Thursday, December 25, 2008

PEE-AEYAN AND PEE-DRAVIDIAN


PEE-AEYAN AND PEE-DRAVIDIAN
IN INDIA



PRINTED BY BHDPENDBALAL BANRBJEB
AT THE CALCUTTA UNIVERSITY PREBB, SENATE HOUftR, CALCUTTA.



K. No. 854B. June, 1929 K



PREFACE

As the following articles have opened up a new field
of enquiry, I have found it necessary to make them
accessible to our scholars and students. I am thankful to
the authors, Professors Sylvain Levi, Jules Bloch and Jean
Przyluski, for kindly permitting me to publish these
translations in the form of a book. The introduction
is meant for our students, and it does not pretend to be
exhaustive. A number of similarities between In do- Aryan
and Austric words have been suggested by my friend and
colleague Dr. S. K. Chatterji, and such suggestions
have also occurred to me. As some of the words noted
by us are popular vernacular terms, not usually registered
in dictionaries, it was thought that their discussion might
be h propos to the subject so brilliantly inaugurated by
the eminent French scholars.

In spite of my best endeavours, some possible slips
in translation and transliteration, have crept in : for
these I crave the indulgence of the authors as well as
the reader. But these, I hope, will not detract from the
merit of the original papers, which are singularly valuable
for the reconstruction of the foundations of our history
and culture.

I have gratefully to acknowledge the assistance E
received from Dr. A. C. Woolner for having gone through
portions of the translation in MS., and also from
Dr. S. K. Chatterji for his constant active interest in
the publication of the work.

P. C. BAGCHI.

THB UNIVERSITY,

CALCUTTA :
1st May, 1929.



TABLE OF CONTENTS

PAGE

INTRODUCTION ... ... ... i

Some more Austric Words in Indo-Aryan
(S. K. Chatterji and P, C. Bagchi) ... xix

PART I

Non-Aryan Loans in Indo-Aryan ( J. Przyluski) S
kadall, p. 4; data, kambala, 8imbala y p. 6;
lahgaldy languid, linga, p. 8 ; Names of betel,
p. 15 ; Bengali numeration and Non-Aryan
Loans, p. 25.

PART II

Sanskrit and Dravidian (Jules Blooh) ... 35
PART III

Pre- Aryan and Pre-Dravidian in India (fylvain
Lvi) ... ... ... 63

APPENDIX TO PART I

Further Notes on Non-Aryan Loans in Indo-
Aryan (J. Przyluski) ... 127

matanga, p. 129 ; maytora, mayuka, maruka,
p. 131 ; Names of Indian Toiras in the
Geography of Ptolemy, p. 188 ; kodumbara-
odumbara, p. 149.

APPENDIX TO PART III

I. Faloura-Dantapura (Sylvain Lvi) ... 168
II. Note on Tosala and Dhauli (P. C. Bagchi) 176

ADDITIONS AND CORRBOTIONS ... ... 179



INTRODUCTION. 1

AUSTRO-ASIATIC.

A linguistic unity in the Mon-Khmer group (then
called Mon-Annam) was first pointed out by Logan and
others and established on a sure footing by Keane in 1880,
Forbes in his work, Comparative Grammar of the Languages
of Further India (1881)9 incontestably proved the exis-
tence of this linguistic unity. In 1888 Miiller * continued
the same study further. Euhn in 1889 while emphasising
on this linguistic unity remarked 3 " what is more
striking is the relation with Annamite ; there is undeniable
relation of the monosyllabic group : Khasi-Mon-Khmer
with Kolh, Nancowry, and the dialects of the aborigines
of Malacca,"

Schmidt pushed the work of Kuhn further 4 and estab-
lished the relation between the languages of the Malay
peninsula and the Mon-Khmer group. He studied also the
correspondence in the vocabularies 5 and the phonetic
laws of those languages. He then applied these laws to

1 The first part of the Introduction is based on (and partly
translated from) the French translation of Pater Schmidt's article.
Of. BEFEO VII, pp. 217ff.

1 Qrundriss der Sprachwissenschajt, Vol. IV (Appendix).

8 Beitr&ge zur Sprachen-kunde Hinterindiens, Sitz. der K. bayer
Akad-der wissensch. phil. hist. KL 1889, I, p. 219.

* Die Sprachen der Sakei und Semang auf Malakka und ihr Verh&l-
tnis zu den Mon-khmer Sprachen, Bijdragen tot de Taal Landen
Volkenkunde van Nederl. Indie, 6th series, Fart VIH.

B Qrundz&ge einer Lautiehre der Mon-khmer Sprachen (1905),
Denksohriften der Kaiserl. Akad. d. Wiss. in Wien (phil. hist,), EL f
Vol



ii INTRODUCTION

the study of Khasi. 1 In an appendix to his study of
Khasi he examined Palong, Wa and Riang of the middle
valley of the Sal ween. Palong was already connected with
the Mon-Khmer family by Logan and Euhn. Grierson
in his Linguistic Survey (II, pp. 1, 88ff.) added Wa
and Riang to it. Wa and Riang extends almost to the
same latitude as Khasi.

Schmidt next studied ' the Nikobarese and by a study
of its phonology proved that it belongs to the Mon-Khmer
family and is related to other languages which belong to
the same group. There is resemblance even in particular
details of vocalism and consonantism. It has the same
development for the roots in ya and ioa as in the Mon-
Khmer languages, the same for the mode of the production
of palatals. As to the morphology, it presents some
earlier phases of morphological development in many cases
and gives us the key for explaining a series of forms in
Mon-Khmer. 8 Nicobarese is not pollysyllabic as often
said; the roots are monosyllabic like other Mon-Khmer
languages and are developed by infixes and prefixes.
There are besides suffixes in it which are completely
missing in other Mon-Khmer languages. Most of these
suffixes indicate direction (as it happens in the languages
of islands) meaning cardinal points. But there are a few
which have purely grammatical function. Hence Niko-
barese is a link between the MundS (or Kol) and the great



einer Lautlehre der Khati-Sprache in ttiren Be*i-
ehungin u derjtnigen far Mon-kJimer Sprachcn (1905) : Abhandlnogen
der kflnigl. Bayer Akad. d. Wies. (I. EL, Vol. XXII, III).

* C/. Gr. Mon-khmer Sprachen, 199 ff. and 225 ff., Qr.Khati.
Sproohe, J 151ff.

> Ct. Appendix to Die Monomer V Biker, tin Bindeglied Zwischen
V&lke Zentralatient und Auetronetitnt, 1906 (cf. French TrsnilatioD,
BBFBO, yn, pp. 251ff.).



INTRODUCTION iii

Austronesian group which also possess suffixes besides
prefixes and infixes.

Sten Konow (Linguistic Survey, IV, p. 11) established
the relation between the MundS and the Mon-Khm$r
groups on phonological grounds. " Both families possess/ 9
he says, " aspirated hard and soft letters. Both avoid
beginning a word with more than one consonant The most
characteristic feature in Munda phonology are the so-called
semi-consonants : &', *bok-karek, or from
bid bSh- karek. [B 21, 22, 35.]

Bat : Bengali badnd (Tf?^) = *bad + affix-w$a -cla.
Cf. hftpSt, sapet (Hes. Sep. A. I.), hompet (Bes. Songs.),
samet, hamet (Bahnar), kawet, kowet (Sem. Stev.), kawed,
kauid (Sem. Pa., Max., Sem. Bukb., Max.), kawat, ganat
(Sem. K. Ken), k&t (Kaseng), kawa <*kawat (Mon); not
(Stieng) ; w&t-da, wat-da, wat (Andamanese). [B 74,
76].

Bird: Hindi cl(jiy5, (f^f^rr)=c?-^-t5. Cf. cam, cem
(in some Sakai and Semang dialects) ; ka-cim (Mon) ; bim
(Cham); kcim (CharS), sem (Bahnar) ; sim (Palaung) ;
si m= cock (Santali). Also camfa (Santali, Mahle,
Munffai, etc,). [B 216]. ;.



aii INTRODUCTION

Breast (Female) : Skt. cucuka : cf. Malay susu, milk :
may be onomatopoeic. [ B 387].

Charcoal : Skt. augar, Hindi ingel : ef. anggu (Sem.
Jur. New); jeng-kfi,, jengkat (Sak.), nying-kah (Sen.) ;
embers : engong oft, ingung us, etc. (Sem.) ; Firewood :
api (Jak.) ; Fire logs : anggeng (Bes.) ; burning embers :
rftngok (Khmer). [C 77.]

Cheek : Skt. kapola : cf. kSbang (Sem. Fa. etc.) ; kapft
(Sak.) ; Face : kapo, kapau (Sak.) ; tapoa (Nicobar,
Central and South) ; Cheek : thpeal < thbal (Khmer).
The Skt. word may be Austric in origin ka-pola, pola
representing the original root. [C 81] Compare kapala head,
Bengali ^*tt*I &*j05a= forehead, which has been suggested
as " Oceanian " by P. Rivet, 'Le group Oceanian/ MSL.,
1927, p. 149.

Coconut : Skt. narikela : cf. Malay niyor (coconut),
niyor (Sak. and Sem.) ; fruit : pie, phlei, etc., kolai (Tareng) ;
(cf. infra, p. 54), kolai (Kontu) ; nankela may be derived
from equivalents of niyor (coconut) and kolai (fruit),
combined. [C 197 ; F 282.]

Cloth : Bengali kani ( ^tfa ) meaning a rag. Cf. Malay
kain.

Crab : Skt. kamatha, karkata, Bengali kMa ( *tfcl )>
ke(e (C^I?)
as it occurs frequently in the Austro-Asiatic languages.

For the origin of these forms we can, therefore,
suppose a root *bala-, *bulu meaning "hair, wool."
We have in Sanskrit bala; vala-, vdra which have the
same sense. The word is ancient, and vara is already
. found in Rg Veda, 2,4,4. The presence of I in bala, vala
points out to a popular form and the analogy with the
non- Aryan words, which we have just now referred to,
suggest an Austro-Asiatic origin. The loan can be
explained by the importance of wool and hairs in
magic and popular religion. It is not at all doubtful

1 Of. Bengali pSZofc, the feather or down of a bird. The word bai
been pointed out to me by Dr. P. C. Bagchi.



NON-ARYAN LOANS IN 1NDO-ARYAN 7

that in this respect the Indian thought has been always
influenced by the beliefs of the aboriginal populations.
One can, for instance, refer to the cult of the hairs of
Buddha. In the legend of Rama, Ball, the famous
monkey-king, the brother of SugrTva, owed his name to the
fact that he was born from the hairs of his mother.

However, the Austro-Asiatic origin of bala- would
remain doubtful if this Sanskrit word had not formed the
part of a group of which the other elements are certainly
non-Aryan. We will see just now, that kambala, sambala
are inseparable from bala and foreign to Indo- Aryan. 1

The Austro-Asiatic root *bala was susceptible to have
prefixes like /kz-h nasal, in order to form a substantive
meaning " a shaggy being." We should not therefore,
be astonished to find in Sanskrit a word kambala, which
means a kind of deer and more precisely, " a sort of deer
with a shaggy hairy coat," according to the definition
of Monier Williams. As one finds, the description of
the animal conforms to the etymology of the name.

From this the use of the word kambala for woollen stuff
can be easily explained, Kambala in the sense of " woollen
stuff " occurs in the Atharva-Veda (XIV. 2, 66, 67).
This word, is doubtless of non-Aryan origin, and has
been introduced in the Sanskrit vocabulary prior to the
redaction of the Atharva-Veda.

