INFORMATION TO USERS

This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy sutxnitted. Thus, some thesis and dissertation copies are in typewriter ^ce, while others may be from any type of computer printer.

The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction.

In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.

Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand comer and continuing from left to right in equal sections with small overlaps.

Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6” x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order.

Bell & Howell Information and Leaming 300 North Zeeb Road, Ann Arbor, Ml 48106-1346 USA 800-521-0600

____ ® UMI

THE EVOLUTION OF THE SLAVIC ‘BE(COME)’- TYPE COMPOUND FUTURE

DISSERTATION

Presented in Partial Fulfillment of the Requirements for

the Degree Doctor of Philosophy in the Graduate

School of The Ohio State University

By

Marika Lynn Whaley, M.A.

*****

The Ohio State University

2000

Dissertation Committee: Approved by

Dr. Daniel E. Collins, Adviser

Dr. Charles E. Cribble Adviser

Dr. Brian D. Joseph Department of Slavic and East European Languages and Literatures UMI Number 9983007

Copyright 2000 by Whaley, Marika Lynn

All rights reserved.

UMI

UMI Microform9983007 Copyright 2000 by Bell & Howell Information and Leaming Company. All rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code.

Bell & Howell Information and Leaming Company 300 North Zeeb Road P.O. 80x1346 Ann Artx)r, Ml 48106-1346 Copyright by Marika Lynn Whaley 2000 ABSTRACT

Among the Slavic languages, the standard languages of Russian, Ukrainian.

Belarusan, Polish, Kashubian, Sorbian, Czech, Slovak, and Slovene, as well as some

Serbo-Croatian dialects, can all express futurity with a construction comprised of an auxiliary derived from the nonpast of *byti ‘be(come)’ in combination with either the infinitive or the /-participle. In addition, the standard language and many dialects of

Serbo-Croatian use this type of construction to express future-perfect meaning—a usage also known from Old Church Slavonic and Rus'ian. Despite the wide distribution of this type of construction in Slavic, a 'be(come)-type future cannot be reconstructed for the proto-language Common Slavic.

The question of how a 'be(come)-type future developed in Slavic is difficult to answer with certainty. Previous scholarship has sought the source of this future in either non-Slavic material or as a development internal to Slavic. Problems with previous theories are encountered with synchronic issues concerning the morphosyntactic constraints on be-future constructions and the semantics of tense and aspect, as well as diachronic issues concerning textual evidence and plausible paths of grammaticalization.

Thus the subject, though much studied, is still unresolved.

This study draws on recent general-linguistic work on the development of

u grammatical categories to provide a fresh perspective on this problem. It reevaluates the commonly known hypotheses on the development of the Slavic be-fiiture and presents a new, comprehensive analysis of the evolution of the construction. Its approach is based on the idea that semantics and semantic change motivate syntactic change. By systematically analyzing the semantics of the change-of-state verb *byti using reductive paraphrases, this study shows that the verb most likely grammaticalized into a future auxiliary autochthonously. Moreover, a comparison of the reductive paraphrases of change-of-state and inceptive verbs shows that the coUigability constraints on many

"be(come)'-type futures can be explained by positing that the change-of-state verb underwent a semantic shift into an inceptive. This new perspective provides a more satisfactory synchronic and diachronic description of this type of future than has been presented previously.

m To Steve, ljubimyj

IV ACKNOWLEDGMENTS

First and foremost, I would Like to acknowledge my husband Steve Fridella, my sine qua non.

I also express my sincerest thanks to my adviser Daniel Collins for all the time, energy, and ink he has devoted to my endeavor. He has always challenged me to become a better scholar and I have benefited greatly from his guidance. All uses of intercalated

“however” in this work, however, are entirely my responsibility.

1 am also grateful to those who found the time to discuss and evaluate my work:

Way les Browne. Joan Bybee, Bernard Comrie, and the other members of my committee,

Charles Cribble and Brian Joseph.

Finally, 1 wish to thank the many people who lent their support and encouragement to me whUe I was in the process of writing this dissertation: my family, my peers in the Slavic Department graduate program, and all my friends who knew better than to ask me for a detailed explanation of my subject matter. VITA

March 11. 1973 ...... Bom - Stow, Ohio, USA

1991...... B.A. History, College of William and Mary

199 4...... M.A. Slavic and East European Languages and Literatures, The Ohio State University

1993-2000...... Graduate Teaching and Research Associate, The Ohio State University

PUBLICATIONS

1. Marika Whaley. 1999. Tracing the origins of the Slavic imperfective be-future. Ohio State University Working Papers in Linguistics 52, 159—71.

FIELDS OF STUDY

Major Field: Slavic and East European Languages and Literatures

VI TABLE OF CONTENTS

Page

Abstract ...... ü

Dedication...... iv

Acknowledgments ...... v

V ita ...... vi

List of Tables...... x

List of Figures...... xi

Chapters:

1. Introduction ...... I

1.1 Overview of the problem ...... I 1.2 The be-future in S la v ic ...... 2 1.2.1 Future-tense marking in Slavic ...... 4 1.2.2 The status of the imperfective future tense ...... 6 1.3 Outline of the study ...... 11 1.3.1 Analysis of the data ...... 11 1.3.2 Evaluation of previous scholarship...... 14 1.3.3 Theoretical framework ...... 15 1.3.4 A new theory of the Slavic be-future ...... 18

2. The Slavic Data ...... 20

2.0 The Proto-Indo-European future ...... 20 2.1 The etymology and meaning of*bqd~ ...... 21

Vll 2.2 The origins of the Slavic be-future...... 23 2.2.1 Old Church Slavonic ...... 23 2.2.2 Placing the be-future within the context of chronology...... 25 2.3 Slovene ...... 26 2.3.1 Usage ...... 26 2.3.2 Chronology...... 27 2.4 Serbo-Croatian...... 28 2.4.1 Usage ...... 28 2.4.2 Chronology...... 31 2.5 C z ec h ...... 34 2.5.1 Usage ...... 34 2.5.2 Chronology...... 35 2.6 Slovak ...... 37 2.6.1 Usage ...... 37 2.6.2 Chronology ...... 38 2.7 Sorbian ...... 39 2.7.1 Usage ...... 39 2.7.2 Chronology...... 40 2.8 PoUsh...... 41 2.8.1 Usage ...... 41 2.8.2 Chronology...... 43 2.9 Kashubian ...... 50 2.9.1 Usage ...... 50 2.9.2 Chronology...... 53 2.10 Belarusan ...... 53 2.10.1 Usage ...... 53 2.10.2 Chronology...... 54 2.11 Ukrainian ...... 58 2.11.1 Usage ...... 58 2.11.2 Chronology...... 59 2.12 Russian ...... 61 2.12.1 Usage ...... 61 2.12.2 Chronology...... 62 2.13 Concluding rem arks ...... 72

3. Previous Scholarship ...... 76

3.1 Introduction ...... 76

vni 3.2 Extemal-source theories...... 77 3.2.1 The putative unusualness of a ‘become'-type future ...... 81 3.2.2 Chronology and the emergence of the be-future ...... 86 3.2.2.1 Reliability of the textual evidence ...... 86 3.2.2.2 The chronology in Russian ...... 88 3.2.2.3 Ukrainian evidence ...... 89 3.2.2.4 The chronology of Germanwerden ...... 90 3.2.3 Infinitival versus participial complements...... 91 3.2.4 The Slavic be-future and aspect ...... 92 3.2.5 Considerations of language contact...... 96 3.2.6 Conclusions ...... 98 3.3 Intemal-source theories ...... 99 3.3.1 The role of the impersonal obligative construction...... 99 3.3.2 The role of the future perfect...... 102 3.3.2.1 Textual evidence ...... 105 33.2.2 The semantics of the future perfect ...... 109 3.3.2.3 Conclusions ...... 116 3.3.3 The role of phase v e rb s...... 117 3.3.4 The role of lexical nuances ...... 119 3.4 Conclusions ...... 125

4. The Grammaticalization of*bqd-: From Change-of-State to Future ...... 127

4.1 Introduction ...... 127 4.2 The nature of syntactic c h a n g e...... 129 4.2.1 Grammaticalization ...... 129 4.2.2 Mechanisms of syntactic change...... 132 4.2.3 Syntactic and semantic change ...... 133 4.3 The development of the Slavic be-future ...... 135 4.3.1 Definitions ...... 136 4.3.2 Change-of-state and inceptive meaning ...... 137 4.3.3 Supporting evidence ...... 141 4.4 The Slavic be-future as a manifestation of drift ...... 143 4.5 The participial complement ...... 144 4.6 Conclusions ...... 151

5. Conclusion...... 152

Bibliography...... 156

IX LIST OF TABLES

Table Page

1.1 The expression T (masc. sg.) will write” in the Slavic languages which have a become-type future ...... 3

1.2 The nonpast tenses of Russian as demonstrated by the verbs pisat’ / napisat’ ‘write’...... 6

2.1 Polish future complements through the fifteenth century...... 44

2.2 Polish future complements through the sixteenth century...... 47

2.3 Kashubian future complements in the twentieth century: written d a ta ...... 52

2.4 Kashubian future complements in the twentieth century: spoken d ata ...... 52

4.1 Correspondence of languages with the be-future to those with pluperfects ... 147 LIST OF FIGURES

Figure Page

2.1 Schematic map of the be-future in the Slavic languages: Earliest dates of attestation/earliest dates of extant tex ts...... 74

2.2 Schematic map of the be-future in the Slavic languages: T reatment of complement and asp ect...... 75

XI CHAPTER 1

INTRODUCTION

1.1 Overview of the problem

For over one hundred years, scholars have debated the origin and development of the Slavic compound future constructions formed with reflexes of the nonpast form of the

Common Slavic verb *byti 'be(come)'. Whereas the earliest attested Slavic languages show future constructions formed with verbs such as ‘want’, have’, and begin’ in combination with an infinitive, many modem Slavic languages express futurity through an auxiliary derived from the future of *byti be(come)’ in combination with either the infinitive or the resultative participle (also called the /-participle).

The question of how these previously used future expressions were replaced by a

‘be(come)’-type future (hereafter shortened to “be-future”) is difficult to answer with certainty. Problems are encountered with synchronic issues such as the morphosyntactic constraints on be-future constructions and the semantics of tense and aspect, as well as diachronic issues concerning textual evidence and plausible paths of grammaticalization.

Thus the subject, though much-studied, is still ultimately unresolved. Although there is merit in many of the arguments that have been presented in the literature, none is completely satisfactory. Moreover, previous studies have not been able to take advantage

1 of recent general linguistic research on cross-linguistic historical syntax and on grammaticalization, the development of words into markers of grammatical categories like tense.

The current study has two main goals: to reevaluate the commonly known hypotheses on the development of the Slavic be-future in light of new general theories on the development of grammatical categories, and to present a new, comprehensive analysis of the evolution of the construction. The fresh perspective on language change provided by recent studies in general linguistics serves as a foundation upon which previous work on the be-future can be criticized and a new hypothesis presented. Given its general linguistic theoretical background, this work wUl be relevant not only to Slavists, but also to linguists who seek to study the development of future-tense constructions from a cross-linguistic perspective.

1.2 The be-future in Slavic

A majority of the Slavic languages have a future-tense expression that is formed with an auxiliary whose origins can be traced back to the Common Slavic verb *byti

(present stem *bqd-) ‘be(come)’.' These languages include Slovene, dialectal Serbo-

Croatian, Czech, Slovak, Sorbian, Polish, Kashubian, Belarusan, Ukrainian, and Russian.

Additionally, literary and dialectal Serbo-Croatian employ a construction, structurally identical to the be-future, to express futiirum exactum (future perfect) meaning. Only the

• For llic purposes of this study, the form^bqdq should be understood as a cover term for its reflexes in the individual Slavic languages. Where the reconstructed form is specifically referenced, the form will be explicitly identified as such. extinct West Slavic language Polabian and the East South Slavic languages Macedonian and Bulgaiian do not have any kind of be-future.

In addition to variations among the Slavic languages as to the meaning of constructions with the ‘be(come)’-type auxiliary, there is also variation with regard to the type of complement used with the auxiliary. Table 1.1 illustrates the types of complement allowed in each language.

+ infinitive + /-participle South Slavic Slovene bom pisai Serbo-Croatian budem pisati* budem pisao West Slavic Czech budii psdt’ Slovak budem pisat’ budem pisal* Sorbian budu pisac / biidu pisas Polish b^dç pisac b^d^ pisai Kashubian bçdç pisac b^d^ pisai East Slavic Belarusan budu pisac Ukrainian budu pysaty budu pysav* Russian budu pisat’

"only in nonliterary language

Table 1.1. The expression “1 (masc. sg.) will write*’ in the Slavic languages which have a become’-type future.

Although infinitival complements are more prevalent in literary languages, the use of participial complements in dialects is not insignificant; the importance of the complement has often been overlooked in the previous literature. Table 1.1 also illustrates the wide range of Slavic languages which use a

‘become'-type future. It is important also to note that the distribution of the construction is not restricted to any of the three traditional branches of Slavic; this suggests that an explanation for its presence cannot be attributed solely to an innovation specific to a single area of the Slavic linguistic territory. As such, any study of the be-future must examine the problem from a broad, pan-Slavic perspective.

1.2.1 Future-tense marking in Slavic

While a comprehensive description of Slavic tense and aspect is beyond the scope of this study, a brief description of the Slavic verb is an essential preliminary to this work.

As such, this section will present a summary of the Slavic verbal system as it is relevant to an analysis of the be-future.

As is well known, the verbal system in Slavic languages distinguishes aspect as well as tense; these categories are independent of each other. The majority of Slavic verbs can be considered part of an aspectual pair, both members of which share the same lexical meaning but different aspect. Each verb is considered to be of either imperfective or perfective aspect, depending on its semantics; this assignation is often reflected morphologically by the presence or absence of certain prefixes or suffixes.

The difference between perfective and imperfective verbs has been succinctly described by Comrie (1976: 16); “perfectivity indicates the view of a situation as a single whole, without distinction of the various separate phases that make up that situation; while the imperfective pays essential attention to the internal structure of the situation.” This means that the present tense cannot be described by perfective aspect, whose semantics is more compatible with describing punctual, completed, or otherwise total actions. In many Slavic languages, this has led to a situation where the present-tense forms of perfective verbs convey primarily future-tense marking. As such, the present tense in Slavic is often termed the "nonpast", as a reflection of the ambiguity these forms can have with regard to tense reference.-

This treatment of tense and aspect was inherited into all Slavic languages from

Common Slavic, and has continued with minimal refinement in all North Slavic languages. In the South Slavic languages (Slovene, Serbo-Croatian, Macedonian, and

Bulgarian), however, the future-tense marking of nonpast perfectives is particularly weak.

As a reflection of this, verbs of both aspects form the same periphrastic future in these languages and nonpast forms typically express present-tense meaning.^

The matter of aspect is significant for the current study because of the close interplay of aspect and tense on one hand, and of aspect and verbal semantics on the other. For example, in the North Slavic languages, use of the be-future is restricted to imperfective verbs. This is in contrast to some of the future-tense constructions used earlier in Slavic, as well as other constructions in early Slavic that have employed reflexes of *bqd- in some capacity; such a disparity must be explained. Moreover, scholars such as Rosier (1952) propose that the be-future was borrowed into Slavic; a comparison of

- For more detailed discussion of aspect in general, see also Forsyth (1970) and Chung and Timberlake ( 1985); regarding aspect in Slavic, see Thelin ( 1978) and Galton ( 1976). ^ Slovene appears to be somewhat transitional between these aspectual systems: this is discussed in detail in 2.3 l^low. the role of aspect in the putative donor language and in the Slavic languages helps to evaluate such claims.

1.2.2 The status of the imperfective future tense

The nature of the Slavic aspectual system has led to a tense system in many North

Slavic languages that appears asymmetrical. For example, imperfective verbs have both present and future-tense forms, while perfective verbs have one form that morphologically resembles a present-tense form. The only exception to this paradigm is for the verb ‘be*, whose future is formed with a simple nonpast form and whose present, at least in the East Slavic languages, is deficient. The nonpast-tense forms of Russian and their usual tense reference, given in Table 1.2, are typical of North Slavic.

“present** “future ** Imperfective pisii budu pisat' Perfective {napisii) napisu

Table 1.2. The nonpast tenses of Russian as demonstrated by the verbs pisat’ / napisat’ ‘write*.

The typical system found in South Slavic languages is quite different. The nonpast forms of perfective verbs more typically express present tense (a possibility that in North Slavic exists only in a few marked contexts), and the future tense of both aspects is expressed with the same periphrastic construction. In this way, the nonpast tenses of the two aspects appear more “symmetrical’' than the North Slavic system shown in Table

1.2 /

The apparent isolation of the North Slavic imperfective future as a periphrastic formation has led some scholars to question the very existence of a future tense in Slavic, or at least the status of the imperfective future as a necessary formal and conceptual component of the verb tense system. For example, in his influential description of the

Russian verbal system, Jakobson (1932/1984: 6) presents the system as one of binary opposition, between the past and nonpast tenses. Of the imperfective future tense,

Jakobson says only that such “‘compounded’ forms .. . stand outside the morphological verb system ” (ibid.: 4) and does not discuss it further. In keeping with this idea, both

Ferrell (1953) and Pettersson (1970) present arguments against the idea that the imperfective future is a tense at all.

From a more general perspective, the conception of the Russian typ>e of verbal system as a binary structure is well-grounded. In Ultan’s (1978: 88-9) description of the universals of the future tense, he finds that the verbal systems of many languages can be viewed as a binary tense opposition, with the present and future often falling together in opposition to the past tense (as in the Russian system above—so-called “prospective ” languages), or the past and present falling together in opposition to the future (termed

“retrospective” languages by Ultan). The question remains, however, as to the role of the

North Slavic imperfective future within such a system.

-* Tlierc is also more complexity to be found in South Slavic with regard to the expression of past tense; this is irrelevant for the current study. As was mentioned above, one possible solution to the problem is to eliminate the imperfective future from the tense system. Both Ferrell (1953) and Pettersson (1970) argue that the Russian construction is simply a combination of a perfective verb (budu) plus an infinitive, much like the phase verb construction nacat’ ‘begin* plus the infinitive.

It remains to explain the future of byt’ itself, which is not formed by a periphrase.

Pettersson (ibid.: 102) concludes that one would expect the future tense of byt’ to be budu byt\ but acknowledges that the form is ungrammatical and the matter unresolved. On the other hand, Ferrell (1953: 375) proposes that the underlying future tense of byt’ be* is budu byt ’, and resolves the ungrammaticality of the construction by proposing a surface deletion of the infinitive.

Ferrell s (1953) and Pettersson*s (1970) approach allows one to eliminate the category of the imperfective future from Russian entirely. However, there is significant evidence that the imperfective future cannot be excluded from the tense system, and that the constructionbudu byt’ cannot be posited as existing even in deep structure. For example. Grenoble (1995) demonstrates that the imperfective future differs from other verbs that combine with infinitives in that it cannot form double negatives (la—b) or be split by the negative particle ne (2a-b):

( la) Ja ne mogu ne kurit’ I cannot ngt smoke* (Grenoble ibid.: 188)

(lb) *Ja ne budu ne kurit’ I will not not smoke* (ibid.: 191)

(2a) Ty mazes’ rwplakat’l ‘Can you not cry* (ibid.: 188)? (2b) *Ja budu ne, kurit ‘I will not smoke' (ibid.: 189)

The proscription on double negation and on negation of the infinitive component of the imperfective future construction shows that it should be viewed as a monoclausal combination of auxiliary plus complement, rather than a biclausal combination of finite verb plus dependent infinitive. The monoclausal behavior of the imperfective future leads

Grenoble to conclude that the imperfective future must be considered a full-fledged tense construction, separate from other constructions of finite verb plus infinitive, on both syntactic and semantic grounds.^

What then to make of the analytic expression of the imperfective future? Grenoble

(1995: 184) concedes that the Russian tense system as presented in Table 1.2, with its inclusion of the imperfective future, is not particularly “neat” compared with Jakobson s efficient binary oppositions, but it is nevertheless more accurate as a model of the system.

The larger issue here is in fact the perceptual distinction between grammar as expressed through morphology and grammar expressed through syntax. Judging from what is found among the world's languages, verbal systems need not conform to a pattern of marking tense exclusively by morphological or syntactic means. While the category of tense is often marked synthetically, through inflectional morphemes (Bybee 1985:13), the future tense is expressed analytically in over half of the languages surveyed by Bybee and

Dahl (1989: 56). This may in part be due to the fact that the future overlaps with modal/irrealis categories in ways that the past and present tenses do not (cf. the discussion

5 For further criticism of Fetters son's views, see Thelin (1978: 57-65). in 1.3.3 below). Moreover, as Bybee argues, given that past and future tense forms tend

to evolve from very different sources, it is not unusual for the expressions to differ

formally (1985: 162). In sum, the periphrastic structure of the imperfective future is not a

particularly distinctive or unusual feature with regard to future-tense constructions in general.

Although the Russian be-future has not undergone morphologization and is not even fixed with regards to the order of auxiliary and complement, Grenoble’s demonstration of the inadmissibility of double negation on the construction strongly suggests that the combination ofbudu plus the infinitive is at least partly grammaticalized.

For other languages, evidence of grammaticalization can be found by different means.

For example, in nonstandard Lower Sorbian, the future tense of ‘be’ itself can be formed with the auxiliary budu; e.g. ja budu bys I will be’ (Mucke 1891/1965: 605).*’ This usage suggests that the auxiliary budu is merely a future marker, conveying no lexical meaning. In other languages, such as Ukrainian, scholars consider the be-future a grammaticalized construction because the finite component of the construction conveys no meaning other than future-tense marking (Rusanivs’kyj 1971: 249). Finally, in

Kashubian, the auxiliary has in fact undergone phonological reduction; alongside the full forms like b^dq_ or bqd^, one also finds md^ or bdq (Stone 1993b: 777; see also Lorentz

1925: 172).

6 Cf. section 2.7.1 below.

10 1.3 Outline of the study

Analyzing the development of the be-future in Slavic is a task of considerable complexity, and although many studies have sought to describe its path, none is entirely satisfactory. The task of the current study is to address the problems that have hindered a clear understanding of this issue, and then to present a more convincing description of the be-future’s evolution. The following subsections present a detailed description of how this study will accomplish its goals.

1.3.1 Analysis of the data (Chapter 2)

The core of this work is a collection of data from various historical periods of the ten Slavic languages which utilize some form of the be-future, some of which have been only poorly studied. It is important for the data to be as complete as possible so that an accurate chronology for the be-future can be estabUshed and the path of its development revealed. This is especially crucial given the lack in many previous works of fine-grained analysis of the data.

There are some inherent problems in gathering the data necessary for this study.

For example, the number of future-tense constructions attested in premodem texts is by nature very limited. Many of the best attested genres of early Slavic literature, such as chronicles and hagiographies, rarely have contexts where the future tense is expected.

Moreover, many of the languages lack extant texts early and numerous enough to provide sufficient data. These facts have hindered previous attempts to study this problem in

11 detail. Nevertheless, by examining data from a wide array of texts in as many of the languages as possible, a satisfactory level of analysis can be reached.

Where possible, this study draws on secondary literature; other scholars have produced careful analyses of early texts in languages such as Polish and Russian, and their work need not be repeated. Such studies can take the form of an analysis of a specific text or author, such as Cemyx's (1953) analysis of the Russian Ulozenie of 1649 or Groschel's (1972) study of the writings of the Ukrainian author Ivan Vysens’kyj (b.

1550), or surveys of texts in a particular period, such as Wandas’ (1966) examination of

Polish texts written during the reign of Kazimierz the Great (1447-92) or Stieber’s

( 1954) analysis of Polish texts of the fourteenth and fifteenth centuries.

Where such secondary sources are not available. 1 have performed my own surveys of collections of early literature. This was necessary for some of the lesser- studied languages, such as Slovak and Slovene. The earliest stages of other languages remain largely or completely unexplored due to a lack of textual evidence; this is the case for Kashubian as well as Sorbian. For these languages, the absence of data renders it nearly impossible to explore the early history of their be-future constructions.

As the earliest attested Slavic language and the literary model for many of the

Slavic languages. Old Church Slavonic (CCS) is an essential source of information. In light of the relationship between CCS and the Church Slavonic used in Rus’ and in other areas of Slavic, it is important to consider the usage of various future constructions in

CCS texts. With this in mind, H. Bimbaum’s (1958) work on DCS infinitival constructions is an important source, as are works on the verbal systems of OCS (Dostal

12 1954: Bunina 1959; Havrànek 1939) and the reconstructed protolanguage. Common

Slavic (Stang 1942).

A particularly rich source of data on this problem is premodem grammars of individual languages. Although linguistically naïve in comparison to modem grammars, these works provide valuable insight into earlier stages of many of the Slavic languages.

Some, such as the grammar of Polish by Statorius (1568/1980) and the grammar of

Slovene by Bohoric (1584/1970). were published as early as the sixteenth century. These grammars are a useful addition to the data and provide additional perspective on the issues that concem the be-future's development.

Careful attention also is paid to the state of the be-future in the modem languages, especially in dialects. There is often disparity between what is found in literary languages and in spoken dialects. With their adherence to prescriptive norms and their propensity for conservatism, standard languages often do not reflect the linguistic variation that one sees in dialects. As such, dialect data is cmcial for a complete understanding of the use and distribution of be-futures in Slavic.

Generally, data on the modem standard languages are readily available through synclironic grammars and other descriptive works, e.g. FaBke’s (1981) grammar of

Upper Sorbian or Cemysev’s (1970) analysis of analytic constmctions in modem

Russian. As for dialects, most of the work focuses on phonological and lexical description. However, dialect atlases and other surveys often report variations among future constmctions in dialects. For example, the Ukrainian dialect atlas (Zakrevska et al.

1984) contains maps illustrating the distribution of the three different future-tense

13 constructions that are attested for the language's dialects. The modem state of the be- future provides the known endpoint to the evolution of the be-future from which the historical path of the construction can be traced backwards.

1.3.2 Evaluation of previous scholarship (Chapter 3)

As is mentioned above, there is an extensive body of literature in Slavistics on the evolution of the be-future and future-tense constructions in general. This study devotes considerable space to a discussion of this literature.

Historical grammars of individual Slavic languages, such as Rospond (1971),

Dumovo (1924/1962), Lehr-Sptawihski (1957), or Bezpal'ko et al. (1957), include historical data regarding the development of future-tense constructions in those languages and can serve as a useful starting-p»oint to an analysis of the be-future. By nature, however, such works are designed to describe all facets of the language, and thus do not usually provide more than a brief description of future-tense forms.

Of primary importance to this study are those works which are specifically concerned with examining the development of the be-future in Slavic as a whole; such works include those of Bonfante (1950), Rosier (1952), Knikova (1960), and Kravar

(1978). These works focus on analyzing the data and positing a path of development for the be-future that conforms to that data.

Much of the third chapter is devoted to the evaluation of the theories presented by previous scholarship. This analysis is facilitated by the recent contributions of general linguistic works towards a more systematic description of syntactic and semantic change.

14 Through this lens, the shortcomings of previous works become significantly clearer and the validity of their reasoning more easily assessed.

Although this study concludes that previous studies of the development of the

Slavic be-future are unsatisfactory, they are not without merit. Rather, they provide an essential foundation to the current study by exploring different possible origins for the be- future and how it came to be used in so many different languages. As such, this work is not a refutation of previous literature but rather a refinement and improvement upton it.

1.3.3 Theoretical framework (Chapter 4)

This study is distinguished from previous literature in that it examines the evolution of the be-future from a broader, more general linguistic perspective. This perspective allows for a rigorous evaluation of previous scholarship, as well as a more convincing description of the be-future's path of development.

This work shall proceed from a few basic assumptions concerning time and the

future tense. A number of works provide important perspectives on time and tense in general (e.g. Jesperson 1924/1992; Fleischman 1982; Comrie 1985) and the future tense

in particular (Ultan 1978; Bybee et al. 1987, 1991).

Of particular importance to any study of future-tense constructions is what is

meant by future tense. Within the traditional model of tense, events are characterized

relative to three absolute tenses: past, present, and future.’ Despite the implication that

^ The term “absolute” here refers to those verb tenses which require only two points of reference, the moment of speech and the moment of the described event. This is in contrast to the so-called "relative" tenses, where the temporal reference of the event is regarded from the perspective of a third point in time (cf. Comrie 1985: 56).

15 these three tenses are conceptually equivalent, dividing the timeline into two areas on either side of a designated moment of speech, the timeline cannot be described as a construct where the past is a conceptually symmetrical counterpart to the future. Due to the existential fact that time progresses for us in a single direction, the description of a future event can be only a guess or prediction of what will occur, one can, after all, never be entirely certain what will occur.

The uncertainty that is inherent to the future precludes an entirely objective description of reality that takes place within its scope. This inherent uncertainty results in a linguistic expression that in many ways is as much modal as it is temporal. In other words, a speaker’s description of a future event must be colored by that person’s perspective; there can be no true statement of fact, only a prediction. The subjectivity of our perspective of the future implies that linguistic expressions involving the future tense cannot be said to describe a truly factual reality. Thus, the semantics of future-tense constructions often contains components of modality and/or irrealis.

The modal/irrealis nature of the future tense as a grammatical category and the concomitant reflection of such modality in linguistic forms is well described in the literature. With regard to the Romance languages in particular, Fleischman (1982: 133) uses this conceptual link to explain the affinity between future-tense expressions and linguistic forms which express the modalities of obligation, volition, and intention. From a broader perspective, general studies on the development of future-tense auxiliaries

16 reveal that verbs expressing such modalities often grammaticalize into such forms/

Indeed, one finds evidence of this path of grammaticaiization in the history of the Slavic languages.

Scholars who have concerned themselves with the historical development of future-tense expressions, both in general and in the Slavic languages, have approached the problem from many different theoretical perspectives. Rather than utilizing a single theoretical approach, this study strives to take advantage of the diversity of the scholarship. Proponents of grammaticaiization theory, for example, have produced valuable research. Works such as Bybee and Dahl (1989) and Reischman (1982) make important claims about the nature of future tenses that can be applied to the Slavic question. Of perhaps the most value are the efforts of scholars like Bybee to explore the potential lexical sources for future tense constructions in a broad, cross-linguistic framework.

It would be inadequate, however, to rely solely on the works of proponents of grammaticaiization theory. Harris and Campbell (1995) and Joseph (2000), for example, approach historical syntax from a different perspective, and explicitly argue that there is no need for a separate theory of grammaticaiization. Rather than choosing sides in what is often a philosophical debate, this study will make use of the best scholarship from all sides.

* In this work, the term "grammaticaiization" is used in the most general sense: see section 4.2.1 for more detailed discussion.

17 Through the analysis of random samples of languages, studies dealing with the development of future-tense constructions have drawn several general conclusions that allow for the establishment of a typology. For example, it is clear that future-tense auxiliaries typically evolve from only a small number of lexical som^ces, which share certain common semantic features that lead to grammaticaiization. The most frequently used lexical sources are verbs denoting desire, obligation, or movement (Bybee and Dahl

1989: 58; By bee et al. 1994). Another common lexical source identified by Bybee et al.

(1994) is ‘be’ or ‘become’, with languages as diverse as Classical Latin, Modem

Icelandic, and Yessan-Mayo (an Asian-Pacific language) employing futures that developed from these verbs.

Despite the discussion of become-type futures in the general literature, Slavic is rarely mentioned as a source of data. For this reason it is all the more important for this study to make the data available in as comprehensive a collection as possible. The richness of the data collected in this study, both of the modem languages and from earlier stages of their development, allow for a fine-grained analysis of the path of development for become-type verbs into future auxiliaries.

