Sussex R., Cubberley P. The Slavic Languages

Подождите немного. Документ загружается.

to the removal of one, and this may explain the absence or loss of /s

/ in West Slavic

or that of /dz

/ almost everywhere, though the picture outside West Slavic is untidy.

Consider the reflexes of Proto-Slavic *noge

‘foot, leg’ LocSg:

(35) PSl noge

‘leg’ LocSg

ESl: OR no(d)ze

(Rus noge

) Ukr nozı

Bel naze

SSl: OCS nodze

Blg noze

[Pl] Mac noze [Pl] B/C/S no

zi (Sln no

gi)

WSl: Cz noze (Slk nohe) Pol nodze Sorb noze

Polish is now the only language with /dz/ from this source, and it is clear that many

languages converted /dz/ to /z/ only later.

/t/ and /d/ þ /j/ Matching the de-affrication of /dz/ to /z/ is that of /dzˇ /to/zˇ /,

which has occurred also in many areas. The many reflexes of /tþ j/ and /dþ j/

indicate that they were reached after the Proto-Slavic period, which is why it is

simpler to say that late Proto-Slavic still had (unchanged) palatal stops /t’/ and /d’/,

which only after the break-up of Slavic unity gave their modern results, which are

not always a related articulatory pair. (The affrication of these to /c

/-/dz

/ is also

possibly Proto-Slavic.) A second source of the same reflexes as /tþ j/ are the clusters

/kt/ and /gt/ before front vowels, occurring mainly on the boundary between (velar)

verb stem and infinitive ending or deverbal suffix. These had both become /t’/ in

Proto-Slavic. Reflexes are unified for East (36a) and West (36b) Slavic, but not for

South Slavic (36c–f ). There is some dispute about whether the last examples of

(36a) in Ukrainian and Belarusian represent an old /dzˇ / or a new one, since this

form appears only in deverbal forms. On balance, the ‘‘old’’ argument is rather

more convincing (see Carlton, 1990: 140).

(36) PSl tj/dj ( > t’/d’)

PSl sve

-tja ‘candle’ xot-jo˛ ‘I want’ pek-ti ‘to bake’ (or pek-t{ ‘oven’)

mog-ti ‘to be able’ (or mog-t{ ‘power’) med-ja ‘boundary’

sad-ja ‘soot’ xod-jo˛ ‘I go’

(36a) ESl: /t’/ > /c

/ and /d’/ (> /dzˇ /) > /zˇ /

Rus svec

xoc

pec

(N, V) moc

mezˇa

zˇa xozˇu

Ukr svic

upe

i (mohty

) mezˇa

zˇa xodzˇu

Bel sve

ka xac

pjac

(mahc

) mjazˇa

zˇa xadzˇu

(36b) WSl: /t’/ > /c/ and /d’/ > /dz/ ( > /z/)

Cz svı

ce chci pe

ci moci mez(e) saze (chodı

Slk svieca chcem piect’ m

oct’ medza sadza (chodı

Pol s

wieca chce˛ piec mo

c miedza sadza chodze˛

3.2 Historical evolution and modern equivalences 139

USorb swe

ˇca

chcu pjec mo

c mjeza sazy chodz

´u

LSorb swe

ca cu pjac moc mjaza caze chojz

(36c) B/C/S: /t’/ > new /c

/ and /d’/ > new /dz´ /

B/C/S sve

axo

upe

imo

ime

dja (d. sadja )(ho

dı¯ m)

(36d) Sln, B/C/S (Kajk): /t’/ > /c

/ and /d’/ > /j/

Sln sve

aho

em pe

imo

ime

ja sa

je (ho

dim)

(36e) OCS, Blg: /t’/ > /s

t/ and /d’/ > /zˇ d/

OCS sve

ta xos

to˛ pes

ti mos

ti mezˇda (sazˇda) xozˇdo˛

Blg sves

ta pes

t (N) mos

t (N) mezˇda

zˇda (xo

dja)

