, 44, 58, 59, 73, 127, 152, 274, 285 and mode of reporting 109, 125–26, 129, 130, 212, 250, 289–90, 292, 297, 298 and tense 228, 234 of verbatimness 52–54, 58, 65–66, 73 Relative constructions 12, 49, 59–60, 123, 148, 199–200, 239, 250, 310, 315, 316 Repairs 52, 84 Reportability 72, 92, 126, 127 Reported speech communicative subordination of 2, 65, 70 context-sensitivity of xiv, 21, 85, 286, 287
S I as continuum 12–13, 104, 134, 311 creative character of 2, 11, 20, 57, 65 deictically ambiguous see Nondirect speech grammatically integrated 9, 37, 118, 123, 173–74, 189, 194, 201, 208, 233, 254, 289, 295, 309; see also Fused reported speech inexplicitly marked 1, 2, 5, 122, 134–35, 142, 276–78 as inset in frame 1–2 less-than-direct 37, 104, 283, 289, 292, 294, 295, 296, 297; see also Compact strategies mediated by mental images 2–3, 4, 73–74 as object clause 11, 304 open-endedness of 13, 15 perceptual autonomy of 1–2; see also Heteroglossia pervasiveness of 1 as phenomenon of discourse 11, 16 as semantic motif 285, 318 submerged 124, 174, 248, 249–50, 252 terminology for xiii, 304 as universal 1, 5, 9, 49, 203, 303 within reported speech see Layered reports Reported thought xiii, 4–5, 10, 52, 135, 138, 143 Reportedness 4–5, 12, 109, 122, 133, 135, 137–38, 143, 144, 145, 247, 263–64, 276–77; see also Averrals Reporter-commitment model of direct speech 49, 51–54, 55, 56, 70, 71, 307 Reports, preliminary 29, 178, 180, 194, 197–201, 294 Reproaches 162 Reproductionist approach to direct speech 49, 50–55, 64–65, 104, 107, 306
375
Requesting, verbs of 159, 160–61, 272 Requests for amplification 212–13, 214, 225, 228, 234, 239, 240, 315 for reaction 213–14, 228 Requotations see Layered reports Residual forms see Prototypical speech events, departures from Responsive understanding 3, 20 Retractions see Recantations Ritardando see Pacing Rjadnaja gramota 46 Rjazan’ 136, 139 rosprositi 315 rule 249 Rus’, Northeastern 29, 34, 45, 46, 48, 91, 245; see also Muscovy Rus’, Northwestern 13, 29, 45–46, 113, 192, 245, 250, 308, 311–312, 316; see also Dvina, Novgorod, Pskov Russian, Contemporary Standard deictically ambiguous reports in 12 direct speech in 5 discourse themes in 96, 266 free direct speech in 14–15 free indirect speech in 133, 135–36, 310 imperfective in 160 indirect speech in 5, 11 intercalated tags in 8 introducers in 5–6, 301 postposed tags in 8 reflexives in 118 report-clause types in 5, 9–10, 11, 233 Russian, Old direct speech in 9, 11, 35, 49, 161, 306 discourse theme in 96 evolution of reported speech in 288, 301 free direct speech in 301
376
R V
fused reported speech in 116–17, 122–23, 237 historical present in 91, 309 hypotaxis in 72, 291 imperfective in 120 indirect speech in 9, 11, 103–4 intercalated elements in 7–8, 9, 76, 200, 238, 265–66 multiple attributions in 9 narrative norm in 15, 27, 45–46 narrative reports of speech acts in 9, 131 negated commands in 128 null subjects in 77, 96 parataxis in 291 postposed tags in 7–8, 43 preposed tags in 7, 43 punctuation in 46, 305 quotative particles in 11, 76 reflexives in 118 report-clause types in 3, 6–7, 9–10, 15 slipping in 9, 13–14, 200, 264, 304 small clauses in 9–10, 116–17, 122–23, 237 types of tag in 6, 11, 76, 85, 161, 301 verba dicendi in 85–86, 91 word order in 37, 43, 77 118 S Saints’ lives 37 Salience see Foregrounding Salience, plot 123 Saliency Hierarchy 318 say 5, 10, 43 Saying, verbs of see Verba dicendi Schematic Language Representation, Theory of 48 Scribes 20, 21, 22, 27, 28, 29, 38, 39, 40, 42–43, 47, 50, 57–58, 61, 71–72, 86, 91, 92, 104, 115, 132, 136, 146, 158, 177, 181, 182, 189,