1,3.4 A new theory of the Slavic be-future (Chapter 4)

In the same way that recent general linguistic scholarship on language change can be used to reveal the flaws of previous hypotheses about the be-future, such work can also provide a solid framework for a new hypothesis. Thus, this study culminates in the presentation of a new, more convincing description of the development of the Slavic be-

18 future. In many respects, this hypothesis is very different from those which have been presented before; it has been shaped by the consideration of a wide array of types and sources of data, and omits no Slavic language from its scope.

This theory rests on the fundamental idea advanced by Wierzbicka (1978, 1988) that the semantics of lexical items dictates how those items interact on the syntactic level.

From this starting-point, one can then argue that syntactic change is motivated by semantic change. The theory presented in this study addresses the problem of the Slavic be-future s development from this perspective, analyzing the semantics of the Common

Slavic verb *bqdq and demonstrating how it underwent a shift. This semantic shift then triggered a change in the verb's government, or coUigability, as well as the verb’s grammaticaiization into a future auxiliary.

The new theory presented in this study makes a valuable contribution to both

Slavistics and general linguistics; it accounts more fully for the available data and examines the development of the be-future from a broader perspective than has been previously attempted. In this way, the development of the be-future becomes a better- understood phenomenon, and the grammaticaiization of become-type verbs into future markers is more clearly described.

19 CHAPTER 2

THE SLAVIC DATA

2.0 The Proto-Indo-European future

Before beginning an examination of the data regarding the future tense in the

Slavic languages, we must determine whether the source for the be-future can be found in

Proto-Indo-European (PIE).

Some IE languages or language groups, such as Greek, Old Latin. Baltic, and

Indo-Iranian, have evidence of a formant *-s- that can be associated with future marking.

In some of these languages, forms that appear to contain this*-s- convey meanings that can be linked closely to futurity, such as a desiderative form found in Indo-Iranian.

However, it is difficult to confirm that such forms all correspond to the same formant in

PIE, although the similarities of the forms are striking (cf. Szemerényi 1990; 307-10).

An 5-future is the standard future formation in Baltic, including both Latvian and

Lithuanian (Endzelins 1971: 231—32; Stang 1942: 202-04). This is notable, as Baltic is often thought to have formed a subgroup of PIE with Slavic. It has been claimed that evidence of this *-s- future can be found in Slavic as well, although only as an isolated, lexicalized remnant. There is a rarely-attested participial form found in some Russian

Church Slavonic texts that appears to be formed by adding the Indo-European *-s- future

20 formant onto the stem of the verbbytv. bysqst-lbys^st- (with s < *sj). Due to the presence of the *-.v- and an analogous future form in Baltic, these forms are often termed a “future participle"."^ The original meaning of the form, however, could more easily be described as resultative, or change-of-state, in keeping with the original semantics of the stembqd- and the presence of an 5-formant with aorist meaning in Slavic (see Aitzetmiiller 1968 and the discussion in section 2.1 below).

In summary, there is no evidence that the Slavic language group inherited a specific morpheme from PIE that can be associated with marking futurity, nor is there evidence that a 'be(come)-type future can be reconstructed for PIE. Thus the formation of future-tense forms and constructions in the Slavic languages must be explored within a purely Slavic context.

2.1 The etymology and meaning*bqd- of

An integral part of the Slavic be-future is, of course, the auxiliary. This section will discuss the etymology and semantics of the verb that developed into this future-tense auxiliary.

Scholars generally agree that the reconstructed Common Slavic verb with the present stem *bqd- is descended from the Late PIE stem *bhii-n-d. The nasal consonant

(which in syllable-auslaut combined with the preceding vowel to form a nasal vowel) is an infix inherited from PIE. In Slavic, this infix is found in a small class of verbs marking ingressivity or inchoativity; other members of this class are represented by the

See, for example. H. Bimbaum 1958: 15-16; D. Bimbaum 1995.

21 OCS verbs s^dq ‘sit (down)’, l^gq ‘lie (down)’, stanq ‘stand (up); become’ and prefixed

verbs based on -rqstq: sTsr^stq ‘meet’ and obr^stq ‘find’ (Stang 1942: 53—54; see also

Szemerényi 1990: 290-93). The nasal infix is restricted to the nonpast form of these

verbs; hence the infinitive counterpart to *bqd- is *byti (derived from *bhü-tei).

The presence of the nasal infix and the aspectuality of other (Slavic) verbs in the

same class suggest that the verb is perfective. If this is the case, one can interpret the

meaning as change-of-state, i.e. become’. The existence of suppletive present tense

forms like *jesmb, however, complicates the issue. Dostal (1954: 146) proposes a

solution to this problem, arguing that byti is biasp>ectual, with a stative imperfective

present *jesmb and a change-of-state perfective present *bqdq. This pairing is reflected

in past-tense forms as well, e.g. the imperfective aorist*bë versus the perfective aorist

stem *by-. A similar solution is propxased by Schooneveld (1951: 103); the Slovmk jazyka staroslovënského also indicates that there is evidence from Old Church Slavonic

(hereafter “OCS”) texts supporting the idea that a change-of-state meaning existed for byti

(SJaS 1961: s.V. “Ayn”’).

The change-of-state meaning for this verb is not restricted to the nonpast form.

Aitzetmiiller (1968: 12—13) finds this meaning in many uses of the so-called “future

*bysqst-!bysqst-, and according to Stang (1942: 53), it was present in the

perfective aorist *byxb as well. These forms do not show the nasal infix because aorists

and participles of this class were formed from the root (see also Schenker 1993:98). The

2 2 change-of-state meaning of the aorist is documented for OCS by Dostal ( 1954: 149-50); moreover, it is found in Rus'ian (Old East Slavic).

The original change-of-state meaning of the stem*bqd- is often overlooked by

Slavists. Nevertheless, the original semantics must be considered when studying the development of *bqdq into a fiiture-tense auxiliary. As is discussed in 3.3.4, knowing the lexical origins of the auxiliary allows for a clearer definition of the path of grammaticaiization.

2.2 The origins of the Slavic be future

Since Proto-Indo-European cannot be claimed as the source of the Slavic be- future. one must then move forward in time to consider the Slavic proto-language.

Common Slavic. The oldest attested Slavic language. Old Church Slavonic, closely resembles the reconstructed protolanguage, although it does show dialectal variants of some features. This section will examine the evidence from OCS and comparative evidence to determine whether the be-future can be reconstructed as an innovation that developed during the Common Slavic period.

2.2.1 Old Church Slavonic

H. Bimbaum ( 1958: 7-8) describes several analytic constructions in OCS that are used to express future tense. In this context, the nonpast forms of several verbs are

See 3.3.4 for funlrcr discussion of Ihe situation in Rus’ian.

23 attested in combination with infinitives: imamb ‘have (to)’ or is (destined) to’, xostjq

‘want’, and prefixed forms of -cbnq ‘begin’.

The nonpast perfective ofbyti be(come)’ is found in combination with predicate

nouns and adjectives, as well as various participles: the present active, the present passive,

the past passive, and the /-participle. In the earliest period of Slavic, Bimbaum (ibid.:

21-22) finds no evidence that these constructions were grammaticalized, but considers them to be constructions of copula plus predicate.

The construction withbqdq plus the /-participle is 2 l futurum exactum or future perfect that is rarely attested in OCS, occuring only seven times. It is generally considered to be part of a system of perfect tenses inherited from Common Slavic that are formed with an auxiliary based on different tense-forms of byti plus the /-participle. The

/-participle is often referred to as a resultative participle; thus the future perfect describes a situation that is or has been in effect prior to a future reference point. For a more detailed discussion of the relation between the futiu-e perfect and the be-future, see sections 3.3.2 and 4.5 below.

In OCS one also frnds examples of the perfective nonpast third person singular form bqdetb in combination with an infinitive and a dative object (the understood agent) as part of an impersonal construction (ibid.: 23—24). Infinitives can be either im­ perfective (la) or perfective (lb).

(la) i mbnë bodetbstradatipodruzija i volq svojejq ...and I will have to be deprived of marriage and my [own] freedom’ (Supr. 237, 1-2)

(lb) ...ty bo emuze nexostesi bodetb ti priieti...

24 “...[As fori you, that which you do not want [to endure] you will have to endure...' (Supr. 131. 18-19).

Although this expression resembles the be-future superficially, infinitives may be of either aspect, and the meaning is one of obligation imposed on the dative object. The lack of a nominative subject in these constructions suggests that they should be seen as a future-tense impersonal expression of obligation rather than an indicative future. ' ‘

Given the data, it appears that OCS did not have a 'be(come)-type future of any kind, except for the nonpast perfective bqdq itself. This suggests that the development of the be-future should be placed at a point after the breakup of Common Slavic, after the differentiation of the individual Slavic languages.

2.2.2 Placing the be future within the context of chronology

As was shown in Table 1.1, the be-future is found in a broad and diverse group of Slavic languages, with its distribution cutting across the traditional divisions of East,

West, and South Slavic. The wide distribution of this construction has motivated some scholars to suggest that the expression can be reconstructed for the protolanguage.

Common Slavic. This is proposed by Townsend and Janda (1996), for example, in their description of Common Slavic. The aforementioned lack of any attestations of the be- future in OCS, however, suggests that the construction cannot be easily placed in the

Common Slavic p)eriod.'“

'1 See 3.3.1 below for more discussion of this construction. '- This conclusion is supported by the situation in Rus'ian. another language of early attestation (see 2.10-2.12 below).

25 With these issues in mind, it becomes important to examine the data from the individual languages. Analysis of the earliest evidence of the be-future in these languages lends further supp>ort to the argument that the be-future cannot be considered part of

Common Slavic, and moreover that the path of development of the construction is quite varied across languages. The following sections present what is known about the development and attestation of the be-future in tlie languages that employ this type of construction, including information about the chronology of the be-future in each language. All Slavic languages except Bulgarian, Macedonian, and Polabian—the languages in which no be-futures are attested—will be discussed. The data not only prove useful for the question of whether the be-future can be reconstructed for Common

Slavic, but also are essential for evaluating existing theories and formulating new theories regarding the development of the construction.

2,3 Slovene

2.3.1 Usage

In literary Slovene the standard future is formed with a shortened reflex of *bqd-, e.g. 1 sg. bom, 2 sg. bos, plus the /-participle (Priestly 1993: 417), a construction that is identical in form to the Common Slavic future perfect (see 2.2.1 above). This is the means of forming the future tense for verbs of either aspect, although Lencek (1982: 192) indicates that the nonpast of perfective verbs may also convey future-tense reference, as is typical in North Slavic languages. Among dialects, Lencek (ibid.) reports that the Upper

Camiola dialects of central Slovenia prefer nonpast forms for perfective future, while the

26 northeast Styria dialects, which border the Kajkavian dialect region of Croatia, favor the use of the be-future construction.

2.3.2 Chronology

The be-future is absent in the earliest text of Slovene recension, the Freising fragments, which is dated around the late tenth century. Future tense is expressed most frequently in the text with nonpast perfective verbs; there is also one example of the auxiliary ‘begin’ with an imperfective infinitive (FD 1968). Other than this early piece of negative evidence, no data are available until the sixteenth century. The earliest attestation of a be-future is a catechism dated 1551, which has bod- plus the /-participle(Rosier

1952: 120). By the latter part of the sixteenth century, the form must have been fairly widespread; it is found in an early grammar by Bohoric (1584/1970), who describes the future as formed with the shortened forms of the auxiliary—e.g. bom, bos, bo {1,2,3 sg., respectively). According to Bohoric, a so-called “conjunctive” future used in conditional clauses is formed in the same way; the Latin equivalents he has supplied are future perfects. Bohoric makes no comment on the relationship of aspect to the formation of the future, and his sample verbs are all imperfective. However, his lack of distinction between the behavior of imperfective and perfective verbs suggests that aspect did not affect the formation of the future. The early nineteenth-century grammar by Kopitar

(1808/1971: 311) states more clearly that the be-future constmction was possible with verbs of either aspect.

27 In general, there appears to be no evidence for a be-future with infinitival complement in Slovene (Rosier 1952: 120—21).

2.4 Serbo-Croatian

Although there are political arguments for considering Serbian. Croatian,

Bosnian, and Montenegrin as separate languages, the main dialect divisions of the Slavic language territory of former Yugoslavia cut across political boundaries.’^ Thus, discussion of the languages in this territory has been combined into a single section under the rubric “Serbo-Croatian”, and organized by major dialect group.

2.4,1 Usage

The modem literary languages have a ‘want’-type future, while become-type futures are found only in dialects. Both of these future constmctions are used with verbs of either asp>ect. There is also a “future II” orfutiirum exactum in the literary language. used in dependent clauses (1):

( 1 ) Kad hiidemo eovorili s Marijom. sve ce biti jasno. 'When/if we speak with Marija, everything will be clear’ (Browne 1993:331).

This participial constmction is used typically with imperfective verbs; in the same context, perfective verbs typically appear in the nonpast (ibid.). Note that in the main clause of

For a discussion of ihc major dialect groupings of Serbian and Croatian, see Browne 1993: 382-ff.

28 this example, the future tense is expressed with the literary standard ce biti, with the

want-auxiliary.

In Stokavian dialects, which comprise the eastern and central area of former

Yugoslavia, including Serbia, Bosnia, and much of Croatia, a monolectic construction with infinitive plus the be-auxiliary is found as a futuriim exactum, e.g. imadbudem ‘I will have had'. According to Belie (1965: 81,152), this form is found only rarely, but he gives no sp>ecific information regarding its distribution. In addition, according to Rosier

(1952: 109), budem plus the infinitive is found in southern areas of Stokavian as remnants found in folk songs and tales; he does not specify whether these expressions convey pure future orfuturum exactum . In her description of the dialect in the region of

Imotska Krajina and Bekija, Simundic (1971: 198) reports constructions with budem in combination with either the infinitive or the /-participle being used as a futurum exactum, but she gives no details regarding possible contextualizations for the usage. She also reports that either aspect is possible for both types of complement.

Rosier (1952: 119) also reports that in Cakavian dialects, spoken in Western

Dalmatia and Istria, budem plus the /-participle is commonly found as a future construction; he makes no mention of aspect with regard to the complement. In general.

Rosier argues that budem plus the infinitive is more widespread in main clauses in

Cakavian than in other dialect groups. In response to this, Kravar (1978: 261) argues that budem plus the infinitive of either aspect is actually widespread in modem dialects, including not only Dalmatia and the coastal areas of Croatia, but also in islands in

29 (Stokavian) Bosnia and Slavonia. In usage it is typically less an expression of pure futurity than a relative future used in dependent clauses (2):

(2) Dat cu ti kad budem imati. 1 will give [it] to you when 1 have [it]' (Kravar 1978: 262).

As in the example from the literary language found above, future-tense reference in the main clause here is indicated with the standard literary 'want-auxiliary.

Studies of specific Cakavian dialect areas reveal some evidence of the main-clause attestations ofbudem plus the infinitive that Rosier mentions. For example, in the Istrian dialect spoken around Orbanici, Kalsbeek (1998: 298) reports that although the construction is used primarily in dependent clauses, main-clause examples do exist. In her corpus, all examples are negated, have non-personal 3 sg. subjects, and are formed with the same verb, cut ‘hear, be heard' (ibid.: 284). For example, the following statement (3) was surrounded by a discussion of the quantity of sound material her tape was thought to contain.

(3) Nic ne buode tako cut. Svi so mucaii. There won’t be anything to hear this way. Everybody was silent' (ibid.).

The modem Kajkavian dialects, located in Northwest Croatia, do not use budu/ budem plus the infinitive at all; rather, constructions with budemlbodem (or bumlbom) plus /-participles of either aspect are used as futures or futuraexacta, in alternation with a

‘want’-type future (Rosier 1952: 118; see also Magner 1966: 40). This mix of features is not surprising in an area that is transitional between Croatian and Slovene.

30 2.4.2 Chronology

The earliest Serbo-Croatian texts, dated from the twelfth to the thirteenth centuries, are heavily influenced by OCS and show no evidence of a be-future (Kravar 1978: 260).

Rather, future tense is expressed either with nonpast forms of perfective verbs, or constructions with ‘want* or ‘have’ in combination with the infinitive (Belie 1965:8 i-82;

Malic 1972: 175-76).

The only date mentioned by Belie (1965: 81) is that from very early on, i.e. the thirteenth century, budem plus the /-participle could replace the infinitival construction in the meaning of the futurum exactum. This statement is confusing in light of what we know about the history of these forms; moreover. Belie gives no textual evidence of budem plus the infinitive existing at all in early texts and thus gives no justification for his statement. The issue remains unresolved without a more careful analysis of early Serbo-

Croatian texts, and this analysis has yet to be done.

A search of available sources does reveal a small amount of data: two examples of budem plus imperfective infinitives (1) can be found in an Old Croatian text, Pisan svetogo Jurja, poj ljudem razumno (Malic 1972). It is a religious poem about St.

George’s encounter with the dragon, and the text is dated circa 1380.

( 1 ) Krai's vlasteli pace tako vecati: “Ucinemo drakunu po d’ voe dobit(‘)ka dati, da budet’ dobitak' pri ezere stati. Pozrv ga, drakun’ budet’ se vracati." ‘The king began to speak to his subjects: “Let us give the dragon two sacrifices each, so that the sacrifice will be standing by the

31 lake. Having eaten it, the dragon will return [to the lake]”’ (ibid.: 34).M

Although both constructions withbudem plus the infinitive are in a dependent clause in

(I), they appear to be pure futures and not future perfects. In light of this evidence.

Belie s claim that infinitival constructions predated participial constructions in early

Serbo-Croatian appears even less credible.

In Ragusan, or Old Dubrovnik, a mixed Cakavian-Stokavian dialect whose texts are from the fifteenth—sixteenth centuries, examples of budem plus the infinitive are found. They occur both in main and dependent clauses, sometimes with perfective verbs

(Rosier 1952: 112). Examples ofbudem plus the /-participle are also found. In the works of the Ragusan author Dominko Zlataric (late sixteenth century). Vaillant (1979:

44—45) finds cases where budem plus the infinitive is replaced by the 'want-future in later revisions of the same text. Constructions withbudem can be formed with complements of either aspect, although Vaillant (ibid.) also finds nonpast perfectives used to express future meaning.

In eighteenth-century Dalmatian texts, budem plus the infinitive is found alongside the 'want' future, the former being attested especially in dependent clauses

(Rosier 1952: 113). The concentration of these constructions in dependent clauses suggests that they were used as a relative future or futurum exactum. A similar usage is found in nineteenth-century texts from Kosovo-Metohija and a text from Prizren dated to

Although it might appear that the verb stati in this example is the perfective verb meaning ‘stand up’, context suggests that it is in fact a contracted form of imperfective stajati be standing’. Such contraction is attested for this verb in Croatian (Ivekovic and Broz 1901: s.v."stâtr).

32 1871. with infinitival or participial complements attested in combination with budem to form afuturum exactum (Vukicevic 1978: 87-88,150).

For Cakavian, Kravar (1978: 259-60) reports that budem plus the infinitive is well-attested from the second half of the fourteenth century. This suggests that it appeared even earlier in the spoken language, but was not reflected in the conservative,

Slavonic-influenced texts of the earlier period. Texts from the fifteenth and sixteenth centuries attest either (/io)cM ‘want’ orbudu plus the infinitive as a pure future. Resetar’s

(1898: 193—94) fifteenth-century texts attest both nonpast perfective verbs and imperfective infinitives combining with budu to express futurity; he notes that later copies often show both types of construction replaced with ‘want-type futures. This situation is described by Popovic (I960: 517) as well, who reports that the future in premodem

Cakavian dialects was expressed by either nonpast perfectives or budu plus imperfective infinitives. For the same period. Rosier (1952: 114—15) finds the occasional use ofbudu plus the /-participle, which he says is more frequently attested than the infinitival complement. According to Rosier, both typies of complement can occur with verbs of either aspect (ibid.); this is contradictory to what is reported by Resetar and Popovic.

In support of what is attested, a grammar written by Cassius (1604/1977) at the turn of the seventeenth century describes only a want-type indicative future (ibid.: 99), but he also describes a “conjunctive future” that is formed with budu plus the /-participle

(ibid.: 95). Judging from the terminology employed by Cassius, this latter form is most likely being described as ti futurum exactum; examples are given with the Latin future perfect as equivalent (ibid.: 108).

33 Based on his own data, Kravar (1978: 257—58) concludes that budem plus the infinitive should not be considered a Cakavian-Ragusan dialect feature but should be seen as a construction that is found in dialects across a large part of the Serbo-Croatian linguistic territory. This appears to be supported by the other data presented in this section.

Kajkavian seems to have budem (orbum) plus the /-participle as a possible future construction since at least the sixteenth and seventeenth centuries, alongside a ‘want’-type construction (Aleksic 1937: 49). Rosier notes that this expression could be formed with verbs of either aspect (1952: 118); this contrasts with Popovic (1960: 365), who describes the dialect as having an aspectual opposition in the future tense between nonpast perfective verbs and bomlbum plus imperfective /-participles.

As was the case for Slovene (see 2.3.2 above), there appears to be no evidence that a be-future with infinitival complement ever existed in Kajkavian.

2,5 Czech

2.5,1 Usage

A be-future with infinitival complement is used in the literary language for imperfective verbs only (Short 1993a: 481). Most dialects reflect the same usage as the literary language, although in the northeastern areas bordering Poland, use of an /- participle complement is also possible (Belie 1972: 198). For example, Kellner (1946:

174) reports that in the Eastern Lach dialects, masculine singular subjects typically form the future with participial complements, while other persons tend to use infinitives.

34 2.5.2 Chronology

The earliest examples of the be-future in written Czech are found in a text dated circa 1300, which is the period of the earliest Czech texts (Rajsans 1903: 118—20; see also Gebauer 1909: 425). They are among some marginalia found within theMnichovské zlomky. a codex in Latin (also called Tiilecsv. Bonaventury):

( 1 ) Nemochi budefbiti wchifle mezi dywcami ydcherami xpi ‘You will not be able to be in the number among the young women and daughters of Christ’ (Rajsans 1903: 118, In. 24b)

This example is one of the earliest be-futures to be attested in Slavic as a whole.

Five be-futures are the only imperfective future constructions found in another text from circa 1300, the Zaltdr Musejnf (Rajsans 1903: 129-42). For example, one finds the following (2), from Psalm 145:

(2) chvâlitihudu hofpodina ... ‘I will praise the Lord.. . ’ (Rajsans 1903: 130)

Be-futures are also found in the Hradecky manuscript from the mid-fourteenth century, the Legenda o sv. Katefiné of the later fourteenth century, as well as other texts dated to the fourteenth centiuy (Krîzkovâ 1960: 93). Based on these examples and given the later attestation of be-futures in South Slavic and East Slavic, Czech appears to have developed a be-future within an extremely early time-frame for Slavic.

Alongside these be-futures are somewhat rarer uses of ‘want’ or have’ plus infinitives (Gebauer 1909:425—26). Knzkova (1960: 94) argues that these constmctions had a primarily modal, not temporal usage. However, given the fact that such

35 constructions were used as future expressions in other early Slavic texts, one cannot assume that constructions with 'want or have' did not convey future-tense meaning, nor can one rule out the jx>ssibility that the be-fuiure was in alternation with these expressions as future-tense constructions in the earliest Czech texts. Gebauer (ibid.) considers some of these constructions to have conveyed future-tense meaning; the evidence appears to support his view. For example, consider this example from the mid-fourteenth-century

Hradecky manuscript (3), from O sevci'ch, a satirical piece on tailors. The example is quoting a wife who is fed up with her husband’s behavior:

(3) Czo mi uczynyti ftobii? Racz JlyJJieti, moy miiy muzy, wjak widif mu i Jvi’U niizy! Ze chczes wzdy w krrczmye lezieti! Dietky chtie hiadem zemrzieti. “'What am I to do with you? So you listen, my dear husband, but you see my and your needs! You always want to hasten to taverns! The children will die of hunger'” (Porak 1979: 89).

While one can argue that the first use of want' in this example conveys primarily desiderative meaning, the second use clearly has no such nuance.

Raw data from texts after the fourteenth century are not provided in the secondary literature. Lehr-Splawihski ( 1957: 141), for example, merely mentions that the be-future became more widespread after the fourteenth century; this does not illuminate whether other future constructions became rarer or whether be-futures became better attested.

Without more data, one cannot determine the exact path of development for the be-future in Czech.

36 In sum, as Kfîzkovà (I960: 93-94) points out. the early data indicate that Czech has textual attestations of the be-future circa 1300, before any other Slavic language.

Moreover, the time-frame for development of the be-future extends prior to this point if one considers that the appearance of the be-future in the written language most likely reflects an earlier development in the spoken language. The rarity of attestation for the future constructions inherited from Common Slavic also suggests that the be-future developed and spread quite early in Czech.

2.6 Slovak

2.6.1 Usage

A be-future is standard for imperfective verbs in the literary language. Dialects in extreme Eastern Slovakia, which in other features show strong East Slavic influence, also attest futures with ‘want* or ‘have*. There is also a future formed with go' that has been termed an “immediate” future, found especially in Central and Eastern Slovak (Stanislav

1958: 418). Also in the East, the be-future with an /-participle complement is found

(Stanislav ibid.; see also Stoic 1994: 109), These areas are adjacent to regions in Poland and Ukraine where the /-participle complement is also used (see 2,8.1 and 2.11.1 below, respectively). For example, the Saris area in northeast Slovakia (which borders Poland to the north) has both the infinitival and the participial be-futures (Buffa 1995:158).

37 2.6.2 Chronology

In the case of Slovak there are scarcely any historical data available. Rosier

(1952). for example, devotes a section to the be-future in Czech and Slovak but makes no specific mention of Slovak. Knzkova (1960), on the other hand, has compiled data relevant to both Czech and Slovak but labels the data “Czech”, Neither approach is entirely satisfactory, although they reflect the close affinity of these languages and the predominance of Czech as the literary language in the earliest period.*^

An additional hindrance to collecting data is the fact that there are no extant Slovak texts prior to the end of the fourteenth century. There are, however, earlier Old Czech texts that contain Slovak linguistic features; these have been identified by Stanislav (1957:

11).

An examination of the early texts found in Stanislav (ibid.) reveals relatively frequent use of be-futures. One example (I) is found in the earliest text, Dialogi

Bohemarii, dated 1379:

( 1 ) yakoz czlowyek syege, takez bude z\eczi As a man reaps, so he shall sow’ (Stanislav 1957: 124-26)

According to Stanislav (ibid.: 126), this text contains features of Moravian Slovak, a

West Slovak dialect, and can be considered one of the earliest texts with Slovak features.

Be-futures with infinitival complements continue to appear relatively frequently throughout texts of the fifteenth century, with attestations increasing towards the turn of

A more detailed description of the relationship between Czech and Slovak in the earliest period can be found in D’urovic (1980).

38 the sixteenth century. Constructions with ‘have’ are also found in many texts, although primarily in combination with perfective verbs. By the seventeenth century, the be-future appears to be quite widespread, and other future constructions are not attested. In the eighteenth century, Bemolak’s (1790/1964) grammar, the first Slovak grammar and one based on West Slovak dialect, confirms the use of the be-future and gives no alternative constructions (ibid.: 207). This is also the case in the grammar by Stur (1846/1943), which favors the Central Slovak dialect and provides the model for the modem Slovak literary language.

In contrast to the data found for the infinitival be-future, there is no historical evidence for the participial type. This may be due to the lack of historical sources for East

Slovak dialects, which are not represented by the early grammars of Slovak.

2.7 Serbian

2.7.1 Usage

A be-future with infinitive complement is standard for imperfective verbs in the literary language. In some dialects of both Upper and Lower Sorbian, the be-future can also be formed with perfective verbs (FaBke 1981: 253; Mucke 1891/1965: 604). The spread of the be-future into the realm of perfective verbs is a situation which has been gradually making inroads into the literary language (FaBke 1981: 253).

In Lower Sorbian, one can even find the future tense of be’ expressed analytically, e.g., ya budii bys I will be’ (Mucke 1891/1965: 605). In criticizing this usage, Mucke (ibid.) erroneously translates this example as the nonsensical ich werden

39 sein zu sein "I will be to be’; it is more likely that this is in fact evidence of the further grammaiicalization of biidu into a tense-marker that no longer conveys any lexical meaning.

2.7.2 Chronology

According to Knzkova (1960: 103), the earliest attestations of a be-future are from the sixteenth century. Examples can be found in texts as early as Miklawus

Jakubica’s 1548 translation of the New Testament into Lower Sorbian. For example, this excerpt from John 16:13(1):

(1) ... Pfchetoafch won wot foebe fame go nebucze molwitz. krome to zcfoch flifchatcz biicze. to won molwitczbucze...... For he will not speak of himself, but whatever he will hear he will speak’ (Schuster-Sewc 1967: 174).

The earliest attested texts also date from the sixteenth century, so there is no way of determining exactly when the be-future came into use, or if other future forms were ever in use.

The use of perfective verbs with the be-future is clearly an unusual development for North Slavic. One can speculate that this relaxing of the asp>ectual constraint on the construction is due to the almost universal bilingualism in the area between Sorbian and

German.'^ German has no aspectual constraint on its own future construction with werden.

Cf. Stone (1993a); sec also 3.2.4 and 4.3.2 below for more discussion of perfective complements in Sorbian.

40 2.8 Polish

2.8,1 Usage

A be-future is standard for imperfective verbs in the literary language, and an infinitival complement is in alternation with plus the /-participle. Much literature has been produced exploring the distribution of these forms in modem Polish, as well as the motivating factors behind the choice of form.

The choice between the participial and infinitival complements is typically seen by scholars as a conflict between economy and expressiveness. The participle conveys gender and number—by concord, although if the pronoun is omitted then the participle is the sole form marking these categories—whereas the infinitive is one syllable shorter than all participial forms besides the masculine singular and thus more “economical” in its expression.

Nitsch (1956: 195) argues that the forces of economy versus expressiveness explain the distribution found in modern-day dialects. He reports that generally there seems to be complementary distribution, so that masculine singular forms have the participial complement, while other forms have the infinitival. His use of the term

“economy” refers to the fact that the use of participial forms with nonmasculine subjects would add another syllable to the expression. However, the use of altemating complements to express the same meaning can hardly be called economical with regard to the larger system, even if the utterance itself is shorter by one syllable. Moreover, the relative frequency of the participial construction with masculine singular subjects can only be viewed as a correlation rather than a cause-and-effect relationship.

41 Nitsch‘s use of the term “expressiveness” is also suspect; ostensibly, he is arguing that the gender marking of the participle, absent on the infinitive, adds to the expressiveness of the participial construction. However, this gender marking merely reflects the gender of the subject, and no categories are expressed by the participial form that are lacking in the infinitival construction.

Regarding the tendency for nonmasculine subjects to appear with infinitive complements in be-futiure constructions, Mikos (1985: 454) has found results that support this view in his study of speakers of the standard language. According to his data, the most important factor that triggers use of the infinitive is a plural subject, followed by non-masculine gender and nonpersonal subjects. Again, one cannot assume a cause-and-effect relationship here, merely the presence of a correlation between gender and type of complement.

Both Nitsch and Mikos appear to support the idea that the choice of form is motivated by stylistic considerations. This view implies that there is no semantic difference between the two types of complement, a position that is criticized by F*roeme

(1991). Proeme argues that the main motivating factor is semantic. According to his theory, the construction with the infinitival complement conveys a nuance of expectation that is absent from the participial future; his examples, however, do not convincingly support this assertion.

Proeme rightly excludes the modal verbs chiec ‘want*, moc able (to)*, miisiec

‘obliged (to)*, and miec have (to)* from his argument, because these verbs always appear in participial form in future constructions (ibid.: 195). These verbs all can take infinitival

42 complements of their own, so one presumes that there is a syntactic constraint present in the language that discourages constructions with two infinitives. Historical evidence supporting the presence of this constraint is discussed below.

The scope of this study does not allow for a detailed evaluation of Proeme's theory, although the data presented by Mikos (1985) call into question his conclusion that choice of complement is motivated by semantic factors. Based on the historical data discussed below, however, it appears that the differing usage of the two forms might be linked to the different age of the constructions.

