Sussex R., Cubberley P. The Slavic Languages

Подождите немного. Документ загружается.

pronunciation also at the end of words in formal speech, and at slow tempi.

Elsewhere the nasal quality is reduced (see 3.3.2.7).

3.3.2.4 Other vowel phonemes

Other vowel phonemes found in individual languages are:

(53a) Slk /a

/[Q]or[f]: pa

t’ ‘5’

(53b) Sln /e

/[eø]: pe

t [peøt] ‘5’

cf. Sln /e

/[fø]: zˇ e

na [¨dføna] ‘woman’

(53c) Sln /o

/[oø]: po

t [poøt] ‘road, way’

cf. Sln /o

ø]: no

zˇ [n

] ‘knife’

(53d) Blg and Sln /

/: Blg pa˘t ‘road’

Sln pe

s ‘dog’

3.3.2.5 Prepalatalization

From a phonological point of view, in some Slavic languages the vowels /i/ and /e/

are preceded only by palatal or palatalized consonants. Orthographically this is

sometimes interpreted as the vowels ‘‘prepalatalizing’’ the consonants:

/i/ [i] (orth. i) prepalatalizes: all consonants in hard/soft pairs:

Ukrainian, Sorbian

labials, velars and /l/: Polish

velars and /l/: Bulgarian, Macedonian

[i] (orth. i) (of /i/) indicates a preceding soft phoneme: Russian,

Belarusian

/e/ prepalatalizes: all consonants in hard/soft pairs (native words):

Russian, Belarusian

dentals (which become palatal): Slovak (with

some exceptions)

velars and /l/: Bulgarian, Macedonian

orth. e

(/e/) indicates a preceding palatal phoneme: Czech

/ prepalatalizes: all consonants in hard/soft pairs: Sorb

3.3.2.6 Stressed (or pitched) vs unstressed vowels

Stressed vowels are usually louder, longer and higher in pitch than comparable

unstressed vowels, very much as in English. Unstressed vowels may show modifi-

cations of vowel quality. As a general rule, the greater the contrast of energy

between stressed and unstressed syllables, the more the unstressed vowels will be

‘‘reduced’’. The fixed-stress languages have relatively weaker stress. Their vowels

tend to preserve their quality in unstressed syllables. The free-stress languages tend

3.3 Modern vowel systems 159

to have a stronger contrast between stressed and unstressed syllables, and

the unstressed vowels tend to be reduced (Bulgarian, Slovenian, Russian,

Belarusian). Ukrainian is an exception to this generalization: unstressed vowels

suffer only a minor reduction in quality. The same applies to B/C/S, in this case no

doubt related to the pitched nature of its stress.

Reduction of unstressed vowels may occur anywhere in the word, or may be

conditioned by the position of the unstressed syllable with respect to the stressed

(‘‘tonic’’) syllable of the word. Such unstressed vowels result in the neutralization of

oppositions in the vowel system, and in a reduction in the number of distinctive

vowel sounds. Reduced vowels are marked in the Belarusian orthography, but in

the other languages one must predict the vowel quality on the basis of the known

phonetic properties of the vowel system.

The reduction of unstressed vowels in the four languages creates a second vowel

system (a sub-system) with fewer contrastive phonemes (table 3.7).

In Slovenian the close versions /e

/ merge with the open /e o/ in unstressed

position:

(54) Sln ko

st ‘bone’ [koøst]

kostı

[GenSg] [k

¨stiø]

sti [DatSg] [¨k

østi]

Unstressed /e/ is raised in Bulgarian (sub-standardly) and Russian (after soft

consonants) to [h], in Russian to [



i] after hard consonants, by a process known by

its Russian name ı

kan

e (3.2.1.5):

(55) Rus sem

‘7’ [s

] semı

[GenSg] [s

h¨m

cen ‘price’ [GenPl] [t



sfn] cena

[NomSg] [t



i¨na]

But in Belarusian unstressed /e/ after hard consonants is lowered to /a/ (akan

e):

(56) Bel c

rap ‘skull’, [Pl] c

arapy

Table 3.7. Stressed vs unstressed vowel systems

Slovenian: Stressed system: i e

ou(r

)

Unstressed system: i e

aou(r

)

Bulgarian: Stressed system: i e

aou

Unstressed system: i (e)

Belarusian: Stressed system: i (y) e a o u

Unstressed system: i (y) (e)

Russian: Stressed system: i (y) e a o u

Unstressed system: i (y) a u

160 3. Phonology

Unstressed /e/ after soft consonants is also lowered to /a/, but only in pretonic

syllables ( ja

kan

e):

