SSLA_A_417719

[Texas

Sep

Scando-Slavica 55 (2009), 147-165.

O n t h e D i a l e c t o f t h e M o r a v i a n C r o a t s 1

Peter Houtzagers

Department of Slavic Languages and Literatures, University of Groningen, P.O.B. 716,

N L-9700 AS Groningen, The Netherlands. h.p.houtzagers@rug.nl

Abstract

The article deals with the northernm ost variety o f what is traditionally called

Burgenland  Croatian  (with  ‘Burgenland’  used  in  a very broad  sense).  Because  o f its
marginal geographic position, the dialect is potentially o f great im portance to  anyone
interested in Burgenland Croatian. However, the inform ation on the dialect to be found
in the linguistic literature is  scarce.  This is partly due to the fact that after the  Second

World War the speakers, who originally lived in three villages in southern Moravia, were

forced to  leave  their hom es  and were  spread  over  a large  num ber  o f villages,  m ostly
in  northern Moravia.  After the political  changes  in  Czechoslovakia  and its  successor
states it becam e possible to trace the M oravian Croats  and to investigate their dialect,
but  until  now  only  one  article  has  appeared  (Lončaric  1998).  The  present  author,
in  contradistinction  to  Lončaric,  is  o f the  opinion  that  the  dialect  has  a  three-way
accentual opposition in stressed syllables: vowels can be long or short, long vowels can
be rising or falling. There is a length opposition in the first posttonic syllable. A  num ber
o f characteristics  distinguish the dialect from  the neighbouring Cakavian dialects, the
m ost  im portant  being  (a)  consistent  lengthening  o f originally  stressed  short  a;  (b)
no  stress  retraction  from   internal  syllables  to  a  preceding  long  syllable.  Both  these
characteristics are relevant for the historical dialectology o f Burgenland Croatian.

K e y w o rd s: C akavian, C roatian dialectology, Serb o-C roatian dialectology, B u r
genland C roatian, C roatian h istorical linguistics, historical dialectology, Slavic

accentology.

1. Introduction

Until the middle of the twentieth century there were three villages in southern
Moravia where a Croatian dialect was spoken. They were the northernmost

I am very grateful to Andrej Novik (Prague) for travelling with me to the Moravian

Croats in Šternberk, Frielištof and Vienna and introducing me to them. I also wish to thank
all the Moravian Croats I spoke with for their hospitality, time, patience and support, espe
cially Jo sef Hubený (Šternberk) and Jo sef Lawitschka (Vienna). My fieldwork was finan
cially made possible by the Center for Language and Cognition Groningen (C LC G ).

DOI: 10.1080/00806760903175466

[Texas

Sep

148

Peter Houtzagers

remaining representatives of a belt of several hundreds of villages with a

Croatian-speaking population that arose as a result of mass migrations of Croats

to the north

. 2

This belt of Croatian villages stretched out from the south of the

present-day Austrian state of Burgenland, through Lower Austria (Nieder

österreich) to the south of Moravia, including adjoining areas in Hungary
and Slovakia

. 3

Throughout this area the Croats were surrounded by speakers

o f one or more other languages - almost always more prestigious ones - and
over the centuries they have been linguistically assimilating to their respective
environments, either by incorporating linguistic elements o f the surrounding

languages into their Croatian or by germanizing, magyarizing, slovakizing

and bohemizing altogether. In the twentieth century the number of villages

with at least a substantial Croatian-speaking minority had fallen to fewer than

a hundred. The total assimilation by giving up Croatian altogether had been
especially strong in the northern part of the belt just mentioned, viz. in Lower

Austria, Slovakia and Moravia. Nonetheless there were three villages in southern

Moravia where Croatian survived until after the Second World War: Frielištof,
Nova Prerava and Dobro Polje

near Mikulov, close to the border with Austria.

Between 1946 and 1950, however, this was put to a sudden end when the

Croatian-speaking communities of southern Moravia were intentionally and

successfully destroyed

. 5

The Czechoslovak authorities considered the presence

o f this Croatian minority on the border with (enemy) Austria a security risk

. 6

These migrations took place from the sixteenth century onward and were caused by the

Ottoman invasions on the Balkans. On the geographical diffusion of the Croats of the "old
diaspora” see Breu 1970. On their dialects see Neweklowsky 1978. On the history o f the
Moravian Croats see Jeřábek 1991 and Pavličevic 1994.

In the following I shall refer to the Croatian dialects spoken in this entire belt as “Bur

genland Croatian”, although some locations are at a considerable distance from the Burgen
land in the administrative sense.

I have adopted here the Croatian names used in Pavličevic 1994 and Lončaric 1998.

My informants said Frielištof, Prerava and Gutfelt. The Czech names are Jevišovka (Fréli-

chov until 1950), Nový Přerov and Dobré Pole, German Fröllersdorf, Neuprerau and Gut-

tenfeld.

According to Pavličevic (1994, 143), the forced displacements took place in two phas

es, viz. 1946-47 and 1948-50.

In 1938 (Munich Agreement), parts of Czechoslovakia, including the southernmost

part of Moravia, were ceded to the German Reich and all men were forced to serve in the

German army. The authorities in post-war Czechoslovakia considered this treason. Also

the  Communist leaders who took over power in  1948 had their doubts  about how much
sympathy the  Croats  could be  expected to have towards  the new ideology.  These doubts
were inspired,  among other things, by the results  of the elections  of 1947  (see Pavličevic

Scando-Slavica 55, 2009

[Texas

Sep

The Dialect of the M oravian Croats

149

Therefore they moved almost

the entire Croatian population out of the three

villages and spread them over a considerable number of villages, most of them

in northern Moravia

. 8

This led to instant linguistic assimilation in the sense that

almost none o f the descendants of the Moravian Croats born outside the three

villages speaks any Croatian

. 9

It goes without saying that, because of its marginal geographic position,

the dialect o f the Moravian Croats (henceforth “M C ”) is potentially very im
portant to anyone interested in Burgenland Croatian.

