A Variational Model of the Indoeuropean Consonant System

Author(s): Gevorg B. Djahukian

Source: Historische Sprachforschung / Historical Linguistics , 1990, 103. Bd., 1. H.
(1990), pp. 1-16
Published by: Vandenhoeck & Ruprecht (GmbH & Co. KG)

Stable URL: https://www.jstor.org/stable/40848925

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.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

Vandenhoeck & Ruprecht (GmbH & Co. KG) is collaborating with JSTOR to digitize, preserve
and extend access to Historische Sprachforschung / Historical Linguistics

This content downloaded from

180.176.178.34 on Sun, 23 Jan 2022 14:02:23 UTC
All use subject to https://about.jstor.org/terms
 A Variational Model of the Indoeuropean
Consonant System

1. To reconstruct a parent language is to elaborate such a model

from which the models of the daughter languages can be regularly
derived. The elaborated model must correspond, on the one hand, to
the real object, the parent language, and, on the other hand, to the
models of daughter languages which are the organic continuation of
the parent language. It is the degree of the two types of correspond-
ence that determines the degree of exactness of the model.
To understand and interpret the relation of the reconstructed
model and the real parent language several conventional stages have
been distinguished in the history of comparative linguistics.
Since the first attempts of the language reconstruction, the ques-
tion of such relationship has not been raised for a considerable pe-
riod of time: the conventionality of the reconstructed language
model was not sufficiently perceived, and no distinction was made
between the model and the real object, though in search of ways to a
more precise reconstruction intuition suggested that there should be
some difference.
This intuition has developed into a principle when new facts and
new discrepancies in the traditional reconstruction were disclosed,
chronological stratification was used and archaic strata revealed.
The traditional reconstruction has been shown to be extremely con-
ventional and some scholars even entertained the idea that it was im-
possible to conceive the reality concealed behind the reconstructed
data. A. Meillet considered the correspondence between the attested
languages to be the only reality dealt by comparative grammar. The
conventionality of the reconstructed system has become more evi-
dent with the rise of the laryngeal theory based on the Saussurian
conception of sonant coefficients, and the disclosure and interpreta-
tion of the facts of the Hittite language. Brugmann's traditional re-
construction of the Indoeuropean phonetic system, particularly the
consonantism - almost all the phonemic series of the reconstructed
system - started to seem doubtful.
The new approach to the relationship of the model and the real
Hist. Sprachforsch. 103, 1-16, ISSN 0935-3518
© Vandenhoeck & Ruprecht 1990

This content downloaded from

180.176.178.34 on Sun, 23 Jan 2022 14:02:23 UTC
All use subject to https://about.jstor.org/terms
2 Gevorg B. Djahukian

object has become indispensable in the period

ment of structural linguistics, which interpre
sufficing system, and, particularly, in the er
theory of models. A certain fetishism of mo
determining the degree of affinity of relate
gree of their archaism, preference was given
dependent of their realizations, the latter be
relationship of the model and its component
components. That preference, for instance, d
tion of the Armenian language in the Indoeu
the affinity of its special models with the r
pean model was considered a matter of parti
out sufficient consideration of their realizat
in all the empty slots of the system was obs
The typological evaluation of the results of
other principle introduced as a reaction to th
attention was paid to the definite nature of
and series of the reconstructed Indoeurop
case analogues of the reconstructed model
known languages along with their actual in
ment was intensified after R. Jakobson's well
at the Eighth International Congress of Ling
there too, as is often the case, a certain extre
typological probability is presented as reality
the cause of the system's naturalness the me
of analogous systems in current languages is
2. If we consider only the stops, leaving asi
Brugmann's fricatives of the fi, fih, ct, dh t
comparative linguistics, by "traditional" re
imply mainly the following systems (for de

(1) the Brugmann-Szemerényi system, which

labial, dental, palatal, velar, and labiovela
voiceless, as well as aspirated and non-aspir
bh dh gh gh g«h
b d g g g«
p t k R k«
ph th kh th k»h;
(2) Meillet's system in which the series of p
on the ground that not a single Indo-E
poses of palatal, velar and labiovelar sto

This content downloaded from

180.176.178.34 on Sun, 23 Jaon Thu, 01 Jan 1976 12:34:56 UTC
All use subject to https://about.jstor.org/terms
 A Variational Model of the Indoeuropean Consonant System 3

bh dh gth gwh
b d gl g-
p t kj kw
ph th kth kwh;

(3) Lehmann's system in which the series of

moved having reflexes only in a few langu
ment of the combination of plain voiceless
quent laryngeals:

bh dh gh gwh
b d g g"
p t k kw.

