Australia and New Zealand Slavists’ Association

New Zealand Slavonic Journal

The development of declensional endings in nasals in Common Slavonic

Author(s): ROBERT ORR
Source: New Zealand Slavonic Journal, (1986), pp. 83-91
Published by: Australia and New Zealand Slavists’ Association
Stable URL: http://www.jstor.org/stable/40921297 .
Accessed: 12/06/2014 14:19

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

Australia and New Zealand Slavists’ Association and New Zealand Slavonic Journal are collaborating with
JSTOR to digitize, preserve and extend access to New Zealand Slavonic Journal.

http://www.jstor.org

This content downloaded from 195.78.108.40 on Thu, 12 Jun 2014 14:19:17 PM

All use subject to JSTOR Terms and Conditions
 NewZealandSlavonicJournal
1986

ROBERT ORR
The developmentof declensionalendings
in nasals in Common Slavonic

I. Introduction:the phonologicalframework
In this shortarticleit will be shownhow the development of
nasalvowelsin CommonSlavonic(CS) mighthavehad an effecton
theevolutionof thenominalmorphology ofCS. The articleis inten-
ded to complementThe effectof the loss of final*-s on the
evolutionof CommonSlavonicnominalmorphology', NZSJ, 1985,
referred
pp. 17-32, (hereafter to as S).

Althoughthequestionof theevolutionofnasalvowelsinCS has

been discussedin greatdetailforovera century,withseveralingen-
ious hypothesesbeingadvanced,no definitivesolutionhasyetbeen
proposed.For a briefselectionof someoftherelevant see
literature,
Shevelovl971:52fn.
In thispapera new pointof departure willbe taken,usingthe
framework outlinedin Orr 1983, and further
developedin subse-
quentarticles.
Mostof theauthoritiescitedacceptwithoutapparentquestionfor
CSthetraditionally
pwposedAuslautgesetze(soundchanges peculiarto
thefinalsyllableof a word)*-om> *-u and * -5ns> *-u(-y). I have
attemptedto show (Orr 1983:73-103) thattheseAuslautgesetze
are unnecessary of CS nominaland verbal
for the reconstruction
morphology.
At thispointwe shouldoutlinethereconstructivephonological
analysiswhichwillbe used as a basisfortheargumentation
in this
paper.Two pointswillbe discussedin detail:
1) Fronting ofbackvowelsafter'j' exceptCS *a
*-ja > *-je-,e.g.,*palje< *palja
*-ju-> *-#,e.g.,*jiga< *juga
but
*-ja-> *-ja-,e.g.,*zemlja< *zemlja

83

This content downloaded from 195.78.108.40 on Thu, 12 Jun 2014 14:19:17 PM

All use subject to JSTOR Terms and Conditions
Thereis a greatdeal of evidenceto suggestthatCS *a did not,in
fact,undergofrontingafter"j":

a) The syllable*-ja- is extremely commonin CS declen-

sional endings,and also in initialand internalposition,e.g., the
suffix*-janina}
b) The vowel*£ is normallyassumedto havebeen- tense,
whereasthevowel*a is assumedto havebeen+ tense.Tensevowels
wouldbe lessproneto fronting in thisinstance.
c) CS *& < IE*e becomes*a afterC,5, 2, j, e.g.,*kedit>
*Wtt > *Zadii;*beg+ e + tT> *bMM > *b&ati; *stttj + e + tT>
*stojati.
d) In theGlagoliticalphabet'ja' and '8' aredenotedbythe
samesymbol,i.e,/'
As partof the discussionof 1), I wouldlike to outlinea hypo-
thesison thereconstructionoflongdiphthongs in CS, especiallythe
fate of *-VN combinations in word-finalsyllables.Some scholars
(e.g., Schmalstieg1965, repeatedby him in numeroussubsequent
articlesand books) would arguethatall originallylongdiphthongs
wereshortenedin the Balto-Slavonic period,whereasotherswould
reconstruct long diphthongsfor some word-finalsyllables(e.g.,
Shevelov1964: 24-5). Mathiassen 1974: 4-5,following Stang1966:
186, adducessomeevidencewhichseemsto showthatshortening of
long diphthongs took place independently in Balticand Slavonic.
I would like to suggestherethatsomelongdiphthongs (especi-
ally *-VN sequences)werepreserved in CS untilthedevelopment of
nasalvowelsand themonophthongisation of *-iand *-u diphthongs,
whenany remaining long diphthongs would havefinallycoalesced
withtheirshortcounterparts (e.g.,* -SNand♦-aNwouldhavemerged
in *-p). I would also like to suggestthatall examplesof *-j<g(j +
backnasalvowel)can be eithertracedbackto CS *-jaN orexplained
as theresultof theworkings of specificmorphological or analogical
factors.2

