The Common Slavic Vowel Shifts

Hennins Andersen

0. Introduction

In the course of the twentieth century students of comparative Slavic phonol-

ogy have tended toward a consensusthat the prehistorical development of CS
comprises three typological vowel shifts (Andersen 1986).t
The first of these-the First CS Vowel Shift (VS,)-relates the earliest vocal-
ism we can reconstruct from within Slavic (with four short and four long vowels
and oral, nasal, and liquid diphthongs) to the successorvowel system with CS
*l for PS u, and *1, *A, *
0 )u for the Proto-Slavic oral diphthongs (that is, with
five long and four short vowels, but no oral diphthongs); cf. Furdal 196l:41,
'Tonality
Lamprecht 1987:70.2 shifts' such as this are documented in the history
of many languages.They include subtypes with r.r> el (and o > u, as in Gallo-
Romance), r ry (and (or r) o I u, as in CS), i >y (and € > i, asin Ukrainian), and
| > i, ti >r* (and t > i, u > u, as in Faroese).
The SecondCS Vowel Shift (VSz) follows a qualitative differentiation of short
and long vowels and gives rise to LCS vowel systems similar to that of Old
Church Slavonic (tensei, !, u, d, a; lax i, il, e, o); cf . Jakobson 1929U971]:33-36,
Lamprecht 1987 71-74. Some parallels to this vowel shift from other languages
will be mentioned below (Section 2). In the light of such parallels, the different
outcomes of VS, in different CS dialects can be interpreted as minor variations
'protensity
on one and the same subtype of shifts'.
The Third CS Vowel Shift (VS,)-with Isadenko's apt term, the Jer-Shift
'sonority shifts', in which certain
(1970)-exemplifies one of the chief subtypesof
lax and/or high vowels are reanalysed as zero in specific, usually metrically
defined,environments;cf. Jakobson 1929U9711:55,Lamprecht 1987:137-44.The
divergent development of jers in certain weak and strong positions in different
LCS dialects lose their appearanceof idiosyncrasy when the Jer Shift is com-
pared with similar sonority shifts in other languages,such as Latvian, French,
Danish, or English.
Although the basic facts surrounding the CS vowel shifts are well known, there
are still many details concerning each of them that remain to be clarified and
interpreted. In the following pagesI will focus on VS, (Section 1) and discussits
relation to two other CS developments, the elimination of liquid diphthongs
(Section 2), and the emergenceof new quantitative relations (Section 3). My aim
is to define more clearly some of the issuesof interpretation that are raised by
this vowel shift'
Americart contributions to th.e Tw,elflh
International Congress of Slattists. Craco',r), Au.g.-
Sept. 1998.Literature.Lirtgulsrrcs.Poetics,ed. R.
A. Maguire & Alan Timberlake. Bloomingtc-rn,
Ind.: Slavica,1998.
240 HENNING ANDERSEN

