Sussex R., Cubberley P. The Slavic Languages

Подождите немного. Документ загружается.

dominance of Moscow as an ideological (as well as cultural and

linguistic) model, and also served to reunite the linguistic traditions

of Eastern Europe with the mainstream of Western linguistics. Russian

was no longer the favored foreign language, and was not the official

language of trans-national bodies like Comecon and the Warsaw Pact.

Everywhere throughout the former Soviet Union and Eastern Europe

Russian is being supplanted by English, and to a lesser extent German

and French, in the educational, commercial and cultural domains.

This is not to say that Russian has been totally replaced: Russia itself

can be expected eventually to return to a position of greater economic

and cultural influence, and the linguistic infrastructure of decades of

working with Moscow are still evident. In addition the 1994 elections

in Ukraine and Belarus, for instance, showed a clear reversion to

more Moscow-aligned policies in politics and economics, as did even

more strongly the 2002 economic pact between Russia and Belarus.

Nevertheless, the new orientation Westwards is having some profound

effects on the modern Slavic languages.

Several tendencies can be discerned from the point of view of the current standing

of the literary languages. First, with the exception of Sorbian, the smaller languages

have benefited by being removed from the authority of neighboring larger

languages into new language-states of their own. As we have seen, Belarusian,

Slovak, Slovenian and Macedonian are the winners. Ukrainian will benefit from

the removal of intrusive Russian models and pressures. The middle-sized languages

already with sufficient independence – Polish, Czech, Bulgarian, Croatian and

Serbian – will probably not gain a lot from their new status, though in the case of

Croatian and Serbian we can expect greater differentiation as the two states diverge

politically and culturally, coupled with a decline in the use and status of Serbo-

Croatian and the emergence of Bosnian.

There remain Russian and Sorbian. Russian loses from the new political

dispositions. Its status as a supra-national language in the USSR has gone, and

its role as a trans-national and international language has been reduced. Within

Russia, however, there is (as elsewhere throughout the Slavic area) a revival of

literature in new, less regimented and more colloquial genres, and the language is

anything but languishing. Sorbian, however, is potentially in peril. Some national-

istic pressures in contemporary Germany favor Germanness at the expense of

multiculturalism, and it is not at all clear to what extent or in what form Sorbian

will continue, the language policies of the European Community notwithstanding,

as a fully viable language (or languages) beyond the next generation.

2.5 Overview 109

Phonology

3.1 Introduction

This chapter begins by examining the equivalences for vowels, consonants, groups

of sounds and suprasegmentals with reference to Proto-Slavic (3.2). We then

discuss the systems of vowels (3.3), consonants (3.4) and suprasegmentals (3.5)

of the modern languages. For conventions of transcription and transliteration,

see appendix B.

3.2 Historical evolution and modern equivalences

The first problem in studying the sounds of modern Slavic is one of equiva-

lence: which sound corresponds to which other sound or sounds in the eleven

languages? Some sounds show wide variations, like the root vowel in the word

for ‘white’:

(1) B/C/S be``o, Rus be

lyj, Pol biały,Czbı

, Slk biely, Blg bjal

But not all Russian words with e correspond to Slovak words in ie, Polish words in

ia, Bulgarian words in ja or Czech words in ı

, cf. the word for ‘day’:

(2) Rus den

, Blg, Cz, Slk den, Pol dzien

, B/C/S da

We therefore need a common basis for comparison of equivalent or cognate sounds

and their typological description.

We shall use Proto-Slavic as the reference-point. It is neutral with respect to all

the other Slavic languages. Using it in this way is linguistically informative, since

many of the sound changes which have formed the modern Slavic languages, like

palatalization and assimilation, are still living as phonological and morphological

processes in the modern languages. And the equivalences for a sound in any

language can be recovered by checking in the various equivalence tables in which

110

it occurs. In other words, we are able to discuss both the historical development of

Slavic (its genetic evolution) and the internal phonological processes of the modern

languages (their synchronic typology) in comparable terms. Combining diachronic

and synchronic analyses works well for Slavic. It is very much in the traditions of

Slavic linguistics. And it is particularly useful in phonology, where we find wide

variations in the data.

