This page intentionally left blank
AReference Grammar of Russian
AReference Grammar of Russian describes and systematizes all aspects of the
grammar of Russian: the patterns of orthography, sounds, inflection, syntax,
tense-aspect-mood, word order, and intonation. It is especially concerned with the
meaning of combinations of words (constructions). The core concept is that of the
predicate history: a record of the states of entities through time and across
possibilities. Using predicate histories, the book presents an integrated account of
the semantics of verbs, nouns, case, and aspect. More attention is paid to syntax
than in any other grammars of Russian written in English or in other languages
of Western Europe. Alan Timberlake refers to the literature on variation and
trends in development, and makes use of contemporary data from the internet.
This book will appeal to students, scholars, and language professionals interested
in Russian.
alan timberlake is Professor of Slavic Linguistics at the Department of Slavic
Languages and Literatures, University of California at Berkeley. He is the author of
The Nominative Object in Slavic, Baltic, and West Finnic (1974) and editor of The Scope of
Slavic Aspect (with M. S. Flier, 1985), American Contributions to the Eleventh International
Congress of Slavists (with Robert A. Maguire, 1993), and American Contributions to the
Twelfth International Congress of Slavists (with Robert Maguire, 1998).

AReference Grammar of Russian
ALAN TIMBERLAKE
University of California at Berkeley
cambridge university press
Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, São Paulo
Cambridge University Press
The Edinburgh Building, Cambridge cb2 2ru, UK
First published in print format
isbn-13 978-0-521-77292-1

isbn-13 978-0-511-16446-0
© Alan Timberlake 2004
2004
Information on this title: www.cambrid
g
e.or
g
/9780521772921
This publication is in copyright. Subject to statutory exception and to the provision of
relevant collective licensing agreements, no reproduction of any part may take place
without the written permission of Cambridge University Press.
isbn-10 0-511-16446-7
isbn-10 0-521-77292-3
Cambridge University Press has no responsibility for the persistence or accuracy of urls
for external or third-party internet websites referred to in this publication, and does not
guarantee that any content on such websites is, or will remain, accurate or appropriate.
Published in the United States of America by Cambridge University Press, New York
www.cambridge.org
hardback
eBook (EBL)
eBook (EBL)
hardback
Contents
1 Russian 1
2 Sounds 28
3 Inflectional morphology 92
4 Arguments159
5 Predicates and arguments 270
6 Mood, tense, and aspect 371
7 The presentation of information 444

Bibliography 473
Index 493
v

1
Russian
1.1 The Russian language
1.1.1 Russian then and now
The present study is a comprehensive description of all aspects (except word
derivation) of modern standard Russian: its sounds, spelling, grammar, and
syntax.
Russian has resulted from a long evolution that can be traced back to the first
millennium of our era. From the fifth century on, speakers of Slavic established
settlements over a vast area of Central and Eastern Europe, from the Danube in
the south to the Elbe in the northwest. In the east, they moved north from the
Dnepr valley to the Gulf of Finland and the Upper Volga, gradually displacing
or assimilating the previous Baltic and Finnic inhabitants.
1
Russian developed
from the dialects of Slavic spoken in the north of this East Slavic territory. In
the ninth century, the East Slavic area came under control of Scandinavian
merchant-warriors. The Christianization of this land in 988 was followed by
subjugation to “the Mongol yoke” from the thirteenth century into the fifteenth
century. As the favored agent of the Golden Horde, the once small principality
of Moscow brought ever more land under its control. By the end of the fifteenth
century, when the Mongol yoke was definitively removed, Moscow had become
the political and ecclesiastical center of the East Slavic lands, and the center of
the Russian language area.
Russian is not only a spoken language, but a written language used for all
cultural purposes. The modern form of Russian took shape over the course of

the eighteenth century. The morphology and phonology is based on the dialect
of Moscow. In its vocabulary, syntax, and rhetoric, Russian, while relying on
native Slavic elements, has a long history of adapting and internalizing foreign
Byzantine, French, and most recently English models.
Parenthetically, it could be noted that the modern word héccrbq ‘Russian’ is an
adjective deriving from the noun Hécm ‘Rus’. According to a venerable etymology,
1
See Sedov 1982 on the complex archeological record of the East Slavic area.
1
2 AReference Grammar of Russian
Hécm wasadescriptive name for Scandinavians that is based on the Germanic et-
ymon ‘to row’, the Scandinavians being above all oarsmen.
2
In East Slavic lands,
Hécm was used initially for the Scandinavian overlords and their principality
of Kiev. Over time it was extended to all East Slavic lands. Muscovy appropri-
ated the name for its political identity, culture, and language as it consolidated
power.
Russian is the first language of approximately 150 million people. According
to an estimate for 2002 the Russian Federation had a total population of 145 mil-
lion people, among whom 81.5 percent, or 118 million, were ethnic Russians.
3
In
the mid-nineties, there were an additional 25 million Russians in the newly in-
dependent countries that emerged from the breakup of the Soviet Union (Novaia
Rossiia 1994). Together that would make 143 million ethnic Russians. To that
figure could be added a substantial though indeterminate percentage of the
remaining 27 million members of other nationalities residing in the Russian
Federation. According to recent statistics, the rate of population growth in the
Russian Federation is negative (−0.33%), from which it would follow that the