1. VS2
According to the scholarly tradition, the period after the monophthongization
of oral diphthongs and the CS *1,, *, change-in short, after VS,-is charac-
terized by such changesas the origin of the jers, CS *d ) LCS *4 the develop-
ment of CS *e, the rise of new quantity oppositions, and changesof *e > *o and
*d > *a (cf. Stieber 1979).But there has long been an understanding that several
of thesechangesare parts of a single,comprehensivereorganizationof the vowel
system (cf. Jakobson 1929 |9711:33-36). This understanding recognizesa vowel
shift comprising a sequenceof innovations, (a) an allophonic differentiation of
long and short vowels into, respectively,tenseand lax and (b) a subsequentrein-
terpretation of the allophonic differencesin protensity (tenseness)as phonemic.
This is a common type of vowel shift. The allophonic phase of such vowel
shifts is often observed;it is illustrated by Modern Czech; cf. Figure I (Lehiste
197031). The classicexample of the phonemic reinterpretation such a qualitative
differentiation can lead to is the loss of distinctive quantity in (varietiesof) Late
Latin (see Figurc 2), which was undoubtedly mediated by realizations of the
original long and short vowels similar to those of Modern Czech(Lausberg 1963:
rs2).
There is a fair amount of variation in the outcomes of such shifts to phonemic
protensity. High and low vowels may be differentiated in tandem, but often low
vowels are differentiated earlier or more than high or mid vowels. In triangular
systems,when the c-vowels are differentiated, the tensenessof the long vowel
may be manifested either in fronting or in backing and/or rounding. In Dutch,
for instance, the opposition latl vs. lal is realized as [a:] vs. [o], tending towards
[re:] vs. [o]. In Swedish, by contrast, la'./ vs. lal is realized as [or] vs. [a]. In
rectangular systems,the long low vowels may be raised, as in Lithuanian: OLi.
d v s . e > e [ e : ]v s . e [ a ] - [ a r ] ; O L i . d v s . a > o [ o r ] v s . a l a l - [ a r ] ( w i t h a n e w
quantity opposition); or the short low vowels may be raised, as in Bulgarian: CS
*d vs. *r, Bg. lel, lal (and lel - laD vs. /e/; CS *c vs. *a > Bg. lal vs. /o/; or front
and back vowels may go separateways, as in most CS dialects,where CS *e and
*a were raised, and*e and*d, lowered (CS *avs. *e > R dial. /G/ vs. lel,CS *a
vs. *a > R dial. lal vs. lttol or lol).

2. Liquid diphthongs and VS,

VS, bears an interesting chronological relation to the elimination of liquid
diphthongs and CS innovations in vowel quantity, which has often been over-
looked in the past.3
In the South Slavic and the Pre-Slovak-Czechdialects of CS, the apparent
metathesisof low-vowel liquid diphthongs (CS TART sequences)occurred before
VSr. Subsequently,in VS2, the secondary,"metathetic" two-mora vowels CS *a
and *E changed to LCS dial. *a and *4' cf. Figure 3. In the Pre-Sorbian and Pre-
Polish-Kashubian dialects,the qualitative differentiation of long and short vowels
occurred before the development of secondaryvowels (CS TART ) LCS TORT);
hence the metathetic (or "pleophonic") two-mora vowels were LCS dial. *d and
*e-,'seeFigure 3. As the display shows, both north and south of the LCS TR1T
 THE COMMON SLAVIC VOWEL SHIFTS 241

F2(Hz)

'Tl

=
N

Figure l: Phonetic Realizations of Long and Short Czech Vowels

Circlesrepresent of long vowels,filleddots,thoseof short
averageF1 and F, frequencies
vowels.5

Classical Latin e e a a o o u u

Late Latin e a a o u

Figure 2: Regular Diachronic Correspondencesin the Development of the Latin

Vocalism

ll TRAI isogloss,in dissyllables,initial circumflex long vowels were shortened;

similarly, final acute long vowels. In Pre-Polish-Kashubianand Pre-Slovak, also
initial acute long vowels were shortened. These are changes that affected all
original (CS) long vowels in these dialects. It looks as if the shortenings may
have occurred after the metathesisof liquid diphthongs.
It is difficult to ignore the fact that the contrasting long and short West Slavic
reflexes of low-vowel liquid diphthongs have exact counterparts in East Slavic.
Due to a peculiarity in the development of Ukrainian, this contrast can only be
illustrated with wordforms in which the pleophonic vowel preceded a LCS weak
jer and underwent compensatory lengthening; and in Russian only LCS TORT
groups with LCS *o have illustrative, distinct vowel reflexes. Sample corre-
spondences are displayed in Figure 4. Here the Ukrainian wordforms have
identical pleophonic vowels in acute and circumflex ToRoT groups, but different
242 HENNING ANDERSEN

