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: andreas.jucker@anglistik.uni-giessen.de

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
 TM The paper used in this publication meets the minimum requirements of
8

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 00-051913
ISBN 90 272 5104 5 (Eur.) / 1 58811 023 0 (US) (alk. paper)
© 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
1.1 Reported speech as intention and as perception . . . . . . . . . . . . . . . . 1
1.2 A “thick description” of reporting . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 Some shortcomings of nonpragmatic reductionist approaches . . . . . . . 10
1.4 A method for historical-pragmatic analysis . . . . . . . . . . . . . . . . . . . 16
1.5 Previous applications of the “Method of Residual Forms” . . . . . . . . . 23

C 2
The text-kind: A pragmaphilological overview . . . . . . . . . . . . . . . . . . 27

C 3
Testimony . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.1 The standard reporting strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.2 Strategic choices in the standard tag . . . . . . . . . . . . . . . . . . . . . . . . 42
3.3 Direct speech: Verbatimness or relevance? . . . . . . . . . . . . . . . . . . . 49
3.4 A convention of explicit typicality . . . . . . . . . . . . . . . . . . . . . . . . . 59
3.5 From conduit to collaborator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
viii T  C

C 4
Residual forms in testimony . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
4.1 Participial tags and foregrounding/backgrounding effects . . . . . . . . . . 75
4.2 Alternative tag verbs and different participant status . . . . . . . . . . . . . 85
4.3 Free direct speech and cohesion effects . . . . . . . . . . . . . . . . . . . . . . 92
4.4 Complementized reports: presupposition and grounding effects . . . . . 103
4.5 Fused reported speech and streamlining effects . . . . . . . . . . . . . . . . 116
4.6 Narrative reports of speech acts . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
4.7 Free indirect speech and the boundary with narrative . . . . . . . . . . . . 133

C 5
The question framework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
5.1 The general function of the judges’ reported speech . . . . . . . . . . . . . 151
5.2 Dyadic questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
5.3 Reduction of the tag clause and rapid-fire exchanges . . . . . . . . . . . . 162
5.4 Verba-dicendi tags: turning the tables on the judge . . . . . . . . . . . . . . 165
5.5 Free direct speech and crossplay within a dominant encounter . . . . . . 167
5.6 Stage-managing: The dative with infinitive after speech-act verbs . . . 172

C 6
Reporting from judicial-referral hearings . . . . . . . . . . . . . . . . . . . . . . 177
6.1 The nature of judicial-referral hearings . . . . . . . . . . . . . . . . . . . . . . 177
6.2 Verifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
6.3 Falsifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
6.4 Preliminary reports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197

C 7
Layered reports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
7.1 Preliminary remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
7.2 Additional layers of direct speech: narratives and dyads . . . . . . . . . . 205
7.3 Complementized indirect speech: rejoinders, hearsay, and hypothetical
reports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
7.4 Additional layers of uncomplementized indirect speech . . . . . . . . . . . 228
7.5 Additional layers of fused reported speech . . . . . . . . . . . . . . . . . . . 233
7.6 Two functions of intercalated tags: Cohesion and disclaimers . . . . . . 237
7.7 Narrative reports of speech acts . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
 T  C ix

C 8
Reporting the verdict . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
8.1 Preliminary remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
8.2 Reported speech in the adjudication . . . . . . . . . . . . . . . . . . . . . . . . 248
8.3 Reported speech in the judges’ ratio decidendi . . . . . . . . . . . . . . . . 252
8.3.1 Indirect and nondirect speech . . . . . . . . . . . . . . . . . . . . . . . 253
8.3.2 Fused reported speech . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259
8.3.3 Direct speech . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261
8.3.4 Intercalated verba dicendi and other quotation markers . . . . . 265
8.3.5 Narrative reports of speech acts . . . . . . . . . . . . . . . . . . . . . 270
8.3.6 Free reported speech in the ratio decidendi . . . . . . . . . . . . . 274

C 9
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285
9.1 The purposiveness of reporting in trial transcripts . . . . . . . . . . . . . . 285
9.2 The main functions of the reporting strategies in the corpus . . . . . . . 288
9.3 Some methodological implications . . . . . . . . . . . . . . . . . . . . . . . . . 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 Interna-
tional 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 Languag-
es and Literatures, which provided financial support to enable me to partici-
pate. 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 pragma-
philology 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 pragmatic-

ian studying reported speech, an eavesdropper in time who must find a way
to reconstruct the language behavior of long-silenced reporters and interpret-
ers, 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 typolog-
ical 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, concomi-
tantly, 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 contextu-
alizations 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 recep-
tion 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 chroni-
cles 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/Vološi-
nov’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 discus-
sion). 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 diffi-
culty. 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 consecu-
tive stages in my analysis. First, I set out to establish the norms of distribu-
tion 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 distin-
guishes my investigation from many other studies of reported speech, which
concentrate on belletristic texts that do not feature recurring contextualizat-
ions 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 individu-
ally 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 transliter-
ated according to the ISO/R9 (diacritic) system. The following system is
employed for premodern Cyrillic:

Transliteration Cyrillic Transliteration Cyrillic

a (not after j) n
b o
c p
c¦ r
d s
e š
e¦ šc¦
f t
g u (not after j)
i v
ja x
ju y
k z
ks z¦
l ’
m ”

In excerpts from diplomatic editions, superscript letters are brought down to

the line; abbreviations are expanded, with the tilde (titlo, a diacritic indicat-
ing 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 paragraph-
ing 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 discus-
sion. 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., rešči and reči ‘speak, say’, respectively, the latter is preferred. Verbs are
generally mentioned in their infinitive form. Variable elements in phraseo-
logisms 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 utteran-
ces…” (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 accor-
dance with culturally specific conventions (cf. J. Collins 1987: 71; Silverstein
1985: 134–35, 1993: 55).4 Indeed,  encapsulates three of the major con-
cerns 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 interpreta-
tion 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 inten-
tional and creative character. No matter how closely a report may approxi-
mate 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, conden-
sation 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 introspec-
tion.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 avoid-
ing 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 appropriate-
ness 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 compe-
tence 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 genre-
based “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; Hick-
mann 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). Cross-
linguistically, 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-/’), report-
ers 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 introduc-
ers”, 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 — soobščit’ ‘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 realign-
ments in the dialogue: vstupit’ ‘enter in’ (2×), načat’ ‘begin’, prodolžat’
6 R V

‘continue’, and obratit’sja ‘turn to.’ Finally, there are verbs of mental
activity — predpoložit’ ‘surmise’ (2×) and zaključit’ ‘conclude’; manner-of-
speaking verbs — vskričat’ and voskliknut’ ‘exclaim’; and an attitudinal verb
— nadut’sja ‘pout, sulk.’
It is inadequate to explain diversity of this kind as mere “elegant varia-
tion” (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 afore-
mentioned Semejka was killed in the Skewbald Horde (1648;
Tokarev (ed.) 1970: 896).
b. [V”spros” Inokentiev: “Gosudar’ Pafnotej!
[question Innokentij-]- [lord Pafnutij]-
Poveli… napisati zavěščanie 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 Lixačev (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 buděte mně v” s(y)ny i
and you-. be-.2. me- in sons-. and
dščeri 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 text-
kinds, including legal-administrative writing, when the complements are
presupposed — e.g., from preceding stimuli-questions.) By contrast, interca-
lation 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-.
čto, 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 recol-
lection. (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 complica-
tion 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’evič’ skazyvaet”: na čem” k”
[prince grand Ivan Vasil’evič]- say-.3 on what- to
vam”, k” [svoej otčině, rekl”, po tomu
you-. to [. patrimony]- speak- by that-
vas”, skazyvaet”, žaloval”…
you-. say-.3 favor-
Grand Prince Ivan Vasil’evič 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 predi-
cate, 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… čto [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 synony-
mous. 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 dis-
course, 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 contextuali-
zations (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 self-
quotations 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 individ-
ual 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, follow-
ing 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;
Padučeva 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 — approxi-
mately, “in re ‘X’”.
12 R V

(7) čto dei po těx 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 dě(i) (in the
reduced form dei), a grammaticalized third-person singular nonpast or
imperative of the archaic  děti or dějati.16 The information topicalized in
the č’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 distin-
guishing 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, third-
person 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 sjlj7́dujuw 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 intonat-
ional 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 com-
pound 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 “incorpo-
rated” 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 third-
person 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 peo-
ple [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). Approach-
es 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 pragmati-
cally 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” (Ander-
sen 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 historical-
syntactic 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 explana-
tions, 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 opera-
tive 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 spoken-
language 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 institution-
alized 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 under-
stand 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 informa-
tion 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 represent-
ed 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 estab-
lish 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 construc-
tional 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 informa-
tion, 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 strate-
gies 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 conven-
tional 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 enfranchise-
ment 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 success-
ful 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 strate-
gies 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 methodolo-
gy 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 partici-
pants 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 low-
frequency 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 conceiv-

ably be rational in the context.23 Indeed, generally speaking there would
seem to be no particular justification for blaming residual forms on incompe-
tence unless there are other, independent signs of it, e.g., startling departures
from the diplomatic norms of the text-kind or (perhaps) grammatical mis-
takes 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 frequen-
cy 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 Bloom-
field (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 unex-
pected 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 analy-
sis of premodern discourse, the only way to gauge whether strategies were
automatized is to find what seem to be inappropriate uses (infelicitous exten-
sions 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 Silver-
stein, is its use of quasi-Praguean binary oppositions (“marked” vs. “un-
marked” 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 Silver-
stein’s approach. I know of no way in which to measure the relative simplic-
ity 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 gram-
mar, 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 foreground-
ing 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 text-
kind. 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 (“individual-
ity”), 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 testimo-
ny 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 word-
for-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 estab-
lishing 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 docu-
ments 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 primari-
ly 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 them-
selves 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 identify-
ing 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 indicat-
ing 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 tran-
scripts 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 tran-
scripts 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 investi-
gated 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 Fedorovič(’).
[this trial]- judge- [Mixajlo Fedorovič]-
† Mixailo Fedorovič [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 presup-
posed 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 plain-
tiff’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 rěčku v Vašku, lovjat, g(ospodi)ne,
and in stream- in Vaška- fish-.3. lord-
i… saki v rěčkě sačjat, 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 fish-
ing, lord, and dragging the stream with… nets, without informing
me or my steward” (ibid.)
b. I [Mixailo Fedorovič(’) vsprosil Vašuty i eg(o)
and [Mixajlo Fedorovič]- ask- Vašuta- and his
tovariščov: Otvěčaite! I Vašuta tak rek:
fellows-. respond-. and Vašuta- thus speak-
Doselě 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 ozerě i na rěčkě; a nyniča,
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 rěčku v Vašku ne vstupaemsja;
his nor in stream- in Vaška-  trespass-.1.
a to, g(ospodi)ne, vědaem, čto [to ozero
and that- lord- know-.1. that [that lake]-
da rěčka… c(e)rkovnoe…
and stream- church-
 T T-K 33

And Mixailo Fedorovič 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 judicial-
referral 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 Fedorovič(‘) [arximandrita Feofana
and by that- [Mixajlo Fedorovič]- [abbot Feofan]-
opravil, a Vašutu… obinil i prisudil
find.right- and Vašuta- find.wrong- and award-
34 R V

ozero i [rěčku Vašku [s(vja)t(o)mu Mixailovu

lake- and [stream Vaška]- [holy Michael-
Čjudu [arximandritu Feofanu.
miracle]- [abbot Feofan]-
And, in accordance with that, Mixailo Fedorovič [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 docu-
ments the ’s are elaborated by a statement of the judge’s ratio deci-
dendi. 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 text-
kind. 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 second-
person 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 report-
ing 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 proper-
ty; it is elicited by the judge(s) in a face-to-face exchange, presumably in
some standard configuration determined by judicial custom. The prototypic-
ality of this speech event stems from the fact that, under ordinary circum-
stances, 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 poly-
syndetic 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 text-
kinds 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  reči ‘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 Ysačkom s
litigate- [Savko Mandakov]- with Isačko- with
[Vaskovym s(y)n(o)m” Obnorin(a)… To, g(ospodi)ne, [otvotčiki
[Vas’ko- son]- Obnorin-  lord- [rentmen
kirilovskie… otveli u menja, u Savki,
of.Cyril]-. take.away-. at me- at Savko-
počinok k” [svoei pustoši… x Kočevinskoi…
clearing- to [. empty.land]- to Kočevinskaja
Savka Mandakov… sued Isačko, son of Vasko Obnorin… “The rent-
men of Cyril [Monastery]… took away from me, Savka, a clearing
[and annexed it] to their uncultivated property of Kočevinskaja…”
(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 repre-
sented 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” tovarišči “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
Čekmakovskoj s [Putilovskoju zemleju meži.
Čekmakovskaja]- 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 Čekmakovskaja proper-
ty 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 pronom-
inal subjects. The referent of the pronoun in the underlined clause is ambigu-
ous; 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 partici-
pation 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” [vsěx” kr(e)st’jan
thus speak- Stepanko-  in [all peasants]-.
město…
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 mežě, 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 sudě…
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 witness-
es 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 exam-
ple, 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
 reči could take an indirect object (or, in ecclesiastical writing, a preposi-
tional 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 contribu-
tions 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 them-

selves. 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 “presup-
posed” (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 preemp-
tively 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 orienta-
tion 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 čto [tě pustoši izstariny za šest’desjat lět
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, čto [tě 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 ondrěevskie.
them- for [prince- lands]-. for Andrej-..
… because those fields of old, over sixty years, belonged to the vil-
lage 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
reči), 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 Novgorod-
ian 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 tran-
scripts, 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, descrip-
tive, 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 Čerepnin 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 Gvozdanović 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 institu-
tional 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 third-
person 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.
 T 49