Leaving aside different inte

aspirates, as well as the quan
and plain voiced stops, one m
consonant system began ju
system. A. Martinet (1953)
rise of typological difficulti
removed from the system, si
supposes the presence of the
The second important groun
"traditional" systems was th
system. The typological prob
plain voiced stops was presen
labial member *b has been
ble only for the series of
stops: there is a considerab
American Indian languages, i
p' is absent. With regard to
quency, an attempt was mad
which, as T. V. Gamkrelidze
"in the systems of stops hav
labial phoneme /b/ is functi
relationship is established on
of the /b/ phoneme as comp
language systems (Gamkrel
This is when the "glottalic t
pioneer, appeared. Stating th
postulated the "empty slot
voiced stops as the result of
talized stops that had lost th

This content downloaded from

180.176.178.34 on Sun, 23 Jan 2022 14:02:23 UTC
All use subject to https://about.jstor.org/terms
4 Gevorg B. Djahukian

3. The number of different reconstructions, a

of these difficulties and connected with the st
rates and voiced labial stops, increased in the
T.V.Gamkrelidze and V.V.Ivanov, O.S.Shirok
Hopper, J. Rasmussen, A. Bomhard, A. Haudric
lowed in close succession. Three-stop-system
common for all of them: the series of voiced s
stituted by that of glottalics (Gamkrelidze-Iv
per, Bomhard, Haudricourt, Normier), the seri
was left unchanged whereas the series of voi
substituted by that of voiceless aspirates (Gamk
first version, Emonds, Normier), or vice versa,
stops remained unchanged, and the series of
instead of the series of voiced aspirates (Shir
per called lai "murmured" marking it Idi). B
noted that (i) Rasmussen achieved typological
what another way, substituting the series of vo
of voiced stops, the series of voiced stops by th
the series of voiceless stops by that of emphatic
geal/tense) consonants, thus acquiring T, t, d in
Gamkrelidze and Ivanov afterwards modified the reconstructed
system, considering the feature of aspiration of voiced and voiceless
stops to be facultative:
I II III
(p') p/ph b/bh
t' t/th d/dh
k' k/kh g/gh
k'« k»/kh g^/g^h1)
Thus it follows that the problem of the status of *b in the IE con
sonant system acquires a special importance in Modern Indo-Eu
pean Linguistics.
4. Of the two reasons for the "revolution", the first, according
which there cannot be a system having voiced aspirates and not h

1) In fact, if we put into brackets the other members with an occurrence f

quency lower than that of *b (Jucquois 1966), we shall have the followin
system:
I II III
(p') p/ph b/bh
t' t/th d/dh
k' k/kh g/gh
k'y (k«/k«h) (g»/g«h)

This content downloaded from

180.176.178.34 on Sun, 23 Jan 2022 14:02:23 UTC
All use subject to https://about.jstor.org/terms
 A Variational Model of the Indoeuropean Consonant System 5

ing voiceless aspirates, proved to be incorrect, and was a result of a

hasty generalization. The scholars who had begun the revision did
not pay attention to Martinet's and Jakobson's tentative style while
establishing the universais, or to the articles of A.B.Stevens (1968:
22, 24 et al.) and R.A.Blust (1973: 309-324), who pointed out the
existence of such a system in Madurese and Kelabit.
The second reason, the functional "weakness" of the b phoneme
in Indoeuropean, brings about serious objections among other scho-
lars (Back 1979; Djahukian 1982, 1986; Diakonoff 1982; Haider
1983; Szemerényi 1985; Pisowicz 1987 et al.), whom the representa-
tives of the "glottalic theory", especially Gamkrelidze and Ivanov, do
not give reasonable answer trying to make cavils.
The main objections refer to the following.
At first the predecessors and the representatives of the "glottalic
theory" themselves did not differentiate between the low frequency
and the absence of the labial member in the series of IE voiced stops.
It is true that in the anlaut the phoneme *b is "weak", but to say that
it does not exist at all in this position would not be correct. If, for in-
stance, H. Pedersen considers the well-known example of *b in the
anlaut -OInd. balam "force, strength", Gk. ßeXxi(ov "better", Lat. dë-
bilis "weak", OSlav. bolje "more" - not to be convincing, disregard-
ing the Greek and Latin forms without a valid reason, M. Mayrhofer
(1963: 417) sees an undoubtful relationship between these forms; we
leave aside the possible Iranian, Armenian and Albanian parallels
suggested by different scholars.
But even if it is agreed that *b does not occur in the anlaut, the
"strength" of a phoneme is not determined by that position only; in
any position other than initial, the presence of the traditional pho-
neme *b does not meet any serious objections. O. Szemerényi (1970:
137), for instance, definitely mentions the normal frequency of *b in
the inlaut. Likewise, I have pointed out, as a typological parallel, the
status of r in the Armenian phonemic system: in the initial position it
occurs only in the borrowing rope "minute" and its dérivâtes, while
by its frequency it is the fourth among all the phonemes and the sec-
ond among the consonants, including the sonants.
The analysis of the data of some languages, in which the IE oppo-
sition *b vs. *bh has been preserved one way or another, provides
with sufficient examples for reconstructing the IE *b. For lack of
space we shall not analyse all the examples but rather cite a few con-
clusions of an article by I. B. Khlebnikov and V.I.Kolodko (1982)
concerning the status of the Germanic p (fromIE *b). Here is what
they write:

This content downloaded from

180.176.178.34 on Sun, 23 Jan 2022 14:02:23 UTC
All use subject to https://about.jstor.org/terms
6 Gevorg B. Djahukian

"It is improbable that the voiced aspirate bh could occ

that it might be of a later origin, not pertaining to the
ally the correspondences refer both to the Western and
lects. The supposition that the voiceless p in Germanic l
the combinative devoicing b < bh seems to be relevant f
considerable number of units having p that correlates w
117). And further: "Despite the fact that not everyth
cited parallels, and there is inconsistency in the treatme
ferent dictionaries, we may speak with sufficient confid
b > p regular correlation by the law of the first conson
sition." (p. 118).

In J.Pokorny's "Indogermanisches etymolog

(1959) *b occurs initially in root morphemes 30
tions 43 times. Out of this number b is alterna
*bh only in 23 cases or it may be identified as
sides, as a determinative it occurs another seve
The Indo-European stops. According to Jucq
quency of *b in the initial position makes 1.2%
cluding the vowels, sonants and fricatives), i. e
quency of V (1.1%), *¿* (0.9%) and *&>/* (0.6
the root, 0.3%, i.e. higher than the frequency o
(0.1%).
Thus it can be stated that, first of all, the thesis about the "weak-
ness" of *b should be accepted rather relatively: *b has a lower fre-
quency than most of the stops, but not all of them. The phonemes of
any language make a definite frequency scale; hence, the "weakness"
of a certain phoneme can be accepted only relatively. In this respect
the historical data of certain investigations are characteristic: as has
been mentioned, Gamkrelidze and Ivanov, at first, substituted the
series of IE plain voiced stops by that of glottalized voiceless stops,
later considered aspiration as an irrelevant feature for the other
series and postulated the series *b/bh, *p/ph, etc. (Gamkrelidze-Iva-
nov 1981; 1984), i.e. they included the series of plain voiced stops in
the IE consonant system but only as variants of voiced aspirates, ad-
mitting the facultative aspiration of the latter. However, this is of
little importance because, after all, *b is alternated not only with
*bh, but also with *p.
The thesis about markedness of b opposed to g as a universal is
based on hasty conclusions resulting from the insufficiency of the
facts considered. It is noteworthy that the relatively higher fre-
quency of b with respect to g is determined by Melikishvili (1974),
on whose data Gamkrelidze and Ivanov base their conclusions, tak-

This content downloaded from

180.176.178.34 on Sun, 23 Jan 2022 14:02:23 UTC
All use subject to https://about.jstor.org/terms
 A Variational Model of the Indoeuropean Consonant System 7

ing into consideration only the arithmetic mean, while in some of

them the reverse relationship is present: the Khevsur dialect of Geor-
gian (1.90:1.91), Old Georgian (1.50:1.98), Usigulian dialect of Sva-
nian (1.46:2.40), Megrelian dialect of Zanian (1.80:2.36), Abkhazian
(1.70:2.13), Osset (1.69:1.78); the relationship of b and g is likewise
in Hungarian (2.35:2.60) (Melikishvili 1974: 96, 100).
So, if systems with the reverse relationship are typologically possi-
ble, why then Indo-European could not be similar to them? The
relationship of b and g in Indo-European, according to Jucquois
(1961: 61), is 31:70; Gamkrelidze and Ivanov bring other figures too:
78 for b and above 250 for g.
5. The suggested models of IE consonant system by the followers
of the "glottalic theory" as a substitute to the traditional ones are far
from being flawless either typologically or otherwise.
The Indo-European system of stops with the globalized series is
not justified by the comparative historical data: the globalized series
is not represented in either of the oldest Indo-European languages;
the glottalics present in certain new languages and dialects appar-
ently have a secondary nature. The works of Kortlandt (1983; 1985)
who tries to reveal the traces of glottalization in the oldest IE lan-
guages, are, to say the least, inconvincing.
I would like to mention that I did not deny the possibility to re-
construct glottalized voiceless stops on the whole: I admit their exis-
tence not in the late Indo-European (on the eve of its disintegration)
but in its early period, and I think that "the data of nostratic linguis-
tics would indirectly testify in favour of such an admission" (Djahu-
kian 1982: 65).
The consonant systems reconstructed by the glottalists have other
drawbacks too. For instance, the model of Gamkrelidze and Ivanov,
in the first, simplified version, where the question of palatalized and
labialized velars was not considered, presented a system with only
marked series: voiceless glottalized and aspirated stops, both voice-
less and voiced. Besides, if one could find some examples of the
transition of voiceless glottalized stops into plain voiced and voice-
less ones, the possibility of the transition of voiceless aspirates into
plain voiceless stops in the majority of IE languages was not
supported by any convincing examples and proofs. The authors later
interpreted aspiration as a facultative feature, partially to overcome
the difficulty and to account for the existence of voiced aspirates in
many Indoeuropean languages.
Many representatives of the "glottalic theory" do not care about