(36f) Mac: /t’/ > /k

/ and /d’/ > /g

Mac svek

e – mok

(N) meg

asag

a (odam)

It would be attractive to see the Macedonian reflexes – which are those of its

northern dialects – as old, as this would mean that they had moved least of all

the languages, since /k

/ and /g

/ are in effect palatal stops, and this approach is

taken by some (e.g. Friedman, 1993: 256). The traditional view, however, is that

they are, in fact, new, that is, they have developed later, from either: (1) affricate

forms (/c

/-/dz

/or/c

/-/dz´ /, seen as the initial common South Slavic reflex), which

then again lost the affrication (e.g. Berns

tejn, 1961: 228; Shevelov, 1964: 215); or

(2) the /s

t/-/zˇ d/ of the southern dialects through influence of Serbian /c

/-/dz´ /,

ending up in the central dialects in a lowest common denominator form

(Koneski, 1967: 81; Carlton, 1990: 139). Variant (2) is based on the evidence of

Old Macedonian (OCS) texts as well as the dialectal results (both centering on the

Ohrid area). However, it seems more plausible that at least some dialects never

went through the Bulgarian pattern, but either retained the original stops or went

through the affricate stage (variant [1] ). The result in B/C/S (C

akavian) – /t’/ > /t’/

and /d’/ > /j/ – indicates a similar effect for /t’/ at least.

Labials þ /j/ The most common, and possibly universal, result of this group was

the appearance of an epenthetic /l’/: the palatal element of the labial articulation

apparently produced a liquid element strong enough to be generalised. The extra

/l/ is now a regular part of the morphophonology of all East Slavic and the western

South Slavic languages (B/C/S and Slovenian), as can be seen in (37a):

(37) Labial þ /j/

PSl zem-ja  zem-{sk-/zem-{n- ‘earth’ [NomSg]  Rel Adj stem

ljub-iti  ljub-jo˛ ‘love’ [Inf  1Sg]

(po)stav-iti  (po-)stav-jen- ‘put’ [Inf ]  PPP stem

140 3. Phonology

(37a) (i) with /l’/

Rus ( þ Ukr/Bel) zemlja

msk- ljubı

ljublju

posta

vit

posta

vlen-

B/C/S (þ Sln) ze

mlja ze`` mn- sta%viti sta%vljen-

(37b) (ii) without /l’/:

Blg (þ Mac) zemja

zemn- posta

vja (Blg [1Sg] ) posta

ven-

Cz (þ Slk) zeme

zemsk- postavit postaven-

Pol (þ Sorb) ziemia ziemsk- lubic

lubie˛ postawic

postawion-

The case for the /l’/ having been general, and then removed in some areas, rests on a

few root forms, including place-names, in group (ii). The most consistent lexeme is

that for the verb ‘spit’, with the Proto-Slavic root *peu

- > pju-, with the reflex plju-

in every area:

(38) PSl root peu

- > pju- ‘spit’

(38a) Rus, Ukr, Bel, B/C/S, Sln pljun-

(38b) Blg pljuv- Mac pluk- Cz plivn-

Slk pl’u(v)- Pol plun-

Sorb plun-/pluw-

The age of this inserted /l’/ is perhaps evidenced in its appearance in non-/j/ forms,

already in the Proto-Slavic form *pl{v- (Rus pleva

, Pol plwac

), which might

suggest that the epenthetic /l’/ is here something other than the effect of /j/, possibly

a sort of onomatopoeia. But this could also simply be a later derived form, since it

occurs only in Russian, Belarusian and Polish, the rest having pljuv-.

The only other common lexeme with a general /l’/ is b(l )judo ‘dish’, a borrowing

from Germanic, which must have acquired the /l’/ in early Proto-Slavic, and was

possibly borrowed into East Slavic first (Rus blju

do, Sorb blido , Blg, Mac blju

do).

Perhaps the relevant factor in both of these is the word-initial position, since it

has no alternation with a simple consonant, and so the trend towards /l’/ was

unhindered (Shevelov, 1964: 221).