Table 1. Distribution of the vocative g(o)s(podi)ne by clause type

# Vocatives Total %
All clauses 1224 1787 68.5
Initial clauses 0434 0448 96.9
Noninitial clauses 0790 1339 59.0
Second conjunct, same-subject 0157 0240 65.4
Second conjunct, different-subject19 0532 0716 74.3
Preposed subordinate clauses 0031 0033 93.9
Main clause after preposed subordinate clause 0020 0031 64.5
Postposed relative clauses 0004 0047 08.5
Postposed temporal, clausal, adverbial clauses 0002 0014 14.3
Complement clauses after cognition verbs 0012 0086 14.0

3.3 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 text-
kind-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 reči), 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 docu-
ments (see Chapters 6–8), occasionally even in the trial record (see Chap-
ter 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 “repro-
ductionist” (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” (non-
verbatim) form.24
Memory constraints do not, of course, exclude the possibility that
reporters, in using , are operating under an illusion or pretense of verba-
timness. 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 utter-
ance. 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 abbreviat-
ed 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 co-
constructed 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 treat-
ments 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 nontech-
nical 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 interpreta-
tion 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 poten-
tially issue denials). This latitude suggests that “what counts as the same” is
not only culture- but also genre-specific. Even in text-kinds where punctili-
ous 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 verba-
timness (or commitment to form and content) is not a fundamental principle
but itself a reading convention for  in certain written genres (e.g., quota-
tions 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 represent-
ed 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 report-
er-commitment model; there can be contextual motivations not only for
emphasizing verbatimness but also for ruling it out as a possible interpreta-
tion.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 “faith-
fulness 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 conceptual-
ization 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 instruc-
tions (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-the-
board commitment to fidelity in form and content. The authoritative religious
texts of Eastern Orthodoxy provided models for wording changes and para-
phrases 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) Napečatano: [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 technologi-
cally 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 out-
pacing them.
Consequently, the scribes who wrote the trial records must have (re)con-
structed 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”
černoj prjamoj). Both of the trial judges denounced this as a forgery, and the
grand prince refused to withdraw his original verdict (N. P. Lixačev (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 Ameri-
can 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 59

3.4 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 , purport-
edly 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 proper-
ly. 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 tovarišči tako rkli… I
and Mamon- with fellows-. thus speak-. and
 T 61

[knjazja velikogo znaxori [Ontipa desjatckoj s

[prince grand]- witnesses- [Ontipa tenthman]- with
tovarišči 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 sense-
making 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 česti; 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(’)evič(’) vozrěl v gramotu, i v gramote
and [Ivan Vasil’evič]- look- in writ- and in writ-
62 R V

pišet…
write-.3
And Ivan Vasil’evič 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 narra-
tion 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 informa-
tion 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 predict-
ability 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 Andrěevič(’) v”zrěl” v” [Olešinu
and [prince Mixajlo Andreevič]- look- in [Oleša-
danuju gramotu; ož v gramotě pišot tak:
donation writ]-  in writ- write-.3 thus
[Tě svoi zemli Oleša dal [živonačalnoi
[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 lěsy, i
village]- and [Gridina village]- and with forests-. and
s požnjami. A poslus(i) v [danoj gramotě:
with fields-. and witnesses-. in [donation charter]-
[kn(ja)z’ Semen Fedorovič(’) Starodubskoj, da [Timofěi
[prince Semen Fedorovič 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 Andreevič 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 Fedorovič 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 counter-
example 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 devia-

tions, 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 subordi-
nated” (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 some-
thing outwardly given, but a free project of the spirit: the apparent ‘reproduc-
tion’ (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. Lixačev 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, “Produc-
tion 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 function-
oriented studies that have correctly emphasized its capacity to create vivid-
ness, immediacy, or involvement, nor, indeed, does it rule out the fidelity-
based (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 complemen-
tized , 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 sanc-
tioned 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 demon-
strations 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 neces-
sarily 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 coher-
ence).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 utter-
ance.” 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 terminolo-
gy, 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, “engross-
ment” — 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 undistin-
guished, 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 back-
ground to interpret the same way they do” (Clark and Gerrig 1990: 793).
Thus Brown and Levinson (1987: 122) treat  as part of the positive-
politeness 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 hearer-
based 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-Coult-
hard 1994: 304). This concept of narratorial interference or control has been
applied not only to reports of actual anterior utterances but also — presum-
ably 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; howev-
er, 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 partici-
pant 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 quota-
tion 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 address-
ee 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 responsi-
bility” (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 ecclesiasti-
cal 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 convey-
ing 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 “report-
ability” (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 super-
vised 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 by-
product of the effort) as by providing the means by which an adequate
image of the trial could be (re)constructed — one may almost say, reconsti-
tuted — 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 signifi-
cant, 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 represent-
ed 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 accor-
dance with culturally determined interpretive conventions, allows the
audience to participate both in the new event (the re-creation) and, vicarious-
ly, 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 produc-
ing 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  reči ‘speak, say’ (1), instead
of the preterit found in the standard tag:
(1) a r(’)kuč(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 con-
veyed 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(’)kuči is so common that
it has been treated as a lexicalized quotative (Barnet 1965: 104–6; SRJa, s.v.
“arkuči”). There are reasons to doubt this claim if it is meant imply that a
r(’)kuči was a citation marker divorced from the paradigm of the verb reči.
Univerbation cannot be proven, given the use of scriptio continua in ; the
putative quotative does not occur in modern dialects. Moreover, r(’)kuči
tends to be preposed, whereas clear-cut  quotative particles like dě(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’kuči could be modified by the adverb tak(o); adverbial modifi-
cation would be unlikely if it were already a quotative particle or other non-
predicative element. Second, r’kuči 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’kuči 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 introduc-
tory 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 newswor-
thy. The combined effect may be to suggest a lesser degree of newsworth-
iness 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 contextuali-
zation, 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 boundary-
setting 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 tempo-
rarily 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 event-
like 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 tovaryšči napravo Kolyčevskoju
... with fellow-. to.right [Kolyčevskaja
dorogoju, da z dorogi nalevo [poperečnim putem, a
road]- and from road- to.left [transverse road]- and
rkuči: 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, čto
 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 tovarišči
witnesses.. [El’ka S’janov]- [. fellows]-.
sud’i tako rekli: Čto esmja, gospodine,
judge. thus speak.. what ..1. lord-
veli… ino napravo zemlja i les Andreeva
lead-.  to.right land- and wood- Andrej-.
Okljačeeva, [Isxodckogo sela, a nalevo zemlja, i
Okljačeev- [Isxodskoe village]- and to.left land- and
les, i pašnja, i požni, i selišča
wood- and field- and fields-. and hamlets-.
Prečistye [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 Kolyčevskaja Road, and from the road to the left by a trans-
verse 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 Okljačeev’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’kuči and once, at the very end, with
tak(o) rek(li) (AFZX 225–26, no. 258). An earlier boundary-setting descrip-
tion 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’kuči,
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 background-
ing, 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 rkuči 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 recon-
venement … 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 potential-
ly 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 čel”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 čelom [kn(ja)zju Mixailu Andrěevič(ju) [igumen
lo hit- forehead- [prince Mixailo Andreevič]- [Abbot
Pasěja [živonačalnye Troici Sergěeva
Paseja]- [life-giving Trinity]- [Sergius-
monastyrja, da igumen [Kirilova monastyrja
Monastery]- and Abbot- [Cyril- Monastery]-
Nifont”, a rkuč(i) [igumen Pasěja 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-
čelom: A u nas v [Kirillově monastyrě [togo
forehead- but at us- in [Cyril- monastery]- [that
Oleši [ot(e)c’ Ofonasei postrig”sja v černci…
Oleša]- [father Ofonasij]- be.tonsured- to monks-.
[1] Abbot Paseja of the Life-giving Trinity-Sergius Monastery peti-
tioned Prince Mixail Andreevič, 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 čel”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 plain-
tiff’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 tenden-
cy 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 present-
ing 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 necessi-
tates 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-linguisti-

cally 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 reči. 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 reči, which appears in the standard tag clause ((1b) in Sec-
tion 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 reči that appear in the earliest version are regularly replaced by
other ’s in the later redactions (Pokrovskij (ed.) 1971: 27). If the assess-
ment of the first foreign grammar of Russian (Ludolf 1696/1959: 5) is
accurate, reči (ecclesiastical-register rešči) 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 tran-
scripts 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 instabili-
ty and hence perhaps of relaxed conventions in legal language.7 Although reči
continued to appear in conservative documents, s(”)kazati came to predomi-
nate 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 reči

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 reči. However, two facts suggest that this explanation, while not necessar-
ily 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
reči 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
reči 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-polite-
ness 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
cělovan’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  reči. 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 Menčakova vyšli…
we- lord- from Menčakovo- 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 Menčakov…”, etc.
(ASÈI 2: 534, no. 492).
The first of the responses, in (6) [1], illustrates the fact that the standard 
reči 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 metaprag-
matic 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 reči 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 state-
ment 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’evič(’) poslal k” mitropolitu
and [prince Ivan Jur’evič]- send- to metropolitan-
[s(y)na svoeg(o) kn(ja)zja Ivana, velěl ego sprositi,
[son . prince Ivan]- bid- him- ask-
on li posadil [těx mužikov, Okulika da
he-  settle- [those peasants]-. Okulik- and
Olferka, na tom na [Šiškinskom selcě, 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’evič(u) tak: [To
Zosima]- answer- [prince Ivan Jur’evič]- thus [that
seliščo Šiškinskoe — zemlja [Simanovskog(o) manastyrja…
village Šiškinskoe]- land- [of.Simonov monastery]-
And Prince Ivan Jur’evič 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’evič as follows: “That village Šiškinskoe is land of Simonov Mon-
astery…”, etc. (ASÈI 2: 411, no. 402).
Here the judge, verifying an assertion made during the hearing, is communi-
cating 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 continua-
tion 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 reči rarely does so in the transcripts investigat-
ed (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 purpo-
sive use of language rather than scribal ineptitude; there is a strong correla-
tion between departures from the genre-specific narrative norm and contextu-
alizations 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 reči 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 foregroun-
ding, 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 connota-
tions 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 tran-
scripts 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 would-