191, 203, 204, 207, 220, 243, 245, 247, 249, 252–53, 274, 285, 286, 288, 290, 291, 293, 294, 297, 298, 304 Scriptio continua 46, 76, 205 Scripture 56 Self-quotations 10, 210, 218, 227 sei/sii/ses’ 181, 182, 183, 197; see also Demonstratives Semi-uncial (poluustav) 57 ses’ see sei/sii/ses’ Setting see Orientation Settlements, amicable 28, 37, 145–46, 190, 293, 305 Settlement charter (rjadnaja gramota) 46 sii see sei/sii/ses’ Silence, reports of 209, 226, 242, 256, 274, 277, 316; see also m”lcˇati/zam”lcˇati “Silencing of prose” 310 Simonov Monastery 183 skazat’ 5, 85, 300 s(”)kazati/s(”)kazyvati as default verbum dicendi 5, 88, 300 displacing recˇi 85, 115, 300 with fused reported speech 117, 120, 233, 259 as intercalated tag 268, 317 in judicial-referral records 182–84, 187–92, 195, 196, 198–200, 208 with layered reported speech 212, 213, 215, 218, 220, 221, 223, 228, 229, 233, 238, 314, 315 with “omitted speech” 243, 316 in other text-kinds 146 in ratio decidendi 254–56, 259, 266, with testimony 87–88, 103, 110, 310 s”kazatisja 261 Skazki see Depositions s”latisja/pos”latisja 130–31, 225, 270, 315 Slavic languages 7, 49, 60, 91, 300–301
S I Slipping 9, 13–14, 67, 264, 304; see also Quotations, incorporated and coherence 264–65 of direct speech into (free) indirect 137–39 in extended reported speech 264 of indirect/nondirect speech into direct 13–14, 108, 109, 195, 199–200, 263–65; into narrative reports 273 and foregrounding 265 of fused reported speech into direct 121 of narrative reports into indirect/nondirect speech 141, 146, 251–52, 294–95; into direct speech 29, 250, 294 and free indirect speech 138 intentionality of 14, 138–39 slovo v” slovo 55, 57 slyxati 221, 228, 229, 233, 315 Small clauses 9, 117, 122, 123, 213, 233, 234, 237, 259–60, 288; see also Fused reported speech so to speak 59 Socialization, linguistic 1, 5 Song of Prince Igor’s Campaign xv Sound-laws 23 Speaker-based strategies 6, 68, 72, 132, 211, 243, 280, 283, 290, 294, 296–97 Speech genres 18, 19, 21, 22, 128, 204, 212, 213, 227, 235, 298, 302 Speech will 4, 42, 50, 65, 288 Speech-act verbs (’s) see also Narrative reports of speech acts definition 310 bound forms after 9–10 tagging direct speech 5, 83, 85, 86, 161, 205, 301 in judges’ reports 172–75 lexical-reductionist approaches to 15 as speaker-based strategy 43, 72, 125
377
in layered reports 209, 213, 215, 225, 226, 228, 233, 240–43, 296, 315 in verdicts 34, 248–49, 270–74 Speech acts, indirect 272 Spokesmen 38–39, 40, 101, 114 s”prašivati 206 s”prošati 160 s”prositi 156–60, 181, 206, 225, 311 s”rok” 175, 199, 200, 313 Ssylocˇnye pamjati see Referral memoranda Stage-managing, reports of see Commands, procedural Staging 136, 138–39 Starožil’ci see Longtime residents Statejnye spiski see Diplomatic dossiers Stopping-points see Episode boundaries Streamlining effects 122 Style indirect libre see Free indirect speech Stylistic factors in reporting 15, 16, 35, 50, 54, 55, 91, 133, 145, 149–50, 157, 165, 289, 302; see also Drift, stylistic; Emotive elements; Narration, covert style of Subjectless constructions 62 Subjects, null 8, 112, 189, 236–37, 246, 256, 258, 317 Subversion in layered reports 214, 228, 234, 314, 315 in ratio decidendi 247 255, 265, 267, 279 sud” 198, 199 Sudebnik see Law Codes, Muscovite Sudnye spiski see Trial records Summaries 42, 43, 45, 95, 114, 148, 186, 197, 218, 258; see also Syntheses of documents 116, 186, 288, 313 and free indirect speech 140, 148, 150 and indirect speech 114, 186, 218
378
R V