(57) Bel zjamlja

‘earth’ [Pl] ze

mli

bjazzu

by ‘toothless’ bezadka

zny ‘irresponsible’ po

le ‘field’

Unstressed Bulgarian /a/ merges with unstressed /

/as[

(58) Blg brat ‘brother’ bra

t ‘the brother’ [¨brat

grad ‘city’ grada˘

t ‘the city’ [gr

¨d

Unstressed Bulgarian /o/ fuses with unstressed /u/ as [U]:

(59) Blg oko

‘eye’ [U¨k

] uxo

‘ear’ [U¨x

]

But in Belarusian unstressed /o/ merges with unstressed /a/ as /a/ (‘‘strong akan

e’’):

(60) Bel zo

lata ‘gold’ zalaty

‘golden’ [NomPl]

In Russian, unstressed /o/ and /a/ after hard consonants are pronounced [ˆ]in

absolute initial or first pretonic position, but [

] elsewhere (also akan

e):

(61) Rus moloko

‘milk’ [m

lˆ¨k

]

karanda

‘pencil’ [k

rˆn¨da

]

okno

‘window’ [ˆk¨n

]

abriko

s ‘apricot’ [ˆbrh¨k

slo

vo ‘word’ [¨sl

]

ma ‘Mum’ [¨mam

]

These four languages give the clearest evidence of the interaction of stress and

reduction in vowel quality. However, the other languages may also show some

variation in vowel quality in conjunction with stress or length. In first pretonic

position, Ukrainian /o/ takes on something of the [U] quality of unstressed

Bulgarian /o/ (especially before stressed /u/):

(62) Ukr holu

bka ‘little dove’ [œU-]

Sorbian /o

/[¨uo] or [o] is normally pronounced [

] in unstressed syllables:

(63) USorb rozpo

znac

‘recognize’ [¨r

znat



 ]

cf. po

znac

‘recognize’ [¨puoznat



 ]

And long Czech and Slovak vowels tend to be more extreme (higher, fronter, etc.)

than their corresponding short vowels, and especially in stressed (initial) position

(3.3.2.1). The most perceptible is /i/, whose short version tends to [



i] or [I], but these

and similar phonetic variations are generally not as perceptible as the major

3.3 Modern vowel systems 161

variations listed above. B/C/S long and stressed /e/ likewise tends to be higher than

short or unstresed /e/ (3.3.2.1).

3.3.2.7 Position in the word

Vowels may also be affected by their position in a word (other than in relation to

stress). Except in careful speech, the Polish nasal vowel /e˛/ normally loses its nasal

quality in word-final position:

(64) Pol prosze˛ ‘please’ [¨pr

And the nasal vowel /a˛ / may either preserve its nasal quality or lose it (both

standard), or give rise to a final -m (non-standard):

(65) Pol s

piewaja˛ ‘they sing’ [-j



], [-j



], [-j

Before consonants other than fricatives, the nasal vowels are regularly followed by

an epenthetic homorganic nasal consonant (3.4.2), and the vowel may lose its

nasality:

(66) Pol be˛de˛ ‘I will be’ [¨bf



ndf], [¨bfndf]

Other properties of word position which affect vowel phonetics involve adjacent

and surrounding consonants, and features of secondary articulation. Palatal or

palatalized consonants in particular may alter vowel quality, and not only in Czech

ehla

ska, or Polish and Sorbian vowel mutation in C

sequences, which are

matters of phonotactics (3.4.3). Even in Russian, surrounding palatalized conso-

nants can cause fronting of a vowel (3.2.1.5):

(67) Rus pja

tyj ‘fifth’ [¨p



ij] pjat

‘5’ [p

]

And in Slovenian: short /e/ > [e] before final /-j/: je

j [jej] ‘eat’

short /o/ > [o] before final [u

]: sto

l [stou

] ‘table’

The spoken language often diverges considerably from the official standard in

vowel pronunciation, and not only in dialects. Educated colloquial Prague

Czech, for example (11.3.2), regularly makes the following changes:

/

/ (orth. y

) > [ei

]: velky

mly

n ‘big mill’ [¨vflkei

mlei



e/ > [iø]: dobre

ho mle

ka ‘good milk’ [GenSg] [¨dobriøœ

¨mliøka]



u/ > [

]: u

ad ‘office’ [¨

at]

(in the written standard /ou/ is the regular morphophonemic partner of short /u/

in all but initial position, where long u



u/) is the only possibility, even for

etymological long /



o/).