Until relatively recently, however, little was known about it. There were

only a few remarks and fragments o f dialect texts published by Herben

(1882), Milčetic (1898) and Vážný (1934).10 In his standard work on the

Croatian dialects o f the Burgenland and its wide surroundings, Neweklowsky
confines himself to summarizing the scarce information available to him and

remarks: “die Mundarten sind heute vermutlich nicht mehr untersuchbar, da
ihre  Sprecher nach  dem  Zweiten Weltkrieg ausgesiedelt wurden  und  daher
kaum noch  aufzuspuren  sind”  (Neweklowsky  1978,  90).  Until the political
changes in Czechoslovakia and its successor states from  1989 onward, it was
probably true that it was not easy to trace the Moravian Croats and investigate
their language, although I do not know o f any attempt to do so.

After 1989, however, not only were two monographs published on the

Moravian Croats (Jeřábek 1991 and Pavličevic 1994) which concentrated on
history, ethnography and culture, and in which a great deal o f older material
was republished, but there was also some renewed interest from the part of
linguistics. In 1998 an article appeared by Lončaric, with part of the material

1994, 139-148). In 1999 the Czech government officially expressed its regret about the

treatment of the Croatian minority after the war (see Novik 2005, 11).

Some of the inhabitants of the three villages fled abroad and a very small number es

caped their fate through luck and managed to stay.

This was one out of a number of ethno-political measures taken by the communist

government in post-war Czechoslovakia. Others were the expulsion of Germans from the
country and the re-immigration of Czechs and Slovaks from such countries as Bulgaria.
Most of the houses that the Croats moved into had belonged to expelled Germans, and
their houses were mostly occupied by re-immigrated Czechs and Slovaks.

My informants knew of one single member of a younger generation with some active

command of Croatian.

1 0

I agree with Neweklowsky (1978, 90) and Koschat (1978, 156) that the sermon pub

lished by Miklosich (1890) is not characteristic for the dialect under discussion. The only
typically Moravian Croatian phenomena to be found in the text are the word maja ‘mother’

and the final -v in czév ‘whole’ (Miklosich 1890, 317). About this final -v see the last para
graph o f section 4 below.

Scando-Slavica 55, 2009

[Texas

Sep

150

Peter Houtzagers

he had gathered for the Hrvatski dijalektoloski atlas, and the Prague Slavist An

drej Novik has been active in linguistic fieldwork among the Moravian Croats

from 2001 on (see Novik 2005).

In the summer o f 2008 I had become curious to hear the language o f the

Moravian Croats myself, and I spent a few days carrying out fieldwork in
some o f the places where the speakers can now be found. My material con
sists o f about

1 0

hours o f audio-recorded conversation with

informants, all

born in Frielistof and Nova Prerava. In the following I shall present my find
ings, which differ in a few respects from those of Loncaric (1998), the most
important being accentuation.

2. Language Situation

It is clear that this variety of Croatian is on the verge o f extinction. My
informants estimated that there were fewer than fifty speakers left, who, with

very rare exceptions are over seventy years o f age. They do not form any real

speech-community, since they are dispersed over a large number o f places in
Moravia  and abroad and the number o f different persons in  each place that
speaks Croatian is very limited. Some of them still speak Croatian with their
spouse  or  other  family  members  o f their  own  generation  and  sometimes
they are healthy and socially active enough to  entertain contact with Croats

outside their direct environment. In many cases, however, the circumstances
are less favourable and these people hardly ever speak their native language.

The informants I spoke with all communicate in Croatian on a regular ba

sis.  It  is  difficult  to  say whether  the language  they speak differs very much
from the language they were brought up with in the three villages that they
were forced to leave, but I have no reason to think that it does. They seemed
not to  suffer from a “diminished fluency” in  Croatian and the language that
they speak is clearly the same language, without great differences between the
informants. I hardly noticed any difference between the speech of the infor
mants from Nova Prerava and those from Frielistof, therefore I shall describe
the  dialect  as  a  single  one

. 1 1

O f course the dialect is influenced by Czech,

but as can be seen from the text specimens from the nineteenth century this
has been going on for more than a hundred years. Also there is a lot of code

11 There is one lexical difference that is always mentioned by the speakers: iertipli

(Frielistof) vs. kumplri (Nova Prerava) ‘potatoes’. It is also possible that the optional final -i

in the infinitive is restricted to Frielistof (see section 7 below).

Scando-Slavica 55, 2009

[Texas

Sep

switching and “spontaneous borrowing" o f words, including Czech phoneti-
cal and morphological features, but not more than can be expected consider
ing the circumstances.

3. Vowel Inventory and Accentuation

The Dialect of the M oravian Croats

151

The dialect has the following accented vowels, including a syllabic resonant:

Vowels

Long

Short

Falling

Rising

uô

Examples: p lsa f ‘write’ LPm

, 1 2

diede ‘grandfather’, dali ‘give’ LPplm, muore

‘sea’, kUpif ‘buy’ LPm, pise ‘write’ PR3sg, siela ‘village’ Npl, láni ‘last year’,
kuoni ‘horse’ GIpl, l'udi ‘people’ GIpl, hod'ili ‘go’ LPplm, sestra ‘sister’, lapti
‘plot o f land’ Npl, mogaf ‘be able’ LPm, drugde ‘some place else’, mrvu ‘a little’.

As can be seen from the table and the examples, I assume that there is a tone

distinction on long vowels. This is an important difference between Loncaric’s
analysis and mine. I shall discuss this matter separately in section

I provisionally assume that there are no length or tone oppositions on syl

labic r and that there is no tone opposition on long monosyllables

. 1 3

12 Abbreviations: N, G, D, A, I, L nominative, genitive, dative, accusative, instrumental,

locative; sg, pl singular, plural; m, f, n masculine, feminine, neuter; INF, PR, IMP, LP,

PP infinitive, present, imperative, l-participle, passive participle; 1-3 first person-third

person.

13 I have heard no clear examples of long syllabic r, but this could be due to the lim

ited amount of material. Both Baumgarten and Koljnof have length and tone oppositions
on syllabic  r  (see  Koschat  1978,  64-66  and Houtzagers  2008a,  253-254).  The  Haci  and
Poljanci  dialects  are mostly assumed to  have  no  tone  distinction  on monosyllables  (see
Neweklowsky 1978, 71  and 99,  and Koschat  1978, 74). For the Koljnof dialect I have ex
pressed some doubt  about  the  absence  of this  distinction  (Houtzagers  2008a,  253),  and
also in the case of the M C dialect I am not certain. I shall look into both these matters on a
later occasion.