be principal. (Uninformative or otherwise uncooperative speech can, of
course, also be tagged by the neutral  reči, 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 fore-
grounds 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” vsěx”
and judge- ask- Stepanko-  in [all
kr(e)st’jan město: Komu ž to u vas
peasants]-. place- who-  that- at you-.
vědomo, [dobrym ljudem, starožilcem? Vědomo
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
vědomo 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 seventeenth-
century 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 polěs’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 města,
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 [mitropoliče 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 metro-
politan’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. napravě — zemlja (NP-) NP-/, a nalevě —
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 boundary-
setting 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 boundary-
setting descriptions. While uninterrupted themes are typically, though not
invariably, indicated by zero in , a theme that is interrupted by  (includ-
ing ) 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 subdivi-
sion 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 move-
ments as manifestations of a single continuous instantiation of testimony.
The same kind of grouping effect may be seen in a further contextuali-
zation of , which also involves what can be viewed as continuous multi-
modal 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, lěs
and Fegnast- thus speak- for that- lord- wood-
u nas [gramota danaja [kn(ja)zja Mixaila Ondrěeviča…
at us- [charter donation]- [prince Mixail Andreevič]-
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
spiskě pišet: [text] I tot, g(ospodi)ne, lěs…
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 Andreevič… [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 multi-
modal 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, č’e to seliščo, 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 lět… 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 lět… 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-
lět… A [Jurka Melexov tako rek: A jaz,
years-. and [Jurka Melexov]- thus speak- and I-
g(ospodi)ne, pomnju za polpjatadesjat(’) lět… I
lord- recall-.1 for 45- years-. and
sud(’)ja v”sprosil Oleksěika…
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 co-
witnesses (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
tovarišči: Skažite v [bož’ju pravdu, č’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
tovaryšči 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 [Plotniča sela [Bortnikovy
of.old of.archangel- [Plotniče 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, čto [ta
know-.1. for that.many- years-. that [this
zemlja arxangel’skaja [Plotniča sela [Bortnikovy
land]- of.archangel- [Plotniče 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 Plotniče. [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 Plotniče…” [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 ambigu-
ity 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 cohe-
sion: 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 distribu-
tion 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 multi-
modal 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 boundary-
setting 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 confronta-
tion (130–32, 138–41). Most of the cases of  in trial transcripts may be
viewed conversely as non-confrontational, especially given that the represent-
ed 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 contextuali-
zations (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 contextu-
alizations. All of the examples feature the complementizer č’to ‘that’ (usually
spelled čto 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 back-
shifting of tense or automatic (matrix-dependent) change of modality.
Explicit subordination, though common, is neither a necessary nor a suffi-
cient 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 less-
than-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 interpre-
tation, corollary to the reproductionist approach to  (see 3.3). The state-
ments presented as  were by and large as accessible to the scribes’ observa-
tion 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 nonproto-
typical speech events and unconventional (or less conventional) strategies as
in other residual forms. For example, one of the tokens represents a state-
ment 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 Ivanovič 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 Borisoviča, a gramotu mi
[donation writ]- [Vasilij Borisovič]- 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, čto 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 Borisovič, 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 contextu-
alization 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 Ikonniče zemle. I Elizar
Timoška- ask- about [Ikonniča land]- and Elizar-
skazal, čto [ta zemlja Ikonniče… po [Spasskuju dorogu
say- that [that land Ikonniča]- up.to [Spasskaja Road]-
protivu Gorbova [zemlja mitropoličja Kulikovskaja, a
opposite Gorbovo- [land of.metropolitan Kulikovskaja]- and
kupil [Ivan Fedorovič u Mixaila do moru,
buy- [Ivan Fedorovič]- at Mixajlo- before plague-
a v knigax pisana v dannyx [Danila
and in books-. write- in of.donations-. [Danilo
Ivanovičja bojarskaja zemlja, a ne [velikogo knjazja.
Ivanovič]- [boyar land]- and  [grand prince]-
And the judge sent Ivan Turab’ev and Timoška to Elizar to ask about
the Ikonniča property. And Elizar said that that property Ikonniča… as
far as the Spasskaja Road opposite Gorbovo belongs to the metropol-
itan’s property of Kulikovskoe, and Ivan Fedorovič bought [it] from
Mixajlo before the Plague, and in the donation books of Danilo
Ivanovič 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 (Town-
send and Bever 1977: 5). In corroborations, as opposed to ordinary testimo-
ny, 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 predictabil-
ity of corroborations obviates one of the main functions of  in the tran-
scripts — 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 tran-
script, 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 Mikulič 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: čto 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, ezdjači, [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 corrobo-
ration 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 state-
ment 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, čto 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, grant-
ed 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 busi-
ness (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 state-
ment 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 overlap-
ping information — what the witnesses remember — is presented collective-
ly 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 Čej 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, čto
Semen]- and [priest Ivan]- say-. that
pomnjat to do Suzdalskogo boju za
recall-.3. that- before [of.Suzdal battle]- for
desjat’ let, čto [tot ogorod i [ta požnja
ten- years-. that [that garden]- and [that field]-
[troeckago dvora. A Isačko skazal, čto
[of.Trinity farmyard]- and Isačko- say- that
pomnit za 50 let, a Rodivon
recall-.3 for 50- years-. and Rodion-
skazal, čto pomnit za 60 let, čto [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 Monas-
tery’s farm. [2a] And Isačko 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 cross-
kissing oath” (see 4.2), clerics “by the priesthood” (po svjaščen’stvu ‘by
priesthood-’).
The underlined passages in (17a) are in unambiguous , with the
subjects of the tag clauses coreferential with the implicit subjects of third-
person 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
economical reporting mode (cf. Čumakov 1975: 19, 26) because it eliminates
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 witness-
es’  — 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 represent-
ed 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 Ste-
panko 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;
Martysevič 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 comple-
mentizer č’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 complemen-
tized  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 ‘foregroun-
ded’… 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 pre-
ferred 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 testimo-
ny presented as  in (17b), which could be characterized as “hyper-fore-
grounded” 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 interpret-
ed differently — as the beginning of a long-term stylistic change, like the
cases of the  s(”)kazati where the standard tag verb reči 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 ssyločnye
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 state-
ments 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 expect-

ed 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 context-
ualizations 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 boundary-
marking 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 homo-
phonous, 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 infor-
mative (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 [mitropoliči 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 Kučeckogo,
Konstantin- with [Ivan- children]-. Kučeckij-
a Fedor da Kostjantin otvečivali 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 Kučeckij; 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 posses-
sive adjective that modifies it (“svoego”) is reflexive because it is coreferen-
tial 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 informa-
tion 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 [mitropoliči xristiane [Karinskogo
rick- and [metropolitan- peasants]-. [Karinskoe
sela [Senka dvornik da [Gračko Paxomov
village]- [Senka tenant]- and [Gračko 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 Gračko 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 descrip-
tions. 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, čto 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 prosečeno. 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; similar-
ly 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 back-
grounded 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, čto [te lugi
who- be.known that [those meadows]-.
tjanut k selu k Menčakovu? I
belong-.3. to village- to Menčakovo- and
Dym skazal [svoix starožilcev, menčakovskix
Dym- say- [. old.residents of.Menčakovo
xrest’jan [Trofimka Fomina… na tex sja,
peasants]-. [Trofimko Fomin]- to those-. 
gospodine, xrest’jan šljusja, čto [te
lord- peasants-. refer-.1 that [those
lugi tjanut k [selu Menčakovu
meadows]-. belong-.3. to [village Menčakovo]-
And the judges asked Dym, “Whom do you have that knows that
those meadows belong to the village of Menčakovo? And Dym said
his witnesses were the Menčakovo peasants Trofimko Fomin [et al.]:
“I call those peasants, lord, [as witnesses] that those meadows belong
to the village of Menčakovo” (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 Wierz-
bicka’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. Structur-
ally 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 inter-
preted 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 informa-
tion 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 repor-
tedness in an explicit manner. The lack of “gear-shifting” (Page 1988: 33)
between  and narrative increases the pace of interpretation. This acceleran-
do 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 con-
struction 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 “informa-
tion 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 dis-
course 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, plot-
advancing 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 particu-
lar,  — 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 them-
selves 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 narra-
tive. 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 imped-
ed by grammatically more independent and hence more interruptive strate-
gies. 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 () velěti ‘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 velěti
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-.
zapečatat’… 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 them-
selves appeared…, but the second haystack was not on the stubble-
field; 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”ezščoj
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 rozvodščiki… 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 rozvodščiki 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 mon-
astery’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’, čto [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- [.
čeloveku krest pocelovat’. I kak na sroce k celovan’ju,
man]- cross- kiss- and when on term- to kissing-
i Dmitrej [čeloveku 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 thick-
et 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 suffi-
cient, 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 incommensu-
rate 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 testimo-
ny-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 metro-
politan; the litigants are proxies for the prince of Belozersk and the archbish-
op 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 učinil 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 tax-
es?” 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 učiniti dovod” (approxi-
mately, ‘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 expres-
sions are rare in the trial record — predictably, given the covert style of
narration; they occur mostly in accounts of no-shows.33 Given its ecclesiasti-
cal 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 iščeiny,
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 presup-
posed 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 (čto…)
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 strate-
gies, ’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 (admit-
tedly 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 informa-
tion 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 133

4.7 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 counter-
examples — as a primarily or exclusively literary phenomenon of post-
medieval 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 subse-
quent 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 ostensi-
ble features of the strategy. (The most frequent English alternative, repre-
sented 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 “parenthet-
ical” (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; Slem-
brouck 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  (Padučeva 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” (Thomp-
son 1996: 514). This likens  to proverbs and allusive (“hidden”) quota-
tions; 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 knowl-
edge 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 nine-
teenth 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 errone-
ously, 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 . Padučeva (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 uncom-
plementized, 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 tran-
script 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 pčely, i on de s [temi pčely [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, čto u menja pčely
men-.  deny- that at me bees-.
ukral; a [te pčely evo. A muži,
stole-. and [those bees]-. his and men-.
gospodine, moi i Fedorcevy, Fedorca v
lord- my-. and Fedorec-.. Fedorec- in
[moix pčelax 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 subject-
controlled reflexive possessive in the predicate and by subsequent statements
of Fedorec and the witnesses (ibid.). Reportedness is indicated by the
quotative particle dě(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 predica-
tive 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 docu-
ment 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 conven-
tion 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 polič’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 present-
ed 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  confes-
sion 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
reči 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, čto on u Kuzmy
them-  deny- that- that he- at Kuz’ma-
pčel ukral, i s [temi pčelami
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 rěči ‘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 immedi-
ately 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 ‘de-
nounced’, 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 recollec-
tion of the represented speakers) rather than “narrative preterit” (cf. Padučeva
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 ostensi-
ble transition at all.
Free indirect or nondirect speech can also be used to integrate hetero-
glossic 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 čern’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 observ-
able; 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).
Padučeva’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 boundary-
setting descriptions — the culmination of the tendency to integrate  into
the narrative in this contextualization.46
(29) I poveli [xr(e)st’jane černye [Nazarik Senkin
and lead-. [peasants black]-. [Nazarik Senkin]-
s tovarišči… 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 [selišča

stump]- on.right land- meadow- [village
Kermedinovskog(o) i [selišča Čovyrevskogo [černaja
Kermjadinovskoe]- and [village Čevyrevskoe]- [black
zemlja, čto ěe [simonovskye xr(e)st’jane
land]-  her- [of.Simonov peasants]-
pašut; a naleve [zemlja černaja [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 compan-
ions 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 Čevyrevskoe, 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 land-
marks 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; Haber-
land 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 custom-
ary trial procedures and the conventions of the text-kind; they could identify
the source of the evaluations from the context, which concerns the move-
ments 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 individ-
ual 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). Evident-
ly 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 agree-
ment. 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 partici-
pants (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 preposi-
tional 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’evičem o [cerkovnoju zemlju, čto
[Feodor- son Jur’evič]- about [of.church land]- 
 R F  T 147

pridal [Bogdan Kurunbeev na rečki na Tjuševki

give- [Bogdan Kurunbeev]- on stream- on Tjuševka-
[Velikomu Bogojavlen’ju. I [Feodor Jur’evič [tu zemlju
[great Theophany]- and [Feodor Jur’evič]- [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’evič [tu zemlju
leave- [this world]- [Feodor Jur’evič]- [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’evič sprosil Dmitrija:
Feodor- and [Feodor Grigor’evič]- ask- Dmitrij-
Otvečaj!
respond-
The Theophany [Monastery] abbot Iona sued Dmitrij, Feodor
Jur’evič’s son, over church land that Bogdan Kurunbeev had donated
to Great Theophany on the Tjuševka River. And Feodor Jur’evič
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’evič 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’evič 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 prop-
erty 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 condensa-
tions 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 unimpor-
tant 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. Padučeva 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 proper-
ties 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 bound-
ary-setting descriptions (29) or incipits (30)–(31) — or because it is unim-
portant per se — as probably in (28), where semi-predictability may also be
a factor. The omission of attribution and ongoing-event perspective informa-
tion 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” (Klei-
mola 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 empow-
ered to render verdicts; in many cases they were restricted to gathering
evidence, which they had to refer to a court of higher instance (see Chap-
ter 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 tran-
scripts, 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’eviča ‘Prince Boris Vasil’evič’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 undoubt-
edly 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 reči,
rassprosnye reči) 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 sep-
arating 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’evič(’) sprosil” [manastyr’skix
and [prince Ivan Jur’evič]- ask- [of.monastery
154 R V

znaxarei… I [Timoxa Denisov tako rek…

witnesses]-. and [Timoxa Denisov]- thus speak-
A Levon tako rek… A [Ivaško Medvěd’ da
and Levon- thus speak- and [Ivaško Medved’]- and
Foka tako rkli… I [knjaz(’) Ivan Jur’evič(’)
Foka thus speak-. and [prince Ivan Jur’evič]-
sprosil tex žo [manastyr’skix znaxarej…
ask- those-.  [of.monastery witnesses]-.
I Timoxa i eg(o) tovaryšči tak rkli…
and Timoxa- and his fellows-. thus speak-.
And Prince Ivan Jur’evič 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’evič 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 higher-
court 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 ubiqui-
tous 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 possibil-
ity 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 interpret-
ers, 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 topical-
ized in the questions became quickly processed givens in the answers; this
focused attention on the less expected parts of the replies, the new informa-
tion (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 limita-
tions 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 tran-
scripts (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 dě(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 uncomple-