2) The development of all *-VN clustersto *-y,withsimultan-

eous obliteration
of any lengthdistinctionthatmay haveexisted;
e.g., *-aN (*-8N) > *-o; *oNs (*-SNs) > *-p; *-jsN > *-jo;
*-SN(s)>*-g; *-jSNs>*-jg.
Nearlyall the authoritiesrely on some sortof morphological
explanationto accountforall the formsactuallyattested.Bruckner
1910: 369 describesthe situationas "Ein RattenkonigvonWider-
spriichen",and goes on to pointout that,as an example,forfive

84

This content downloaded from 195.78.108.40 on Thu, 12 Jun 2014 14:19:17 PM

All use subject to JSTOR Terms and Conditions
formsattestedin Greekwithan ending-on, OCS has a different
endingforeach,e,g.,Gk Vferon,OCS vedti(1 sg.); Gkgferon, OCS
vedp (3 pl.); Gkferon,OCS vedy(neut.nom.sg.p.p.a.); Gk zugon,
OCS jrgo(*-tf-stem neut.nom./acc.sg.);Gk Son, OCS jaje (♦-jo-
stemnom./acc.sg.).

Althoughsomeof the formscitedby Bruckner can be explained

as the resultof regularphoneticand phonologicaldevelopment, it
would seem morelogicalthatmorphological and analogicalfactors
shouldbe used to explainmanyof theapparently anomalousforms
cited above. As will be discussedbelow,thepicturebecomeseven
more complicatedwhen we examinethe formswith morpheme-
initial'j'

II. The development of *-VN endingsin CS nominalmorphology

In thissectioncertainendingswillbe discussedin pairs,following
the formatintroducedin S. I will attemptto demonstrate that,
withineach pair,the endingswouldhavediffered fromeach other
onlyin one feature,and thatthe obliteration of thisfeaturecould
havesetin traina morphological/analogicalrestructuring.
Shevelov1971: 52 fn. calls fora "studywhichwould takeup
involvedfora thorough
each of the categories separateexamination,
undertaken, however,againstthe broadestpossiblebackground of
the phonemicand morphologicalsystemsof the time,without
mixingup phenomena whichbelongto different epochsandwithout
atomising the material".
Although thispresent is byno means
article
intendedto fulfilsucha role,an attemptwillbe madeto accountfor
thesevenproblematic formslistedby Shevelov1971: 51 withinthe
framework usedhere:

a) Ace. sg.of * -ja-stems,

e.g.,OCS zemljg
b) *
Ace. pl. of -j5-stems,e.g.,OCS mgig
c) Gen.sg.// nom/acc.pl. of * -ja-stems,e.g.,OCS zemljg
d) 1 sg.ofverbsof classIII, e.g.,OCS koljq
e) 1 sg.ofverbsof classIV, e.g.,OCS moljq
f) 3 pl. ofverbsof classIII, e.g.,OCS delajgtu
g) Nom.sg.masc./neut. of p.p.a. ofverbsof ClassIII, e.g.,
OCS kryft
As d), e), f)>and g) belongto conjugation
ratherthanto declen-
siontheywillbe treatedonlybriefly here.The formsa), b), andc)
abovewillbe discussedin muchgreaterdetail,in so faras theyare
an integralpartoftheproblemunderdiscussion in thisarticle.

85

This content downloaded from 195.78.108.40 on Thu, 12 Jun 2014 14:19:17 PM

All use subject to JSTOR Terms and Conditions
 The formsto be examined are as follows:

1) The accusative singular endings of the masculine *-8 and

*-jd stemsand thefeminime*-a- and *-fa-stems

*gVrd-d-m(*-X-m) *gen-a-m
*vod-j8-m (*-jX-m) *zeml-ja-m

This group of endings has been the topic of much discussion in

Slavonic historical phonology and morphology. Although many
scholars(e.g., Shevelov 1964: 156-7) have not treatedthem as part
of one group, it is essential for the purposes of thisarticleto do so.

The firststep is to establish a reconstructiveevolution for these

endingsin CS. The followingscheme is proposed:

Stage I: Early CS/Late NorthIE (already splitby variousisoglosses).