2.8,2 Chronology

The earliest extant texts in Polish date from the fourteenth century. According to some scholars, both forms of the be-future were used “from the earliest period”

(Klemensiewicz et al. 1964: 374; see also Rospond 1971: 308). Until midway through the fifteenth century, however, constructions with the infinitival complement predominate

(Cyran 1961: 223), while forms with the (imperfective) participial complement are very rare in the earliest texts (Gorecka and Smiech 1972: 13). Unfortunately, there appears to be no data regarding the existence of constructions with a perfective /-participle in any

Polish texts, which would lend insight into the fate of the future perfect construction inherited from Common Slavic.

The infinitival complement was preferred overwhelmingly until the late fifteenth century. In terms of dialect distribution, Mazovian, which comprises the northeast area of

43 Poland, shows the most frequent use of the participle at this earliest stage (Gorecka and

Smiech 1972: 14).

In Table 2.1 below, data taken from Slieber ( 1954: 231-32) show the distribution of the types of be-future in fourteenth- and fifteenth-century texts, while data from

Kowalska (1976: 127) show percentage distributions of the possible complements.

Although Stieber and Kowalska both separate the data for the infinitival be-future according to the order of the construction s components, variation in word order is more relevant to stylistic studies and will be ignored in the current analysis.

/-participle infinitive Kazania swû^tokrzyskie 2 Kazania gnieznienskie 3 1 Psaiterz florianski 1 443 Psaiterz pidawski 11 443 BibliaZofii 9 579 Kodeks Swi^tosiawa 31 75 total {%) for singular 6 94 total (%) for plural 1 99

Table 2.1. Polish future complements through the fifteenth century.

Based on this data, Stieber concludes that secular texts like the Kodeks Swiqtosiawa were more likely to use the participial complement (ibid.: 233), but regardless of register, the participial complement was clearly used much less often than the infinitive.

44 It is possible that at least some of the early participial constructions can be identified as future perfects. For example, the one attested participial form in the Psaiterz florianski is as follows (I):

( 1 ) Gonic b 0 d 0 nieprzyjaciele moje, ipoiapi0 je: ani sie obroc 0 , alii zgin 0 . Ztami0 je, ni b&d0 moeii stac: padn 0 pod nogi moje. ‘I will pursue my enemies and overtake them; 1 will not turn back until they are consumed. I will break them so that they will not be able to stand: They will fall beneath my feet’ (Ps. 17,41^2).

It is not clear from context whether the Polishbpdp mogli stac should be interpreted as a future or a future perfect here. However, the equivalent passage from the Vulgate Bible

(2), which is the likely source for the Polish translation, has a future perfect.

(2) Persequar inimicos meos, et comprehendam iilos: Et non convertar donee deficiant. Confringam ilios nee po te runt stare: Cadent siibtus pedes meos (BS 1965: Ps. 17, 38-39).

The form poterunt is truncated from potuerunt, an unequivocal future perfect. One should also note that in this context, the English translation is not an accurate indicator of the tense in Latin or, indeed, in Polish; in the Vulgate, there are many cases where a Latin future perfect is used that is best translated with an English present (Plater and White

1926: 106). Clearly, there is a great deal of ambiguity surrounding this example, and it is possible that Old Polish constructions withbqdq_ plus the /-participle may indicate future-

45 perfect tense rather than pure future.’’ A more detailed study of the earliest examples is needed.

According to Kowalska (1976: 127), the participial complement continued to be found more in secular texts than church writings through the sixteenth century; Kowalska argues that the more influence of the spoken language, the more likely the use of the participle (ibid.: 128). Whereas earlier use of the participial future occurred primarily with masculine singular subjects, in the sixteenth century the participial complement began to appear more frequently with other genders in the singular (ibid.: 129; see also Gorecka and Smiech 1972: 29).

We can see evidence of these trends in the grammar of Statorius (1568/1980). At various points in his work, he describes the future as either with the participle, e.g. pisaf (ibid.: 108), or with the infinitive (ibid.: 136). In an example of conjugation (using robic ‘make, do*), he uses only the /-participle for masculine and plural forms, but gives both the participle and the infinitive as variants for the feminine and neuter singular (ibid.:

138).

Table 2.2 shows data from sixteenth-century texts (Stieber 1954: 232-33), as well as percentage totals for the various constructions from Kowalska (1976: 127). When compared with the data from Table 2.1, these data clearly reflect an increase in the use of the participial complement, especially in the singular.

See section 3.3.2.1 for more discussion of this example.

46 l-participle infinitive Kodeks Diiaiynski 7 10 Modlitwy Waciawa 1 66 Kodeks Stradomskiego 19 25 Kodeks dzikowski 36 35 Historia Aleksandra 18 25 Wielkiego Kazania Paterka 1 14 Modiitewnikdla kobiet 6 1 Zywot iw. Eufraksji 8 2 total (%) for singular 23 77 total (%) for plural 1 99

Table 2.2. Polish future complements through the sixteenth century.

In the seventeenth century, use of the participial complement was increasing, especially in the Eastern dialects of Mazovia and Little Polish (Gorecka and Smiech 1972:

27). It is for this period that scholars begin to mention that the participle was favored in contexts where the entire construction took an infinitival complement (Kowalska 1976:

132; see also Nitsch 1956: 196), a syntactic constraint on the construction (see 2.8.1 above).

This situation continued through the eighteenth century (Gorecka and Smiech

1972: 18—19). By the end of the Middle Polish period, use of the participle was beginning to increase significantly; the earlier stylistic distribution of the two complements, whereby the infinitive was preferred in higher-styie religious texts and the participle was preferred in secular texts, appears to be no longer in effect.

47 There appears to be a further increase in the frequency of the participle and a decrease in the infinitive in the nineteenth century (Bajerowa 1992: 182), a process that

had already begun in the eighteenth century (ibid.: 185). Especially notable is the advance of the participial form into the plural. The modal verb miec ‘have (to)* continued to appear in the participial form as a complement (Bajerowa 1992: 183), in keeping with the

syntactic constraint on double infinitives mentioned above. Gender also continued to

influence the choice of form: the masculine singular had the greatest increase in use of the participle, while the plural consistently favored the infinitive (ibid.: 184).

Grammars of the nineteenth century generally mention both types of future construction: e.g. bçdçrobii and robic b^d^ ‘I will work*. The construction with the participle preceding the auxiliary is referred to dismissively as a dialect form, and robic is often referred to as a Germanism (and thus frowned upon by grammarians throughout the nineteenth century (Proeme 1991: 182)). An 1897 grammar by Malecki mentions both participial and infinitival types as possible (ibid.: 185).

One reason for the shift from the infinitival to the participial complement is suggested by Nitsch (1956: 194). He argues that it was due to analogy to the concomitant development of the past tense into a synthetic form of the /-participle with suffixed personal endings formed from the present tense of ‘be*.** However, there are problems with this argument. For example, there is little evidence that the participial future expression with the /-participle preceding b^d^ was a viable form. Without such a form in use, then, there is no direct formal correspondence to be found between the past and

I* See also ihe discussion in Andersen (1987: 26—27).

48 future tense formations. Moreover, functions as a clitic in the past-tense construction, but not in the future tense; thus the two constructions do not appear to be analogous. Finally, given the inherent semantic differences between the past and future tenses in languages in general, there is no reason to assume that past- and future-tense forms would resemble each other or develop in tandem.'^

Cyran (1961: 223) offers another explanation for the increasing use of the participial complement, arguing that the greater “expressiveness” of the participial form, with its marking of gender, made it a more desirable form. He adds, moreover, that there might have been a perception among speakers that the infinitival complement too closely resembled Russian (ibid.: 224). However, it cannot be argued that gender marking on the participle adds to the expressiveness of the construction, since the participle merely agrees with the gender of the subject. As for Cyran"s claim regarding the perception of speakers, this is an entirely speculative conclusion.

The trend of the spread of the participle and the retreat of the infinitive continues into the twentieth century (Cyran 1961). Again, masculine singular forms are most likely to use the participle (ibid.: 223).

In summary, the Polish data reveal many interesting developments, although gaps in the data leave the situation in the earliest period unknown. Without earlier data, one cannot gain a complete picture of the relationship between the infinitival and participial

For discussion of this point from a synchronic perspective, sec Bybce ( 1985: 162). For a diachronic perspective, sec Pappas (1999).

49 constructions. The exact nature of the relationship between these forms in the modem

language also remains a contested issue.

Textual evidence does indicate that the participial constmction was attested only

rarely in the earhest extant texts, then spread in use and now appears to be the favored construction. It also appears that the spread of the participial future has taken place hierarchically, from the least-marked masculine singular into more marked forms such as plurals and feminine and neuter singulars. This pattern of diffusion might suggest that the participial construction to be innovative; for more detailed analysis of this point, see 3.3.2 below.

2,9 Kashubian

Some comments should be made regarding the status of Kashubian with regard to

Polish. According to Topolihska (1974: 17-ff.), Kashubian is most properly considered a dialect of Polish, but one whose geographic location led to its developing many unique features that appear in writings of the area. However, spoken Kashubian is considered difficult, if not impossible, to understand by many Poles (Stone 1993b: 759-60). With these factors in mind, this study will consider the Kashubian data separate from Polish.

2.9,1 Usage

Generally speaking, Kashubian resembles Polish in its expression of the future tense. The be-future in Kashubian can be formed with either an infinitival or participial

50 complement. The infinitival complement is more widespread in the south, in areas closer to Polish (Lorentz 1925: 172).

In his grammar, Lorentz ( 1919/1971:44) reports that the be-future is formed with

imperfective complements only. Rosier ( 1952: 127) presents one example, removed from context, of a pterfective complement,pozqdac bqdzes ‘you (sg.) will desire’. However,

Kfizkova (1960; 107) questions Rosier’s example and suggests that the verb in question is actually biaspectual. Indeed, in modem Polish, the verb pozqdac is considered imperfective (SPF 1994: 561). A dictionary of Kashubian contains only a nonprefixed form of the verb, considered to be equivalent of the Polish pozqdac (Sychta 1973: s.v.

"zqdac"). There is no entry for pozqdac or pozqdac, this opens up the possibility that

Rosier s example is more Polish than Kashubian. Without more definitive examples, there is no reason to assume that Kashubian allows the be-future to combine with perfectivecomplements.

Regarding the frequency of use of the respective complements, Gorecka and

Smiech ( 1972: 23—26) examine samples of both written and spoken Kashubian. The data from their study are reproduced in Tables 2.3 and 2.4.

51 /-participle infinitive masc. sg. 87 13 fern. sg. 33.3 66.7 neut. sg. 41.7 58.2 masc. pers. pi. 22.8 79.2 other pi. 62.5 37.5 Totals 56.7 43.3

Table 2.3. Kashubian future complements in the twentieth century: written data. (Numbers are percentages)

/-participle infinitive masc. sg. 96.3 3.7 fern. sg. 88.6 11.4 neut. sg. l(X) masc. pers. pi. 94 6 other pi. not enough data Totals 94.4 5.6

Table 2.4. Kashubian future complements in the twentieth century: spoken data. (Numbers are percentages)

The data from these tables demonstrate the striking dominance of the participial complement in spoken Kashubian. It is even more widespread than in spoken Polish, where data from Gorecka and Smiech (1972) indicate that the participial complement is used in an average of 52.4% of cases overall.

52 2.9.2 Chronology

Written works showing uniquely Kashubian features can be found from the sixteenth century, but only in the nineteenth century is there a literature that can be considered “undiluted” Kashubian (Stone 1993b: 761). For this reason, there is little opportunity for a historical analysis of the be-future. Within the Polish dialect data collected by Gorecka and Smiech ( 1972) there is Kashubian data (given above), but it is restricted to the twentieth century and thus insufficient for a diachronic analysis.

2.10 Beiarusan

2.10.1 Usage

A be-future is standard for imperfective verbs in the literary language.

Southwestern dialects also have a synthetic future with the suffix -mu ‘have/take’-® added onto the infinitive (DABM 1963: map 166; Jankouski 1983: 200). This form is also found in neighboring Ukrainian (see section 2.11 below); southwest Beiarusan dialects show many affinities with Ukrainian, including shared phonological and morphological innovations (Mayo 1993: 943^44).

20 There is some uncertainly regarding the semantics of this form. At a very early period, the verbs imantb 'have (to)’ and itnu 'take’ were confused in texts (Kuznecov 1959: 236; Kiparsky 1967: 234), most likely due to the fact that they are formed from the same stem (cf. Townsend and Janda 1996: 215-16). It is unclear which verb developed into the synthetic future found in Beiarusan and Ukrainian. For the purposes of this study, the question is not particularly relevant, and these forms shall be described as 'have/take’-futures for the sake of expediency.

53 2.10.2 Chronology

According to Jankouski (1983: 200), the fourteenth—sixteenth centuries show widespread use of maju plus the infinitive, a “have/take'-type future similar to that found in Old Church Slavonic as well as Ruslan Church Slavonic. A be-future is attested as early as the fourteenth century, although only sporadically in the earliest texts.

The earliest known attestation of a be-future identified by scholars is from a 1375 donation charter by Aleksandr Korijatovic, a Lithuanian prince, who is donating a mill and the surrounding property to the Smotrickij monastery ( I ). In the charter, Korijatovic states that despite the change in ownership for the property, the townspeople will not be exempt from providing defensive support to the donated land:

(1) Ale to sto z kali vsi bojare i zemljane biidut’ gored" tverditi. togdy tii Ijiidi tako ze iméiiit’ tx'erditi eorod" smotric'. ‘But if all boyars and inhabitants will fortifv the city, then these people also will fhave toi fortifv the city Smotric* (Pescak 1974: 49-53, no. 24).

It is interesting to note the contrast in this example between the be-future and the expression with have/take* in the following clause; the latter clearly conveys a modal nuance of obligation. Later on in the document, a have/take* expression is found in a similar construction to the one with the be-future above (2):

(2) ize to sto z koli vsi zemljane imut’ davati dan’ ii tatary, to serebro imëjut’ tako ze tii ljudi dati. ‘And if all inhabitants will give tribute to the Tatars, those people will also [have to] give that silver* (ibid.).

54 Note that imut' (etymologically from ‘lake’) in the first clause of this example is more fully grammaticaiized than iméjiit' (from have ) in the second clause; the latter conveys a stronger nuance of obligation. These early examples show that with these verbs, there was a fine line between pure future-tense reference and a modal sense of obligation.

A be-future is also found in a gramota from 1388, an oath of fealty written by

Dmitrij Koribut, a prince of Novhorod-Sivers’kyj, to King Wladislaw of Poland-

Lithuania (3):

(3) ...obécujemy nasa prisjaga i nasa vëra i c(es)t’ju ize o(t) ty(x) mest’ s nasimi dëtmi cistaja vërnost’. i polna budem” derzait) tomu istomu korolevi i jego korolici i ix” detem” i korunë poi’skoi a nikoli ni v odno veremja ne o(t)stavati a ni o(t)stupiti i na vëki i k” ix” dobromu radiit) a lixogo vërnë ostereeati. ‘...we promise our oath and our faith and honor that from this time forward, with our children, we will hold pure and full faith to that same king and his queen and their children and the Polish crown, and at no time fwill wel leave it or violate it for all time, and fwe will! look after their good and faithfully defend [them] against ill’ (Pescak 1974: 84, no. 44).

In this example, the auxiliary has several infinitival complements (underlined above) that follow in a long passage of text.

These early texts are also considered by scholars of Ukrainian to be indicative of the earliest Ukrainian examples of the be-future, although Shevelov (1980: 146-^7) considers the language of these texts to be primarily Beiarusan. At any rate, the labeling of these early texts as “Beiarusan”, “Russian”, or “Ukrainian” is anachronistic; the time period was early enough that differentiation of the East Slavic languages was minimal. It is with this in mind that some scholars refer to the language of this early period simply as

55 Rusian (or Rus'ian), a term that contains no anachronistic overtones of nationality.-' As for distinguishing between Beiarusan and Ukrainian, the problem is more difficult; the

Grand Duchy of Lithuania spanned both of these linguistic territories as well as parts of

Poland, and prior to that period there is other evidence linking southern Beiarusan with northern Ukrainian. Thus, until approximately the sixteenth century, it is perhaps best to employ a term such as Western Rus’ian when describing these linguistic areas.^

Stang (1935, 1939) provides additional data from Polack (northeast Beiarusan) and the chancery language of Lithuania. In texts from Polack, ‘have/take’ futures are found in the fourteenth century (1939: 35), while be-futures are found consistently throughout the fifteenth century (ibid.: 83, 124). In chancery texts of the Old Beiarusan period, be-futures are attested from the reign of Kazimierz Jagiçllohczyk (1447-92) through the sixteenth century (Stang 1935: 49, 112). Stang’s data confirms that one also finds ‘have/take’ {imu as well as (i)maju) as auxiliaries in the earlier period (ibid.).

Additional sixteenth-century evidence comes from the works of Francysk

Skaryna. including his Psalter of 1517 and Apostol of 1525. In his works, be-futures as well as futures with ‘have/take’ are attested (Bulyka et al. 1990: 114). Bulyka et al. (ibid.:

115) argue that instances of ‘have/take’ in Skaryna express modal meaning, and that in general the use of ‘have/take’ in this period is rare. Nevertheless, the modal nuance of

‘have/take’ would certainly not preclude its conveying future meaning, and this evidence does not speak to the issue of the be-future.

Cf. the discussion on this subject in Wcxlcr (1977: 58) and Lunt (1992: 466). -- Cf. the detailed discussions on this subject by Shevelov (1953: §1, 17. 22, 50. §8).

56 The Lithuanian Statute of 1529 also has examples of the be-future. In a previous study. I have demonstrated that the be-future is semantically differentiated in the text from the construction with ‘have/take’. This can be shown in the following example (4):

(4) a xto biidet’ iska\tM zemli ...ne maet' inoho nic{o]ho... ‘and who[ever) solicits lands [in the future] mav not [hold] anything else...’ (XIX.52)23

Here, the construction with have (to)’ clearly has obligative meaning, whereas the preceding be-future is used to express pure futurity (Whaley 1995; 34). Additional evidence that the ‘have/take’-type future in the 1529 Statute contains a lexical nuance of obligation can be seen in the apparently free alternation of the ‘have/take’ auxiliary with the adjective povinen obliged’ (ibid.).

It is clear that in the period of Old Beiarusan. examples of the be-future are attested consistently, intermixed with the ‘have/take’ construction. Jankouski (1983:

200-1) argues that the be-future and the ‘have/take-type future were in complementary distribution within the Beiarusan linguistic area throughout the Middle Ages, with individual dialects employing one expression or the other. Only with the modern-day codification of the literary language was the have/take’ future officially relegated to substandard status.-^ Textual evidence from Old Beiarusan presented above indicates that a complementary distribution of the two constructions was not necessarily in effect at the time; both are found in the same text. The brief analysis presented above suggests, moreover, that in some contexts, ‘have/take’ plus the infinitive conveyed an obligative

23 The context surrounding this example suggests that the complement of m aet', which should be the infinitive derzati. is missing. 2-1 See also Birala et al. (1957: 231).

57 meaning rather than simple future-tense reference. Nevertheless, judging from the use of the ‘have/take* future in modem Beiarusan dialects and the Ukrainian literary language, it is clear that ‘have/take’ was being grammaticaiized as a future marker at the same time that the be-future was undergoing the same process.

2.11 Ukrainian

2.11.1 Usage

A be-future with infinitival complement is standard for imperfective verbs in the literary language; it is in alternation with a synthetic future derived from ‘have/take', which is postposed to the infinitive, e.g. cekatymu versus budii cekaty I will wait’

(Shevelov 1951:306).

The ‘have/take* future is widespread throughout all dialect groups. Exceptions to this are some areas in the southwest that show only be-futures (see below) and small, north-central islands between Cemihiv and the Beiarusan border, where only the infinitival be-future is found. The Transcarpathian dialects also have the ‘have/take’ future, although the auxiliary remains separable from the infinitive (Zakrevska et al. 1984: maps 1:263.11:244).

Many dialects also have budu plus the /-participle as a future expression, and this construction is also found occasionally in literary texts (Rusanivs’kyj 1971: 250). The

Ukrainian dialect atlas (ibid.: map 11:244) reports that budu in combination with the

/-participle is found in a large area of southwestern Ukrainian dialects, extending from the

Polish border in the west deep into the areas associated with the Dnister and Podillja

58 dialect groups. Most of this area also usesbudu with an infinitival complement as a variant, but there are significant islands around L viv, Temopil', and south of

Xmel'nyc kyj where only the participial be-future is found. In addition, some areas around the Dnister show a further development of the participial future, in which the masculine singular complement is used for all genders; e.g. my budem rubiv ‘we will work' rather than the expected *my budem rubili (Zylko 1966:101-2).^

Data specific to the L viv region is found in Bandrivs’kyj (I960: 72-73), who reports that the participial future is most prevalent in the north and northeast areas of this region. For the area around Temopil’, Dejna (1957: 111-12) reports that the participial future dominates, with infinitival futures occurring only in careful or very formal speech.

Such a stylistic marking suggests that use of infinitive complements is due to influence of the literary language.

2.11.2 Chronology

Bevzenko (1960: 324-ff.) gives several examples of budu in combination with the infinitive from the late fourteenth-early fifteenth century, although he does say that the number of examples increases sharply in the 16th—17th centuries. He does not indicate the source for any of his examples, but the earliest, at least, are from texts that are also cited as evidence for Beiarusan. This overlap is not surprising, given the close affinity of

25 Final -/ vclarized in Ukrainian lo [v| or [wj; despite the similarity of the resulting form to the Slavic past active participle, it is improbable that rubiv is historically a past active participle given the rarity and limited use of such forms.

59 Beiarusan and Ukrainian al the time, and their mutual status as territories of the Grand

Duchy of Lithuania until the end of the seventeenth century.-^

From the sixteenth century, the Krexivs’kyjApostol of 1560 has examples of the be-future, along with the use of other auxiliaries such as ‘want’, begin’, and ‘have/take’

(Ohijenko 1930: 385). Other evidence of the be-future is found in the afterword to Ivan

Fedorov's primer of 1574, such as the following example (1):

( I ) i imëti hudesi v” poslëdnjaja dni nadezdu 'And you will have hope in the last days' (Jaskevic 1996: 103-4).

Fedorov, whose primer was published in L viv, uses the be-future several times in the primer's afterword. Within the main text of the primer, there is no mention of the be- future; only a future passive with have/take’ is mentioned (ibid.: 81). From the evidence in Fedorov's afterword, however, it is clear that the be-future was well-established at the time. In contrast to the evidence from Fedorov’s primer and the Krexivs'kyj Apostol, grammars from the sixteenth century are notably conservative in their descriptions. The anonymous grammarAdelphotës (1591/1988), for the example, describes futures with

have/take' and ‘want’—the latter being clearly Slavonic in character. In this grammar, the standard future form for imperfective verbs is the perfective counterpart. By contrast,

Zyzanij (1596/1972: 55) describes the future as formed with either have/take' or budu.

Grbschel (1972: 232) provides analysis of the early seventeenth-century writings of Ivan Vysens'kyj, in which there are 40 examples of the be-future. Grbschel (ibid.:

234-35) also finds seven examples of futures with want’ and one with have/take', but

See also the discussion in 2.10.2 above.

60 these occur only in religious passages where their use was most likely due to Church

Slavonic influence.

Descriptions of the imperfective future in seventeenth-century Ukrainian grammars of Church Slavonic suggest that the be-future was well-established by that time. Uzevic (1645/1996), for example, mentions no other possibilities for the imperfective future. Smotric’kyj ( 1619/1974) describes only nonpast perfective forms as future counterparts to the corresponding imperfective verbs, but his grammar is overwhelmingly Slavonic rather than vernacular in character.

2.12 Russian

2.12.1 Usage

A be-future is standard for imperfective verbs in the literary language, and its presence in dialects is described as “universal” by Mescerskij (1972: 191). In some

North Russian dialects, such as those of the Velikij Ustjug, Grjazovec, Kadnikovo, and

Cerepovec regions, one also finds ‘have/take’-type futures (Cemyx 1957: 229; Cemysev

1970: 256). Other regional dialects such as that in the Vologda area have stanu

‘become/begin’-^ plus the infinitive as an imperfective future (Mescerskij 1972: 188,

191).

Occasionally, one finds stanu in texts considered to be literary Russian, for example this oft-quoted example from Lermontov’sKazacjakolybel’najapesnja (1):

-7 See section 4.3.2 below for discussion of the semantics of this verb.

61 ( 1 ) Stanu skazyvat’ ja skazku, / Pesenkii svoju... I will (begin to) tell a story / my own song’ (Lermontov 1957: 38).

This usage is doubtless an imitation of folk language; for more information concerning the future with stanu, see section 4.3.2 below.

According to Cemysev (1970: 233—34), examples of the be-auxiliary in combination with perfective infinitives are found in dialects and folkloric contexts. This might appear to contradict the aspectual constraint on the infinitive that is present in the literary language, but Cemysev points out that all such examples occur in impersonal obligative constmctions. In this context, the copula and infinitive are not part of a compound future, but rather a biclausal constmction identical to what is found in OCS.“*

2.12.2 Chronology

A be-future is not found in the earliest Russian texts. In the absence of be- futures. Old Russian attests several strategies for expressing futurity. Besides nonpast forms of perfective verbs, auxiliaries based on the verbs imamb ‘have/take’ and xocu

'want’ are attested with infinitives of either aspect, while -cbnu begin’ and stanu

become’ combined with imperfective infinitives only. Non-be-future constructions are attested as late as the seventeenth century.

There are isolated cases of stanu and ‘begin’-type auxiliaries combining with apparently perfective infinitives. They appear in only a very limited context, occurring

28 See 2.2.1 above for discussion of OCS, and 3.3.1 below for general discussion of the impersonal obligative construction.

62 only with negation. As such, they appear to be a contextually restricted exception to the usual colligability of these auxiliaries. These have received little attention in the literature besides a brief discussion in Krîzkovâ (1960; 150-53); a more detailed examination of their use and function is beyond the scope of this study

Birchbark documents from Novgorod, which date from the eleventh to the fifteenth centuries, express the imperfective future primarily with an auxiliary based on the verb pocnii ‘begin’ (Zaliznjak 1995). Also found are examples of budii in combination with an /-participle; such constructions express future-perfect meaning.

However. Zaliznjak (ibid.: 159) claims that one example conveys pure future meaning (1).

In the document, the author is inquiring of another man, Nester, about a helmet that

Nester is selling to him.

(1) poslal” jesm' ... .k. bel” k tobë a ty Nestere pro cicjak” prisli ko mni gramotu s kim” budes" poslal” I sent twenty squirrel pelts ... to you, and you, Nester, send me a note about the helmet [telling me] with whom you will send / will have sent [it]’ (Zaliznjak 1995:455, no. 358).

Zaliznjak prefers the first translation ofbudes poslal ”, but based on the limited context of this example, his interpretation is not the only possible one. It might also be the case that

Nester has already sent the helmet, thus justifying the use of the resultative-oriented /- participle. The semantics of the future perfect construction in Slavic is fully compatible with the second interpretation; even though such an interpretation ostensibly resembles a past-tense formation, the point at which the author will receive Nester’s note is in the

Sec also the discussion in 3.3.2 below.

63 future. Often, uses of the future perfect conform to this type of context, describing the future acquisition of knowledge about a past event.^’

The earliest putative example of a be-future in Russian is from Article 99 of the law code Russkajapravda (2), the earliest edition of which is dated circa 1280;

(2 ) aze budiit’ v domu deti maly a ned[ii]zi sja budut' sami pecalovati. a mati im” poidet’ za muz . to to k[t]o im” blizei budut", tomu ze dati na rucé i s" dobyt"kom’ i s" domom”, donelé ze v" zmogut’... Tf there are small children and they are incapable of caring for themselves, and their mother marries, then whoever is closest to them, to that person [they] should be entrusted both with the property and the house until they are able...’ (PR 1940: 132; cit. in Obnorskij 1934: 765).

The status of this example as a genuine be-future is contested. Kuznecov (1959: 244) argues that there are two possible interpretations of the phrase, neither of which is an imperfective future. He suggests that the second budut' relates to the predicate neduzi

unable’, or that its presence is due to the influence of the preceding budut’. The former interpretation is the most reasonable for the context; the use of future-tense expressions in the protasis of conditional clauses is quite frequent in Old Russian, and can be seen, moreover, in the first clause of the example. Further support for the idea thatbudut’ forms a predicate with neduzi can be seen in examples like (3) from the Primary

Chronicle (s.a. 6479):

(3) my neduzi protivii vam" stati ’we are unable to stand against you’

For more discussion of the future perfect, see section 3.3.2 below.

64 In this example, it is clear that neduzi functions as a predicate adjective, which then takes the infinitival stati ‘stand’ as its complement. In other words, the infinitival form in (2) could be the complement not of the existential verb, but rather of the predicate adjective.

All editions of the Russkaja pravda have a version almost identical to the above example, save one from the Puskin group of texts, dating from the second half of the fourteenth century. In this example (4), the budut’ in question has been replaced by pocnut’ begin (3 pi.)’, and both the predicate adjective neduzi and the reflexive particle have been deleted.

(4) Aze budut’ v domu dëti maly, a ne pocnut' sami soboi pecalovati... (PR 1940: 290) If there are small children and they will not [begin to] care for themselves...’

The replacement of budut’ by pocnut’ results in a phrase that certainly does not capture the sense of the earlier version or make much sense in its context. One possible explanation for this puzzling evidence is that pocnut’ is in fact a hypercorrection; in other words, the copyist interpreted the phrase budut’ pecalovati as a future-tense expression, but knew that the be-future was proscribed by literary norms of the time. However, future-tense constructions with ‘begin’-type auxiliaries were acceptable; in fact, all other future-tense constructions in theRusskaja pravda are with ‘ begin ’ -auxiliaries (Obnorskij

1934: 765). Thus the copyist made a replacement that appeared appropriate.

This line of reasoning is entirely speculative, of course; there can be no re-creation of the original thought processes behind the creation of the Puskin-group example, and more data are certainly needed. Nevertheless, it provides an explanation for the

65 unexpected use of pocnut’ in (4). This example thus opens up the possibility that the be- future was already appearing to some degree in the spoken language at this early point in

Russian.

According to Krîzkovâ (i960: 175), the be-future construction appears in later fifteenth-century diplomatic correspondence between Rus and Poland. In at least one case, she has mistaken impersonal obligative constructions for be-futures, such as in this example from a document from 1492 (5).

(5) .../ cto budet" at” nas” vasej milosti molviti...... and if there will be something for us to sav to your grace...’ (ibid.).

Otherwise, Krîzkovâ s data do show the occasional occurrence of be-futures in the period. Nonetheless, Muscovite authors seem to favoriicnit ‘begin’ for future-tense constructions in this genre.

Additional early evidence can be found in two examples of the be-future found in the letters of Metropolitan Iona (d. 1461), an author whose language shows no influence of Western Rus’ian. For example, (6a) is taken from his letter to Prince Mixail Andreevic dated sometime after 1450, and (6b) is from a letter to the inhabitants of Pskov dated to the mid-fifteenth century.

(6a) ...a ne oboronis mene ty, moi syn", ... a jaz" biidii sia ot” nix" boronit’ zakonom" Boziim". ...and (if) you do not protect me, my son, ... I will be defended from them with God’s law’ (AI 1841: 99, no. 50).

6 6 (6b) ...molimsja celovëkoljubcu Bogu. ... da vospriimete ... milost' i storicniijii m'ldu, vremenno ze i budusce. ... po vasemu ot” Saga zelaniju i po vasej pravosti, kak" biidete kazdyj po svoemu xristijan’st\'u ziti. kak” to poslo ii vas”, ta vasa dobraja starina. ot" Velikogo Knjazja Aleksandr a. ..we pray to the philanthropic God ... that you will receive ... grace and a hundredfold reward, now and in the future, ... according to your desire from God and your rectitude, if each of you will live according to his own Christianity, as was your custom, that your good custom, from [the time of] Grand Prince Aleksandr' (ibid.: 107, no. 60).

