TOSHIKI OSADA

N O T E S ON T H E P R O T O - K H E R W A R I A N VOWEL SYSTEM*

0. INTRODUCTION

This paper re-examines the vocalic system of the Kherwarian languages

(including Santali, Mundari, Ho, etc.) which are spoken in the Chotanagpur
plateau area in eastern India. My main concern has been the reconstruction
of the Proto-Kherwarian vowel system on the basis of the comparative
method since I read a paper on this subject (Osada 1984) at the 88th
meeting of the Linguistic Society of Japan at Keio University, June 1984. In
that paper I proposed a five-vowel system for Proto-Kherwarian (Osada
1984) instead of the seven-vowel system set up by several scholars (Pinnow
1959, Zide & Munda 1966, Munda 1968).
In 1989 Mr. Minegishi, Associate Professor in the Institute for the Study
of Languages and Cultures of Asia and Africa (ILCAA), Tokyo, Japan, did
field research on the Singhbhum dialect of Santali under the auspices of the
ILCAA -- CIIL (Central Institute of Indian Languages) joint research
project. According to him, the Singhbhum dialect has a six-vowel system:/i,
e, a, ~, o, u/(Minegishi 1990) while the dialect described by Bodding has
an eight-vowel system:/i, e, e, a, ~, a, o u/(Bodding 1929, Sebeok 1943).
When Mr. Minegishi showed me this data, it struck me that the vowel
correspondences between the two dialects of Santali should be examined,
since nobody has used this dialect for the reconstruction of Proto-
Kherwarian. I have been able to collect the relevant data from Mr. Ganesh
Murmu, who had worked as an informant for the Singhbhum dialect with
Mr. Minegishi.
In this paper I will present the new data on the vowel correspondences
between three Kherwarian languages, viz. two dialects of Santali, and
Mundari. As a result I propose a new hypothesis on Proto-Kherwarian
vocalism.

1. THE VOWEL SYSTEMS IN KHERWARIAN

The Kherwarian languages consist of Santali, Mundari, Ho, Asur, Bhumij,
Birhor, Korwa and Turi. 1 The Kherwarian languages are divided into two
groups, viz. Santali and Mundari-Ho. The latter group also includes Asur,

lndo-lranian Journal 39: 245-258.

~)1996 Kluwer Academic Publishers. Printed in the Netherlands.
246 TOSHIKI OSADA

Bhumij, Birhor, Korwa and Turi. 2 These languages are traditionally

considered to be minor dialects of Mundari. Santali, Mundari and Ho have
been well-studied mainly by the missionaries. 3 The data on the other
Kherwarian languages are still scanty.
As far as the vowel system is concerned, the languages belonging to the
Mundari-Ho branch of Kherwarian, including the so-called minor dialects,
have a five-vowel system (Bhattacharya 1975: 40--58), while Santali as
given by Bodding has an eight-vowel system.
The Santals are one of the largest tribes in India. According to the
Census of India 1981 the Santali speakers are more than four million. The
information on Santali dialects is still very poor. Fortunately we have been
able to collect the data on the Singhbhum dialect thanks to Mr. Minegishi.
The vocalic system in the Kherwarian langanges is as follows:
(1) The Santali described by Bodding: henceforward Sa(B) (Sebeok 1943).
/i, e, ~, a, ~, 3, o, u/.
(2) The Singhbhum dialect of Santali: Sa(S) (Minegishi 1990).
/i, e, a, o, o, u/.
(3) Mundari:Mu (Osada 1992).
/i, e, a, o, u/.
(4) Ho (Deeney 1975).
/i, e, a, o, u/.
(5) Korwa (Bahl 1962)
/i, i, e, a, fi, o, u, 11/. 4
(6) Birhor (N. K. Sinha 1964, Osada 1993)
/i, e, a, o, u/.
(7) Turi (Ponette: personal communication, see Osada 1991)
/i, e, a, o, u/.