On the other hand the name of " silk-cotton tree "
or Bombax Hepiaphyllum is in Pali Simbalt or Simbala and
in Sanskrit Salmali or lialmala. One can recognise here

1 In Tibetan, bal means the hair of certain animals, the wool.
A composite adjective of this word has been formed 'bal-'bal which
Saratohandra Das translates by " shaggy." It is difficult to decide
if these forms are in their origin Tibeto-Bnrmese, or if bal has been
borrowed from Indo-Aryan.



8 PRE-ARYAN AND PRE-DRAVID1AN

the root *bala which forms a part of the name of cotton
or cotton-tree in some Austro-Asiatic languages.

In Pali the root has the prefix **, as i Q wwbala, simbalt.
Simbala already exists in Vedic and means, according to
Sayana "the flower of cotton tree" (cf. GBLDNBR, Teditche
Stndien, 2, 159), that is to say, its bud having the appea-
rance of a big white flower while it is still in its covering.

Skr. salmala, aalmall correspond to Pali simbala,
simbali, and equally means the Bombax Heptaphyllum.
These words, however, cannot be the same. Salmala
seems to be the Sanskritisation of another form. In the
Austro-Asiatic languages, between a root *bala and a
prefix sa-j si-, one might have intercalated a nasal and a
liquid. Simbala contains the nasal m. Salmala which has
got the liquid / seems to be the Sanskritisation of *salbala<

The Sanskrit words bala, kambala, salmala form a
series in which the idea of hair or wool can be discovered
all through. They are differentiated only by the prefixes,
i.e., by a process which is foreign to the morphology of
Indo- Aryan. It is therefore the Austro-Asiatic root *bala,
which we should suppose to be the origin of all these words.

II 1

SANSKRIT langala, languid, linga.

The plough is designated by the following terms in
the principal Mon-Khmer and Indonesian languages :

Khmer an/cal

V

Cam lanan, lanal lanar

Khasi Tta-lynkor

Tembi fengala.

\

1 Cf. Memoires de la Socie'te' de Linguistique, XXII, p. 205 ft.



NON-ARYAN LOANS IN INDO-ARYAK 9



Batak
Makassar



tmgala, langala
lingala
uankala.



How to explain these different forms ? One can
suppose either that they have been borrowed from Indo-
Aryan (c/. Sauskrit laiigalam), or that they are all derived
from an ancient Austro- Asiatic word of which the begin*
ning and the end might have undergone several modi-
fications while the middle part remained more stable.

The first explanation is subject to serious difficulties.
The word langalam has no etymology in Indo-Aryan and
is certainly not Indo-European. Besides, the counterpart
of the words quoted above is found in Annamite, i.e.,
amongst a people which has never been Indianised
like their western neighbours.

In Annamite the word edy (pron. kdi) is both a verb
meaning " to plough " and a noun signifying the " plough."
It is possible that in ancient time this word was longer,
AS we know that in Annamite the tendency towards
monosyllabism has been strongly active from early times.
Previous to the modern form kdi, an ancient one *kdl, can
be supposed. In fact, the final I, replaced by i in Anna-
mite, is preserved even to-day in several Muong dialects :





Annamite


Muong


" tree "


kai

*\


kol


u to be hungry "


dot


tol \


"two"


hai


V '

h*l 1


"to fly (of bird)"


bai

*


paltptt



10 PRE-ARYAN AND PRE-DRAVIDlAN

Annamite *kal " tbe plough," and " to plough " reduced
to one syllable, is very similar to the Austro-Asiatic forms,
with this difference that they can be separated, and as
the Indian influence here is out of question we are
brought to suppose that the Mon-Khmer and Indonesian
names of the plough have not got an Indo-Aryan origin.
Laiigalam is found already in the Rg Veda, but the two
/-s in the word indicate a vernacular form of it.

The only alternative left to us is to admit that
l^hgahm has been borrowed from the non-Aryan peoples
of the East since the Vedic times. The same conclusion
is inevitable, if one handles a problem of a different kind.

Besides " the plough/' the Sanskrit word laiigalam
designates also the "penis." On the other band,
specially in the Sotras and in tlie Mahabharata, a form
laiigula is found to mean both the " penis" and "the
tail " (of an animal). If the equivalence laiigala-langula
is authorised, then the semantic evolution of the word
would be easily understood. From " penis " one can pass,
without difficulty, to the sense of " plough " and " tail/'
There are evident analogies between copulation and the act
of ploughing by which one digs up the earth for depositing
the seeds. The problem becomes more complicated from
tbe fact that, almost inevitably, the word linga which
strongly resembles the two other words and has the
meaning of " penis " comes in.

Such equivalence is phonetically impossible as long
as we are in the Indo-Aryan domain, but they are
fully justified in the neighbouring groups. In *0am,
for instance, the scolopendra is called In pan or Upon.
In the same lan^nag^, kalik and Tculik, kayan and kuybu,
tabal and kubnl are equivalent forms [E. Aymonier and
A. Cabaton, Dictionnaire Cam-frangait]. In the Malaya



NON-ARYAN LOANS IN INDO-ARYAN 11

Peninsula, the tree "palai" according to Skeat and
Blagden is denoted by the following words :

tingku

lengkal

tengkol

V V

tangkal
v
tengkvl.

w / ^ w

Tangkal is to tenfold and tingku without final is to
tengkul what langala is to lahgTda and libgv to laiigala.

One is thus led to suppose that these multiple and
suspicious forms, lihga, taiigala, laiiyaln, langu'a, langjila,
represent diverse aspects of the same word, borrowed
by Indo-Aryan from the Austro-Asiatic languages.
This hypothesis would be still strengthened if it can be
shown that lihga in the sense of " penis " has equivalents
in the non-Aryau languages of the East.

Here are the principal names of the sexual organs in
the Austro-Asiatic languages :

Malay Peninsula lalc % la, lo.

Stieng klan

Bahnar k-lao

Khasi t-loh

Santali loc

Ho lojf

Mundari loc\ l

All these forms appear to be derived from lak still
found in the Malaya Peninsula. The final k is sometimes
palatalised into h and sometimes disappears completely
with the result that the vowel is changed into a diphthong.

1 Bev. P. 0. Bedding writes to me : The word 7oo' IB by the
Santals considered indecent and is cot used before women. Then i&
anotber word cf the same root lie 1 , used about the organ oi touull
boys, but also oooiiderad



1$ PRE-ARYAN AND PRE-BRAVIDIAN

Here also the hypothesis of a loan from Indo-Aryau
is excluded on account of two reasons. The vowel
i of linga is never found isolated in any of the Austro-
Asratic words derived from a form in a. Besides the
name of " penis " can be found in Annamite with the

V V

word kale (Quoc-ngu : cac) which doubtless comes from
an ancient *k-lak. We know that the initial consonantal
groups were all reduced in Annamite, some before the
XVlIth century and the others much later.

On the whole everything tends to show that an
ancient A astro- Asiatic root *lak has given rise to the
nominal derivatives ending in -ala-, -via-. The existence
of a final with vocalism is not solely attested in Indo-
Aryan and only by the word lahgida. Skr. lagitda, lakuta,
appear to be copied from laftgula, and its meaning of
"stick M can be very well derived from " penia." Parallel
to Skr. tingtia "tail" (of an animal) we find Malay
ctor, and in the Malay Peninsula, ikul, tkur, ekor, kur,
with the same meaning.

A certain number of forms which we have examined
contain a nasal element which seems to have been inserted
in the root. Now we know that in the majority of
the Austro-Asiatic languages, the infix n seems to form
the names of instrument (cf. Father. W. Schmidt, Lcs
peuples M on- Khmer, French translation in B.E.F.E.O.,
1307, p. 237ff.). I will quote only one example, similar
to the case studied by me : Khmer canMut " helm "
derived, by adding an infix from, *Mdut " to obstruct,
to move against the helm " [E. Aymonier, Dictionnairc
RJimfr*fr p. xvi]. Hence it is to be noted that
amongst the non-Aryan words quoted above, the nasal
infix is wanting in those which designate a part of the
body : " penis " " tail " (of an animal), while it is found



NON-ARYAN LOANS IN INDO-ARYAN 18

in tbe names of instrument, for instance in the names
of "plough." On the other hand, as to be expected
in the case of loan words, Indo-Aryan has no regularity
at all in this respect. The contrast laguda-langula has
no morphological value at all.

Nasal infix and suffix in -ul(a) seem to co-exist in
Khmer. In this language, boh signifies "to drive in
(a post) " and bdnkvl means the "post." If from Khasi
t-loh " penis " one goes back to the root *lak from which
lynkor " plough " is derived, one can also go back from
boh " to drive in " to a root *bak which explains bdnkti
" post." The first root *lak is not however entirely hypo-
thetical. One can recognise an alternation of it in Khmer
luk " to drive in (the hand or the finger) " (Dictionnairt
Tandart). Besides, the Santals have a common word la
" to dig or make a hole." The derivatives like lai^galam,
etc., express the penetration of plough into female earth.
The names of " penis " and of " plough " therefore, signify
respectively in the languages in question : "the limb which
one drives in " and " the instrument which one drives in."

The insertion of an infix in the body of the root
has tbe effect of lengthening the word, of making it
stand wear and tear. The length of the non-Aryan
names of plough can be thus explained by reference to
other words of the same group derived from the same
root. Compare for instance :

Malay : Ifagala "plough," ekur " tail."
Khasi : kaJynkor " plough," ' l-loh "penis."

1 Judo-Aryan has borrowed even the prefix fcA of Kbasi ka-lyftkor ?
ID the Mahfihbirata in. 642, kaldngala appears to detignate a kind of
weapon. This way of uiing the sharp end of plough is not tbe
only instance in tbe epic. Balarima is armed with JdnyoJam and fox
this reason, named



U PBE-ABYAN AND PBE-DRAVIDIAN

It may seem strange that the In do- Aryans have
borrowed so many words from the A astro-Asiatic
languages. Various circumstances have contributed to
this result. Some Austro- Asiatic peoples use even to-day,
not a plough to furrow but a simple pointed stick
for digging holes in which they place the seeds (Skeat
and Blagden, Pagan Races of the Malay Peninsula, I, p.
348). There the analogy between the penis and the
farming instrument is as clear as possible. Profs. Hubert
and Mauss point out to me that in Melanesia and
Polynesia the farming stick has often the form of a penis.
In some Polynesian languages the same word designates
the penis and the " digging stick " (ef. Tregear, Maori
Comparative Dictionary, under ko and Violette, Dictionnaire
Samoan-frangais, under off a). It is possible that the
aborigines of India, at first, knew the use of this stick and
that the name of the instrument for digging the soil
has not changed after the introduction of plough.

The persistence of old notions helps us in explain-
ing the legend of the birth of Slta. In the Bamayana I,
66, it is by furrowing the earth with a plough that Janaka
gave birth to Slta. The names are transparent here :
Janaka signifies "procreator" and Sltft means a
"furrow," The farrow has been personified since the
Vedic times. In the Mahsbharata, VII, 105, 3,945,
Sltft is a goddess of the harvest, The legend of the
birth of Sltft conceals the ancient myth about the produc-
tion of grain. The same forces are manifested there and
the sole action which gives play to them is the penetration
of the plough-penis in the female earth.

On the other hand the phallic cults, of which we know
the importance -in the ancient religious of Indo-China,
are generally considered to have been derived from Indian



NON.ARYAN LOANS IN INDO-ARYAN 15

Saivism. It is more probable that the Aryans have
borrowed from the aborigines of India the cult of lifoga
as well as the name of the idol. 1 These popular practices,
despised by the Brahmans were ill-known in old times. If
we try to know them better, we will probably be able to see
clearly why so many non- Aryan words of the family of llnga
have been introduced into the language of the conquerors.