Scando-Slavica 55, 2009

[Texas

Sep

152

Peter Houtzagers

The long mid vowels are opening

1 4

diphthongs in the sense that their

quality is more closed at the onset than at the end o f the vowel. There is con
siderable freedom as to  the prominence o f the respective parts  o f the  diph
thong. In most cases the first, more closed part is more or less as prominent
as the second one:  one hears a closed vowel  [i]  or  [u]  which more or less in
the middle opens to  [e]  or [o]. The first part can also be less prominent than
the  second,  resulting in  [ie],  [uo],  and  sometimes  the  first  part  is  glide-like

([-e], [uo]). The timbre o f the second part is always high-mid. In this respect

the  dialect is  distinct  from  a  number  o f Haci  and  Poljanci  (that  are  closest
to the M C )  and other Burgenland dialects, where the second component of
the  diphthong can be realized  as very open  (see  Koschat  1978,  42;  Newek-
lowsky 1978, 62 and 95; Houtzagers 2008a, 255). The diphthongs, especially
if they have a rising pitch, can be realized as rather short. However, there is no
chance o f confusion with short e and o, which are always monophthongal

. 1 5

Nasalization o f vowels in the direct vicinity o f m or n is rare and, if present,

less strong than in a number o f Haci and Poljanci dialects (see Koschat 1978,
41; Neweklowsky 1978, 63; Houtzagers 2008a, 255).

Loncaric (1998, 110) rightly signals rounding o f i before laterals, resulting

in  [u], i.e.  [nedul’a]  ‘Sunday’ but states “Analognu pojavu kod vokala e nisam
zabiljezio.”  I  have  attested  several  cases  o f e before  a  lateral  being  rounded
to  [o],  e.g. in Gutfielta Gsg, velike ‘large’ Apl.  One informant often produced
rounded realizations  of / i /   (somewhere between  [i]  and  [u])  after palatals,

e.g. in vicisti ‘clean’ PR3sg.

In the first posttonic syllable there is no distinctive pitch, but the same length

distinctions are present as in stressed position, except on syllabic resonants:

Vowels

Syll. Res.

Long

Short

14 I have avoided the term "rising diphthong” because it could cause confusion between

the level of vowel timbre and that of pitch.

15 In a few cases I heard short e and o realized with a slight diphthongization in the op

posite direction ([ei],[ou]).

Scando-Slavica 55, 2009

[Texas

Sep

The Dialect of the M oravian Croats

153

Examples: ctovlk ‘person", devedesiet ‘ninety", lapat ‘plot o f arable land",

krizuom ‘cross" Isg, tisuc ‘thousand" Gpl, brali ‘take"LPplm, buhe ‘flea" Npl,
blaga ‘livestock" Gsg, spatno ‘bad" ADV, Uru ‘hour" Asg.

From the above it is clear that I do not agree with Loncaric "s observation

that, except for the diphthongs, original posttonic lengths were shortened in
the dialect (Loncaric 1998, 109). I shall discuss this further in section

In unstressed syllables other than the first posttonic there is no length distinction:

Vowels

Syll. Res.

Examples:  (pretonic) strilali ‘shoot" LPplm, nedil’a ‘Sunday", razumili ‘under
stand"  LPplm,  nosila  ‘carry"  LPf,  udielaf ‘do"  LPm;  (posttonic)  obisili  ‘hang
up"  LPplm,  uodale  ‘marry"  LPplf,  nuovoga  ‘new"  Gsgn,  vieliko  ‘large"  Nsgn,
zacinaju ‘begin" PR3pl.

In German loanwords syllabic resonants other than r can occur in posttonic

syllables. I attested only l and n, and only in the first posttonic syllable: majsns
‘mostly" kepl ‘utensil used for treshing", but the dialect might very well have
other syllabic resonants such as l and m, and also in other posttonic positions.

4. Consonants

The dialect possesses the following consonant phonemes

: 1 6

Stop

Fric.

Affr.

Nas.

Lat.

Trill

Glide

Voice:

Lab.

Dent.

Pal.

Vel.

Lar.

16 The use of the symbols d, n, l’, c and c agrees with common practice in Croatian dialec

tology. I shall not go into the discussion about their mono- or biphonemic status. I have
not attested any proof of distinctions d’ vs. dj, n vs. nj, etc.

Scando-Slavica 55, 2009

[Texas

Sep

154

Peter Houtzagers

As almost everywhere in Burgenland Croatian, word-final consonants that

participate in the opposition voiced vs. voiceless are voiced before any voiced
phonemic unit, including m, n, n, r, l, l’, j, voiced allophones of h and vowels.

Word-final consonants are voiceless only before a pause or before a voiceless

consonant. Examples: prek trih liet (h [fi]), ciekat as ‘wait until’ ( t [d]), ces

ju poznat ‘you will recognize her’ (s ^ [z]), duojt rano ‘arrive early’ (t ^ [d]),

kat bis imaf ‘if you would have’ (t ^ [d]; s ^ [z]), nek malo zitka ‘only a little

wheat (k [g]), nasa maja su fu rt hlibepiekli ‘our mother baked loaves all the
time’ (t ^ [d] before [fi]).