For Birhor, Korwa and Turi we have information only on the vowel
phonemes. These languages, and Asur and Bhumij have not yet been
adequately described. Further, we do not use the Ho material, notwith-
standing the rich data available, because there is no difference in vowel
correspondences between Mundari and Ho.
Strictly speaking we cannot reconstruct the Proto-Kherwarian vowel
system until we have sufficient data on all the Kherwarian languages.
Although we do not have that as yet we can at least supply new data on the
Singhbhum dialect of Santali. This is important because our hypothesis is
supported by the data from this dialect.
We will, therefore, reconstruct the vowels of Proto-Kherwarian by
 NOTES ON THE PROTO-KHERWARIAN VOWEL SYSTEM 247

comparing the two dialects of Santali and Mundari. We will examine the
vowel correspondences between them in detail in the next chapter.

2. VOWEL CORRESPONDENCES

2.1. The lssue

First of all, we will illustrate the vowel correspondences between Santali

and Mundari, and the vowels reconstructed by Pinnow (1959), Zide &
Munda (1966) and Osada (1984) for Proto-Kherwarian (PK) in the follow-
ing table:

Santali i e e.e a a o.a o u

Mundari i i e a a o u u
Pinnow *i *e *e *a *~ *o *u
Zide & Munda *i *l *e *a *o *v *u
Osada *i *e *a *o *u
As is shown in the chart, the vowel correspondences between Santali e/
Mundari i, and Santali o / M u n d a r i u are the problem. What should be
reconstructed for Proto-Kherwarian? From these correspondences Pinnow
(1959) and Zide & Munda (1966) have reconstructed *e, *o and *l *u
respectively. I have, however, reconstructed *i and *u for Proto-Kherwarian.
This reconstructed system is almost identical to the Mundari vowel system.
I have also shown that the vowel developments from *i and *u in Proto-
Kherwarian to e and o in Santali can be explained by several plausible
phonological changes. In conclusion, I have suggested that Santali developed
the phonemic distinction between the closed vowel/e, o / a n d the open
vowel/e, a/probably under the influence of the neighbouring Indo-Aryan
languages, notably Bengali.
In the Singhbhum dialect of Santali the phonemic distinction between the
closed vowel and the open vowel is not found. If we want to adopt a seven-
vowel system for Proto-Kherwarian, we should explain why this phonemic
distinction was reduced in this dialect. On the other hand my hypothesis on
Proto-Kherwarian vocalism is still valid. The phonological change posited
by me is also applicable to this dialect with some modifications.
We present some examples below.
248 TOSHIKI OSADA

2.2. Vowel correspondences among the two dialects of Santali and Mundari

2.2.1. Mu. u~ Sa/B) o~ Sa(S) u PK *u

2.2.1.1. Monosyllabic Words

Mu. Sa(B) Sa(S)
bu ~ bhok ' bhu ~ 'to bark' P:V-248, M:p.75
ku ~ khok ' khu ~ 'to cough' P:V-236, M:p.75
du ~' dok' du ~ 'rescue' M:p.75
su ~ sok' su q 'to thrust hand into a hole' M:p.74
lu lo lu 'to ladle out' (Mu) P:V-400, M:p.74
(lu 9 by Pinnow) 'to draw, fetch' (Sa)
ud ot' ud 'mushroom' P:V-237, M:p.74
tud tot' tud 'to pick up'
In a previous paper we proposed that the following phonological change
can be observed from these correspondences; *Cu ~, *Cud in PK > Cok',
Cot' in Sa(B), Cu ~, Cud in Mundari and Sa(S). We have not found a
correspondence between Cuk' in Santali(B) and Cu ~ in Santali(S) and
Mundari. That is to say, there is no exception to the former phonological
change. One exception to the latter phonological change is ud 'to swallow'
in Sa(B), Sa(S) and Mundari. We have, however, considered it as an
expressive.5 As is well known, expressives are resistant to normal phono-
logical change.8
The following cases are problematic:
Mu. Sa(B) Sa(S)
dul dul dul 'to pour'
tul tul tul 'to rise (the sun)'
bul bul bul 'to be drunken' P:V-105
kul kol kul 'to send' M:p.75
ul ol ol 'to mould' M:p.75
rul ral rol 'to take off fire' M:p.105
ol al ol 'to write'
The form * Cul in PK has been basically kept in all the Kherwarian
languages. Further the vowel correspondences Mu o/Sa(B) a~ Sa(S) o are
very common. The other examples are not explainable at this point, but
they have the structure CV/. If we had the data on the other Kherwarian
languages we might be able to solve this problem.
 NOTES ON THE PROTO-KHERWARIAN VOWEL SYSTEM 249