One other argument for the generality of epenthetic /l’/ lies in OCS texts from the

Bulgarian/Macedonian area, where /l’/ does appear, and more often in older than

in younger texts. However, the earliest texts are still from the eleventh century, by

which time Russian and Serbian versions would have had the /l’/, so that this

evidence is also unreliable.

A probable explanation (Shevelov, 1964) is that there was hesitation or inter-

changeability in the presence of the /l’/. This was reflected directly in the mixed

OCS evidence, which was then eliminated according to either phonetic motivation

(giving /l’/) or morphological (giving variation by individual word, favoring the

initial position and allowing place-names to remain unaffected).

3.2 Historical evolution and modern equivalences 141

An alternative result for /m/ þ /j/ in non-/l’/ areas is /mn’/, seen in some

Bulgarian dialects in zemnja

(for standard zemja

). This result is also seen in more

modern form in the Czech realization of the same sequence, e.g. me

ˇsto

‘town’ is

phonemically /mjesto/, but phonetically [¨młfst

/r/ þ /j/ Of the three dental sonorants – /r/, /l/ and /n/ – the last two do not present

major problems in respect of a following /j/. They are continuants, and so easily

became palatal without losing their basic quality, and their reflexes throughout

either remain palatal or, much more commonly, join with the positionally soft

equivalents (see below, 3.2.2.2).

/r/ was more complex. The addition of /j/ would have made for a truly complex

sound, one which has not survived as a palatal /r’/in any area. The range of reflexes

is: (1) /r/þ/j/, that is, a sequence of two phonemes (Ukrainian, Sorbian); (2) /r

that is, palatalized alveolo-dental (Russian); (3) hard /r/ (South Slavic, Belarusian,

Slovak); (4) (post-)alveolar fricative /r

/ (Czech), which has been reduced to (5) a

simple fricative /zˇ / in Polish (with the etymology reflected in the spelling rz), both

devoiced to [r

] and [

] respectively in certain contexts (3.4.3.2). Slovenian has a

mixture of (1) and (3), and Sorbian of (1), (3) and (4), including in the last

the unique reflex /s

/ (after /t/ in Upper Sorbian, after /p t k/ in Lower Sorbian)

(ibid.). No distinction was maintained anywhere between palatal /r’/ and palata-

lized /r

/ (39).

(39) /r/ þ FV, /r/ þ /j/

/: PSl re˛dy ‘row’ re

ka ‘river’ tvar{ ‘creature, thing’ tri ‘three’

/r’/: PSl morje ‘sea’ burja ‘storm’ le

k-arj{ ‘doctor’ (agentive suffix)

B/C/S/Sln re

dre

ka tva

rtr

re/morje

ra/bu

rja lje

kar/leka

Blg/Mac red reka

/re

ka tvar tri mo

re bu

rja/bura le

kar

Rus rjad reka

tvar

tri mo

re bu

rja le

kar

(arch.)

Ukr rjad rika

tvar try mo

re bu

rja lı

kar

Bel rad raka

tvar try mo

ra bu

ra le

kar

Cz r

eka tva

i [r

] mor

e bour

ele

kar

Slk rad rieka tva

r tri more bu

ra leka

Pol rza˛ d rzeka twarz trzy [

] morze burza lekarz

Sorb rjad re

ka twar tr

i [si] morjo – le

kar

3.2.2.2 Palatalized (soft) consonants

Within the Proto-Slavic syllable, the harmony between consonant and vowel

(1.3.1.3) meant that consonants before front vowels were automatically palatalized

to some extent wherever this was possible (e.g. not palatals, which were inherently

142 3. Phonology

soft; or velars, which had earlier become palatals in such a context). When the

syllable restructure occurred, consonants could occur without a following vowel

and became independent phonemes. At this point the languages could either retain

or not retain the underlying palatalization of consonants previously followed by a

weak front jer. At first there was probably some retention everywhere; then the

status of such consonants was sorted into phonemic or allophonic, or they were

removed entirely. In East and West Slavic phonemic palatalization first developed

strongly, then was reduced in extent in several areas (Czech/Slovak and

Ukrainian), but kept in the rest (Russian/Belarusian, Polish/Sorbian). In South

Slavic it developed weakly in Bulgarian/Macedonian, and not at all in B/C/S and

Slovenian.