M F
*-o- *-jo- *-a- *-ja-

*-a-m *-ja-m *-a-m *-ja-m

Stage II: Vowel frontingafter'j'

*-a-m *-je-m *-3-m *-ja-m

Stage III: Nasalisationand the loss of finalV

*-g *-jg *-? H?
State III above is a representationof what the subsystemdepicted
above mighthave been without any morphological/analogical chan-
ges. We can see, however,that with nasalisationand the concomitant
loss of lengthdistinctions,such an evolution would have destroyed
vital genderdistinctions.When we consider the loss of final*-s(see
S), it seems very probable that these two sound changes between
them would have obliteratednumberdistinctionsalso. For instance,
*-p could have denoted masculine accusative singular,masculine
accusative plural, feminineaccusative singular,and feminineaccu-
sative plural. In order to preservethese distinctions,I propose thata
radical reshuffling of endingstook place.

The ending *-g was generalisedas feminineaccusative singular

ending, while the masculine *-6 and *-jo-stems took the ending
*-tl fromthe *-u-stems.The ending *-jg was thereforeeliminated
from the masculine accusative singular,and specialised as an accu-

86

This content downloaded from 195.78.108.40 on Thu, 12 Jun 2014 14:19:17 PM

All use subject to JSTOR Terms and Conditions
sative plural ending for the masculine *-jti-andfeminine*-ja-stems,
as genderdistinctionsdo not seem to have been nearlyas important
in the pluralas in the singular.

At this stage we shouldjustifythe reconstructionof a long vowel

in the CS formof the accusativesingular.In 1 1) above it was hinted
brieflythatsome long diphthongswere preservedlater than othersin
CS. I would like to suggestthat the feminineaccusative singularof
the *-a-stemsand *-ja-stemswould have been one of those.

There are some quite strongargumentsforreconstructing a short

vowel for this ending; see, e.g., Lehmann 1958: 192. Nevertheless,
most IE languages seem to have generalisedthe long vowel by the
timeof theirfirstattestations.Greekeven has variantforms,possibly
pointing to the coexistence of long and short stem-vowelsat some
period,e.g., time- timen(< *-lm); doxa - doxan (< *-8m).

However, in Slavonic it seems possible to trace an earlytendency

to distinguishthe gender of nouns in the accusative singular(see
Orr 1983: 135-6). Whenwe considerthat IE *5 and *a would have
mergedin Slavonic (there is, indeed, no hard evidence that Slavonic
ever had an* 8/* a distinction),an argumentin favourof the preser-
vation (or introduction)of lengthin thisformappears. The lengthin
the feminine*-a-stemform(*-am) would have servedto distinguish
it from the masculine *-o-stem form (*-am), and thus this form
probablyescaped the generalshorteningof long diphthongsproposed
for Balto-Slavonic by Schmalstieg. Later on, when the distinction
was again threatened,this time by nasalisation,the morphological/
referredto above would have taken place.
analogical reshuffling

In the plural,however,the feminine*-ja-stemsseem to show the

reflex of a short vowel (-jg < *-jeNs < *-jaNs). I propose that the
reason for this is as follows. Gender distinctionsare generallyless
importantin the plural than in the singular(see, e.g., Stankiewicz
1977: 169), and thereforewe mightexpect that the factorswhich
protectedthe stem-vowelin the singularfromshorteningwere absent
in the plural.

A table of the proposed changesis givenhere:

M F
*-o- *-jo- *-a- *-ja-

A.S. *-g *-j£ *-<2 *-jg

API. *-0 *-J£ *-g *-j^

87

This content downloaded from 195.78.108.40 on Thu, 12 Jun 2014 14:19:17 PM

All use subject to JSTOR Terms and Conditions
 New formsafterthemorphological/analogical
changesproposed
above:
A.S. *-u /*-ji *'2 *"J9
(from*-tl-stems)
A.P1. *-y ^ - ■
- **-y
Hg *-jg
singularof the neuter^men-
2) The nominative/accusative
stemsand the nominativesingularof the masculine*-men-stems
*i-men *ka-m£n(*ka-mins)