mentized  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 interpre-
tation. 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) tovariščov:
and judge- ask- Kazak- and  his fellows-
Otvečaite!… I sud’ja vs”prosil [stroitelja Ieva:
answer-. and judge- ask- [abbot Iev]-
Počemu žo ty zoveš(’) [tě 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) tovariščov: A vy
Kazak- and  his fellows- and you-.
zovete [tě požni [velikog(o) kn(ja)zja,
call-.2. [those fields]-. [grand prince]-
komu to u vas vědomo…
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: Otvěčai!… I sudia
and judge- ask- Vasko- answer- and judge-
sprosil [starca Semiona: Počemu 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 počemu

ask- Vasko- and you-. [that field]- why
zoveš” [Gorodišč”skoju požneju?
call-.2 [of.Gorodišče 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 Gorodišče’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 lower-
court 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 Galič (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 Galič 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 ešče!
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 imper-
fective 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 [iščeja 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-.
otvetčikov… 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 judg-
es 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 modali-
ty 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 modera-
tors 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 Chap-
ter 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’ Ikonniče
and judge- ask- Korovaj- call-.2 Ikonniče-
[velikogo knjazja, a kosil ego xto [to dvoriščo
[grand prince]- and mow- it- who- [that farm]-
da i [te luzi pod dvoriščom? I Korovaj
and  [those meadows]- below farm- and Korovaj-
tak rek: kosili ego, gospodine, [ljudi
thus speak- mow- it- lord- [people
mitropoliči. 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 dvorišče da i [te luzi
mow-.3. [that farm]- and  [those meadows]-
mitropoliči, 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 molčiš’, 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 Ikonniče 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 peo-
ple, lord, have been using it.” [2] And the judge asked, “Have they
been using it long?” And Korovaj [spoke] as follows: “The metropol-
itan’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 idiosyn-
cratic 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-
vědomo… I Kur’an” tak rek: Vědomo, 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-
č’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 partici-
pants (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’ Ikonniče zemlja [velikogo knjazja
call-.2 Ikonniče- land- [grand prince]-
počemu? I Korovaj tak rek: vedomo, gospodine,
why and Korovaj- thus speak- be.known lord-
[ljudem dobrym, čto [to Ikonniče zemlja [velikogo
[people good]-. that [that Ikonniče]- land- [grand
knjazja. I sud’ja vsprosil: komu vedomo, čto 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 Ikonniče
is the grand prince’s land?” And Korovaj spoke in this way: “It is
known, lord, to good people that that Ikonniče 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  reči is semantically
broad enough to tag any kind of reported sentence, it is associated proto-
typically 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  reči, 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… selišča [monastyrskie
only-. lord- [these villages]-. [of.monastery
xrest’jane pašut…, pašut [tex
peasants]-. plow-.3. plow-.3. [these
selišč mnogo… i my tebe [te vse
villages]-. many- and we- you-. [those all
selišča 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’ oprič’ [tex dvu selišč… 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 wit-
nesses: “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 situa-
tions 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 reči, 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 respon-
dent, 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 present-
ed 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 prop-
erty, 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: Počemu 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: Počemu 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 Sec-
tion 5.2 and (8) in Section 5.3).
172 R V

5.6 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 orienta-
tion. 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  velěti ‘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 velěti 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 parti-
cipants’ 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 velěti overlaps
with  is in orders initiating boundary-setting procedures. Here the illocuti-
onary intent may involve both eliciting testimony and directing the partici-
pants’ 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 unexpect-
edly 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 tovariščy, 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 seliščem meži
prince]- land- [those villages]-. borders-.
ukazati. I Markel tako rek: Jaz na [te
show- and Markel- thus speak- I- to [those
selišča ne poslan… a s toboju na [te
villages]-.  send- and with you-. to [those
selišča 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 wit-
nesses 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 situa-
tion 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 rečisja 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] stew-
ard, 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 velěti ‘command’ or, when appropriate, after the phrasal verb dati/
učiniti 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 učinil srok stati pered [velikim
witnesses-. make- term- appear- before [grand
knjazem o [tem dele… a čto na [tom selišče
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 selišča 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 Chap-
ter 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 Aleksandrovičem sud(’)i

before [prince Danilo Aleksandrovič]- 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 Aleksandrovič(’), vyslušav spisok, vsprosil [oboix
Danilo Aleksandrovič]- 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-.
čto im [sud takov byl, kak v [sem spisku.
that them- [trial such]- be- as in [this copy]-
[1] Before Prince Danilo Aleksandrovič, the judges Mixajlo Šapkin
and Golova Semenov placed this [trial] copy and produced both liti-
gants… [2] And Prince Danilo Aleksandrovič, 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 proceed-
ings [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 judicial-
referral 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 essen-
tially 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 verifica-
tion 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”kě
X- [such trial]- be- as in [this copy]-
(pisano)
(write-)
X had a trial such as is written in this copy.
b. X- [takov” sud” byl”, kakov” s”pis”k”
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 inani-
mate 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 impos-
sible 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 [iščeja Toropko skazal, čto 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
[otvetčik 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 [iščeja Karpik, i v [brata svoeg(o)
and [plaintiff Karpik]-  in [brother .]-
město v Fedkovo, i, v [starcevo město
place- in Fedko-. and in [elder- place
Aleksandrovo, [Mitja Galasěev skazali, čto
Aleksandr-]- [Mitja Galasěev]- say-. that
im [sud takov byl, kak v [sem spiskě
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 Monas-
tery 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 corrobora-
tion 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 tauto-
subjective 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’ state-
ment is partially presupposed; it was not expected to be more, in essence,
than yes or no. Presupposition often favors complementization and indirect-
ness (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 promi-
nent 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, čto 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 zakaščik 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 Ivanovič(’) velěl pered
you-. and [prince Vasilij Ivanovič]- bid- before
sobou [gramotu pravuju česti. I v gramotě v
. [charter of.law]- read- and in charter- in
pravoi napisano, čto [ta zemlja Noskovo —
of.law- write- that [that land Noskovo]-
troeckaja [Sergieva monastyrja, a dal
of.Trinity- [Sergij- monastery]- and give-
eě k monastyrju [Iev Oběručev.
her- to monastery- [Iev Oberučev]-
[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 Ivanovič commanded the judgment charter to be read be-
fore him. And in the judgment charter [is] written that that proper-
ty of Noskovo belongs to Trinity-Sergius Monastery, and Iev
Oberučev 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 respon-
dent’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 summa-
rized 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 Ivanovič(’) velěl [d’jaku svoemu
and [prince Vasilij Ivanovič]- 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 sobě 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
Ivanovič(ju): Napisano, g(o)s(podi)ne, v [kostromskix
Ivanovič]- write- lord- in [of.Kostroma
knigax [lět(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 [Zalězskom Berezuice v Verxnem, na
monastery]- in [Zalesskij Berezovec]- in upper- in
Ogloblině Ondronko.
Oglobino- Ondronko-
And Prince Vasilij Ivanovič 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
Ivanovič, “[It is] written, lord, in the Kostroma books of the year
’73 [1464/1465] written by Mixajlo Volynskij: the lands of Trini-
ty-Sergius Monatery in Zalesskij Verznij Berezovec — in [the vil-
lage 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 Ivanovič(’) v”sprosil [kn(ja)zja Semena
and [prince Vasilij Ivanovič]- ask- [prince Semen
Romanovič(ja): Dal li esi v” manastyr’
Romanovič]- 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 Romanovič(’) pered [kn(ja)z(e)m” Vasil(’)em
Semen Romanovič]- before [prince Vasilij
Ivanovičem skazal, čto [požnju Kr(e)stci… on
Ivanovič]- say- that [field Krestcy]- he-
dal… v manastyr(’)…
give- in monastery-
And Prince Vasilij Ivanovič asked Prince Semen Romanovič, “Did
you give the Holy Savior Monastery the field of Krestcy?” And Prince
Semen Romanovič said before Prince Vasilij Ivanovič 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, čto [sja duxovnaja
stand- before bishop- and say- that [this spiritual
gramota pered nimi pisana. A [Ivan Vasil’evič’
writ]- before them- write- and [Ivan Vasil’evič]-
skazal, čto [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’evič 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 contem-
porary text-kinds, e.g., surety bonds (poručnye kabaly), purchase deeds
(dokladnye kupčie 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 state-
ments thus presented are presented as dyads and have more content, includ-
ing 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’ verifica-
tions. 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 corrobora-
tions, 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 imagina-
tion 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 recount-
ed 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 Ivanovič (1505–33) and Ivan IV Vasil’evič (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 č’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 sepa-
rately, 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 land-
cadastre 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, čto emu [sud takov byl, kak
and Šalaba- say- that him- [trial such]- be- as
v [sem spiske pisano; a Mixejko skazal, čto
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: Čto 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 accor-
dance 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 accor-
dance 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 on-
site 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
velěti, 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 Ivanovičem [nedelščik Vaska
and before [prince Vasilij Ivanovič]- [bailiff Vaska
Karačev [sudnyx mužej… postavil; a [sud’ja Ivan
Karačev]- [of.court men]-. place- and [judge Ivan
Kuzmin da [starec Efrem stali že. A
Kuz’min]- and [elder Efrem]- stand-.  and
[otvetčika Stepanka Dorogu, — skazal [nedelščik Vaska
[respondent Stepanko Doroga]- say- [bailiff Vaska
Karačov, — dal na poruku, a vyručil ego,
Karačev]- 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 poručniki.
Stepanko- run- [with [. bondsmen]-
[1] And the bailiff Vaska Karačev placed the men of court… before
Prince Vasilij Ivanovič; 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 Karačev, [he] set bail for the respondent Stepanko Doroga,
“and, lord, Ofremko Bulgak, Jakuš’s son, promised me to be his sure-
ty, 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 (topicaliz-
ation) of the corresponding noun phrase otvetčika 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 back-
ground 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(’)evičem
by [this copy]- stand- before [Mikita Vasil’evič
Beklemišovym, … Andrěiko skazal, čto 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, čto
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’evič 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 judicial-
referral 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’evič had
reserved all real-estate lawsuits for his heir, Ivan Ivanovič. 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 given-
ness 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 Ondrěem Vasil(’)evič[em], [sud
sovereign- before [prince Andrej Vasil’evič]- [trial
svoi skazali i spisok položili. I [knjaz(’)
.]- tell-. and copy- put-. and [prince
Ondrěi Vasil(’)evič(’), vozrěv” v spisok, vyslušav sud,
Andrej Vasil’evič]- 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’evič, related their trial and put the [trial] copy
[into evidence]. And Prince Andrej Vasil’evič, 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 Davydovič(’), postavja [pristava Petelju

and [Dmitrij Davydovič]- place- [constable Petelja
nedělščika i [Sidora istca pered [knjazem Andrěem
bailiff]- and [Sidor plaintiff]- before [prince Andrej
Vasil(’)evičem, [sud svoj skazal, i [srok tot
Vasil’evič]- [trial .]- say- and [term that]-
skazal, čto [poselskoi krutickoi starec’ Davyd vzjal,
say- that [steward of.Krutica elder Davyd]- take-
g(o)s(podi)ne, srok sobě 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 otvěčjat(’) i [igumena
stand- before me- respond- and [abbot
svoeg(o) ne postavil i gramot ne položil.
.]-  put- and writ-.  place-
And Dmitrij Davydovič, having placed the constable Petelja the bailiff
and Sidor the plaintiff before Prince Andrej Vasil’evič, 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 comple-
mentizer č’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 complemen-
tized 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 specifi-
cally 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 orienta-
tion (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 back-
ground 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 recursive-
ness 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 evi-
dence 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 interac-
tions 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 partici-
pants’ 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 report-
ers officially witness (1). The onset of the dyad is presented as  tagged by
the  izvěčati ‘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
tovariščom izvečjal ot poselskogo: To, brat[’]e,
fellows-. warn- from steward-  brethren-
les sečete [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-
sečem [knjazja velikogo, a ne manastyrskoj, a
cut-.1. [prince grand]- and  of.monastery- and
seči 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 warn-
ing 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 monas-
tery’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 resem-
bles 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, izvečjal ot poselskogo,
and I- them- lord- warn- from steward-
čtob iz [manastyrskogo lesu iz Dubrovinskogo
that- from [of.monastery forest]- from of.Dubrovino-
[sečenyx 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: seč’

[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 monas-
tery’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 reči,
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 uncoop-
erative, 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, čelom [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 perspec-
tives 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 nočes’ 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 Pre-
sentation, 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 priteči
‘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 “direct-
experience” 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 evi-
dence 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 č’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, čto [tret(’)ego lěta

selves-. say-.2. that [third year]-
pokosili este [to seliščo, 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 synony-
mous 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 idiosyn-
cratic 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 occur-
rence 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 imple-
mented (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 preposi-
tional 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, čto [te
[your- old.residents]-. say-.3. that [those
zemli paxali na monastyr’ let s
lands]-. plow-. for monastery- years-. about
šestdesjat’, i vy o čem monastyrju za
sixty- and you-. for what- monastery- for
tolko let molčali, i [velikomu
so.many- years-. be.silent-. and [grand
knjazju este o [tex zemljax
prince]- ..2. about [those lands]-.
bivali li čelom?
hit-.  forehead-
“Your witnesses say that they farmed those properties for the monas-
tery 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 čto, g(o)s(podi)ne, skazyvaet [černec’ Kirilo, čto
and that lord- say-.3 [monk Kirilo]- that
 L R 215