Ill
THE NAMES o? BETEL,*

We know that the betel-leaf, with some other products,
is used for the composition of a masticatory much appre-
ciated by the Indian and Indo-Chinese peoples. The
following words designate the betel in the Austro-Asiatic
languages :

Alak balu

Khmer mluv

Bahnar bolou

Rongao bolou

Sue' malua

Lave melu

Stieng mlu

Kha bin

Falaung pin.

1 While writing this article for a lin guistic review I have been led
to develop the idea still further. It is clear from what has been
written that the history of a word like Hnga is not without impoitance
for the study of religion. I have begun in 1923 a series of study meant
to prove that a certain number of myths, legends and tales of
Aryan India have been borrowed from the Austro-Asiatic people. The
first two of these articles will shortly appear in the publication of
the Ecole^ FrSrufaite d' Extreme-orient (the Jubilee volume) and in
the Journal Asiatique.

Cf. Bulletin de la 3ociftf de Linguutiquc, XXIV, 8, pp. 255.2(8.



1$ PEE-ABYAN AND PRE-DRAVIDIAN

All these forms can be reduced to one type *mat
often having for the initial the alternation m/d. The long
final id sometimes redoubled into uv t tin, ua. The .vowel
is palatalised into
pp. 30-31, Father W. Schmidt has compared the
following words :



Moo


Khmer


Bahnar


"to throw the ( poh
stones with ]
a bow" (pofc

" this bow " pnoh


44 to throw, shoot to
bask (the cotton)"
boj*

" Card for cotton "
phnol)


"to drawfptowfc
thebow"(pano*



A verbj00A, pah, with infixes gives rise to the following
derivatives : panah, ponah> phnoh pnoh. The derivation
is regular, bat it is not a priori clear why the same
root is used to mean such operations as drawing the
bow and husking the cotton. This remarkable fact
becomes clear if one observes : 1 that in Stieng ak
designates an instrument used for preparing the cotton

1 Cf. Bulletin dc la SocUU de Lingubtique, XXV, 1, pp. 66-69.

The initial which I transcribe here as t, and which Father Schmidt
has written p is an ancient labial ocolnsive intermediate between
p and b which M. ft. Maspero calls " mixed occlusive." (Qrammair*
d$ la langue Khmbe, p. 65).



20 PRE-ABYAN AND PBE-DBAVIDIAN

before the spinning ;, and 2 the same word is a name
of bow or cross-bow in other Mon-Khmer languages
(Danaw ak ; Riang at " bow " ; Alak ak " cross-bow "). 1

On the other hand amongst the Makassar of Celebes,
the word pdna designates the bow for shooting the arrows
and a kind of bow which is also used for washing the
cotton (NiEUWENHuis, Der Oebrauch von Pfeil und Bogen
auf den grown Sunda-Iiiseln, in Internationales Archiv
fur Ethnographic, XIX, pp. 10-11). Sonnerat has observed
and sketched a similar instrument in India ( Voyages aux
Indes et h la Chine, Paris 178*, vol. I. p. 108 and pi. 26).
"The machine for carding the cotton/ 1 says he, "is
extremely simple. It is made of a piece of long wood of
six to seven feet. To each of its ends is attached a strong
string of entrails which, when touched, makes sound
like that of the violin (our hatters also have a machine
almost similar to it called the archet or fiddle-stick)*
The violin is suspended by a string to that of a
bow attached to a plank. The worker holds the violin
by the middle in one hand and in the other, with
a piece of wood with a pad at the end, stretches quickly
the catgut which slips out, strikes the cotton, throws it
out, fills it with wind, separates the dust from it acd
makes it fit for spinning. The elasticity of the bow,
which sustains the violin, affords the worker the facility
of -carrying it from one place to another on the heap
of cotton which they come to thrash." The instrument,
on tbe whole, is formed of two bows superposed, because
the tower part of the violin which Sonnerat compares

* Tbe languages of the Malaya Peninsula have the forms fg, eg,
1? and the equivalent & whioh is preserved in Khmer where it
means the bow fixed against the stag-fly (6k khltng). Cf. also Santali



LOANS Of INDO-ARYAN 2)



with the arcket is essentially a vibrating string attached
to the ends of a |*iece of wood. Sir G. Grierson has
described a similar but more simple machine in Bihar
Peasant Life, pp. 64-65. 1

If the bow for carding the cotton is used in Indo-China
as in the Malaya Archipelago and India, it will be shown
just now that the words of the same origin similarly
designate the shooting, the bow or the arrow, and the
cotton.

The tendency of making words monosyllabic has often
had the effect of reducing the ancient forms :



Mon


Khmer


Stieng


ROngo


MuoDg


Annamite






"to draw


the bow "






V


V


V


V


v


v


pan


barf


pS*


pan


pan


ban



These words differ from Bahnar panah, ponah, by the
loss of the final and by certain alterations of the nasal.
Moreover it is to be noted that the initial of the Khmer
form is a fc, an unstable phoneme, intermediate between
the sonant and the surd and of which the equivalent is
6 in Annamite while most of the other languages have p.

As regards Mon pnoh "bow for throwing stones"
we have :

Guru : panan " bow "

Kon-tu : panen " cross-bow M

Sedang : poneh, monen " cross-bow "

Halang menen " cross-bow."

I /. Sanskrit : tula-Tcarmuka, tula -cap a, tula-dhanus " cotton-bow ,
a bow or similarly shaped instrument used for cleaning cotton "
Monier- Williams). In the Himalayan dialects we have also : in
Ltpe*, tt ofofc "to make cotton fine with bow, to eard."



22 PRE-A&YAN AND PRE-DRAVIDIAN

In the Kol or Muruja languages, Santali banam means
(t violin, to play on violin " ; the last operation requires
the use of an archet or a little bow.

On the other hand the Indonesian forms can be
reduced in a large number to the type of panah. This
word designates the bow in Malaya, and in Java, the
bow and the arrow. Amongst the Dayaks of Borneo,
the bow is called panah. In numerous languages of the
Philippines, pana is the name of the arrow, and in
Mindanao panah is the name of the bow. Lastly in
Madagascar fana, falla, designate both the bow and
the arrow. M. Nieuwenhuis who has studied these
Indonesian forms reasonably admits that panah must
have meant, in ancient times, both the bow and the
arrow in all .parts of the Malaya archipelago (Art.
already referred to, p. 19).

The comparison of the Mon-Khmer forms thus teaches
us that panah is derived from the verb paH, poh "to
draw the bow " by adding an infix. One understands that
the name of instrument, thus formed, designates both
the bow and the arrow, i.e., all that is required
for drawing the bow. The origin of the Sanskrit word
bafta cannot be, therefore, any longer doubted. It is a
loan from the Austro- Asiatic languages and a very
ancient loan because the word can be found in the Rg-
Veda, VI, 75, 17. The sonant initial of ba^a was surely
not used to translate an Austro- Asiatic p in Indo-Aryan.
The b of the Yedic form is therefore, of a nature to
prove the antiquity of ft still found in the Cambodian
writing of our days.

The Aryans however, certainly knew the use of bow
before their entrance into India. Why have they then
borrowed from the Austro-Asiaties a word for the arrow ?



NON-ARYAN LOANS IN INDO-ARYAN 83

Probably the arrow made of bamboo was unknown to them
and this is why they borrowed the name as well as the
instrument itself from the aborigines of India. In fact, in
the Malaya Archipelago, the arrow called panah is made of
bamboo (Nieuwcnhuis, pp, 9 and 28). In the same way
bana designate precisely an arrow of bamboo or
of cane in India,

V 1

SANSKRIT karpasa.

The verbs pah, poA, Wk which have given origin to
the name of the bow and the arrow probably do not
represent the ancient form of the root. In the Austro-
Asiatic languages, a final h rises normally from an
ancient *. In Khmer, for instance, aml&oh "cotton 11
has another form anfbas. One can suppose, therefore,
that the verbs pah, poh, ftoA had originally a root
*ta* which meant the action of handling a bow
either for throwing projectiles or for carding the
cotton.

We now know enough for understanding the formation
of the following names which designate cotton in the
A us tro- Asiatic languages :

Crau paq, bag
Stieng pahi

Khmer cwfoas, anfoah Rade kapa*

Bahnar kopaiA Malayan "I

Sedang Wpl Javanese J kapa *

Kuoi kabas Batak hapa*

Kco kopas Cam kapah.

1 Bulletin, XXV, L pp. 69-71.



24 PRB- ARYAN AND PRB-DRAVIDIAN

At the base of all these forms, whether they have a
prefix or not, one finds the root has, of which the very
unstable initial generally becomes p or b and the final
has been sometimes softened into A with a compensatory
i in some cases. The name of the cotton fibre,
therefore, properly means " that which has been husked,
carded/ 1

In most of the Austro-Asiatic languages, the prefix
is simple : ka or ko. But we know that in this linguistic
family a nasal or a liquid is frequently inserted between
the prefix and the root. This can probably explain
Khmer : (k)amtto*, (k)amftoh of which the initial has
disappeared ; and in the same manner we can account
for Sanskrit karpasa "the cotton tree" which cannot
pal be exined by Indo-European.

Under the form x<*P 7ra with Hindi ko, ke, etc., is
accidental unless it is admitted, on the contrary, to be
a borrowing by Dravidian from Indo-Aryan. Even an
isolated expression, like the use of a word signifying



46 PRE-ARtfAN AND PRE-DRAVlDIAN

" having said " in Indo-Aryan, to mark the subordination
of propositions, is not to be invoked here ; because it is not
only in use in Marathi and Singhalese, languages in
contact with Dravidian, but also in Nepalese and Bengali
and at least in one language of the Tibeto-Burman group,
the Borio (Langue Marathe, p. 272 and Errata).

One is, therefore, ultimately led to search for the
Dravidian elements of Sanskrit only in the vocabulary.
But the history of vocabulary is absolutely different from
phonetic or grammatical evolution, and the loan of words
10 essentially different from the facts of a substratum.
The facts of a substratum result from the unconscious
blending of two systems existing amongst the same
people ; the loan results from a willing effort to add
elements taken from outside to the mass of the voca-
bulary. The loan proves the contact of the two languages
and not the substitution of the one by the other. On
the other hand it is often difficult to recognise in .what
sense the borrowing is made between two given languages
and to make sure that it has not been made by each of the
two languages from a third one, known or unknown.

Lastly, where it becomes clear that Aryan is
the borrower, it is necessary to determine from what
group of Dravidian the loan has been taken and
also to draw from it information for the history of
Dravidian itself. There is no question of undertaking
that work here, which is still impossible, but we only
want to point out by some examples, the interest and
the present aspect of the question.

The Vedic (and Indo-European) name for horse, afva,
is no longer represented to-day in Indo-Aryan except on
the confines of the Iranian world where the corresponding
word it still living (Grierson, Pilac* language, p. 78,



SANSKRIT AND DRAVIDIAN 47

and the list of Ling. Survey, No. 68). The word which
has replaced it in all other parts of the country oocura
in the Srauta Satra of Apastamba a text which appears
to be of southern origin (cf. Biihler, 8BE, II, p. xxx)
under the form gho\a. Mr. J, Charpentier has tried
(KZ, XL, p. 441) to identify this word with German
gaul ; this eqvivalence would be strange by itself ; Prof.
Sommer has shown (IF, XXXI, p. 362) that this
Germanic word has its correspondents in Slavonic and
not in Indian. On the other hand, the similarity of ghota
with some Dravidian forms with the same meaning has
long been recognised : Tel. gurramu. ; Can. Itu'lurc ; Tarn.
Kudirei (Gondi Kora is suspected to be borrowed from
Hindi ghora like Kui goda) ; the Dravidian form which
has preceded the Hindi word amongst the Gonds is
undoubtedly that which accounts for Gadaba Knita and
Savara knrta, alone of their kind in Munda. The Brahui
kulli is out of the question ; on the value of initial
A, cf. on one hand Br. hal "rat," het "goat/' hln "to
deposit "and Tarn. eli,ddu, In \ on the other Br. kur
and Gondi AwrA, kui sto} (cf. Tuttle, Am. J. Phil, XL,
p. 84).