The stops / g / and /k / are consistently realized as velars. The fricative /h /,

however, varies between a voiceless velar and a voiced laryngeal. Before a

voiceless consonant or a pause it is a voiceless velar. Before a voiced consonant

or a vowel it is mostly a voiced laryngeal, but in a minority of cases it can be

realized as a voiceless laryngeal or velar. For example, I heard voiceless velar
realizations in petieh ‘rooster’ and hfili ‘want’ LPplm, laryngeal realizations

(mostly voiced, but not always) in hrvati ‘Croat’ Npl and voiced laryngeal and

voiceless velar realizations in kozuhof ‘part of the national costume’ Gpl. I have

the impression that in positions where there is variation, the variation is free. In
sandhi position before a voiced speech sound a voiced velar realization can be
heard, e.g. ot tih viecihpaurov ‘o f those bigger farmers’ (h [y]). The laryngeal

allophones probably arose under the influence of Czech. The difference with
Czech however, is that Czech has separate phonemes /x / and / fi/ (spelled ch
and h, respectively), whereas the M C dialect has only one phoneme /h / with

velar and laryngeal allophones. As a result of this, Czech loanwords which have

/x / (ch) in Czech can be pronounced with a voiced laryngeal by the Moravian
Croats, which, from the point o f view of Czech, is a change in the phonemic

make-up o f the word

. 1 7

For example, I attested vihuovali ‘bring up’ LPplm and

hlape ‘boy’ Apl with voiced [fi], cf. Czech vychovat, chlap.

Etymological *di, *ti, *ni are sometimes heard as [d’i], [t’i] and [n]. Two

cases must be distinguished: (1) Czech loanwords, e,g, rod’ici ‘parent’ Npl,

zjiscit ‘establish’, hranici “border’ Lsg; (2) Occasional palatalization of /d /,

/ t / and / n / before / i / under the influence o f Czech. The latter phenomenon

was attested, among other forms, in porodit ‘give birth’, Brni ‘Brno’ L sg

. 1 8

It is

17 In positions where in Czech consonants are normally devoiced, e.g. word-finally, this

phenomenon does not occur, because in such positions the opposition between /ft/ vs.

/ x / is neutralized in Czech and the resulting archiphoneme is realized as [x].

18 I did not find any example with [t'] in my material, which I consider accidental.

Scando-Slavica 55, 2009

[Texas

Sep

The Dialect of the M oravian Croats

155

possible that the distinctions /d i/, /ti/, /n i/ vs. / d ’i/, / t ’i/, /n i/ are under
pressure, but in view of (a) the great number o f attestations of unpalatalized

/d /, / t / and / n / before / i / and (b) the fact that one o f my informants cor
rected his [rad’ili] into [radili] ‘work’ LPplm, we may safely assume that the
distinctions are still operative.

Word-final *l

1 9

almost consistently appears as a labiodental fricative, which

is realized voiced or voiceless according to the general rules given above in
this section. I shall write this fricative as f , which agrees with its realization
before a pause

. 2 0

Examples: blf ‘be’ LPm, p isaf ‘write’ LPm, cief ‘whole’ SFm,

provrisaf Gpl ofprovrlslo ‘rope made o f plaited straw’, vuof ‘ox’. The only excep

tion in my material is siel Gpl ‘village’.

5. Diachronic Observations on Vowels and Accentuation

On the basis o f the small amount o f data that he had at his disposal, Newe-

klowsky assumed  that the  dialect was  not  essentially different  from  those
o f the Haci and Poljanci. The Haci and Poljanci dialects, together with the
Dolinci group, belong to the “ikavisch-ekavischen Mundarten mit partieller
regressiver Akzentverschiebung”  (Neweklowsky  1978,  58).  It  is  true  that
the M C  dialect is i/e-kavian, i.e. the reflex o f ja t is a high front or a central
front  vowel,  according  to Jakubinskij’s  law  (1925),  e.g.  sniga  ‘snow’  Gsg,
dite  ‘child’,  dielat  ‘do’,  kolieno  ‘knee’.  However,  the  stress  retraction  that
Neweklowsky describes seems not to have been fully operative.  This must
not  remain  unnoticed,  since  the  stress  retraction  is  one  o f the  primary
criteria in Neweklowsky’s classification o f the dialects o f the “old diaspora”.

What Neweklowky calls the partielle regressive Akzentverschiebung consists

of four components:

(i)

The stress is retracted by one syllable from final syllables.

(ii)

The stress is retracted by one syllable from internal syllables if the

vowel in the preceding syllable was long.

19 This does not apply to syllabic word-final l as in kepl ‘utensil used for trashing’.

20 Loncaric (1998, 32) discusses the question whether this fricative should be analyzed
as an f or a v. However, there is no opposition between f and v in word-final position. Prob
ably the bilabial glide attested in similar positions in other dialects must be seen as a pre
ceding stage. If this is true, the specific M C development can simply be viewed as a change
from bilabial to labiodental.

Scando-Slavica 55, 2009

[Texas

Sep

(iii)

If the vowel that received the stress was long, it became falling.

(iv)

The vowel that lost the stress became long

. 2 1

In the M C dialect (i), (iii) and (iv) apply, cf. the stress retraction and the falling

vowels in mliko ‘milk’ and pitat ‘ask’ and the posttonic lengths in lapât ‘plot of

arable land’ and petieh ‘rooster’. However, the forms pitâla LPf, rukâmi22 ‘hand’

Ipl and utuôrak ‘Tuesday’ show that (ii) does not apply, cf. the following forms
from the Poljanci dialect o f Baumgarten: pitala, rukami, utuorak (all three with
optional length in the first p osttonic syllable, Koschat 1978, 118, 75, and 91).23

The dialect shares a number o f other accentual features with the Haci and

Poljanci and part of the Dolinci: (1) originally stressed (i.e. stressed before

the stress retraction) short *o and *e became long and rising, except in open
final syllables, e.g. duôma ‘at home’, ciêkam ‘wait’ PR1sg

. 2 4

(2) There are many

instances of vowel lengthening in closed syllables, e.g. brât ‘brother’.