2.2.1.2. Disyllabic words

(A) Mu. a-u~ Sa(B) a-o~ Sa(S) a-o

Mu. Sa(B) Sa(S)
katu kato kato 'knife' P:V-334, M:p.106
adu ado ado 'urine' M:p.106
capu capo capo 'touch' M:p.106
halu halo halo 'scoop' M:p. 106
hau hao hao 'red ant' M:p.106
bandu bando bando 'kind of creeper' M:p.116
gandu gando gando 'seat' M:p.106
haku/hai hako hdku 'fish' P:V-64
apu dpu dpu 'father' P:V-25, M:p.120
The progressive assimilation with the low vowel a has occurred in
CVCV. For example, PK *katu > Sa kato, Mu. katu. The last two cases of
the list can be interpreted in the following manner:
(i) haku/hai and hako are composed of ha- and the plural suffix ku/ko or
the singular suffix i. That is to say, this assimilation has taken place
within one disyllabic morpheme but not in a sequence of two mor-
phemes, e.g. PK *atu > Sa. atu, Mu. atu 'to flow' (a- = causative
suffix).
(ii) This kinship term is used not alone but with a personal suffix in both
languages, e.g., apu-~ 'my father', apu-m 'your father', apu-te 9 'his/her
father'. This assimilation has not occurred in CaCuC. For example,
*hati~ 'to share' in PK > hatit3 in Mundari, hati~ in Santali; PK *japid
'to shut eyes' > japid in Mundari, japit' in Santali.
This assimilation is not unknown in the South Asian languages. Bright
(1966) has reported that South Dravidian languages and Bengali have
developed this type of assimilation.
(B) Mu. u-a~ Sa(B) o-a~ Sa(S) u-a, o-a
Mu. Sa(B) Sa(S)
tudka totka tutka 'nape' P:V-276, M:p.105
surta sorta surta 'kind of flower' M:p.105
kuram koram koram 'chest' M:p.105
ludam lodam lodam 'kind o f f lower' M:p.105
putam potam potam 'dove' M:p.105
250 TOSHIKI OSADA

Here, regressive assimilation with the low vowel a has occurred in Santali.
This assimilation does not occur beyond a consonant cluster in the Singh-
bhum dialect but occurs in all cases in the other dialect described by
Bodding. Thus PK *CuCaC > CoCaC in Santali, CuCaC in Mundari; PK
*CuCCa > CoCCa in Sa(B), CuCCa in Mu and Sa(S).
The following instances from Pinnow (1959) do not show this phono-
logical change:
Mu. Sa(B) Sa(S)
kutam kutam kutam 'hammer' P:V-264
lumba ~ lumbak' 'to get dark' P:V-280
nuba ~ ~ubdk' 'dark' P:V-128
tuka tuka tukd 'nest' P:V-389
ula ula ula 'vomit' P:V-312
kundam kudam 'backyard' P:V-367
We shall examine each case below.
(i) As far as kutam, tuka and ula are concerned, we can find the corre-
sponding words in the neighbouring Indo-Aryan and Dravidian
languages, kutdsi 'hammer', kutek 'to pound' in Sadani (Indo-Aryan);
khota 'nest' in Sadani and Kurux (Dravidian), khota, thoka in Kharia
(Central Munda but adjoining to Mundari), which are likely a meta-
thesis; ulti 'vomit' in Sandani. 7
(ii) We can not find the word lumba ~ in the Encyclopaedia Mundarica. It
is probably a variant of nuba ~, which we will discuss below.
(iii) Initial *rio is not found in Santali. So *fiuba 9 did not become *fiobak'
in Santali.
(iv) As for kund. am it is not a loan word. According to Pinnow (1959) we
have the correspondent word in Kharia; i.e., kund. ab. From this form
we may reconstruct the protoform as *kund. am in Proto-Kherwarian.
Thus this phonological change may not occur in a word of CuCCaC
structure.
(C) Mu. u-u~ Sa(B) o-o/Sa(S) u-u
Mu. Sa(B) Sa(S)
cutu coto cutu 'mouse' P:V-132, M:p.75
huru horo huru 'paddy' P:V-244, M:p.76
gucu goco gucu 'beard, pubic hair' P:V-241, M:p.75
uku oko uku 'to hide' P:V-239, M:p.76
supu sopo supu 'the upper part of the arm' P:V-134, M:p.76
tuyu toyo tuyu 'a jackal' P:V-238, M:p.76
 NOTES ON THE PROTO-KHERWARIAN VOWEL SYSTEM 251