and narrative reports of speech acts 125, 128–29, 131, 132, 252 speech 52, 62 Sureties 195, 314 Surety bonds (porucˇnye kabaly) 190 Suspension of action 267–68 Synopses see Summaries Syntactic change 15, 16, 300–301 Syntheses 61, 218, 260, 269, 279; see also Summaries of content 116, 237, 283, 315 T Tag formula 4, 5–9, 10, 14, 15, 21, 62, 102, 288 abridged 164–65, 168 adjuncts as 6, 15, 37, 43, 62, 133, 279, 289 backgrounding of 79, 84, 230 and coherence 15, 62, 68 constatative function 44, 77, 143, 146 and coordinating interpretation 6 delimiting function 44, 72, 94, 95, 101, 102, 111, 153, 238, 263, 268, 291 with documentary evidence 61–64 and evaluation 6, 43, 72, 91–92, 209–10, 212, 213, 224, 259, 296 first-person 227 foregrounding of 43, 77–78, 83–84, 92, 107, 113–14, 184, 185, 222, 230, 289, 293 hedged 52 irrealis 52 in judges’ reported speech 151–52, 153–54, 156–67, 180–81 and narrative framework 27, 42, 44, 46, 59, 72, 153, 290 negated 52 null see Free direct speech in Old Russian 6, 11, 76, 85, 161, 301
position of 7; see also Intercalated tags; Postposed tags; Preposed tags present tense in 105, 109, 212–13, 214, 215, 220, 224–25, 228, 234, 241 reinforced by intercalated elements 9, 105, 109, 238–39, 240, 241, 263, 265–66, 267, 296–97 second-person 227 speech-act verbs (’s) as 5, 83, 85, 86, 161, 205, 301 standard: with testimony 36–37, 39, 40–47, 59, 79, 81, 82, 85–86, 92, 98, 103, 109, 110, 115, 137, 144, 146; in judicial-referral reports 184, 192–93; in layered reports 205, 207, 290, 296; in ratio decidendi 261 tracking function 44, 62, 68, 290 visual icons as 14 tak skazat’ 59 tak(o) and cohesion 59 and complementized reports 108, 231 and foregrounding 155, 210 in gerund tags 76 postposition to verb 77, 78, 86, 90 with quotations from documents 63–64 spelling of 307 in standard tag 36–37, 44 as typicality marker 59–61, 63–64 Tense, choice of 6, 105, 117, 228, 229, 233–34, 241, 251; see also Present tense; Preterit tense Testimony and foregrounding 115 ostensive (gestural) 22, 79, 95, 97, 103, 293 prototypical: defined 36, 291–92; and standard reporting strategy 36, 75, 110, 115, 191, 261; 78, 79, 86, 93, 116
S I Text(ual) conveyors 6, 52, 62, 198, 213 that 259 Theme change of 48, 96, 105, 118, 252, 264, 267, 268 continuity of 96, 231, 233, 246, 252, 306 Thick description 3, 303 to 231 togo deˇlja cˇ’to 252 toi/tyi/tot” 182; see also Demonstratives Topic, discourse see Theme Topic statements see Syntheses of content Topicalization 12, 78, 122, 155, 195, 220, 223, 232, 238, 247, 270, 291; see also Fronting Topicalizing clauses 11, 12, 233, 239, 316 tot” see toi/tyi/tot” Tracking 62, 169, 171, 290 Transitional episodes see Interludes between hearing, accounts of Transmitter-senders 65, 207 Treaties 47 Trial dossiers (sudnye dela) 29, 85, 115–16, 155–56, 246, 249, 283, 297, 309, 310 Trial records (sudnye spiski) 19, 28–29, 31–33, 34 Trial transcripts abridged 36, 148, 152, 305, 307 dates in 30–31 Trinity-Sergius Monastery 64, 158, 313 tyi see toi/tyi/tot” Typicality 2, 48, 53–54, 58, 59–61, 63–64, 179, 306 Typological data, use of 17, 18, 21, 286, 299 U ucˇiniti dovod” 130 ucˇiniti s”rok” 175, 313
379