Elsewhere I have argued that it should not be excluded that rule (1) in the

preceding paragraph may apply to *a as well, in other words that before the
stress retraction all short stressed non-high vowels became long and rising

(Houtzagers 2008a, 260-261). Surprisingly enough, although many Burgen

land  Croatian dialects have instances of long rising a reflecting an originally
short stressed *a,  the  question whether the lengthening rule  included *a  is
not asked in the publications where the lengthening is discussed  (e.g.  New-

eklowsky 1978 and Koschat 1978). This could have to do with the fact that
in the Burgenland Croatian dialects studied so far the results of the lengthen
ing o f *a (if there was such a lengthening) present a much less systematic

picture than those of the lengthening of *e and * o. In the case o f *e and * o, the
rule seems to have applied without any exception, whereas in the case o f *a
we  often find doublet length  or no  length at all. I have proposed a scenario
that explains the apparent lack of systematicity of the reflexes o f the supposed
lengthening o f *a  in comparison to the reflexes of lengthened *e and *o. Ac

cording to that scenario, after the lengthening and before the stress retraction,

21  Elsewhere (2008a, 261-262) I have argued that there is little independent evidence for
such  a lengthening  and that  instances where  this  lengthening is  supposed to have  taken
place can almost always be explained by other lengthening rules. What is essential here is
that the situation is the same for the M C and for the Haci and Poljanci and Dolinci.
22  In both these forms the long rising â was originally short (see (3) in the next paragraph).
23  Koschat  (1978, 260) does not give rukami but assigns ruka to the same accentual para

digm as brazda, Ipl brazdami.

24 The forms diêlat and koliêno given above were also examples of this.

156

Peter Houtzagers

Scando-Slavica 55, 2009

[Texas

Sep

The Dialect of the M oravian Croats

157

length was not distinctive on e, o and a and the phonetic realization of e, o and

a could vary freely with respect to length, especially if these vowels had a ris
ing pitch

. 2 5

As a result o f the stress retraction, new short e, o and a arose (pre

viously pretonic short e, o and a). New short a and the shorter allophones

of old stressed a now occupied the same area, which led to confusion and
reshuffling, whereas new short e and o did not overlap phonetically with old
stressed e and o because the latter two had diphthongized in the meantime

(Houtzagers 2008a, 260-261).

In the M C dialect old stressed short *a is reflected as a long and rising

а, without many exceptions or cases o f optionality, e.g. baba ‘grandmother’,

povídala ‘told’ LPf, našoj ‘our’ Lsgf. Stressed short a reflects old pretonic short

*a and, o f course, old short stressed a in words that were incorporated in the
dialect after the lengthening had ceased to operate, such as loans. Examples:
našlí ‘find’ LPplm, lapat ‘plot of arable land’, najgerík ‘curious’.

I regard these examples as supportive for the hypothesis that in this part of

Burgenland Croatian, before the stress retraction, stressed *a was lengthened
together with *e and *o

. 2 6

The fact that the M C material presents a clearer

picture than the other dialects of the northern Burgenland described so far
is perhaps caused by the marginal geographic position of the dialect with re
spect to the rest of Burgenland Croatian. It is imaginable that the freedom

described above with respect to phonetic length after the lengthening and

before the stress retraction originated somewhere south o f the M C dialect

and did not spread that far to the north.

б. Vowel Length and Pitch Opposition

According to Lončaric, the dialect has no pitch opposition, but it does have

a length  opposition in stressed syllables.  He  also has his reservations  about
the  latter  opposition:  “Ograda  za  kvantitetu  postoji  zbog  toga  što  se  kratki
slogovi mogu  ostvarivati  dugo, jednako  kao  i  dugi slog,  npr.  vuoda, ja:gode,

25 In the present-day Haci and Poljanci dialects the average long falling vowel is one and a

half time as long as its long rising counterpart (see Neweklowsky 1978, 67). I did not point

this out in Houtzagers 2008a.
26 Moreover, if we accept the scenario depicted above, we can conclude (1) that the

lengthening of old stressed *e and *o was primarily a lengthening, not a diphthongization;

(2) that the diphthongization arose after the lengthening but before the stress retraction.

The latter is also proposed by Neweklowsky, but on other - in my opinion insufficient -
grounds (cf. Neweklowsky 1978, 73; Houtzagers 2008b, 310-311).

Scando-Slavica 55, 2009

[Texas

Sep

158

Peter Houtzagers

buoli,pa:met, uoko, stada” (Loncaric 1998, 109).27 Loncaric clearly proceeds

from the assumption that the stressed vowels in these words should be
phonemically short and is surprised to hear them realized long. One can
wonder, however, if these vowels must be analyzed as phonemically short. If
one assumes that the stressed vowel in uoko was originally short, one expects
it to be long and rising in the present dialect by virtue of the lengthening of

o and e in nonfinal syllables (see section 5 above

) . 2 8

The words pâmet and

stâla are also present in my material, both with long and rising a. The length
in these words can be explained by the lengthening o f a in nonfinal syllables

(also section 5 above

) . 2 9

I have a number o f attestations o f vuôda, consistently

with long and rising uo, which leads me to the assumption that the vowel is
long in the present dialect. One would expect shortness on newly stressed

o in this form. A possible explanation is that uô (from old stressed short o)

spread from the accusative singular to the rest o f the paradigm

. 3 0

Loncaric (1998, 109) assumes that in the first posttonic syllable there are

only two long vowels, viz. ie and uo, and that the other posttonic lengths have

been shortened

. 3 1

This does not agree with my findings. I have many attesta

tions o f â, Ï and ü in the first posttonic syllable and the distribution o f these
posttonic lengths is largely consistent, both dialect-internally and historically.

Loncaric’s findings and mine also differ with respect to the tonal distinc

tion on long stressed vowels. I am quite confident that the dialect has a tonal

opposition on long stressed vowels. However, it can be difficult to hear in

2 7

Later on the same page Loncaric seems to doubt whether the present dialect has a

length opposition at all: “Treba pretpostaviti da je jezik mH u 20. st. kada je bio redovno
sredstvo komunikacije hrvatske zajednice poznavao fonolosku opreku po kvantiteti.”

2 8

I have no attestation of this word myself. In Baumgarten the stressed vowel is long and

falling (Koschat 1978, 240), so that there is a chance that also in the dialect of the Mora
vian Croats the stressed vowel was not short to begin with.

2 9

In the lexicon of Koschat 1978 one finds pamet but jagoda/jâgoda and stala/stâla.