If we reconstruct the masculinenominative singularendingfor

this declensionas *-men,it would seem thatthe masculineand
neuterformswould have been distinguished fromeach otheronly
by the featureof length. In this
fact, maywell havebeen another
instanceof preservation of lengthin originallylong diphthongs in
orderto maintainvitaldistinctions of numberor gender.Withthe
development of *-vN to>-Y, thegenderdistinction betweenmascu-
lineand neuterwouldhavebeen in dangerofvanishing in thisstem-
class,and therefore it was rescuedby the introduction of various
distinctively masculine endings. The morphologicalreanalysis
proposedforthisgroupof formsis morecomplicatedthanforthe
*-5- and *-a-stemaccusativesingular endings:thefollowing set of
endingsseemto havebeenused:
a) -my,e.g.,OCS kamy,plamy
b) -mem,e.g.,R kremen' SC pramen
c) -myku,e.g.,?jgczmyk,Bgplamik
d) -mu,e.g.,Ka kam,R stam
wandadvantages,of recon-
In S I consideredthe possibilities,
structingthe masculine endingas *-miensratherthan as *-men.
In a sense,it does not reallymatterwhichone we reconstruct,
as
both would have fallentogetherwith the neuterending*-m£n.
I have alreadydealtwiththeproposedmorphological reanalysis
elsewhere,see Orr1983: 170-183, S, and forthcoming.

III. The verbalformscited by Shevelov

For the sake of completeness, in orderto justifytheproposed

development * -jan> * -J9forCS, therelevant verbalformsshould
be brieflydiscussed.Actually,Shevelovhimselfgivesa justification
for reconstructing the firstsingularof verbsof classesIII and IV

88

This content downloaded from 195.78.108.40 on Thu, 12 Jun 2014 14:19:17 PM

All use subject to JSTOR Terms and Conditions
witha longvowel: he statesthatthe *-N (or possibly*-m RAO)
was introducedinto this form(originallysimply*-o) afterthe
shorteningof originallylong diphthongs.(1971: 52-3).) The
nominative
singular ofthepresentparticiple
masculine/neuter active
can be explainedin conjunctionwith the firstsingularformas
follows:
The first
singularcanbe reconstructed as *-5N,following Shevelov
1971: 52-2. The masculineformof the presentparticiplecan be
reconstructed as *-6n(t)s, and the neuteras *-ont. Afterthe
relevantvowelshad been frontedfollowing 'j', and also takinginto
accountthemerger nasal sonantsin *-N and the
of all syllable-final
simplificationof *-nt to *-N, one can reconstruct the endingsas
*-6N, *-j£Ns,and *-jfcNrespectively forclassIII verbs.Nasalisation
and thelossof final*-s wouldhavecausedthemasculine andneuter
formsto merge,and thismerger wouldhavebeentoleratedbecause
genderdistinctionswouldhavenotbeenas important in theparticiple
as theywerein thenoun.

a fairlyconvincing
Shevelovalso offers morphological explanation
for the thirdpluralendingof class III verbs.This is, indeed,one
instancewheretheexistenceof *-jg-shouldbe ascribedto morpho-
logicalfactorsratherthanthe reconstruction of an originallylong
vowel.Shevelov,briefly,sees the *-jg (insteadof *-jg) in thisform
as comingfromclassesI and II, underpressurefromtherestof the
presenttenseparadigm.

may be summarised
These developments in thefollowing
table:
1stsing. pr.p.a.(m.) pr.p.a.(n.) 3rdpl.
*znajoN(* -IN) *znajdn(t)s *znajdnt *znajtintit
Vowelfrontingafter'j' andlossoff after4N'
*znaj6N *znajeNs* *znajeN *znajentu*
Nasalisation
andlossof final*-s

*znajo *znaj% *znaj% *znajqtiis

It seems,therefore, thatthehypothesis of CS *a notundergoing
frontingafter'j' can be used to explainquiteplausiblytheproble-
maticformscitedbyShevelov.

IV. Summary
and conclusions
It seemsfromtheabovethatthenasalisation
in CS didnothavea
verygreateffecton the development
of the nominaldeclensional

89

This content downloaded from 195.78.108.40 on Thu, 12 Jun 2014 14:19:17 PM

All use subject to JSTOR Terms and Conditions
system,except in the accusative singularof *-S- and *-a-stemsand
possibly in the nominativesingularof * -men-stems.Compared to
the effectof the loss of *-s (see S), the effectof nasalisationwas
much more restricted.

It also seems that thereis strongsupport for the hypothesisthat

* a was not frontedafter*j' The formscited by Shevelov 1971 can
easily be fittedinto thisframework.