These attestations from Metropolitan Iona’s writings and those from Muscovite diplomatic correspondence are strikingly early within the context of the be-future s chronology in Russian. Surprisingly, they are mentioned only by Krîzkovâ (I960: 174).

Aside from the be-futures in Metropolitan Iona's letters, fifteenth-century diplomatic correspondence, and the curious evidence from Russkaja pravda, the earliest examples of the be-future in Russian are scattered examples from the sixteenth century.

Many of these examples are putative, however, because during this period writers who use the be-future tend to be from Western Rus', where the be-future arose (or, at least, is attested) much earlier. For example, one of the earliest sources of examples for the sixteenth century is the work of Ivan Semenovic Peresvetov, an author from Western

Rus' whose texts are primarily petitions to the Czar (Lixacev 1989: II, 178).

Peresvetov's Skazanie o Maxmete Saltana, dated 1547 but preserved only in a copy from the seventeenth century, has eleven examples of the be-future, compared with one construction with begin' and two with ‘want’ (Krîzkovâ 1960: 176). There has been much discussion among scholars about Peresvetov, which is not surprising, given the

67 early date of his texts and the relative frequency of be-futures in them. His biography, however, has led some scholars to argue that his language was influenced by West Slavic and Western Rus'ian and thus not reliable for Russian in a strict sense.^'

One example of a be-future has been claimed from the Domostroj (1549):

(7) A kto na srok" vsjakix" obrokov” ... ne platit”, a ot toga otkiipaetsja, i dvë dani budet", ino uze vdvoe budet" nlatiti ‘And whoever pays no quit-rents within the term, and pays it [the debt] off. and there will be two payments, then [he] will pay twice*

This example is taken from the Sil’vestr edition; the first redaction of the text does not have a verb between vdvoe and platiti (Kolesov and Rozdestvenskaja 1994: 62, 127).

The context, however, calls into question the status of this example as a genuine be- future. There is a clear sense of obligation regarding the subject’s paying of the quit- rents; this context suggests that the notional subject, elided in the final clause, is actually the object of an impersonal construction and would appear in the dative if present (cf.

Nikiforov 1952: 179).

A Moscow gramota of 1588 is identified by Borkovskij (1949: 146) as containing the only Russian example in his corpus, which includes texts from all over

Rus* dated from the twelfth to the seventeenth century. The example is from a letter sent from Nikita Romanovic Trubeckoj, a Muscovite official assigned to oversee Pskov, to

Riga.

(8) A ne otpustite Timoxi v Ljubok", i jaz" o tom" budu pisat’ do gosudarja svoego...

See, for example. Krîzkovâ (I960: 176); Groschel (1972: 232): Borkovskij and Kuznecov (1963: 287).

68 But [if] you do not let Timoxa [go] to Ljubok. I will write about it to my lord... (Nap’erskij 1868: 399, no. 397).

Generally, the language of the text is clearly Muscovite Russian, save for the distinctively

Western use of do ‘to’ found in the passage cited here. Borkovskij’s identification of the text as Russian, rather than the more Western Rus’ian, appears to be well-founded. As such, this is one of the earlier examples of a be-future in a non-Westem Russian text, alongside those of Metropolitan Iona from over a century earlier.

During the seventeenth century, there is considerable evidence that both budu and stanu were used regularly as future auxiliaries. Awakum, for example, uses the be- future in his autobiography, but only as part of a single, repeating phrase: budu bit’ celom" 1 will petition’. All other future expressions from the text are withstanu. In

Awakum’s other works, however, there is a more even distribution of the two forms

(Cocron 1962: 241^2). Other seventeenth-century authors, such as Prince XovanskiJ, have future expressions with both budu and stanu, although the latter predominates (ibid.:

244).

Additional evidence from the seventeenth century comes from Pennington ( 1968), who has compared Russian translations of a Polish text,Dwor cesarza tiireckiego

(1646), which date from 1649 to 1690. According to Pennington, the earliest, most colloquial^- translation favorsstanu, while budu is found only in the presence of a Polish model. Other, later translations have more examples ofbudu, while stanu is also found.

The most “literary ” text (in Permington’s own words) usesimam’ most often (1968: 40).

Pennington uses this tenn because the text is relatively free of Church Slavonicisms.

69 Though hardly statistically significant, her data lends further evidence of an alternation

between sranu and budu in this period.

Also dating from the seventeenth century, Vesti-kuranty (documents consisting of summaries of Western newspapers, as well as diplomatic correspondence) of the period appear fairly progressive for their time. Clark's (1978: 35) analysis of a selection of these texts from the 1620’s shows four be-futures, in contrast with a single begin-future. A study of private correspondence from the same period by Vontsolos (1978) shows that both budu and stanu were used as future-tense auxiliaries, with the same verb capable of combining with either auxiliary (ibid.: 91).

At the end of the seventeenth and into the eighteenth century, we see the use of both stanu and budu described in grammars. Ludolf (1696/1959: 30). for example, describes the future as formed with either auxiliary, although the dialogues he provides contain only examples withstanu.

In the eighteenth century, the alternation ofbudu and stanu continues to appear in grammars, including those of Adodurov (1731/1969) and Barsov (1784-88/1981). In the original Russian edition of his grammar, Lomonosov (1755/1972: 130) describes the

“future indefinite " (i.e., imperfective future) as formed with budu. Later, however, he describes the relationship between stanu and budu as future auxiliaries, pointing out that stanu, unlike budu, can be used only with infinitives, and not in combination with other forms such as passive participles (ibid.: 199—200). The expanded, German edition of

Lomonosov’s grammar mentions only that bothbudu and stanu are used as future auxiliaries (1764/1980:201).

70 Very little work has been done to clarify the relationship between budu and stanu as future auxiliaries in Russian. For example, Krîzkovâ s (1960: 146) only comment regarding usage is that contexts where stanu are attested are scattered across different genres. Based on examples from seventeenth-century texts. Cocron (1962: 244) argues that stanu is used in more “personal” and “concrete” contexts thanbudu, but these terms are vague. In Pennington’s (1968: 42) data, stanu is not found in combination with the infinitives imét’ ‘have’ or svidéteVst\'ovat’ ‘bear witness’; she argues that this restriction in the use of stanu indicates it was not a fully grammaticalized future auxiliary at the time

(ibid.: 41). Without a more systematic investigation of patterns of attestation, however, differences in the usage of stanu and budu in this period can only be guessed at.

It is in the nineteenth century that grammars begin to present descriptions of the imperfective future that resemble those of the present day. Whereas Lomonosov only hinted at the idea that stanu and budu were not employed as functional equals, now one finds the perspective that budu is the only “pure ” future auxiliary. Such a position is found in Buslaev’s (1959) grammar, originally published in 1881. In it, he lists several possible future-tense auxiliaries, including those based onxocu want’, îme/w have/take’, nacnu ‘begin’, as well as stanu and budu. In his description of these various auxiliaries, he argues that whereas stanu and nacnu convey an inceptive meaning, and xocu and imeju convey intention or inclination,budu has no other semantic nuance besides futurity (ibid.:

361-62),

Thus with the grammars of Lomonosov and Buslaev, one sees the beginning of the modem point of view regarding the status of the be-future in Russian. Despite the

71 overwhelming dominance of the be-future in modem Russian, however, there remain traces of an imperfective future with stanu. not only in dialects and the admittedly colloquial example from Lermontov given in 2.12.1 above, but also in other literary works, e.g.. those of Goncarov, Turgenev, and Kuprin (Kuznecov 1959: 243-44). Other future auxiliaries found in Old Russian, such as ‘begin’- and have-types, have fallen completely out of use.

2.13 Concluding remarks

As has been shown above, in many of the Slavic languages (including the early- attested East Slavic and Serbo-Croatian) there is a significant gap of time between the end of the Common Slavic period and the first appearance of the be-future in texts, with many languages attesting alternative constructions prior to and contemporaneously with be- futures. A schematic map of the relevant languages, showing the earliest date of attested be-futures and the earliest date of extant texts, is found in Figure 2.1 below. Based on the data shown in Figure 2.1, it app>ears unlikely that the be-future can be reconstructed as part of Common Slavic.

The data also indicate that the Slavic be-future is not a monolithic construction with a single usage and chronology across all Slavic languages. As Figure 2.2 below shows, differences abound with regard to different treatments of participial and infinitival complements and different degrees of restriction on the aspect of the complement. These differences, combined with additional evidence of widely diverging dates of the constmction’s earliest appearance as well as variation among dialects, suggest that

72 although each language inherited the verb *bqdq and its potential to develop into a future- tense auxiliary, the development of the be-future took distinctly different paths within individual languages.

If the be-future cannot be reconstructed for Common Slavic, one necessarily must argue that the origin of the be-future be placed in the post-Common Slavic period. This is by no means a novel idea; many scholars have argued in favor of this point when theorizing about the development of the be-future. What has not been attempted previously, however, is a reconciliation of the raw data involved in this problem with general linguistic theories regarding the nature of language change. An examination of the problem from this perspective in fact leads to a new analysis of the development of the

Slavic be-future.

Before this new analysis can be presented, it is necessary to first evaluate the merits of the work that has previously been done on this problem. The following chapter will present a description and discussion of previous scholarship on this topic, including commentary on how well the theories that have been presented can be reconciled with the data presented above.

73 West Slavic East Slavic

Russian 16th c. (16th c.) Mid-15th c. 16th c. ( 16th c.) Consistent attestation: 17th c. (1 1th c.)

German Belarusan 14th c. (12th c.) Late 14th c. (Sorbian) (Mid-14th c.)

Ukrainian Czech Early 14th c. 14th c. (1 1th c.) (Earlv 14th c.) Slovak

South Slavic 14th c. (14th c.)

Slovene ^ Serbo-Croatian 14th c. (Cakavian); 16th c. 15th c. (Old Dubrovnik) (1 0 th / 16th c.) (12th c.)

Figure 2.1. Schematic map of the be-future in the Slavic languages: Earliest dates of attestation / earliest dates of extant texts (in parentheses).

74 West Slavic East Slavic

•Inf. or part. Russian •Inf. only •Impfv. only •Eilheraspecl •Inf. only •Impfv. only

German Belarusan Polish Inf. only Inf. or part Impfv. only f.Sorbian) Impfv. only Ukramiap •LL: inf. (Part, possible in western dialects) Slovak Impfv. only

•LL: inf. only Part, possible in eastern dial. (Part. poss. in E) South Slavic •Impfv. only Impfv. only

■Slovene ^ Serbo-Croatian •LL: part, used as futurum exact urn •Part, only (Inf. possible in dialects, as is pure •Eilhcraspect future meaning) •Eitheraspect

Key = languages or dialect regions with □participial complement possible = languages or dialect regions with only participial complement LL = literary language

Figure 2.2. Schematic map of the be-future in the Slavic languages; T reatment of complement and aspect.

75 CHAPTER 3

PREVIOUS SCHOLARSHIP

3.1 Introduction

The data in Chapter 2 suggest that the development of the be-future must be placed at a point after the Common Slavic period. Thus, a satisfactory description of the be-future should serve as an explanation of how the construction was innovated in the individual Slavic languages that now employ it. There have been many theories regarding the exact details of this development, and they can be divided into two categories: those which identify the source of the be-future as external to Slavic and introduced into the

Slavic languages via borrowing or calquing (hereafter, “extemal-source” theories), and those which identify a source from linguistic material native to Slavic (“intemal-source” theories).

The purpose of this chapter is twofold: to summarize the literature that has been produced regarding the be-future and its development, and to demonstrate that a satisfactory description of the development has not yet been presented. These goals will be achieved by analyzing both external- and intemal-source theories, addressing the advantages and shortcomings of both as revealed by the data presented in Chapter 2 and what is known about morphosyntactic change in general.

76 3.2 Evternal'source theories

Many scholars who have explored the nature and origin of the Slavic be-future have concluded that the construction arose as a result of the influence of a non-Slavic language. German is considered the source language by most scholars, although Latin and Greek are also mentioned.

Attempts to find connections between Slavic and non-Slavic future-tense expressions appear in the literature very early on. One of the earliest studies conceming the origins of the be-future is that of Schleicher (1855), whose work concems the parallels that appear to exist in the future formations of Slavic and Gothic. These parallels include the absence of the PIE future suffix *-s-, and the use of change-of-state verbs and verbs of obligation as future auxiliaries. However, Schleicher stops short of arguing that a language-contact situation is responsible for the similarities between the futures in

Gothic and Slavic; rather, he believes that Germanic and Slavic formed a single PIE subgroup and that differences between them are due to later dialectal differentiation.

Although more recent scholarship indicates that Germanic and Slavic formed no such subgroup, Schleicher is nevertheless the first scholar to suggest a connection between the

Germanic and Slavic futures.

Polivka (1888) also sees similarities between the be-futures in Germanic and

Slavic. When comparing Biblical texts in the original Greek and translations in OCS, he notes the diversity in the translation of future-tense expressions and concludes that in

OCS there was no single grammaticalized future expression (ibid.: 193). He argues that a similar situation existed in Germanic (i.e.. Gothic), as attested in the Wulfila Bible, and

77 concludes by suggesting that a comparison of OCS and Gothic would shed light on the problem. Again, like Schleicher. Polivka does not suggest that borrowing is responsible for the similarities in Germanic and Slavic, but states only that such similarities should be explored further.

Polivka s study and, indeed, any study which concems the future-tense expressions of OCS, does not directly concern the be-future, since such a futine was absent from OCS (see section 2.2.1 above). Nonetheless, these studies are relevant for the matter of the be-future; they involve the larger question of the development of periphrastic futmes in Slavic, and indicate that scholars have long sought a cotmection between the Germanic and Slavic futures, even at a point prior to the appearance of the be-future.

In a more recent work that identifies borrowing as the mechanism behind the development of the be-future, Bonfante (1950) argues that all periphrastic futures in

Slavic are of Greek origin. He reaches this conclusion based on the fact that Proto-Slavic, in his words, “had originally no real future” (ibid.: 96). As a result of this deficiency,

Bonfante argues, when Greek texts were translated into the newly created OCS, Greek periphrastic futures were imitated in OCS texts by calqued periphrases.

Kravar (1986) also sees the be-future as an inherently non-Slavic construction, arguing that it is a caique of the VMm futurum exactum of the late Middle Ages. In support of this position, Kravar argues that Latindixero ‘I will have spoken', formed with the future of ‘be’ attached as a suffix to the perfect stem of a verb, corresponds to a hypothetical Slavic caique reci budu, later reversed to budu reci according to Kravar's

78 iheor>' (ibid.: 279—80). His identification of Latin as the source language is motivated by the fact that the earliest attestations of a be-future are in West Slavic, a group whose

languages were influenced by Latin throughout the Middle Ages.

Whereas Bonfante (1950) sees the source of the be-future as Greek, and Kravar

(1986) argues for Latin, Rosier (1952) argues for Middle German. The work of Rosier

(1952) remains the most detailed argument for an extemal-source hypothesis, and has been cited by several sources, including Cocron (1962) and Galton (1976, 1979, 1981), as the most convincing explanation for the presence of be-futures in Slavic.

Rosier argues that German developed its future construction with werden (like

*bqdq. a verb with change-of-state meaning) plus the infinitive in the eleventh-thirteenth centuries, from a construction usingwerden plus the present participle. According to his theory, this werden-ïwWxxc was then borrowed into Czech, which had close contact both politically and linguistically with Germany throughout the Middle Ages. This borrowing took place before the appearance of the earliest texts written in Old Czech. Rosier estimates the time frame of the borrowing to be the late thirteenth century (ibid.: 142), which is also the time of the earliest surviving Czech texts (see section 2.5.2).

From Czech, Rosier's theory continues, the be-future was borrowed into Polish.

According to Rosier, the earliest Polish texts, dated to around the middle of the fourteenth century, show be-futures with the infinitive. In his view, their presence is indicative of

Czech influence, a language which had considerable influence over Polish in the earliest stages of the Polish literary language (ibid.: 125).

79 Rosier next proposes that the construction was borrowed from Polish into East

Slavic. According to Rosier's data, the be-future construction begins to appear in westernmost East Slavic circa the late fourteenth century, spreading into Ukrainian by the sixteenth century (ibid.: 135). It is a simple matter for Rosier to demonstrate Polish influence over Old Belarusan and Ukrainian, given that the territory occupied by speakers of all three languages was part of a single political unit, the Grand Duchy of Lithuania, and there is considerable linguistic evidence of contact between Polish and Western

Rus'ian.

From the western reaches of East Slavic territory, the theory continues, the construction then spread eastward and ultimately was adopted into literary Russian.

Given that Rosier believes that the first be-futures do not appear in Russian until the seventeenth century. Rosier argues that he has a sufficiently late endpoint to his proposed sequence of borrowings.^^

While the arguments of Bonfante, Kravar, and Rosier differ in many ways, including the proposed source for the Slavic borrowing, they have much in common.

Several factors inherent in the Slavic situation have conspired to persuade these scholars that the be-future cannot be an entirely native Slavic construction. The following sections will explore these factors, why they have been perceived as legitimate by proponents of the borrowing theory, and why they ultimately fall short, rendering the borrowing theory a nonviable hypothesis.

33 Although Rosier does mention Slovene in his discussion of the Slavic data, he does not describe its place (or lack thereof) in his proposed theor) .

80 3.2.1 The putative unusualness of a become-type future

Of all the arguments presented by scholars in support of an extemal-source hypothesis, the most frequently encountered assertion is the idea that it is rare for a change-of-state verb like “become’ to develop into a future auxiliary—rare enough for its evolution in so many Slavic languages to be considered an unnatiual development. A more careful analysis of the data and a broader, more cross-linguistic perspective reveal that the development of a change-of-state future in Slavic need not be motivated by external influence, and that such a development is not particularly unusual.

A review of the future tense expressions of languages in Europe reveals only a small number of languages or language groups that have developed ‘become'-type futures: Classical Latin (where a synthetic future has developed), Celtic, Germanic

(including High German), and Slavic. The geographic proximity of German, as well as

Latin, to the Slavic linguistic area appears to lend support to the idea that change-of-state future constructions might have developed from a single source within Indo-European.

In order to assess the validity of this view, one must determine not only whether a

“become’-type future is actually rare, but also whether there is such diversity in sources of future constructions that two similar constructions are likely to be related. These questions have been resolved by studies of the grammaticalization of future-tense auxiliaries. For example, in a series of cross-linguistic studies of the future tense. Bybee and her colleagues (1987, 1989, 1994) examine a representative sample of the world’s languages in order to analyze the most typical kinds of future tense formations. By their assessment, there are in fact only a small number of lexical sources for future tense

81 constructions, and one of those sources is verbs of existence or coming into existence

( 1987: 111). Besides the languages identified above, Ultan (1978:110) reports that such futures are also found in some Celtic languages. Cuna. and Upper Chinook.

If there are in fact only a limited number of possible lexical sources for future tense expressions, then it becomes more likely that two languages would develop futures from similar sources independently. In many cases, of course, geographical and genetic distance between two languages greatly reduces the possibility that they share the same source for a future construction; it is clear, for example, that the English futurewill is not related to the Bulgarian future marker ste, despite the fact that they are both derived from desideratives. In the case of Slavic, there are neighboring languages that do share a similar lexical source for their future-tense constructions, but the data of grammatical­ ization scholars indicate that similarity of lexical source alone is not sufficient to justify an argument of direct relation between future constructions.

Another factor which Rosier, Kravar, and Bonfante employ to bolster their arguments is the fact that the be-future carmot be reconstructed for Common Slavic.^

This leads Kravar (1986: 276) to conclude that the construction is “only partly Slavic” and to seek a non-Slavic origin for the construction. In a similar vein, Bonfante (1950;

96-97) argues that Slavic borrowed a future construction from Greek into OCS because

Common Slavic “had originally no real future”.

Kravar’s argument seems to imply that the Slavic be-future could not have developed without external influence. However, there is nothing unusual or particularly

See section 2.2 for more a detailed discussion of Common Slavic and the be-future.

82 atypical for Slavic in the structure and formation of the be-future construction. In fact, the be-future appears similar in structure both to other periphrastic future-tense constructions and to many past-tense constructions in early Slavic, and in no way stands out as a

"misfit" in the Slavic verbal system.

As for Bonfante’s argument, his conclusions are based on faulty premises. OCS in fact had several strategies for expressing future tense, including nonpast perfective verbs as well as several periphrastic constructions (see 2.2 above); one cannot argue that no "real future” existed in Common Slavic. One should also recall Ultan’s observation

(1978: 88) that it is possible in many languages for present-tense forms to describe future events, such that tense expressions in these languages can be viewed as distinguishing a binaiy opposition between “past” and “nonpast”.^^ In this way, a system such as that of

OCS cannot be considered to contain a gap that would need to be filled.

It is certainly true that the periphrastic structure of the be-future is unlike that of other Slavic nonpast-tense forms; however, this does not necessarily mean that the construction is a borrowed form. One possible diagnostic for determining if such forms were borrowed is a comparison of early Slavic translations of foreign texts with the original texts. If the presence of analytic future-tense constructions in Slavic were due only to foreign influence, one would expect to find such constructions in Slavic texts used more consistently, if not exclusively, in the presence of foreign models.

According to data presented by H. Bimbaum (1957, 1958), such a correspondence is not found. After comparing OCS texts to equivalent texts in Gothic,

^5 Sec also 1.2.2 above.

83 Bimbaum (1957) concludes that the analytic future constructions in Slavic cannot be considered caiques of the equivalent constructions in Gothic. For example, in his comparison of the Wulfila Bible and the OCS Gospels, he finds that of the 88 examples in OCS of future constructions using the ‘have’ auxiliary iméti, only one corresponds to the similar Gothic periphrasis with skiilan have to; be obligated’. For the other 87 cases, an equivalent is either lacking or corresponds to a monolectic construction (ibid.: 79). For the four instances where OCS uses a construction with a prefixed form of-c^ti begin’ plus the infinitive, only one corresponds to an expression in Gothic withduginnan

begin’, and Bimbaum considers both to be imitations of the Greek original (ibid.: 80).

Although Bimbaum identifies some analytic constmctions in OCS as caiques of

Greek, his data shows that there is no evidence to assume that all such constmctions in

OCS are imitations. This is made clear by his subsequent (1958) work, in which he compares OCS analytic future expressions to the Greek originals. In this work,

Bimbaum finds no evidence of a one-to-one correspondence between these forms. In fact, OCS future expressions do not always correspond to any kind of periphrastic constmction in the Greek (ibid.: 197,199). If periphrastic constmctions in Slavic actually had their origins in Greek, one would exp»ect a more consistent correspondence between

OCS and the Greek originals; Bimbaum has shown that this is not the case.

Cross-linguistic evidence also does not support the idea that it is unusual for a verbal system to have a mix of synthetic and analytic verb-forms; in fact, many of the world's languages utilize both synthetic and analytic tense constmctions. Regarding future constmctions specifically, analytic constmctions are far from rare, as Bybee and

84 Dahl ( 1989) show in their cross-linguistic study. Moreover, there is no reason to assume that an analytic tense construction is unusual even in languages such as Slavic, which favor inflection as a means of indicating grammatical relationships.^^ In the history of

Slavic, analytic future constructions have always existed. According to many scholars, including Benveniste (1968), Andersen (1987), Bybee et al. (1994: 19—21) and Heine

(1993: 54-58), in many cases analytic and synthetic verbal constructions can be seen as representing different diachronic stages of development.^^ From this perspective, synthetic forms can be viewed as a later stage of grammaticalization for a particular construction, reflecting the phonological reduction and agglutination that often characterize forms that are more fully grammaticalized. In Ukrainian, for example, one can trace the origins of the synthetic future with ‘have/take’ back to an earUer analytic construction of verb plus infinitive, which originally conveyed purely lexical meaning (cf.

Bevzenko et al. 1978: 294-96).

Thus, in the case of Slavic, it might be best to view analytic future constructions not as unusual from the standpoint of the Slavic verbal system, but as typical from the standpoint of future-tense constructions in the world's languages. In sum, one cannot argue that a construction in Slavic is of foreign origin based on the fact that it is analytic.

Sec also the discussion in 1.2.1 above. Many more scholars who subscribe to this view are cited by Hopper and Traugott ( 1993: 18-30).

85 3.2.2 Chronology and the emergence of the be-future

According to Rosier (1952), when the dates of the construction's emergence in each Slavic language are compared, a geographical and chronological pattern of development becomes evident, revealing a progression from twelfth-century Czech in the west to sixteenth-century Russian in the east. Kravar ( 1986) subscribes to a similar view, with Latin as the source language rather than German.

Both Rosier and Kravar are making an assumption that the synchronic evidence of the earliest date of attestation in each language can be correlated to a diachronic pattern of the be-future construction's dispersal throughout Slavic. There are, however, problems with this understanding of the relationship between chronology and the modern-day distribution of the be-future; moreover, additional data for some languages do not fit in with the theory proposed by Rosier. These problems and data will be discussed in the following subsections.

3.2.2.1 Reliability of the textual evidence

The main supporting evidence employed by Rosier to back his claims is textual evidence, a source which I argue is problematic at best. Whaley (1999) presents ways in which textual evidence falls short in providing explanations of the path of development of the be-future; these arguments are reproduced here. For example, when considering a particular linguistic feature in the earliest extant text, there is no way of knowing whether an earlier, non-extant text would contain the same feature. It is unlikely that all texts that have ever existed have survived until the present day, and thus there must be gaps in the

86 written record. In Polish, for example, the earliest extant texts are dated especially late, and early data are unavailable.

The copying and recopying of texts over time also opens the possibility of later contamination of the language of earlier works—we have seen examples of such textual variation in the early Russian evidence, for example (see section 2.12.2 above). A gap also exists between the more normalized and conservative language of written texts and the contemporaneous spoken language. It takes time for a innovation that originates in spoken language to manifest itself in written texts.

At the same time, it is possible for “old” words to be attested relatively late, further weakening the notion that earliest date of attestation can be considered a reliable indicator of a form’s age. This can happen even in languages with a well-represented textual history; e.g., Janda and Joseph (forthcoming: §1.2.1.2) report that the Greek word

éor daughter; relative; kinfolk’ is attested only in the fifth century A.D., but appears to be descended from the PIE *sw’és(o)r ‘sister’. Thus, in some cases, textual evidence provides little insight into the chronology of a form.

Rosier’s description of the spread of the be-future across North Slavic territory becomes unfeasible in light of the problems associated with relying on written evidence.

In order for the construction to become available to each consecutive language in his proposed chain of borrowings, it would have to become established in each language at a speed that seems unusually rapid. Moreover, Rosier’s chronology does not perfectly fit the data; the earliest written attestations of the be-future in Belarusan and Ukrainian, for example, are found in fourteenth-century texts—roughly contemporaneous with the

87 earliest examples in Polish (see 2.10.2 and 2.11.2. respectively). Outside the scope of

Rosier s theory, such data are not problematic. Within the confines of his argumentation, where written attestations are paramount, the data do not support his conclusions.

3.2.22 The chronology in Russian

In general, the Russian data are not compatible with Rosier"s theory. As shown in 2.12.2, examples of the be-future are attested sporadically beginning in the fifteenth century, but the construction does not become widely attested until the seventeenth century. It is this later date that Rosier uses for his chronology.

In establishing a chronology for the be-future in Russian, one cannot discount the few examples that do exist from an earlier period. It is true that many early examples of be-futures in Russian are from texts whose language shows features of Western Rus’ian, where the be-future was attested more consistently an an earlier period. However, examples such as those from the correspondence of Metropolitan Iona are surrounded by unequivocally Muscovite Russian and are not significantly later than the earliest examples found in Western Rus'ian texts. One should also note that the purpose of Iona’s correspondence was to conduct official business, and the language of his letters is correspondingly full of the more conservative Slavonicisms characteristic of high style.

Considering the narrow window of time between the first examples of the be-future in

Western Rus" and the appearance of the be-future in the official writings of Metropolitan

Iona, it appears unlikely that the appearance of the be-future in Russian could be attributed solely to borrowing from Western Rus".

88 3.2.23 Ukrainian evidence

Rosier's theory also fails to find supporting evidence in the data regarding

Ukrainian, a problem that is discussed in previous work (Whaley 1999: 165). Rosier

(1952: 144—45) argues that the borrowing of the be-future from Polish into Ukrainian was a manifestation of the linguistic dominance that Polish had over Ukrainian in the medieval period. However, the relationship between Polish and Ukrainian is far more complex than Rosier suggests. As Shevelov (1952: 348—49) argues, Polish influence over Ukrainian was not dominant untU after the sixteenth century. During the earliest period of contact between Polish and Ukrainian peoples, from the tenth to the fourteenth centuries, this prestige relationship was not yet in effect, and Ukrainian influence on

Polish also can be identified.

The earliest data regarding the be-future in Western Rus* also do not appear to support Rosier s hypothesis. For example, the be-future is attested in this area as early as the late fourteenth century, although Rosier’s earliest examples from Ukrainian date from the sixteenth century.^* As was mentioned in 2.10.2 above, it is anachronistic to assign language from the fourteenth century a label such as “Ukrainian” or “Belarusan”.

Nonetheless, given that the earliest Polish examples date from the same period, it is premature to assume that the be-future appeared first in Polish and then was borrowed into Western Rus’.

Data from modem Ukrainian dialects (presented in 2.11.1 above) also suggest that

Rosier s characterization of events is inaccurate. There are islands wherebudii plus the /-

This error is also pointed out by Groschcl (1972: 232).

89 participle is the sole future formation, and these are in areas contiguous to Poland. This suggests that if any future construction was borrowed from Polish into Ukrainian, the more likely construction to have been involved appears to bebudii plus the /-participle.

Given the nature of the relationship between Polish and Ukrainian described by

Shevelov (1952), as well as the evidence for the early attestation of the be-future in

Ukrainian, it appears highly unlikely that Ukrainian borrowed the construction of budii plus the infinitive from Polish, especially considering the narrow window of opportunity upon which Rosier’s theory depends.

3.22.4 The chronology of German werden

The problems of the borrowing hypothesis with regard to chronology are not restricted to the Slavic side of the equation. In his review of Rosier’s article, Kurz (1952:

154) is the first to point out the error regarding the development of future in German with werden: he argues that the be-future was already attested in Czech prior to the first occurrence of the German construction.

Leiss (1985) corroborates Kurz’s statements regarding the chronology of the werden-ïwVüxc, arguing that Rosier is incorrect when he states that thewerden-fwX\xx& was available for borrowing into Slavic as early as the thirteenth century. In reality, Leiss argues, examples of werden plus the infinitive are almost nonexistent before the thirteenth century, and the scattered examples from earlier texts have been characterized by some scholars as scribal errors (ibid.: 257). With the earliest Czech examples dating from the

90 end of the thirteenth century, it is impossible to argue that the construction ofwerden plus the infinitive was available for borrowing at a sufficiently early point.

3.2.3 Infinitival versus participial complements

Another problematic aspect of the theory proposed by Rosier is that it overlooks the significance of the nonfinite constituent of the future periphrase. In order to begin the chain of borrowing with German and end with the correct result for Russian, one must assume that Rosier is concerned with the future constructiono f *bqdq plus the infinitive.

This poses no problems for the proposed transmission from German into Czech, or indeed the proposed borrowing of the construction into East Slavic (notwithstanding

Ukrainian dialect evidence presented above). Polish, however, employs a participial complement as well, and Rosier's argument does not consider this factor.

The data for Slovene are similarly problematic for Rosier's hypothesis, yet he makes no attempt to reconcile them. If one supposes that the German future withwerden plus the infinitive was calqued into Slovene, one would assume that the resulting construction would be an infinitival be-future. However, as the data in section 2.3.2 indicate, there is no evidence that Slovene ever had a be-future construction with the infinitive.