Another context in which palatalization could become phonemic prior to the

loss of the jers was where the front nasal was denasalized to [a

]. This was then seen

as an allophone of /a/, meaning that there was an opposition between ‘hard

consonant þ old /a/’ and ‘soft consonant þ new /a/’. However, while the syllable

was still open – which it was until the loss of the jers – the true opposition

can be interpreted as between hard and soft syllables, in this case between

‘hard consonant þ [a]’ and ‘soft consonant þ [a

]’. So while the denasalization

provided a potential case for phonemically soft consonants, they really became

established only with the loss of the jers. In a similar way, where /e

/ became /a

/-/a/ –

i.e. in the areas which did not denasalize – /e

/ would have provided the

same impetus, but would have remained unrealized before the loss of the jers,

and after the loss, would simply have reinforced the new situation. The extent

of the hard/soft opposition in different areas will be taken up in the ‘‘modern’’

section (3.4.1).

3.2.2.3 Velars (g/˜)

The articulation of the voiced velar consonant in Proto-Slavic is traditionally said

to have been initially occlusive [g], which then became fricative [˜] or laryngeal [œ]

in many areas (‘‘lenition’’). We cannot say whether this is so, or whether some areas

always had the fricative version, but the areal division is certainly very old, and

looks as follows:

[g]: ESl: N-Rus WSl: Pol, LSorb SSl: all standard languages

[˜]: ESl: S-Rus, Bel SSl: NW-Sln, NW-C

[œ]: ESl: Ukr WSl: Cz, Slk, USorb

One interpretation of this picture is that the lenition was a feature of the centre of

the Slavic area, perhaps earliest in the third ( [œ] ) group, which may have reached

the laryngeal [œ] via the velar [˜].

3.2 Historical evolution and modern equivalences 143

Within the system, there was no pressure either way, since the velar set was

tripartite, with /k  g  x/. From one view-point, /g/ was odd by being voiced,

leaving /k/ and /x/ as a voiceless pair opposed as stop  fricative; from another, /x/

was odd, with /k/ and /g/ opposed as voiced  voiceless. The articulation of /g/

therefore simply promoted one or other of these equally valid scenarios, going

better with /k/ if a stop, better with /x/ if a fricative.

Given that the fricative variant was early, in relation to the palatalizations of the

velars, it is more likely that its presence was the reason for the different results for

/g/ – either /dzˇ /or/zˇ / for PV1, /dz

/or/z

/ for PV2 and PV3. The fricative version

would have produced directly a fricative palatal /zˇ / matching the result of /x/ (/s

/),

and the stop version an affricate /dzˇ / matching that of /k/ (/c

/). The uncertainty

in the (Old Russian) textual evidence may reflect the presence of the two variants

in important areas like N-Rus vs S-Rus þ Ukr þ Bel. In other words, we do not

have to posit the traditional ‘de-affrication’ of /dzˇ / and /dz

/ to account for the

modern /zˇ / and /z

/ (see also 1.3.1.5).

3.2.2.4 Labials (v/w)

A similar division occurs across the area in respect of the bilabial vs labio-dental

articulation of /v/ (which range we shall refer to as /w-v/ ). Again an early division is

certain, and we cannot be sure that there was initially a unified version which split.