In 3.2 we describe the main changes in vowels (3.2.1.), consonants (3.2.2.)

and sound-combinations (3.2.3.) which have formed the modern languages from

Proto-Slavic; in 3.2.4 we describe the changes in suprasegmental features. For

details of the Proto-Slavic system and its features, see 1.3.

3.2.1 Development of the Proto-Slavic vowel system

The late Proto-Slavic vowel system had the following shape (see also table 1.3):

Front Central Back

High i y u

Mid-high {y

Mid-low e e˛ oo˛

Low e

Plus fronted allophones of /u/, /o˛ / and /a/ after palatal consonants

Features of special interest are the development of (1) the high-mid vowels {, y

( jers) (3.2.1.1); (2) the low-mid nasal vowels (e˛ o˛)) (3.2.1.2); and (3) low front /e

( jat

) (3.2.1.3), followed by (4) vowels in word-initial position (3.2.1.4); (5) quality

changes (free and contextual) (3.2.1.5); (6) contraction (3.2.1.6); and (7) diphthongs

(3.2.1.7).

3.2.1.1 The jers ({y)

(On the name ‘jer’, see 1.3.1.6) These two vowels (derived from IE short i˘ u

respectively) are often referred to as ‘‘reduced’’ or ‘‘ultra-short’’ vowels. However,

it is unlikely that there were three degrees of length in the system (ultra-short, short

and long). More likely, the location of these short vowels in the central area made

them weak in the same way as the neutral vowel (schwa) is typically weak in many

systems, especially those with relatively strong stress. Their disappearance as

phonemes was in principle a Proto-Slavic phenomenon, since it occurred in all

areas. However, the realization of the disappearance varied widely across the

whole Slavic area, indicating that it began right at the end of the Common

Slavic period (1.1).

3.2 Historical evolution and modern equivalences 111

Throughout the Slavic area the jers were in principle lost in ‘‘weak’’ position,

and virtually throughout fused with other existing vowels when in ‘‘strong’’ position.

The two contexts may be summarized as follows:



a jer was weak in word-final position, or where the following syllable

contained a vowel other than a jer;



a jer was strong where the following syllable contained a weak jer.

Complications to this simple rule of alternating strong and weak positions

arose from resistance to the consonant clusters which were potentially formed by

the loss of a weak jer, in which context all the languages retained such a jer to some

extent or other. The extent of retention is a partial indicator of the relative

consonantism vs vocalism of each language (for example, the extremes are

Sorbian (consonantal) and B/C/S (vocalic), see further below 3.2.3.1).

In addition to creating new and more common consonant clusters, the loss of the

jers signals the end of the open-syllable structure which had prevailed in Proto-

Slavic, since words, and thus syllables, could now end in a consonant. It further

signals the start of the assimilation in clusters (especially of voice, but also place

and, to a lesser extent, manner of articulation) which occurs in all the languages,

and of the devoicing of final voiced obstruents which occurs in all but two

languages (B/C/S and Ukrainian) (3.2.3.2).

A development related to the resistance to new clusters was – again to a greater

or lesser extent by language – the introduction of fill vowels to obviate the

appearance of (even existing) consonant clusters in the new final position. In all

cases the fill vowel has the same form as the strong jer, meaning that this develop-

ment pre-dates the completion of the loss of the jers, the assumption being that the

loss in final position was early and was compensated for by an inserted jer while the

jers were still in the system. In a few cases, initial clusters have been obviated by

vowel prothesis (3.2.1.4).

In two modern languages (Bulgarian and Slovenian) the strong position for at least

the back jer is reflected in a schwa ([

] ). However, there is plenty of evidence that

schwa was the first result in most areas, subsequently to be itself merged with other

vowels. Also in many of the Slavic areas the jers merged together first (in schwa)

before the final step, while other areas kept them apart.

Since the loss of internal jers was dependent on word shape, the loss gave rise to

the vowel  zero alternation which is one of the most typical morphophonological

features of Slavic (4.2.1).

The realization of the strong jers by language is summarized in table 3.1. For

Bulgarian and Macedonian we give plural or definite article forms, since they have

no declensions.