Kravar (1986) errs on this same point in his argument that the Slavic be-future with infinitival complement is a caique of the later Latin future perfect construction. As is mentioned above, the Latin future perfect is formed with the perfect stem plus an inflectional ending derived from the future of ‘be*. Kravar argues that by the time of the

91 borrowing into Slavic, the perfect stem had been reinterpreted as a nonfinite form analogous to the Slavic infinitive (ibid.: 279-80).

Kravar's argument is not supported by the data. These show that, in contexts where Latin future perfects were translated into Slavic, the Slavic future perfect, i.e. the construction with*bqdq plus the /-participle, is the form that is attestedIn other words, it appears that the Slavic equivalent of the Latin future perfect was a construction with the

/-participle, not the infinitive. Evidence of this can be found in Polish (cf. 2.8.2 above).

In sum, the issue of the participial be-future is not addressed satisfactorily by either

Kravar or Rosier.

3.2.4 The Slavic be-future and aspect

In proposing a connection between Slavic and non-Slavic futures, one must also consider the ramifications of aspectuality with regard to the behavior of the constructions.

Any discussion of aspect is absent from the arguments of most scholars, including Rosier

(1952) and Kravar (1986), except to specify that the be-future is used with only imperfective verbs in North Slavic.

Leiss (1985) does consider the role of aspect in the formation of be-futures in

German and Slavic, although some of her argumentation is incorrect. She argues that werden could not have developed independently into a future marker because of its aspectuality, whereas Czech biidii, an imperfective verb, is a perfect candidate for

A more detailed description of the relationship between the Common Slavic future perfect and the participial be-future can be found in section 3.3.2 below.

92 grammaticalization into a future auxiliary. Leiss concludes her argument by positing that

German in fact borrowed its future construction from Czech (ibid.: 264-65).

Leiss is incorrect when she argues that perfective verbs are less suitable future auxiliaries; recall, for example, the prefixed begin-type future auxiliaries in many Slavic languages, as well as stanii in Russian—all perfective verbs that have developed into auxiliaries. Leiss also misconstrues the aspect ofbiidii, which is generally understood to be perfective in early Slavic (see section 2.1 above). These considerations do not, however, contradict her assertion that the be-future was likely not borrowed from German into Czech.

Although Leiss discusses only the aspect of the auxiliary in her analysis, the aspect of the complement plays an important role in the evaluation of extemal-source theories, including those of both Rosier (1952) and Kravar ( 1986). One must remember that, for most of the Slavic languages, the be-future can be combined with only imperfective verbs, whereas aspect of the complement is not an issue for Germanwerden or the Latin future perfect. The aspectual constraint on the Slavic construction is an important factor that must be explained by any theory that seeks to describe its origins. In fact, the existence of the aspectual constraint in Slavic suggests that the only possible direction of borrowing regarding the be-future is from Slavic into another language.

The importance of the role of Slavic aspect on verbal borrowings can be illustrated by comparing hypothetical scenarios of borrowing. If one were to speculate, for example, that Czech calqued the German future construction using the native equivalent of werden, one would have to propose that the borrowing and assimilation was a two-step process:

93 1 ) calque of w erden-iy^ future construction using native biidii 2) imposition of aspectual restriction on complement so that only im­ perfective infinitives can be used

Evidence of this process would likely include the attestation of perfective futures in Slavic

both with *bqdq plus the infinitive and with nonpast forms. Such a situation is not

reflected in textual evidence, even in languages with extant texts from a very early period.

On the other hand, the following scenario describes the necessary steps that

would occur if German borrowed the construction from Czech:

1 ) caique ofbudu-type future construction using native werden 2) generalization ofwerden-Vypc future so that any verb may be used as complement, regardless of aspectuality (i.e., Aktionsart)

German verbs do not have the same type of aspectual distinction as do Slavic verbs.

Indeed, it is most likely that a German sf>eaker with an imperfect knowledge of Czech

who hears the Czech future construction would not recognize or adopt the aspectual constraint on its use. Thus, the constraint would go unnoticed and would not be

borrowed into German. The latter scenario eliminates the entire problem of explaining

why Czech would create a special restraint constraining the colligability of a

newly-borrowed construction from either German or Latin.

The above analysis does not support Leiss’ ultimate conclusion that the construction ofwerden plus the infinitive could not have developed independently and

was borrowed from Czech. For the purpose of this study, however, the importance of the conclusion is clear: the proposal that Slavic borrowed its future construction from

German or Latin is an unlikely possibility.

94 It might appear that the situation in Serbian. Slovene, and Serbo-Croatian is

incompatible with this argument; in these languages, verbs with either aspect can serve as complement to the be-future. The history of aspect in Slavic, however, suggests that the treatment of aspect in these languages reflects innovative development. It is hypothesized that the individual Slavic languages each inherited an aspectual system from Common

Slavic where nonpast perfective verbs tended to convey future-tense reference; such a situation is already seen in OCS (Lunt 1974; 135; H. Bimbaum 1958; 17; Schenker 1993:

94—95). Thus, although Serbo-Croatian, Slovene, and Serbian all inherited the same treatment of aspect from Common Slavic, they have innovated in their use of the category.

Til is is well-documented for Serbo-Croatian; Belie (1969; 133), for example, confirms that in earlier stages of Serbo-Croatian, nonpast perfective verbs often conveyed future meaning. Only later, possibly due to the influence of neighboring languages and/or the generalization of a 'w ant-type future, did the aspectual system undergo revision.

Slovene, a language heavily influenced by German, appears to reflect a transitional state of affairs. As was mentioned in 2.3.1, both types of aspectual system are found among dialects and, to some extent, in the literary language; either nonpast perfective verbs convey future-tense reference, or a be-future is combined with complements of either aspect. Thus, by preserving the usage of nonpast perfectives, Slovene patterns like the North Slavic languages, but its generalization of the be-future to all aspects is more typically South Slavic (cf. Galton 1979,1981).

Sorbian lacks the historical data that would give evidence for an older situation where the aspectual constraint on the be-future was still intact. The Germanization of

95 Sorbian speakers has been ongoing since the eleventh century (Polanski 1980). and the

modem state of the aspectual system has certainly been affected by the interference

between the two languages due to the nearly universal bilingualism of Sorbs (cf. 2.7.2

above). An earlier situation with a more typically North Slavic aspectual system can be

assumed, however, based on the comparative Slavic evidence, the fact that perfective be-

futures are not found in all Sorbian dialects, and the fact that normative grammars of

Sorbian proscribe the use of perfective infinitives with the be-future (Stone 1993a: 637;

FaBke 1981: 253), Such a treatment suggests that the use of the be-future with both

aspects is acceptable in the more innovative spoken language of some dialects, while the

conservative norms of the literary language continue to resist the development.

Given the innovative nature of the treatment of aspect and the future tense in these

languages, one can argue that nonpast perfective verbs conveyed future-tense meaning in

the earliest stages of all Slavic languages; thus, the issue of the be-future complement's

aspectual constraint is relevant to North and South Slavic languages alike.

3.2.5 Considerations of language contact

When arguing that Slavic borrowed or calqued the be-future construction from

German, Rosier assumes that a language contact situation between (West) Slavic and

German existed such that a borrowing could feasibly take place. There is clear evidence that German exercised some level of influence on Slavic, especially on Czech and Polish.

A more detailed analysis of the nature of this contact situation, however, reveals that the relationships between the languages were more complex than is immediately apparent.

96 Thus, it becomes more difficult to argue that the appearance in Slavic of a syntactic construction such as the be-future could be the result of borrowing.

As reinforcement for his idea that the Czech be-future was borrowed from

German. Rosier (1952: 142, fn. 58) cites evidence of the prestige held by the German language over Czech in Bohemia, This evidence includes the borrowing of numerous loanwords from German into Czech and the renaming of Bohemian properties with

German names. Leiss (1985: 259-61), however, presents another view. She argues that the situation in Bohemia is best characterized not by Rosier's description of German prestige and Czech imitation, but by mixed ethnic groups living intermingled. Indeed, one can imagine that Czech must have enjoyed a prestige status at least in the fourteenth century, since the Holy Roman Emperor, Charles IV, was himself Czech and actively promoted the Czech language (cf. Auty 1980: 165—ff.). In these circumstances, it is highly likely that linguistic influence flowed in both directions. Moreover, Leiss points out that the lack of evidence for such mutual influence, e.g. evidence of Czech influence over the local dialect of German, is not unexpected; such local-level linguistic interaction would hardly be reflected in literary German of the time (ibid.: 262).

Leiss supports her characterization of the German-Czech contact situation by presenting evidence that the werden-iuXuxQ first appears in the spoken language of eastem

Middle German (ibid.: 265-ff ), the area that she describes as populated with a mix of

German and Czech peoples. This fact, coupled with the fact that the spread of the werdenAuiuxQ in German travels from east to west (ibid.: 265-66), suggests to her that the be-future actually originated in Czech and was in fact borrowed into German.

97 The goal of the present study is not to determine whether the German werden- future was borrowed from Czech, and thus a detailed analysis of Leiss’ arguments on this point will not be presented. However. Leiss’ description of the Czech-German linguistic contact situation, combined with the evidence presented in other parts of 3.2. adds further weight to counterarguments to the extemal-source hypothesis.

3.2.6 Conclusions

The data in Chapter 2 reveal that the be-future construction in Slavic is in fact a type of construction that has a multitude of different expressions. In fact, in many ways the auxiliary *bqdq is the only element of the construction that can be considered universal: differences abound in the use and distribution of the construction in the various languages, and there is variation not only in synchronic factors like type of complement and treatment of asp>ect, but also in diachronic issues like the date of appearance and the presence or absence of other future-tense constructions at earlier periods in the history of the languages.

The extemal-source theories that have been proposed do not satisfactorily explain the diversity of the be-future constmctions in Slavic, but ratlier tend to view it as a single, monolithic constmction with the same type of manifestation across all Slavic languages.

A more detailed analysis of the data leads to a greater understanding of the nature of the

Slavic be-future, and shortcomings of the theories discussed above are revealed. Issues of chronology, the constraints of Slavic aspect, and general linguistic knowledge about

98 the typicalities of future-tense constructions and their origins all contribute to weaken the plausibility of these theories.

Extemal-source theories cannot describe the evolution of the be-fiiture satis­ factorily; it thus becomes necessary to seek a source or sources for the be-future that are native to Slavic. A theory that proposes an internal source for the be-future avoids, or explains, many of the issues that extemal-source theories cannot. Many scholars have analyzed early data and the reconstmcted protolanguage for linguistic material that would prove a likely source. The following section presents the description and evaluation of the most important of the intemal-source theories that have been proposed.

3.3 Internal-source theories

3.3.1 The role of the impersonal obligative construction

As in OCS (see 2.2.1 above), Rus’ian attests examples of an impersonal obligative construction using a 3 sg. copula plus infinitive and a dative object. The copula could appear in any tense, including the future, and the infinitive could be of either aspect.^ Thus, with an imperfective infinitive, the future-tense obligative constmction appears superficially similar to the be-future, e.g. this example from a Muscovite immunity charter of 1546(1):

( 1 ) A komu hudet” cego iskati na Grisé Hi na ego prikazsciké, ino ix” suziijaz” knjaz’ velikijHimojbojarin” wedenoj. ‘And if anyone has to seek any [legal redress] from Grisha or his agent, 1, the grand prince, will judge them [i.e., the parties in the suit], or my boyar-delegate [will]’ (cit. in Borkovskij 1968: 157).

■*0 For additional description of the construction and examples, see Borkovskij (1968: 152-58).

99 This similarity has led some scholars to propose that the modem Russian be-

future developed out of this impersonal construction. For example. Vaillant (1966: 108)

argues that it is a small jump for bqden> vidéti ‘will have to see’ to develop intobqdq

vidéti ‘I will see’ (although he does not specifically describe the path for such a

development). The same argument is found in Groschel (1972: 233) and Vlasto (1988:

165).

H. Bimbaum (1958: 270) argues against this hypothesis by emphasizing the

modal meaning of this impersonal obligative. In his view, it was the temporal, not modal,

meaning of * bqdq that conditioned its adoption in Slavic as a future auxiliary. However,

his conclusion is faulty; far from precluding a meaning-shift into a future auxiliary, some

types of modal verbs are frequently grammaticaUzed into future auxiliaries (Bybee et al.

1989; 22-25). Moreover, there are reasons why it should be assumed that *bqdq retained

its lexical meaning in order to develop into a future auxiliary; this will be discussed in

3.3.4 below.

Bimbaum’s arguments notwithstanding, several factors cast doubt on Vaillant s proposal. First, the obligative constmction could be formed with infinitives of either aspect; e.g. this example with perfective verbs (2) from a 1350-51 treaty by the

Muscovite Grand Prince Semen Ivanovic.

(2) A gdé mi budet v”sësti na kon’, vsësti vy so mnoju. ‘And wherever I am to ride [lit. “get on a horse”], you are to ride with me’ (DDG 1950/1970: 13, no. 2).

1 0 0 Like the hypothesis that posits the borrowing of Germanwerden plus the infinitive into

Slavic (see 3.2.4 above). Vaillant s proposal supplies no motivation to explain why the be-future developed only out of the obligative construction with imperfective infinitives.

Vaillant's proposal also appears unlikely given the structure of the impersonal obligative construction. First, the phrase is biclausal; the future copula fills the verb slot, and the infinitive is part of a dependent clause. This view of the structure is corroborated by the fact that when the impersonal obligative construction is used in the present tense, the infinitive remains while the present-tense copula {esm’ in Rus'ian, null in modem

Russian) occupies the verb slot. In contrast to this, in the be-future construction*bqdq is an auxiliary and occupies the verb slot only in combination with the infinitive.

In her work on the Russian be-future, Grenoble (1995: 189-91 ) demonstrates con­ vincingly both the biclausal structure of the impersonal obligative and the monoclausal structure of the be-future. Using the acceptability of negation and double negation as a diagnostic, she shows that only impersonal obligatives allow negation of the infinitive and negation of both the auxiliary and the infinitive, while the be-future allows only single negation of the auxiliary. For example, compare (3a) (also (2b) in 1.2.2 above) and (3b):

(3a) *Ja budii ne kurit’ I will not smoke’ (ibid.: 189)

(3b) Nekotorym budet sovsem ne vviti iz tramvaja Some will be completely obliged to not exit off the tram' (ibid.: 191)

The different treatment of negation is a clear indication that the be-future and the impersonal obligative are different types of construction.

1 0 1 Moreover, this obligative construction is syntactically related to other impersonal obligative constructions with a dative object plus infinitive, such as those with nado(bé) and nuzno. and is still found in modem Russian (see (3b) above as well as 2.12.1). In general, these impersonal constructions have remained quite stable in Slavic; there is no independent evidence of a dative object acquiring nominative (i.e., grammatical-subject) case marking in Slavic. There is thus a lack of evidence that Vaillant s proposed change was likely to have taken place.

3.3.2 The role of the future perfect

As was mentioned in 2.2.1 above. Common Slavic had a future perfect construction that was formed with*bqdq plus the /-participle. The ostensible similarity between the Common Slavic future perfect construction and the later future constructions with the same auxiliary in several Slavic languages has led some scholars to propose a direct link between the constructions. In his discussion of the Russian future, for example, Dumovo (1962: 325) argues that the Russian infinitival be-future developed from the future perfect, presumably with an intermediate stage where both infinitival and participial complements could combine with biidu. He sees the Polish situation as representative of this intermediate stage, citing the similarity between the Polish future with bqd^ plus the /-participle and the Common Slavic future perfect. The same conclusion is presented by other scholars of Russian, including Potebnja (1899/1958:

290), and is an assumption also made by some historical grammars of Ukrainian as an

1 0 2 explanation of the origins of the participial be-future found in dialects (cf. Bezpai'ko et al.

1957: 344; Bevzenko et al. 1978: 293—94).^'

A more detailed theory which subscribes to this view is found in Lomtev (1952).

In his analysis, he argues that there are two possible sources of the Russian be-future.

The first possible source is the future perfect construction, with the be-future developing via a replacement of the /-participle by an infinitive. The second possible source is the future constructions which were in use in Old Russian with ‘have/take’, want', or

begin-type auxiliaries plus the infinitive. The required change would be a replacement of these auxiliaries by the auxiliary budu (ibid.: 251—52).

According to Lomtev (ibid.), only the first proposed path of development produces, or can produce, the proper outcome for modem Russian. In his analysis, the second proposed path is not possible. His argument states that, whereas the earlier auxiliaries could combine with infinitives of either aspect, the be-auxiliary is attested with only imperfective infinitives. If the second path were the correct one, an intermediate stage is predicted where the be-auxiliary is attested with infinitives of both aspects. Since this intermediate stage is not found, Lomtev argues that the source for the modem

Russian be-future must be the future perfect construction.

Lomtev s hypothesis has several shortcomings.^^ For example, Lomtev rejects the second possible path of development for the be-future on the basis that it fails to account for the modem constmction s restriction to imp>erfective infinitives, yet this is

Interestingly, scholars concerned with Polish do not subscribe to this view; this is discussed in 3.3.2.1 below. ■*- These arguments were first presented in Whaley ( 1999: 166-67).

103 also a problem for the path of development that he prefers. In contrast to modem North

Slavic be-futures, the Common Slavic future perfect could be formed with verbs of either aspect. In order to explain the presence of the aspectual constraint on the modem be- future. Lomtev adds a stipulation to his hypothesis that only imperfective /-participles were replaced by infinitives to form the modem be-future (ibid.: 252). He presents no evidence, however, to justify this position, save the outcome that can be observed. This a priori reasoning is not convincing.

Moreover, when describing the behavior of Old Russian future constructions that existed prior to the be-future, Lomtev fails to consider the fact that phasal verbs, including the future auxiliary based on the verb ‘begin’, combine with perfective infinitives only in extremely limited contexts and are semantically incompatible with colligated perfectivity.'*^

The proposed path of development that Lomtev rejects on asptectual grounds, thatbudu replaced other future auxiliaries and combined with an infinitive, is not subject to his counterargument if one argues thatbudu replaced a ‘begin'-type auxiliary. Lomtev argues that the lack of any intermediate stage where budu combined with perfective infinitives indicates that this path of development is unfeasible. However, if budu can be seen as patteming with phase verbs with regard to colligabUity, such an intermediate stage would not be expected or required.^

GRJa (1960; II, 215-16), Kuznecov (1959: 235), and H. Bimbaum (1958: 192, fn. 93) all report that phase verbs combine with imperfective verbs only. There appears to be one rare context where dati ‘give’, a perfective, appears in combination with phase verbs under negation; this usage remains unexplored. However, the restriction on the colligability of phase verbs holds true for the vast majority of cases. A similar criticism is presented by Kfi'zkovd (1960: 172).

104 From a more general perspective. Lomtev has not presented a mechanism to explain the path of his proposed development. By arguing that the /-participle of the future perfect was replaced by the infinitive to form a be-future, Lomtev merely describes what appears to have happened. He does not identify a factor that would have motivated the kind of analogous change he is proposing, and thus he does not explain how or why such a replacement of the complement would take place.

A careful analysis of the evidence demonstrates that, for most languages, it is in fact unlikely that the modem infinitival be-future is directly descended from the Common

Slavic future perfect. Of particular importance to this question are the textual evidence and the semantics of the future perfect construction, including the latter’s role in determining the aspect of the complement. These issues will be discussed in the subsections below.

3.32.1 Textual evidence

Evidence from early Slavic texts provides evidence that the infinitival be-future cannot be linked to the Common Slavic future perfect. Early texts which lend insight into the use of future perfects can be found in Old Polish and Rus'ian.

The construction of plus the /-participle (with imperfective verbs) is rarely attested in the earliest Polish texts. This observation is made by Gorecka and Smiech

(1972: 13), who also argue that if the participial be-future in Polish were really a continuation of the Common Slavic future perfect, then one would expect more attestations ofb^d^ in combination with the /-participle to appear in the earliest Polish

105 texts. However, the use of future perfects was typical of only a limited number of contexts, and thus the rarity of participial forms in early texts does not provide any insight into the origins of the participial be-future.

The distribution of participial and infinitival futures in the earliest Polish texts heavily favors the latter, although Stieber (1954: 233) notes that participial forms, when attested, are found more frequently in secular texts. This could indicate that such forms were actually future perfects rather than futures; secular legal or administrative texts are more likely to describe situations that require future-perfect tense marking. However,

Stieber (ibid.: 234) also points out that written secular texts tend to reflect contemporaneous spoken language more closely than do religious texts. In this case, the significantly higher frequency of the participial future in secular texts might suggest that the participial construction that conveyed pure futurity was innovative and had not yet found its way into the more conservative language of religious texts. Such conclusions are, however, speculative.

In order to determine whether the Old Polish constructions with plus the /- participle are future perfects or innovative futures, the semantics of the earliest examples must be analyzed more carefully. One such example (1), also discussed as (1) in 2.8.2 above, can be interpreted with either future or future perfect meaning.

( 1 ) Gonic b0d0 nieprzyjaciele moje, ipoiapi0 je: ani sie obrocp, alii zgin0. Ziami0 je, ni b 0 d 0 moeii star: padn0 pod nogi moje. ‘I will pursue my enemies and overtake them; I will not turn back until they are consumed. I will break them so that thev will not be able to stand:

106 They will fail beneath my feet* (Ps. 17.41^2).

In the same context, the Latin source has a future perfect form, although the semantics of the Latin future perfect in this period are in question. Plater and White (1926: 105) indicate that future perfects in the Vulgate convey a meaning not far removed from the construction's original semantics, but Fleischman (1982: 34) indicates that this form could convey pure future (see also below).

In general, in the literature there is a lack of analysis regarding combinations of

plus the /-participle in Old Polish. On one hand, we caimot assume that combinations of plus perfective /-participles did not exist; it might be that scholars simply have not mentioned them in their discussions of the Polish be-future s development because such forms do not resemble the modem construction. On the other hand, it might be the case that many or all of the earliest attestations of plus the /- participle are in fact future perfects, but have been misinterpreted as pure futures by scholars who assume that the semantics of the modem construction can be applied to the

Old Polish. Without more data, such ambiguities caimot be resolved; the collection of such data is beyond the scope of this study.

Kfizkovd (1960: 106) supports the idea that the participial be-fiiture in Polish was innovative. Using Biblical texts from different centuries, she demonstrates that early texts often utilize infinitival complements where later texts have participial forms in the same passages. This evidence might suggest that the infinitival be-future is the older form, or it

107 might simply illustrate that the usage of the participial form became less restricted, or more normative, over time.

Perhaps more supportive of the argument that the participial construction in Polish is innovative is the apparent pattern of the spread of the construction from primarily masculine singular subjects into other singular and plural subjects. This type of hierarchichal pattern, from the least-marked gender to the most-marked, would not be expected if the future perfect were the source for the construction, because the future perfect was certainly capable of combining with subjects of any gender.

In sum, the Polish data indicate that combinations of plus either the /- participle or infinitive are both found in the earliest extant texts. Without data from earlier texts, it is not possible to determine whether the sparse attestation of the participial construction is the result of its being a recently innovated construction or simply a form with restricted usage.

In Old Russian, and in fact throughout Rus'ian. the future p>erfect is attested fairly often, especially in legal texts where it appears frequently in conditional clauses

(Kuznecov 1959: 248; see also Karskij 1956: 288-89). There are also several examples from Novgorodian birchbark texts (Zaliznjak 1995: 159); Zaliznjak (ibid.) argues that one of the examples conveys pure future meaning, but a careful semantic analysis suggests that this is not necessarily the case (cf. the discussion of (1) in 2.12.2 above, also (4) below). The construction falls out of use in texts by the end of the sixteenth century

(Kuznecov 1959:251).

108 Since the be-future is attested consistently in Russian only in the latter part of the seventeenth century. Nikiforov (1952: 182-83) argues that this is further evidence that there is no direct connection between the future perfect and the be-future in Russian.

However, his conclusion relies too heavily on textual evidence, and assumes that one can establish a firm terminus post quern for the be-future that postdates the disappearance of the future perfect. Such a date cannot be established, given the earlier attestations of be- futures that are found as early as the fifteenth and sixteenth centuries (see 2.12.2 above).

The problems encountered with textual evidence in Polish and Russian indicate that such data alone cannot tell us if the Slavic be-future developed from the future perfect. However, other evidence, such as the semantics of the constructions themselves, lends more insight into the matter.

3.3.22 The semantics of the future perfect

It is not impiossible for future perfect constructions to develop into constructions that indicate pure futurity; such a change took place in later Latin/early Romance dialects, for example (cf. Fleischman 1982: 34). However, the semantics of the Slavic future perfect construction is also amenable to re interpretation into past-tense or irrealis meaning.

Kuznecov (1959: 252), for example, argues that in Old Russian as early as the fourteenth century, the so-called future perfect construction often expressed past, rather than future, meaning. Thus, a shift into pure future-tense meaning is not the only possible outcome for the expression.

109 Indeed, a brief analysis of instances of the future perfect from early Slavic texts

reveals that many examples of so-called future perfects do not express future-tense

reference per se; they express, rather, a displaced perception of state, the realization of

which occurs at a moment posterior to the moment of speech. The event for which the realization is relevant, however, most often takes place anterior to the moment of speech.

What is meant by displaced perception of state can be illustrated by the following examples.

The first example (2) is taken from Codex Suprasliensis, a reading menaion in

OCS. The context is that Mary has recently been informed by an angel that she will give birth to Jesus, and is unsure whether to tell Joseph.

(2) I Marija v’ sebë si razmysljase povédë ii se losifu Hi pace shkryjq tainoje se jeda bodefb s'bl'beal'bprixodivxi ‘And Mary pondered to herself, “Shall 1 tell this to Joseph, or hide this secret, in case the one who had come [i.e., the angel] will have lied " (Suprasliensis fol. 239,11. 23-25)

The future tense marked by bqdeti> can only refer to the p>osterior moment where Mary would realize that the angel has lied; in other words, the speech-act of lying is actually prior to the moment of Mary’s represented thought, whereas her realization of that lie will take place in the future. Moreover, the entire clause is entirely within the realm of irrealis

(marked by jeda in case’), since the statement about the angel lying is merely supposition from Mary’s perspective; we know, after all, that within the context of the story the angel did not in fact lie. The Greek original for this text uses an aorist indicative in the same context.

1 1 0 Another example (3) comes from a Novgorodian birchbark document. In this gramota, the author Anna is asking her brother to help convince a third party that she did not give surety for her son-in-law at an earlier time. She tells her brother what to say

(referring to herself in the first person):

(3) oze budu Ijiidi pri komo budu dala ruku za zjate to te ja vo vine If there are people before whom I gave surety for [my] son-in- law, then I am at fault’ (Zaliznjak 1995:344, no. 531)

Once again, the future perfect in this example describes a future perception of a past state, rather than an actual future event. The people of which Anna speaks, if they exist, will be discovered in the future, but the act of giving surety of which Anna has been accused is already in the past. As in (2), the entire scenario is located within irrealis, since Anna is in fact claiming that she did not give any surety.

Another example (4), also discussed as (I) in 2.12.2 above, is also from a birchbark gramota.

(4) postal” jesm’ ... .k. bél" k tobé a ty Nestere pro cicjak” prisii ko mni gramotii s kim” budes” postal” ‘I sent twenty squirrel pelts ... to you, and you, Nester, send me a note about the helmet [telling me] with whom you will send / will have sent [it]’ (Zaliznjak 1995:455, no. 358).

Zaliznjak (1995: 159) claims that the underlined construction unequivocally conveys future-tense reference. In fact, the tense reference is ambiguous, since we do not know whether the helmet has already been sent at the time of the note’s writing; the lack of context makes the issue one of principled uncertainty.

Ill The above examples and discussion are certainly not adequate if the semantics of the future perfect is to be fully understood; a more thorough analysis is needed, which is beyond the scope of this study. Nevertheless, even these few examples show how the semantics of future perfects overlaps with both past-tense and irrealis meaning. The only tense reference implied by such constructions is that which places the state described and the perception of that state in a chronological order. The actual time-frame of the state, with regard to the moment of speech, is not indicated; for this reason, these states can be in the past or even exist only as irreal potentialities. Thus, future-perfect constructions often convey future-tense reference only as a secondary meaning, and there appears to be significant evidence that many uses of the future perfect at an older stage of Slavic are altogether incompatible with pure future-tense reference. This conclusion weakens the idea that future perfects typically develop only into future-tense constructions.

Analysis of the semantics of the future perfect also reveals that future-perfect meaning is more compatible with perfectivity than imperfectivity. The /-participle was originally resultative; in other words, the /-participle focused on a past, present, or future state resulting from an earlier action. As such, the meaning of the participle was more compatible with perfective verbs, which emphasize the result of an action or view the action as a totality. The affinity between future perfects and perfectivity is demonstrated by the OCS evidence, where out of seven examples of the future perfect, only one /- participle is imperfective (cf. Vyskocil 1956). A review of the examples given by

Kuznecov (1959: 152-55) and statistical data provided by Kfizkova (1960: 158) also show a greater proportion of perfectives than imperfectives.

1 1 2 The strong tendency for future-perfect constructions to be formed with perfective

/-participles is in sharp contrast to the modem North Slavic be-future constructions, including the participial be-future found in modem Polish, Kashubian. Slovak, and

Ukrainian. These constructions combine with imperfective participles only.^^ Previous scholars, such as Lomtev (1952), have overlooked this disparity between the future perfect and later be-futures.^

By not accounting for the aspect of the be-future complement, these scholars make the same mistake that proponents of extemal-source hypotheses do (see 3.2.4 above). To propose that the future perfect developed into the be-future, one must find a way to justify the latter's restriction in coiligabilit>' to imperfective complements where the source construction did not have one. Without an explanation for this restriction, the hypothesis cannot be validated.

It appears that the data for Serbo-Croatian and Slovene indicate that the issue of the aspectual constraint does not apply to South Slavic; we have seen in Chapter 2 that the be-future in these languages can be formed with verbs of either aspect. Without the problem of aspectual restrictions, the future perfect construction app>ears to be a more plausible source for the Serbo-Croatian and Slovene be-futures and futiira exacta formed with the auxiliary in question.

Nevertheless, for Serbo-Croatian, at least, the historical evidence suggests that the colligability of budu/budem with infinitival complements might have been restricted

■*5 The situations in modem Slovene and Serbo-Croatian differ, they are discussed below. A rare exception is Kftzkovd (1960: 105-106), although her ultimate conclusion is unsatisfactory (sec 3.3.3 below).

113 originally to imperfectives. For example, the earliest Croatian examples ofbudu plus the infinitive are with imperfective infinitives, and the earliest evidence of a be-future with perfective infinitives is attested only in the fifteenth-sixteenth centuries (see 2.4.2 above).

It should be noted, however, that early textual evidence is scarce.

As for the Serbo-Croatian construction ofbudu/budem plus the /-participle, it appears reasonable to propose that the participial be-future is indeed a continuation of the

Common Slavic future perfect. The construction can be formed with /-participles of either aspect, conveys futiirum exactum meaning in the modem literary language and in many modem dialects, and historical evidence documents the existence of similar colligability in earlier stages of the language (see 2.4.2 above). Thus, while one might propose on good grounds that the infinitival constmction is innovative, the participial construction does not necessarily demand such an argument.

However, one must account for the tendency in the modem literary language for the futurum exactum to be formed with imperfective verbs. This is entirely opposite the situation in OCS and early Slavic in general, where the semantics of the future perfect seems to have favored the use of perfective /-participles. If the modem construction of budu/budem plus the /-participle evolved out of the Common Slavic future perfect, it is unclear how to explain the modem preference for imperfective verbs.

A possible solution to this problem is to propose that an intermediate stage existed where /-participles of both aspects could combine with budem in dependent clauses, and that perfective futura exacta were later replaced by nonpast perfectives in some contexts.