162 3. Phonology

3.4 Modern consonant systems

3.4.1 Phonemes

The Slavic consonant phonemes show less diversity than the vowels. Table 3.8

presents a classification by manner and place of articulation, based on oppositions

in minimal pairs, of all the phonemes which appear across the standard languages;

sounds which are never phonemes, but only allophones, are enclosed in square

brackets (including semivowels). Many pairs of sounds in this set are mutually

exclusive, e.g. a language does not have both dental and uvular /r/, or both velar /g/

and glottal /h/. Foreign sounds are not indicated here, as there are no common

sounds which are foreign to the whole Slavic area; they are shown in table 3.10.

/r/ and /l/ may also be vocalic phonemes in many systems. /m/ and /n/ may also

be vocalic allophones in a few systems (3.3.1), and all four of these may be devoiced

(3.4.3.4).

Nineteen of the phonemes are common to all the languages (table 3.9).

Table 3.8. Slavic consonant phonemes

Labial Dental

Post-

Alveolar

Alveolo-

Palatal Palatal

Velar/

Uvular Glottal

Stop p p

t’ k’ k k

[?]

d’ g’ g g

Nasal m m

n’ [l]

Fricative f f

zˇ zˇ

z´ ˜˜

(or Semivowels) w w

]j[i» ]

Affricate c c

dz dz

dzˇ dzˇ

dz´

Liquid r r

r’ [

] 

]

l’

Table 3.9. Consonant phonemes common to all Slavic languages

Labial Dental/Alveolar Post-Alveolar Palatal Velar Glottal

Stops p b t d k (g)

Nasals m n

Fricatives (v) (w) s z s

zˇ jx(˜) (h)

Affricates c c

Liquids r 1

3.4 Modern consonant systems 163

Table 3.10. Non-common phonemes; allophones; distribution of /w-v/ and /g-

-h/

Labial Dental/Alveolar Post-Alveolar Alveolo-Pal Palatal Velar Glottal

Blg v; H/S all; f H/S all; (dz) [dzˇ ]g(˜)(l)

(Some descriptions interpret soft as ‘palatalized þ j’, see 3.4.1.1) H/S k/g

Mac v f l

(dz) [dzˇ ]k’g’n’g(˜)(l)

B/C/S v [f] (dz) [dzˇ ]g(˜)(l)

Sln v [f] (u

) (dz) [dzˇ ]g(˜)(l)

Pol v f; H/S all (dz) (l

) [dzˇ ]s

z´ c

dz´ n’ g (˜)(l)

w (H only) (< PSl l)

)(g

)

USorb w; H/S all H/S n/c; (r > uvular) dzˇ (l) H/S  h

(w < v/l) [v] [f] (dz) (˜) [g] (?)

LSorb As USorb H/S n/l/r (dz) (dzˇ )s

z´ c

dz´ g(˜)(l)

)(g

)

Cz v f (dz) r

] (dzˇ )t’d’n’(˜)(l) [g] h (?)

Slk v [f] (u

)dzdzˇ t’ d’ n’ l’ (˜)(l) [g] h

Rus v f; H/S all H/S all exc. c (dz) (dzˇ )g(˜) H/S all

Bel v [f] (u

) H/S n/s/z/c; H/S dz dzˇ ˜ [g] H/S all

H/S all exc. u»

Ukr w (v) [f] H/S all; [dz] dz

dzˇ [g] H/S all h

H/S all

Notes:

‘H/S’ ¼ ‘Hard/Soft’, i.e. Palatalized; ‘all’ ¼ the common sounds listed in table 3.9 in that place of articulation; round brackets indicate

allophones, square brackets foreign

The brackets in table 3.9 mean that all languages have either /v/ or /w/; and either

/g/, /˜/ or /h/. The distribution of these and the main remaining parts of each

language’s system are shown in table 3.10.

All Slavic languages contain sizeable quantities of foreign words, which have

been integrated into the native Slavic phonological systems to differing degrees.

Sometimes, as with the Slovak diphthongs /au eu/ (3.3.1), they occur only in

obviously foreign words, and so remain present but marginal in the language’s

phonemic system. But there comes a point where the quantity of such words makes

this treatment insufficient. Determining this point often depends on distribution.

We discuss these phonological processes in 3.4.3 below, and illustrate some of the

problems by the feature of palatalization.

3.4.1.1 Palatalization

The most highly palatalized languages in Slavic are East Slavic, Sorbian and Polish,

and in some descriptions Bulgarian. There are increasingly severe limitations on the

distribution of palatalized consonants as we move from Russian to Belarusian and

Ukrainian, and further to Bulgarian, especially the ability of palatalized conso-

nants to occur in final position (3.4.3.4).