While earlier Proto-Slavic, and probably PIE, would have had [w], traditionally, [v]

is said to be the starting-point, at least in Late Proto-Slavic, reducing – again by a

sort of lenition – (back) to [w] (probably actually [B], a bilabial fricative, but we will

follow the convention and use [w] ). It is true that in some modern languages an

underlying [v] alternates directly with [u

] (e.g. Belarusian), but one might equally

suggest that it is the [v] which is new (e.g. in this case through the influence of

Russian). The distribution is as follows (for details of distribution and allophones

in each language, see 3.4.1):

[v] devoiced to [f ] before a word boundary or voiceless obstruent

(see 3.4.3.2): Blg, Mac, Rus, Cz, Pol

[v] not devoiced: B/C/S

[v] becoming [u

] before a word boundary or consonant: Sln,

Bel, Slk

[v]/[w] becoming [u] initially before a consonant (Bel) or between

consonants (Ukr)

[w] becoming [u

] before a word boundary or consonant: Ukr, Sorb

[w] becoming [v

] before [i]: Ukr

144 3. Phonology

Examples showing the open (pre-vocalic) form vs the closed (pre-consonantal or

pre-pausal) form are given in (40) with the /w-v/ in phonetic transcription. The

orthography frequently disguises the pronunciation, some using v to represent [w],

others w to represent [v], etc.:

(40) /w-v/

PSl volja ‘will’ v{dova ‘widow’ gne

vy ‘anger’

prav{da ‘truth’ lavyka ‘bench, counter’

Blg/Mac [v]o

lja [v]do

[v]ı

ca gnja[f ]/gne[f ] pra

[v]da la

[f ]ka

B/C/S [v]o%lja [u`` ]dova gn( j)e

[v] pra

[v]da –

Sln [v]o

lja [w]do

[v]a gne

] pra

]da la

]ka

Rus [v]o

lja [v]do[v]a

gne [f] pra

[v]da la

[f ]ka

Ukr [w] o

lja [u]do[w]a

hni[u

] pra

]da la

]ka

Bel [v]o

lja [u]da[v]a

hne[u

] pra

]da la

]ka

Cz [v]u

le [v]do[v]a hne

[f ] pra[v]da la

[f ]ka

Slk [v]

ol’a [v]do[v]a hne[u

] pra[u

]da la[u

]ka

Pol [v]ola [v]do[v]a gnie[f] pra[v]da ła[f ]ka

Sorb [w]ola [wu]do[w]a hne

] pra[u

]da ła[u

]ka

(See also the examples of prothetic /w-v/ noted in 3.2.1.4.)

3.2.2.5 Quality of /l/

Again as a result of the phonemicization of consonants and vowels out of the

earlier syllabic units (1.3.1), the new hard and soft versions of /l/ caused problems.

The nature of the ‘‘hard syllables’’ of Proto-Slavic was probably not simply

neutral (vs palatalized for the ‘‘soft’’ ones), but a positive (labio-)velarization.

In the subsequent development of the hard  soft opposition, the hard members

became the unmarked ones, with palatalization being the marked feature. But in

/l/ the velarization remained in many areas, probably because a neutral [l] – the

basic ‘‘European’’ one – was not distinct enough from the soft. So where there is

still an opposition, it is normally between a strongly palatalized and a strongly

velarized version (East Slavic) or between a medium [l] and a velarized one (Blg,

Mac), the latter having sometimes gone on to lose its lingual contact and become

[w] (Pol, Sorb). The opposition may also be between medium and palatalized

(or palatal), whether as separate phonemes (Slk, B/C/S, Sln) or allophones (Pol).

In some areas there is a single medium /l/ from both sources (Cz). Where the /l/

is velarized, it may be reduced by context only to [u

] (Ukr, Bel, Slk, Sln), or even

to vocalic [

] (B/C/S).

In the examples in (41), we again use phonetics for the articulation of the former

/l/: [l

‘

] for velarized, [l

] for palatalized, [·] for palatal:

3.2 Historical evolution and modern equivalences 145

(41) /l/

PSl lyby ‘forehead’ le

sy ‘forest’ listy ‘letter, page’

ljud{je ‘people’ Pl stoly ‘table’