After all, grammars and descriptions of Serbo-Croatian characterize the use ofbudem

114 plus imperfective /-participles as a tendency, not a prescribed norm. If this intermediate stage existed, then one could argue that the participial be-future in Serbo-Croatian followed a path similar to the construction in Slovene (see below). However, evidence to support this hypothesis must be found.

In sum, the Serbo-Croatian evidence does not refute the arguments against the future perfect as a source for the modem infinitival be-fiiture that have been proposed in this study. However, more detailed textual evidence is needed to clarify the situation as it applies to Serbo-Croatian specifically.

The be-future in Slovene is more easily reconciled with an argument like that proposed by Lomtev (1952) for Russian. No infinitival complements are ever attested with the verb bo(de)m, and so it is not necessary to propose that the Slovene be-future ever underwent some kind of replacement of an infinitival complement by the /-participle.

Moreover, there is no evidence that the be-future was restricted to imperfective /- participles at any stage of the language. Thus, there is no reason to assume an aspectual constraint ever existed for the construction in Slovene, and the Common Slavic future perfect is a plausible source for the construction.

Evidence that supports this argument can be found in the grammar by Bohoric

(1584/1970). In his work, he describes the construction bod- plus the /-participle not only as an indicative future, but also a “conjunctive” future. In other words, the construction was used both in main clauses, expressing pure futurity, and in dependent clauses like conditionals, expressing relative futurity. This usage might indicate that the future perfect inherited from Common Slavic did generalize in Slovene to become a pure

115 future. The examples in Bohoric could be considered representative of a transitional period in the development of the be-future.

The Kajkavian dialects of Serbo-Croatian share features with both Serbo-Croatian and Slovene. In them, as in Slovene, there is no evidence of an infinitival be-future ever having existed, nor is there any historical evidence of a constraint on the aspect of the participial complement. As such, one may assume a development of the be-future along the same path as Slovene.

S.3.2.3 Conclusions

Based on the discussion in this section, one concludes that the future perfect is an unlikely source for the modem infinitival be-future, at least in North Slavic. The chief stumbhng-block for many of the arguments presented in the sections above is the aspectual constraint on the be future. Most scholars cannot or do not explain its presence, or else their theories fail to consider its significance. In fact, the aspectual constraint on the be-future is an important factor, which must be understood in order for the path of development for the construction to become clear. This idea is explored more fully in

Chapter 4.

It is also clear that the development of infinitival and participial be-future must be considered separately. Despite scholars’ attempts to link the constructions to a single source, their arguments are difficult to justify in light of a more careful analysis of the data and a more sophisticated understanding of language change. For example, one cannot

116 explain that the infinitival be-future was formed by the replacement of the /-participle by the infinitive without identifying a plausible motivation for such a change.

Another issue that this section brings to light is the fact that future-perfect constructions have more than one possible path of development. Besides the more familiar path of development into a future-tense construction, as in Latin, the evidence from OCS and Old Russian presented above indicates that irrealis is also a possible outcome for these forms. The semantics of the future perfect is complex, so it is not surprising that multiple paths of grammaticalization lead out from its source.

This conclusion alone certainly does not rule out the possibility that the Slavic future perfect developed into a be-future; indeed, the data for Slovene and Serbo-Croatian suggest that this hypothesis is plausible. However, for the other Slavic languages, the aspectual constraint on the be-future complicates the issue considerably.

3.3.3 The role of phase verbs

As was mentioned in 3.3.2 above, future-tense constructions with auxiliaries derived from inceptive phase verbs were formed with only imperfective verbs. This constraint on the colligability of phase verbs is not considered by scholars such as

Lomtev (1952) in their description of the development of the be-future, despite the fact that, in many of the languages that employ it, the be-future has the same constraint.

Kfizkova (1960) was the first to point out that one can see an analogous relationship between phase verb constructions and the be-future. She uses the identical colligability patterns of the two constructions to argue against Rosier"s extemal-source

117 hypothesis, pointing out that the aspectual constraint on the Slavic be-future has no analogue in the German future with werden. With regard to Lomtev s assertion that earlier future-tense constructions in Russian combined with infinitives of either aspect, she points out that the "begin-type future was constrained for aspect (ibid.: 172).'*^

Given the similar colligability of phase verbs and the modem (North Slavic) be- future, Kfizkova argues that the following sequence of events led to the genesis of the be-future. First, the present tense forms of "begin-type verbs came to be used as future auxiliaries. Such verbs retained their lexical meaning in other tenses and in some nonpast contexts, however, and thus Kfizkova argues that this made them unsuitable for use as purely grammatical markers. She claims that the be-auxiliary then came to be used because it made a more suitable future auxiliary, already free of other nuances (ibid.: 101).

Kfizkova's insight in finding a connection between phase verbs and the be-auxiliary is essential to formulating a satisfactory theory, and is an important component in the hypothesis presented in the following chapter of this study. However, her argument presumes that *bqdq was free of lexical nuances by the time it came to be used as a future auxiliary. As is discussed in detail in 3.3.4 below, textual evidence contradicts this presumption.

There are other problems with Kfizkova"s hypothesis. In light of more recent scholarship regarding the development of future markers, one particular error in her reasoning stands out: her argument regarding the suitability of the future auxiliaries used in Slavic. According to Kfizkova, *bqdq came to be the regular future auxiliary because

The current study presents similar arguments: see 3.2.4 and 3.3.2, respectively.

118 it was a more suitable auxiliary in that it carried no lexical nuances. Knzkova's error on this point is one that is shared by many scholars who favor an intemal-source hypothesis for the be-future. In the following section, this idea is discussed in more detail.

3.3.4 The role of lexical nuances

For Kfî^ovâ, as well as for many other scholars, the relevant question regarding the be-future is not only how the construction developed, but also why it ultimately replaced other analytic future constructions. This issue has been of special interest to scholars of Russian, a language in which many different future auxiliaries are attested in premodem texts, and in which the be-future emerges as the norm in the language relatively late. From this perspective, many scholars have attempted to identify a motivation for the shift from other possible future constructions to the be-future.

In order to explain the replacement by the be-auxiliary of other future auxiliaries, scholars have attempted to isolate the factors which cause the former to appear more suitable than the latter. One idea posed quite frequently in the scholarship is that, in contrast to other future auxiliaries, the be-auxiliary conveyed no lexical nuances. It is clear, for example, that many future auxiliaries found in languages are in a state of polysemy with their fully lexicalized counterparts. For example, constructions withstat’

‘become' in modem Russian can convey pure future meaning in some contexts, but lexical meaning in many others (SSRJa 1963: s.v. "“stat’"). However, many scholars have assumed that the be-auxiliary had no such lexicalized counterpart but rather

119 expressed only pure futurity, and thus it was favored by speakers over auxiliaries that could be used in other contexts.

For example, Knzkova (1957; 1960: 179) and Lomtev (1952: 252) express this view regarding Russian budu. Also, Kiparsky (1967: 235) says that Russian budu was

“completely grammaticalized” and thus more suitable as a future auxiliary, and Nikiforov

(1952: 179) says that unlike other auxiliaries, budu had “no modal or aspectual nuances”.

Both Cemyx (1957: 229) and Kuznecov (1959: 254) identify budu as the “most abstract” future auxiliary found in Old Russian. Wytrzens (1953: 27) argues that one can conclude that it had already lost its “lexical function”, based on the fact that budu was found in Old

Russian in impersonal constructions and the future perfect.

This reasoning is erroneous. From the most general perspective, it is marred by the underlying notion that the innovation of a be-future construction arose out of a replacement by *bqdq of other future auxiliaries such as those based on ‘have/take' or

begin'. In reality, however, these constructions no doubt coexisted, and their functions most likely overlapped to some extent. As is clear from the evidence presented in Chapter

2, even in the modem Slavic languages there are often multiple futiu’e-tense constructions in use contemporaneously. Moreover, from a broader point of view. Bybee et al. (1994:

243) point out that such a situation is quite frequent among the world’s languages, with different future-tense strategies evolving along different paths of development. It may be the case that diachrony reveals one construction to be in decline while another is

1 2 0 increasing in use, but the idea of one auxiliary “replacing” another is often an overly simplistic description of the changes taking place/*

More specifically, the idea that *bqdq prevailed as future-tense auxiliary because of its lack of nuances is flawed because it is based on the a priori assumption that because *bqdq is the auxiliary that is found most often in the modem Slavic languages, it must have been the best suited for the task from the outset. This implies that *bqdq existed at some abstract level as a future-tense auxiliary prior to its actual use in that function. This viewpoint also assumes that previously used auxiliaries, like stanii, or those based on ‘have/take’ or ‘begin’, were perceived as flawed by speakers at the time, whereas *bqdq was not—the flaw being their polysemy; they had retained lexical meaning in non-future contexts, and were not fully grammaticalized future markers.

In fact, general linguistic scholarship on syntactic change and grammaticalization demonstrates that it is precisely the lexical content of a verb that determines the possibility of its development into a grammatical marker. As Bybee et al. ( 1994: 255-ff.) show, the lexical (i.e., semantic) content of certain verbs contains features which imply a temporal orientation towards the future via nuances of intention or prediction. It is the implicature of futurity in such verbs that can then be foregrounded in the semantics of the verbs, and thus enables them to be grammaticalized into future auxiliaries.

This characterization of grammaticalization has certain implications which further weaken the argument that the be-future developed because *bqdq was the most suitable auxiliary. First, it implies that grammaticalization does not, and in fact cannot, occur

Cf. the discussion of "layering" in Hopper and Traugolt (1993: 124—ff.).

121 without a semantically motivated impetus to do so. This view invalidates the notion that

'have/take' or ‘begin'-futures in Old Russian did not continue as future auxiliaries in

Modem Russian because their continuing use as fully lexical verbs precluded it. Indeed, these verbs existed in a state of polysemy for centuries, serving to some degree as both future auxiliaries and lexical items, and similar situations are attested in many of the world's languages; e.g. the English construction ‘be going to’. Such examples are further evidence that arguments such as Knzkova’s are incorrect.

An example of this can be seen in the Old Russian Ulozenie of 1649, in which iicmi begin’ plus the infinitive functions as the primary expression of imperfective futurity in the text (Cemyx 1953: 347^8). Although iicnu is one of the verbs considered ultimately unsuitable for grammaticalization because it could still convey lexical meaning, its lexical nuances clearly did not interfere with its use as a future auxiliary.

Another implication of the grammaticalization perspective is thatbqdq * must have retained a lexical nuance compatible with future implicature up to and beyond the point at which it came to be used in be-future constructions. In other words, a major issue to be understood with regard to the development of the be-future is that it in fact retained lexical meaning, despite its evident use as a future copula and as a component of other constructions. This issue is especially important for Russian, where the be-future is attested only at a relatively late date.

There is a great deal of evidence that *‘bqdq retained lexical nuances well into the period during which the be-future was evolving. As was shown in section 2.1 above, the verb *byti was polysemous in the oldest stages of Slavic, with a imperfective stem

122 meaning be' and a perfective stem meaning ‘become'. This meaning can be shown to extend to the individual Slavic languages; the best explication of this can be found in the rich textual evidence of Rus'ian.

According to both Sreznevskij (1893—1912: s.v. “Z?yr/”) and the SRJa (1975: s.v.

^'byiF'), forms ofbyti continued to express change-of-state meaning in some contexts in

Rus'ian, with examples attested as late as the seventeenth century. This change-of-state meaning can be seen in some uses of the nonpastbudu, such as in the following excerpt

(I) from Xozdenie Zosimy k” Raxmanam", a seventeenth-century copy of an earlier text translated from Greek.

(I) ...ni est' v" nas” nikogoze. ize poimaet” zenii sebé, doneléze budet” u nego [neju] dvoe cad”, i potom” razlucitasja drug” ot” druga i prebyvajut” oba v” cistotë.... “...nor is there anyone among us who takes a woman to himself, until there will be two children by them (i.e., they will have two children), and then they divorce from each other and both remain in purity...” (Tixonravov 1863: 87).

In this example, the perfectivity of budet” is indisputable, given that perfective verbs are obligatory with dowe/éfe until', a word with an inherent meaning of telicity.

Change-of-state meaning was present not only in nonpast forms, but also in aorist forms such as byx". This is revealed by several examples found by Schooneveld (1959), who rejects the possibility of polysemy for this form but ultimately demonstrates it with copious textual evidence. For example, the following entry (2) from the Old Russian chronicle Povest’ vremmenyx” let” (s.a. 1026) illustrates the change-of-state meaning expressed by the aorist:

123 (2) / raidélista po Dnépr” Riis'skuju zemlju: Jaroslav" prija sjii storomi. a M ’stislav" onu. / nacasta zlti mirno i v bratofjiib'stvé. i lista iisobica i mjatez , / b\st' tisina velika v zemli. "And they divided the Russian land along the Dnepr: Jaroslav took this side and M’stislav the other. And they began to live in peace and brotherly love, and quarrel and upheaval subsided, and there was great peace in the land” (Cited in Schooneveld ibid.: 74; his translation).

Schooneveld explains the use of the perfective aorists such asrazdélista, nacasta, and byst’ in this entry as expressing “actions . . . all reaching the limit of their development"

(ibid.). Nonetheless, a more effective interpretation of the function ofbyst’ in this example is as an expression of a change of state, from the “quarrel and upheaval” that had existed to the peace that reigned from that point on. In addition, this change-of-state meaning is corroborated in the earlier phrase nacasta ziti mirno [they] began to live in peace

Based on the textual evidence, it appears that claims that *bqdq was purely grammatical prior to the development of a be-future are incorrect. Such anachronistic explanations are, moreover, based ona priori reasoning that cannot be justified within a more rigorous conception of language change. Ultimately, they misconstrue the real issues that must be resolved in an analysis of the development of the be-hiture—that of the polysemy of the auxiliary and the presence of the aspectual constraint on the complement. It is not necessarily possible to answer the question ofwhy the be-future developed, replacing other future constmctions; it is only possible to suggesthow the developmentoccurred.

See also the discussion of a similar example by Collins (no date).

124 3.4 Conclusions

Based on several issues, including chronology and aspect as well as larger issues regarding the development of future-tense constructions, it appears that the development of the Slavic be-future cannot be attributed to the influence of a non-Slavic source. Only an intemal-source hypothesis can explain many of the factors involved in the development of the construction.

However, the intemal-source theories that have been proposed in the literature also fall short in their descriptions. They fail to account for the presence of the aspectual constraint on the be-future or the fact that *bqdq must have retained lexical meaning in order for it to become grammaticalized. Moreover, the status of the participial be-future has been poorly understood; based on the arguments presented in 3.3.2 above, one concludes that the constmction is in fact innovative, rather than a continuation of the

Common Slavic future perfect. This idea has been addressed briefly by Knzkova ( 1960), although her study does not state this conclusion outright.

Kfizkova’s study does address the central problem of aspect with regard to the be- future’s development; she was the first to suggest a conceptual link between the colligability of phase verbs and that of the be-future. Moreover, Lomtev (1952) is aware of the importance of explaining the aspectual constraints on the be-future, although his solution does not succeed in doing this. It is clear that an understanding of the aspectual constraint and the relationship between phase verbs and the be-future s auxiliary both prove essential to understanding the development of the be-future. In the following chapter, a new hypothesis is proposed regarding this development that seeks to refine the

125 idea first proposed by Knzkova and to ground it within a broader, more general linguistic framework.

126 CHAPTER 4

THE GRAMMATICALIZATION OF FROM CHANGE-OF-STATE TO FUTURE

4.1 Introduction

It is clear from the discussion presented in the previous chapter that the origins of the be-type future in Slavic have not been satisfactorily investigated. Although many theories have been put forth, each has shortcomings that are revealed by rigorous examination of both the empirical data and the argumentation behind each conclusion.

For example, it has been determined that extemal-source hypotheses do not present a viable description of the development of the be-future. On the other hand, intemal-source hypotheses that have been presented up to now are inadequate as well.

One of the chief problems with previous theories regarding the development of the be-future is that of aspect. For the North Slavic languages, the be-future is formed with only imperfective verbs,^° whereas all of the sources proposed for the construction, e.g. the impersonal obligative, the future perfect, or a non-Slavic source such as the

German future with werden, can be formed with verbs of either aspect or are unmarked with regard to aspectuality. The discussion in the previous chapter demonstrated that the

The situation in Serbian diverges from the rest of North Slavic (see section 2.7.1); it is given special attention in 4.3.1 below.

127 presence of this kind of aspectual constraint is a critical factor in the study of the Slavic be-future.

A more endemic problem with previous scholarship on the be-future is the lack of connection that exists between discussions specific to Slavic and more general linguistic scholarship concerned with language change. This problem surfaced repeatedly in the discussions in Chapter 3. Conversely, it is rare to find Slavic data included in general linguistic discussions of future-tense constructions and their evolution.

In this chapter, a new hypothesis is presented that describes more accurately the mechanisms and changes behind the development of the be-future. This hypothesis is centered around understanding the presence of the aspectual constraint on the North

Slavic be-future, as well as the heterogeneity of the construction’s use and development across all the languages which use the construction. In short, 1 argue here that it is possible for the change-of-state verb *bqdq, inherited from Common Slavic into all Slavic languages, to grammaticalize into a future-tense auxiliary independently and at different times in multiple areas; this argument has been rejected by previous scholars. In order to demonstrate the feasibility of this hypothesis, 1 show how a semantic analysis of*bqdq, combined with a more modem understanding of syntactic change, leads directly to this conclusion. Moreover, this semantic analysis also demonstrates how it is possible for

*bqdq to combine in many languages with only imperfective verbal complements.

To this end, the following sections present a summary of the theoretical framework upon which this hypothesis is based. By proceeding from a solid theoretical

128 background, this study can be more easily incorporated into the larger body of scholarship on syntactic and semantic change.

4.2 Thenature ofsyntactic change

In recent years, much work has been done to formulate a systematic theory of syntactic change. Such a theory is essential to this study, a primary goal of which is to create a description of the development of the Slavic be-future that is consistent with more general principles of change. Such an approach allows not only for the formulation of a hypothesis that is valid within a broad, cross-linguistic context, but also allows scholars interested in the development of futme tenses in other languages to utilize the Slavic data for comparative purposes.

In seeking to define basic mechanisms of change, as well as what kinds of change are likely to happen, recent studies of diachronic syntax have achieved a new level of insight into this area of linguistics. The following sections describe different approaches to the study of syntactic change that are of particular relevance to the current study.

4.2.1 Grammaticalization

Of special interest to this work is the growing body of research on grammatical­ ization, or the process by which lexical items become used as purely grammatical markers. The study of this process has been inspired by the fact that many grammatical morphemes can be analyzed as developing historically from words that originally served a purpose other than conveying grammatical meaning.

129 Future-tense auxiliaries have attracted special attention from scholars interested in grammaticalization. because their lexical origins are often relatively transparent and thus more easily analyzed. Specific and detailed analyses of future-tense constructions have been produced, such as Fleischman’s (1982) study of the future in Romance. Scholars like Bybee and her colleagues (1987, 1989, 1991, 1994) have presented cross-linguistic studies that address the typology of future-tense expressions and their regular paths of grammaticalization.

The grammaticalization of change-of-state verbs into future markers has been discussed to a small extent in the literature. For example. Bybee et al. (1994) describe

‘be(come)’ as one of the more frequently encountered lexical items to undergo such development. Heine (1993: 35) also identifies change-of-state verbs as candidates for grammaticalization into tense markers, although he says that such a path of development is rare.

In discussing examples of change-of-state verbs grammaticalizing into future markers. Bybee et al. (ibid.) list German werden, Danish blive, and Modem Icelandic verda', they do not mention any Slavic evidence. Thus, there is comparative support for the idea that the semantics of *bqdq is amenable to grammaticalization. Their proposed path of development for this type of verb, however, does not appear to be supported by the Slavic situation. For ‘be(come)’ they propose that the semantic catalyst for grammaticalization is a nuance of either obligation or predestination (ibid.: 262-63); cf. the construction in English with ‘be (to)’, e.g.He is to present his paper tomorrow. As I

130 show below, Slavic * bqdq follows a different path of development—a path that has not been mentioned in previous scholarship on grammaticalization.

The term “grammaticalization” has become closely associated with the theory of the same name that argues for the universality of principles like the unidirectionality of change. There is substantial debate, however, about whether the theory is viable, effective, or even necessary in larger efforts to define and understand language change.

Harris and Campbell (1995), for example, argue that the types of language change that grammaticalization theory seeks to describe can be subsumed a larger, more general theory of the mechanisms of language change.^* Thus, a specific theory of grammatical­ ization is considered by Harris and Campbell, among others, to be unnecessary (ibid.:

20).

It is not the intention of this work to take a stand for or against the validity of grammaticalization as a separate and single process adhering to specific principles; rather, this work recognizes that scholars working within grammaticalization theory have made valuable strides in understanding some aspects of syntactic and morphological change, and that their efforts can be applied to the problem at hand. Indeed, since one of the major areas of study of scholars operating within the framework of grammaticalization theory is the origins and evolution of future tense constructions, such works provide an important cross-linguistic perspective on the Slavic data, which have been largely ignored in past scholarship in Slavistics.

51 See also section 4.2.2 below.

131 Moreover, the term “grammaticalization” should by no means be dismissed as superfluous. Even though the linguistic changes that result in grammaticalization may be best described in a larger, more general framework such as that proposed by Harris and

Campbell (1995), there is an advantage of efficiency in employing this term as a shorthand. For example, such a term is quite useful when analyzing a complex series of linguistic changes that result in a typical outcome; one certainly sees this in the recurring pattern of changes that results in the development of a future-tense marker. It is in this spirit that the current study employs the term.^-

4.2.2 Mechanisms of syntactic change

Harris and Campbell (1995) argue that it is possible to show that all kinds of syntactic change are motivated by three basic mechanisms. In this sense, their theoretical standpoint is from an even broader perspective than those studying grammaticalization, who are concemed with what they consider to be a single kind of syntactic change. As we shall see, however, it is possible to mesh these two approaches into a single, unified set of ideas about syntactic change.

Harris and Campbell’s (ibid.: 50—51) three basic mechanisms of change are reanalysis, extension, and borrowing. Reanalysis refers to any change that affects “deep structure”—in other words, changes that alter the grammatical underpinnings of a language’s syntax without changing the overt expression of the language. Extension results in surface changes and does not affect deep stnictme. The third mechanism.

52 A similar viewpoint is expressed by Harris (1997).

132 borrowing, occurs in situations of language contact when a syntactic construction

develops in one language that is based on a model from another. By relating linguistic

changes to these three types, Harris and Campbell argue that it becomes possible to

compare different syntactic changes more effectively and to establish general assumptions

about what changes are possible and most typical.

4.2.3 Syntactic and semantic change

An important factor in understanding syntactic change is the semantics of the

lexemes involved. In her work on semantics, Wierzbicka argues that “. . . grammatical

distinctions are motivated (in the synchronic sense) by semantic distinctions. . .” (1988:

3). To support this statement, her book shows how differences in surface manifestations

of syntax (“grammar” in her terminology) can be seen to reflect differences on a semantic

level.

In order to represent fine-grained distinctions of semantics, Wierzbicka (1988:

9-12) schematizes the semantic content of individual lexical items or morphosyntactic categories using formulae to paraphrase their meaning. These paraphrases are restatements of the semantic composition of a lexeme using a semantic metalanguage consisting of what she believes to be only the most fundamental components of expressions and ideas.

Wierzbicka’s reductive paraphrases and semantic metalanguage provide a lowest- common denominator by which lexical and grammatical items can be systematically compared. Without such a system, she argues, scholars have been able to analyze

133 grammatical distinctions only with intuitions and guesses applied in piecemeal fashion, or

in terms of arbitrary symbolic markers that have no intuitive relation to natural language.

Reductive paraphrases, on the other hand, allow for a systematic analysis of semantics

and its effect on syntax.

Wierzbicka is primarily interested in the synchronic perspective. Her view on the

relation of semantics to syntax, however, can be constructively applied to the diachronic

level as well. It can be argued, in fact, that syntactic changes are initially motivated by

changes in meaning. This is certainly addressed by the theory of Harris and Campbell

(1995); the two intemal-source mechanisms, reanalysis and extension, are mechanisms

that are facilitated by meanings that are ambiguous or otherwise affect the processing of

the language. Grammaticalization scholars also recognize the central importance of

semantics with regard to syntactic change; their theories rest entirely on the principle that

the reinterpretation of compatible lexical meanings leads to the creation of new linguistic markers of grammatical relationships.

Given the centrality of the role of semantics in syntactic change, it is imperative for a study of the Slavic be-fiiture to rigorously analyze the semantics and semantic development of *bqdq ‘be(come)’ in order to accurately describe the path of the verb’s development into a future auxiliary. Often, previous scholars have not undertaken this step because of their misconception that *bqdq has always been, or was long ago grammaticalized into, a marker of pure futurity that lacked other nuances. We have seen, however, that this view is not correct and in fact cannot be correct if one is to explain its

53 See her criticism of various approaches in Wierzbicka (1978) and (1988; 1-20).

134 evolution into a future auxiliary (see 3.3.4 above). Proponents of extemal-source

theories, on the other hand, view the development of change-of-state verbs into future

markers as unexp>ected or otherwise unusual. Typological studies on the

grammaticalization of future-tense auxiliaries, however, suggest that such a development

is not without precedent.

The semantics of Common Slavic*bqdq and its reflexes in individual Slavic

languages has been discussed in 2.1 and 3.3.4 above. A more in-depth analysis of the

semantics of this change-of-state verb is found below as part of the new hypothesis being

presented by this study. By means of this analysis, it becomes possible not only to define

the exact path of grammaticalization for this verb, but also to explain the constraint on the

aspect of the complement that exists in so many of the languages.

4.3 The development of the Slavic be-future

As was discussed in 3.3.3, Knzkova (1960) has taken an important step towards understanding the development of a 'be(come)-type future by pointing out the similarities between the government of the modem be-fiiture and that of phase verbs; the same conclusion has been reached in the current study as well. Relating the be-fiiture to phase verbs provides a means of explaining the innovations in govemment-relationships found in so many languages: the presence of the aspectual constraint, as well as the ability of

*bqdq to combine with an infinitival complement. Up until now, however, the precise path of grammaticalization for*bqdq has not been mapped.

135 In keeping with the idea that grammatical distinctions are motivated by semantic distinctions (see section 4.2.3 above), a satisfactory description of the development of the be-future should demonstrate that the common colligability pattern of the change-of-state verb *bqdq and phase verbs can be traced back to an earlier semantic similarity. Once this similarity has been identified, it becomes possible to explain how *bqdq was able to develop into a future-tense auxiliary capable of combining with infinitives. This goal can be achieved by theoretical means and supported with empirical evidence. Section 4.3.2 will elucidate the theoretical approach, while section 4.3.3 will address the supporting evidence.

4.3.1 Definitions

Before proceeding with a semantic analysis of the verbs that are involved in the current study, a word should be said regarding the terminology that is used to describe them. Scholars studying the semantics of the verb and verbal aspect have employed a diverse array of terms to identify different types of Aktionsarten. Labels such as

“inchoative”, “inceptive”, and “ingressive” are aU used at various times to refer to verbs such as ‘begin’ and become’. This section will discuss the relative merits of the terms that have been used and describe the terms which are employed in the current study.

In analyzing the development of the be-future in Slavic, scholars are compelled to describe and categorize the semantics of the verbs which have served as future-tense auxiliaries. The meaning of some of these verbs is straightforward: xotéti ‘want’ is a desiderative, for example, and its grammaticalization into a future auxiliary, found in East

136 South Slavic languages such as Bulgarian, is comparable to the development of English will.

Other verbs, such as those meaning ‘begin’ and ‘become’, have been given different semantic labels by different scholars. For example, Connie (1976:19) identifies verbs that describe the beginning of an action as “ingressive ”, and uses the same term to refer to verbs meaning become’ (ibid.: 20). In addition, the index to his book contains an entry for “inchoative” that refers to the entry for “ingressive ”. Similar definitions are proposed by Binnick (1991: 154) who states that “ingressive ” verbs signify “the commencement of an action or state”; he also appears to consider the terms “ingressive ” and “inchoative” synonymous.

The theoretical framework of this study requires a fine-grained analysis of the semantics of verbs meaning begin’ and become’, a task which calls for a terminology that is unambiguous and adequately descriptive. For this reason, the terms “ingressive ” and “inchoative” are avoided altogether, their use in the hterature is inconsistent and often overlaps. Instead, the term “ inceptive ” will be used to describe begin’, and “change-of- state” will be used to describe verbs meaning become’.

4.3.2 Change-of-state and inceptive meaning

Using the method of reductive paraphrases described by Wierzbicka (1988) and discussed in 4.2.3 above, it is possible to demonstrate that the semantic distance between change-of-state and inceptive meaning is quite small.^'* Inceptive verb constructions, as

5-* These paraphrases were first presented in Whaley (1998).

137 shown in (1), can be represented by a reductive paraphrase consisting of two

components:

(1) ‘I begin to X’: a. I am not X-ing b. I am X-ing where (b) is temporally subsequent to (a)

The reductive paraphrase for a change-of-state verb like ‘become’ also has two

components, shown in (2), which are quite similar to those for expressions with

inceptives.

(2) ‘Ibecom eX ’: a. lam notX b. I am X where (b) is temporally subsequent to (a)

Both types of expressions inherently involve a change of state, represented by the transition between the stages described by the first and second components of the paraphrase. Moreover, both expressions require the second state to exist at a time subsequent to that time when the first state exists. Indeed, these reductive paraphrases show that inceptive and change-of-state verbs inhabit very similar semantic spheres.

The semantic closeness of these verbs is evident in the parallels that exist between their reductive paraphrases; both verbs describe a change of state, and both verbs have a generalized meaning in that they do not imply any additional information about the second state other than that it began. In this way, the semantic similarity of these verbs offers a means of describing how the Slavic change-of-state verb *bqdq could have been reinterpreted as an inceptive verb, adhering to the colligability constraints that have always existed for phase verbs in Slavic. By positing that this re interpretation took place in the

138 history of the individual Slavic languages, one can explain both innovations of the construction: the newly present ability for reflexes of*bqdq to combine with infinitive complements, and the constraint that such infinitives must be imperfective.

These reductive paraphrases suggest not only a semantic affinity between types, but also how both these types of verb could grammaticalize into a future marker. At the beginning point of such a process, when these verbs convey purely lexical meaning, futurity is implied only by context via the predictive inference inherent in the verbs' semantics. In other words, the verbs convey no explicit reference to the future tense, but futurity is built into their semantic nature as reflected by the fact that the second state (in terms of the paraphrases) must happen after the first. This is certainly reflected in the use of*bqdq in irrealis constructions, such as the uses of the future perfect discussed in 3.3.2 above. Given the passage of time, however, the future implicature could be reanalyzed as the core meaning of the verb; indeed. Bybee et al. (1994) consider future implicature to be the final stage prior to complete grammaticalization of any lexical source into a future marker.

This schematic of the relatedness of the semantics of inceptive and change-of-state verbs provides us with the means to understand the major issues that concern the development of the be-future—issues which have not been resolved by previous scholarship. Most importantly, it can explain the constraint that*bqdq combines with only imperfective infinitives, while other auxiliaries such as ‘want’ and ‘have/take’ are attested with verbs of either aspect. Scholars such as Lomtev (1952) and Kfizkova

( 1960) have not been able to explain this constraint satisfactorily.

139 This approach also lends weight to the argument of grammaticalization scholars that the lexical semantics of individual verbs is an important catalyst to trigger the development of a verb into an auxiliary. As was shown in section 3.3.4, it is erroneous to argue that the be-auxiliary became the primary future auxiliary, rather than one of the other future auxiliaries, due to a preexisting lack of lexical nuances. These reductive paraphrases illustrate that one can identify an inherent semantic quality of both inceptive and change-of-state verbs that serves as the catalyst for a process of grammaticalization.