number of speakers of Russian will not increase in the foreseeable future.
1.1.2 Levels of language
Russian is a spoken language and a written language. In its written form Rus-
sian has long been highly codified: grammars, dictionaries, and manuals define
standards for usage that are enforced in the educational system and through
editorial practices in publication. Although the Russian tradition is quite clear
about what usage counts as standard, it does acknowledge the existence of a
range of varieties, or registers,from archaic to bookish to standard (normative)
to colloquial (hfpujdjhyfz htxm)tosubstandard and uncultured (ghjcnjhtxbt).
The grammar recorded here is the normative grammar of standard, written
Russian, which is the culturally privileged, and also the most accessible, form
of Russian. Occasionally, there are asides on usage in less-than-standard or oral
language, but this study cannot treat colloquial Russian with the same attention
as the works of E. A. Zemskaia and colleagues,
4
which have documented the sig-
nificant differences between spontaneous spoken Russian and formal, written
Russian.
2
Possible candidates are Ro
þ
er, Ro
þ
in,former names for Sweden’s Uppland region, and ro
þ
s- ‘oar’,
thegenitive form used in compounding (Thomsen 1879:99 104, also Vasmer 1986 87:s.v. Hecm,de
Vries 1962: s.v. rj´
ð
r,Schenker 1995:57 60). A form of this etymon was adopted into West Finnic

languages (Finnish ruotsi ‘Sweden’) and into Slavic, and then found its way into Greek (␳
<
ς ) and
Arabic (r
¯
us) sources from the ninth and tenth centuries.
3
At: />4
Zemskaia 1973, 1978, 1983; Zemskaia and Shmelev 1984; see also Timroth 1986.
Russian 3
Russian has undergone some change since the political and economic tur-
moil of the late eighties and early nineties, but it is difficult to assess how
much. Most tangibly, there have been changes in vocabulary.
5
Borrowing and
native derivational processes have produced many new words and word combi-
nations, leading to macaronic texts: ytqk-fhn ‘nail-art’, WEB-lbpfqy ‘WEB-design’,
Htrbq ,bhvbyutvcrbq lb-l;tq gj bvtyb Graham Mack lb-l;tbk ct,t, lb-l;tbk,
lf nfr b ljlbl;tbkcz, xnj c hflbj eitk ‘A certain Birmingham DJ, named
Graham Mack, DJ-ed, DJ-ed, and so DJ-ed out, that he had to leave the radio
station’. This internationalized vocabulary now dominates the linguistic land-
scape, just as Soviet-speak used to dominate language a half century ago. Along
with these changes in vocabulary has come a less quantifiable but still palpable
change in the mores of language. Unedited, informal texts of written Russian
of a type that would never have become public during the days of active So-
viet censorship are now available in print and especially electronic form. And
yet, despite political changes and a loosening of speech manners, contemporary
Russian in its grammatical structure remains Russian.
1.2 Describing Russian grammar
1.2.1 Conventions of notation

The notational conventions employed here are those of Table 1.1.
In the body of the text, Cyrillic words and phrases will be given in italics,
and English translations in single quotation marks. Stress is marked in citation
formsofwords or short phrases; stress is not marked on vowels in fragments of
text cited in the text or in set-off numbered examples. In numbered examples,
italics and quotations are not used.
1.2.2 Abbreviations
The abbreviations used in this study are listed in Table 1.2.
1.2.3 Dictionaries and grammars
The definitive dictionary of Russian in Russian is the Slovar

sovremennogo russkogo
literaturnogo iazyka,aseventeen-volume dictionary published over 1950 65. Self-
evidently it does not include the numerous new words from the last sev-
eral decades. Shorter Russian-language dictionaries are fully useful, notably
Ozhegov’s one-volume classic, which conveniently lists grammatical forms with
stress. More than adequate bilingual dictionaries are the Oxford dictionary (both
directions) and now the Novyi Slovar

(Russian to English), the most up-to-date
5
Zemskaia 2000.
4 AReference Grammar of Russian
Table 1.1 Conventions used
notation interpretation
nom sg ntnhƒlm grammatical gloss and Russian word
ntnhƒlm
<nom sg>
alternative grammatical gloss of Russian word
lj

<\gen>
grammatical form conditioned by another word (preposition or
verb)
≤=≥ spelling of letter (or word) in Cyrillic, when spelling is at issue
[ƒ] sound (from narrow phonetic through broad phonetic to
phonemic)
{ƒ} or {ƒ : ø : ə} vowel series,orset of stressed and unstressed vowels related
by etymology and/or synchronic alternation
{-ej} or -tq morphological unit
{X:Y} any relation of elements, notably two stems of verbs,
{CVC-a-
<pst/inf>
: CVC-aj-|e|-}
<prs>
}
X ∼ Ytwo forms potentially available in the same context
jcnƒnmcz/jcnfdƒnmcz aspect pair: perfective and secondary (derived) imperfective
vf[ƒnm\vf[yénm aspect pair: simplex imperfective and semelfactive perfective
(gj)ghjc∫nm or
ghjc∫nm\gjghjc∫nm
aspect pair: simplex imperfective and prefixed perfective
√
/ ± /?/
∗
hierarchy of acceptability judgments: neutral, acceptable,
frequent / less preferred option / restricted, marginal /
dubious, ungrammatical
dictionary available. A selection of dictionaries Russian only and bilingual is
available on the web.
Russian dictionaries, unlike many dictionaries of English, do not give infor-

mation about etymology, for which one should consult the dictionary of Max
Vasmer (in its original German edition of 1953 or the Russian edition of 1986 87
revised by O. N. Trubachev), nor about earlier usage, for which one should use
Srevnevskii’s “materials” for a dictionary of Old Russian from 1893 1912 (and
later reprints), Slovar