tupu topo 'to put under water'

tupu P:K-531, M:p.76
gu.ru god. o 'a rat'
gud. u
hupu hopo 'a kind of tree'
hupu M:p.76
uyu oyo 'a boil'
uyu M:p.75
utu oto 'a kind of vegetable'
utu M:p.76
lupu 9 lobok' 'bran, husk'
lubu ~ P:V-245, M:p.77
mulu ~ molok' 'a new moon'
mulu ~ P:K-212, M:p.76
'to become visible'
sutu ~ sotok' sutu 9 'following after' M:p.77
From these correspondences, two types of syllabic structure, C V C V and
C V C V 9, should be examined.
The following phonological change can be posited: PK CuCu > Sa(B)
CoCo, Sa(S), Mu CuCu. The exceptions to this phonological change are
given below:
Mu. Sa(B) Sa(S)
buru buru buru 'mountain' P:V-278
bulu bulu bulu 'tight' P:V-145
nunu fiufiu fiufiu 'breast' P:V-112
ruku ruku ruku 'to shake'
uru uru uru 'beatle'
We will examine each case in the following way:
(i) As for nunu/fiufiu, it is a reduplicated form of nu/fiu 'to drink'. This
reduplication is productive in Santali and Mundari. Further PK *flu
has never changed to *rio in Santali(B) diachronicaily. If *fiufiu in PK
would have become *~ofio in Santali(B),fiufiu could be derived from
flu synchronically.
(ii) We have found the expressive form ruku, rukun in Santali (Bodding:
1936). As we have seen above, expressive forms tend to resist
phonological change.
(iii) We have found ura 'beatle' in Kurux, a Dravidian language, which is
adjacent to Mundari and Santali. I think it is difficult to decide whether
uku is borrowed or not.
(iv) The word buru originally means 'a mountain, hill' in Santali and
Mundari. The other meaning of buru is 'God' as marao buru in
Santali. It seems to me that buru may be a taboo word which is,
therefore, an exception to the phonological change.
We can not find a plausible reason why *bulu in PK did not change to
*bolo in Santali(B).
252 TOSHIKI OSADA

Moreover, this phonological change can be supported by the following

explanation.
The open [o] and closed [o] were circumstantially allophones in Pre-
Santali; e.g.,/CoCo/was realized as [CoCo]. When this distinction was
made phonemic in Santali(B) under the influence of Bengali, the form
[CoCo] was phonemically interpreted a s / C o C o / i n a systematic way. As the
systematic gap, i.e. 'holes in the pattern' (Martinet: 1952, p. 19), 8 i n / C o C o /
should be filled, the form CuCu became CoCo. Since Santali(S), on the
other hand, has a five-vowel system, *CuCu in Pre-Santali has never
changed to CoCo in Santali(S).
As far as the form Cucu ~ is concerned, the following phonological
change occurred: *CuCu ~ > Coco ~in Sa(B), CuCu ~in Sa(S) and Mu.
The following example is seemingly exceptional to this phonological
change:
busu ~ busup' busu 9 'straw' P:V-120
We have, however, found bhusu in Sadani (Blain: 1975) and busfi in
Kurux (Bleses: 1956). This form may have been borrowed. 9

2.2.2. Mu. i/ Sa(B) e~ Sa(S) i

2.2.2.1. Monosyllabic
Mu. Sa(B) Sa(S)
bil bel bil 'to spread' P:V-221, M:p.72
ril rel 'winnow' M:p.72
til tel 'to distribute' M:p.72
This vowel correspondence is very rare in a monosyllabic word. All cases
have the structure CV/. As we have seen in 2.2.1.1., additional study is
required in these cases. I have suggested the following tentative phono-
logical change; PK Cil > Sa(B) Cel, Sa(S) Mu Cil.