By describing the semantics of these verbs in terms of reductive paraphrases, we can see confirmation of what grammaticalization theory has predicted:*bqdq is not, in the early period, an empty copula; rather, its change-of-state meaning allowed it to share a common semantic sphere with inceptive verbs, and thus it could have patterned as an inceptive phase verb. Moreover, the paraphrases demonstrate that there is an implicature of futurity present in both inceptive and change-of-state verbs, allowing grammatical­ ization toOCCIU’.

The results of this analysis also reinforce the counterarguments to previous scholarship presented in Chapter 2 above. For example, scholars have attempted to seek a monogenetic origin for this type of future construction in the Indo-European languages in general, and within the smaller framework of Slavic. By demonstrating how an implicature of futurity is inherent to change-of-state verbs, one obviates the need to seek a single point of origin for a change-of-state future. According to this theory, a change-of- state verb in any language could undergo such a development; thus no special impetus or outside motivation for the change is necessarily required.

140 The languages which attest perfective infinitives as complements, Serbian and

Serbo-Croatian, apparently do not adhere to the description of the development of the be-

future as described above. Without a constraint on the colligability of *bqdq, it appears

that one need not semantically link the be-future with phase verbs. However, as was

discussed in 3.2.4 and 3.3.2.2 above, there is evidence that a constraint on the colligability

of the be-future did exist at an earlier stage of these languages. Thus, the development of

the (infinitival) be-future in these languages was first restricted to imperfective infinitives; only later did the constraint loosen and perfective infinitives began to combine with the auxiliary.

4.3.3 Supporting evidence

There are empirical data to supp>ort the hypothesis that change-of-state and inceptive verbs are so similar semantically that meaning shifts from one type to the other can occur. In all three East Slavic languages, a similar shift can be discerned for another change-of-state verb: stat’ (Russian) / stac’ (Belarusan) / jfary (Ukrainian).^^

According to Flank (1987: 312), the modem Russian verb stat' can have two meanings, ‘become’ or begin’, and only the verb with the meaning of become’ has an imperfective counterpart. In other words,stat’ is polysemous. There is a perfective verb stat’ with an imperfective counterpart stanovit’sja, that has the meaning become’; cf.

Russian Ja xocu stat’ vracom I want to become a doctor’. In addition, there is a perfective verb stat ’2 that has no imperfective counterpart, and its meaning is begin’.

55 In the interests of space, this verb will be referred to with the Rus’ian forms stati or stanu.

141 Flank’s characterization of Russian stat’ is supported by Ozegov’s (1960) dictionary of

Russian, the SSRJa (1963), as well as other sources/^

In all the modem East Slavic languages, the perfective-only statij can be used as a future auxiliary. There is also evidence of this in the historical data for Russian, where future constructions with rran? &re attested from the second half of the sixteenth century and continue, in limited contexts, in the modem language (see 2.12.2 above).

An examination of the path of development that has been documented for stati reveals many similarities to the path that this study proposes for*bqdq. First, there is the evidence that the verb stati has become polysemous over time, with stati^ retaining change-of-state meaning and stati 2 developing into an inceptive phase verb. Proposing such a meaning-shift for *bqdq allows one to explain the colligability of the modem auxiliary. Stati 2 became a verb which takes infinitive complements, as has * bqdq.

Moreover, the shift ofstati 2 from change-of-state to inceptive verb has indeed led to a verb which takes infinitives of only imperfective verbs. Second, rrarn also (or perhaps only subsequently) underwent at least partial grammaticalization into a future auxiliary, with inceptive meaning giving way to future meaning.^*

The propensity for inceptives to grammaticalize into futures in Slavic, especially

East Slavic, is revealed not only by the behavior of stati 2 and *bqdq but by OCS nacbnu and Old Russian ucnu and pocnu as well. Thus, once the semantic shift of*bqdq into an inceptive phase verb occurred, its grammaticalization into a future auxiliary was hardly an

56 Sec also Dickey (1998: 28—32). 57 For Belarusan, see BRS (1989: s.v. "stac’"); for Russian, see Ozegov (1960: s.v. "stat"') and SSRJa (1963: s.v. "stat’"); for Ukrainian, see SUM (1978: s.v. "stavaty"). On the chronology of the development ofsta t’, see 2.12.2 above.

142 unusual occurrence. In general, the semantic shifts and grammaticalization of the verb stati 2 in East Slavic lend compelling support to the hypothesis that change-of-state verbs can develop into an imperfective future auxiliary.

4.4 The Slavic be-future as a manifestation of drift

The reductive paraphrases presented above indicate that *bqdq, a verb inherited by every Slavic language, is well-suited to grammaticalization into a future auxiliary. Many scholars have assumed that the development of a be-future in Slavic must have been a unique event, with use of the grammaticalized form of the verb then spreading to encompass its modem distribution. One sees this assumption especially as part of the extemal-source hypotheses described in 3.2 above.

However, facts of chronology and language contact militate against the idea that

*bqdq developed into a future auxiliary in one single area and subsequently spread to its current distribution. Rather, it must have developed independently in at least five linguistic areas: Russian, Belarusan-Ukrainian, PoUsh-Kashubian, Czech-Slovak, and

Slovene-Serbo-Croatian. Nevertheless, the presence of be-futures in so many Slavic languages is certainly striking, especially since the development of this type of future took place after the end of the Common Slavic period.

The possibility of drift, in which related languages develop along similar paths even after they have become separate, has been explored by scholars from Sapir (1921) onwards, and cases of drift comparable to the evolution of the Slavic be-fiiture have been identified in many languages. For example, LaPolla (1994) argues that a large number of

143 Tibeto-Burman languages have developed morphemes that mark what he terms “anti-

ergativity”, although no such form can be reconstructed for the protolanguage (ibid.:

66-67). In New Zealand English, Trudgill et al. (forthcoming) explore several

phonological changes that have taken the same path as changes in “English” English, even after speakers arrived in New Zealand. Within Slavic, both the evolution of the category of animacy and the development of the Common Slavic short vowels called “jers” appear to be examples of drift.^^

These examples of drift are quite similar to the path of development that the current study proposes for the be-future. By arguing that the semantics of *bqdq itself facilitated its grammaticalization into a future auxiliary, the possibility for be-futures to develop independently in many different Slavic languages is entirely feasible and by no means implausible. Thus, it is reasonable to consider the development of the Slavic be- future an example of drift.

4.5 The participial complement

The hypothesis presented in 4.3 allows for a more satisfactory explanation of the origins of the infinitival be-future; by positing a meaning-shift from change-of-state to inceptive, one can show how the verb *bqdq might have begun to combine with im­ perfective infinitival complements. However, this argument cannot be easily applied to the participial be-future that is found in several Slavic languages.

59 For further discussion of drift in Slavic, see Andersen ( 1990).

144 For Slovene and Serbo-Croatian, the presence of a participial future can be explained satisfactorily by proposing that the construction developed from the Common

Slavic future perfect. Unlike the North Slavic languages, these languages show no evidence of a clear-cut prohibition against perfective complements; this reflects the colligability of the future perfect, which allowed participles of either aspect.

It is the origins of the North Slavic participial be-future that remain unclear. By drawing a connection between the semantics of the verb *bqdq and phase-verb con­ structions, it becomes possible to explain the presence of the aspectual constraint on the formation of the infinitival be-future that is attested in North Slavic. This constraint is present for the North Slavic participial future as well, but the combination of*bqdq plus an imperfective participle cannot be viewed as analogous to constructions with phase verbs.

It is beyond the scop*e of the current study to present a complete description of how the North Slavic participial be-future came into existence; at the present time, the data needed to present such a description are not available. However, some factors can be identified that might have facilitated such a development. These include the areal distribution of the modem participial be-future, the semantics and usage of the /-participle in the languages with participial be-futures, and the relation between asp>ect and the semantics of the future perfect.

The data presented in Chapter 2, and especially the consolidation of that data in

Figure 2.2, reveal that the North Slavic languages (or dialects) which attest a participial be-future form a contiguous area. Western Ukrainian, Polish/Kashubian, and East Slovak

145 all converge at a central point, and the Lach dialects of Czech show many affinities with

neighboring Polish. Moreover, all these languages and dialects show a similar pattern of

spread for the participial future, with masculine singular forms appearing either earliest or

in isolation. In this way, one might hypothesize that the North Slavic participial be-future

is in fact an areal development. Without more historical data, however, such an idea is

speculative.

In all the North Slavic languages and in West South Slavic, a reform of the past-

tense system inherited from Common Slavic took place by which the older, synthetically

formed past tenses fell out of use. Concomitantly, the present perfect, formed with the imperfective nonpast form of “be’ plus the resultative /-participle, came to serve as the primary means of marking past tense. In East Slavic, the auxiliary/copula was lost and the past tense is now marked only by the participle. Thus, in East Slavic the /-participle has become entirely verbal in function, with its morphology the only vestige of its original nominal character.

Several languages, however, do preserve a pluperfect tense in some capacity, which is formed with the /-participle in combination with a different copula. Most interestingly, all languages that have some use of the participial be-future also have at least some evidence of a pluperfect construction using the same complement. This correlation, if it can be termed thus, can be seen in Table 4.1 below.

146 be+inf. be+part. pluperfect

Croatian + 4- 4-

Slovene — 4- 4- Czech + ——

Slovak + 4- 4-

Sorbian + — 4-

Polish + 4- 4-

Kashubian 4- 4- 4-

Russian 4- — -

Belarusan 4- — 4-

Ukrainian 4- 4- 4-

Table 4.1. Correspondence of languages with the be-futnre to those with pluperfects.

In Croatian, the be-future forms themselves are oftenfittura exacta rather than pure

futures; also, such forms are used in only limited contexts in Slovak (Short 1993b: 554), considered archaic in Polish (Rothstein 1993: 711), possible but rare in Kashubian (Stone

1993b: 777—78), and colloquial in Belarusan (Mayo 1993: 913).

In languages or dialects where the pluperfect is found, the /-participle by itself must be considered to be unmarked for tense; otherwise, it would not be possible for the participle to combine with different tenses of the copula. In this way, the presence of a pluperfect indicates that the participle conveys no tense meaning; this is in contrast to languages like Russian, where the /-participle itself conveys past-tense meaning and there is no pluperfect. Thus, for languages where the /-participle is unmarked for tense, the participial be-future appears structurally similar to other constructions of copula plus

147 participle, such as the passive future construction with *bqdq plus the past passive

participle.

Although one can identify a correlation that languages with participial be-futures

also tend to have an /-participle unmarked for tense, it is not clear how this situation is

relevant to the question of how the participial be-future developed. One can perceive an

analogous relationship between the participial be-future and the passive future in these

languages, but the former is subject to an aspectual constraint while the latter is not. A

more careful analysis of this question is needed.

It might be possible to hypothesize that the future perfect is the source for the

modem participial future if one can isolate some kind of factor that facilitated the

development of only imperfective future perfects into a be-future in North Slavic. Let us

return to the semantics of the future perfect, which was discussed in 3.3.2 above. In

particular, the relation of the semantics of the category to aspect might provide a clue

towards understanding the development of the participial be-future in North Slavic.

As is mentioned above, the future perfect describes the displaced perception of a

state, with futurity inherent in the fact that the perception takes place at a point after the

moment of speech. Not only is this perception at some future time, but it is understood

not to exist at the moment of speech; thus the change-of-state, nonpast perfective form

*bqdq is used as the copula.

The action or state that is being perceived can be either imperfective or perfective; this is expressed by the /-participle form. As is mentioned above, the semantics of the future perfect lends itself more easily to perfective meaning, and thus a majority of early

148 Slavic future perfects are perfective. However, the imperfective is by no means impossible; one such form occurs in OCS, and there are examples of *bqdq in combination with imperfective /-participles in Old Polish and Rus’ian texts as well.

Given the semantics of the future perfect in general, is it possible that the semantics of the imperfective future perfect specifically lent itself to reinterpretation into a pure future, whereas the perfective future perfect did not? Can we justify proposing that there was a split in the development of the future perfect along aspectual lines? If such a thing were possible, one could then argue that only imperfective future perfects developed into be-futures in North Slavic. This would account for the aspectual constraint that exists on the modem construction.

A detailed exploration of this question is beyond the scope of the current study.

However, some preliminary ideas can be presented here that suggest how one might support this line of reasoning.

Using reductive paraphrases, one can schematize the semantics of perfective and imperfective future perfects. The perfective future perfect is schematized in ( 1 ):

( 1 ) a. There comes into being after now b. it will be perceived c. the state of affairs in which d. an action has been completed

In this schematic, (a) represents the copula, which estabUshes future perspective and conveys change-of-state meaning, (b) represents the (displaced) perception of the action expressed by the /-participle, (c) represents the stativity/resultativity of the /-participle,and

(d) represents the perfectivity of the /-participle.

149 By contrast, the imperfective future perfect conveys a different aspectuality with

regard to the action. As with all imperfectives, the verb is unmarked for completeness or

boundedness; i.e. it is atelic. This can be seen in (2).

(2) a. There comes into being after now b. it will be perceived c. the state of affairs in which d. an action is happening without termini

Thus, the imperfective future perfect differs from the perfective only in the component which reflects the aspect of the participle (d).

These reductive paraphrases reveal that there is in fact very little semantic difference between the imperfective future perfect and the imperfective future. In fact, as the reductive paraphrase of the imp>erfective future (3) demonstrates, only the presence of components (b) and (c) in (2) distinguishes the future perfect from the future.

(3) a. There comes into being after now b. an action is happening without termini

In this way, it seems quite plausible for imperfective future perfect constructions to be reanalyzed as simple futures, with an elimination of the second reference point expressed in (c) and the displaced perception in (b) downgraded to implicature.

This analysis of the semantics of future perfects is only preliminary, but it does suggest a more principled means of understanding the development of imperfective participial be-futures in North Slavic. Further investigation is needed to verify this hypothesis; this must be presented in later work.

150 4.6 Conclusions

The reductive paraphrase for ‘become’ illustrates the future implicature that is

inherent in the semantics of the verb. Indeed, a general conclusion reached by this study

is that reductive paraphrases allow for the identification of the precise semantic content of

change-of-state verbs that serves as a catalyst for their grammaticalization. This

conclusion is relevant not only to the Slavic situation, but to any language where a future

marker has developed from a change-of-state verb.

However, in the case of Slavic *bqdQ, this conclusion is incomplete. In the

course of its grammaticalization, the verb adopted a new colligability pattern that must be

explained. It is at this point that the importance of examining Slavic phase verbs becomes

apparent. Their colligability is identical to that of the be-future, and there are numerous

examples of such verbs grammaticalizing into future auxiliaries in various Slavic

languages. By revealing the semantic link between phase and change-of-state verbs

through reductive paraphrases, it becomes possible to argue that a change-of-state verb

could undergo a semantic shift into a phase verb and, in the course of grammaticalization

into a future auxiliary, adopt the same colligability pattern as phase verbs. Thus, this

theory allows the development of the Slavic be-future to be described in a way that explains the presence of the aspectual constraint that is found in many of the languages.

In sum, the theory presented in this chapter provides a description of the development of the be-future that better accounts for the data and is more compatible with general linguistic theories of language change.

151 CHAPTER 5

CONCLUSION

The data presented in Chapter 2 reveal that what has been identified by the cover

term “be-fiiture” is in fact a set of constructions with diverse meanings, uses, and paths of development in the individual Slavic languages. The major flaw of previous theories is their failure to account for this diversity among be-future constructions; they tend to view the be-future as a single, unified type of construction for which a single point of origin is assumed.

The theory presented in the current study addresses the data more satisfactorily by positing that the be-future could have evolved independently. This view is supported by the fact, demonstrated in detail here, that the semantics of the change-of-state *bqdq inherently convey a future implicature that could be reanalyzed as the verb's core meaning. This implicature is revealed through the schematicization of the verb’s semantics by means of a reductive paraphrase. Since *bqdq was inherited into all Slavic languages, any of its reflexes would have been able to grammaticalize into a future auxiliary. Thus, this path of development is possible without any external influence.

By proposing that the be-future could develop independently, this theory avoids many of the problems encountered by previous scholarship. For example, with this

152 approach, there is no need to coordinate the chronologies of be-futures in neighboring

languages, since borrowing is not a necessary mechanism of the construction’s

appearance in the various Slavic languages. This theory also adds suppK)rt to the

argument proposed by Leiss (1985) that German borrowed its change-of-state future

from Slavic, since the construction is aspectually motivated in Slavic but not in German.

It also becomes possible to view the development of infinitival and participial be-futures

separately, thus avoiding the difficulties involved with attempting to isolate a single source for these different types of be-future.

This theory also provides a means of understanding the relationship between the be-future and aspect by explaining why be-futures in many of the Slavic languages combine with only imperfective complements. This constraint on colligability is identical to that of phase verbs, and a comparison of the reductive paraphrases of these verbs reveals that they are very similar semantically. Given the similar semantics of the two types of verb, one can argue that *bqdq underwent a semantic shift into a phase verb prior to its grammaticalization into a future auxiliary.

This theory is the most satisfactory explanation of the be-future’s aspectual constraint that has been presented, and there is empirical evidence that supports it. The verb stanu ‘become; begin’ is polysemous in all East Slavic languages, expressing either change-of-state or inceptive meaning. This verb has also served and continues to serve in some contexts as a future auxiliary, subject to the same colligability constraints that one typically finds for the be-future. Thus the semantic shift of a change-of-state verb into an

153 inceptive is not only hyp)othetically possible, it is actually attested in Slavic and its

chronology is comparable to that of the be-future.

The theory presented in this study best explains the development of infinitival be-

futures; regarding participial be-futures, the path of development is not as clear. Many

scholars have argued that the source for these constructions is the future perfect inherited

from Common Slavic, but the presence of the aspectual constraint on many modem

participial be-futures requires additional explanation. It may be the case that differences in

semantics led to a dual fate for the future perfect, with only imperfective future perfects

undergoing a shift into a construction expressing pure future. However, without further

investigation, this hypothesis remains speculative. A better understanding of the

semantics and development of the future p>erfect is a central goal of future research.

The situations in Slovene and Serbo-Croatian also merit further analysis. In

Slovene, the relationship between the use of the be-future and of nonpast perfective verbs

in future-tense contexts is unclear. In Serbo-Croatian, there is much variation among

dialects not only in the use of infinitival and participial complements, but also in the

meaning of constructions withbudem. Moreover, the history of these constructions

remains largely unexplored. A more detailed investigation of these languages and their

dialects might lead to a clearer understanding of the functions of these constructions and of the development of constructions using the become-type auxiliary.

This study demonstrates the rewards of examining problems in Slavistics from a more general linguistic perspective. Conversely, one sees that the Slavic languages are a relatively unmined source of data for cross-linguistic research on grammaticalization and

154 semantic and syntactic change. For example, the grammaticalization of change-of-state

and inceptive verbs into future auxiliaries has been little studied until now. Such verbs

develop into future auxiliaries less often than desideratives or modals, yet the Slavic

languages seem to attest this path of grammaticalization with disproportionate frequency.

Nonetheless, analogous developments are found in a diverse group of the world’s

languages. Thus, the analysis of the Slavic data presented in this study provides a

launching-point for a more general study of the grammaticalization of tliese verb-types.

This study also contributes to general linguistic study in that it demonstrates a new

use for a diagnostic tool for semantic analysis. Although reductive paraphrases were first

proposed by Wierzbicka as a means of analyzing semantic relationships on a synchronic

level, this study reveals the utility of these paraphrases in analyzing semantic change.

This study shows how reductive paraphrases can identify the precise semantic content of a verb that serves as a catalyst for its grammaticalization; by contrast, previous works on grammaticalization have dealt with semantics in a less systematic fashion.

In sum, this study greatly enhances our understanding of the development of a

be(come)’-type future in Slavic. It presents a more comprehensive collection of data than has been attempted previously, and combines general linguistic theories of language change and semantic analysis to propose a more satisfactory explanation of how the be- future evolved. Moreover, it uses reductive paraphrases in a new way, using them to systematize the analysis not only of semantics, but of semantic change. This approach leads to a more precise identification of the motivations behind semantic change and can be applied to a broad range of studies of grammaticalization and syntactic change.

155 BroUOGRAPHY

Adodurov, V. E. 1731/1969. Anfangs-Griinde der russischen Sprache. In Drei russische Grammatiken des 18. Jahrhunderts, ed. B. O. Unbegaun. Munich: Wilhelm Fink.

AI. 1968. Akty istoriceskie, sobrannye i izdannye arxeograficeskoju kommissiejii. Vol. 1. St. Petersburg: Tipografija Èkspedicii zagotovlenija Gosudarstvennyx bumag.

Aitzetmiiller, Rudolf. 1968. Das angebliche s-Futurum des Slavischen. In Studien ziir Sprachwissenschaft und Kulturkunde. Gedenkschrift fiir Wilhelm Brandenstein, ed. Manfred Mayrhofer, 11—16. Innsbrucker Beitrage zur Kulturwissenschaft 14. Innsbruck: Institut fiir Vergleichende Sprachwissenschaft der Leopold-Franzens- Univ.

Aleksic, R. 1937. Prilozi istoriji kajkavskog dijalekta. Juznoslovenskifllolog 16, 1—99.

Andersen, Henning. 1973. Abductive and deductive change. Language 49,165-93. ------. 1987. From auxiliary to desinence. In The Historical Development of Aicxiliaries, eds. Martin Harris and Paolo Ramat, 21-51. Berlin: Mouton de Gruyter. ------. 1990. The structure of drift. In Historical Linguistics 1987: Papers from the 8th International Conference on Historical Linguistics 8. (ICHL), eds. Henning Andersen and Konrad Koemer, 1-20. Amsterdam: Benjamins.

Adelphotës. 1591/1988. Grammatica dobroglagolivago ellinoslovenskago iazyka. Soversénnago iskustva osmi castéj slôva. Ko nakazânïju mnogoimenitomu Rossijskomu rôdu. In Adelphotës, ed. O. Horbatsch. 2nd ed. Specimina Philologiae Slavicae 76. Munich: Sagner.

Auty, Robert. 1980. Czech. The Slavic Literary Languages: Formation and Develop­ ment, eds. Alexander M. Schenker and Edward Stankiewicz, 163-82. New Haven, Conn.: Yale Concilium on International and Area Studies.

Bajerowa, Irena. 1992. Polski j^zykogôlny XIX wieku: Stan iewolucja. Vol. 2: Fleksja. Katowice: Uniwersytet Slgski.

156 Bandrivs’kyj, D. G. 1960. Hovirky pidbuz’koho rajonu l'vivs’koî oblasti. Kiev: AN URSR.

Barsov, A. A. 1784—88/1981. Obstojatel’naja rossijskaja grammatika. In Rossijskaja grammatika Antona Alekseevica Barsova, eds. B. A. Uspenskij and M. P. Tobolova. Moscow: MGU.

Belie, Aleksandr. 1965. Istorija srpskohrvatskog jezika. Vol. 2. 2nd ed. Belgrade: Naucnaknjiga.

Belie, Jaromir. 1972. Nâstinceskédialektologie. Prague: Stâtni pedagogieké nakl.

Benveniste, Émile. 1968. Mutations of linguistic categories. In Directions for Historical Linguistics: A Symposium, eds. W. P. Lehmann and Yakov Malkiel, 83—94. Austin: University of Texas Press.

Bemolak, Anton. 1790/1964. Grammatica Slavica Auctore Antonio Bernolak ad Systema Scholarum Nationalium in Ditionibus Caesareo-Regiis introductum accommodata. Editio prima in Pannonia. In Gramatické Dielo Antona Bernoldka, ed. Juraj Pavelek. Bratislava: SAV.

Bevzenko, S. P. 1960. Istorycna morfolohija ukrains'ko'imovy. Uzhorod: Zakarpats'ke oblasne vyd.

Bevzenko, S. P., A. P. Hryscenko, T. B. Lukinova, et al. 1978. Istorija ukrains’koj movy. Morfolohija. Kiev: Naukova dumka.

Bezpal'ko, O. P., et al. 1957. Istorycna hramatyka ukrai'ns'ko'i movy. Kiev: Radjans'ka Skola.

Binnick, Robert I. 1991. Time and the Verb: A Guide to Tense and Aspect. New York: Oxford University Press.

Birala, A. Ja., et al. 1957. Narysy pa historyi belaruskaj movy. Minsk: Dzjarzaunae vucebna-pedahahicnae vyd. Ministerstva asvety BSSR.

Bimbaum, David J. 1995. The Church Slavonic future participle from a comparative perspective. Die Welt der Slaven 40, no. 1: 76—92.

Bimbaum, Hetuik. 1956. Zum analytischen Ausdruck der Zukunft im Altkirchen- slavischen. Zeitschriftfurslavische Philologie 25, 1—7. ------. 1957. Zum periphrastischen futiuum im Gotischen und Altkirchenslavischen. Byzantinoslavica 18,77-81.

157 ------. 1958. Untersuchungen zu den Zukunftsumschreibungen mit dem Infinitiv im Altkirchenslavischen, ein Betrag zur historischen Verbalsyntax des Slavischen. Études de philologie slave 6. Stockholm: Almquist & Wiksell.

Bohoric [Bohorizh], Adam. 1584/1970. Arcticae Horulae Succisivae, de Latino- carniolana Literatura, ad Latinae Linguae Analogiam Accomodata, unde Moshoviticae.... In Arcr/cae//orn/ae, ed. BrankoBercic. Monumenta Literarum Slovenicarum VU. Ljubljana: ZGP Mladinska knjiga.

Bonfante, G. 1950. The origin of the Russian periphrastic future.Annuaire de l’Institut de Philologie et d ’Histoire Orientales et Slaves 10, 87-98.

Borkovskij, V. I. 1949. Sintaksis drevnerusskix gramot. Prostoe predlozenie. L vov: L'vovskij gosudarstvennyj univ. ------. 1958. Sintaksis drevnerusskix gramot. Sloznoe predlozenie. Moscow: AN SSSR. ------. 1968. Sravnitel’no-istoriceskij sintaksis vostocnoslavjanskix jazykov: Tipy prostogo predlozenija. Moscow: Nauka. ------. ed. 1978. Istoriceskaja grammatika russkogo jazyka. Sintaksis: prostoe predlozenie. Moscow: Nauka.

Borkovskij, V. I., and P. S. Kuznecov. 1963. Istoriceskaja grammatika russkogo jazyka. Moscow: AN SSSR.

Boiras, F. M., and R. F. Christian. 1971. Russian Syntax: Aspects of Modem Russian Syntax and Vocabulary. Oxford: Clarendon Press.

Braun, Maximilian. 1947. Grundziige der slawischen Sprachen. Gottingen: Vanden- hoeck & Ruprecht.

Brisard, Frank. 1997. The EngUsh tense-system as an epistemic category: The case of futurity. In Lexical and Syntactical Constructions and the Construction of Meaning: Proceedings of the Bi-Annual ICLA Meeting in Albuquerque, July 1995, eds. Marjolijn Verspoor, Kee Dong Lee, and Eve Sweetser, 271-85. Amsterdam : Benjamins.

Browne, Wayles. 1993. Serbo-Croat. In The Slavonic Languages, eds. BemardComrie and Greville G. Corbett, 306—387. London: Routledge.

BRS. 1989. Belaruska-ruski sloûnik. Minsk: AN BSSR.

Brycyn, M. Ja., Zovtobijux, M. A., and Majboroda, A. V. 1978. Porivnjal'na hramatyka ukrains’koi i rosijs’koi mov. Kiev: Vysa skola.

158 BS. 1965. Biblia Sacra iuxta Vulgatam Clementinam. 4th ed. Matriti: Biblioteca de Autores Cristianos.

Buffa, Ferdinand. 1995. Sarisské nârecia. Bratislava: Veda.

Bukatevic, N. I., et al. 1958/1969. Ocerki po sravnitel’noj grammatike vostocno­ slavjanskix jazykov. Slavistic Printings and Reprintings 137. The Hague: Mouton.

Bulyka, A. M., A. I. Zurauski, and U. M. Svjazynski. 1990. Mova vydannjaü FrancyskaSkaryny. Minsk: Navuka i tèxnika.

Bunina, I. K. 1959. Sistema vremen staroslavjanskogo glagola. Moscow: AN SSSR.

Buslaev, F. I. 1959. Istoriceskaja grammatika russkogo jazyka. Moscow: Gosudar- stvennoe ucebno-pedagogiceskoe izd-vo.

Bybee, Joan L. 1985. Morphology: A Study of the Relation Between Meaning and Form. Amsterdam:Benjamins.

Bybee, Joan L., and Osten Dahl. 1989. The creation of tense and aspect systems in the languages of the world. Studies in Language 13, 51—103.

Bybee, Joan L., and William Pagliuca. 1987. The evolution of future meaning. In Papers from the 7th International Conference on Historical Linguistics, eds. A. G. Ramat, O. Camiba, and G. Bernini, 109—22. Amsterdam: Benjamins.

Bybee, Joan L., William Pagliuca, and Revere Perkins. 1991. Back to the future. In Approaches to Grammaticalization, eds. Elizabeth Gloss Traugott and Bemd Heine, 17—58. Amsterdam: Benjamins. ------. 1994. The Evolution of Grammar. Chicago: University of Chicago Press.

Cassius, Bartholomeus [Bartol Kasic]. 1604/1977. Institutionum linguae illyricae libri duo. Vol. 2. In InstitutionesLinguaeIllyricae, ed.ReiiiholdOlesch. Slavistische Forschungen 21. Cologne: Bohlau.

Cemysev, V. I. 1970. Opisatel’nye formy naklonenij i vremen v russkom jazyke. Izbrannye trudy v dvux tomax. I: 230—59. Moscow: Prosvescenie.

Cemyx, P. Ja. 1952. Istoriceskaja grammatika russkogo jazyka: Kratkij ocerk. Moscow: Gosudarstveimoe ucebno-pedagogiceskoe izd-vo. . 1953. Jazyk Ulozenija 1649 goda. Moscow: AN SSSR.

159 . [Tschemych] 1957, Historische Grammatik der russischen Sprache. Haile (Saale): VEM Max Niemeyer.

Chung, Sandra, and Alan Timberlake. 1985. Tense, aspect, and mood. In Language Typology and Syntactic Description, ed. Tim Shopen, 3: 202—58. Cambridge: Cambridge University Press.

Clark, James Michael. 1978. A Linguistic Analysis of Ten Russian Texts o f the Seventeenth-Century Vesti Kuranty. Unpublished Master’s thesis. The Ohio State University, Columbus, Ohio.

Cocron, Friedrich. 1962. La langue russe dans la seconde moitié du XVlle siècle (Morphologie). Paris: Institut d ’Études Slaves.

Collins, Daniel E. n.d. The perfective aorist of ‘be’ in the 1028 Chronicle entry: A Rus’ian case of no news being good news. Unpublished squib.

Comrie, Bernard. 1976. Aspect. Cambridge: Cambridge University Press. ------. 1985. Tense. Cambridge: Cambridge University Press.

Cyran, Wladyslaw. 1961. Dlaczego gina w jçzyku polskim formy czasu przysztego zlozone z bezokolicznikiem?J^zykpolski 41, no. 3: 223—4.

DABM. 1963. Dyjalekalahicny atlas belaruskaj movy. Minsk: AN BSSR.

Dahl, Osten. 1985. Tense and Aspect Systems. Oxford: Blackwell.

Damborsky, Jirf. 1967. Participiuml-ové ve slovanstinë. Warsaw: PWN.

DDG. 1950/1970. Duxovnye i dogovornye gramoty velikix i udel'nyx knjazej XfV—XVI w. Slavica Reprint 40. Düsseldorf: Briicken-Verlag.

Dejna, Karol. 1957. Gwary ukrainskie tarnopolszczyzny. Wroclaw: PAN.

Dickey, S. M. 1998. Expressing ingressivity in Slavic: The contextually-conditioned imperfective past vs. the phase verb stat’ and the procedural za-. Journal of Slavic Linguistics 6, no. 2: 11—44.