Upper Sorbian Polish

Acute Neoacute & Circumflex Acute Neoacute & Circumflex
pretonic pretonic
TROT TROT TROT TROT TROT TROT
hrdch drdn, -a hl6d, -oda groch dial. dion, -u gl6d, -odu
bl6to ml6t, -a hl6s, -osa bloto dial. ml6t, glos
brdza brj6d, -oda brzoza dial. mlii wrzbd, -odu
wrbna w16bel, -bla irjewo wrona wrobel strzewo
kruwa mloko drje *-o krowa dial. mliko drzewo
drdha irddlo zloto droga irbdlo zloto
pluwa br6zda strona plewa br6zda strona
kl6da str6ia broda kloda str62a brodu
sl6ma pr6ca hlowa sloma glowa

Czech Slovak
Acute Neoacute & Circumflex Acute Neoacute & Circumflex
pretonic pretonic
TRAT TRAT TRAT TRAT rRAT TRAT
hrach diin hlad hrach driei hlad
blato mlit hlas blato mliec hlas
biIza vied breza vred
vrdncr dial. vrabel slfevo vrana Iop. Vrabel' trevo
krava mliko dievo krava mlieko drevo
drdha znato zlato draha iriedlo zlato
pldva brdzda strana pleva strana
klada strdi brada klada brada
slama prdce hlava slama praca hlava

Figure 3: West Slavic Reflexesof CS TART Sequences

reflexesunder the neoacute (cf. Polish and Slovak). Russian,on the other hand,
has identical reflexesof acute and neoacute ZoRoT groups-stressed pleophonic
vowel, R dial. 6 for LCS *o-distinct from those with a circumflex (cf. Upper
Sorbian and Czech).
The Ukrainian data exemplified in Figure 4 were first interpreted by Bulaxov-
s'kyj (1949,1961).They have sometimesbeen consideredcontroversial,but no-
thing can be gained by continuing to question their value as evidence.4Ukrainian
specialistshave recognizedfor some time that the dialects of Ukraine present a
significant spatial gradation in TORT reflexes: TOR1T reflexes are largely
limited to original neoacute wordforms in the extreme southwest (as in Figure
4), but are attestedmore and more commonly in both neoacuteand acute word-
forms as one progressestoward the north and northeast. This statistical grada-
 THE COMMON SLAVIC VOWEL SHIFTS 243

Ukrainian Russian
Acute Neoacute & Circumflex Acute Neoacute & Circumflex
pretonic pretonic

TOROT TOROT TOROT TOROT TOROT TOROT

moroz borid gen.pl. pdrox mor6z bor6d gen.pl. pbrox
sordk gen.pl. storin gen.pl. v6ron sor6k stor6n gen.pl. vdron
kol6d gen.pl. holiv gen.pl. h6lod kol6d gol6v gen.pl. gdlod
nahordd gen.pl. borfn gen.pl. k6los ogor6d bor1n gen.pl. kdlos
dol6n' dolit gen.pl. xblod ladbn' dol6t gen.pl. x6lod
bol6t gen.pl. uolik pst.m. s6lod bol6t vol6k pst.m. sdlod
por6v pst.m. vdroh por6l pst.m. dial. vorog
pol6v pst.m. pol6l pst.m.
mol6v pst.m. mol6lpst.m
kol6v pst.m. kol6l pst.m.
'a
bor6vs pst.m. bordls'a
pst.m.

merez zberfh pst.m. bdreh meroz b'er'69 pst.m. b'ir'eg

sterih pst.m. vdred st'er'69 v'er'ed
vdres pst.m. v'ir'es
idreb iir'eb
siren s'ir'en
bdrest b'dr'est

sordika boridka bdroSno sor6ika bor6dka dial.

kol6dka storfnka kol6dka stor6nka bdroino
meriika holivka m'er'6ika gol6vka
ter[dka ier'6dka