2.2.2.2. Disyllabic
(A) Mu. a-i~ Sa(B) a-e~ Sa(S) a-e
Mu. Sa(B) Sa(S)
caki cake cake 'heap of something' M:p.103
capi cape cape 'wash, overflow' M:p.103
dari dare dare 'strength' M:p.103
 NOTES ON THE PROTO-KHERWARIAN VOWEL SYSTEM 253

gati gate gate 'companion' M:p. 103

sari sade sade 'sound' M:p. 103
mari mare mare 'old, ancient' M:p.103
tasi tase tase 'spread dry' M:p.103
]argi ]arge ]arge 'rainy season' M:p.104
The progressive assimilation with low vowel a has occurred in these
cases in Santali. For instance, PK *caki > Sa cake, Mu caki.
The exceptions to this phonological change are as follows:
(i) *ali~ in PK > ali~ in Santali, aliv in Mundari 'we two exclusive', alio
is a full form; liv is a short (bound) form. A sequence of two mor-
phemes may not be subject to this phonological change.
(ii) This phonological change didn't occur in *CaCiN (where N is a nasal
consonant); e.g. hativ in PK > hati~ in Santali, hatio in Mundari 'to
share'.
(iii) This phonological change did not occur in the kinship term given
below:
a]i a]i a]i 'sister' P:V-38
The kinship term is used not alone but with a personal pronominal
suffix in Kherwarian, e.g., a]i-h 'my sister', a]i-m 'your sister'. *(C)aCiN
in PK is not applied to this phonological change.
This type of assimilation is also common in South Asian languages, as we
have seen in 2.2.1.2. (See Bright 1966).
(B) Mu. i-a~ Sa(B) e-a~ Sa(S) e-a
Mu. Sa(B) Sa(S)
sirma serma serma 'sky, year' P:V-223, M:p.102
sida sedae sedae 'previous' P:V-224, M:p.102
kiwa kewa kewa 'chin'
Here regressive assimilation with the low vowel a has occurred in Santali;
e.g. *sirma in PK > serma in Santali, sirma in Mundari.
We should explain the following cases as exceptions:
Mu. Sa(B) Sa(S)
italia~ irM irM 'eight' P:V-92.
irilia
nida hinda nida 'night' P:V-43.
(i) According to Zide (1975), "(North Munda) '8' and '9' are more
254 TOSHIKI OSADA

mysterious innovations" (p. 2). Further, we have found iril in Ho,

which is very closely related to Mundari. Thus we can not reconstruct
*iral in Proto-Kherwarian.
0) We don't find the form *he in the initial position in Santali. Thus
*fii(n)da didn't become *fienda.

(C) Mu. i-i~ Sa(B) e-e/Sa(S) i-i or e-e

Mu. Sa(B) Sa(S)

bili bele bili 'ripe, egg' P:V-232, M:p.73
tili tele tili 'lice' M:p.73
tiki teke tiki 'to boil' M:p.73
birid beret' birid 'get up' P:V-231, M:p.73
diriv derefi derefi 'horn' P:V-347, M:p.73
fili~ jelefi jilifi 'long' P:V-340, M:p.73
sipio sepefi sepefi 'hold on palms' M:p.73
tisiO teheh teheh 'today' P:V-349, M:p.73
sibil sebel sibil 'taste' P:V-257, M:p.74
tiril terel tiril 'ebony' P:V-227, M:p.74
iril erwel iril 'younger sister/brother-in-law' M:p.73

These cases are the most problematic for our hypothesis. We may
suggest the following phonological change: PK * CiCi > Sa(B) CeCe, Sa(S),
Mu CiCi. We find, however, the following items which are not subject to
this phonological change.