Of the paired hard/soft consonants in final position, Polish has only the palatal

nasal /n’/. But Polish shows a more advanced state of phonetic palatalization than

any of the other languages. This fact is mirrored in the analyses of S

aumjan (1958)

and Koschmieder (1977), who give Polish a full double series of hard and soft

consonants:

Paired: Hard: p b t d k g f v s z n m r l c

Soft: p

Unpaired: Hard: s

zˇ xc

dz dzˇ

Soft: s

z´ c

dz´

for which Koschmieder supplies minimal pairs like:

(68) Pol wara ‘beware!’ wiara ‘belief’

mara ‘apparition’ miara ‘measure’

wiza ‘visa’ wizja ‘vision’

pana ‘man’ (GenSg) piana ‘foam’

The opposite analysis is proposed by Jassem (1964: 335) and Bidwell (1969), who

re-analyse most of the paired series as ‘consonant þ /j/’, with the sole exception of

/. Bidwell cites phonetic data which show a pronunciation of ‘consonant þ [j]’

for historically palatalized consonants:

(69) Pol piasek [¨pjasek] cf. Rus peso

k [p

h¨s

k] ‘sand’

3.4 Modern consonant systems 165

The phonetic data do not square with Koschmieder’s palatograms, which show

clearly the palatalizing influence of the /j/ on preceding consonants, and so favor

the /C

/ analysis. From the typological point of view, the analysis in two paired

series makes Polish look more like East Slavic. But the analysis with ‘consonant þ

/j/’ makes it easier to relate Polish progressively to Sorbian, and thence to Czech

and Slovak, on a scale of declining palatalization. The unpaired soft phonemes

noted by Koschmieder (s

z´ c

dz´ ), while phonetically palatal, are the soft alternants

of the hard /s z c dz/ (e.g. gro

d  grodu  grodzie ‘castle’ [NomSg  GenSg 

LocSg]: /t/  /d/  /dz´ /); the same applies to /n’/, phonetically palatal, but the soft

partner of /n/ (e.g.

zona 

zonie ‘wife’ [NomSg  LocSg]: /n/  /n’/). In this view,

written forms with j (e.g. wizja) – all foreign in origin – represent a sequence of

‘consonant þ /j/’, in which the consonant is allophonically palatalized.

3.4.2 Phonetics

At the phonetic level – that is, including allophones – we can measure each

language as a sub-set of the maximum number of phones occurring across the

whole group, expressed in rounded percentage terms, with the following results

(since different descriptions will produce different numbers of phonemes, this

comparison is possible only for the phonetic level):



the range is 34–59 percent;



lowest is Slovenian, highest Ukrainian (and Russian in one standard

variant (with both soft and hard post-alveolar fricatives, see 11.3.1 –

noted below as Russian

; Russian

has only hard post-alveolar

fricatives);



under 40 percent: Slovenian (34 percent), B/C/S (35 percent), Slovak

(39 percent);



40–50 percent: Macedonian (40 percent), Upper Sorbian (41 percent),

Czech (44 percent), Lower Sorbian (49 percent);



over 50 percent: Polish (51 percent), Bulgarian (54 percent),

Belarusian/Russian

(56 percent), Ukrainian/Russian

(59 percent).

South Slavic – except for Bulgarian – and Slovak are lowest, and East Slavic highest.

The main reason for the differences lies in the extent of the feature of palatalization.

Beyond the basic phonetics of Slavic sounds, which are indicated sufficiently by

the phonemic descriptions in tables 3.8–3.10, the process of palatalization and the

palatal region itself are the area of greatest differentiation.

As noted above, the palatalization of non-palatal consonants may vary, on the

one hand, between ‘‘genuine’’ co-articulation palatalization – as in Russian and

166 3. Phonology

Belarusian, and, on the other hand, articulations closer to a sequence of

‘consonant þ /j/’ – as in Ukrainian, Bulgarian and Sorbian. In the latter areas,

palatalization is largely allophonic, occurring only (and automatically) before

front vowels and /j/ (as for Polish in foreign borrowings, above), and before back

vowels it may be interpreted as phonemic or as a sequence. Elsewhere, it is often

very limited in its range of activity, e.g. only in dentals, or not in final position.

In the palatal region, stops are rare, confined to Czech and Slovak in West Slavic

(with /t’/, /d’/, /n’/, as well as /l’/) and Macedonian in South Slavic (with /k

/ and /g

/),

the latter alone reflecting the Proto-Slavic palatal stops resulting from /t, d þ j/

(3.2.2.1). As for the fricatives, the Proto-Slavic ‘‘palatals’’ /s

/ and /zˇ / are reflected

everywhere as post-alveolars (also called ‘‘palato-alveolars’’), mostly with low

tongue position (i.e. ‘‘hard’’, e.g. East Slavic, Polish), sometimes high (i.e.