russkogo iazyka XI XVII vv., or Slovar

russkogo iazyka XVIII
veka.Lubensky (1995) should be consulted for Russian idioms.
Forgrammatical information, the “grammatical dictionary” of A. A. Zalizniak
(1977[a]), with 100,000 entries arranged in reverse alphabetical order, is defini-
tive. Entries of the dictionary are indexed with paradigm numbers; excep-
tions are marked. The 142 introductory pages list paradigms with accentual
contours.
Avariety of grammars is available, including two compact grammars in En-
glish (Unbegaun 1957, Wade 1992), which, however, do not treat syntax exten-
sively, as well as the multiple generations of “academy grammars” (for example,
Russian 5
Table 1.2 Abbreviations used
abbreviation interpretation
C/C/ /R/W set of consonants / obstruents / sonorants / {[v v˛]}
C¸/C
0
set of palatalized consonants / set of non-palatalized consonants
V/V
!
/V
*
set of vowels / stressed vowels / unstressed vowels

P/T/K/
ˇ
S consonant articulations: labial / dental / velar / alveo-palatal
C
0
/ C
j
/ C
i
/ C
i
␣
/ C
i
␣

/ C
i
␤
consonant grades (§2.5.2)
[z˛]/[rü]/[r3] palatalized [z] / voiceless [r] / voiced [r]
[´a5 ]/[´a5]/[´a55][a] fronted in initial transition / final transition / both transitions
ˇ
rrü articulation in which one feature changes over duration of
segment
nom / acc / gen / dat /
loc / ins
nominative / accusative / genitive / dative / locative / instrumental
gen1 / gen2 // loc1 /loc2 primary / secondary genitive // primary / secondary locative
nom=acc/acc=gen syncretism of nominative and accusative (“inanimate accusative”) /

syncretism of accusative and genitive (“animate accusative”)
sg / pl / du singular / plural / dual
msc / fem/nt masculine / feminine / neuter
an/in animate / inanimate
pv predicative (= “short”) adjective
nn / qu / adj / pss noun / quantifier / adjective / possessive
Declension
<I>
first declension: Declension
<Ia>
and Declension
<Ib>
Declension
<Ia>
first declension (masculine type with nom sg {-∅}: ,j´,)
Declension
<Ib>
first declension (neuter type with nom sg {-o -e}: cnƒlj)
Declension
<II>
second declension
Declension
<III>
third declension
Declension
<IIIa>
third declension (feminine with nom sg {-∅}: gkj´oflm)
Declension
<IIIb>
third declension (neuter with nom sg -z: dh†vz)

Declension
<IIIc>
third declension (masculine with nom sg {-∅}: génm )
R / E / A / F / T / M stress paradigms stress on: root / ending / classificatory suffix
(verbs) / antethematic syllable / thematic syllable / mobile stress
prs / pst / fut / inf /
imv / irr / rls / pcl /
dee/psv
present / past / future / infinitive / imperative / irrealis / realis /
participle / adverbial participle (lttghbxfcnbt)/passive participle
if/pf//dt/id imperfective / perfective // determinate (imperfective) /
indeterminate
1sg /1pl /2sg /2pl /3sg /
3pl
first-person singular / first-person plural / second-person singular /
second-person plural / third-person singular / third-person plural
dim diminutive
intg interrogative
/ B / ↔ /
B ↔B /
↔B
address by ns / address by ds / mutual address by ns / mutual
address by ds /asymmetric address, one speaker using ns,the
other ds /
У / И / О / Ф diminutive name / first name / patronymic / surname
s
jyf
v
dpzkf
o

vtyz
d
yf
,fpfh
m
,kfujgjkexyj
word order: subject verb object domain manner
6 AReference Grammar of Russian
RG 1980). The four-volume “functional grammar” is superb (Bondarko 1991 96).
Good grammars exist in other European languages (for example, Garde 1980 in
French, Isa
ˇ
cenko 1975 in German). The discussion below, though it is informed
by this tradition of grammatical analysis, does not cite them in the interests of
avoiding a clutter of references.
1.2.4 Statistics and corpora
To characterize how likely some construction is, it is often useful to cite statistics
of usage. At the same time, it is important to acknowledge the limitations on
statistical statements. The likelihood of using some or another morphological
form or syntactic construction is really the likelihood of using the context in
which the form or construction is appropriate; statistics ultimately measure how
likely people are to say a whole context. For example, if we find that the com-
bination e ytq is less frequent than e ytt, what we have really found is that the
contexts in which e ytq is appropriate occur less frequently than those in which
e ytt is appropriate. Any statistical statement, even one that appears to deal with
morphological variants, is a measure of the frequency of the contexts in which
these variants are appropriate. When the discussion below cites statistical obser-
vations, it is usually to say, informally and without pretense of scientific rigor,
that a certain construction occurs surprisingly often or not particularly often,
relative to what one might expect. The limitations on what statistical statements