Figure 4: East Slavic Reflexesof CS TART Sequences

tion cannot be the result simply of analogical leveling. It is unmistakable

evidenceof an ancient prosodic isoglossthat cut acrossUkraine in the past, but
*kol6d,
has been thoroughly blurred over the centuries, e.9., OldU SW-dial.
*bordd, *hdlod ll N-dial. *kol6d, *bordd, *h6lod (Nazarova 1975).The statistical
gradation in Ukraine forms an obvious counterpart to the loose bundles of lexi-
cal isoglossesthat now reflect the ancient, phonological isoglossfor Acute Short-
ening between Czech and Slovak (see BElid 1972:133-37 and maps 21,22).
But not only do the East Slavic correspondencesetsin Figure 4 correlate with
the similar West Slavic sets in Figure 3. As we shall see, their geographical
distribution is consistentwith other CS dialect features of similar kind and age.

3.0. VS2 and the Quantitative Differention

The correspondencesets in Figures 3 and 4 highlight the shortenings of CS
two-mora syllabics that are traditionally enumerated in the handbooks (e.9.,
Shevelov 1965:506-24).In the following I want to comment on these;however,
due to space constraints I limit myself to observations on the northern CS
244 HENNING ANDERSEN

dialects. As a consequence,I will not comment on the shortening of initial two-

mora syllabics in unaccented(circumflex) disyllables("Circumflex Shortening"),
which appears to set North Slavic apart from South Slavic.

3.1. Acute Shortening

It is well known that CS acute long (two-mora) syllabics have short-vowel
reflexesin all the modern Slavic languages,except that Czech and Upper Sorbian
have long-vowel reflexesin the initial syllable of disyllabic words, and Upper
Sorbian, possibly also in polysyllabic words (Dybo 1963). This diachronic
correspondenceis usually accounted for by hypothesizing a shortening of acute
long vowels ("Acute Shortening") some time after VS, in parts of CS (Timber-
lake 1993,Bethin 1997).But considerthe changesof low-vowel liquid diphthongs
reflected in Figures 3 and 4.
The Figures imply an isogloss that delimits a central, East-Lechitic-Slovak-
Southwest Ukrainianareaagainst peripheral Sorbian-Czechand North Ukrainian
Belarusian-Russianareas.This is a familiar sight to the historical dialectologist:
Acute Shortening was a central CS innovation. The fact that this central area is
bisectedboth by the TR7Tll TRAT isoglossand by the TR1T TRATII TOR1T
isoglosssuggeststhat Acute Shortening may have precededthe changesin liquid
diphthongs. Sincethe metathesisof liquid diphthongs precededVS, in part of the
area (Pre-Slovak), we can infer that Acute Shortening may have preceded VSr.
This conclusion does not necessarilyconflict with the observation made in
Section 3.0, that Acute Shortening appearsto have taken place after VSr. On the
contrary, these two inferencesmerely show it may be useful to distinguish be-
tween (a) allophonic innovations in vowel duration (including a reducedduration
assignedto acute vowels) prior to the changesin liquid diphthongs and VS, and
(b) a subsequentphonemic reinterpretation of these allophonic differences in
duration, at the time of VSr, by which qualitatively identical vowels of different
duration were analysedas quantitatively distinct. I will return to this suggestion
in Section 4.

3.2. Polysyllabic Shortening

Once we recognizethat CS Acute Shortening was a central innovation, we are
in a better position to interpret two other CS dialect differences.
Most CS dialects show evidenceof a fairly general shortening of long vowels
in words of three or more syllables ("Polysyllabic Shortening"). This putative
change is usually explained in phonetic terms in agreement with the common
observation that the more syllablesa word has, the lessduration is assignedto
its individual syllables.But there is evidencein Upper Sorbian that suggeststhat
Polysyllabic Shortening did not occur here, or that, if it did, it did not affect
trisyllabic words; see(1).
(1) a. US mr6iit 'darken'(P mroczyt,Cz mraiil r, Sk mraiit', Bg mrdia beside
dial. mra\d, Sn mraiiti, mraiim, but U moroiyty, R moroiit);
b. US pl6iit'frighten' (P ploiyt, Cz plaiiti, Sk plaiit'sa, Sn pldiiti, but SC
pldiiti, U poloijtty,poloiyt', R poloiit', poloift);
c. US tl\tit 'press'(P tloczyi, Cz tlaiitt, Sk tlaiit', Sn tldiiti, tldiim, but SC
tldiiti, tldiim, u toloijtty, toloiyt" R toloitt" toloitt, but also toloiit"
 THE COMMON SLAVIC VOWEL SHIFTS 245