‘‘soft’’, e.g. Upper Sorbian). Only Polish and Lower Sorbian have developed new

alveolo-palatal fricatives (/s

/, /z´ /), from previous palatalized /s

/, /z

/. The affricates

are slightly different, partly because the voiced one (/dzˇ /) is frequently missing.

They are also normally post-alveolar, but more often have high tongue position

(‘‘soft’’, e.g. Russian), though low tongue position (‘‘hard’’) is also found

(Belarusian, Ukrainian). Polish and Lower Sorbian have developed the alveolo-

palatal affricates /c

/, /dz´ / from palatalized /t

/, /d

/. B/C/S also has alveolo-palatal

/, /dz´ /, but from /t, d þ j/ (/t’ d’/).

3.4.3 Phonotactics and limitations on distribution

Limitations on the appearance of particular phonemes are most often related to

clusters and word- or syllable-final position. Only a few are related to CV contexts,

and we treat them first.

3.4.3.1 Palatalization

Among the languages with a hard/soft opposition, this opposition is often neutral-

ized before certain vowels, usually the front vowels /i/ and /e/. In the case of /i/, it is

normal for such languages to have two audible variants: [i] (/i/) after soft (palatal-

ized) consonants, and [



i] (/i/ or /y/) after hard consonants. This allows the hard/soft

consonant opposition to operate (as in East Slavic, Polish, Sorbian). The absence

of the second variant tends to prevent it (as in Bulgarian), though some languages

oppose hard/palatal: Czech /t d n/, Slovak /t d n l/ and Polish /n/ before /i/, and

Slovak also before /e/. In most cases, and normally in the case of /e/, the soft

member of the pair occurs here automatically (as in Russian, Belarusian, Polish,

Sorbian). Bulgarian and Ukrainian have only hard consonants before /e/, and

Bulgarian also before /i/ (in each case the velars are exceptions, the soft variant

3.4 Modern consonant systems 167

occurring before both vowels). But the soft velars are rarely freely opposed to hard

velars, being normally positional variants, even in languages like Russian with a

strongly developed hard/soft opposition.

3.4.3.2 Assimilation

The general rule is that consonants within a cluster will adopt the voice and place of

articulation of the final member of the cluster. This situation arose at the end of the

Proto-Slavic period with the reappearance of closed syllables and the formation of

new clusters (3.2.3.1), both of which were brought about by the loss of the jers

(3.2.1.1). Thus, while assimilation represented a broad trend, not every area or

language was certain to follow it.

1. Assimilation of voice

Regressive assimilation of voice among obstruents is the norm throughout:

voiced obstruents are devoiced before voiceless obstruents, and voiceless

obstruents are voiced before voiced obstruents (for word-final position, see

3.4.3.4). The single exception in the standard languages is Ukrainian, in which

the second rule applies, but not normally the first:

(70a) Rus, Bel no

zˇka ‘foot’ [¨n

Ska]; Ukr nı

zˇka [¨n

idka]

(70b) Rus, Bel, Ukr pro

ba ‘request’ [¨pr

ba]

Ukrainian optionally devoices only the final /z/ of prefixes before a voiceless

root-initial obstruent, e.g. ro

z-kvit ‘daybreak’ [-skv-] or [-zkv-]. The prefix z- is

normally devoiced: z-c

ˇy

styty ‘to clean’ [StS



-] (assimilation of voice and place).

It is in this general situation that allophones otherwise not phonetically

present may appear, e.g. [dz], [dzˇ ], [˜] as voiced versions of the voiceless

phonemes /c/, /c

/, /x/ (table 3.10).

Sonorants do not participate in these rules of assimilation: normally they do

not themselves devoice and they do not cause voicing of preceding obstruents.

The common sonorants /r l m n/ stand apart in every language except Slovak,

at least within a word. In Proto-Slavic, clusters with sonorants second were

common, and they continue to be so and to show no assimilation in the

modern languages:

(71) Rus tri ‘three’ [tr

i] sloj ‘layer’ [sl

j] sneg ‘snow’ [sn

fk]

In Slovak, /m/ and /n/ at least may cause voicing where internal morpheme

boundaries are clear:

(72) Slk tak-mer ‘almost’ [¨tagmfr] nas

-mu ‘our’ [DatSgMasc] [¨nadmu]

168 3. Phonology