112 3. Phonology

Table 3.1. Realization of strong jers

PSl examples: syny ‘sleep’ NomSg  syna GenSg; d{n{ ‘day’ NomSg  d{nja GenSg;

dyzˇ d { ‘rain’ NomSg  dyzˇdja GenSg; myxy ‘moss’ NomSg  myxa GenSg; p{sy ‘dog’

NomSg  p{ sa NomSg; l{ny ‘flax’ NomSg  l{na GenSg

Language Back jer (y) Front jer ({) Examples (NomSg  GenSg)

Blg /

/ /e/ sa˘n  sa˘nja

t (Def); den  dni (Plur);

da˘zˇd  da˘zˇdove

; ma˘x  ma˘

xove (Plur);

pes  pe

sove (Plur); len le

na˘t (Def)

Mac /o/ /e/ son  sonis

ta/snis

ta; den  denove/dni;

dozˇd  dozˇdovi; mov  movot;

pes  psa (dial. pesovi); len  lenot

B/C/S /a/ /a/ sa`` n  sna`` ; da

n  da

na;

da`` zˇ d  da

zˇda; ma

x  ma`` xa;

pa`` s  psa`` ; la`` n  la`` na

Sln /a

//a

/ se

n  sna

; da

n  dne

;

zˇ  dezˇja

; ma

h  ma

ha;

s  psa

; la

n  la

Rus, Bel /o/ /

e/ Rus, Bel son  sna; den

 dnja;

dozˇd

 dozˇdja

;

Rus mox  mxa, Bel mox  mo

xu or imxu

(then e > o, 3.2.1.5): Rus, Bel pe

s  psa;

Rus le

n  l

na, Bel le

n  le

nu or l

nu or il

Ukr /o/ /e/ son  sna; den

 dnja;

dos

 dos

; mox  mo

xu;

pes  psa; l

on  l

Pol /e/ /'e/ sen  snu; dzien

 dnia;

deszcz  deszczu/d

zu; mech  mchu;

pies  psa; len  lnu

USorb /e/ /e/ dz

´en

 dnja; des

 des

a; len  lena

(then e > o before a hard

consonant): so

n  sona; moch  mocha; pos  psa

LSorb /e/ /e/ z

´en

 dnja ; des

 des

a; mech  mecha;

(then e > o or a before a

hard consonant): son

 sni; pjas  psa; lan  lana

Cz /e/ /e/ sen  snu; den  dne;

des

t’  des

; mech  mechu;

pes  psa; len  lnu

Slk /e o a/ /e/ sen  sna; den

 dn

zˇ  dazˇd’a; mach  machu;

pes  psa; l’an  l’anu

(Central dialects and standard; West and East dialects as in Czech)

3.2 Historical evolution and modern equivalences 113

From table 3.1 we can see that most of West and South Slavic probably first

merged the jers into [

]. West Slavic at first retained the hard or soft quality of the

preceding consonant, a feature which thereby became phonemic, but was later

lost in Czech and Slovak. In most areas the later result was the merging of /

/ w ith

the low-mid vowel /e/. Only B/C/S and partly Slovenian fused the schwa with /a/.

Macedonian, with the same results as East Slavic (/o/ vs /e/) – though without the

hard  soft distinction – may reflect what is often said to be the result in

Old Church Slavonic, but the latter is more likely to be seen in the Bulgarian

retention of schwa. It is unclear whether Bulgarian passed through the ‘‘merged’’

schwa stage, though this possibility is supported by many examples of [

]from

front { (e.g. pa˘n ‘tree-stump’ from PSl *p{ny). Macedonian, on the other hand,

has maintained the original quality distinction like East Slavic, and so cannot in

the interim have merged the two (though its northern dialects behave like

Bulgarian).