3 0

I assume that buoli is the PR3sg of the verb ‘hurt’. I cannot explain the long vowel in the

form attested by Loncaric. I have one attestation:  boli. Loncaric also gives four examples of
words in which he heard a short vowel instead of an expected long one: “mùs za mu:s, vràta
za vra:ta, ràstu za ra:stu, zldi za zi:di” (Loncaric  1998,  109). In the case of mus and zidi I can
understand the confusion, because long and falling i and u  are sometimes hard to  distin
guish from short i and u. I cannot explain the shortness Loncaric heard in vràta and ràstu (I
attested vrâta  and râstie). I also do not understand why he expected a long stem-vowel in
rastu, since Koschat (1978, 257) has rasie PR3sg.

3 1

A few sentences later Loncaric gives the form lona:c as one of the examples for his as

sumption that the history of the accentuation in this dialect is essentially the same as that
in the dialect of the Poljanci.

Scando-Slavica 55, 2009

[Texas

Sep

The Dialect of the M oravian Croats

159

an individual instance whether a given vowel is rising or falling, because the
sentence intonation also has its consequences for the pitch movement

. 3 2

is possible that this is why Loncaric heard the stressed vowel of one and the
same word sometimes as rising and sometimes as falling and concluded that
pitch was not distinctive on word level

. 3 3

Because this is a matter o f perception and not argumentation, it is im pos

sible to prove here that I am right and Loncaric is wrong. Therefore I have
made a web page on which one can hear series of different realizations of one
and the same word-form, e.g. ten realizations o f form x with vowel a, ten re
alizations of form y with vowel a, etc. When listening to the different realiza
tions of x, one gets an acoustic picture o f the limits between which a varies

within different circumstances of sentence intonation. When one listens to y,

one gets an acoustic picture of the variation of a, and its becomes quite clear
that, in all their variety, a and a are not the same. Presentation on the web
also has the advantage that a larger number o f colleagues can form an opinion
on the material. The web page also includes similar series of forms with and

without length in the first posttonic syllable. The U RL is http://www.let.rug.

nl/houtzage/tones.html.

7. Remarks on M orphology

The Gpl ending -of is not restricted to masculine and neuter paradigms, but
has also been attested for feminine nouns, e.g. bacvof ‘barrel’, Hizof ‘house’. In
contradistinction to Loncaric I did not attest a Gpl ending -af for feminine nouns.

The infinitive occurs both with and without final -i, e.g. dielat(i) ‘do’, duojti

(

3 4

come

. 3 4

Like Czech, the dialect has the ending -me in the first person plural o f the

present tense and the prefixes vi- and roz-. Examples: grieme ‘go’ PR1pl, go-
vuorime ‘say’ PR1pl, vibrali ‘choose’ LPplm, vidrzali ‘endure’ LPplm, rozbit

‘break’ PPm, rozdilili ‘divide’ LPplm. The prefix raz- occurs in razumis ‘under
stand’ PR2sg.

3 2

Koschat (1978, 77) writes about her own native dialect: “Deshalb ist es bisweilen recht

schwer, den eigentlichen Wortakzent herauszufinden”

3 3

The perception of the pitch opposition in Cakavian is further complicated by the fact

that on the phonetic level the distinction is not between a falling and a rising tone but be
tween a falling and a flat tone (see Langston 2006, 40).

3 4

I have the impression that infinitives in -i are only used by Frielistof informants, but I

have to check this.

Scando-Slavica 55, 2009

[Texas

Sep

In addition to the second person polite form that is common for Slavic

languages, there is also a third person polite form

. 3 5

If the subject of a clause

contains a singular noun that refers to an elder relative or another person that
deserves respect (an older person, a priest, etc.), the corresponding finite verb
form or l-participle has a plural ending. As in the second person polite form,
if the person referred to is a woman, the LP ending is not -e (feminine) but -i

(normally masculine or mixed gender). A predicatively used adjective is usu

ally also in the plural (in -i), but singular also occurs. Examples: To su nas caca
imali jako radi ‘Our father liked that very much’, M aja su bili ur stari ‘Mother

was already old’, Kakof su bili ta tvoj susiet? ‘What was that neighbour o f yours
like?’ A person belonging to the category described above, either man or wom

an, can also be referred to by oni ‘they’ (masculine or mixed gender).

My material is too limited to be able to give a picture of the role of accen

tuation (stress, length, pitch) in morphology, but such forms as kravami ‘cow’
Ipl (with unexpected stress on the ending) suggest that also in this respect
the dialect deserves attention.

A number of other morphological characteristics are discussed in section

8. Other Characteristics o f the Dialect

At the end o f his standard work on the Croatian dialects o f the Burgenland

in the widest sense, Neweklowsky (1978,  after page 376)  presents 52 maps,
which  give  a  picture  of  the  most  important  isoglosses  of  the  area.  The
dialect  of the Moravian  Croats is not included in the maps and most o f the

characteristics that the maps are about are missing in the part o f the book
dedicated to it. In this section I shall try to fill up this gap by giving the

missing information  about  the  dialect  as  far  as  I  have  it  myself.  I  shall  use
Neweklowky’s map numbers.  I shall state about each characteristic whether
it agrees with the two northernmost dialect groups o f Burgenland  Croatian,

viz. the Haci and Poljanci and the Dolinci dialects (= H P /*H P ; = D o l/*D o l).

Sometimes I shall include other Burgenland dialects in the comparison.

1. The interrogative pronoun ‘what’ is ca, =H P =Dol.

2. The reflex ofja t is i/ekavian, =H P =Dol.

35  Neweklowsky (1978, 248-249)  also briefly mentions this. His  examples differ on mi
nor points from what I found in the M C  dialect. The third person polite form is absent in
Burgenland Kajkavian (described in Houtzagers  1999).

160

Peter Houtzagers

Scando-Slavica 55, 2009

[Texas

Sep

The Dialect of the M oravian Croats

161

3. Jat is reflected a in gnazdo, =H P =Dol.

4. There is a length opposition in stressed and posttonic syllables, =H P

=Dol.

5. There is a partial stress retraction, as Neweklowsky calls it, o f “type

1" (žena, otac, dite, ženie, pedesiet), =H P =Dol, but the stress is not

retracted from medial syllables to a preceding long vowel (see section

5 o f this article), *H P *D o l

. Length of (once) stressed final closed syllables: lapat, clovik, =H P

=Dol.

7. Diphthongization of (i) old long *e, *o and (ii) old short *e, *o with

original stress, not if word-final, =H P =Dol.