This content downloaded from

180.176.178.34 on Sun, 23 Jan 2022 14:02:23 UTC
All use subject to https://about.jstor.org/terms
8 Gevorg B. Djahukian

the plausibility of the typologically reconstruc

cases, when typological parallels are pointed out
rect. So Gamkrelidze claims that the system rec
attested in Armenian dialects, with globalize
and voiced aspirates as allophones, and plain
voiceless aspirates as allophones" (Gamkrelidz
specialist of Armenian dialects, I do not know s
kian 1972; cf. Pisowicz 1976). Here we deal ra
tion of elements of different dialects; the poss
ments can be compatible in one system is rathe
firstly, the glottalized stops are present only in
spoken on the territory of Georgia and are obv
ing the result of the Georgian influence; seco
with voiced aspirates, where the plain voiced sto
dependent phonemes, they are, in fact, the pos
phones, of the plain voiceless stops and genetic
the voiced aspirates, but to the plain voiceless s
turn, trace back to IE plain voiceless stops;
Mush, tun "house" (from IE *dom-) and adeli
from IE *od-)' thirdly, in the other dialects w
the latter are opposed to the plain voiceless sto
positional variants; fourthly, in either of the di
pirates are not the positional variants of the
only in a few dialects the voiceless aspirates are
tional!) variants of the plain voiceless stops; the
the majority of cases are the reflexes of IE voi
ditional Old Armenian b, d, g), and not of the
It should also be mentioned that there are
either in New Armenian or in Old Armenian li
the former this is testified by the data of the f
tics of stops, as well as by experimental data (K
tian 1972). One can judge about Old Armenian
likishvili (1983: 222), whose works are hig
Gamkrelidze and Ivanov. She writes:

"The typological data at our disposal about the place of glottalized phonemes
in the hierarchy of stops can present a sufficient ground for the verification of
the hypothesis about the glottalic character of the series of plain voiceless stops
of Old Armenian. We have obtained the following frequency characteristics of
Old Armenian. Here are the absolute frequencies of phonemes:

This content downloaded from

180.176.178.34 on Sun, 23 Jan 2022 14:02:23 UTC
All use subject to https://about.jstor.org/terms
 A Variational Model of the Indoeuropean Consonant System 9
Plain Voiceless Plain
voiceless aspirated voiced
/p/ 149 /ph/ 35 /b/ 171
A/ 381 /th/231 /d/ 172
/k/323 /kh/302 /g/ 139

It is clearly seen that the series of phonemes to whi

tion has a functional characteristic judging by wh
marked. The glottalized stops cannot have such func
archies of three-stop systems of the mentioned type
Consequently, the functional data do not agree with
glottalization of this series. In terms of functional t
tiated phonetic interpretation of the Old Armenian s
ing: plain voiceless -voiceless aspirated - plain voic
Armenian phonological system did not belong to the
phonetic or functional viewpoints."

One may rightfully ask the question: if fo

of Indo-European stops the glottalists canno
cal parallels, what about the complete system
and labialized velars are considered too and where we have the
unique combination of palatalization and glottalization? The fact
that Gamkrelidze and Ivanov usually present the system of Indo-Eu-
ropean velars in a simplified way, without the palatalized and labial-
ized velars gives rise to one-sided interpretations of the Indo-Euro-
pean system of stops (Melikishvili 1983: 227).
6. The author of the present paper has not stated anywhere that
the traditional reconstruction of Indo-European system of stops is
impeccable. For instance, Jucquois's data about the frequency char-
acteristics of plain voiceless stops and voiced aspirates of Indo-Euro-
pean show a relationship reversely corresponding to the state of
daughter languages (cf. Jucquois 1966; Sigurd 1968; Melikishvili
1980). How can it be explained? To my mind the cause of this is that
the glottalists, following the tradition, attempt to bring the whole
diversity of stop systems of Indo-European languages to a single
monolithic system, without considering the principle of language
variation as a necessary quality of language. In Indo-European par-
ent language there did not, and could not exist a single pronuncia-

2) By the way, in the book "Ocerki pò istorii dopis'mennogo perioda arm-

janskogo jazyka" (Djahukian 1967: 75) I speak about a similar system of stops in
Old Armenian (if we do not consider the possibility that the voiced stops too
might be aspirated in some areas).