Rus [l

‘

]ob [l

]es [l

]ist sto[l

‘

][l

]udi

Ukr [l

‘

]ob [l

]is [l

‘

]yst sto[u

][l

]udy

Bel [l

‘

]ob [l

]es [l

]ist sto[u

][l

]udzi

Cz [l]eb(ka) [l]es [l]ist stu

[l] [l]ide

Slk [l]ebka [

]es [

]ist sto[l] (coll. [u

]) [

]udia

Pol [w]eb [l] as [l

]ist sto

] [l]udzie

Sorb – [l]e

s [l]ist sto[u

] [l]udzio

Blg/Mac [l

‘

]ob [l]es [l]ist sto[l

‘

][l

]ude/[l]ug

B/C/S – [l]es [l]ist sto

(< oo) [l

]udi

Sln (arch) [l]eb [l]e

s [l]ı

st sto

] [l] judje

3.2.2.6 Syllabicity of /r/, /l/

In 1.3.1.7 we opted for the view that Proto-Slavic developed syllabic liquids, and so

now we must consider the extent to which these were retained or restructured after

the break-up.

As syllables again became closed through the loss of the jers, all areas had

to cope with the new clusters previously separated by a weak jer, including

those with /r/ or /l/ (types CRyC, CR{C, where ‘R’ means /r/ or /l/). Many

areas extended existing syllabic /r

/ and /l

/ to these contexts, while others did

not allow the weak jer to disappear. The first group thus strengthened the

range of the syllabic liquids, while the second tended to abandon even the

existing range. But there is, in fact, quite a range of treatments between these

two extremes.

The extreme ‘‘syllabic’’ type is seen in South Slavic generally and in (Central)

Slovak, with only two results from the twelve possibilities (four each of syllabic,

strong jer and weak jer – see (42) ). At the other extreme are Ukrainian and

Belarusian, with eleven results from the twelve. In between, it appears that the

rest of West and East Slavic maintained some syllabic forms for a time, and then

abandoned them. The Lekhitic group (Polish, Sorbian) accepted consonantal

clusters (involving consonantal /r/ and /l/), and both they and Czech converted

the syllabic segments into a range of vowels dependent on the surrounding context.

Later on, the range of syllabic forms was reduced in different ways even in the

‘southern’ group, especially /l

/, which survived only in Czech and Slovak; /r

continues not only in Czech and Slovak, but also in B/C/S and Macedonian, and

146 3. Phonology

structurally also in Bulgarian and Slovenian, though here it is phonetically (and

probably phonemically) [

] þ [r] (3.4.1–2). In (42a) we find:

a. the example of Pol wierzch, in which the rz reflects a former soft or

palatal /r/, is believed to indicate that initially Polish – and thus

perhaps others – solved this problem by inserting a vowel after the

/r/, in the manner of syllables with a full (i.e. non-jer) vowel

(cf. polnoglasie, 1.3.1.7.). However, the retention of the softness of

syllabic /r

/ could also account for this effect;

b. the various inserted vowels in Polish and Czech are related to the

articulation of the surrounding consonants;

c. /c

/ > /c/ before /r

/ is a special feature of B/C/S and Macedonian;

d. Bulgarian normally inverts the /r/ and /

/ in (new) closed syllables (to

avoid final clusters).

(42) Syllabic /r

/, /l

(42a) CR

C (or C þ jer þ R þ C):

PSl: gr

dlo ‘throat, neck’ vr

xy‘summit’ c

ny(j{) ‘black’

sty ( j{) ‘fat’ vl

ky ‘wolf’ zˇ l

ty( j{) ‘yellow’

Blg ga˘

rlo vra˘x c

ˇe

ra˘n tla˘st va˘lk zˇa˘lt

( [Def] va˘rxa˘

Mac grlo vrv crn – volk zˇolt

B/C/S/Sln gr`` lo/gr

lo vr

h/vr

hcr

n/c

ntu`` st/to

lst vu

k/vo

lk zˇu

t/zˇo

Rus go

rlo verx c

ˇe

rnyj to

lstyj volk zˇe

ltyj

Ukr ho

rlo verx c

rnyj tovsty

j vovk zˇo

vtyj

Bel ho

rla verx c

ˇo

rny to

sty vou

kzˇo

Cz hrdlo vrch c

erny

tlusty

vlk zˇluty

Slk hrdlo vrch c

ierny tlsty

vlk zˇlty

Pol gardło wierzch czarny tłusty wilk _zo

łty

USorb hordło wjerch c

orny tołsty wjelk zˇołty

LSorb gjardło wjerch carny tłusty wjelk zˇołty

In (42b) we see:

a. the Belarusian examples with /ry/-/ly/ in unstressed syllables (and

Ukrainian has similar ones, as well as /lu/ in jabluko) are taken to

indicate a preceding stage of syllabic /r

b. there are examples in East Slavic of a third possible strategy for

avoiding undesirable clusters: to elide the sonorant as well as the

weak jer , e.g. the town name ORus Pl{skovy has become Pskov.

3.2 Historical evolution and modern equivalences 147

(42b) C þ R þ jer þ C (strong/weak sets)

PSl: kryv{/kryvi ‘blood’ [NomSg/GenSg] kr{sty ‘cross’

kr{stiti ‘to baptize’

plyt{/plyti ‘flesh’ [NomSg/GenSg] (j)ablyko ‘apple’

sl{za ‘tear’

(Note: /r

/ (phonemic) ¼ syllabic; [r

][r

] (IPA) voiceless; [r

]

(IPA) ¼ syllabic)

Blg kra˘v kra˘st kra˘

stja pla˘t – – ja

ba˘lka sa˘lza

Mac krv – krst krsti – – jabolko solza

B/C/S kr

vkr`` vi kr`` st kr`sta pu`` tpu`` ti ja`` buka su

Sln krı

krvı

st kr

sta po

lt po

lti ja

bolko so

lza

Rus krov

kro

vi krest krestı

plot

plo

ti ja

bloko sleza

Ukr krov kro

vi xrest xresty

ti plot

plo

ti ja

bluko sl

oza

Bel krou

kryvı

– chryscı

ploc

plo

ci ja

blyk sljaza

Cz krev krve [r

] kr

est [r



] kr

tit [r



] plet’ pleti jablko slza

Slk krv krvi [r

] krst krstit’ plet’ pleti (lit.) jablko slza

Pol krew krwi [r

f] chrzest [

] chrzcic

[

] płec

płci jabłko [pk] łza

USorb krej krwe

[rw] –kr

[

] – – jabłuko sylza

LSorb ks

ej ks

–ks

(arch.) – – jabłuko łdza

In summary, of the twelve contexts we find the following distribution of results

(for further details see below, 3.4.1, modern systems):

South Slavic and Slovak:

1. /r

/ for all forms with /r/ and /l

/ for all forms with /l/;

2. Slovenian and Macedonian convert /l

/ to /ol/; B/C/S converts it to /u/;

3. Bulgarian converts both to sequences of ‘consonant þ /

/’, the order

being controlled by word shape: avoidance of syllable-final clusters

causes /

rCC-/ to become /r

CC-/, and /

lCC-/ to become /l

CC-/.

Czech: retains /r

/; converts /l

/ to /lu/ except in the limited context after labials

(e.g. vlk); retains the distinction between strong and weak jer, with the latter giving

syllabic forms, except that soft /r

/ > consonantal /r

/aftervelarsandlabials(e.g.kr

tu).

Polish and Sorbian: retain the distinction between the two sets: the first > ‘vowel þ

liquid’ (except Pol /l/ after dental, e.g. tłusty); the second > ‘liquid þ /e/’ when the jer is

strong (3.2.1.1) (e.g. krew),andsimpleliquidwhenitisweak(e.g.krwi, płci).

East Slavic: retains the distinction between the two sets: the first > ‘vowel þ

liquid’; the second > ‘liquid þ vowel’; Ukrainian and Belarusian distinguish

between strong and weak jer, the latter > /y/, /i/ (e.g. Bel kryvı

, ja

blyk); stressed

jers are here strong (Ukr kro

vi, plo

ti).

148 3. Phonology