Uncooperative responses see Cooperation, conversational ukazati 174, 316 Ukaznye gramoty, ukazy see Edicts Uneigentliche direkte Rede see Free indirect speech Unexpected events, representation of 42, 78, 110, 128–30, 172, 210, 211, 224, 292, 298 Uniformitarian Hypothesis xv, 288 unquote 263 V veleˇti 125, 172, 175, 194, 242, 243, 251–52, 313, 316 Verba dicendi (’s) 5, 9, 10, 15, 24, 221, 309; see also deˇ(ja)ti, govoriti, kazati, m”lviti, recˇi, s(”)kazati/s(”)kazyvati alternative, with testimony 85–92 in Contemporary Standard Russian 5; see also govorit’; skazat’ contrasted with speech-act verbs 43, 72, 85, 243, 301 default 5, 10, 255 with grammatically integrated reports 9, 117, 119, 233, 304 as hearer-based strategy 72, 85, 243, 301 in Old Russian 9, 16, 85, 301 omission of 164 Verba percipiendi 62, 117, 307 Verbatimness and direct speech 49–58, 60, 64–65, 67, 265, 306, 307; see also Reporter-commitment model of direct speech; Reproductionist approach to direct speech and indirect speech 104 marked character of 55 in medieval Russian culture 56–57 as percept of interpreter 66, 67, 71, 74
380
R V
as purport 51–52 in printed language 55–56, 203 in quotations of written records 61, 62–63, 64, 71, 96, 203 and relevance 52–54, 58, 65–66, 73 Verdictive speech acts 249 Verdicts directed 34, 247, 253, 260, 263–64, 274–75 executed 34, 253, 260, 263–64, 275 Verification procedure 28–29, 177–89, 190–94, 196, 197, 198, 200, 232, 254–55, 293–94, 317; see also Corroborations; Falsifications v”imenovati 130, 273 vinu položiti 141 Visual icons 14 Vividness 66, 68, 72, 91, 92, 102, 297; see also Drama; Involvement Vocatives see also Appellatives; g(o)s(podi)ne position of 8 in direct speech 35, 291, 306 in testimony 35, 97 in judges’ questions 154–55, 168, 171, 309 v”prositi/v”prašati 159, 312 v”sprositi 157–59, 165, 181, 311, 313 vyl”gati 243 v”zveˇstiti 301 W Wackernagel’s position 8, 48, 200, 238 Warnings 205, 207, 210 Wills 189–90
Word order 7, 37, 43, 118 inverted 31, 37, 77, 119, 222–23, 232, 258, 259, 317 marked 231–32, 258, 259, 295 neutral 31, 232, 258, 259 Writer-based strategies see Speakerbased strategies Writing role of 29, 30, 47, 55, 57, 58, 198, 203, 249, 308; see also Orality technology of 55, 57 verb of see pisati (pišet”) Y Yoruba 310 Z žalobnicy see Petitions žalovannye gramoty see Immunity charters žalovannaja pravaja gramota see Immunity judgment charter zam”lcˇati see m”lcˇati/zam”lcˇati zaperetisja 141, 315 zaslyšeti 233, 315 že 61, 220, 238 Zero quotatives see Free direct speech Zero reference see Subjects, null znaxar’stvo 215 z”vati 208, 213, 228, 314, 315
In the PRAGMATICS AND BEYOND NEW SERIES the following titles have been published thus far or are scheduled for publication: 1. WALTER, Bettyruth: The Jury Summation as Speech Genre: An Ethnographic Study of What it Means to Those who Use it. Amsterdam/Philadelphia, 1988. 2. BARTON, Ellen: Nonsentential Constituents: A Theory of Grammatical Structure and Pragmatic Interpretation. Amsterdam/Philadelphia, 1990. 3. OLEKSY, Wieslaw (ed.): Contrastive Pragmatics. Amsterdam/Philadelphia, 1989. 4. RAFFLER-ENGEL, Walburga von (ed.): Doctor-Patient Interaction. Amsterdam/Philadelphia, 1989. 5. THELIN, Nils B. (ed.): Verbal Aspect in Discourse. Amsterdam/Philadelphia, 1990. 6. VERSCHUEREN, Jef (ed.): Selected Papers from the 1987 International Pragmatics Conference. Vol. I: Pragmatics at Issue. Vol. II: Levels of Linguistic Adaptation. Vol. III: The Pragmatics of Intercultural and International Communication (ed. with Jan Blommaert). Amsterdam/Philadelphia, 1991. 7. LINDENFELD, Jacqueline: Speech and Sociability at French Urban Market Places. Amsterdam/Philadelphia, 1990. 8. YOUNG, Lynne: Language as Behaviour, Language as Code: A Study of Academic English. Amsterdam/Philadelphia, 1990. 9. LUKE, Kang-Kwong: Utterance Particles in Cantonese Conversation. Amsterdam/Philadelphia, 1990. 10. MURRAY, Denise E.: Conversation for Action. The computer terminal as medium of communication. Amsterdam/Philadelphia, 1991. 