It is easy to reconstitute the common prototype of
all these forms. *ghutr . In the same process one
gets some important data for the history of Dravidian
phonetics :

1st. The consonantal group has been eliminated in
Telugu by total assimilation, in Tamil and Canarese by
vocalic insertion.

2nd. In the last two languages, the intervocalic surd
is changed into a sonant. In Tamil, at any rate, the
date of this alteration is rather late, cf. MSL, XIX, p. 89 ;
for Canarese an indication is to be fouud in the fact that



PRB-ARYAN AND PBE-DBAVIDIAN 48

the name of Maski, the village where an inscription of
Afioka has been discovered is still Piriya-masangi in a
Calukya inscription (H. Krishna Shastri, The Maski Hock
Edict, p. 1).

3rd. In the same languages the initial consonant is
changed into a surd. Here from the Dravidian stand-
point the rule is not clear : there are two series of corres-
ponding forms. In fact M. Subbaya in his articles in the
Indian Antiquary, 1909 (where he always attributes
wrongly the surd to common Dravidian) has given a
series of equivalent forms : Tarn, k Can. Tel. g
(pp. 05, 217; c/\ for the dental p. 200). But in his
Dictionary of Canarese, Eittel gives a good number of
examples of the Tarn. Can. k, Tel. g similar to that in
the name of horse: thus Tarn. Can. kadal " love/ 1 Tel.
gadiln ; Tarn. Can. farn " to scratch," Tel. giiu ; Tarn.
Can. km "sheep," Tel. "fforre" ; Tarn. Can. Kuli "hole,"
Tel. " goyyi" The interpretation of the facts is difficult ;
but the antiquity of the sonants in Dravidian remains
undisputable.

If it were certain that the Sanskrit word was borrowed
from Dravidian one could have rightly deduced at once
a fourth observation, more important than all the previous
ones. In that case the most ancient Dravidian, in fact,
would have had aspirate consonants, either a dialect in
contact with Indo-Aryan having developed aspirates
in some cases or the aspirates having belonged to
common Dravidian. There is nothing inadmissible in this
view ; the interval is extremely long between the epoch
when ghota was admitted into Sanskrit and the
late date very likely the 5th century A.D. when the
alphabets of the North were borrowed by the principal
Dravidian languages : in fact it is known that the



SANSKRIT AND DRAVIDIAN 49

characters which mark the aspirates in In do- Aryan are
wanting in these alphabets. But in this case it must
be asked if Dravidiau itself is not a language brought
to the Dekhan, its presort area : because the loss of
aspiration is one of these typical facts which immediately
makes one think of the action of the substratum : this
substratum could not have been Muuda which possesses
aspirates. Dravidian, the language of the Dekhan,
therefore, would have been at first a language of the
North and the horse, in fact, is in India really an animal
of the North : it has been discovered in a fossil state
in the Siwalik mountains ; and the Vedas specially
mention the horses of Sind and the Saras vat i (cf. Crodke,
Things Indian, p. 253 ff. ; Macdooell-Keith, Vedic Index,
under agva), Thus one would again fall back on a
hypothesis, similar to the one already mentioned, about
the contact of two languages in the pre-historic period
in anterior Asia ; but it will have this time another
degree of historical probability ; the history of ancient
India can be explained to a great extent by the successive
floods of invasions of which the first is only an anticipated
consequence of the second : the Dravidians might have
preceded the Aryans, as the Sakas preceded the Kusanas
and as later on the Ku^auas again preceded the Huns.
The difference would be this that the Dravidians and
Aryans imposed their languages on India.

Thus questions that are brought forward are
important at least in the hypothesis that ghota was
taken from Dravidian. But the name of the horse is
essentially a name subject to renewal and no one can
foresee whence the new name would be taken. One may
think of ross, pferd and gaul without speaking of mdhre
and of stute and, in another domain, of caballm and

7



50 PRE-ARYAN AND PRE-DRAVIDIAN

mannus. If Dravidian be the borrower, or if both
languages took the same word, perhaps along with
the specimens of a particular breed, for instance, of
Iran or of Arabia 1 the entire edifice, phonetic as
well as historical, will collapse.

The name of the " ass " suggests a problem analogous
to that of the "horse." The identity of R, V. gardabha
(on the suffix Skr. -b/ta Gr. -o-, etc., of the names of
animals, see Brugmann, Qrundriss, II, I, p, 389), Hindi
yadha, etc. (borrowed freely in Dravidian, in Munda and
in Assam by the Khasi ; see the lists of Ling. Survey, 74)
on the one hand, and Tel. gadide, Can. kalte, kattc. Tarn.
kaludei on the other hand is evident (Kurukh gadrarna
** to bray," Is it Dravidian or Ar^an? In the Celebes
the language of the Bug tribes has a form borrowed
from the Dravidian, kaledde) ; a prototype *gard accounts
for all the forms if only because the rule concerning
the consonantal groups previously stated admits of
an alteration in the case of a liquid preceding the
occlusive instead of following it. In fact one finds
the use of the Can. kalte, katte, in the word for "rice,"
Can. flfefet, Tarn, artfi, forms of which the comparison
is sufficient to suggest an ancient *arki, or in the word for
the "cat" (admitted in Sanskrit at the time of the
epic, Skr. bid- a la-, bil-ala-* bir-dla-, Ka^ra brar, Syrian
Gypsy blarij Hindi bilari, i,ilaya> billt, etc., from which
secondarily Can. etc. pilli ; see the lists of the. Ling.



1 M. Autran would like to explain the Egyptian word fytr which
means the carriage and the horse as a loan from an unknown
language of Southern Arabia. We know that horse appeared in Egypt
only towards the 16th oen. B.C.



SANSKRIT AttD DRAVlDlAtf 5l

Survey , No. 71), Can. berku, bekku, Kur. ber.ra, Qondi
boka. Tarn, verugu.*

What is the origin of this *gard common to Indo-
Aryan and Dravidian ? The presence of this word in
fygveda has led etymologists to search for an Indo-
European origin. Some connect it with the Romance
word for " male " admitted very late into Latin (v. Walde,
under burdo ; cf. Ernout, Elem. dial, du vocab. latin, p.
13&) ; Prof. Wackernagel has thought of English colt,
which 'primarily designates the little ones of an animal,
and particularly in the Bible and in Middle English
the young one of a camel or an ass ; agreements
which are very poor and far-fetched the ass has no
Indo-European name. The ass is an animal of Asia ; it
is rare in India except in the Western regions (cf.
Levi, BEFEO, IV, p. 568). The Mediterranean name

of ass, Gr. OFOS etc., appears to come from Western Asia ;
khara , which is wanting in the most ancient Sanskrit
texts, is known only in India and in Iran ; hence it
is not at all astonishing that the Vedio words gardabha
and rasabha have not any corresponding Indo-
European terms, just as it is natural that Brahui has
a name, which as far as we know, belongs only to it,
(bis). The probabilities are, therefore, in favour of *gard

1 Tamil has another word pwuet, p$0t ; one is inclined to connect
pS0u f Can. pwtu, and Tel. p*yu, " to smear " with one another : the
Semantic relation will recall classical Skr. marjara (which has the same
suffix as bir-ala) ; but we find in Munda pSri, in Tibetan pisi (beside
byila borrowed from Indo- Aryan, cf. Lanfer, Tibetan Loan Words, n.
64), in Afghan pto, in Persian pusek ; in the North* West of India
put, and 6uit (Grierson, Pit. Lang., p. 66) in Brahui pisi. The
words are independent of each other and are remits of onomato-
poeia; the same is found in Europe, puts, etc.



5B PRE-ARYAN AND PRE-DRAVIDIAN

being a local word existing on the confines of India and
Iran. Hence we are again faced by the fundamental
problem : has Dravidian supplied the word to Aryan
and is it the first language that the Aryans met with
in India ? Or have both Dravidian and Sanskrit borrow-
ed the name of the ass from a third language which,
at any rate (to judge by the lists of the Linguistic
Survey), can be neither Munda nor a language related to
the mysterious Burusaski ? Or lastly is it not Dravidian
which took the word from Sanskrit ? It is impossible
to give an answer for the time being.

That the two families have been io contact with
each other for long, there is no room for doubting.
There are facts which prove it but which raise new
problems too.

One has identified (6, A. Jacob, J.R.A.8. 1911. p.
510 ; D. R. Bhandarkar, Anc. Hist, of India, p. 26) matacl
found in the Chandogya-Upanis.ad with Can. midice
" grass-hopper." The relation between Skr. ma- and Can.
mi- is not without analogy ; it is, for instance, difficult
to separate the different words for " black pepper/' Skr.
marica, Tarn, milagu, Can. melasu from each other. But
one is led to ask if a family of Dravidian words express-
ing size is not entirely borrowed from Aryan, Skr. ma ha,
Can. i*ige, Tarn, migei " abundance," Tarn. Can. Tel.

V M

minctt minju " greatness, excellence," Can. mikku
"excess," etc. (cf. Caldwell, Compar. Gramm.*, p. 602);
Kur. mecha "high," megro "elder." If it be so, then
of the two languages Aryan may be considered as
having the most prestige and very likely as being
the least open to borrowing and all the more to the
phonetic and morphological influence of a Dravidian
substratum.



SANSKRIT AND DRAVIDIAN to

There are cases in which a language would submit
to the influence of another without borrowing complete
words. It can be asked whether the word for " wheat "
which is found from the Yajurveda onwards, e.g., godhumah
(in the singular in the Satapatha BrShmana) does not owe
its form to such an influence. This word apparently signi-
ficative but having an absurd signification (" smoke of
the cow "), cannot be separated from the Iranian gandum,
which being in no way significative, is necessarily the
most ancient. Cannot the deformation undergone by -the
word in India be explained by the presence of a word
with the same meaning in Dravidian, Can. godi, Tarn.
kodi, Toda kodj '( One would be inclined to explain, by
an inverse contamination, the double aspect in classical
Sanskrit of the word for " fan " vljana and vyajana
alternating in an abnormal way ; it looks as if that a wcrd
expressing the instrument derived from the root of Can.
btxu, Tarn, vifu, Tel. mcu, vUaru, and vivu "to swing,
to fan, to blow " was at the time of its introduction into
Sanskrit, formed on the model now of vij. and now of



These diverse aspects, presented by the problem of
loans are not the only ones. There are others in which
non-Munda languages must be counted.