This content downloaded from

180.176.178.34 on Sun, 23 Jan 2022 14:02:23 UTC
All use subject to https://about.jstor.org/terms
 10 Gevorg B. Djahukian

tion norm; there were territorial distinctions in

as well.
The tendency to idealization and maximum ge
the reconstructed consonant systems groundles
air" (Tronskij 1967). It is known that in late
siderable dialectal differentiation existed. Every
takes the reconstruction of a certain state of lan
all make it clear whether the object of reconstr
lectal area, superdialectal (standard) language, o
and generalized state. Though in the second ver
struction Gamkrelidze and Ivanov give the vari
is the positional variants or allophones that the
the definite areas of distribution of variants.
If the dialectal differentiation of Indo-European should be taken
into account, one has to admit the incorrectness of the reconstruc-
tion of a unified consonant system for Indo-European. We must
speak of coexistence of several consonant systems or, more precisely,
of the dialectal variation of the generalized system to be recon-
structed. In other words, the reconstructed general model of IE con-
sonant system was in fact realized only by partial models (cf. Djahu-
kian 1982: 74). In this connection we shall point to the features
common for the IE dialects, as well as the features differentiating
them from one another (not the latest states of IE languages, but
their reconstructable oldest state (Urzustand). The common features
for the examined stops are those characterizing their series (i) by the
place of articulation: labiality, prelinguality, postlinguality; and (ii)
by the participation of vocal cords: voicedness (resp. voicelessness).3)
The distinguishing features are: (i) aspiratedness (aspiration) and (ii)
labialization or/and palatalization of the postlinguals; there must
have been transitional zones in which (i) the features of labialization
and palatalization were combined (cf. Albanian; partially, Arme-
nian), (ii) palatalization did not affect all the words (cf. Slavic, Baltic
languages).
Thus it is possible to distinguish two main coexistent dialectal var-
iants in the system of stops by the feature of aspiration, and three
variants by the features of palatalization and labialization of the vel-
ars.

3) We can speak of general characteristics only relatively; thus, for in

is known that the areas in Indo-European differed by the degree of
which was the main reason that the so-called consonant shifts (Laut
bung) took place.

This content downloaded from

180.176.178.34 on Sun, 23 Jan 2022 14:02:23 UTC
All use subject to https://about.jstor.org/terms
 A Variational Model of the Indoeuropean Consonant System 1 1

By these features it is possible to set up the following areas in late

Indo-European:
(1) the area with aspiration and palatalization: *k, *k', *kh, *k'h, *g, *g', *gh,
*g'h (Proto-Indian, Proto- Armenian);
(2) the area with aspiration and labialization: *k, *k«, *kh, *k«h, *g, *g», *gh,
*g«h (Proto-Greek, Proto-Italie);
(3) the area without aspiration, with palatalization: *k, *k', *g, *g' (Proto-Ira-
nian, Proto-Slavic, Proto- Baltic);
(4) the area without aspiration, with labialization: *k, *k«, *g, *g« (Proto-Cel-
tic);
(5) the area without aspiration, with palatalization and labialization: *k, *k',
*k«, *g, *g'> *g« (Proto- Albanian).4)

For all these special models the following common model can be
reconstructed:

(p) or (p) and (ph) (b) or (b) and (bh)

(t) or (t) and (th) (d) or (d) and (dh)
(k) or (k) and (kh) (g) or (g) and (gh)
(k') or (k') and (k'h) (g') or (g') and (g'h)
and/or or and/or or
(k») or (k») and (k»h) (g») or (g») and (g«h).

Using the signs of mathematical logic and denoting conjun

("and") by the sign A and disjunction ("or") by the sign V5), we
have:

(P) V (p) A (ph) (b) V (b) A (bh)

(t) V (t) A (th) (d) V (d) A (dh)
(k) V (k) A (kh) (g) V (g) A (gh)
(k') V (k') A (k'h) (g>) V (g') A (g'h)
A/V V A/V V

(kv) V (k») A (k»h) (g») V (g«) A (g»h).