One further position of the jers needs to be mentioned. In the ‘‘tense’’ position,

before a [ j], the jers tended to be raised rather than lowered. In all areas except

Russian these jers merged with high /i/ and /y/. In Russian these jers, along with

/y/ and /i/, were treated as normal jers, becoming /o/ and /e/ when strong, and lost

when weak. In this case, the position of the stress was a significant factor: in the

Proto-Slavic forms *s

ja ‘neck’ and *my

jo˛ (or *myVjo˛ ) ‘I wash’, the vowel was

treated as strong, even though clusters of ‘consonant þ j’ would not have been

unacceptable in most areas: compare PSl *pijo˛

/p{jo˛

‘I drink’, in which the

unstressed jer was allowed to disappear (Rus p

ju). In PSl *moldyVj{ ‘young’

[NomSgMasc] (and all such long adjectival endings, e.g. *do

bryj{ ‘good’), the

first jer was strong regardless of stress, and never disappeared. In Old Russian

such vowels always became /o e/, though subsequently, under the conditions of

unstressed vowels (see below 3.3.2.6), only the stressed ones have remained /o/, and

the others are now treated as /y/ and /i/ (the RusChSl result). These forms are

reflected as follows:

(3) Rus s

ja mo

ju p

ju molodo

j do

bryj

All other areas have s

ja, my

j-u/em (or s

ja, mı

j-u/em in languages where /y/ does

not exist). Where the long adjectival endings have survived, the forms are -yj/ij

or contracted -y

/ı

The fill vowels (above) also belong here, since they are the same as the results

of the strong jers. Indeed, we might include here those weak jers which did

not disappear (see 3.2.3.1), since the motivation for the two is identical. The clusters

always involve a final sonorant, since ‘obstruent þ sonorant’ were the typically

acceptable Proto-Slavic clusters, based on the principle of rising sonority

114 3. Phonology

(see 1.3.1.1). The fill vowel appears only when the cluster is final, that is, where the

ending is zero:

(4) Fill vowels

PSl ogn{ ‘fire’ ve

ˇtr

y ‘wind’ mysl{ ‘thought’

o˛ gl{ ‘coal’ osm{ ‘eight’ (all NomSg)

Blg o

ga˘n vja

ta˘r mı

sa˘l (va

glen) o

sem

Mac ogan vetar (misla)(jaglen) osum

B/C/S o

ganj ve``tar m

sao u``galj o`` sam

Sln o

genj ve

ter mı

sel o

gel o

sem

Rus ogo

ter mysl

gol

sem

Ukr voho

vı

ter mysl

hil

vı

sim

Bel aho

ter mysl

hal

sem

Pol ogien

wiatr mys

lwe˛giel osiem

Sorb wohen

tr mysl (wuhlo) wosm

Cz ohen

vı

tr mysl uhel osm

Slk ohen

vietor mysel

uhol

osem

The inconsistency of the fill vowels in such noun contexts in East and West Slavic is

readily seen, and only South Slavic, as the most vocalic group, regularly inserts a

vowel. A common verbal context for such new clusters was the past participle

active NomSgMasc, with the suffix -l-: psenes-ly ‘carry’, mog-ly ‘be able’. Such forms

are normally regularized morphologically, with or without the inserted vowel:

in East Slavic the norm is removal of the suffix /l/ (e.g. Rus ne

s [Fem nes-la

mog [Fem mog-la

] ). In languages with a syllabic /l

/ no change is necessary (e.g.

Cz nes-l, moh-l, both bisyllabic), though a vowel may still be inserted (e.g. Slk nies-

o-l, moh-o-l). Elsewhere either the cluster is accepted (e.g. Pol nio

s-ł, mo

g-ł ) or not

(e.g. B/C/S tre

s-a-o ‘shake’ (Fem tre

s-la), mo`` g-a-o, where final /l/ has become /o/;

Blg ne

s-a˘-l, mog-a˘

-l).

3.2.1.2 The nasal vowels (e˛,o˛)

At the time of syllable opening the nasal vowels of Proto-Slavic had appeared

from vowels followed by nasal consonants within the same syllable (‘‘tautosyl-

labic’’), with front and back versions according to the quality of the original vowel

(see 1.3.1.7). This meant that they typically alternated with nasal consonants, and

were probably not well integrated into the vowel system (cf. Shevelov, 1964: 583).

Simultaneously with the reappearance of closed syllables – caused by the loss

of the jers – the nasal vowels began to lose their nasality. For one thing, new

clusters with tautosyllabic nasal consonants began to appear, albeit rarely,

3.2 Historical evolution and modern equivalences 115

but they nevertheless became structurally possible (e.g. PSl *ton{k-yj{ ‘thin’ >

tonk-yj). This, combined with the weak integration and articulatory complexity of

the nasal vowels, led to their demise in most of the area fairly early: probably

in the tenth century in all East Slavic, in all but Polish and Polabian in West

Slavic, and in B/C/S, but somewhat later in the rest of South Slavic. Traces

may still be found in some dialects of Bulgarian, Macedonian and Slovenian (10.2).