11. LUONG, Hy V.: Discursive Practices and Linguistic Meanings. The Vietnamese system of person reference. Amsterdam/Philadelphia, 1990. 12. ABRAHAM, Werner (ed.): Discourse Particles. Descriptive and theoretical investigations on the logical, syntactic and pragmatic properties of discourse particles in German. Amsterdam/Philadelphia, 1991. 13. NUYTS, Jan, A. Machtelt BOLKESTEIN and Co VET (eds): Layers and Levels of Representation in Language Theory: a functional view. Amsterdam/Philadelphia, 1990. 14. SCHWARTZ, Ursula: Young Children’s Dyadic Pretend Play. Amsterdam/Philadelphia, 1991. 15. KOMTER, Martha: Conflict and Cooperation in Job Interviews. Amsterdam/Philadelphia, 1991. 16. MANN, William C. and Sandra A. THOMPSON (eds): Discourse Description: Diverse Linguistic Analyses of a Fund-Raising Text. Amsterdam/Philadelphia, 1992. 17. PIÉRAUT-LE BONNIEC, Gilberte and Marlene DOLITSKY (eds): Language Bases ... Discourse Bases. Amsterdam/Philadelphia, 1991. 18. JOHNSTONE, Barbara: Repetition in Arabic Discourse. Paradigms, syntagms and the ecology of language. Amsterdam/Philadelphia, 1991. 19. BAKER, Carolyn D. and Allan LUKE (eds): Towards a Critical Sociology of Reading Pedagogy. Papers of the XII World Congress on Reading. Amsterdam/Philadelphia, 1991. 20. NUYTS, Jan: Aspects of a Cognitive-Pragmatic Theory of Language. On cognition, functionalism, and grammar. Amsterdam/Philadelphia, 1992. 21. SEARLE, John R. et al.: (On) Searle on Conversation. Compiled and introduced by Herman Parret and Jef Verschueren. Amsterdam/Philadelphia, 1992.
22. AUER, Peter and Aldo Di LUZIO (eds): The Contextualization of Language. Amsterdam/Philadelphia, 1992. 23. FORTESCUE, Michael, Peter HARDER and Lars KRISTOFFERSEN (eds): Layered Structure and Reference in a Functional Perspective. Papers from the Functional Grammar Conference, Copenhagen, 1990. Amsterdam/Philadelphia, 1992. 24. MAYNARD, Senko K.: Discourse Modality: Subjectivity, Emotion and Voice in the Japanese Language. Amsterdam/Philadelphia, 1993. 25. COUPER-KUHLEN, Elizabeth: English Speech Rhythm. Form and function in everyday verbal interaction. Amsterdam/Philadelphia, 1993. 26. STYGALL, Gail: Trial Language. A study in differential discourse processing. Amsterdam/ Philadelphia, 1994. 27. SUTER, Hans Jürg: The Wedding Report: A Prototypical Approach to the Study of Traditional Text Types. Amsterdam/Philadelphia, 1993. 28. VAN DE WALLE, Lieve: Pragmatics and Classical Sanskrit. Amsterdam/Philadelphia, 1993. 29. BARSKY, Robert F.: Constructing a Productive Other: Discourse theory and the convention refugee hearing. Amsterdam/Philadelphia, 1994. 30. WORTHAM, Stanton E.F.: Acting Out Participant Examples in the Classroom. Amsterdam/Philadelphia, 1994. 31. WILDGEN, Wolfgang: Process, Image and Meaning. A realistic model of the meanings of sentences and narrative texts. Amsterdam/Philadelphia, 1994. 32. SHIBATANI, Masayoshi and Sandra A. THOMPSON (eds): Essays in Semantics and Pragmatics. Amsterdam/Philadelphia, 1995. 33. GOOSSENS, Louis, Paul PAUWELS, Brygida RUDZKA-OSTYN, Anne-Marie SIMONVANDENBERGEN and Johan VANPARYS: By Word of Mouth. Metaphor, metonymy and linguistic action in a cognitive perspective. Amsterdam/Philadelphia, 1995. 34. BARBE, Katharina: Irony in Context. Amsterdam/Philadelphia, 1995. 35. JUCKER, Andreas H. (ed.): Historical Pragmatics. Pragmatic developments in the history of English. Amsterdam/Philadelphia, 1995. 36. CHILTON, Paul, Mikhail V. ILYIN and Jacob MEY: Political Discourse in Transition in Eastern and Western Europe (1989-1991). Amsterdam/Philadelphia, 1998. 