[tě pustoši dal [Danilo Blin, a skazyvaet

[those fields]- give- [Danilo Blin]- and say-.3
sem(’) lět 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-
[těx pustošei do [sěx měst”.
[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 present-
tense 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 otvěte
self- lord- [prince Ivan]- say- in response-
pered toboju, čto [to seliščo Zelenevo pašet
before you-. that [that village Zelenevo]- plow-.3
poltret(’)jatcat(’) lět, a znaxori, g(ospo)dine,
twenty-five- years-. and witnesses-. lord-
216 R V

[znaxar’stvo svoe skazali pered toboju, čto

[testimony .]- say-. before you-. that
[tu zemlju Zelenevo selišče pašet [knjaz(’)
[that land Zelenevo village]- plow-.3 [prince
Ivan tritcat(’) lět… ino, g(ospo)dine, eg(o) [znaxorju
Ivan]- thirty- years-.  lord- his [witness
Gridkě Aleksěeveu tritcati lět ot rodu
Gridka Alekseev]- thirty- years-. from birth-
nět… obyščy lět 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 sebě u Timofěja u Mixailovičja
get.falsely- . at Timofej- at Mixajlovič-
[gramotu lgotnuju na [tu rozsěč(’), a skazal,
[writ of.privilege]- for [that clearing]- and say-
g(ospodi)ne, to lěs [velikog(o) kn(ja)zja… I my,
lord-  forest- [grand prince]- and we-
g(ospodi)ne, bili čelom [Timofěju Mixailovičju;
lord- hit-. forehead- [Timofej Mixajlovič]-
i Timofěi… obyskal [ljudmi dobrymi [vseju
and Timofej- seek.out- [people good]-. [all
volostiju, čto, g(ospodi)ne, Onikei da Senka skazali
region]- that lord- Onikej- and Senka- say-
Timofěju, čto [tot Savka popaxal [našu zemlju
Timofej- that [that Savka]- plow- [our land
monastyrskuju kočevinskuju; i Timofěi, g(ospodi)ne,
of.monastery Kočevinskaja]- 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 počal staviti
fall- he- lord- [that house]- begin- build-
čerez [Timofěevo slovo…
across [Timofej- word]-
“… [Savka] obtained, under false pretenses, a writ of privilege for that
clearing from Timofej Mixajlovič, and said, lord, that [was] a forest of
the grand prince… And we, lord, petitioned Timofej Mixajlovič, 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 Kočevinskaja, belong-
ing 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 self-
quotation; 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 presuppo-
sition 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’evičem v [tom dele
before [prince Andrej Vasil’evič]- 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 [prince-
Ondreeva suda? I peredo mnoju esi nyneča
Andrej- trial]- and before me- ..2 now
skazal, čto [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, čto 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’evič?
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 testify-
ing, which the defendant Gridja has violated. The juxtaposition of Gridja’s
past offense [1] with the present charge against him [2] (cf. nyneča ‘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 back-
ground 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 testimo-
ny; as presupposed reports, it is not surprising to find them in complemen-
tized . 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 repre-
sented 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, čto 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 partici-
pants (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, čto 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 comple-
mentized  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 — nynečja my
[father my]- me- say-.3 thus now we-
[te luzi kosim za
[those meadows]- mow-.1. as
mitropoliči, potomu čto ešče 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
mitropoličix [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 report-
ing 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 princi-
pal’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 foreground-
ing, 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 non-
coreferential 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 embed-
dings, while a switch to the narrator’s indirect orientation would have caused
referential complications; for example, the phrase i oni to u nas u mitropoličix
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 hypo-
thetical , which falls into two subtypes. The first is projected statements;
 L R 225

these include first-person utterances with the ’s cělovati 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, čto 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 č’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 address-
ees could be expected to have performed before the trial (hence the preterit)
in accordance with normative procedure (14a).
(14) a. Bivali li este čelom o [toi
beat-.  ..2. forehead- about [that
zemlě [g(o)s(u)d(a)rju velikomu kn(ja)zju, čto 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
naězdom?
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 čelom
‘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 kupčaja 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”lčati ‘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 extrane-
ous 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 effi-
ciently 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 perspecti-
val 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 second-
person 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 address-
ee; 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 uphold-
ing), 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’ oppo-
nents.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 proce-
dure (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 promot-
ed it was pacing, since its general effect is to smooth transitions by eliminat-
ing 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 čto, g(o)s(podi)ne, skazyvaet [černec’ Kirilo, čto
and that lord- say-.3 [monk Kirilo]- that
[tě pustoši dal [Danilo Blin, a
[those fields]- give- [Danilo Blin]- and
skazyvaet sem(’) lět 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 sebě u Timofěja u Mixailovičja
get.falsely- . at Timofej- at Mixajlovič-
230 R V

[gramotu lgotnuju… a skazal, g(ospodi)ne, to

[writ of.privilege]- and say- lord- 
lěs [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 direc-
tion 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) Počemu 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-
lět [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 lět tu
zemlju… pašut, with the  intercalated in the usual position for parenthet-
icals. 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 interpreta-
tion) 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
tě, g(o)s(podi)ne, pustoši votčina [g(o)s(u)d(a)rja
those- lord- fields- patrimony- [liege
moeg(o) Gněvaševa…
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 con-
struction, 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 subordi-
nate (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 uncom-
plementized and complementized  reflects pragmatic or semantic differenc-
es, 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; conceiv-
ably, 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 compari-
son to . Fused reports do not occur in dyadic contexts or as plot-advancing
events in narrative sequences; they tend to appear as presupposed back-
ground 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) Počemu 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, nalevě [zemlja manastyr’skaja Vasilovskoe seliščo
 lord- on.left [land of.monastery of.Vasilovo village
Kuznecovo, i ta, g(o)s(podi)ne, [zemlja Malečkinskaja
Kuznecovo]- and that- lord- [land Malečinskaja
manastyr’skaja žo, čto 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 Malečinskaja 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 increasing-
ly archaic; in a second variant, extant in a seventeenth century copy, the
 L R 235

passage is altered to čto 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
[Ykonničskoju zemleju ot rečki ot Sendegi… po
[of.Ikonniče 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 fore-
grounded  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 [tě
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 molčali

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, [tě naši starožilcy… a Suxovo,
lord- [those our witnesses]-. and Suxoj-
g(ospodi)ne, n(y)neča doma nět, a
lord-, now at.home be-. and
skazyvajut ego v zadělě.
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 non-
action 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 orienta-
tion 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 contextuali-
zation 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 preced-
ing 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 bound-
ary — 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 sur-
rounding 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 proper-
ties.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 reči ‘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 preced-
ing 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 seliščax
I- lord- bid- them- in [those villages]-.
dvory staviti, a mne, gospodine, skazali
farms-. build- and me- lord- say-.
ivašovskie ž xristiane… čto [te selišča
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 Du-
brovka 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 intercalat-
ed tags, without implying the occurrence of a new , specify the informa-
tion 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 dě(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 convey-
ing 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 vědomo, čto [tě derevni…
whom- that- be.known that [those villages]-.
[vaši zemli monastyr’skie [Vasil(’)evskie
[your- lands of.monastery]-. [Vasil’evskaja
slobodki, čto vam, skazyvaete, dal
slobodka]-  you-. say-.2. give-
[Roman Ivanovič(’)?
[Roman Ivanovič]-
“To whom is it known that those villages… are your (the monas-
tery’s) properties pertaining to Vasil’evskaja Slobodka, which you,
you say, were given by Roman Ivanovič?” (ASÈI 2: 202, no. 288).
b. A vy celuete li krest na tom,
and you-. kiss-.2.  cross- on that-
čto 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 intercalat-
ed (parenthetical) and preposed (embedding) tags. Its placement is actually
ambiguous, since the complementizing conjunction č’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 ex-
pressed by using report markers “pleonastically”, where the fact of hetero-
glossia is evident, e.g., after the negative-presuppositional  l”gati ‘lie’:
(27) tot, gospodine, [Ivaško Vnuk lžet, čto 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 individu-
al 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 molčali. Kak
we- lord- them-  be.silent-. when
počali oni paxati siloju, i my,
begin-. they- plow- force- and we-
gospodine, bili čelom 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”lčati ‘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 čel”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 velěti ‘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
nyně 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  velěti in (29a), like biti čel”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 rela-
tions 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 subjec-
tive 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 exam-
ined, the dyadic question-and-answer structure found in the trial record is
abandoned for a concise third-person, past-tense narrative devoted exclusive-
ly 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 Grigorievič(’) v [těx zemljax
and [Konstantin Grigor’evič]- in [those lands]-.
246 R V

[Simanovskog(o) ma[na]styrja [poselskog(o) starca Ignat(’)ja

[Simonov monastery]- [manager elder Ignatij
Travina opravil, a iščei… obvinil,
Travin]- find.right- and plaintiffs-. find.wrong-
potomu čto 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]-.
iščei i starožilci; i prisudil [tě
plaintiffs-. and old.residents-. and award- [those
zemli [Simanovskomu manastyrju.
lands]-. [Simonov monastery]-
[1] And Konstantin Grigor’evič 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 differ-
ent 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., Lixačev
(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 men-
tioned 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 subse-
quent 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(’)evič(’)… [Nestera, starca
and therefore [prince Jurij Vasil’evič]- [Nester elder]-
[Simanovskog(o) monastyrja… opravil, a Ondrěika
[of.Simonov monastery]- find.right- and Ondrejko-
da [Ivaška Strelu [Lopotovyx detei obvinil
and [Ivaško Strela]- [Lopot- children]- find.wrong-
i prisudil [zemlju Malečkinskuju k Verznev’skomu
and award- [land Malečkinskaja]- to Verznevskoe
selu [Simanovskogo manastyrja.
village]- [of.Simonov Monastery]-
And therefore Prince Jurij Vasil’evič… 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 Malečinskaja 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 implementa-
tion of the judgment; they are “verdictives”, in Austin’s (1975) classification,
’s like acquit, convict, rule, which “consist in the delivering of a find-
ing…” (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. Seventeenth-
century 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’evič’… [Danila Stromilova i
and [prince grand Ivan Vasil’evič]- [Danilo Stromilov]- and
[Osifa čern’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, opriče [bortnič’ix dereven’, čto
old.usage- except [beekeeper- villages]-. 
otnjal u nix [otec moj knjaz’ velikij Vasilej
take.away- at them- [father my prince grand Vasilij
Vasil’evič’ u manastyrja u Simonova.
Vasil’evič]- at monastery- at Simonov-
And Grand Prince Ivan Vasil’evič… 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 bee-
keepers’ villages that my father Grand Prince Vasilij Vasil’evič 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’evič’, 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 Andreevičju [Kirilova
Rus’]- [prince Mixajlo Andreevič]- [Kiril-
monastyrja igumena suditi po starine… oprič’
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
Andreevič 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 velěti ‘bid’, which takes an infinitival comple-
ment with a dative for the recipient of the command (if explicit). (In prohibi-
tions, 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 Dmitreevič(’)… [těx zeml’
and [prince grand Vasilij Dmitrievič]- [those lands
c(e)rk(o)vnyx… ne velěl 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, tě[x] velěl [veliki [knjaz’] [arximandritu
 obey- those- bid- [grand prince]- [abbot
mixailov’skomu vo[n] metati.
of.Michael]- thence drive-
And Grand Prince Vasilij Dmitrievič… 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 velěti 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 velěti 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 č’to or,
in a few cases, togo dělja č’to ‘because’) or by the all-purpose
complementizer č’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 narra-
tive 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 continu-
ous 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 complemen-
tized indirect or nondirect speech (6). The only complementizer attested is
the all-purpose conjunction č’to ‘that’. This strategy is attested, sometimes
alongside other methods, in 34 examples in 25 documents.7
(6) I [knjaz(’) velikii Ivan Vasil(’)evič(’)… [Vaska Danilova
and [prince grand Ivan Vasil’evič]- [Vasko Danilov]-
velěl obiniti… potomu čto skazal Vasko, čto [sud
bid- find.wrong- because say- Vasko- that [trial
takov byl, da i to sam že skazal, čto
such]- be- and  that- self-  say- that
za sorok lět o [toi požně molčal, da
for forty- years-. about [that field]- be.silent- and
n(y)něča [tu požnju pokosil.
now [that field]- mow-
And Grand Prince Ivan Vasil’evič… 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-
nyneča pokosil jaz; a požnja to [velikog(o)
now mow- I- and field-  [grand
knjaz(ja) [Gorodišč”skogo sela.
prince]- [of.Gorodišče 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
Gorodišče” (ASÈI 2: 383, no. 383).
b. I Vasko tak rek: … [ta požnja byla
and Vasko- thus speak- [that field]- .
Pašin pokos; a poslě 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, bolě soroka lět… a poslě
already lord- more forty- years-. and after
ix [tog(o) pokosa ot [těx měst”…
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 (infor-
mative 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 comple-
mentizer; 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… [selčan
by [this copy]- [grand prince]- [villagers
voislavskix Larivonka s tovarišči
of.Voislavskoe]-. Larionko- with companions-.
velel obiniti, potomu čto [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, čto za
Leša]-  with peasants-. say-. that for
dvatcat’ let [ta zemlja [knjažščina
twenty- years-. [that land]- [prince’s.land
Voislavskaa, a paxali ee [te
of.Voislavskoe]- and plow-. it- [those
selčane…
villagers]-.
In accordance with this copy, Grand Prince Ivan Vasil’evič… 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 čto 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 interpre-
tation 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
cělovaniju ‘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’evič(’) Naumov… [Okiša Oljunova da
and [Vasilij Grigor’evič Naumov]- [Okiš Oljunov]- and
260 R V