Mu. Sa(B) Sa(S)

diri dh iri dh iri 'stone' P:84, M:p.55
idi idi idi 'take away' M:p.55
rici rici rici 'hawk' M:p.55

We cannot give a clear explanation for these exceptions at present.

Since some irregular correspondences are found in the limited circum-
stances we shall propose a five-vowel system for Proto-Kherwarian. If we
would adopt the seven-vowel system, we would have to explain why the
following difference occurred:

Mu. Sa(B) Sa(S)

dirio dereh dereh 'horn' P:V-347, M:p.73
filig jelefi jili~ 'long' P:V-340, M:p.73

Further, some cases can be explained by the following phonological

 NOTES ON THE PROTO-KHERWARIAN VOWEL SYSTEM 255

change: PK *CiCeC > Sa(B) CeCeC, Sa(S), Mu CiCiC. For example, *sileb
in PK > selep in Sa(B), silib in Mu, Sa(S). We have found the form sileb
in the Naguri dialect of Mundari in the Encyclopaedia Mundarica.

2.3. Mu. u-i/Sa(B) o-e~ Sa(S) u-i

Mu. i-u/Sa(B) e-o~ Sa(S) i-u
Mu. Sa(B) Sa(S)
bugi boge bugi 'good' P:V-229, M:p.74
filu jel fil 'meat' P:V-228, M:p.72
tiju tejo tiju 'insect' P:V-258, M:p.77
tiogu teogo ti~gu 'to stand' P:V-258, M:p.77
diku deko diku 'outsider' M:p.77
situo setoti situ~ 'heat' M:p.72
We have the following data described by E. G. Man (1867):
buge 'good', ti~go 'to stand'
If these forms are accurate, we can set up the phonological change: PK
*buge > Sa(B) boge, Sa(S), Mu bugi; PK *ti~go > Sa(B) teogo, Sa(S), Mu
tiogu. That is to say, progressive assimilation occurred in Santali (Singhbhum
Dialect) and Mundari while regressive assimilation occurred in Santali
(described by Bodding). This tentative phonological change can be sup-
ported by the case of diku/deko, which is divided into di-/de- and ku/ko,
plural suffixes. Since ko is a basic form *di-ko is a plausible form in Proto-
Kherwarian. Thus *di-ko in PK > deko in Sa(B), diku in Sa(S) and
Mundari.

3. P H O N O L O G I C A L CHANGE

As is seen above, e, o in Santali have, in all likelihood, developed from *i,

*u respectively in Proto-Kherwarian. That is to say, there was a split in
Santali in each case. We can explain this split through the following phono-
logical changes:
(1) Progressive assimilation with the low vowel a. See 2.2.1.2. (A) and
2.2.2.2. (A).
(2) Regressive assimilation with the low vowel a. See 2.2.1.2. (B) and
2.2.2.2. (B).
(3) Lowering u before glottal stop in Santali (B). See 2.2.1.1.
(4) Lowering u in CVCV in Santali (B). See 2.2.1.2. (C) and 2.2.2.2. (C).
256 TOSHIKI OSADA

(5) Regressive assimilation in Santali(B) and progressive assimilation in

Mundari and Santali(S), in PK *CVCV. See 2.3.
We have carefully examined the exceptions to these phonological
changes.
(i) Several exceptions happen to be loanwords.
(ii) Expressives are resistant to these phonological changes.
(iii) As for (1) and (2), these assimilations took place within one morpheme
with two syllables but not in a sequence of two monosyllabic mor-
phemes.
(iv) Further these assimilations did not occur in CVCVC structures.
*This paper originated from my M.A. thesis, which was submitted to
Hokkaido University in 1984. I would like to express my sincere thanks to
Professor Jiro Ikegami, who was my supervisor for my M.A. thesis.
I also owe many thanks to Professor F. B. J. Kuiper, one of the legendary
figures in Indology, who, in spite of a very busy schedule and old age, gave
me comments upon an earlier draft.
I would also like to acknowledge the encouragement and helpful advice
of Professor Norman Zide of the University of Chicago, and Professor
Rocky Miranda of the University of Minnesota.
My special thanks are due to my friends, Mr. Makoto Minegishi, Tokyo
University of Foreign Studies, and Mr. Ganesh Murmu, Ranchi University,
who gave me information about the Singhbhum dialect of Santali.