37. CARSTON, Robyn and Seiji UCHIDA (eds): Relevance Theory. Applications and implications. Amsterdam/Philadelphia, 1998. 38. FRETHEIM, Thorstein and Jeanette K. GUNDEL (eds): Reference and Referent Accessibility. Amsterdam/Philadelphia, 1996. 39. HERRING, Susan (ed.): Computer-Mediated Communication. Linguistic, social, and cross-cultural perspectives. Amsterdam/Philadelphia, 1996. 40. DIAMOND, Julie: Status and Power in Verbal Interaction. A study of discourse in a closeknit social network. Amsterdam/Philadelphia, 1996. 41. VENTOLA, Eija and Anna MAURANEN, (eds): Academic Writing. Intercultural and textual issues. Amsterdam/Philadelphia, 1996. 42. WODAK, Ruth and Helga KOTTHOFF (eds): Communicating Gender in Context. Amsterdam/Philadelphia, 1997. 43. JANSSEN, Theo A.J.M. and Wim van der WURFF (eds): Reported Speech. Forms and functions of the verb. Amsterdam/Philadelphia, 1996. 44. BARGIELA-CHIAPPINI, Francesca and Sandra J. HARRIS: Managing Language. The
discourse of corporate meetings. Amsterdam/Philadelphia, 1997. 45. PALTRIDGE, Brian: Genre, Frames and Writing in Research Settings. Amsterdam/Philadelphia, 1997. 46. GEORGAKOPOULOU, Alexandra: Narrative Performances. A study of Modern Greek storytelling. Amsterdam/Philadelphia, 1997. 47. CHESTERMAN, Andrew: Contrastive Functional Analysis. Amsterdam/Philadelphia, 1998. 48. KAMIO, Akio: Territory of Information. Amsterdam/Philadelphia, 1997. 49. KURZON, Dennis: Discourse of Silence. Amsterdam/Philadelphia, 1998. 50. GRENOBLE, Lenore: Deixis and Information Packaging in Russian Discourse. Amsterdam/Philadelphia, 1998. 51. BOULIMA, Jamila: Negotiated Interaction in Target Language Classroom Discourse. Amsterdam/Philadelphia, 1999. 52. GILLIS, Steven and Annick DE HOUWER (eds): The Acquisition of Dutch. Amsterdam/ Philadelphia, 1998. 53. MOSEGAARD HANSEN, Maj-Britt: The Function of Discourse Particles. A study with special reference to spoken standard French. Amsterdam/Philadelphia, 1998. 54. HYLAND, Ken: Hedging in Scientific Research Articles. Amsterdam/Philadelphia, 1998. 55. ALLWOOD, Jens and Peter Gärdenfors (eds): Cognitive Semantics. Meaning and cognition. Amsterdam/Philadelphia, 1999. 56. TANAKA, Hiroko: Language, Culture and Social Interaction. Turn-taking in Japanese and Anglo-American English. Amsterdam/Philadelphia, 1999. 57 JUCKER, Andreas H. and Yael ZIV (eds): Discourse Markers. Descriptions and theory. Amsterdam/Philadelphia, 1998. 58. ROUCHOTA, Villy and Andreas H. JUCKER (eds): Current Issues in Relevance Theory. Amsterdam/Philadelphia, 1998. 59. KAMIO, Akio and Ken-ichi TAKAMI (eds): Function and Structure. In honor of Susumu Kuno. 1999. 60. JACOBS, Geert: Preformulating the News. An analysis of the metapragmatics of press releases. 1999. 61. MILLS, Margaret H. (ed.): Slavic Gender Linguistics. 1999. 62. TZANNE, Angeliki: Talking at Cross-Purposes. The dynamics of miscommunication. 2000. 63. BUBLITZ, Wolfram, Uta LENK and Eija VENTOLA (eds.): Coherence in Spoken and Written Discourse. How to create it and how to describe it.Selected papers from the International Workshop on Coherence, Augsburg, 24-27 April 1997. 1999. 64. SVENNEVIG, Jan: Getting Acquainted in Conversation. A study of initial interactions. 1999. 65. COOREN, François: The Organizing Dimension of Communication. 2000. 66. JUCKER, Andreas H., Gerd FRITZ and Franz LEBSANFT (eds.): Historical Dialogue Analysis. 1999. 67. TAAVITSAINEN, Irma, Gunnel MELCHERS and Päivi PAHTA (eds.): Dimensions of Writing in Nonstandard English. 1999. 68. ARNOVICK, Leslie: Diachronic Pragmatics. Seven case studies in English illocutionary development. 1999.