Let us at first come back to the names of animals.
A name which has a good chance of being Indian is that
of the " peacock " and it would be in no way astonishing
if in fflce of Rv. mayura and in the forms supplied by
Asoka, mora at Girnar, majura in the North- West,
majula at Kalsi and Jaugada, we find a group of Dravidian
forms : Tarn, mayil, Can, maylu and navil, Tel, mail
Gondi mal. The identity of the names is evident ; but
it is difficult to determine the ancient form. If 'it is



54 PRE-AKYAN AND PllE-DRAVlDIAN

admitted with Mr. T. Michelson (J.A.O.S., XXX, p. 84,
n. fi) that the -j, of the inscriptions of the North- West
is " Magadhism " one still remains embarrassed by the
co-existence of the forms with -I- and -r-. Should one say
that the contact took place between Dravidian and the
Eastern dialects of Sanskrit ? It would be a further
definition of great value. But Eastern Munda possesses
a word of similar appearance, with r ; e.g. Savara, mara,
Santali maralt ; and this word re-appears in Indo-China :
M5n mrak, Bahnar mra (to tell the truth, Father Schmidt
connects these two forms with Skr. Pali bar hi, derived
from Aflr^a -another word of unknown origin). One does
not know if the two series should be put together or not.

Is the Tamil word pa\am rt ripe fruit " copied from
or the original of the Vedic phctta ? Here the difficulty
is manifold. One can imagine the Indo-European
etymologies {cf. Uhlenbeck, s,v. ; Wackernagel Altind. gr.,
I, pp. 120, 128; M. Meillet proposes Old Slav, plodu
" fruit '0- But one can also refer to Can, j?an, Tel. pondu
Kur. panjna "fruit," possibly even to Brahui pin
" to swell up; " the nasal does not cause any absolute
difficulty, Oanarese has menasn by the side of me\asu quoted
above as the designation of " pepper " ; it gives unaJce
in face of Tarn, nlakkei, Gondi usbal, Toda wask- " pestle/ 1
If the connection were proved it would be most probable
'that phala was borrowed from Dravidian. But " fruit "
is called in Khmer phl$, in Kaseng plei, in Bahnar plef,
in Stieng pUi ; and Prof. Przyluski who communicates
these words to me adds that, in his opinion, they could
not have been borrowed from India, because Annamite,
in which there is no Indian influence, has trai which
goes back to blai attested in the 17th century by Father
Bhodes.



SANSKRIT AND DRAVID1AN 55

It is very carious that the same problem arises about

a word which is the name neither of an animal nor a

plant, nor the name of any ordinary article. Of the

ancient Indo-European word for " mouth " occurring in the

Rgveda under the forme, as-, asan-, asfflya, there remains

no trace to-day except in the dialects of the mountainous

regions of the North-West (cf. Grierson, Pis. Lang.,

p. 75 ; and the lists of the Ling. Survey, No. 36). Besides

this word and the mysterious prdty andm I, 52, 15, 37

(from which the word anlkam "face" is derived) the Rgveda

offers some examples of a new word mukJia , the

use of which appears to have been already current :

it is applied to the author of a hymn IV, 39, 6 ; to

Agni VIII, 43, 10 (cf. YigidtomMa, I, 97, 8;

X, 81, 3) ; to the Purua X, 90, 11 ; it designates the

point of the arrow VI, 75, 15 j in a comparatively late

hjmn I, 162, 2 muk/iatah is translated "by the bridle";

which presupposes that mukha was used for the mouth

of the horse. Whence comes this word which is used

everywhere in Indo-Aryan to-day (except in Sindhi in

which there is a representative of Vaktra-) and which

the Afghan has borrowed (max) ? The Indo-European

words which are usually referred to, Lette mute

Got. munps, old High German mula (and even Skr. mula-

"root" if the conjecture of Prof. Wackernagel is

accepted. Sitzbet. Beilin, 1918, p. 410) are of known

formation ; but one would search in vain for -kha- amongst

the normal suffixes in Sanskrit (mayftkha- "nail," "peg"

is solitary and recalls modern Iranian, Persian mex, etc. [See

the works of P. Horn, and Hiibschmann under No. 1005]

without it being possible to propose a common ancient form).

Now, if we admit that Indo-European of India had

any derivative of original *nu- then its deformation might



56 PRE-ARYAN AND PKE-DRAVIDIAN

be attributed to local influences. By a still simpler
process, the old word as might have been replaced by a
popular form borrowed from the native tribes. One will
therefore, be inclined to accept with slight modification,
the identification already proposed by Gundert and Eittel
of mukha- with the Dravidian words for " nose." (Jan.
Mrityn, along with mu, Tel. mukku, Tarn, mukku, Gondi,
massor, Malto mutoth, Brahui bamus (where da is the
Dravidian term for " mouth" ; see the list of Linguistic
Survey, No. 36 ; for the words for " nose," No, 34), Kui
mungeli* ; these names appear to be authentic because
they are connected with all the words expressing the idea
of "in front" (Can. Tel. mu, Tarn, mnn Kur. nund ,
Brah. man "in front," Can. muti. "face, mouth," Toda
mun " face," Tarn, mudal, Kur. muddh " first " etc.). That
the term for " mouth " or " face " would be subject to
renewal, is not at all astonishing ; mukJia- itself has
in modern Indo-Aryan another rival in : Mar. tond, Guj.
Beng. lund, Singh, tufa Ma; this word was previously
applied to animals ; in Pali and in Sanskrit tunda
designates " trunk, beak, snout " ; it is evidently the
same as Tarn. luwiLi "beak," Gondi tuddi "mouth,
face " ; probably Malto toroU " mouth " (on the contrary
Tel. 'ofidflww "trunk" appears to be a loan word, and
Can. tnti " lips " recalls too much Bena;. thomt, deforma-
tion of the term for " lips," Mar. etc. ott, Skr. ont/ia-
to be taken into consideration).

In the first place, therefore, the probabilities would
stand for mukha- being a loan word from Dravidian.

1 It is carious that Kft$miri muk means "short and flat (nose) "
while Skr. miifc- (Mar. mukd etc.) "dumb" is related to the family of
Qr. fivx^i Arm. ftiimj, Lat,



SANSKRIT AND DfUVIDIAN 57

ID such a case one would be convinced that Dravidian
had certainly, as the history of ghoia- made us suspect,
aspirate occlusive ; and hence one would be justified in
suggesting new equivalences for it. 1

But it is not confirmed that mukha- comes from
Dravidian. Let us consult the Munda. lists of the
Linguistic Survey. On the one hand the North-Eastern
group gives for " mouth " a word mo<& ; we cannot
say in the present state of our knowledge if it has any
thing to do with Vedic mukha- but it curiously reminds
us of the modern names of " moustache " in two other
families : H. mufaki mvcv, Mar. iwi, and Can. mse, t
Tarn, migei On the other hand, the word for nose is
everywhere mu or m& ; and Prof. Sten Konow has pointed;
out in his Introduction, p. 13, that Bahnar has muh ;
and Prof. Przylaski communicates to me the following
list : Khmer ciamuh, Stieng tromuh, Anna mite mui (tbc
substitution of i for an ancient final is regular in
Annamite), Mon and Bahnar muti, Sedang wh, and lastly*
Guru and Semang (the last of the Malaya Peninsula), muk
which very likely preserves the most ancient form* We
CAB scarcely see how to classify all these forms. Besides,,



1 For instance modern In do Aryan (Gypsy included) -pfctr. " turn,
change," it of unknown etymology (what is said in Langue Marathv
p. xii, and in the erratum on p. 181 is hardly more satisfactory
than before). There might be relation between Can. peta, Gondi pt/ja
"behind, in the back " (Can. peratege " to draw back, 11 ' to come
back"), Tel. perafu "facade in the back,*' Tarn, pira, Tel, per,
Brahuipen "other" (Brah. per, "roll up" must be rather related
to Knr. pes " pick up "). Of course there exists in Tibetan an adverb
phyir ''newly, re," but Prof . Przyluski falls back on LepGb*6yt7
" recommence " and other analogous forms, and suggests that a toot'
bal or byd, meaning " to repeat " most haye been at the basil.

8



58 PRE-ARYAN AND PRE-DRAVIDIAN

it is good to bear in mind that at the root there is
an onomatopoeic word, on account of which the agreements
are possible. One knows indeed the difficulties presented
by the etymology of the words like Or. fiuiSos, Lat. mugio,
m&tva, French mustau, etc.

The conclusions which are drawn from all that has
been said and which it is necessary to formulate in order
to oppose a tendency to which one has been tempted
hitherto to yield too easily are above all negative. In
the present state of our knowledge, there is nothing which
permits us to affirm that the aspect assumed by Aryan
in India is due to its adoption by a population speaking
Dravidian languages. If there is any substratum at all,
it can be searched for equally well in other families,
especially in Munda.

On the other hand the vocabularies furnish a proof
of very ancient relations between the populations speaking
Sanskrit and Dravidian. But in what did these relations
consist : superposition and substitution from Sanskrit to
Dravidian, direct contact or indirect exchanges ? It is
impossible to determine that. So far as there is a
chronology of the Sanskrit texts these relations can be
dated at the earliest by the end of the Vedic period and
would be localised at first in Northern India. One
would like to ascertain which dialects, Dravidian or
Indo-Aryan, weie involved in it : unfortunately the
evidences are confusing. The initial b- of bidala confirm-
ed by Ka?miri and Syrian Gypsy is to-day in Dravidian
the characteristic of the Canara-Kurukh- Brahui group ;
the v of Vijana-ryaiana- (if the interpretation suggested
above is taken into consideration) characterises the Telega-
Gondi-Tamil group ; as the division of b and v bet ween
the Western and Eastern dialects in Dravidian corresponds



SANSKRIT AND DRAVIDIAN 59

with that in Aryan, one could say that here is a proof
of the two ways of exchange : it is possible, because
these loans do not count among the most ancient ones.
On the other band the name of the "peacock," for
instance, would give the occasion for a discussion on the
alteration I \ r \ but it has been seen that Eastern Munda
contradicts Dravidian.

Perhaps the principal interest for ourselves in the
study of ancient loans (and it would be necessary to try
both ways since Dravidian has borrowed much from
Aryan) would be to form an idea of prehistoric Dravidian ;
because even those Dravidian languages which have a past
are only attested in a definite way, for the first time, a few
centuries after the Christian Era. Moreover the complica-
tions we have met with, suggest that Dravidian
like Sanskrit may have taken loans of vocabulary from
Munda, which must be at least as ancient as Dravidian
in India. As far as the borrowings made by Sanskrit
are concerned, we have seen that the notions formed up
till now are to be either revised or further defined and
with the advancement of research new snares and problems
do arise. If it is no reason for giving up this research
it is one for bringing into it much caution and for
leaving necessary room for possibilities to which hitherto
too little attention has been paid.



PART III



PRE-ARYAN AND PRE-DRAVIDIAN
IN INDIA



BY

SYLVAIN LEVI



Pre- Aryan and Pre-Dravidian
in India 1

The geographical nomenclature of ancient India
presents a certain number of terms constituting almost
identical pairs, differentiated between themselves only by
the nature of their initial consonants. I propose to
examine some of them here.

1 . Kosala-Tosala. The name of Kosala is familiar to
the Sanskrit epics. The RamSyana begins with the eulogy
of the country of Kosala, on the banks of the Sarayu ;
Dalaratha, the father of Kama, is king of the country
of Kosala ; the mother of ll&ma is Kausalya " the
Kosalian " ; the city of AyodhyS, the capital of the
kingdom of Kosala, is commonly designated as Kosalft.
The MahabhSrata often mentions the people and the
city ; it associates the Kosalaus with Kadi, Matey a,
Karusa, Cedi, and Pu Q 17. Of the Kosalans, the Maha-
bhSrata distinguishes those of the East (Pfirva , Prak)
and those of the North (Uttara ) ; the Rftmftyana die tin-
guishes those of the North (Uttara ) as the Kosalans par

1 Journal Atfatique, Tome ocjii (1928), pp. 1-57.



64 PRE-ARYAN AND PRE-DRAVIDIAN

excellence (VII, 107, 7). Later on, Kosala proper
(Kosala-defia) or Great Kodala (Maha) received the
designation of Southern Kosala (Dak$iua) ; it is under
this name that Kosala is frequently mentioned in
the epigraphy of the Middle ages. While the Northern
Kosala is the country of Oudh to the North of the Ganges,
the Southern Kosala extends on one side up to Berar
and Orissa and on the other up to Amarakantak
and Bastar. The region of Chhattisgarh along the upper
course of the MahSnadi is its nucleus.