Fomku… ob’vinil potomu čto skazali u sobja

Fomka- find.wrong- because say-. at .
[starožilca mertvovo, aprič(’) tog(o) ne skazali u
[old.resident dead]- except that-  say-. at
sobja starožilca…
. old.resident-
And… Vasilij Grigor’evič 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) Vědomo, 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(’) lět, kak ego ne
that- lord- ten- years-. as him- 
stalo; aprič(’) 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’evič(’)… [kn(ja)zja Ivana Kostjantinovič(a)

and [prince Ivan Jur’evič]- [prince Ivan Konstantinovič
Obolenskog(o) velěl obiniti potomu čto 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’evič… bid [the trial judge] to decide against
Prince Ivan Konstantinovič 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’evič’ velel… [Andreja Okljačeeva
and [prince Ivan Jur’evič]- bid- [Andrej Okljačeev]-
obviniti… potomu čto 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 Čjubaru da Ondreju
forest- thus speak-. [judge Čubar]- 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 otvečav ničego…
Ofonasij- and [elder Poluekt]-  answer- nothing-
And Prince Ivan Jur’evič bid [the trial judge]… decide against Andrej
Okljačeev… because Andrej and his witnesses went astray… [they]
led through the land of Ugrešskij Monastery’s village of Konstanti-
novo; and the Ugrešskij cleric Ofonasij Morž and the monk Poluekt,
coming upon them in the forest, spoke in this way to the judge Čubar
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 wit-
nesses 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 other-
wise 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 men-
tioned 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. Conse-
quently, 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 interpreta-
tion 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 dě(i), which generally appears in intercalation after the first phono-
logical 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 děi, dei, dě, 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 Aleksandrovič(’)… [Martynka Kosjaka
and [prince Danilo Aleksandrovič]- [Martynko Kosjak]-
264 R V

velěl obviniti potomu čto ego že znaxori, na

bid- find.wrong- because his  witnesses-. to
kotyryx [sic] sja on poslal, pered sud(’)jami
..  he- refer- before judges-.
skazali, čto rozvod mež [těmi d(e)r(e)vnjami
say-. that demarcation- between [those villages]-.
ne byval, a my dei rozvoda ne
 be- and we-  demarcation- 
vědaem.
know-.1.
And… Prince Danilo Aleksandrovič… 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 unambiguous-
ly coreferential with the original speech situation, there is no way to deter-
mine 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 subdivi-
sions of discourse tend to be organized around a single participant’s perspec-
tive 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 dě(i),
which made the status of the second clause explict and reasserted the
reporter’s perspective. In this sense, the seemingly pleonastic use of dě(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 height-
ened 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 dě(i) in
(14) may be a complex strategy for highlighting the aggravating circum-
stance 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 dě(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 dě(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 č’to, so that it conforms to the most
common pattern of reporting (see 8.3.1):
(15) I [knjaz’ velikij Dmitrej Ivanovič… [Ivaška Izbina
and [prince grand Dmitrij Ivanovič]- [Ivaško Izbin]-
velel sud’e obviniti potomu čto Ivaškov
bid- judge- find.wrong- because Ivaško-.
že [starožilec Vas’ko Roskorjaka pered sud’eju
 [old.resident Vas’ko Roskorjaka]- before judge-
skazal, čto 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 [mitropoliči
fields]-. say-.3 mow-. [metropolitan-
xristiane za pusto…
peasants]-. as empty
And Grand Prince Dmitrij Ivanovič… 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 cohe-
sion-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 Ivanovič(’)… Stepka da Oksenka
and [Prince Jurij Ivanovič]- Stepko- and Oksenko-
velěl obviniti… potomu čto vošli sami
bid-. find.wrong- because go.in-. selves-.
na [tě počinki 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 [tě počinky ix,
[local peasants]-. in [those clearings]-. them-
skazyvajut, ne sažyvali, ni gramot(y) im
say-.3.  settle-. nor writ- them-
lgotnye na [tě počinki ne davyvali…
of.privilege- to [those clearings]-.  give-.
And Prince Jurij Ivanovič… 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 dě(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(’) Andrěi Vasil(’)evič(’)… [igumena krutickog(o)
and [prince Andrej Vasil’evič]- [abbot of.Krutica
Galasěja i eg(o) poselskog(o) velěl obvinit(’)
Galasij]- and his manager- bid- find.wrong-
togo dělja, čto poselskoi… vzjal sobě [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
otvěta [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’evič… 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 (dě(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 dě(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 dě(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 informa-
tion, 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 Ivanovič’… [Stepanka Dorogu
and [prince Vasilij Ivanovič]- [Stepanko Doroga]-
velel obviniti, potomu čto [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
poručniki…
bondsmen-.
And Prince Vasilij Ivanovič… 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 lězem 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 Aleksandrovič(’)… velěl Efrěma
and [prince Danilo Aleksandrovič]- bid- Efrem-
obviniti potomu čto 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 Aleksandrovič… 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 renar-
rations 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 illocuti-
onary 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 rěči ‘say what litigant says’
(Gorčakov 1871, appendix [separate pagination]: 65, no. 4.5). Another
perlocutionary-based  centers on očistiti ‘prove ownership of’ (literally,
‘clear’) in (20):
(20) I [kn(ja)z’ veliki… Kirila velěl obviniti,
and [prince grand]- Kirilo- bid- find.wrong-
potomu čto [Danilo Blin [těx zemel’… po gramotam
because [Danilo Blin]- [those lands]-. by writs-.
ne očistil…
 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  očistiti 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”lčati ‘be silent’ (e.g., AFZX 107, no. 114); however, more
specific negated ’s like bivati čel”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 repre-
sent a second layer of reporting within the rationale, i.e., to convey requota-
tions 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 čto 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 velěl obviniti,
and [prince grand]- [elder Kirilo]- bid- find.wrong-
potomu čto… u kn(ja)ži u [Mixailovy gramoty u
because at prince-. at [Mixajlo- writ]- at
žalovalnye pečjat’ ne [knjaža Mixailova, —
of.immunity- seal-  [prince- Mixajlo-]-
potpisal i zapečjatal [gramotu knjažemixailovu
sign- and seal- [writ prince.Mixajlo-]-
bojarin eg(o) [Grigorei Fedorovič(’)…
boyar- his [Grigorij Fedorovič]-
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 Fedorovič 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 Grigorievič(’)… iščei…
and [Konstantin Grigor’evič]- plaintiffs-.
obvinil, potomu čto… [sami oni
find.wrong- because [selves they]-
iščei i starožilci…
plaintiffs-. and old.residents-.
And Konstantin Grigor’evič… decided against the plaintiffs…
because… they themselves [were] plaintiffs and longtime resi-
dents (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 informa-
tion conveyed by third-person cognitive verbs must be reported or inferred
(cf. Austin 1946/1989: 113–14). The possibility of inferences about knowl-
edge can be ruled out in this context; thus, the reportedness of knowledge
predicates is presupposed and can be omitted.
(24) I… [Dmitrij Volodimerovič’… Efimka velel
and [Dmitrij Volodimerovič]- Efimko- bid-
obvinit’, potomu čto 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 molčal; 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 Volodimerovič… 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 Romanovič’… Martynka… da Ermolku…
and [Grigorij Romanovič]- Martynko- and Ermolka-
obvinil”, potomu čto iskali [těx” zemel’ za
find.wrong- because seek-. [those lands]-. after
sorok” lět”; da i starožilci ix”…
forty- years-. and  old.residents-. their
togo ne upomnjat”, [koj mitropolit” [tě
that-  recall-.3. [which metropolitan]- [those
zemli poimal” za sebja; a v” [Mixajlovyx”
lands]-. take- under . and in [Mixajlo-
knigax” Volynskogo [těx” zeml’ [starye
books]-. Volynskij- [those lands]-. [old
derevni pisany mitropoliči, a [novye
hamlets]-. write- metropolitan-.. and [new
derevni stavleny promež” těx” že zeml’…
hamlets]-. place- amidst those-.  lands-.
And Grigorij Romanovič… 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 metropoli-
tan’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; interpreta-
tion 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 inseparabili-
ty 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 authori-
tative 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 Vasilievič(’), požaloval
lo I- [prince grand Ivan Vasil’evič]- grant-
 R  V 281

esm’ [kn(ja)zja Vasilia Ivanovič(a). Čto mi

-.1 [prince Vasilij Ivanovič]- that me-
bil čelom o tom: moi de [selčane
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”zorščika Semena Nakvasu,
give- him- [. surveyor Semen Nakvasa]-
a velel es[mi em]u [tě zemli
and bid- -.1 him- [those lands]-.
obyskati [tamošnimi st[arožil’cy n]a [o]bě
investigate- [local witnesses]-. for [both
storony bez polja.
sides]-. without field-
I, Grand Prince Ivan Vasil’evič, hereby grant a boon to Prince
Vasilij Ivanovič, [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 čel”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 dě(i) (see Section 8.3.4). While the complaint, as a
nondyadic statement, is not prototypical testimony, it can still have evidentia-
ry value in the trial setting; thus  is strongly preferred in that contextuali-
zation. 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 mně… spisok], a
and [Semen Nakvasa]- bring- to me- copy- and
v nem pisano, čto 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
mežě; i oni pered nim” skazali čto
border- and they-. before him- say-. that
mežy [těm zemljam ne vědajut. A
border- [those lands]-.  know-.3. and
[glumovskix starožilcev v”sprašival… i oni
[of.Glumovo witnesses]-. ask- and they-
skazali mežu [dobromer’skoi zemlě 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] seličany za nimi po
and with [my villagers]-. after them-. along
[toi mežě ne poěxali. 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]kvasině spiskě 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 prop-
erties. [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 de-
scription 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 affirma-
tive, 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 cultural-
ly 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 context-
sensitive 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 institutional-
ized 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 scribe-
writer 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 func-
tionally 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 tradition-
al 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 increas-
es 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 diffuse-
ness 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 recon-
structive abilities. Thus the standard reporting strategy matched the organiza-
tional 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 
reči, 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 attribu-
tion. 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 contextuali-
zation is relatively close to the prototype, the standard strategy predominates.
By contrast, in certain other nondyadic situations, there is a distinct prefer-
ence 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 evi-
dence 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 indetermi-
nate) 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 testimo-
ny. 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 verifica-
tion 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-than-
direct (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 institution-
al 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 presup-
posed 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 back-
ground 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 pre-
ferred, 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 context-
ualizations do not fully coincide; there are divergences linked with differen-
ces 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 interpreta-
tions 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 complemen-
tized 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 corrobora-
tions 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 associat-
ed 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 investi-
gated 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 strate-
gies 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 accor-
dance 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 recon-
struct, without speculation, collective intentions to which individual writers
had to subscribe as part of cooperative communication. Subsequent applica-
tion 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 demon-
strates 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 contextu-
alization. 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 fine-
grained 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 miscella-
neous 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  reči, 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 reči 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 reči; 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 dia-
chronic “competition” among the alternatives — in other words, that it is not
purposive. This is simply fallacious, a form of begging the question; co-
occurrence 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 (incipi-
ence 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 Metro-
politan Filipp in 1471 (AI 512–14, no. 280) includes  tagged by not only
by ’s but also by prikazyvati ‘charge’, v”zvěstiti ‘proclaim’, napisati
‘write’, and several nominal introducers, and  tagged by biti čelom’
‘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 long-
term 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. “děti, děju” and “dějati”; 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 Čerepnin 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 Čerepnin 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; Čerepnin 1956: 564–69; Kotkov and Popova
1986: 3.
14. On this word, customarily abbreviated as (>n or (c>n, see Čerepnin 1956: 246; Zaliznjak
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;
Xaburgaev 1974: 425–26; Čumakov 1975: 16; Longacre 1983: 132; Li 1986: 39–40.
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., Čumakov 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;
Sprinčak 1960–64, 2: 119; Otin 1969; Lopatina 1979: 442; Rinberg 1985: 53; Schmücker-
Breloer 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; Čumakov 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 Nemčenko
1980: 152–54; Stola 1958.
2. One complication is the aspect of the gerund. Since the verb reči was perfective by the fifteenth
century, r’kuči 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. Lixačev (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. Lixačev (ed.) 1894: no. 2 (1583), 3 (1589); N. P.
Lixačev (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.
Lixačev (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, ssyločnye pamjati (“referral memoran-
da”; 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 counter-
example.
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; Gorčakov 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; Gorčakov 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.
Lixačev (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
1973: 33–36; Čumakov 1975: 38–48.
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; Padučeva 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 Padučeva (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 (Sprinčak 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; Gorčakov 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 (Gorčakov 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; Gorčakov 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 rečisja, 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 rečisja 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 velěti, 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 učiniti
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; Gorčakov 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 reči. 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. Lixačev (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; Gorčakov 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 non-
judgment).
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
cělovati 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 dě(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”lčati ‘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”lčati is attested in
ibid.: 1: 247, no. 340.
31. For further examples of biti/bivati čel”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 velěti, 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 nyněča ‘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; Gorčakov
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-načala 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 — načala XVII veka: Sbornik dokumentov.
Vol. 1. Moscow: Arxeografičeskij Centr.
Arnovick, Leslie K.
1999 Diachronic Pragmatics: Seven Case Studies in English Illocutionary Develop-
ment [Pragmatics and Beyond, New Series 68]. Amsterdam: John Benjamins.
Arxeografičeskaja Kommissija
1836 Akty, sobrannye v bibliotekax i arxivax Rossijskoj Imperii Arxeografičeskoju
Èkspedicieju Imperatorskoj Akademii Nauk. 1836. Vol. 1. St. Petersburg:
Arxeografičeskaja Kommissija.
1838 Akty juridičeskie, ili sobranie form starinnogo deloproizvodstva. St. Peters-
burg: Arxeografičeskaja Kommissija.
1841 Akty istoričeskie, sobrannye i izdannye Arxeografičeskoju Kommissieju.
Vol. 1. St. Petersburg: Arxeografičeskaja Kommissija.
320 R V