mean should always be kept in mind.
As a corpus for making statistical observations, I initially used the “Uppsala
Corpus.” The corpus, assembled by the Slavic Institute of Uppsala University and
mounted on the web by the University of T
¨
ubingen,
6
offers a balanced selection
of styles of texts through the 1980s; it has its own search. As time went on, I
made use of the broader resources of the web. The address “ />has a vastly larger number of (belletristic) texts. By using a powerful search en-
gine (such as Google, Zndex, or Rambler), it is possible to search this site or
the whole web for words or phrases, and produce quantities of Russian larger
by orders of magnitude than the Uppsala Corpus. For example, in the Uppsala
Corpus, the target ins sg nsczxtq produced no tokens, the target e ytq five
tokens. In contrast, a search of (with Google, <20.X.02>)pro-
duced 233 hits for nsczxtq and 796 for e ytq; and on the whole web (with Google,
<20.X.02>), there were 8,790 hits for nsczxtq and 25,900 for e ytq. The new elec-
tronic resources, then, offer the possibility of vast quantities of Russian, most
of it very contemporary.
6
At: The description (<.
se/korpdesc.htm>)states that the corpus is based on 600 Russian texts, one million running words,
of informative (late 1980s) and literary texts (1960 88).
Russian 7
There are, however, some negatives, which grow in proportion to the size of
the corpus and the frequency of the target word or phrase. Unlike the Uppsala
Corpus, which was designed to serve as a corpus and has a balanced selection
of genres of texts, the web was not designed to serve as a corpus for linguistic
investigation. The web has properties that make it less than ideal as a corpus:
(a) the relative weight of genres www.libr.ru is heavy on literary texts and trans-

lations (if one has hesitations about translations), while the web as a whole has
arandom mix of commercial writing, personal travelogues, detailed histories
of the repair records of automobiles, journalism, and religious texts; (b) the
quality of Russian, which includes translations, sites from outside Russia, and
informal personal writing and commercial writing that is no longer subjected
to the same editing as was Russian printed in the Soviet era; (c) the fact that
many of the texts show up on more than one site, undercutting the value of
statistical observations; (d) instability the sites are not stable over time, im-
peding replication and verifiability; (e) the number of positive hits, which can
be so large that the finite amount of time it takes to evaluate any token makes
it difficult to examine all the data. The enormous volume of Russian available
now is a mixed blessing.
7
Allow me to cite cautionary tales. With respect to repetition: the phrase e;t
jnrhsdfk
<pst.if>
jryj ‘[he] already opened the window’ a familiar phrase in
aspectology gave a modest forty hits on the whole web (<20.XII.01>). But every
one of them was the same sentence from a text by A. Tolstoy. With respect to
stability, I searched the web for the expressions hfymit ytuj ‘earlier than him’
and d jnyjitybb ytuj ‘inrelation to it’, and came up with 1,590 and 5,490 to-
kens, respectively (<20.XII.01>). The same search nine months later (<15.IX.02>)
yielded 2,080 and 7,190 tokens an increase of 17 percent. With respect to quan-
tity: I searched the web (<20.X.02>)fortokens of nsczxtq 8,790 hits and
nsczxm/ 10,800 hits with the goal of finding out in crude terms the relative
frequency of these two forms of the instrumental case of nsczxf.Itwould take
perhaps eighty hours to evaluate all that data, if a modest fifteen seconds were
devoted to each token. In short, the investigator has no control over the web and
no way of determining what its properties as a corpus really are. The Uppsala
Corpus, though smaller, offers a more balanced corpus.

In light of such difficulties, it is important to emphasize the limitations on
citations from the web. All statistical statements made on the basis of the web
should be taken for what they are: informal characterizations of frequency over
unstable, often repetitive, collections of Russian assembled for other (commer-
cial, etc.) purposes than to serve as a corpus for linguistic investigation. The
corpus is not stable and one cannot control for repetition.
7
Browne 2001 explores the problems of using the web as a corpus.
8 AReference Grammar of Russian
In the same vein, it is also important to register the disclaimer that there is
no guarantee that specific websites, referred to occasionally below, will remain
valid.
1.2.5 Strategies of describing Russian grammar
The discussion of Russian below follows an unsurprising sequence: after these
preliminaries, ending with the writing of Russian, the discussion goes from
sound to morphology (grammar in the traditional sense) to syntax first argu-
ments, then predicates, then predicates in context (tense, aspect, modality)
and finally, selected discourse operations that apply to the presentation of in-
formation. Obviously there are many topics that belong in two places tense in
participles is a question of morphology and of predicate semantics in context;
the second genitive is a question of morphology, of arguments, and of predicates
(since the use of the second genitive depends on the syntactic context) and it
was necessary to make decisions about where to put discussion. Cross-references
are provided.
Aword about the philosophy of grammar invoked here. Modern linguistics has
prided itself on identifying basic, primitive elements (phonemes, morphemes,
constituents of sentences) and their rules of combination. For some researchers,
the ultimate goal is to characterize which sentences are possible, which impos-
sible, and to state the rules of combination. My experience in assembling this
grammar has led in a different direction. Repeatedly I found that what was

significant was the construction the pattern, the configuration, the template
(nhfafhtn
8
). Patterns include all manner of linguistic knowledge: constituent
elements; typical lexical items that participate; strategies of interpreting the
meaning, or value, of the pattern in discourse; stylistic value in short, pat-
terns include all kinds of linguistic knowledge. The semantic, pragmatic, and
stylistic values of a construction are not entirely predictable from its primitive
elements and rules of combination, and though any construction certainly con-
tains smaller entities, it is not always possible (or important) to identify the
primitive elements. It becomes more important to say in what contexts, and
with what meaning, a construction can be used. The whole is often greater
than its parts. For example, the free (dative) infinitive construction (yfv
<dat>
yt
vbyjdfnm
<inf>
ub,tkb ‘itisnot for us to avoid disaster’, ,tp htdjk/wbb yfv
<dat>
yt lj,bnmcz
<inf>
cjdthitycndf ‘without a revolution it is not for us to achieve
perfection’) has recognizable parts: an infinitive, a dative that would be the sub-
ject if the infinitive were a finite verb, and the other argument phrases governed
by theverb. There is no overt finite verb; no form of ,ßnm ‘be’ is used in the
8
Zhivov and Timberlake 1997.
Russian 9
present tense. The meaning of this construction it makes a prediction about the
possibility of an imagined event cannot be computed just from its constituent