Of the standard languages, only Polish has retained the nasal vowels, at least at

the structural level, and in the orthography. In practice they are usually realized as

an oral vowel plus nasal consonant, the articulation of the latter corresponding to

that of the following consonant (be˛de˛ ‘I will be’ [bfndf] ). Only in a few contexts is a

nasal vowel the norm: before spirants (mie˛so ‘meat’) and in final position for the

back nasal (sa˛ ‘they are’). Otherwise true nasal vowels are heard only in formal,

careful speech (3.3.2.3). Such articulations are identical to new (foreign origin)

sequences of ‘vowel þ nasal consonant’, which may also produce nasal vowels in

the pre-spirant position (kwadrans ‘quarter of an hour’). In this same context

standard Bulgarian may also have nasalized vowels, though they need not subsume

the nasal consonant (o

nzi ‘that’ [

zi] or [

nzi] ).

The quality of the Polish nasal vowels is interesting etymologically, for they have

clearly gone through a stage of merging and then one of separation, in which the

new quality does not correspond to the old. In stage one the two merged into a

central nasal vowel (probably schwa, but perhaps rounded), but the previous

quantity distinction was retained, as was the hard or soft quality of the preceding

consonant. In stage two, with the loss of phonemic quantity in Polish, the long

version (either originally long or newly long through contraction or compensation)

moved back (the modern letter a˛). The short version moved forward (the modern

letter e˛), while the consonant palatalization remained phonemic.

In the other languages the quality distinction was usually retained and the

denasalized vowels merged with others in the system:

o˛ > u and e˛ > a

: Rus, Ukr, Bel, Cz, Slk, Sorb

o˛ > u and e˛ > e: B/C/S

o˛ > o and e˛ > e: Sln

o˛ >

and e˛ > e: Blg

o˛ > a and e˛ > e: Mac

So all East and West Slavic except for Polish had a common result for both nasal

vowels ( > u/a

), and all South Slavic had a common result for the front nasal vowel

( > e). Subsequent changes to the [a

] have separated East and West Slavic some-

what, since [a

] became /a/ or /e/ in Czech and Sorbian (and even /o/ in the latter),

116 3. Phonology

and /a/ or /a

/ in Slovak, but always /a/ in East Slavic. Some varieties of Bulgarian

and Macedonian may well have initially merged the quality of the two into a

nasalized schwa and then separated them again in the same way as Polish.

In table 3.2 we give the reflexes of the four Proto-Slavic forms (back  front and

long  short).

Table 3.2. Nasal vowel reflexes

a. Proto-Slavic

1. short back (o˛):

ro˛ka ‘hand’ zo˛ba ‘tooth’ [GenSg] mo˛ka ‘flour’ ido˛ ‘I go’

2. long back (o˛):

so˛dy ‘court’ zo˛by ‘tooth’ mo˛ka ‘torture’ ido˛t{ ‘they go’

3. short front (e˛ ):

pe˛t{ ‘five’ me˛so ‘meat’ je˛zyky ‘tongue’ se˛ ‘self’ [AccSg]

4. long front (e˛ ):

pe˛tyj{ ‘fifth’ re˛dy ‘row’ te˛gno˛ti ‘to pull’ ljube˛t{ ‘love’/prose˛t{

‘ask’ [3 Pl]

(Note: short and long here include later results of suprasegmental processes.)

b. Modern

Blg/Mac 1 ra˘ka

/raka za˘

bi/zabi (Pl) – ı

]/(idam)

2 sa˘d/(sud) za˘b/zab ma˘

ka/maka ı

da˘t/idat

3 pet meso ezı

k/jazik se

4 pe

ti/petti red te

gn[

]/tegne lju

bj[

]t/ljubat

B/C/S/ Sln 1 ru

ka/ro

ka zu`` ba/zo

ba mu

ka/mo

ka (ı

dem)