1846 Dopolnenie k Aktam istoričeskim. Vol. 1. St. Petersburg: Arxeografičeskaja

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: Formal-
ist 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: Universi-
ty 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”. Founda-
tions 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. Cam-
bridge: Cambridge University.
Barnet, Vladimír
1965 Vývoj systému participií aktivních v ruštině [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 reči v staroslavjanskom i v
drevnerusskom jazykax”. In Očerki po russkomu jazyku i stilistike, 192–201.
Saratov: Saratovskij Gosudarstvennyj Pedagogičeskij Institut.
Borkovskij, V. I.
1981 Sintaksis skazok: Russko-belorusskie paralleli. Moscow: Nauka.
Borkovskij, V. I., and P. S. Kuznecov
1965 Istoričeskaja 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: Cam-
bridge 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 Učebno-pedagogičeskoe
Izdatel’stvo.
1958 Istoričeskij 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.
Čerepnin, L. V.
1948–51 Russkie feodal’nye arxivy XIV — XV vekov. Moscow: Akademija Nauk
SSSR.
1952 “Akty konca XIV — XV vv., ix raznovidnosti i značenie v kačestve
istoričeskogo istočnika”. In Akty social’no-èkonomičeskoj istorii severo-
vostočnoj Rusi konca XIV-načala XVI v., B. D. Grekov and L. V. Čerepnin
(eds.), 1: 638–50. Moscow: Akademija Nauk SSSR.
1956 Russkaja paleografija. Moscow: Nauka.
Čerepnin, 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 Evidenti-
ality: 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 Con-
scious Experience in Speaking and Writing. Chicago: University of Chicago.
 R 323

Chatman, Seymour
1975 “The structure of narrative transmission”. In Style and Structure in Litera-
ture: 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 Bertuccelli-
Papi (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 Strate-
gies 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 proščal’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: Explora-
tions 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.
Čumakov, G. M.
1975 Sintaksis konstrukcij s čužoj reč’ju. Kiev: Vyšča 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: Istočniki dlja ego
izučenija, S. I. Kotkov (ed.), 151–88. Moscow: Nauka.
1980 “Ob istoriko-stiličeskom 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 Docu-
ments [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, otnosjaščiesja k istorii tjaglogo naselenija v Moskovskom gosudarstve.
Part 2: Gramoty i zapisi. Jur’ev: Tip. K. Matisena.
Dmitriev, L. A., and D. S. Lixačev (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 R V

Forsyth, J.
1970 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.
Gorčakov, 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. Čerepnin (eds.)
1952–64 Akty social’no-èkonomičeskoj istorii severo-vostočnoj Rusi konca XIV —
načala 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.
Gvozdanović, 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
Učebno-pedagogičeskoe 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 Func-
tions of the Verb [Pragmatics and Beyond, New Series 43], Theo A. J. M.
Janssen and Wim van der Wurff (eds.), 289–98. Amsterdam: John Benja-
mins.
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 čestina 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 develop-
mental aspects”. In Reflexive Language: Reported Speech and Metaprag-
matics, 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 Dis-
course 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, Vjač. Vs.
1965 Obščeindoevropejskaja, praslavjanskaja i anatolijskaja jazykovye sistemy:
Sravnitel’no-tipologičeskie očerki. Moscow: Nauka.
 R 329

1973 “Značenie idej M. M. Baxtina o znake, vyskazyvanija i dialoge dlja

sovremennoj semiotiki”. Učenye 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 Lan-
guage, 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 mestničestva. 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

Kalačov, Nikolaj (ed.)

1857–84/1970 Akty, otnosjaščiesja do juridičeskogo byta Drevnej Rossii, izdany
Arxeografičeskoju Kommissieju [Slavistic Printings and Reprintings
252]. 3 vv. The Hague: Mouton.
Karpov, G. O. (ed.)
1887 Pamjatniki diplomatičeskix snošenij drevnej Rusi s deržavami inostrannymi
[Sbornik Imperatorskogo Russkogo Istoričeskogo Obščestva 59]. Vol. 2. St.
Petersburg: Imperatorskoe Russkoe Istoričeskoe Obščestvo.
Kaštanov, S. M.
1970 Očerki 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 sočinenija. Leningrad: Akademija Nauk SSSR.
Keller, Rudi
1985 “Towards a theory of linguistic change”. In Linguistic Dynamics: Discours-
es, 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”. Canadian-
American 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 peredači čužoj reči v russkom jazyke”. Učenye Zapiski Lenin-
gradskogo Gosudarstvennogo Pedagogičeskogo Instituta 104: 107–72.
Koreckij, V. I.
1969 “Vnov’ otkrytye novogorodskie i pskovskie gramoty XIV-XV vv.”. Arxeo-
grafičeskij ežegodnik za 1967 god, 275–90.
Kotkov, S. I. (ed.)
1990 Pamjatniki južnovelikorusskogo narečija: Konec XVI-načalo XVII v. Mos-
cow: 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 Očerki po sintaksisu južnovelikorusskoj pis’mennosti XVII veka. Moscow:
Nauka.
Kozlovskij, Pavel
1890 “O sočetanii predloženij prjamoj i kosvennoj reči v russkom jazyke”.
Filologičeskie zapiski 4–5: 1–7.
Krotov, M. G. and S. I. Smetanina (eds.)
1987 “Istočniki 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”. Hokkaidȧo Daigaku
7: 67–89.
Kuz’mina, I. B., and E. V. Nemčenko
1980 “K istorii kratkix form dejstvitel’nyx pričastij v russkom jazyke”. In
Obščeslavjanskij lingvističeskij 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. Amster-
dam: John Benjamins.
Larin, B. A.
1961 “Razgovornyj jazyk Moskovskoj Rusi”. In Načal’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 Očerki 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-èkonomičeskoj istorii severa Rossii konca XV — XVI v.:
Akty Soloveckogo monastyrja. 2 vv. Leningrad: Nauka.
Lips, Marguerite
1926 Le style indirect libre. Paris: Payot.
 R 333

Lixačev, D. S.
1946/1986 “Russkij posol’skij obyčaj XI — XIII vv.”. In Issledovanija po
drevnerusskoj literature, 140–53. Leningrad: Nauka.
1952 Vozniknovenie russkoj literatury. Moscow: Akademija Nauk SSSR.
Lixačev, N. P. (ed.)
1894 Mestničeskie 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 peredači čužoj reči v ramkax složnogo predloženija”. In
Istoričeskaja grammatika russkogo jazyka: Sintaksis: Složnoe predloženie,
V. I. Borkovskij (ed.), 422–47. Moscow: Nauka.
1985 “Sredstva organizacii teksta pri peredače čužoj reči: Na materiale sudebnyx
dokumentov XVII v.”. In Vostočnye slavjane: Jazyki, istorija, kul’tura, J. N.
Karaulov (ed.), 57–64. Moscow: Nauka.
1987 “Zakonomernosti sintaksičeskoj struktury teksta v pamjatnikax narodno-
razgovornoj 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. Cam-
bridge: 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 českého. 2d ed. Prague: Academia.
Martysevič, I. D.
1951 Pskovskaja sudnaja gramota: Istoriko-juridičeskoe 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 Linguis-
tic 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 sintaksičeskix konstrukcij dlja peredači čužoj reči v russkom
jazyke: Konstrukcija prjamoj reči”. Učenye 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”. Arxeografičeskij 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 mıæněti (sja)”. In American Indian and Indo-
european Studies: Papers in Honor of Madison S. Beeler [Trends in Linguis-
tics 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 Novo-
sibirskogo Universiteta.
Otin, E. S.
1969 “K istorii razvitija formy kosvennoj reči v drevnerusskom jazyke”. Naučnye
Doklady Vysšej Školy. Filologičeskie Nauki, 1969, no. 3: 54–64.
Padučeva, E. V.
1996 Semantičeskie 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 Čet’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 Nineteenth-
Century European Novel. Manchester: Manchester University.
Pearce-Higgins, Lucinda
1989 Goldilocks and the Three Bears. Austin, Texas: Steck-Vaughn.
336 R V

P’ecux, Vjačeslav
1990 Novaja moskovskaja filosofija: Roman. In Vjačeslav P’ecux, Ja i pročee,
218–333. Moscow: Xudožestvennaja Literatura.
Peškovskij, A. M.
1956 Russkij sintaksis v naučnom osveščenii. 7th ed. Moscow: Gosudarstvennoe
Učebno-pedagogičeskoe 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 istočniki po istorii černosošnogo zemlevladenija v Rossii XIV —
načala 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: Učpedgiz.
Preobraženskaja, M. N.
1983 “Specifika složnyx sintaksičeskix konstrukcij v uslovijax drevnerusskogo
teksta”. In Struktura predloženija v istorii vostočnoslavjanskix 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 reči i svjaznogo teksta v istoričeskom
sintaksise”. In Vostočnye 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: Učpedgiz.
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 Peredača prjamoj reči v delovyx tekstax XVII v.: Po materialam arxiva
Požarskix. Istorija russkogo jazyka i lingvističeskoe istočnikovedenie, 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. Albuquer-
que, 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 Linden-
berger, and Madeline Sutherland (eds.), 41–81. Stanford, California:
Stanford University.
Sprinčak, Ja. A.
1960–64 Očerki russkogo istoričeskogo 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 sintaktičeskogo celogo v drevnerusskom
tekste: Na primere konstrukcii s prjamoj reč’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 deepričastie 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 [Stock-
holm Slavic Studies 14]. Stockholm: Almqvist and Wiksell International.
Švedova, N. Ju
1952 “K voprosu ob obščenarodnom 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 Lan-
guage, 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 molest-
ed (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: Cam-
bridge University.
Tarabasova, N. I.
1963 “Ob odnom frazeologizme v častnoj perepiske XVII veka”. In Issledovanija
po lingvističeskomu istočnikovedeniju, S. I. Kotkov and O. A. Knjazevskaja
(eds.), 144–55. Moscow: Nauka.
1964 “O nekotoryx osobennostjax jazyka delovoj pis’mennosti”. In
Istočnikovedenie 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 jasačnogo 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 Con-
ceptual 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 Dis-
course 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 čelobitnyx 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 Institu-
tional 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 Slavon-
ic Literary Languages, Gerald Stone and Dean S. Worth (ed.), 233–41.
Columbus, Ohio: Slavica.
Xaburgaev, G. A.
1974 Staroslavjanskij jazyk. Moscow: Prosveščenie.
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”. Čtenija v Imperatorskom
Obščestve Istorii i Drevnostej Rossijskix pri Moskovskom Universitete 3 (7):
85–122.
Zaliznjak, A. A.
1986a “Novgorodskie berestjanye gramoty s lingvističeskoj točki 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, Vjač.
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 Juridičeskoj
Literatury.
Zipf, George Kingsley
1949 Human Behavior and the Principle of Least Effort. Cambridge, Massachu-
setts: 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 Pravaja gramota (with referral, 1462–64 [1525–50]
114 Pravaja gramota, end of the 15th century [1525–50]
117 Pravaja gramota (with referral, 1498 [1525–50]
125 Pravaja gramota (with referral, 1462–1505 [1525–50]
129 Pravaja gramota, 1495–99 [1525–50]
140 Pravaja gramota, end of the 15th century [1525–50]
157 Pravaja gramota, 1499 [1525–50]
204 Pravaja gramota, 1492 [1525–50]
248 Pravaja gramota (with referral, 1493–94
249 Pravaja gramota (with referral, 1473–89 [1525–50]
254 Pravaja gramota, ca. 1501–2 [1525–50]
258 Pravaja gramota, ca. 1501–2 [1525–50]
259 Pravaja gramota (with referral, 1498 [1525–50]
261 Pravaja gramota (with referral, 1473–89 [1525–50]
308 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 Sudnyj spisok, 1462–73 [mid-16th century]
340 Dokladnoj sudnyj spisok, ca. 1464–78
344 R V