The following conclusions can be drawn from the cons

some other data:
1. In the Indo-European consonant system, as in any other struc-

4) It should be mentioned that not all the Albanists support H. Pedersen's con-
tention that the difference in the development of labialized and non-labialized
velars before the front vowels reflects the state of parent language, i.e. three
series of back consonants simultaneously existed in Indo-European.
5) JK. Jakobson s contention that the presence ot voiced aspirates is determined
by voiceless aspirates (which was later shown not to have a universal nature) can
be presented as a logical implication ("if-then").

This content downloaded from

180.176.178.34 on Sun, 23 Jan 2022 14:02:23 UTC
All use subject to https://about.jstor.org/terms
12 Gevorg B. Djahukian

turai division of Indo-European, there was no c

pean norm: the Indo-European language structure
by areal variation. Despite the outstanding achieve
guistics which disposes of typological analogues in
guages varying structurally, socially and territor
ative linguistics still operates with monolithic com
pean linguistic models in the reconstruction of the
as other systems. Generally, it should be borne in
ity is an absolute characteristic in language, while
tive. The first is peculiar to reality, the second to
2. The reconstruction of the series of voiceles
replacing the traditional plain voiced stops for all
pre-divisional Indo-European, i.e. for the period
comparative data, does not seem to be justified
eover, it is theoretically faulty. There was no com
in that period. Naturally, one cannot rule out t
there existed an isolated area with glottalized co
3. Though the present paper does not deal with
should be mentioned that the examination of the available data and
the use of the principle of variability inherent in language render it
possible to state that in the late (pre-divisional) Indo-European there
was no common laryngeal norm either: the number, nature and dis-
tribution of the laryngeals varied territorially. In the atmosphere
where everybody was carried away by the laryngeal theory the re-
construction of different sets of laryngeals shows that (1) laryngeals
are reconstructed for all the lost consonants, and laryngeals are
found where other explanations are not valid; (2) the facts of some
languages are generalized for all the Indo-European areas and the
reconstruction of archetypes with an obligatory laryngeal element is
in vogue, which is obvious redundancy.
4. For all the common, as well as special and specific features of
daughter languages that are not accounted for by external factors
(influences), the explanation should be looked for in the state of par-
ent language. In other words, the embryos of intralinguistic changes,
apparent in the daughter languages not as a result of an exterior in-
fluence, took place in the period of community when all the struc-
tural levels and elements of the parent language varied. Thus, the in-
tralinguistic chronological changes can and should be projected onto
the parent language level in its territorial and social variability.
5. Many comparativists consider the typological orientation of
Modern Comparative Linguistics to be its main feature. Having

This content downloaded from

180.176.178.34 on Sun, 23 Jan 2022 14:02:23 UTC
All use subject to https://about.jstor.org/terms
 A Variational Model of the Indoeuropean Consonant System 1 3

acquired in typology an important criterion for the evaluation of the

reconstruction, i.e. the degree of correspondence of the model to the
object, however, it still remains a descriptive science. For it to make a
step to become an explicative science, it is necessary not only to find
real examples for the reconstructed model, but also substantiate why
a certain system is plausible and another is not.

Institute of Linguistics Gevorg B. Djahukian

Armenian Academy of Sciences
15 Abovian st., Yerevan
375001 USSR

References

Adjarian, H. (1971-1979): Hayeren armatakan bararan. 1-4. Erevan.

Allen, W.S. (1950): Notes on the phonetics of an Eastern Armenian speaker:
Transactions of the Philological Society.
Back, M. (1979): Die Rekonstruktion des idg. Verschlußlautsystems im Lichte
der einzelsprachlichen Veränderungen. KZ 93.
Blust, R. A. (1974): A double counter-universal in Kelabit. Papers in Linguistics 7.
Bomhard, A. R. (1975): An outline of the historical phonology of Indo-Euro
pean. Orbis XXIV, 2.
- (1979): The Indo-European phonological system: new thoughts about its re
construction and development. Orbis XXVIII. 1.
- (1988): Recent trends in the reconstruction of the Proto-Indo-European con-
sonant system. HS 101, 1.
Brugmann, K. (1904): Kurze vergleichende Grammatik der indogermanischen
Sprachen. Berlin.
Cantineau, J. (1952): Mufaxxama - the emphatic phonemes in Arabic.
Semiotika. 4.
Chatterji, S.K. (1964): Glottal spirants and the glottal stops in New Indo-Aryan
In honor of Daniel Jones. Londen.
Diakonoff, I. M. (1982): O prarodine nositelej indojevropeiskix dialektov. 1-2
Vestnik drevnej istorii 3-4.
Djahukian, G.B. (1967): Ocerki pò istorii dopis'mennogo perioda armjanskogo
jazyka. Erevan.
- (1975): Die Bedeutung der ersten (indogermanischen) und der zweiten (inter-
armenischen) Palatalisierung für die Konstituirung des armenischen Konso-
nantensystems. KZ 89, 1.
- (1978): Obsceje i armjanskoje jazykoznanije. Erevan.
- (1982): Sravnitel'naja grammatika armjanskogo jazyka. Erevan.
- (1986): O tak nazyvajemoj glottal'noj teoni" v indojevropeistike. Voprosy
drevnej istorii 3.
Lmonds, J. (1972): A reformation of Grimms law. Contributions to generativ
phonology. London.
Frisk, H. (1936): Suffixales -th- im Indogermanischen. Göteborgs Högskolas
Àrsskrift, = Kleine Schriften, 1966, 139-182.