NOTES

1 According to Pinnow (1959) Dhangor, Karrnali and Mahle are included in the Kherwarian
languages. Dhangor, as far as I know, refers to the ethnically Oraon people, whose language
belongs to the Dravidian family. Karmali speak Sadani, an Indo-Aryan language. Mahle is
socially different from Santali. From a linguistic point of view, both languages are the same.
2 See Bodding (1929, 1932--36) on Santali, Hoffmann (1903, 1930--50) on Mundari and
Deeney (1975, 1978) on Ho.
3 See Zide 1966, p. 218.
4 According to Zide (personal communication), the long vowels/i, ~, f i / b y Bahl (1962) are
dubious.
5 Emenean (1980) has defined expressives as follows: "'expressives' is the most inclusive
term for a class with semantic symbolism and distinct morphosyntactic properties;
'ideophones' are a subclass in which the symbolism is phonological; 'onomatopoetics' are
ideophones in which the reference of the symbolism is acoustic (i.e. imitatives of sound)."
(p. 7)
6 See Annamalai (1968).
7 According to Professor Kuiper (personal communication) "There are some problems with
 NOTES ON THE PROTO-KHERWARIAN VOWEL SYSTEM 257

your argumentation via Sadani and Kurux. Sadani or Sadri, as you know, is the name for
every local IA dialect that has adopted many foreign elements. In fact, it is apparently Indo-
Aryan as spoken (originally) by non-Aryans. The occurrence in Sadani of specific words that
also occur in Munda cannot, therefore, be used as an argument, since they have most likely
been borrowed from Munda into local Indo-Aryan. The same is true of Kurux and Malto.
They are the two Dravidian dialects that are spoken in the neighbourhood of Kherwarian
and the many words which they have in common and which do not occur in the neighbouring
Munda languages are in all likelihood borrowings from Munda. The proof in such cases is
the fact that the words have no cognates in other Dravidian languages. In that case they do
not occur in the index of Kurux words in Burrow-Emeneau, Dravidian Etymological Diction-
ary, 2nd revised ed. (DEDR, 746--751). Thus Kurux khota 'nest' is not Dravidian, nor is
urd. You may be fully right in supposing that there is something wrong with the exceptions
but for the next time, at least, we will have to accept that this cannot be proved."
8 According to Martinet (1952), "This attraction exerted by a closely knit pattern on
marginal phonemes have been referred to as the filling of 'holes in the pattern' ". The term
'holes in the pattern' is found in K. L. Pike's Phonemics, 1947. p. 117.
Further Ferguson (1969) has reported "an example of such a filling of a gap in the
pattern by borrowing" (p. 114) in Syrian Arabic.
9 Professor Kuiper wrote to me on "busu ~ 'straw': "This has been adopted by Indo-Aryan at
a comparatively early date: classical Sanskrit busa-, Prakrit busa-, bhusa- 'chaff'. See Turner,
A Comparative Dictionary of the lndo-Aryan Languages No. 9293 (pt II, p. 526). The final
vowel of Sadani bhusti shows that this is not the Indo-Aryan word but a borrowing from a
Munda language. As often, the evidence seems to point to the conclusion that this is an early
borrowing from Munda, but we cannot rule out the theoretical possibility that also Munda
has borrowed from some unknown language that has now died out. Pinnow (1959: 93) is not
quite correct and mixes busa- up with a different word."