The name of Tosala has not acquired the same
celebrity as that of Kosala. It is met with, coupled
with the name of Kosala and probably saved from oblivion
through the prestige of its twin, in Atharva-Teda Pari&ata,
Chap. 56, in a list of people connected with the South-
Bast ; the Kosala of this passage is, therefore, Dakgina-
Kosala ; it appears in the same way in the geographical
lists of some of the PurSnas (Matsya P. 118, 58;
Markanfcya P. 57, 54 -7ayu P. 45, 183 : Tofialsh
Kodalfth) ; it is still the same even in the curious resume of
Indian geography introduced by Yagbhata in the commen-
tary on his art of Poetry (Kavyanu&sana, ed. Kavyam&lS,
p. 4, 4) : Varanasyah, par at ah purvadesah \ Yatr Ahga

Kalinga Kosala Totala-Otkala ; Hemacandra has

reproduced the same list in his treatise on the same
subject, which bears the same title (Kavyanu'sasana, ed.
Kfivyamala, adhy. 3, p. 127). Tosala or Tosalaka, "the
native of Tosala/' is the name of a wrestler vanquished
by Krsna (Harivatfn&a, II, 80, 50 ; 48 ; 55 ; Vi$y,upuranfi,
Iran. Wilson 5 , Vol. V, p. 89). Tosaliputra, Prakrit
Tosallpntta, " the son of the Tosalian, " is a Jaioa Acarya,
who was the teacher of Arya Rakajta or Kaksitasvfimin,
disciple and successor of Vajra, the last of the Dasapurvin



PRE-ARyAN AND PRE-DRAVIDLAN IN INDIA 65

(Avasyaka ; -nijjutti 8, in Ind. Stud., XVII, 68 ;
Hemacandra, Parmstaparvan, XIII, 88 j. The name of
Tosall (in the feminine) is hardly known to Indian is ts
except from the inscriptions of ASoka ; two of the
different edicts, of Dhauli, are addressed to the Kum&ra
and the Mahamatras at Tosall (Tonaliyam Mahamata
nagaravit/ahalakai \ ; Tosaliyam Knmale Mafiamata ca, 2).
The name of Tosall must have been applied to a
region, because we find the mention of Northern Tbsall
(Uttara-Tosali) and Southern Tosall (Daksina-Tosall) ;
the King of Orissa Subhakaradeva, who reigned towards
the end of the 8th century, while presenting to the
Emperor of China his own copy of the GaniLavyuha in
795, issues a document conveying the gift from Uttara-
Tosall (Ep. fnd., XV, p. 8). A deed of gift by Sivaraja in
288 (Gupta) i.e., 601 A.D. (Ep. Ind., IX, 286), discovered
at Patiakella, mentions Daksina-Tosall in a rather obscure
context, either as the residence of his suzerain, which is
the way the editor of the inscription, Mr. Banerji, takes
it or as the district in which Vortanoka whence the
document issues, was situated. Both of these documents
have been discovered in Orissa in the district of Cutback.
It is also in this district that Dhauli is situated at a
distance of 4 miles South- South- West of Bhuvanesvar ;
the inscriptions of Asoka are engraved there on a rock
called Aswastama, near the summit of a low bill. It is
therefore evident that Tosall occupied almost the same site
as that of the Dhauli of to-day. There is no accounting
for the indication furnished by Ptolemy who places Tosalei
or Tosale in India beyond the Ganges, at 150 East and
23 20 X North, on the way from the Ganges to the
peninsula of Gold (Khruse Khersooesos), in the vicinity
of the Kirata (Kirrhadia, Tiladai), in the centre of a,

9



66 PRE-ARYAN AND PRE-DRAVIDIAN

region which corresponds to modern Sylhet and Manipur.
To add to our confusion, Ptolemy places at 5 South and
4 East of Tosalei, a city called Trilingon or Triglypton,
which he qualifies as jfturt'Actov " the royal residence," and
which may very well be Trilinga, of which we shall have
to speak later on, a region situated in fact to the South of
Tosall, rather to the South-South- West, along the Western
coast of the Bay of Bengal, in cis-Oangetic India according
\ to the division adopted by Ptolemy. The other towns enu-
merated by Ptolemy in the same paragraph have not yet
been identified : Rhandamarkotta, where there is an
abundance of nard ; Athena gounon, Maniaina (Maniataia),
Tosalei, Alosanga, Adeisaga, Kimara, Parisara, Tougma
which is a capital (metropolis), etc. For the first of these
names a suggestion may be offered en peasant, which may
lead to its definite identification by discarding at any
rate all the previous identifications which McCrindle has
summarised in an important note (Ind. Ant.> XIII, 382).
" Rhadamarkotta (v. 1. Rhandamarkotta). Saint Martin has
identified this with Rangftmati, an ancient capital situated
on the western bank of the lower Brahmaputra, and now
called Udfipur (Udayapura, city of Sunrise). Yule who
agrees with this identification, gives as the Sanskrit form of
the name of the place, ftangamrtika. The passage about
JNard which follows the mention of Rhadamarkotta in the

majority of editions is, according to Saint Martin (Etude,
p. 352 and note), manifestly corrupt. Some editors
correct iroAAi/, much, into WAcis, cities, and thus Nardos
becomes the name of a town, and Rhadamarkotta the
name of a district to which Nardos and the towns that
come after it in the table belong. On this point
we may quote a passage from Wilford, whose views
regarding Rhadamarkotta were different. He says



PRE-ARYAN AND PRE-DRAVlDlAN IN INDIA 6?

(Alia*. Research, Vol. XIV, p. 4-41), 'Ptolemy has
delineated tolerably well the two branches of the river of Avft
and the relative situation of two towns upon them, which
still retain their ancient, name, only they are transposed.
These two towns are Urathena, and Nardos or Nardon j
Urathena is Rhftdana, the ancient name of Amarapur,

and Nondon is Nartenh on the Kay n-d ween ' He

says that ' Nartenh was situated in the country of Rhanda-
markota, literally, the Fort of Randamar, after which
the whole country was designated.' All the exegetista
appear to me to have gone wrong ; Wilford, however,
had a glimpse of one part of the solution. The Sanskrit
name of nard is nalada ; a metathesis, always easy in
the case of r in Sanskrit, has given rise to Ian (a) da and
then randa. It may be as well noted that the aspirate which
accompanies the initial r of rhando* or rhado* is a purely
Greek feature, and does not imply any aspiration in the
original word. As to the alternation of / and r in the
name of nard we have a sure trace in the gana kisaradi
on Panini, IV, 4, 58 ; the grammarian prescribes that
for designating the merchants of certain perfumes one
must have a derivation in ika from the name of the
perfume. The ganapafba gives immediately after Kisara
the words narada and nalada ; Bohtlingk, P.W*, under
narada, does not hesitate to recognise in it the name of
nard. I find that Candragomin in the corresponding
gana (ad III, 4, 55) has omitted narada and retained
only nalada. Thus the annotation which accompanies
the name of the locality in Ptolemy is occasioned by the
name itself, which it explains. I do not know how one
should restore the final syllables. The nalada in botanical
nomenclature is Nardostachys Jatamansi or Nardus
Indious ; Khory and Katrak (Matcria, II, 344) indicate



68 PRE-ARYAN AND PRE-DRAVIDIAN

the alpine Himalaya as its habitat ; Yule and Buraell
likewise (Hobton-Job&on, under nard) indicate that the plant
Nardostachys Jatamansi is "a native of the loftier
Himalaya." If Rhandamarkotta abounds in nard, it
must, therefore, be situated either in the Himalayan
heights or must be in such a vicinity as to be able to
serve the market. Rhandamarkotta, therefore, leads us
towards upper Bengal ; we may ask what error of informa-
tion could have Ted Ptolemy to locate Tosall (Tosalei,
Tosale) and Triliriga (Trilingon) to the east of the Ganges.
And yet Ptolemy was not ignorant of the importance of
Tosall, for he has himself termed it a capital (metropolis).

However that may be it remains certain that Tosall
was situated in the district of Guttack, in Orissa, and
that the present village of Dhauli stands on a site near
to, or identical with that of Tosall. It can be then
asked whether the very name of Dhauli does not represent
the ancient name Tosall ; the two names sound so
strangely alike that mere chance seems out of the
question. The transformation of Tosall into Dhauli is
not a phonetic impossibility. The intervocalic sibilant
of Sanskrit can, and in certain cases must become a
simple aspirate in Prakrits (Pischel, 264), for example
diaha^divasa and still better dukala beside dftsara " un-
fortunate " =duhsara. If Tosall could likewise develop
into Tohall, 1 this unintelligible name could suggest Dhauli
" the white." Nevertheless, it must be admitted that the
widening of the intervocalic * is frequent only in the
North- Western group : Sindhi, Panjabi, Kashmiri it

1 Let me point out, without any intention of drawing any
argument from it, that Varaha Mihira, Brh. S, XIV, 27, classes
amongst the populations of the North, beside the Huna, the Eohala
for which the commentator Utpala substitutes Rofiala*



PRE-ARYAN AND PRE-DRAVIDIAN IN INDIA 69

is already rarer in Gujrati and in Rajputana (Jules Bloch
Langue Marathc, 160). But the phonetics of place-
names leave the gate widely open to fancy.

A text which has not yet been mentioned will perhaps
help the solution of the problem of the site of Tosall ;
I have found it in the Ganclavyuha. The Gamlavyuha
is a Sanskrit-Buddhist work preserved in Nepal and not
yet published. Raj. Mitra has given an analysis of it
in his catalogue, The Sanskrit -Buddhist Literature of
Nepal, p, 1,0. Its extent is considerable. In reality,
however, it is only a fragment ; it forms the last
part of the vast collection which bears the title
of Avatamsaka, the entirety of which is preserved
in the Chinese and Tibetan versions. On account of
its importance the Avatamsaka has been, on two
occasions, completely translated into Chinese under
the direction of Buddhabhadra between 398 and 421 ;
and under the direction of oiksSnanda between 695 and
699. The section which forms the Gan with
a palatal sibilant has also been adopted for general use;
this had the advantage of avoiding th e difficulty ; it had
still more appreciable advantage of connecting this
embarrassing ethnic with a family of common words,
Kosa > Kusa, Kn'sala, which contain the palatal sibilant.
Tosala has been no less affected ; it has been attracted
by the analogies of the words tosa, etc., which express
satisfaction; we have therefore more often Ityala, but
sometimes Tosala also as Kosala.