2 su

d/so

dzu

b/zo

bmu`` ka/(mu

ka)



u/(dial. ido

)

3 pe

t/pe

tme

so/meso

zik/je

zik se

4 pe

tı¯/pe

ti re

d/re

d Cr te

gnuti lju

be/lju

bijo

Rus/Ukr/Bel 1 ruka

ba muka

idu

2 sud zub mu

ka idu

3 pjat

mja

so jazy

k sja

4 pja

tyj rjad/Bel rad tjanu

lju

bjat

Pol 1 re˛ka ze˛bu me˛ka ide˛

2 sa˛ dza˛ bma˛ ka ida˛

3 pie˛c

mie˛so je˛zyk sie˛

4 pia˛ ty rza˛ d cia˛ gna˛ c

lubia˛

Cz/Slk 1 ruka zubu/zuba muka jdu/(idem)

2 soud zub mouka/mu

ka jdou/idu

3 pe

ˇt/pa

t’ maso/ma

so jazyk se/sa

4 pa

/piaty r

d/rad ta

hnout/

tiahnut’

prosı

/prosia

U/LSorb 1 ruka zuba muka du

2 sud zub – du

3 pjec

/pe

mjaso/me

so jazyk/je

zyk so/se

4 pjaty rjad c

ahnyc

ˇgnus

lubja/lubje

3.2 Historical evolution and modern equivalences 117

3.2.1.3 jat

ˇ)

The table of Proto-Slavic vowel phonemes in 3.2.1 makes /e

/ look well integrated as

a front partner to /a/. But the reality is that the [a

] position became unstable after

the developments which led to phonemicization of palatalization. In the former

opposition of hard  soft syllables, it was the vowel which was dominant, for

example in the opposition of, say, /ta/ to /t

/. The back  front distinction resided

mainly in the vowel, supported by the hard or soft nature of the consonant. After

the loss of the jers and the closing of syllables, the bond between the consonant and

vowel was weakened and the single segments became phonemic. In this situation it

was the consonants which became dominant within the system, since they were

‘‘independently’’ phonemic, for example in final position where no vowel followed.

In the new syllables /t-a/ and /t

-a

/ the quality of the vowel was secondary, and since

in the context of simple phonetics the distance between /a/ and /a

/ is small and fragile,

/ readily came to be seen as an allophone of /a/. Thus Proto-Slavic /e

/ found itself

losing independence. But in several areas it appears to have ‘‘defended itself’’, so to

speak, in the short term by shifting to a higher position, in the extreme case (Ukr)

ending up as /i/, while original /i/ there moved back to /y/. Former /e

/ is now nowhere

a separate phoneme, and is directly traceable only in the jekavski (Jekavian) variant

of B/C/S, where it is reflected in the phonemic sequences /j-e/ or /i-j-e/. Elsewhere it

has coalesced with some other phoneme – most often /e/, sometimes /a/. In the case of

Ukrainian it was joined in the /i/ position by new /i/ from other sources (3.2.1.5).

A different view about the systemic weakness of /e

/ is that, unlike the other

phonemes, it was phonetically still a complex sound, effectively a rising diphthong

of the /ea-ia/ type (Shevelov, 1964: 585). Its complexity within an otherwise ‘‘pure

vowel’’ system was therefore its weakness. This does make the actual range of shifts

easier to understand. The phonetic nature of /e

/ at the end of Proto-Slavic is open

to debate, and the phonemic scenario described in the preceding paragraph can

equally well account for its subsequent fate, since the end point of push–pull

activity is not necessarily phonetically predictable. So we shall simply note the

fact of its weakness and observe its reflexes, which must also be related to the rest of

each given vocalic system (examples are given in table 3.3).

/ > /e/: Rus, Bel, B/C/S(Ekavian), Mac

> /ie/: Cz, Slk, Sorb, B/C/S(Jekavian), Sln

> /a/(via /a

/): Pol, Blg

>/i/: Ukr, B/C/S(Ikavian) (On the variants of B/C/S

see 1.5.3 and 10.2.2.)

It is probably no coincidence that the areas which initially ‘‘retained’’ /a

/ (but

where it later became /a/) were those which also retained nasality, on the grounds

118 3. Phonology