Diels, Paul. 1963. AltkirchenslavischeGrammatik. 2nd ed. Heidelberg: C. Winter.

Dostal, Antonin. 1954. Studie o vidovém systému v staroslovénstinë. Prague: Statni pedagogickénakl.

160 Dumovo, Nikolaj. 1924/1962. Ocerki po istoriji russkogo jazyka. Slavistic Printings and Reprintings 22. ‘S-Gravenhage: Mouton.

D'urovic, L’ubomir. 1980. Slovak. In The Slavic Literary Languages: Formation and Develoment, eds. Alexander M. Schenker and Edward Slankiewicz, 211—28. New Haven: Yale Concilium on International and Area Studies.

Endzeüns, Jânis. 1971. Jànis Endzetins’ Comparative Phonology and Morphology of the Baltic Languages. Translated by William R. Schmalstieg and Benjamips Jëgers. The Hague: Mouton.

FaBke, Helmut. 1981. Grammatik der obersorbischen Schriftsprache der Gegenwart: Morfologie. Bautzen: VEB Domowina-Verlag.

FD. 1968. Freisinger Denkmaler. Brizinski spomeniki. Monumenta frisingensia. Geschichte, Kultur und Geisteswelt der Slowenen 2. Munich: Rudolf Trofenik.

Ferrell, James. 1953. On the aspects of byt’ and on the position of the periphrastic imp>erfective future in Contemporary Literary Russian. Word 9, 362—76.

Flajsans, V. 1903. Nejstarsipamdtkyjazyka ipisemnictvi ceského. Pt. I: Prolegomena a texty. Prague: Frant. Backovsky.

Flank, Sharon. 1987. Phase subdivisions and Russian inceptives. Die Welt der Slaven 32, 310-16.

Fleischman, Suzanne. 1982. The Future in Thought and Language. Cambridge: Cambridge University Press.

Forsyth, John. 1970. A Grammar of Aspect. Cambridge: Cambridge University Press.

Gallon, Herbert. 1976. The Main Functions of the Slavic Verbal Aspect. Skopje: Macedonian Academy of Sciences and Arts. ------. 1979. Some peculiarities of verbal aspect in Slovene. Slovene Studies I, 52-60. ------. 1981. The specific position of Slovene in the Slavic verbal aspect. Slovene Studies 3,49—58. ------. 1987. From Indo-European perfect to Slavic perfect to Slavic preterite. In Papers from the 7th International Conference on Historical Linguistics, eds. Anna Giacalone Ramat, Onofrio Carruba and Giuliano Bernini, 251-65. Amsterdam:Benjamins.

Gasihski, Tadeusz. 1971. On the functional load of the contemporary variants of the Polish future tense. The Polish Review 16, 108—13.

161 Gebauer, Jan. 1909. Historickâ mluvnice jazyka ceského. Vol. 3: Tvaroslovi. Prague: Unie.

Georgieva, V. L. 1968. Istorija sintaksiceskix javlenij russkogo jazyka. Moscow: Frosvescenie.

Giacalone Ramat, Anna. 1986. On language contact and syntactic change. \n Linguistics across Historical and Geographical Boundaries: In Honour of Jacek Fisiak on the Occasion of his Fiftieth Birthday. Vol I : Linguistic Theory and Historical Linguistics, eds. Dieter Kastovsky and Aleksander Szwedek, 317-28. Berlin: Mouton de Gruyter.

Givôn, Talmy. 1980. The binding hierarchy and the typology of complements. Studies in Language 4, no. 3: 333—77.

Gorecka, Janina, and Witold Smiech. 1972. Czas przyszly zloZony w jçzyku polskim. Rozprawy Komisji J^zykowej Lodzkiego Towarzystwa Naukowego 18, 11—38.

Grenoble, Lenore. 1989. Tense, mood, aspect: The future in Russian. Russian Linguistics 13, 97—110. ------. 1995. The imperfective future tense in Russian. VPorJ46, 183—205.

GRJa. I960. Grammatika russkogo jazyka. 2 vols. Moscow: AN SSSR.

Groschel, Bernhard. 1972. Die Sprache Ivan Vysenskyjs: Untersuchungen und Materialien zur historischen Grammatik des Ukrainischen. Cologne: Bohlau.

Harris, Alice C. 1997. Remarks on grammaticalization. In Proceedings of the Lexical Functional Grammar Conference (LFG97), eds. Miriam Butt and Tracy Holloway King. Stanford, CA: CSLI Publications. Available only online: http://www.-csli.stanford.edu/publicationsUFG2/harris-lfg97.html

Harris, Alice C. and Lyle Campbell. 1995. Historical Syntax in Cross-Linguistic Perspective. Cambridge Studies in Linguistics 74. Cambridge: Cambridge University Press.

Haugen, Einar. 1950. The analysis of linguistic borrowing. Language 26, 210-31.

Havranek, Bohuslav. 1939. Aspect et temps du verbe en vieux slave. In Mélanges de linguistique offerts a Charles Bally sous les auspices de la Faculté des lettres de l'Université de Genève par des collègues, des confrères, des desciples reconnaissants, 223—30. Geneva: Georg et cie, s a.

162 Heine, Bemd. 1993. Auxiliaries: Cognitive Forces and Grammaticalization. Oxford: Oxford University Press.

Heine, Bemd, Ulrike Claudi, and Friederike Hiinnemeyer. 1991. Grammaticalization: A Conceptual Framework. Chicago: University of Chicago Press.

Hewson, John, and Vit Bubenik. 1997. Tense and Aspect in Indo-European Languages: Theory, Typology, Diachrony. Amsterdam: Benjamins.

Hock, Hans Henrich. 1986. Principles of Historical Linguistics. Berlin: Mouton de Gruyter.

Hopper, Paul, and Elizabeth Closs Traugott. 1993. Grammaticalization. Cambridge: Cambridge University Press.

Isacenko, Aleksandr [Issatchenko, Alexander]. 1940. Tense and auxiliary verbs with special reference to Slavic languages. Language 16,189—99. ------. 1962. Die russische Sprache der Gegenwart. Pt. 1: Formenlehre. Halle (Saale): VEB Max Niemeyer.

Ivekovic, F., and Ivan Broz. 1901. Rjecnik hrvatskoga jezika. 2 vols. Zagreb: Karl Albrecht.

Jakobson, Roman. 1932/1984. Stmcture of the Russian verb. In Roman Jakobson, Russian and Slavic Grammar, Studies I93I—I98I, eds. Linda Waugh and Morris Halle, 1—14. Berlin: Mouton. ------. 1957/1984. Shifters, verbal categories, and the Russian verb. In Roman Jakobson, Russian and Slavic Grammar, Studies 1931—1981, eds. Linda Waugh and Morris Halle, 41—58. Berlin: Mouton.

Janda, Richard D. and Brian D. Joseph. Forthcoming. On language, change, and language change—or, of history, linguistics, and historical linguistics. In Handbook of Historical Linguistics, eds. Richard D. Janda and Brian D. Joseph. Oxford: Blackwell.

Jankowski, F. 1983. Histarycnaja hramatyka belaruskaj movy. Minsk: Vysejsaja Skola.

Jaskevic, A. A. 1996. Starabelaruskijahramatyki. Minsk: Belamskaja navuka.

Jesperson, Otto. 1924/1992. The Philosophy of Grammar. Chicago: University of Chicago Press.

163 Joseph, Brian D. 2000. Is there such a thing as “grammaticalization?” Language Sciences, forthcoming issue.

Kalsbeek, Janneke. 1998. The Cakavian Dialect of Orbanici near 2minj in I stria. Studies in Slavic and General Linguistics 25. Amsterdam: Rodopi.

Karabowicz, Kazimiera. 1994. Fleksja rzeczownikowa i czasownikowa w gwarze ukraihskiej wsi Stary Brus. Slavia Orientalis 43, no. 1: 85—99.

Karskij, E. F. 1956. Belorusy: Jazyk belorusskogo naroda. Part 2—3. Moscow: AN SSSR.

Keller, R. E. 1961. German Dialects: Phonology and Morphology, with Selected Texts. Manchester: Manchester University Press. ------. 1978. The German Language. Atlantic Highlands, NJ: Humanities Press.

Kellner, Adolf. 1946. Vychodolasskd ndrecl. Vol. 1. Bmo: Dialektologicka komise pn Matici moravské.

Kiparsky, Valentin. 1967. Russische historische Grammatik. Vol. 2. Heidelberg: C. Winter.

Klein, Wolfgang. 1994. Time in Language. London: Routledge.

Klemensiewicz, Zenon, T. Lehr-Splawihski, and S. Urbahczyk. 1955. Gramatyka historyczna jazyka polskiego. Warsaw: PWN.

Kolesov, V. V., and V. V. Rozdestvenskaja, eds. 1994. Domostroj. St. Petersburg: Nauka.

Konzal, Vaclav. 1994. Latinske participium futuri v staroslovënském pfekladu. Slavia 63,193-205.

Kopitar, Bartholomaus [Jemej]. 1808/1971. Grammatik der slavischen Sprache in Krain, Kdrnten und Steyermark. Ljubljana: Cankaijeva zalozba.

Kowalska, Alina. 1976. Ewolucja analitycznychform czasownikowych z imiesiowem na -I w j^zyku polskim. Katowice: Uniwersytet Sl^ski.

Kravar, Miroslav. 1978. Zur Futurperiphrase budem + Infinitiv im Serbokroatischen. In Slavistische Studien zum VIII. Internationalen Slavistenkongress in Zagreb 1978, eds. Johannes Holthusen, Wolfgang Kasak, and Reinhold Olesch, 255—63. Cologne: Bohlau.

164 ------. 1986. Zur Herkunft der slavischen Futurperiphrase + Infinitiv. In Fest­ schrift fiir Herbert Braiier zum 65. Geburtstag am 14. April 1986, eds. Reinhold Olesch und Hans Rothe, 273-83. Cologne: Bohlau.

Krizkova, Helena. 1957a. Jestë k staroslovénskému sbtvorilb bqdq. Slavia 26,500-05. ------. 1957b. K vyvoji zàpadoslovanského futura. Sbornik Vysoké Skoly Pedagogické v Olomouci: Jakyk a Lite ratura 4, 21—4%. ------. 1960. Vyvoj opisného futura vjazyclch slovanskych, zvlâstév rustiné. Prague: S tatni pedagogické nakl.

Kuznecov, P. S. 1953a. Istoriceskaja grammatika russkogo jazyka: Morfologija. Moscow: MGU. ------. 1953b. K voprosu o genezise vido-vremennyx otnosenij drevnerusskogo jazyka. Trudy Instituta Jazykoznaniia AN SSSR 2,220-53. ------. 1959. Ocerki istoriceskoj morfologii russkogo jazyka. Moscow: AN SSSR.

Lamprecht, Amost, Dusan Slosar, and Jaroslav Bauer. 1986. Historickâ mluvnice cestiny. Prague: Statm pedagogické nakl.

LaPolla, Randy J. 1994. Parallel grammaticalizations in Tibeto-Burman languages: Evidence of Sapir’s ‘drift’. Linguistics of the Tibeto-Burman Area 17, no. 1: 61-80.

Lehr-Splawihski, Tadeusz. 1957. Zarys gramatyki historycznej jçzyka czeskiego. 2nd ed. Warsaw: PWN.

Leiss, Elisabeth. 1985. Zur Entstehung des neuhochdeutschen analytischen Futurs. Sprachwissenschaft 10, no. 3—4: 250—73.

Lencek, Rado L. 1982. The Structure and History of the Slovene Language. Columbus, Ohio:Slavica.

Lermontov, M. Ju. 1957. Sobraniesocinenij. Vol. 1. Moscow: Gos. izd-vo xudozest- vennoj literarury.

Lixacev, D. S., ed. 1989. Slovar’ kniznikov i kniznostiDrevnejRusi. Leningrad: Nauka.

Lomososov, M. V. 1755/1972. Rossijskaja grammatika Mixajla Lomonosova. Photomechanic reprint. Leipzig: Zentralantiquariat der DDR. . 1764/1980. Russische Grammatick verfasset von Herrn Michael Lomonosow ... aus dem Russischen iibersetzt von I.-L. Stavenhagen. Specimina Philologiae Slavicae 27. Munich: Sagner.

165 Lomtev, T. p. 1941. Ucenye zapiski. Serija: Filologiceskajall. Minsk: BGU. ------. 1956. Ocerki po istoriceskomu sintaksisu russkogo jazyka. Moscow: Moscow University Press. ------. 1961. SravniteV no-istoriceskaja grammatika vostocnoslavjanskix jazykov. Morfologija. Moscow: Vyssaja Skola.

Lopusanskaja, S. P. 1981. K stanovleniju analiticeskix form vremeni v istorii russkogo glagola. Ceskoslovenskà rusistika 26, no. 5: 213—18.

Lorentz, Friedrich. 1919/1971. Kaschubische Grammatik. Hildesheim: Dr. H. A. Gerstenberg. ------. 1925. Geschichte der pomoranischen (kaschubischen) Sprache. Berlin: Walter de Gruyter.

Los, Jan. 1915. Syntaktyczne u±ycie form gramatycznych. In J^zyk polski i jego historya z uwzglednieniem innych j^zykow na ziemiach polskich Pt. 2, eds. Henryk Ulaszyn, et al. Krakow: Nakl. Akademii umiejçtnosci.

Ludolf, Heinrich Wilhelm. 1696. Grammatica Russica quae continet non tantum praeclpua fundamenta russicae linguae.... In Henrici Wilhelmi Ludolfi Grammatica Russica, ed. B. O. Unbegaun. Oxford: Clarendon.

Liidtke, Helmut. 1987. Auxiliary verbs in the universal theory of language change. In Papers from the 7th International Conference on Historical Linguistics, eds. Anna Giacalone Ramat, Onofrio Carruba and Giuliano Bemini, 349—54. Amsterdam : Benjamins.

Lunt, Horace. 1974. Old Church Slavonic Grammar. 6th ed. ‘S-Gravenhage: Mouton. ------. 1992. Notes on nationalist attitudes in Slavic studies. Canadian Slavonic Papers 34, no. 4, 459—70.

Machek, Vaclav. 1951. S l a v . ‘ich werde sein’. Zeitschriftfiir slavische Philologie 21, 154-58. ------. 1965. Slaw, bqdq, Hilfsverbum zur Bildung des analytischen Futurums. Studia zfilologiipolskiej i siowiahskiej 5 , 67—75.

Magner, Thomas F. 1966. A Zagreb Kajkavian Dialect. The Pennsylvania State University Studies 18. University Park, PA: The Pennsylvania State University.

Malic, Dragica. 1972. Jeziknajstarije hrvatskepjesmarice. 2[agreb: HFD.

Mayo, Peter. 1993. Belorussian. InThe Slavonic Languages, eds. Bernard Comrie and Greville G. Corbett, 887—946. London: Routledge.

1 6 6 Mescerskij, N. A., ed. 1972. Russkajadialektologija. Moscow: Vyssaja skola.

Mikes, Michael J. 1985. Alternative forms of the future imp>erfective tense in Polish. Slavic and East European Journal 29, 448—60.

Miodek, Jan. 1985. Czas przyszly zloiony: norma i czas.J^zykpolski 55,1 2.

Mucke, Karl Emst. 1891/1965. Historische und vergleichende Laut- und Formenlehre der niedersorbischen (niederlausitzisch-wendischen) Sprache. Reprint. Leipzig: Zentral-Antiquariat der DDR.

Nap'erskij, K. E. [Napiersky] 1868. Russko-livonskie akty (Russisch-livldndische urkunden). St. Petersburg: Tipografija imperatorskoj akademii nauk.

Nikiforov, S. D. 1952. Glagol: Ego kategorii i formy v russkoj pis'mennosti vtoroj poloviny XVI veka. Moscow: AN SSSR.

Nitsch, K. 1956. Tajemnice polskiego czasu przyszlego zloZonego.J^zyk polski 36, 190-96.

ObnorskiJ, S. P. 1934. Russkaja pravi/a kak pamjatnik russkogo literatumogo jazyka. Izvestija AN SSSR, VII ser., Otdelenie obscestvennyx nauk, no 10., 749—76. ------. 1953. Ocerki po morfologii russkogo glagola. Moscow: AN SSSR.

Ohijenko, Ivan. [Ilarion, Metropolitan of Winnipeg and All Canada.] 1930. Ukraïns’ka literaturna mova XVI-ho st. i ukraïns’kyj Krexivs’kyj apostoL Vol. 1. Warsaw: Drukamja synodal’na.

Ozegov, S. I. 1960. Slovar’ russkogo jazyka. 4th ed. Moscow: Gosudarstvennoe izd- vo inostrannyx i nacional’nyx slovarej.

Palmer, P. R. 1986. Mood and Modality. Cambridge: Cambridge University Press.

Pappas, Panayiotis. 1999. The development of counterfactuals withthélâ ‘want’ in Early Modem Greek. Ohio State University Working Papers in Linguistics 52, 15—40.

Pennington, A. E. 1968. Future periphrases in 17th-century Russian: Some evidence from translated material. The Slavonic and East European Review 46, no. 106: 31-47.

Pescak, M. M. 1974. Hramoty XIV st. Kiev: Naukova Dumka.

167 Petrovic, Dragoljub. 1978. Govor Banije i Korduna. Novi Sad: Matica srpska; Zagreb : Prosvjeta.

Pettersson, There. 1970. A note on future time and future tense in Russian. Scando-Slavica 16,97—103.

Plater, W. E., and H. J. White. 1926. A Grammar of the Vtdgate: Being an Introduction to the Study o f the Latinity o f the Vulgate Bible. Oxford: Clarendon.

Polanski, Kazimierz. 1980. Serbian (Lusatian). In The Slavic Literary Languages: Formation and Development, eds. Alexander M. Schenker and Edward Stankiewicz, 229-45. New Haven, Coim.: Yale Concilium on International and Area Studies.

Polivka, Jerzy. 1888. Czas przyszly w jçzyku staroslowiehskim. Prace filologiczne 2, 175-94.

Popovic, Ivan. 1960. Geschichte der serbokroatischen Sprache. Wiesbaden: Otto Harrassowitz.

Porak, Jaroslav. 1979. Chrestomatie k vyvoji ceskeho jazyka (13.—18. stoleti). Prague: Statm pedagogické nakl.

Potebnja, A. A. 1899/1958. Iz zapisok po russkoj grammatike. Vols. 1—2. Moscow: AN SSSR.

Priestly, T. M. S. 1993. Slovene. In The Slavonic Languages, eds. Bernard Comrie and Greville G. Corbett, 388—451. London: Routledge.

Proeme, Henk. 1991. On the compound future in Polish. InStudies in West Slavic and Baltic linguistics, eds. A. A. Barentsen, B. M. Groen, and R. Spoenger, 181—271. Amsterdam: Rodopi.

Remneva, M. L. 1988. Oformlenie uslovnyx konstrukcij v pamjatnikax delovoj i knizno-slavjanskoj pis’mennosti XI—XII vv. Filologiceskienauki no. 5: 47—54.

Resetar, M. 1898. Primorski lekcionari XV. vijeka. Rad Jugoslavenske Akademije Znanosti i Umjetnosti 136, 97—199.

Rothstein, Robert. 1993. Polish. InThe Slavonic Languages, eds. Bernard Comrie and Greville G. Corbett, 686—758. London: Routledge.

168 Rosier, Karl. 1952. Beobachtungen und Gedanken iiber das anaiytische Futurum im Slavischen. Wiener Slavistisches Jahrbuch 2, 103—49.

Respond, Stanislaw. 1971. Gramatyka historyczna jazyka polskiego. Warsaw: PWN.

Rusanivs’kij, V. M. 1971. Struktura ukrains'koho dieslova. Kiev: Naukova Dumka.

Sapir, Edward. 1921. Language. New York: Harcourt, Brace and Co.

Saxmatov, A. A. 1941. Sintaksis russkogo jazyka. Leningrad: Ucebno-pedagogiceskoe izd-vo.

Schenker, Alexander. 1993. Proto-Slavic. In The Slavonic Languages., eds. Bernard Comrie and Greville G. Corbett, 60-124. London: Routledge.

Schleicher, A. 1855. Das futurum im deutschen und slawischen. Zeitschrift fiir vergleichende Sprachforschung 4, 187—97.

Scholz, Friedrich. 1986. Zur Entwicklung der grammatischen Kategorie Futurum im Russischen. In Festschrift fiir Herbert Brauer zum 65. Geburtstag am 14. April 1986, eds. Reinhold Olesch and Hans Rothe, 467—78. Cologne: Bohlau.

Schooneveld, C. H. van. 1951. The aspect system of the Old Church Slavonic and Old Russian verbum flnitum byti. Word 7.96—103. ------. 1959. A Semantic Analysis of the Old Russian Finite Preterite System. *S- Gravenhage: Mouton.

Schuster-Sewc, Heinz. 1967. Das niedersorbische Testament des Miklawus Jakubica 1548. Berlin: Akademie-Verlag.

Shevelov, George [Serex, Jurij]. 1951. Narys sucasnoi ukrains’koi literaturnoi movy. Munich: Molode Zyttja. ------. 1952. Problem of Ukrainian-Polish linguistic relations from the tenth to the fourteenth century. Word 8, no. 4, 329—49. ------. 1953. Problems in the formation of Belorussian. Word 9 {Slavic Word 1), 3-100. ------. 1980. Ukrainian. In The Slavic Literary Languages: Formation and Development, eds. Alexander M. Schenker and Edward Stankiewicz, 143-60. New Haven, Conn.: Yale Concilium on International and Area Studies.

Short, David. 1993a. Czech. In The Slavonic Languages, eds. Bernard Comrie and Greville G. Corbett, 455—532. London: Routledge.

169 . 1993b. Slovak. In 7/ie S/avon/c eds. Bernard Comrie and Greville G. Corbett, 533—92. London; Routledge.

Simundic, Mate. 1971. Govor imotske krajine i bekije. Sarajevo: Akademija nauka i umjetnosti Bosne i Hercegovine.

SJaS. 1961. Slovmk jazyka staroslovënského. Vol. 4. Prague: Nakl. ceskoslovenské akademie vèd.

Slonski, Stanislaw. 1953. Historia jqzyka polskiego w zarysie. Warsaw: PWN.

Smotryc'kyj, Meletij. 1619/1974. Hrammatiki siovenskija pràvilnoe syntagma. Specimina Philologiae Slavicae 4. Frankfurt am Main: Kubon & Sagner.

Sobolevskij, A. I. 1907/1962. Lekcii po istorii russkago jazyka. 4th ed. Slavistic Printings and Reprintings 37. S-Gravenhage: Mouton.

SPP. 1994. Slownik poprawnej polszczyzny. 18th ed. Warsaw: PWN.

Sreznevskij, 1.1. 1893—1912. Materialy dlja slovarja drevnerusskogo jazyka. 3 vols.

SRJa. 1975. Slovar’ russkogo jazyka XI—XVII w . Moscow: Nauka.

SSRJa. 1963. Slovar' sovremennogo russkogo jazyka. Leningrad, Moscow: AN SSSR.

Stang, Chr. S. 1935. Die westrussische Kanzleisprache des Grossfiirstentums Litauen. Oslo: J. Dybwad. ------. 1939. Die altrussische Urkundensprache der Stadt Polozk. Oslo: I Kommisjon Hos Jacob Dybwad. ------. 1942. Das slavische und baltische Verbum. Skrijier utgitt av det Norske Videnskaps-Akademi i Oslo. II. Historisk-Filosofisk Klasse. Oslo: J. Dybwad.

Stanislav, Jan. 1957. Dejiny slovenského jazyka. Vol. 3: Texty. Bratislava: Slovenska AkademiaVied. ------. 1958. Dejiny slovenského jazyka. Vol. 2: Tvaroslovie. Bratislava: Slovenska AkademiaVied.

Stankiewicz, Edward. 1984. Grammars and Dictionaries of the Slavic Languages from the Middle Ages up to 1850: An Annotated Bibliography. Berlin: Mouton.

170 Stanojcic, Z. 1983. Die diachronisch-syntaktische Basis fiir eine mogliche Sprach- interferenz am Beispiel einer Verwendung des Futur II im Serbokroatischen. ZeitschriftfürSlawistik 28, no. 4: 641—52.

Statorius, Petrus [Piotr Stojenski]. 1568/1980. Polonicae Grammatices Institutio. Slavistische Forschungen 26. Cologne: Bohlau.

Stecenko, A. N. 1977. Istoriceskij sintaksis russkogo jazyka. Moscow: Vyssaja Skoia.

Stevanovic, Vida. 1986. Oko altemativne upotrebe dvaju oblika u drugom delu slozenog futura. Nasjezikll, 118—24.

Stieber, Zdzislaw. 1954. Czas przyszly niedokonany w zabytkach polskich XIV i XV wieku. Rozprawy Komisji J^zykowej Lodzkiego Towarzystwa Naukowego 1, 231-34.

Stoic, Jozef. 1994.Slovenskâ Dialektolôgia. Bratislava: Veda.

Stone, Gerald. 1993a. Sorbian. In The Slavonic Languages, eds. Bernard Comrie and Greville G. Corbett, 593-685. London: Routledge. ------. 1993b. Cassubian. In The Slavonic Languages, eds. Bernard Comrie and Greville G. Corbett, 759—94. London: Routledge.

Stur, L’udovit. 1846/1943. Nauka reci slovenskej: Vybrané ukâzky. In Ndrecie slovenské alebo potreba pisania v tomto nâreci, ed. Henrich Bartek, 127—290. Turciansky sv. Martin: Kompas.

SUM. 1978. Slovnyk ukraïnskoïmovy. Kiev: Naukova dumka.

Sychta, Bernard. 1973. Slownik gwar kaszubskich na tie kultury ludowej. 7 vols. Wroclaw: PAN.

Szemerényi, Oswald. 1990. Einfuhrung in die vergleichende Sprachwissenchaft. 4th ed. Darmstadt: WissenschaftlicheBuchgesellschaft.

Thelin, Nils B. 1978. Towards a Theory of Aspect, Tense and Actionality in Slavic. Stockholm: Almqvistand Wiksell.

Thomason, Sarah Grey. 1980. Morphological instability, with and without language contact. In Historical Morphology, ed. Jacek Fisiak, 359—72. Trends in Linguistics: Studies and Monographs 17. The Hague: Mouton.

171 Thomason, Sarah Grey and Terrence Kaufman. 1988. Language Contact, Creolization, and Genetic Linguistics. Berkeley: University of California Press.

Timberlake, Alan. 1977. Reanalysis and actualization in syntactic change. In Mechanisms of Syntactic Change, ed. Charles N. Li, 141—77. Austin: University of Texas Press.

Tixonravov, N., ed. 1863. Pamjatniki otrecennoj russkoj literatury. Vol. 2. Moscow: SORJaS.

Tommola, Hannu. 1984. On the aspectual significance of 'phase meanings’. InAspect Bound: A Voyage into the Realm of Germanic, Slavonic and Finno-Ugrian Aspectology, eds. Casper de Groot and Hannu Tommola, 111—32. Dordrecht: Foris.

Topolihska, Zuzanna. 1974. A Historical Phonology of the Kashubian Dialects of Polish. Slavistic Printings and Reprintings 255. The Hague: Mouton.

Townsend, Charles E., and Laura A. Janda. 1996. Common and Comparative Slavic: Phonology and Inflection. Columbus, Ohio: Slavica.

Trudgill, Peter, et al. Forthcoming. The role of drift in the formation of native-speaker southern hemisphere Englishes: Some New Zealand evidence. Diachronica 17, no. 1.

TZG. 1979. Texte zur Geschichte der polnischen und tschechischen Sprache. Specimina Philologiae Slavicae 23. Frankfurt am Main: Kubon & Sagner.

Ultan, Russell. 1978. The nature of future tenses. In Universals of Human Language, eds. Joseph Greenberg, Charles Ferguson, and Edith Moravcsik, 3: 83—123. Stanford, Calif.: Stanford University Press.

Unbegaun, B. O. 1958. Russian grammars before Lomonosov.Oxford Slavonic Papers 8,98-116.

Uzevic, Ioann. 1645/1996. Gramatyka slovenskaja. ... \n Starabelaruskijahramatyki, A. A. Jaskevic. Minsk: Belaruskaja navuka.

Vaillant, André. 1966. Grammaire comparée des langues slaves. Vol. 3: Le verbe. Lyon: LAC. ------. 1979. La langue de Dominko Zlataric: Poète Ragusain de la fin du XVle siècle. Vol 3: Syntaxe. Belgrade: Académie serbe des sciences et des arts.

172 Vinogradov, V. V. 1972. Russkij jazyk: Grammaticeskoe ucenie o slave. 2nd ed. Moscow: Vyssaja Skoia.

Vlasto, A. P. 1988. A linguistic history of Russia to the end of the Eighteenth Century. Oxford: Clarendon Press.

Vontsolos, Nicholas. 1978. A Study of Seventeenth Century Russian (Based on the Bezobrazov Collection). Unpublished Ph.D. dissertation. The Ohio State University, Columbus, Ohio.

Vukicevic, Milosav S. 1978. Pisani jezik na Kosovu i u Metohiji sedamdesetih godina XIX veka. Pristina: Jedinstvo.

Vukovic, Jovan. 1957—58. Futur Il-gi i ekvivalentni glagolski oblici po upotrebu u srpskohrvatskom jeziku. Pitanja knjizevnosti i jezika 4—5, pt. B, 5-28.

Vyskocil, Pavel. 1956. Staroslovënské 5bTvori/b S/ov/a 25, 260-61.

Wandas, Adam. 1966. J^zyk staroruskiego przekladu polskich statutôw ziemskich Kazimierza Wielkiego i Wladysiawa Jagieily. Wroclaw: PAN.

Wells, C. J. 1985. German: A Linguistic History to 1945. Oxford: Clarendon Press.

Wexler, Paul. 1977. A Historical Phonology of the Belorussian Language. Historical Phonology of the Slavic Languages 3. Heidelberg: Carl Winter.

Whaley, Marika. 1995. The Lithuanian Statute of 1529: A Linguistic Analysis. Unpublished qualifying paper. The Ohio State University, Columbus, Ohio. ------. 1998. The origins of the Russian imperfective future constructionbudu + infinitive. Paper presented at the annual meeting of the American Association of Teachers of Slavic and East European Languages, San Francisco, Calif. ------. 1999. Tracing the origins of the Slavic imperfective be-future. Ohio State University Working Papers in Linguistics 52, 159—71.

Wierzbicka, Anna. 1978. Syntax vs. semantics. Theoretical Linguistics 5, no. 1: 115-33. ------. 1988. The Semantics of Grammar. Studies in Language Companion Series 18. Amsterdam.Benjamins.

Willett, Thomas. 1988. A cross-linguistic survey of the grammaticization of evidentiality. Studies in Language 12, no. 1:51-97.

Worth, Dean S. 1983. The Origins o f Russian Grammar. Columbus, Ohio: Slavica.

173 Wytrzens, Günther. 1953. Zur Frage des periphrastischen Futurums im Russischen. Wiener Slavistisches Jahrbuch 3, 22—27.

Zagrodnikowa, Alicja. 1972. Rywalizacja dwu typow czasu pizyszlego zloionego: pisai - b^d^ pisac. J^zyk polski 52, no. 5: 346—58.

Zakrevska, Ja. V., et al. 1984. Atlas ukrainskoi movy v trokh tomakh. (Only two vols, published.) Kiev: Naukova dumka.

Zylko, F. T. 1966. Narysy z dialektologii ukrains’koi movy. 2nd ed. Kiev: Radjans'ka Skola.

Zyzanij-Tustanovs’kyj, Lavrentij. 1596/1972. Grammatica slovenska suversennago iskustva osmi castij slova, i inyx nuzdyx. In Hrammatika Slovenska, ed. Gerd Freidhof. Specimina Philologiae Slavicae 1. Munich: Kubon & Sagner.

174