397 Sudnyj dokladnoj spisok, ca. 1470–78 [mid-16th century]

430 Pravaja gramota, ca. 1474–75
431 Sudnyj spisok (with referral, ca. 1474–75 [mid-16th century]
447 Pravaja gramota, 1474–91 [mid-16th century]
467 Pravaja gramota, 1478–82
485 Pravaja gramota, ca. 1470–1490s [1641]
521 Sudnyj spisok (pravaja gramota?), ca. 1485–90 [mid-16th century]
522 Pravaja gramota, 1485–90
523 Pravaja gramot (with referral, 1485–90
524 Pravaja gramot (with referral, 1485–90
525 Pravaja gramot (with referral, 1485–90 [mid-16th century]
537 Sudnyj spisok (with referral, 1488–90 [mid-16th century]
538 Sudnyj spisok (with referral, 1488–90 [mid-16th century]
539 Sudnyj spisok (with referral, 1488–90 [mid-16th century]
540 Sudnyj spisok (with referral, ca. 1488–90 [mid-16th century]
557 Pravaja gramota (with referral, ca. 1490–95
571 Sudnyj spisok, ca. 1492–94 [mid-16th century]
581 Pravaja gramota, ca. 1495–97 [16th century]
582 Pravaja gramota, 1495–99 [mid-16th century]
583 Dokladnoj sudnyj spisok, 1495–99
584 Sudnyj spisok (2 variants, one with referral, 1495–99 [mid-16th century]
585 Dokladnoj sudnyj spisok, 1495–99 [mid-16th century]
586 Dokladnoj sudnyj spisok, 1495–99
587 Dokladnoj sudnyj spisok, ca. 1495–99 [contemporary copy?]
588 Dokladnoj sudnyj spisok, 1495–99 [mid-16th century]
589 Dokladnoj sudnyj spisok, 1495–99
590 Dokladnoj sudnyj spisok, 1495–99 [mid-16th century]
591 Dokladnoj sudnyj spisok, 1495–99 [mid-16th century]
592 Dokladnoj sudnyj spisok, 1495–99
593 Sudnyj spisok, 1495–99 [mid-16th century]
594 Dokladnoj sudnyj spisok (2 variants), 1495–99 [mid-16th century]
595 Pravaja gramota (with referral, 1495–99 [mid-16th century]
604 Dokladnoj sudnyj spisok, 1496 [mid-16th century]
607 Dokladnoj sudnyj spisok, 1496–98 [contemporary copy]
607a Pravaja gramota (with referral, 1496–98
615 Sudnyj spisok (with verdict), 1497–98 [mid-16th century]
628 Sudnyj spisok (with referral, ca. 1499–1502 [mid-16th century]
635 Dokladnoj sudnyj spisok, ca. 1500–1501 [mid-16th century]
639 Sudnyj spisok (with verdict), ca. 1501–2 [mid-16th century]
640 Sudnyj spisok (with verdict), ca. 1501–2 [mid-16th century]
642 Dokladnoj sudnyj spisok, ca. 1502–4 [mid-16th century]
 A 345

651 Dokladnoj sudnyj spisok, 1504

658 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 Pravaja gramota (with referral, 1494–99 [17th century]

414 Pravaja gramota, 1497–98
416 Pravaja gramota, 1498
418 Podpisnoj sudnyj spisok (with referral, 1498–99 [17th century]
419 Pravaja gramota (with referral, 1498–99 [17th century]
421 Pravaja gramota, ca. 1498–99 [17th century]
422 Pravaja gramota (with referral, 1500
428 Podpisnoj sudnyj spisok (with referral, 1504
450 Pravaja gramota (with referral, ca. 1449–50
458 Pravaja gramota, ca. 1456–64 [18th century]
463 Dokladnoj sudnyj spisok, 1465–66 (tentative dating)
464 Pravaja gramota, ca. 1465–71
465 Pravaja gramota, ca. 1470–85 (probably abridged) [1529]
481 Podpisnoj dokladnoj sudnyj spisok, ca. 1484–90
483 Pravaja gramota (with referral, 1485–1490 [1529]
492 Pravaja gramota (with referral, ca. 1497–98 [18th century]
493 Pravaja gramota (with referral, ca. 1497–98 [1556]
495 Sudnyj spisok, 1503
496 Pravaja gramota, 1453–62
ASÈI 3
31 Pravaja gramota, ca. 1416–17
32 Pravaja “sudnaja” gramota, 1425
35 Pravaja gramota, ca. 1448–52 [15th century]
48 Pravaja gramota (with referral, 1505–6
50 Pravaja gramota (with referral, ca. 1499–1502 [17th century]
55 “Pravaja i bessudnaja” gramota (with referral, ca. 1461–62
56 Pravaja gramota (with referral, ca. 1462–69
105 Pravaja gramota, 1498–99 [17th century]
172 Podpisnoj dokladnoj sudnyj spisok, 1504
173 Pravaja gramota (with referral, 1504
208 Pravaja gramota (with referral, ca. 1495–97
209 Dokladnoj sudnyj spisok, ca. 1495–97
218 Dokladnoj sudnyj spisok, 1505
221 Pravaja gramota, 1501 [18th century]
223 Pravaja gramota, 1504–5
224 Pravaja gramota, 1504–5
250 Podpisnoj dokladnoj sudnyj spisok, 1501 [1576]
251 Podpisnoj dokladnoj sudnyj spisok, 1501 [1576]
276 Pravaja gramota (with referral, 1499–1500
288 Pravaja gramota, ca. 1490–1509
 A 347

319 Pravaja gramota (Rjazanian; paraphrased), 1464–71 [unspecified date]

357 Pravaja gramota (Rjazanian), 1483–1500 [17th century]
364 Pravaja gramota (Rjazanian), 1464–82 [18th century]
477 Pravaja gramota (with referral, 1478–85
478 Sudnyj spisok, 1503–4 [late 17th century]
Gorčakov 1871
5 [Pravaja gramota], 1504 [16th century]
GVNP
92 Pravaja gramota (Novgorodian), 1st quarter of the 15th century
340 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 Bloomfield, Leonard 23
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 Cassirer, Ernst 65
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
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
Čerepnin, 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
Čumakov, G. M. 112, 306, 308, 310
D 307, 308, 311, 318
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 Halliday, M. A. K. 155, 307, 317
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 Issatschenko, A. V. 287
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,
Goody, Jack 203, 307
Goosens, Louis 10, 304
Gorčakov, 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
Gvozdanovićå, 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
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
Istrina, E. S. 15, 304, 306, 309, 310
Ivan Ivanovič, Grand Prince 197
Ivan III Vasil’evič, Grand Prince 197
Ivan IV Vasil’evič, Tsar’ 191
Ivanov, Vjač. Vs. 303, 304
Ivina, L. I. 308, 314
 N I
J L
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 Lee, Benjamin 135
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
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
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
Lixačev, D. S. 6, 65
Lixačev, 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
Martysevič, 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 Rader, Margaret 291, 294
Nemčenko, 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 306
Ong, Walter J. 46, 56, 307
Osipov, B. I. 305
Otin, E. S. 59–60, 306
P S
Padučeva, 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, Vjačeslav 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
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,
Rosen, Harold 67
Rozental’, D. È. 306
Ryle, Gilbert 303
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
Sprinčak, 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 161, 249, 303, 304, 306, 310
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 Ivanovič, 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,
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 Adjournment procedures 33, 81–82,
a 76, 78, 98–99, 101, 186, 192, 229,
246
a r(’)kuči 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; Voca-
tives
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
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 commands initiating 167, 172–73
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 Default-
judgment charters
Binary oppositions 24
Birchbark letters 13–14
biti/bivati čel”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
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
čelobitnye gramoty see Petitions
cělovati kr’st” 225, 315; see also Cross-
kissing oaths
 S I
Challenges 42, 58, 104, 188, 193,
214–15, 216
Chancery writing see Legal-
administrative 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 č’to; koe
Complement type 15, 21
Complementized direct speech see Direct
speech, complementized
Complementized indirect speech see
Indirect speech
Complementizers see also aže (až’);
č’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
č’to 12, 103, 114, 192, 199, 212, 225,
240, 252, 254, 266
Cursive 57
Cyril Belozersk Monastery 145
Czech 310
D Dialogue, complex 173
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
dě(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
dě(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 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 Referred-
judgment 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 88–89, 128, 146, 159, 191
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,
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 compared with nondirect speech
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
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 112, 184, 197, 217, 227, 253
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,
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 Imperatives 12, 154, 160, 225, 251
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 Indiction, dating by 30
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
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
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 reči) 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 dě(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 reči)
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
izvěčati 205
J quotative particles in 296–97, 316
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 Legal-administrative writing 8, 25, 28,
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
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
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 capitalization in 153–54
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”lčati/zam”lčati 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
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 as narrative reports 126, 210, 212,
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
očistiti 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
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 dě(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 č’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
Poručnye 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
kupčie gramoty) 190, 313
Purposive, definition of 3
Q Reductionist approaches 11–16
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 and institutional purpose 27, 30–31,
Raising 122–23, 259
Rapid-fire exchanges 165, 168, 233
raskazati 140
Rassprosnye reči 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
reči
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
rečisja 174, 312, 313
Reconvenements 172, 292
Recoverability see Givenness
Referral formula 33, 174, 190, 243,
312, 313
Referral memoranda (ssyločnye 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
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 with layered reported speech 212,
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”lčati/zam”lčati
“Silencing of prose” 310
Simonov Monastery 183
skazat’ 5, 85, 300
s(”)kazati/s(”)kazyvati
as default verbum dicendi 5, 88, 300
displacing reči 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
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
Ssyločnye 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 (poručnye 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 40–47, 59, 79, 81, 82, 85–86, 92,
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,
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 dělja č’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 in medieval Russian culture 56–57
učiniti dovod” 130
učiniti 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
velěti 125, 172, 175, 194, 242, 243,
251–52, 313, 316
Verba dicendi (’s) 5, 9, 10, 15, 24,
221, 309; see also dě(ja)ti, govoriti,
kazati, m”lviti, reči,
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
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”zvěstiti 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 Speaker-
based 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”lčati see m”lčati/zam”lčati
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 pub-
lished 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/Phila-
delphia, 1989.
5. THELIN, Nils B. (ed.): Verbal Aspect in Discourse. Amsterdam/Philadelphia, 1990.
6. VERSCHUEREN, Jef (ed.): Selected Papers from the 1987 International Pragmatics Con-
ference. 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. Am-
sterdam/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/Phila-
delphia, 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. Amster-
dam/Philadelphia, 1991.
13. NUYTS, Jan, A. Machtelt BOLKESTEIN and Co VET (eds): Layers and Levels of Repre-
sentation 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/Philadel-
phia, 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, func-
tionalism, 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. Amster-
dam/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 Japa-
nese 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 Tradi-
tional 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. Amster-
dam/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 SIMON-
VANDENBERGEN 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 impli-
cations. Amsterdam/Philadelphia, 1998.
38. FRETHEIM, Thorstein and Jeanette K. GUNDEL (eds): Reference and Referent Accessi-
bility. 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 close-
knit 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. Am-
sterdam/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/Phila-
delphia, 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. Amster-
dam/Philadelphia, 1998.
51. BOULIMA, Jamila: Negotiated Interaction in Target Language Classroom Discourse. Am-
sterdam/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 cogni-
tion. 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 Interna-
tional 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 rel-
evance-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 Eng-
lish. 2001.
72. MARMARIDOU, Sophia S.A.: Pragmatic Meaning and Cognition. 2000.
73. HESTER, Stephen and David FRANCIS (eds.): Local Educational Order. Ethno-
methodological 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 Propo-
sitional 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 Commu-
nication. 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 per-
spective. 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.