The evidence,though not limited to the examplesin (l), is somewhat skimpy (see
Dybo 1963). But if it is taken at face value, the peripheral location of Upper
Sorbian immediately suggeststhat Polysyllabic Shortening too may have been a
central innovation. If it was, then perhaps the few East Slavic correspondents
with regular acute reflexes (R, U ToRdT) can be taken as evidence of similarly
peripheral East Slavic dialects in which the change did not affect trisyllabic
words. And the Russian and Ukrainian attestations with non-acute liquid diph-
thongs can be understood as results of Polysyllabic Shortening.

3.3. Final Shortening

Yet another traditionally recognized shortening of CS long vowels affected
word-flnal vowels ("Final Shortening"). There are apparent exceptions to this
change in several of the Slavic languages,but they can be defined in morphologi-
cal terms, and it is not certain they reflect phonological constraints on the gen-
eral Final Shortening.
Final Shortening is usually viewed narrowly as a change affecting original two-
mora vowels. However, if we suppose it too originated as a general allophonic
shortening, then it is reasonable to inquire whether perhaps this shortening
affected not only final (tense) two-mora vowels, but also final (lax) one-mora
vowels. From this point of view it is relevant that there is in fact a central Slavic
area that may reflect an allophonic shortening of final one-mora vowels.
In Andersen 1978 and 1998 I describethe CS *e > *o change that gave rise to
the correspondences in (2) as an allophonic diphthongization ([e] > [S]) that
made it possible for lel realizations to be subsequently reinterpreted as lol, and
I argue that such reinterpretations were possible only in contexts where the diph-
thongs were assignedfull, unshortened realizations(cf. 1998,section 2.1.4). This
hypothesis allows us to interpret the otherwise unexplained absenceof the *e >
*o change reflected in the correspondences in (2) and the -o ll -e and -e ll -o
isoglosses(isogloss 5) in Andersen 1998, Figure 6. They are consistent with a
general allophonic shortening of final vowels and indicate that this shortening,
like those that gave rise to Acute and Polysyllabic Shortening, spread out from
the center of the CS language area.
(2) CS *-e.'LS lico, zboio, buto, buio.foll P lice, zboie, bedzie,bedziecie:lJ tyc6,
moj6,nesett,hl'adyt6;Br S-dial.l'ic6, mojt, n'es'ac'6,hl'a3'ic'e2pl. pres.ll Br
N-dial. l'ico, majo, n'as'ic'6, hl'a3'ic'd; R l'ico, mojo. NE-dial. n'es'it'6,
gl'ed'it'o.

4. Conclusion
The apparently early date of Acute Shortening (Section 3.1) and its geographi-
cal coherencewith Polysyllabic and Final Shortening (Sections3.2-3.3) support
the idea that the CS shortenings of long vowels originated as allophonic innova-
tions prior to VSr.
One can sketch the following hypothesis. Simultaneously with the CS qualita-
tive differentiation of long and short vowels-as the realizations of phonemic
quantity became supported more and more by differences in vowel tenseness-
246 HENNING ANDERSEN