69. NOH, Eun-Ju: The Semantics and Pragmatics of Metarepresentation in English. A relevance-theoretic account. 2000. 70. SORJONEN, Marja-Leena: Recipient Activities Particles nii(n) and joo as Responses in Finnish Conversation. n.y.p. 71. GÓMEZ-GONZÁLEZ, María Ángeles: The Theme-Topic Interface. Evidence from English. 2001. 72. MARMARIDOU, Sophia S.A.: Pragmatic Meaning and Cognition. 2000. 73. HESTER, Stephen and David FRANCIS (eds.): Local Educational Order. Ethnomethodological studies of knowledge in action. 2000. 74. TROSBORG, Anna (ed.): Analysing Professional Genres. 2000. 75. PILKINGTON, Adrian: Poetic Effects. A relevance theory perspective. 2000. 76. MATSUI, Tomoko: Bridging and Relevance. 2000. 77. VANDERVEKEN, Daniel and Susumu KUBO (eds.): Essays in Speech Act Theory. n.y.p. 78. SELL, Roger D. : Literature as Communication. The foundations of mediating criticism. 2000. 79. ANDERSEN, Gisle and Thorstein FRETHEIM (eds.): Pragmatic Markers and Propositional Attitude. 2000. 80. UNGERER, Friedrich (ed.): English Media Texts – Past and Present. Language and textual structure. 2000. 81. DI LUZIO, Aldo, Susanne GÜNTHNER and Franca ORLETTI (eds.): Culture in Communication. Analyses of intercultural situations. n.y.p. 82. KHALIL, Esam N.: Grounding in English and Arabic News Discourse. 2000. 83. MÁRQUEZ REITER, Rosina: Linguistic Politeness in Britain and Uruguay. A contrastive study of requests and apologies. 2000. 84. ANDERSEN, Gisle: Pragmatic Markers and Sociolinguistic Variation. A relevance-theoretic approach to the language of adolescents. 2001. 85. COLLINS, Daniel E.: Reanimated Voices. Speech reporting in a historical-pragmatic perspective. 2001. 86. IFANTIDOU, Elly: Evidentials and Relevance. n.y.p. 87. MUSHIN, Ilana: Evidentiality and Epistemological Stance. Narrative Retelling. n.y.p. 88. BAYRAKTAROGLU, Arin and Maria SIFIANOU (eds.): Linguistic Politeness Across Boundaries. Linguistic Politeness Across Boundaries. n.y.p. 89. ITAKURA, Hiroko: Conversational Dominance and Gender. A study of Japanese speakers in first and second language contexts. n.y.p. 90. KENESEI, István and Robert M. HARNISH (eds.): Perspectives on Semantics, Pragmatics, and Discourse. A Festschrift for Ferenc Kiefer. 2001. 91. GROSS, Joan: Speaking in Other Voices. An ethnography of Walloon puppet theaters. n.y.p. 92. GARDNER, Rod: When Listeners Talk. n.y.p. 93. BARON, Bettina and Helga KOTTHOFF (eds.): Gender in Interaction. Perspectives on feminity and masculinity in ethnography and discourse. n.y.p.