Notes:

1. The suffix *-janitiwas partiallymergedwith another,originallydistinct,

suffix*-enintf;see Shevelov 1964: 259.
2. Shevelov 1964: 262 fn and Darden 1970: 459 state thatthe sequence *jp
is neverattestedin roots, which could possiblylend supportto the hypo-
thesis of a general shorteningof long diphthongsin CS (roots would be
less subject to morphologicaland analogicalpressuresthanendings).There
are a very few examples of roots containing*jo (e.g., pajofina, 'spider's
web1,attestedin Suprasliensis626), but thesecan all be explainedwithout
recourseto the preservationof long diphthongsin internalsyllables.
3. It is possible that the masculineformof the participlepreservedits length
to distinguishit fromthe neuter.In my doctoral dissertation(Orr 1983:
196-7) I proposed thatvariantforms*znajdnt(s) and *znaj&nt(i) existed
at some stage,withthe loss of final*-s servingas a catalystforthe general-
isation of *znaje*nts> *znafe A formsuch as *znajont(s) > *znajo c would
have fallentogetherwiththe firstsingularform.
4. The existence of *znajentu was probably transitory,with *znajontU
, swiftlybeing reestablished.One possible confirmationfor the one-time
existenceof *znajentumay be foundin the behaviourof the oblique forms
of the participle,which normallyhave *jg, e.g., *znaje - genitivesingular
*znajoj?ta.I would like to suggestthat the *J£sequence in the participle
was reintroducedwhen *znajdntii ousted *znajeMtUin the thirdperson
plural. The form*znaj%would have been preservedbecause of the danger
of confusion with other forms if it, too, had reintroduced*jg. In the
participle,however,the recoveryof *jg was neveras fullas it was in the
thirdplural; numerousforms(e.g., bore - poborfitaago) are attestedwith
*e throughout.
5. Barnstead1984 also sees thisformas arisingfroma morphologicallevelling.

Bibliography:
Barnstead,J., 'The Interrelationshipbetween the Development of the
Common Slavic Accentual Systemand EarlyVowel Nasali-
zation in Common Slavic*; paper presentedat the Annual
Meeting of the Canadian Association of Slavists,Guelph,
Ontario.June 1984.
Brlickner,A., 'Drei urslavischeNasalvokale',Zeitschriftfur Vergleichende
Sprachforschung, vol. 43, 1910, pp. 369-73.
Darden,B.J., The Fronting of Vowels after Palatals in Slavic', Papers
fromtheSixthRegionalMeeting,ChicagoLinguisticSociety,
(1970), pp. 459-70.
Lehmann,W.P., 'On Earlier Stages of Indo-European Nominal Inflection',
Language,vol. 34, (1958), pp. 179-202.

90

This content downloaded from 195.78.108.40 on Thu, 12 Jun 2014 14:19:17 PM

All use subject to JSTOR Terms and Conditions
Mathiassen,T., Studien zum Slavischenund Indo-EuropaischenLangvokal-
ismus.Oslo: Universitetsforlaeet. 1974.
Orr,R.A., The Role of the *-tf-stems in the MorphologicalRemodel-
ling of the Common Slavic Nominal Declensional System1,
unpublishedPh.D. dissertation.University ofAlberta. 1983.
, 'Some developmentsin Final Syllables in Slavic - A New
Synthesis (A Comment on Jasanoff, Schmalstieg, and
Kortlandt)' forthcoming.
Schmalstieg,W.R.,'Slavic o- and a-stemAccusatives', Word,vol. 21, (1965),
pp. 238-43.
, MorphologicalConsiderationson the Balto-SlavicProblem',
AmericanContributionsto theNinthInternationalCongress
of Slavists, (Kiev, 1983), Columbus: Slavica, vol. I, pp.
269-76.
Shevelov,G., A Prehistoryof Slavic, Heidelberg:Carl Winter,1964.
, 'On Endings with Nasal Consonants after Palatal and
Palatalised Consonants1,in Teasers and Appeasers: Essays
and Studies on Themes of Slavic Philology, Munich: Wil-
helm Fink Verlag,1971, pp. 50-9.
Stang,C, VergleichendeGrammatikder Baltischen Sprachen, Oslo:
Universitetsforlaget. 1966.
Stankiewicz,E., 'Typological Invariance and Variation of Grammatical
Categories:The Categoriesof the Slavic Noun1,Linguistica
Generalia,vol. I, (1977), pp. 167-71.

91

This content downloaded from 195.78.108.40 on Thu, 12 Jun 2014 14:19:17 PM

All use subject to JSTOR Terms and Conditions