ere were changes in the durational profile of phonological words; the prosodic
template manifestedin every spoken word was adjustedmore closelyto accentual
conditions (accent vs. none, acute vs. circumflex) and metrical conditions (word
length, foot structure). Subsequently,when the vowel system was reinterpreted
in terms of qualitative differences,the shift to phonemic protensity was accom-
panied by a quantity shift: in the new LCS vowel systemsthat resulted from VSr,
the more or less intricately conditioned allophonic distributions of relative
duration becamereflectedas quasi-phonemicquantity. Thesephonetic differences
in length were available to be exploited in new contexts by the development of
the neoacuteaccentfrom non-initial circumflexesand from accentedjers, by con-
tractions, and by the development of compensatory lengthening. Or they could
be abandoned for ever.
It seemsit would be worthwhile elaborating this hypothesis in terms of the
method of phonetic reconstructionillustrated by Timberlake'sadmirable account
of CS compensatorylengthening(1983a,1983b).A full-fledgedhypothetical de-
scription along theselines could perhaps overcome the limitations of Jakobson's
overly schematic theory (1963, 1965). It might be able to side-step a certain
typological weaknessin Timberlake's theory of prosodic shortening (acknowl-
edged at 1983b:307,1986:419).And it might well demonstrate the ultimate com-
patibility of a substance-orientedapproach with the framework of formal
constraints achieved by Bethin (1998).
University of California, Los Angeles

NOTES
t Any discussionof the CS vowelshiftspresupposes a typologyof vowel shifts,which
is part of the theoryof phonologialchange.I cannotpresentsucha typologywithin these
brief remarks,but I must mentionthat I usethe term vowelshiftnot in the generalsense
of chain shift as recentlydone by Labov (1994.113-293), but in the more specificsense
of 'chainshift implyinga shift in vowelsystemtype'.Implicit in this notion of vowelshift
is (a) an understanding (first articulatedby Trubetzkoyin 1928;cf. Jakobson1985:117,
Trubetzkoy 1958 tl962])that vowelsystems represent a smallnumberof typesand (b) an
assumptionthat the developmental possibilities of a givenvowel systemis significantly
conditionedby its type-or groundplan,as Sapircalledit (1921).Hencea shift in type
entailsa shift in developmental possibilities (cf. Andersen1974,Cekman1919).
2 I use the following abbreviationsfor the reconstructedlanguages:CS (Common
Slavic),LCS (Late CommonSlavic),PS(Proto-Slavic). CS refersto the entireprehistori-
cal developmentof Slavic.LCS refersto the last part of this development, the period
from the qualitativedifferentiation of long and short vowels to the fall of the jers. LCS
dialectwordformsand segments arereferredto in termsof their successor languages, e.g.
Pre-Slovenian *q (abbreviatedthus: LCS Pre-Sn.dial. *g). The term PS is usedfor the
reconstructed wordformsand segments that serveaspoint of departurefor thereconstruc-
tion of the chronologicaldevelopment of CS;cf. Andersen1996:183-87. PS segments are
written in smallcapitalswithout an asterisk.
3 In thisexposition,for simplicity'ssake,I follow standardusageand speakof 'metathe-
sis of liquid diphthongs'as if the moderncorrespondences arosethroughsucha change.
Traditionallymany Slavistshavethoughtthat the metathesis of liquid diphthongsin
 THE COMMON SLAVIC VOWEL SHIFTS 247

South Slavic and in the Pre-Slovak-Czechdialects was precededby a vowel lengthening