. Presence of syllabic r, =H P =Dol.

9. Reflex o f weak jer in Isg mmo as an a-like vowel (mânu), =H P =Dol.

10. Reflex of weak jer in *dbnbsb: danas, *H P *D o l. H P and Dol have de-

nas. The same reflex as that of the M C is found in Devínska Nová Ves
and Chorvátsky Grob.

11.  Reflex o f front nasal in *ž$dan: no data
12.  Reflex o f front nasal in *ž$tva: no data.
13.  Development *ra > re in vriebci Npl but not in râslo, krâst, =H P =Dol.
14.  Preposition va, =H P =Dol.
15.  Reflex o f *vl· in *vbzçti: ziet, =H P =Dol.
16.  Accentuation in the present o f the verb morati: no data.
17.  Presence vs.  absence  o f -i in  mâti:  not relevant since  the  dialect has

mâja *H P *D o l (Hidegseg and Fertóhomok have majka).

18. Vowel in tepal/topal: teplo, =H P =Dol.
19. The form o f the verb ‘pull’ is vUcie PR3sg *H P (only attested in one

Dolinci dialect and in the extreme south), but ‘dress’ and ‘undress’ are
obličie and visličie PR3sg =HP.

20. Initial i- in šli ‘go’ LPplm is mostly absent, =H P *D o l, but it can be

present as well.

21. Presence of opposition č vs. c, =H P =Dol.

2 2

. The masculine singular form of the l-participle ends in f , which is un

like any other Burgenland dialect, but is closest to the variant -u of

the HP.

23.  Retention o f the phoneme I' as opposed to j, = HP, *D o l.
24.  Absence o f prothesis o f j-, d'- or dž- before initial i-: *  H P =Dol.
25.  Absence of prothesis o f v- in Učit, =H P =Dol.

Scando-Slavica 55, 2009

[Texas

Sep

162

Peter Houtzagers

26. Retention o f initial h in htlli ‘want’ LPplm, unlike anywhere else in the

Burgenland (elsewhere it is either til or stil). * HP.

27.  Final -m has not become > n, =H P *  Dol.
28.  No -k- in the comparative of velik: vieci, =H P =Dol.
29.  Ordinal numeral trieti or treci: no data.
30.  The  Lsg  ending  o f o-stems  is  -i,  =H P  =Dol.  In  contradistinction  to

the H P and Dol, however, there is no free alternative ending -u along

with -i. I have attested three instances with -u: Znojmu/-i (toponym),

bamku/-i ‘house (informal)’ (loan from Czech) and piru ‘wedding’.

The latter form suggests that this word has a preference for -u, -u in the
two other words could be due to Czech influence. Preference o f indi

vidual words for the ending -u in a dialect where the regular ending is

-i was also found in Koljnof, Hidegség and Fertohomok and in older

written texts (see Houtzagers 1999, 76, 2008b, 305-306, Hadrovics

1974, 192-193).

31.  The Isg ending o f a-stem nouns is -u, =HP, *D o l.
32.  Lpl ending o f neuter pluralia tantum: no data.
33.  Ending -u or -i in po nimski/po nimsku: no data.
34.  Enclitic accusative of neuter personal pronoun (je or ga): no data.
35.  Habitual past with htlli: su htlli govorit, =H P =Dol.
36.  The word ‘ear’ is uho, =H P =Dol.
37.  German ‘bei uns’ is almost always attested u nas,  *   all other Burgen

land dialects; pri has also been attested but less frequently, *H P *D o l.

38.  ‘Always’ is senek, =H P =Dol.
39.  ‘He goes’ is grie and not ide, =H P *D o l.
40.  The word *poredan ‘bad, evil’ was not known to my informants, =H P

=Dol, but there is a word sporiedan ‘good’.

41.  The word for ‘forest’ is luk, *H P  *D o l =HiFe =Stoji.
42.  The word for ‘rooster’ is petieh, =H P =Dol.
43.  The word for ‘Monday’ is pondll'ak, =H P =Dol.
44.  The word for ‘kitchen’ is viezi Lsg (=H P  =D ol).
45.  The word for ‘lazy’: no data.
46.  The  word  for  ‘spring’: protolice, prolice,  etc.  are  not  recognized.  The

M C word is a loan from Czech: jaro *H P *D o l.

47. The word for ‘wedding’ is plr, =H P =Dol.
48. The word for ‘dog’: no data.

Scando-Slavica 55, 2009

[Texas

Sep

The Dialect of the M oravian Croats

163

49. The word for ‘harrow’ is branämi Ipl, *H P (except two villages in the

north o f the Burgenland) = D N V = ChG (see Takac 2004, 28) =Dol.

50.  The word for ‘acre (plot o f arable land)’ is lapât, =H P =Dol.
51.  The word tiek means ‘corn’ =H P *D o l; zitâk is ‘wheat’.
52.  The word for ‘farmer (richer than average)’ is paur, =H P =Dol.

9. Conclusions

The dialect o f the Moravian Croats is on the verge of extinction, but can still

be studied, since there still are a number o f fluent speakers.

The dialect has a “troakcenatski sistem” (" ” ') in stressed syllables and an

opposition between long and short vowels in the first posttonic syllable. It is

very close to the dialects o f the Haci and Poljanci and I agree with Newek-

lowsky’s assumption

3 6

that we can classify it into the same group of Burgen

land Croatian dialects

. 3 7

As can be expected, the dialect has been influenced lexically by Czech,

also on the level o f conjunctions and prefixes, e.g. a ‘and’, nebo ‘or’ and the

prefixes vi- and roz-. Cf. also järo ‘spring’. On the morphological level it took

over the 1pl ending -me from Czech. There are also some lexemes that are
not from Czech but that are probably restricted to the M C dialect, e.g. mäja
‘mother’, roba ‘woman’.