This content downloaded from

180.176.178.34 on Sun, 23 Jan 2022 14:02:23 UTC
All use subject to https://about.jstor.org/terms
14 Gevorg B. Djahukian

Gamkrelidze, T. V. (1978): Language typology and lingui

ceedings of the Twelfth International Congress of Lingui
- (1980): Markirovannost' v fonologii i tipologii fonolo
ceskie osnovy klassificacii jazykov mira. Moscow.
- (1981): Language typology and language universais i
the reconstruction of the Indo-European stop system.
says in historical linguistics in honor of J.Alexander
T. R. and Bomhard A. R. Amsterdam.
Gamkrelidze, T. V., Ivanov, V. V. (1972): Lingvisticeskaja tipologija i rekonstruk-
cija sistemy indojevropejskikh smycnykh. Konferencija po sravniteVno-istorices-
koj grammatike indojevropejskikh jazykov. Moscow.
- (1972): Problema opredelenija pervonacaPnoj territorii obitanija i putej migra-
cii nositelej dialektov obsceindoevropejskogo jazyka. Konferencija po sravni-
tel'no-istoriceskoi grammatike indoevroùeiskix iazvkov. Moscow.
- (1973): Sprachtypologie und die Rekonstruktion der gemeinindogermanischen
Verschlüsse. Phonetica 27.
- (1980): Drevn'aja Peredn'aja Azija i indoevropejskaja problema. Voprosy drev-
nej istorii 3.
- (1980): Problema jazykov centum i satem i otrazenie "gutturaPnykh" v istori-
ceskikh indoevropejskikh dialektakh. Voprosy jazykoznanija 6.
- (1980): Rekonstrukcija sistemy smycnykh obsceindoevropejskogo jazyka.
Glottalizovannye smycnye v indoevropejskom. Voprosy jazykoznanija 4.
- (1980): R'ady "gutturaPnykh" v indoevropejskom. Problema jazykov centum i
satem. Voprosy jazykoznanija 4.
- (1984): Indoevropejskij jazyk i indoevropejcy: rekonstrukcija i istoriko-tipo-
logiceskij analiz prajazyka i proto-kuPtury. 1-2. Tbilisi.
Gimbutas, M. (1973): Old Europe c. 7000-3500 B.C. The earliest European civi-
lisation before the infiltration of the Indo-European peoples. The Journal of
Indo-European Studies. 1, 2.
Greenberg, J. H. (1966): Language universais with particular reference to feature
hierarchies. The Hague.
- (1970): Some generalizations concerning glottalic consonants especially implo-
sives. IJAL 36.
Haider, H. (1983): Der Fehlschluß der Typologie. Acten der 10. Osterreichischen
Linguistentagung (1982).
Haudncourt, A. (1975): Le mutation consonantiques (occlusives) en indoeuro-
péen. Mélanges linguistique offerts à E. Benveniste. Paris.
Hiersche, K. (1964): Untersuchungen zur Frage der Tenues aspiratae in Indo-
germanischen. Wiesbaden.
Hopper, P.J. (1972): Glottahzed and murmured occlusives in lndo-buropean.
Glossa. VII, 2.
- (1982): Areal typology and the Early Indo-European consonant system. The
Indo- Europeans in the Fourth and Third Millenia. Ed. by Polomé E.C. Ann Ar-
bor.
Illic-Svityc, V.M. (1971-1976): Opyt sravnenija nostraticeskix jazykov. 1-2.
Moscow.
Jakobson, R. (1958): Typological studies and their contribution to historical
comparative linguistics. Proceedings of the Eighth International Congress of Lin-
guists. Oslo.