that supposedlydid not occur in Pre-Lechitic-Sorbian;thus Shevelov(1965:408)and more
recently Carlton 1990:145,Schenker 1995:94,Townsend and Janda 1996:94.Bethin corre-
spondingly speaksof "mora preservation" in the south, but not in Pre-Lechitic-Sorbian
(1998, with references).The idea that the southern and central CS dialects lengthened
liquid diphthongs before the metathesisis incompatible with the parallel quantity reflexes
in Lechitic-Sorbian and Slovak-Czech,displayedin figure 3. This understandinggoesback
to Fortunatov (1980); see Timberlake 1986:423;cf. Andersen 1973, Lamprecht 1987:58.
a Often in the past scholarswho
hesitatedto recognizedeep differencesamong the East
Slavic languageswould try to argue these correspondencesaway by presenting them as
the results of contradictory and chaotic analogical levelings; thus most recently Zaliznjak
(1985:161-63).In fact, however, forms that do not fit the regular pattern of reflexes,say,
in Standard Ukrainian, testify to different directions of analogical leveling in different
periods of the history of the language;cf. nom.sg. por[h, gen.pl. ber[2, dorfh, koriv, vorit,
diminutives ber[zka, dor{ika, korivka, vorftcja, motivated by currently productive rules of
alternation and, on the other hand, such old derivatives as hol1dnyj, kor|tkyj, xolodnyj,
perbdnij, seridnij. Such diverse formations cannot be treated on a par, but need to be
examined with due attention to the historical perspective.
5 The diagram is based on the first syllable
of 571 dissyllabic Czech words; see Lehiste
1970:31.

REFERENCES

Andersen, Henning. 7973. "Prosodic Innovations and the Development of Liquid Diph-
thongs in Early Slavic," Siodmy Miqdzynarodowy kongres slawist|w, Warszawa21-27
VIII 1973; streszczeniareferatow i komunikatow, ed. Janusz Siatkowski et al., l0-ll.
Warsaw.
1974. "Markedness in Vowel Systems," Proceedingsof the Eleventh International
Congressof Linguisls, ed. Luigi Heilmann, 1136-41. Bologna.
1978. "Perceptual and Conceptual Factors in Abductive Innovations," Recent
Developmentsin Historical Phonology, ed. Jacek Fisiak, l-22. The Hague.
1986."Protoslavic and Common Slavic:Questionsof Periodization and Terminol-
ogy," Slavic Linguistics, Poetics, Cultural History. In Honor of Henrik Birnbaum on his
Sixtieth Birthday, I3 December 1985(= International Journal of Slavic Linguistics and
Poetics,3ll32), ed. Michael S. Flier and Dean S. Worth, 67-82. Columbus, Ohio.
1996. Reconstructing Prehistoric Dialects. Initial Vowels in Slavic and Baltic
(Trends in Linguistics. Studies and Monographs, 91.) Berlin.
1998. "Dialektnaj a differenciacija obideslavjanskogo jazyja. Paradoks ob56ix ten-
dencij razvitija s razlidnymi lokal'nymi rezul'tatami," American Contributions to the
Twelfth International Congressof Slavists,Cracow,27 August-2 September 1998: Litera-
ture, Linguistics, Poetics, ed. Robert A. Maguire and Alan Timberlake, 565-600.
B6lid, Jaromir. 1972. Ndstin ieske dialektologre. Prague.
Bethin, Christina. 1998. S/avic Prosody: Language Change and Phonological Theory.
Cambridge.
Bulaxovs'kyj, L. A. 1949. "Sxidnoslovjans'ki movy jak dZerelo vidbuvannja spil'noslo-
vjans'koji akcentolohidnoji systemy," Movoznavstvo415:7-17 .
1961. "OtraLen|ja tak nazyvaemoj akutovoj intonacii drevnejiego slavjanskogo
jazyka v vostodnoslavjanskix," Issledovanija po leksikologii i grammatike russkogo
jazyka, 3-31. Moscow.
 HENNING ANDERSEN