But there are more characteristics that distinguish it from the Haci and

Poljanci dialects. Two of them are important from the point o f view o f his
torical accentology:

. There was no stress retraction from internal syllables to a preceding

long syllable. This suggests that what Neweklowsky saw as the partielle

36  In  Neweklowsky's  words:  "Diese  Mundarten  würden  sprachlich  eigentlich  kein  ei
genes  Kapittel verdienen,  da sie  [ ...]   den Mundarten der Poljanci und der Mundart von
D[evinska] N[ova] V[es]  sehr ähnlich sind”  (Neweklowsky 1978, 89-90).
37  There is very little historical evidence as to where the Croatian inhabitants of the three

villages originally come from. Pavlicevic (1994, 29-44) presents a summary of this evi

dence  and his  view  on the  matter.  I  expressed  my opinion  on the problem  of the  prov
enance of the Burgenland  Croats in Houtzagers  2008b,  296-302.  I think that there is no
reason to believe that the original  dwelling place of the Moravian  Croats must be looked
for in another area than that of the Haci and Poljanci. However, as I have tried to show, it is

highly improbable that there is a one-to-one correspondence between areas in the old and

the new homeland of the Burgenland Croats (Houtzagers 2008b, 301-302).

Scando-Slavica 55, 2009

[Texas

Sep

regressive Akzentverschiebung must not be seen as one but several devel
opments, which did not necessarily occur simultaneously.

. The material contains strong evidence that the generally accepted

lengthening o f *e and *o before the stress retraction operated on *a as
well. It is my opinion that this lengthening o f *e, *a and *o took place

not only in the dialect under discussion, but in a larger part o f the Bur
genland Croatian area.

Eight  other  characteristics  have  a  less  systematic  character,  but  must  be
mentioned  all the  same because they play a  role  in  the  classification  o f the
Burgenland Croatian dialects  (cf. the maps in Neweklowsky 1978  after page

376).

.  Prothetic j-, d’- or dž- are absent before initial i-.
.  The habitual past is formed with htili LPplm.
.  Word-final *-l > -f.
.  The weakjer is reflected a in *dbrnsb: danas.
.  The form of the verb ‘pull’ is vucie PR3sg.
.  The Lsg ending of masculine is -i, no free variation with -u.
.  The infinitive ending -i is optionally retained.
.  ‘At our place’ is u nas, ‘wood’ is luk.

The  dialect  also  has  a  few  characteristics  o f its  own  on  the  subphonemic
level:  the phonetic realization o f the  diphthongs and the phoneme  /h /   and
the  relative  rareness  o f the  nasalization  o f vowels  in  the  vicinity  of nasal

consonants.

References

Breu, J. 1970. Die Kroatensiedlung im Burgenland und den anschließenden Gebieten.

Vienna: Franz Deuticke.

H adrovics, L. 1974. Schrifttum und Sprache der burgenländer Kroaten. Vienna:

Ö sterreichische A kadem ie der W issenschaften.

H erben, J. 1882. “Tri chorvátské osady na M oravě”. Časopis Matice moravské 1 4 :1 -2 5 .

H outzagers, Peter. 1999. The K ajkavian Dialect of Hidegség and Fertohomok. Studies

in Slavic and G eneral L inguistics 27. A m sterdam : R odopi.

--------- . 2008a. “O n the Cakavian D ialect o f K o ljn o f near Sopron” In Evidence and

Counter-Evidence. Essays in Honour o f Frederik Kortlandt. Studies in Slavic and

G eneral Linguistics 32, edited b y A . Lubotsk y et al.

164

Peter Houtzagers

Scando-Slavica 55, 2009

[Texas

Sep

The Dialect of the M oravian Croats

165

--------- . 2008b. “O n Burgenland Croatian Iso glo sses”. In Dutch Contributions

to the Fourteenth International Congress o f Slavists, Ohrid, September 1 0 -1 6 ,

2008. Linguistics. Studies in Slavic and G eneral Linguistics 34, edited by Peter

H outzagers et al.

Jakubinsky, L. 1925. “D ie V ertretung des urslavischen ě im C akavischen”. Zeitschrift

fü r slavische Philologie 1 :3 8 1 -3 9 6 .

Jeřábek, R. 1991. M oravští charváti: dějiny a lidová kultura (Antologie). Brno.

Koschat, H. 1978. Die cakavische M undart von Baumgarten im Burgenland. W ien:

Verlag der Ö sterreichischen A kadem ie der W issenschaften.

Langston, K . 2006. Čakavian Prosody: The Accentual Patterns of the Čakavian Dialects

of Croatian. B loom ington: Slavica.

Lončaric, M . 1998. “Prilog istraživanju govora m oravskih H rvata” Radovi Zavoda z a

slavensku filologiju 3 2 :1 0 7 -1 1 5 . Zagreb.

Lawitschka, J. 2005. Lipo naše selo: Paměti jihomoravského Chorvata. Prague: A equitas.

M iklosich, F. 1890. “Eine Sprachprobe des kroatischen D ialektes von N euprerau bei

N ikolsbu rg” Archiv fü r slavische Philologie 1 2 :3 1 7 -3 1 9 .

M ilčetic, I. 1898. “O m oravskim H rvatim a” Vienac 3 0 :4 1 0 -4 1 3 , 4 4 2 -4 4 4 , 4 5 8 -4 5 9 ,

4 7 4 -4 7 6 , 4 8 7 -4 9 2 , 5 2 1 -5 2 3 , 5 5 7 -5 5 8 .

Neweklowsky, Gerhard. 1978. Die kroatischen Dialekte des Burgenlandes und der

angrenzenden Gebiete. Schriften der Balkankom m ission, Linguistische A bteilung

25. W ien: Verlag der Ö sterreichischen A kadem ie der W issenschaften.

Novik, A. 2005. “Chorvati na Moravě. Poválečný vývoj. Jazyk” In Lawitschka 2005, 7-14.

Pavličevic, D. 1994. Moravski H rvati: Povijest - život - kultura. Zagreb: H rvatska

sveučilišna naklada.

Takač, Ferdinand. 2004. Rjecnik sela Hrvatski Grob. Z agreb: Institut za hrvatski jezik

i jezikoslovlje.

Vážný, V. 1934. “M luva charvátských osad v republice Č eskoslovenské”. Českosloven

ská vlastivěda vol. 3, 5 1 8 -5 2 3 .

Scando-Slavica 55, 2009