BIBLIOGRAPHY

Annamalai, E. (1968) "Onomatopoeic Resistance to Sound Change in Dravidian". Studies in

Indian Linguistics: M. B. Emeneau Sastipurti volume. 15--19. Poona and Annamalainagar.
Barker, P. R. (n.d.) The Phonemes of Korowa. Mimeographed. Seattle,
Bahl, K. C. (1962) Korwa Lexicon. Mimeographed. Chicago.
Bhattacharya, S. (1975a) Studies in Comparative Munda Linguistics. Simla: Indian Institute
of Advanced Study.
Bhattacharya, S. (1975b) "Munda Studies: A New Classification of Munda". Indo-lranian
Journal 17: 97--101.
Blain, E. (1975) English-Sadri Dictionary. Jharsuguda, Orissa, India: The Society of the
Divine Word.
Bleses C. (1956) An English-Uraon Dictionary. Ranchi: Dharmik Sahitya Samiti.
Bodding, P. O. (1929) Materials for a Santali Grammar. Dumka.
Bodding, P. O. (1929--36) A Santali Dictionary. Oslo.
Bright, William (1966) "Dravidian Metaphony". Language 42:311--22.
Deeny, J. (1975) Ho Grammar and Vocabulary. Chaibasa: Xavier Ho Publication.
Deeny, J. (1978) Ho-English Dictionary. Chaibasa: Xavier Ho Publication.
Ferguson, Charles A. (1969) " T h e / g / i n Syrian Arabic: Filling a Gap in a Phonological
Pattern". Word 25-1,2,3/114--119.
Hoffmann, J. (1903) Mundari Grammar. Calcutta.
Hoffmann, J. (1930--50) Encyclopaedia Mundarica. Vols. 1--15. Reprint Edition. Delhi:
Gian Publishing House. Reprinted in 1990.
258 T O S H I K I OSADA

Konow, S. (1906) "Munda and Dravidian Languages". G. Grierson (ed.) Linguistic Survey of
India. Vol. 4. Reprint Edition. Delhi: Motilal Banarsidass. 1967.
Mahapatra, B. P. (1980) "Munda Languages". Bulletin of the Department of Comparative
Philology and Linguistics. 4: 22--42. University of Calcutta.
Man, E. G. (1867) Santal and Santalia. Reprinted in 1983.
Martinet, A. (1952) "Function, Structure, and Sound Change". Word 8-1: 1--32.
Minegishi, M. (1990) "Santali English Japanese Wordlist -- A Preliminary Report --".
Journal of Asian and African Studies 39: 69--84.
M. or Munda, Ram Dayal (1968) Proto-Kherwarian Phonemic System. Unpublished M.A.
thesis. Chicago University.
Osada, Toshiki (1984) "On the Proto-Kherwarian Vowel System." (in Japanese) Paper
presented at the 88th Meeting of the Linguistic Society of Japan.
Osada, Toshiki (1991) "Father Ponette's Field Notes on Turi with a Comparative Vocabu-
lary". Journal of Asian and African Studies 42: 175--189.
Osada, Toshiki (1992) A Reference Grammar of Mundari. Tokyo: Institute for the Study of
Languages and Cultures of Asia and Africa.
Osada, Toshiki (1993) "Field Notes on Birhor". T. Nara (ed.) A Computer-Assisted Study of
South-Asian Languages 30--40. Tokyo: Institute for the Study of Languages and Cultures
of Asia and Africa.
P. or Pinnow, H.-J. (1959) Versuch einer historischen Lautlehre der Kharia-Sprache.
Wiesbaden: Otto Harrassowitz.
Sebeok, T. (1943) "Phonemic Systemn of Santali" JAOS 63: 66--67.
Sinha, N. K. (1964) "Some Linguistic Features of Birhor in Contrast to Mundari" abstract in
Stampe (1966) 720.
Stampe, D. (1965--66) "Recent Work in Munda Linguistics I--IV". IJAL 31: 332--41, 32:
74--88, 164--68, 390--97.
Zide, N. H. (1966) "The Proto-Korku-Kherwarian Vowel System". N. H. Zide (ed.) Studies
in Comparative Austroasiatic Linguistics 214--29. The Hague: Mouton.
Zide, N. H. (1968) "Munda and Non-Munda: Austroasiatic Languages". T, Sebeok (ed.)
Current Trends in Linguistics 5:411--30.
Zide, N. H. (1975) Studies in the Munda Numerals. Mysore: Central Institute of Indian
Languages.
Zide, N. H. and R. D. Munda (1966) "Notes on Proto-Kherwarian Vocalism". Abstract that
has appeared in Stampe (1966) 732.