Carlton, Terence R. 1990. Introduction to the Phonological History of the Slavic Lan-
guages. Columbus, Ohio.
eekman, V. N. 1979. Issledovanie po istoriieskoj fonetike praslavjanskogo jazyka.
Tipologija i rekonstrukcija. Minsk.
Dybo, Vladimir Antonovid.1963. "Ob otraZenii drevnix kolidestvennyxi intonacionnyx
otnoSenij v verxneluZickom jazyke," Serbo-luiickij lingvistiteskij sbornik, ed. L. E.
Kalnyn', 54-83. Moscow.
Fortunatov, F. F. 1880. "Zur vergleichendeBetonungslehreder lituslavischenSprachen,"
Archiv fiir slqvischePhilologie 4:575-89.
Furdal, Antoni. 1961. Rozpadjqzyka praslowiaitskiegow iwietle rozwoju glosowego(Prace
Wroclawskiego Towarzystwa Naukowego, seria A, 70.) Wroclaw.
Isadenko,Aleksandr.l970. "East Slavic Morphophonemics and the Treatment of the Jers:
A Revision of Havlik's Law," International Journal of Slavic Linguistics and Poetics
13:73-124.
Jakobson, Roman (ed.). 1985.N ,S.Trubetzkoy'sLetters and Notes (= Janua Linguarum,
SeriesMajor, 47.) Berlin/New York/Amsterdam.
1929ll97l). Remarquessur l'|volution du russe comparbed celle des autres langues
slaves(= Travaux du Cercle Linguistique de Prague, 2.) Repr. Jakobson 1971:7-116.
1963119711. "Opyt fonologideskogopodxoda k istorideskimvoprosam slavjanskoj
akcentologii. Pozdnij period slavjanskojjazykovoj praistorii," American Contributions
to the Fifth International Congressof Slavists,1: Linguistic Contributions. The Hague.
Repr. Jakobson 197l:664*89.
196511971]."Information and Redundancy in the Common Slavic Prosodic Pat-
tern," Symbolae Linguisticae in honorem Georgii Kurylowicz (Polska Akademia Nauk,
Oddzial w Krakowie, Prace komisji jgzykoznawstwa,5), 145-51. Wroclaw, Warszawa,
Krak6w. Repr. Jakobson l97l 693-699.
1971. Selected Writings, l: Phonological Studies [First edition, 1962]. The Hague.
Labov, William. 1994. Principles of Linguistic Change, l'. Internal Factors. Oxford/Cam-
bridge, Mass.
Lamprecht, Arnoit. 1987.Praslovanitina. Brno.
Lausberg, Heinrich. 1963. Romanische Sprachwissenschaft,1: Einleitung und Vokalismus
(= Sammlung Goschen, 1281128a.)Berlin.
Lehiste, Ilse. I 970. Suprasegmentals. Cambridge, Mass./London.
Nazarova, T. V. 1975."Zaminne podovZennjav ukrajins'komu areali na tli sxidnoslovjan-
s'kyx prosodydnyx peretvoren'," Movoznavstvo1975, 6:22-32.
Sapir, Edward. l92l [1949). Language, An Introduction to the Study of Speech.New York.
Schenker, Alexander M. 1995. The Dawn of Slavic. An Introduction to Slavic Philology.
New Haven/London.
Shevelov, George Y. 1965. A Prehistory of Slavic. The Historical Phonology of Common
Slavic. New York.
Stieber, ldzislaw. 1979. Zarys gramatyki por6wnawczejjqzyk|w slowiartskicft.Warsaw.
Timberlake, Alan. 1983a."Compensatory Lengthening in Slavic, l: Conditions and Dia-
lect Geogr&phy," Paperson the Occasionof the Ninth International Congressof Slavists,
Kiev, September, 1983 (= UCLA Slavic Studies, 12), ed. Vladimir Markov and Dean
S. Worth, 207-36. Columbus, Ohio.
1983b. "Compensatory Lengthening in Slavic, 2: Phonetic Reconstruction,"
American Contributions to the Ninth International Congressof Slavists,Kiev, September,
1983, 1: Linguistfcs,ed. Michael S. Flier, 293-319. Columbus, Ohio.
1986. "The Metathesis of Liquid Diphthongs in Upper Sorbian," Slavic Linguis-
tics, Poetics, Cultural History. In Honor of Henrik Birnbaum on his Stxtieth Birthday,