parts, the dative and the infinitive. Moreover, the construction has different
variants, each of which has a specific stylistic value. The variant just illustrated
is folksy, apodictic. Another variant of the construction used in content ques-
tions is neutral and productive, as in, Rfr gjgfcnm d yfxfkj cgbcrf yfqltyys[
cfqnjd yf gjbcrjds[ vfibyf[?‘How [is it possible] to get to the beginning of
the list of sites in search engines?’ Indeed, the initial portion of this question,
Rfr gjgfcnm ‘how [is it possible] to reach . . .’, produced 18,900 hits on
the whole web (<20.X.02>). In general, then, the presentation of Russian gram-
mar below emphasizes whole combinations and their value (including stylistic),
downplaying the task of identifying primitive elements or articulating notations
for encoding rules of combination.
When there are two closely related constructions that differ by one linguistic
form for example, relatives made with rnj´ vs. rjnj´hsq,genitive vs. accusative
with negated verbs, etc. it is an interesting question how speakers choose be-
tween the variants. In a notational approach to grammar, one can always create
different structures that will produce different cases (for example). But because
thestructures will be distinct, there is no way of comparing the properties that
distinguish them the properties of the noun phrases, the discourse import
and such an approach says nothing about how speakers make choices. As an
alternative, one can look for as many tangible variables as possible variables
such as the number of a noun, its position relative to the verb, the aspect of
theverb and measure their statistical contribution. But the result of a variable
rule is only a probability, which does not explain how a speaker works with a
half dozen to a dozen factors and makes a choice that is binary to use one
construction or another. In the following, I assume that speakers operate with
templates (constructions) that have multiple properties lexical to syntactic to
discourse. In any instance, speakers ask which template a given utterance better
matches. This is a holistic decision: in the genitive of negation, perhaps, speak-
ers evaluate a context as being concerned with absence of a situation (genitive)
as opposed to reporting an entity’s properties (accusative). To get to this holistic

judgment, speakers ask which template better fits the context. And to answer
that question, speakers probably have to select one feature to pay attention to,
while others are ignored. In practical terms, this means it is difficult, for many
constructions, to give watertight rules about usage (there are too many variables;
speakers have some freedom in how they rank and evaluate variables). What can
be done is to point out the general, holistic value of a construction, and, often,
some tangible linguistic features that are consistent with that holistic value that
will influence choices.
10 AReference Grammar of Russian
1.2.6 Two fundamental concepts of (Russian) grammar
While each construction, each problem of grammar, requires its own descrip-
tion, some general, recurrent ideas emerged. Two can be mentioned.
One is modality and the related concept of quantification. Every statement is
understood against alternatives. Sometimes there is just a contrast of the mere
fact that some x having one salient property exists at all, in contrast to the
possibility that x might not hold, or that a certain situation ␾ holds in contrast
to the possibility that ␾ might not exist (existential or essential quantification).
Sometimes a specific individual x or property ␾ is contrasted with other possible
x’s or ␾’s (individuated quantification). Modality consideration of alternatives
by an authority pervades grammar.
The other is directionality, dialogicity. An utterance does not exist or have
meaning in isolation, but is manipulated by speakers and addressees in a three-
step process. The speaker invites the addressee to construct a background of
information, taken as given and known (first step). Against this background
the speaker formulates, and the addressee evaluates, the current assertion (sec-
ond step). On the basis of that comparison, the speaker and addressee then
project further conclusions or anticipate further events (third step). Thus the
speaker invites the addressee to engage in a directional process of manipulating
information.
These concepts modality (and quantification) and directionality pervade

thegrammar of Russian and, no doubt, other languages.
1.3 Writing Russian
1.3.1 The Russian Cyrillic alphabet
Russian is written not in the Latin letters used for English and Western Euro-
pean languages but in an alphabet called Cyrillic (Russian rbhbkkbwf). Cyrillic,
with small differences, is also used for other languages Ukrainian, Serbian,
Bulgarian. Cyrillic will be used to write Russian throughout the discussion be-
low, with certain obvious exceptions: in the discussion of sounds and the inter-
nal structure of words, in glosses of Russian words or phrases, and in citations
of scholarly literature. For reference, the version of the Cyrillic alphabet used for
modern Russian is given in Table 1.3. In Column 1 the alphabet is presented in
thelower- and uppercase forms used in printing. Column 2 gives the italic vari-
ants. Column 3 gives longhand forms of lowercase and then uppercase letters as
used in connected, cursive writing (unusual uppercase letters are omitted); the
subsequent discussion, however, will not treat handwriting.
9
The contemporary
9
With thanks to Victoria Somoff for the handwriting sample.
Russian 11
name of the letter is given in Column 4. These names are mostly transparent.
The names of consonant letters have a vowel added to the sound of the conso-
nant. Four unusual letters are referred to by descriptive phrases. For reference,
Column 5 gives the older names of the letters. Column 6 states approximate
sound values of individual Cyrillic letters in English, although there are obvi-
ous difficulties in attempting to state the sound of Cyrillic letters in terms of
English sounds: the closest English sound is not always particularly close; in-
dividual Cyrillic letters do not represent just a single sound (consonants can
be palatalized or not; vowel letters have different value depending on whether
or not they follow consonant letters). The statements of sound value are quite