2. Aiiga-Vanga. These two names are so familiar
throughout Sanskrit literature that they hardly need
explanation. Anga is already mentioned in Atharva-Teda,
V, 22, 14 by the side of Magadha, as the eastern limit
of the Aryan world. Vanga (Bariga) still survives in
the name of Bengal ( = Banga + ala). Anga and Vanga,
most often coupled together, have ordinarily as their
companion Kalinga to which we shall return presently.
All three, with Pundra (and Suhma), which we shall also
have to consider, are represented (Mahabharata, I, 104)
as five brothers born for the benefit of King Bali, from
a union accomplished, at his request, between the queen
Sudesnfi and the blind old rsi Dlrghatamas ; the whole
chapter has such a singular scent of savagery that the
Indian translator, the author of the English version
published by P. C. Roy, has been obliged several times
to take recourse to Latin for the sake of decency. Here
we are doubtless confronted by old local legends which



PRE-ARYAN AND PRE-DRAVIDIAN IN INDIA 78

the study of folklore will discover in the Austro-Asiatie
domain. Anga and Vanga had long remained suspect
to the Aryans of India. Baudhayana, so rich in curious
features, prescribes (I, 2, 14) a sacrifice of expiation after
a travel amongst the Aratfca, the Kftraskara, the Pumjra,
the Sauvlra, the Vanga, the Kalinga, and the Prftn&na
(Arattan Karaskaran Pundran Sauvlran Vanga-Kali'bgan
Pran&nan iti ea gatva punastomena yajeta sarvapf^thaya
va). It will be noticed that Vanga and Kalinga are
united in a compound noun while the other peoples
are mentioned one by one. In the stanza which
precedes this one BaudhSyana had related a verse
which classes the Atiga amongst the half breeds :
Avantayo'hga Magaclhah Burastra DatyinHpathah I Upavyt
Bindhmauvlra ete samktrnfiyonayah. The very reasons
which attributed to these countries a bad reputation
in the Bri&hmamcal society assured them a privileged
rank in the heretical churches. For the Jainas, Anga
is almost a holy land ; Cam pa, the capital, is the
residence of a large number of holy personages of
Jain legend and history. The BbagavatI places
Anga and Vanga at the head of a list of sixteen peoples,
before the Magadha (Weber, Ind, St., XVI, 804). One
of the Upfifigas, the PrajfiapanB, classes Aftga and Vafiga
in the first group of Arya peoples whom it calls the
KJietiariya ; the list begins thus : Rayagiha Magaha,
Campa Aihga taha, Tamalitti Vanga ya (ibid, p. 897).
Buddhism incorporates Anga in the classical list of sixteen
kingdoms; Vanga occupies an inferior position. The
Anguttara-nik&ya makes mention of it only once (I, 213)
in the list of sixteen kingdoms; everywhere else the
place is occupied by the Vaibsa (Sansk. Vatsa) the later
Buddhist literature constantly put together Afiga am)

10



74 PRE-ARYAN AND PRE-DRAVIDIAN

Variga, Ariga corresponds to the district of Bhagalpur
and Vanga to the districts of Birbhum, Murshidabad,
Burdwan, and Nadiya in Bengal.

3. Kalinga-Trilinga "Kalinga comprised all the
Eastern coast between the Utkalas, on the north and
the Telingas on the south. The Vaitaranl flowed
through it ; the Mahendra mountains (the Eastern Ghats)
were within its southern limits. Kalinga comprised
therefore, the modern province of Orissa, the district
of Ganjam and probably also that of Vizagapatam."
(Pargiter, Mark. P., p. 331). We have just seen the
close relationship which binds Kalinga with Anga and
Vanga, and the nature of the reprobation which they
received in common from the BrShmanical schools.
Kalinga had even the honour of having a special verse
devoted to it in the code of BaudhSyana, a traditional
verse which the legislator adopts on his own account
(I, , 15) : " The adage is cited : it is to commit a sin
with the legs to go to Kalifiga ; for its atonement, the
saints prescribe a VaisvSnara libation (atrapy udaharanti,
padbhyam sa frurute papam yah Kalingan prapadyate \ f$ayo
niskrtim tasya prahur raisvanaram havih). The juristic
compilations of the last centuries continue to register, as
an echo of this reprobation, another traditional verse : " If
one goes to Anga, Vaftga, Kalinga, SaurS$ra and
Magadha except for a pilgrimage, it is necessary for him
to receive a new sacrament."

Anga Vanga Kahngequ Sauraqtre Magadhesu ca \ ttrtha-
yatraih vina gacchan punah sartwbaram arhati (cited by
R. P. Chanda, Sir Asutosh Volumes, III, 1, 10, 7).

Regarding Kalinga the Mahabharata presents a
curious hesitation in course of the same canto, at an
interval of some verses in VIII, 44, 2066 ; the Kalifigae



PRE-AR^AN AND PRE-DRAVIDIAN ifr INDIA f 5

are enumerated amongst the tribes whose religion is bad
(durdharma), pgle-m&Ie with the Karaskara, the M&higaka,
the Kerala, the Karkotfaka, and the Vlraka; but in VIII,
45, 2084, they are counted amongst peoples who know
the eternal law (dharmam jananti sasvatam) in the com-
pany of the nations who are the highest of Brahmanism,
Kuru, Paficala, Salva, Mateya, Naimisa, etc. This change
of attitude is undoubtedly due to the importance held by
Kalinga since the time when the Indian civilisation spread
along the Bay of Bengal. We know that the conquest
of Kalinga, at the cost of streams of blood, provoked
the moral crisis from which the Emperor Asoka came
out transformed. After him, under Kharavela, Kalinga
became the centre of a powerful empire of which the
chief assumed the title of Cakravartin. Buddhism had
one of its holy places in Kalinga : this was the capital
of the country, Dantapura, " the city of the tooth " whence
the holy relic was later on transported to Ceylon, Pliny
mentions on several occasions the Calingae (VI, 18 ;
19; 20). Ptolemy enumerates a city of Kalliga (VII,
i, 93) amongst the Maisdloi, between the Kistna and
the Oodavari. Kalingapatam, port of the district of
Gaii jam, still preserves the old name of the region. The
appellation of Kling, applied to the Indians of all origin
all through the Malayan world, attests the brilliant
rdle of the men of Kalinga in the diffusion of the
Indian civilisation in the Far East. (See Hobson-Jodson,
under Kling).

The term symmetrical to Kalitiga appears in the written
documents only at a later date ; it takes diverse forms which
present the terrible perplexity of the scribes in face of
a kind of monster. The P. W. records the forme Trilinga
; the Mirkanrfeya P., 58, 28 and the Vftyu



76 PRE-ARYAN AND PRE-DilAVIDIAN

P., 45, 111 write : Tilanga. We find in tbe inscriptions
also Tilinga (Ep. ltd., XIV, 90), Telumga (ibid, XIV,
271), Kriltoga (ibid, XIV, 361), Trikalinga (ibid, XII,
208 and pass.). The Arab and Persian authors write
Tilang, Tiling, Tilingana ; in the nomenclature of the
languages of India, the language of this country is called
Telngu. An inscription of the 14th century thus traces
the limits of the country : " To the West and to the East,
two famous countries, Mahara^ra and Kalinga; to the
South and to the North, Pancjya and Kanyakubja;
it is that country which is called Tilinga " (pascal
puraslad yisya desau khyalau Alaharastra-Kalinga-samjftau \
avaff udak Pandyaka Kanyakubjau desas sa talrasli
Tilinganama. Srirangam Plates, Saka 1280 in Ep. Ind.,
XIV, 90). The region thus defined covers the greatest
part of eastern India According to the notice on the
Telugu in the Linguistic Survey, Vol. IV, p. 577, "The
Telugu country is bounded towards the East by the Bay of
Bengal from Barwa in the G an jam district in the north to
near Madras in the South. From Barwa the frontier line
goes westwards through Oanjam to the Eastern Ghats
and then South-westwards crosses the Sabari on the border
of the Sunkam and Bijji Taluks in the State of Bastar,
and thence runs along the range of Bela Dila to the
Indrivati ; it follows this river to its confluence with the
Godavari, and then runs through Cbanda cutting off the
southern part of that district and farther eastwards,
including the southern border of the district of Wun. It
then turns southwards to the Godavari, as its confluence
with tbe MSnjira, and thence farther south towards Bidar,
when Telugu meets with Kanarese. The frontier line
between the two forms of speech then runs almost due
south through the dominions of the Nizam. The Telugu



PRE-ARYAN AND PRE-DIIAVIDIAN IN INDIA 77

country further occupies the north-eastern edge of Bellary,
the greater eastern part of Anantapur, and the eastern
corner of Mysore. Through North Arcot and Chingleput
the border line thence runs back to the sea/' If the
Telugu country has such an extension, one understands
why TSrSnatha (p. 264) designates Kalinga as merely
a part of Trilinga. But on the other hand it is astonishing
that the name of the country is not met with till a late
period, only after the year 1000 A.D. By a singular
anomaly Ptolemy is the only guarantee of the name for all
the earlier period. He records the city of Trilingon, the
royal residence which he places in the trans-Gangetic
India (VII, 2, 23), at 154 East x 18 North; the city
is also called Triglypton (var. Triglyphori) ; in the region
where it is situated, " it is said, adds Ptolemy, that the
cocks are bearded, the crows and the parrots are white."
If the white parrots refer to the cockatoos, which is very
probable, the indication can only poiut to the further
regions of the Far East, as " the cockatoos are confined
to the Australian region, to the Philippine?, and Sulti;
the cockatoo galerita which is completely white is peculiar
to Australia and Tasmania." (Cambridge Natural History,
Vol. IX, Birds, p, 372.) The white crows lead in
another direction altogether ; if they refer to the species
called Dendrocitta leucogastra, which " has the top of
the head, the neck, the breastbone, the abdomen and the
covering of the tail white, the species belongs to south
India, particularly to Malabar (Fauna of British India,
Birds, I, p. 81). We would be thus brou
India and to the very borders of th
However, the place assigned to
of Ptolemy is very far from
modern Arakan, in the interior of




78 PRE-ARYAN AND PRE-DRAVIDIAN

of Akyab. The name would not be unexpected there,
because it is still preserved in that region under the
form of Talaing. It is known that the Burmese
designated under this name the Mon race which had
preceded them in Pegu and disseminated there a
civilisation, tributary to India. According to Sir
Arthur Phayre, it can be generally admitted that
Talaing = Telinga : Forchhamer has proposed to replace
this interpretation by another explanation drawn from the
Mon language where talaing signifies " trampled over by
feet ; " the derogatory term might have replaced the proper
ethnical name of the Mons after their defeat (cf. Hobson-
Jobson, s. v. Talaing for the texts and the references).
Phayre himself notes that though Kalinga figures in the
Peguan annals, " the word Telingana is never met with
there." The case is therefore exactly parallel to that of
India ; we have before us a name of very ancient aspect,
which the literature has ignored for a long time. It is
possible, even probable, that the literary usage has preferred
to maintain the old denomination of Andhra, applied by
Brahmanism since the Vedic times (Aitareya BrShmana),
and consecrated by its mere antiquity, rather than to
employ a vocable of uncertain form. The other name
given to Trilingon in Ptolemy, Triglypton or Triglyphon,
appears to be an attempt at interpretation, conforming
. to that which the medieval usage in India had already
furnished.. The term is composed of lre = Sk. tri " three"
+glypton or glyphon, both of which has the meaning
of " chiselled and engraved," the " triglyph " (triglypho*
or triglyphon ; its gender is undetermined) ie a term in
architecture which designates a feature of the frieze in
the Doric entablature; the triglyph is composed of the
parallel grooves grouped by threes, with the "drops"



PRE-ARYAN AND PRE-DRAVIDIAN IN INDIA 79

below, represented by the tips of cones, which symbolise
drops of water flowing from the roof through the grooves
and resting in suspense. Nothing could better recall
to a Greek, by a familiar image, the stone linga decorated
with vertical grooves by which the water of sacred asper-
sions drop down. Ptolemy's informant had picked up
an interpretation which is known even to-day; one
continues to explain Tilifaga, etc., still by Triliiiga and
T riling a would be the country of three Lingas, divine
manifestations of Siva on the three mountains which
mark the frontier of the Telugu country, Kfibfivara,
Srldaila and Bhlmetivara. Kslesvara is situated on the
Kistna, at the entrance of the pass by which it flows
into the plain ; SrlSaila is at the confluence of the
Wainganga with the Godavari in the district of Thanda ;
BhfmeSvara is in the Western Ghats, at the point where
the Telugu country touches the Maratha country and
Mysore. In Pliny also we have another evidence of the
interpretation Tilinga-Trilinga (Pliny VI, 18 Insula in
Gange est Magnoe amplitudinix geniem continem nnam
nomine Modogalingam), if one admits with Campbell
(Grammar of the Teloogoo, Introd.) that Modogalinga
must be analysed as Modoga + linga; Modoga would
represent the Telugu mufyiga, poetical form of the word
mufiu "three." But Caldwell (Compar. Grammar, Introd.,
p. 32) contests this explanation: the use of Mvdvga
would be pedantic, according to him ; the only analysis
which he would accept is Modo = M&du=3, galinga=
Kalmga, i. e. the three Kalingas, the Trikalinga of so
many epigraphic documents of the middle age.