This content downloaded from

180.176.178.34 on Sun, 23 Jan 2022 14:02:23 UTC
All use subject to https://about.jstor.org/terms
 A Variational Model of the Indoeuropean Consonant System 1 5

- (1962): Selected writings. I. Phonological studies. The Hague. Jakobson, R.,

Cherry, E.C., Halle, M. (1953): Toward the logical description of language in
the phonemic aspect. Language 22, 1.
Jakobson, R., Fant, G. M., Halle, M. (1955): Preliminaries to speech analysis.
Jucquois, G. (1966): La structure des racines en îndoeuropeen envisagée dun
point de vue statistique. Linguistic research in Belgium. Wetteren.
- (1977): The typology of Proto-Indo-European segmental inventory. The Jour-
nal of Indo-European studies. 5, 1 .
Khlebnikov, I.B., Kolod'ko V.l. (1982): K problème peredvizenija indojevro-
pejskogo b v germanskom. Voprosy jazykoznanija 6.
Kortlandt, F. (1983): Greek numerals and PIE glottalic consonants. Münchener
Studien zur Sprachwissenschaft 42.
- (1985): Indojevropejskije glottaPnije smycnyje. Voprosy jazykoznanija 4.
Kurylowicz, J. (1956): Etudes indoeuropéennes. Wroclaw.
- (1956): L apophonie en îndoeuropeen. Wroclaw.
Khacatnan, A.A., Ajrapetian, V. N. (1972). Ekspenmental noje íssledovanije sog-
lasnyx armjanskogo jazyka. Erevan.
Lehmann, W. P. (1952): Proto-Indo-European phonology. Austin.
Makajev, E.A. (1967): Znacenije armjanskogo jazyka dlja rekonstrukcii ob-
sceindojevropejskogo sostojanija. Istoriko-filologiceskij zurnal 4.
Martinet, A. (1953): Remarques sur le consonantisme sémitique. BSLP. 49.
- (1970): Economie des changements phonétiques. 3-me éd. Berne.
Martynov, V.V. (1968): Slavjanskaja i indoevropejskaja akkomodacija. Minsk.
Mayrhofer, M. (1962): Kurzgefaßtes etymologisches Wörterbuch des Altin-
dischen. 2. Heidelberg.
Meillet, A. (1937): Introduction à l'étude comparative des langues indoeuro-
péennes. Paris.
Melikishvili, I.G. (1974): K izuceniju ierarxiceskix otnosenij fonologiceskogo
urovnja. Voprosy jazykoznanija 3.
- (1983): O foneticeskoj xarakteristike jazykov kavkazskogo areala: drevnearm-
janskogo, obscekartvel'skogo i obsceindojevropejskogo. Geneticeskije,
areal'nyje i tipologiceskije sv'azi jazykov Azii. Moscow.
Normier, R. (1947): Indogermanischer Konsonantismus, germanische „Lautver-
schiebung" und Vernersches Gesetz. KZ 91, 2.
O'Konnor, J. D. (1952): RP and the reinforcing glottal stop. English studies. Am-
sterdam.
Pedersen, H. (1951): Die gemeinindoeuropäischen und vorindoeuropäischen
Verschlußlaute. Copenhagen.
Pisani, V. (1948): La palatalizazzione armena. Atti del Sodalizio glottologico ita-
liano. I, 1.
Pisowicz, A. (1987): Objections d'un armenologue contre la Theorie gioitale.
Folia Orientalia.
Pokorny, J. (1959): Indogermanisches etymologisches Wörterbuch 1. Bern-
München.
Rasmussen, J. (1974): Haeretica indogermanica. Kobenhavn.
Ringard, K. (1960): Vestjusk stod. Aarhus.
Shirokov, O.S. (1972): Kavkazsko-indoevropejskije ronologiceskie sxozdenija.
Konferencija po sravnitel'no-istoriceskoj grammatike indoevropejskix jazkov.
Moscow.

This content downloaded from

180.176.178.34 on Sun, 23 Jan 2022 14:02:23 UTC
All use subject to https://about.jstor.org/terms
 16 Gevorg B. Djahukian
Smith, S. (1944): St0det in dansk reigsprog. Kobenhavn.
Stevens, A.B. (1968): Madurese phonology and morpholog
Szemerényi, O. (1967): The new look of Indo-European re
pology. Phonetica 17.
- (1970): Einführung in die vergleichende Sprachwissensch
- (1985): Recent developments in Indo-European linguistic
Philological Society.
Trubetzkoy, N.S. (1958): Grundzüge der Phonologic Göt
Tronskij, I. M. (1967): Obsceindoevropejskoje jazykovoje
grad.
Winn, M. M.: Thoughts on the question of Indo-European movements into Ana-
tolia and Iran. The Journal of Indo-European Studies 11, 2.
Zinder, L.R. (1979): Obscaja fonetika. Moscow.

This content downloaded from

180.176.178.34 on Sun, 23 Jan 2022 14:02:23 UTC
All use subject to https://about.jstor.org/terms