approximate.
Because Cyrillic is an alphabet, by establishing correspondences between each
individual Cyrillic letter and one or more Latin letters, it is possible to rewrite, or
transliterate, Cyrillic into Latin letters. Column 7 is the table of equivalences
established by the Library of Congress as used in slightly simplified form in this
study. (Other systems are discussed later: §1.3.7.) The final column gives sources of
the Cyrillic letters. The alphabet given in Table 1.3 is the contemporary alphabet.
The civil alphabet used until the reform of the October Revolution included two
additional letters: ≤î≥ “b ltcznthbxyjt” (alphabetized between ≤b≥ and ≤r≥) and
≤˜≥ “znm” (between ≤m≥ and ≤э≥). Additional letters are found in Russian Church
Slavic.
10
From various people, one often hears that Russian must be a difficult lan-
guage because its alphabet is so difficult. Nothing could be further from the
truth. Whatever the difficulties of Russian, they cannot be blamed on the al-
phabet, which anyone with a modicum of ability in language systems and a
vague acquaintance with the Greek alphabet can learn in half an hour, as will
be demonstrated after a brief introduction to the history of the alphabet.
1.3.2 A brief history of the Cyrillic alphabet
The beginning of writing in Slavic is a fascinating tale that deserves to be told
in brief.
11
The story can be picked up at the end of the eighth century, around
796, when tribes of Slavs from the region of Moravia (in the south of the con-
temporary Czech Republic, along the Morava River) helped Charlemagne rid
Central Europe of the last remnants of the Avars, a confederation of Eastern ma-
rauders. This venture marked the beginning of more active relations between
Moravian Slavs and the West, both with secular political authorities (Charle-
magne until his death in 917, his descendants thereafter) and with ecclesiastical
10

Library of Congress Romanization: ≤î≥ > ≤
¯
ı≥, ≤˜≥ > ≤
ˇ
ıe≥. Russian Church Slavic used also ≤º≥ >
≤
˙
f≥, ≤v≥ > ≤
˙
y≥
11
Dvornik 1970, Vlasto 1970.
Table 1.3 The Russian Cyrillic alphabet
Cyrillic Cyrillic Cyrillic contemporary archaic letter English equivalent Library of Congress etymology
(print) (italic) (longhand) letter name name (very approximate) Romanization of letter
f/F f/F ffpMasha | all aGkA
,/< ,/<
,э ,erb bother bGkB
d/D d/D
dэ dtlb Volga vGkB
u/U u/U
uэ ukfujk guard gGk
l/L l/L lэ lj,hj do dGk
t/T t/T
t [åt] tcnm Pierre | yell eGkE
=/+ =/+
= [åj] Fyodor | yoyo e Cyr t/T
;/: ;/:
;э ;bdbnt azure, French je zh Gl æ
p/P p/P

pэ ptvkz zoo zGkZ
b/B b/B
bbbeat | eat iGkH
q/æ q/Q
b rhfnrjt boy i Cyr b/B
r/R r/R
rf rfrj car kGkK
k/K k/K
эkm k/lb Leeds lGk
v/V v/V
эv vsckbnt Masha mGkM
y/Y y/Y
эy yfi no nGkN
j/J j/J
jjygo | only oGkO
g/G g/G
gэ gjrjq pot pGk
h/H h/H
эh hws rot rGkP
c/C c/C
эc ckjdj sew sGk
Table 1.3 (cont.)
Cyrillic Cyrillic Cyrillic contemporary archaic letter English equivalent Library of Congress etymology
(print) (italic) (longhand) letter name name (very approximate) Romanization of letter
n/N n/N nэ ndthlj toe tGkT
e/E e/E
eerdo | oops uGkY
a/A a/A
эa athn far fGk
[/{[/{

[f [th German ach kh Gk {
w/W w/W
wэ ws tsetse, prints ts Gl Ö
x/X x/X
xt xthdm church ch Gl ÷
i/I i/I if if shallow, fish sh Gl
Ø
o/O o/O of inf Josh should, fish shop shch Gl ù
( (
ndthlsq pyfr (th
∞
) th [boundary marker]

Gl ú
s/S s/S
s, ths
∞
ths pituitary yGlû
m m
vzurbq pyfr (thm
∞
) thm [consonant palatalized]

Gl ü
э/э э/э
э j,jhjnyjt best | Evan e Cyr t/T
//? //?
//cute | yule iu Gl
z/Z z/Z zzDiaghilev | Yalta ia Cyr
b

x | y = pronunciation after consonant letter | pronunciation not after consonant letter
Gk = Greek
Gl = Glagolitic
Cyr = (earlier) Cyrillic
x
∞
= older name still used (SRIa 1.152)
14 AReference Grammar of Russian
authorities. As part of this interaction, missionaries were sent to the Mora-
vians from the Franks (from the relatively new bishoprics of Regensburg, Passau,
Salzburg) and from the Italians (from the bishopric of Aquileia). The conversion
of Prince Mojmír of Moravia (r. 818 46) in 822 was followed by a general baptism
in 831. In this period of missionary activity, churches some in stone were
constructed at sites in Moravia such as Mikul
ˇ
cice.
In 846, Mojmír’s nephew Rostislav took control and began to act with greater
autonomy. After the bishopric of Salzburg had its charter renewed in 860, Ros-
tislav took steps to avoid further ecclesiastical interference from the Franks. In
862, after having been put off by the Pope, he approached the Byzantine Emperor
Michael III with a famous request:
Though our people have rejected paganism and observe Christian law, we have not
ateacher who would explain to us in our language the true Christian faith, so that
other countries which look to us might emulate us. Therefore, O lord, send us such
a bishop and teacher. (Kantor and White 1976:45)
Emperor Michael and Patriarch Photius responded by sending Constantine
(canonized as St. Cyril) and Methodius, two brothers educated in Greek who
spoke a Slavic language, to Moravia to train disciples and translate the liturgy
and the Bible into Slavic. In order to write in Slavic, they devised an alphabet
which is now called Glagolitic. The letters of Glagolitic are stylized combinations

of strokes and loops; for example, the chapter title for Luke 11 (Marianus) reads
in Glagolitic,
‘onthe catching
of fish’).
12
It is still an open question what sources Constantine and Methodius
used for this new alphabet. It has long been assumed that the model was Greek
minuscule,
13
but it may have been cursive of a Latin (specifically Carolingian)
type.
14
Whatever the source of the alphabet, writing in Slavic has its origins in
the “Moravian mission” of Constantine (St. Cyril) and Methodius.
The Moravian mission began auspiciously. It was given papal approval when
thebrothers traveled with their disciples to Rome (867). After Constantine died
in Rome (869), Methodius was appointed bishop of a large missionary area in-
cluding Moravia and Pannonia. In the long run, however, the mission proved vul-
nerable. It was resented by the Frankish bishops, who went so far as to imprison
12
Jagi
´
c1883 (interleaf 110 11, 186).
13
Beginning with Taylor (1880, 1883), who exhibited apparent similarities between individual
Glagolitic letters and Greek minuscule letters.
14
Lettenbauer 1953 (summarizing an inaccessible study, Hocij 1940) cites intriguing pairs of
Glagolitic and Carolingian letters. For example, the Carolingian ≤j≥ is a vertical arc open on
the left, with loops both on the top and at the bottom, hence very similar to the double loop

of Glagolitic ≤î≥;Taylor’s Greek cursive omicron has no loops. Taylor’s Greek cursive ≤l≥ looks
like a modern English cursive ≤l≥, with an internal loop (that is, ≤≥), very unlike the Glagolitic
double-looped ≤ä ≥, which looks like the Carolingian.
Russian 15
Methodius until the Pope secured his release. Rostislav, the Moravian prince who
originally sponsored the mission, was blinded and exiled. When Methodius died
in 885, a hostile bishop (Wiching of Nitra) chased out the troublemakers and
reinstalled the Latin rite. Disciples of Constantine and Methodius were fortunate
to make it to Ohrid and Bulgaria.
In Bulgaria, Tsar Boris, who had initially converted to Christianity in 863,
held a council in Preslav in 893, at which he abdicated, turned over power
to his pro-Christian son Symeon, and appointed Clement, one of the original
Moravian disciples, as bishop. Around this time, conceivably at this council,
15
thepractice was established of writing religious texts in Slavic in letters that
were modeled to the extent possible on Greek majuscule letters.
16
(For Slavic
sounds that had no equivalents in Greek, letters were adapted from Glagolitic.)
This neophyte Christian culture, with sacred texts written in Slavic in this Greek-
like alphabet, flourished in Bulgaria in the first half of the tenth century, until
the time (in 971) when Byzantium defeated Boris II and absorbed the Bulgarian
patriarchate. This tradition of writing was brought to Rus as a consequence of
the conversion to Christianity in 988. The alphabet that was imported was the
direct ancestor of the alphabet in which modern Russian is written, the alphabet
we call “Cyrillic.” As this brief sketch shows, Cyril himself did not invent the
Cyrillic alphabet. But he and his brother did invent the alphabet in which Slavic
was written systematically for the first time, and the alphabet they constructed
did provide the model for Cyrillic.
After having been brought into East Slavic territory, this alphabet was used

in the oldest principalities of Kiev, Novgorod, and Vladimir-Rostov-Suzdal from
the eleventh century on, and then subsequently in Moscow, the principality
that emerged as dominant as the “Mongol yoke” was loosened. This alphabet
has continued to be used with only modest changes until the present day. Peter
theGreat attempted to reform the orthography in 1708 10. His new civil al-
phabet (uhf;lfyrf) had letters of a cleaner, less ornate (more Western) shape.
Peter also proposed that, in instances where more than one letter had the same
sound value, only one letter be preserved, the first of the sets
for the
sound [i],
for [z], for [o], ≤y/U≥ for [u], ≤a/f≥ for [f]; some other
letters with quite specific functions
were also to be eliminated.
17
Al-
though all of Peter’s proposals did not catch on, his initiatives led to modernizing
thegraphic shape of the alphabet and set in motion the process of rationaliz-
ing the inventory of letters. While the general trend has been to simplify the
15
Dvornik 1970:250 52; Vlasto 1970:168 76.
16
The similarity is quite striking between early Cyrillic writing and contemporary Greek Gospels,
forexample Lord Zouche’s gospel text from 980 (Plate IV, Gardtgauzen 1911).
17
Zhivov 1996:73 77.
16 AReference Grammar of Russian
inventory of letters, ≤q э =≥ were introduced in the course of the eighteenth
century.
Russian Cyrillic took its contemporary form in a reform of October 1918, which
built on the results of earlier commissions (most immediately, the commission

of 1917). The notable changes were that remaining duplicate letters were elim-
inated (≤b≥ in place of ≤î≥, ≤t≥ for ≤˜≥, ≤a≥ for ≤f≥) and the “hard sign” ≤(≥
was eliminated from the ends of words after consonant letters, where it had
previously been required. For example, nineteenth-century ≤,˜c(≥ ‘demon’ be-
came ≤,tc≥.Other changes concerned the spelling of specific morphemes (for
example, adjectival msc sg ≤juj≥ in place of ≤fuj≥).
The principles established in 1918 were canonized by the publication of Rules
of Russian Orthography (= Pravila)in1956. The principles and detailed rules have
largely been stable, despite occasional discussions of possible further reforms
of some annoying but in the larger scheme of things, insignificant incon-
sistencies (for example, in 1964).
18
There was uncertainty, and continues to be
uncertainty, with respect to the vexed question of how much to use ≤=≥.Other
unresolved questions include: use of the hard sign ≤(≥ as mark of separation;
spelling of ≤b≥ or ≤s≥ after ≤w≥; spelling of ≤t(=)≥ or ≤j≥ after ≤;ixo≥;
spelling of ≤mj≥ and ≤qj≥ in borrowings; use of ≤э≥ after consonants; use of
double letters in borrowings. At this moment, there is a renewed impetus to
address certain details of writing, notably those involving compounds.
19
1.3.3 Etymology of letters
As noted, most Cyrillic letters were based on Greek upper case (majuscule) let-
ters. Many of the contemporary Cyrillic letters look like Greek letters, and as a
first approximation they can be read as one might expect on that basis. Among
Cyrillic letters for consonants, we observe the following similarities (Greek ma-
juscule prototypes are written in parentheses; the approximate sound value is
recorded in Table 1.3): ≤u/U≥ (Greek ); ≤l/L≥ (); ≤p/P≥ (Z); ≤r/R≥ (K); ≤k/K≥ ();
≤v/V≥ (M); ≤y/H≥ (N); ≤g/G≥ (); ≤h/H≥ (P); ≤c/C≥ (␴/); ≤n/T≥ (T); ≤a/A≥ ();
≤[/X≥ (X). From the single Greek ␤/B, Cyrillic has ≤,/<≥ (a bilabial stop [b]) and
≤d/D≥ (a labio-dental fricative [v]).

The consonant sounds of Slavic that did not have obvious correspondences
have unique symbols without any obvious source in the Greek or Latin alpha-
bets; they apparently derive from Glagolitic, which did have distinct letters for
these sounds: ≤;/:≥, ≤w/W≥, ≤x/X≥, ≤i/I≥, ≤o/O≥. Though these letters are
unfamiliar, sounds somewhat similar to those represented by these letters occur
18
Comrie, Stone, and Polinsky 1996 (ch. 8) gives a comprehensive survey from 1917 forward (see also
Chernyshev 1947). For the eighteenth and nineteenth centuries, see Grot 1873.
19
Proposals and rejected changes were accessible on www.gramota.ru/ <01.XII.01>.
Russian 17
in European languages. The most exotic is the sound spelled as ≤o/O≥,acon-
sonant of double length; it can be approximated by combining two tokens of
the sound written in English as ≤sh≥ in two words: Jo
sh sh
ould, fi
sh sh
op.
Vowel letters are largely based on Greek prototypes. As discussed below, there
are two parallel sets of vowel letters. In the first set (hard-vowel letters) we
find: ≤f/F≥ (Greek ␣/A), ≤э/Э≥ (an innovation, based on older Cyrillic ≤æ≥), ≤j/J≥
(Greek o/O), ≤e/E≥ (Greek ␷ /Y), ≤s≥ (derived from a combination of two letters,
the uniquely Slavic letter ≤(≥ and the Slavic adaptation of Greek ì/I). The sound
corresponding to ≤s≥ is perhaps the single most difficult for non-natives to
pronounce. Some Russians use this sound as a substitute for the vowel of p
i
t
or h
i
p in speaking English. A closer approximation would be a vowel that

changes from an [u]-like vowel to an [i]-like vowel, something like pit
ui
tary or
ph
ooey
, but pronounced as one syllable, not two. In the other set of vowel
letters (soft-vowel letters), two derive from Greek: ≤t/T≥ (from Greek ε), pro-
nounced as [e], and ≤b/B≥ (Greek ␩/H), pronounced as [i]. One has a source in
Glagolitic (≤//?≥ = the sound [u]) and two others arose in the history of Russian
Cyrillic writing (≤z/Z≥ = the sound [a]; ≤=/+≥, derived from Cyrillic ≤t/T≥ = the
sound [o]).
Identifying the etymology of letters does not, of course, explain how the Cyril-
lic alphabet works. But it should make it clear that the majority of the letters, in
their graphic shape and (approximate) sound value, are familiar from a cursory
acquaintance with the Greek alphabet.
1.3.4 How the Cyrillic alphabet works (basics)
The Cyrillic alphabet is a good guide to pronunciation. It is generally clear how a
sequence of letters should be pronounced. One complication is that in every word
in Russian one vowel is strongly stressed, and the remaining unstressed vowels
are pronounced less clearly than the one stressed vowel (unstressed vowels are
“reduced”). Once one knows which syllable is stressed, phonetic reduction is not
difficult for speakers of English. Unstressed vowels are commonly the indistinct
“schwa” vowel; Russian Vƒif is pronounced with [ə]inthe second syllable, thus
[mƒ
ˇ
sə], much as the final vowel of the English version of this name, Masha,is
spoken. However, most writing does not indicate which vowel is stressed. In this
respect, spelling does not give complete information about pronunciation.
To understand how the Russian Cyrillic alphabet works, it is necessary to
mention one fact about consonant sounds. Most consonants can be pronounced

in two significantly different ways: not palatalized, when they are somewhat
similar to consonants in English, or palatalized, when the tongue is raised to-
wardsthe front and top of the mouth, towards the area behind the teeth. The
effect of palatalization is similar to the beginning of English few, pew,or,in one