We have indicated that the position assigned by
Ptolemy to " the royal residence of Trilingon," in modern
Arakau is not impossible, but we have had already



80 PRE-ARYAN AND PRE-DRAVIDIAN

occasion to explain, as regards Tosalei-Tosall mentioned
in the same list, VII, 2, 23, that Ptolemy had carried
by error to the East of the mouths of the Ganges an
itinerary really directed towards the South-West of the
delta. The question must remain open pending further
discoveries.

One is tempted to class side by side with the peoples
of Kalinga and Tiliftga the people of Bhulinga who are
known to us from numerous sources. Pliny, VI, 20
names the Bolingae amongst the series of peoples who
succeed one another up the course of the Indus. Ptolemy,
VII, 1, 69, places the Bolingai to the east of the mountain
Ouindios (Vindhya) with the cities of Stagabaza or
Bastagaza and of Bardaotis, on the right bank of the
Sda, i.e t> the Sona (S6n). The Ganapatha annexed to the
grammar of Pftnini names on different occasions, the
Bhaulingi: on II, 4, 59; IV, 1, 41; IV, 1, 173;
the rule enunciated in the last sutra is applied to the
constituting elements of the tribe of the SSlvas, and
consequently appears in the traditional verse, collected
by the Kadikft and the Candravrtti (on Candra, II, 4,
103) which enumerates the six sections of the Salvas :

Udumbaras Tilakhall Madrakara Yugandharah
BMingah Saradandaa ca Salvavayava tarhjfiitah.

The Sftlvas are well known (cf. Pragiter, Mart. P.
349) ; they inhabited the vicinity of the Kuru and the
Trigarta, at the western foot of the Aravalli. And,
consequently, in the RftmSyana G II, 70, 15 the messen-
gers, sent by Vasistha to recall Bharata back from the
Kekaya country where he was the guest of his maternal
uncle, had to traverse at first the long road which went
from Ayodhyfi towards Kurukgetra and the SarasvatI;



FEE-ARYAN AND PRE-DRAVIDIAN IN INDIA 81

they crossed the sacred river, they next passed the river
Saradanda, and " then entered into the town of Bhuliftga."
The Bengali recension shows here again its superiority
over the two others; the Bombay recension, and the
Southern recension, II, 68, 16, give the city the name
of Kulinga. The name of Kulingfi re-appears this time
in the feminine, in the two recensions of Bombay and
of the South, II, 71, 6, when the poet describes
the itinerary of Bharata returning from Kekaya
to Ayodhyft ; it is there the name of a river which
waters the Doab between the Ganges and the Yamuna",
The Bengal recension has here an altogether different
text. The Mahft Bhfirata does not mention Bhuliriga
as an ethnic name ; the word appears there to desig-
nate a bird which lives on the other side of the
Himalayas and of which the cry " ma sahasam " warns
men to move without precipitation, II, 44, 1545. But
the edition of the South, II, 67, 28, writes the name
of this bird as Kulinga, The Bhulifiga birds are again
mentioned in the great epic, XII, 169, 6826, as " the
birds of the sea, sons of the mountains" (samudrah
parvatodbhavah). In the corresponding passage, the edition
of the South (XII, 168, 9)', substitutes for the bhuliriga
the bhrunda birds.

4. Utkala-Mekala. The two names are connected
together as intimately as Afiga and Vatiga. The
Rftmayana which mentions them only once IV, 41, 9 B. ;
41, 14 G,, refers to them together : Uekalan UttalarM-
caiva, by the side of Kalitiga ; Kgemeudr*, in his resum
(Ram. maftj., IV, 284) combines them still more
intimately ; Mekalolkali'kah. The Mahft Bhftrata does the
same, VIII, 22, 88? : Mekalotlwlah Kaliiigah ; elsewhere
it juxtaposes them : VI, 9, 348, Mtkalab colkalaih saha j

u



82 PRE-ARYAN AND PRE-DRAVIDIAN

VII, 4, 12J Utkala Meltala. Mr. Pargiter in his
translation of the Markandeya P. (p. 327) has a note on
these two peoples which we will reproduce in full because
of the way it seems to anticipate the conclusions which I am
endeavouring to disentangle from this examination of the
facts. " The Utkalas were well-known (though not, often
mentioned in the M.-Bh.) and were a rude tribe of very
early origin, for they do not appear to have had any close
affinities with the races around them, and the Hari-vamsa
throws their origin back to the fabulous time of Ila (X,
631-2). Their territory reached on the east the R. Kapisfi
(Raghu- 7. IV, 38.... and on the west they touched the
Mekalas, for the two peoples are coupled together in the
M. Bh. and the R&mayana and the Mekalas were the
inhabitants of the Mekala hills, i.e., the hills bounding
Chhattisgarh on the west and north. Northward
dwelt the Punrjras and southward the Kalingas. Hence
Utkala comprised the southern portion of Chuta Nagpur,
the northern tributary states of Oripsa and the Balasore
district. Various derivations have been suggested of the
name Utkala but I would only draw attention to some of
the above passages where Utkala and Mekala are placed
together as if their names possessed something in common."
Utkala is still to-day one of the usual designations of
Orissa and the language of Orissa is called at will OfiyS,
Jrl or even Utkali (M*g.-8*rv. 9 F, Part II, p. 367). As
the Survey states clearly : " the Orissa country is not
confined to the division which now bears that name. It
includes a portion of the district of Midnapur in the north.
...OriyS is also the language of most of the district of
Singhbhum, belonging to the division of Chota Nagpur and
of several neighbouring native states which fall politically
within the same division. On the west it is the language



PRE-AHYAN AND PR&DRAVIDIAN IN INDIA 83

of the greater parts of the district of Sambalpur and of a
small portion of the district of Raipur in the Central
Provinces and also of the number of Native states which
lie between these districts and Orissa proper/'

If the name of Utkala has preserved its vitality, the
name of Mekala has survived only as a memory associated
with religion. The heights of Mekala give birth to one
of the most important rivers of India, the Narmada,
which is second only to the Ganges in point of sanctity.
One of her sacred appellations, recorded by Amara and
other lexicographers, is Mekalakanyaka, " the daughter of
the Mekala." But here also the name having no
definitely established form has been attracted by the
analogy of the common word " Mekhala " " girdle " and
the written form oscillates between the two. The com-
mentator on Amara, Sarvananda authorises both (ad Am.
I, 10, 31 . Mekalacalaprabhavatvad Mekala kanyaka
Mekhalakanyaketi kecit. Yan Mekhalad bhavati Mekhala-
bailaputri Hi khakaravan). The obscurity of the name
of Mekala is painfully evident in the edition of the
Maha Bharata published at Calcutta, in which the name
is printed several times as Melaka, under the influence
of the common word meld " fair." The editor can allege
for his justification a distant precedent ; the translator
of the Saddharma-smrtyupasthSna Sutra had already
substituted Melaka for Mekala in his Tibetan version, in
which he had reproduced too faithfully the fault commit-
ted by the scribe of the Sanskrit original ; the author
of the Chinese version had read and transcribed Mekhala
with aspiration (Pour I'Htstoire du Ram., p. 7). In
another passage of the same Sutra which mentions, in
imitation of the Ramftyana, G. IV, 40, 20 " the river Soba
born from the Mekala" (Mekalaprabbavam Sonar*) t



84 PBE-ARYAN AND PRE-DRAVIDIAN

the Tibetan version writes Megalati, the Chinese version
Mecaka (ibid, p. 18). The official geography of British
ladia has collected and saved this ancient name ; under
the name of Maikal Range it designates the chain of
mountains which starts from the sources of the NarmadS
(Amarakantak) and extends towards the South-South-
West up to the district of Balaghat.

The country of Utkala bears still another name from
which the modern appellation of Orissa is derived. Orissa
is Ovjradefia, "the country of Odra." The supposed
Sanskrit original given by Mr. Crooke in the second edition
of Hobson-JobsoH) Odrastra (sic) is monstrous and fantastic,
The first forms used by Europeans, Ulixa, Udeza, Orisa,
etc., are derived directly from 0 St., 16, 8o2,
Pre&navyakarana ; 397, Prajfiapana) and their women in
the conventional lint of royal slaves (ibid, 313, 380, 412).
The Rftma^ana, 4, 43, 1 IB; 44, 12 G, locates them in the



PRE-ARYAN AND PRE-DRAVIDIAN IN INDIA 89

north of India between the Matsya (Mleccha B.)and
the Surasena that is to say between Alvar and Mathura.
The Maha Bharata also considers them Mleccha ; it is
they who are to reign on the earth in the Kali age, with
the Andhra, the Saka, the Yavana and the Kamboja,
III, 188, 12839 ; they are met with again in the same
company, XIII, 83, 2104, among the races fallen to the
rank of Sudra (Frsala) because they did not see B rah mane.
They appear frequently in the great epic, but always
in bad company: Paundra, Yavana, Kirata, Clna and
other Mlecchas, 1, 175, 6685 ; Dravida, Andhra, and other
Mlecchas, V, 160, 5510; Dafiarna, Mekala, Utkala,
VI, 9, 84.7. The condition of sinners in hell is like that
of the Pulinda, XII, 151, 5620; the sinners will be
reborn in the south in the families of Andhra, Pulinda,
Sahara, XII, 207, 7559. The Pulinda derive their
origin from the foam thrown off by the cow of Vasistha,

I, 175, 6685. Bhlma, the conqueror of the east, turns
towards the south, finds them on his way when marching
on Cedi, II, 29, 1068, and reduces their city (nngard).
Sahadeva, who has just reduced eastern Kosala, meets
with them, II, 31, 1120 before fighting with Panrjya,
Kiskindhyft, and Mahismatl. In the legend of Udayana,
elaborated by the author of the Brhatkatha, the Pulindas
are the allies and the auxiliaries of the king of Kaus&mbi,
during his love-affair with V&savadatta (Katha 8. Sag.,

II, 12). Their kingdom is situated amidst the Vindhyas,
on the route which goes from KauSambi to Ujjayinl.
Their king adores the cruel Devi, offers her human
victims, and pillages the caravans (ibid, IV, 22). The
BfhatkathS-dlokasamgraha, always impregnated with a
picturesque realism, which it certainly owes to its model,
draws a striking picture of the Pulinda, VIII, 31 ; a



90 PRE-ARYAN AND PRE-DRAVIDIAN

group of young men leave for the chase ; one of them
tells the party : " I see before us an innumerable army
of these Pnlindas who haunt the caverns of the forest,
appearing like a forest of trunks blackened in the fire.
From their ranks come a fat little man small as a dwarf,
and with copper-coloured eyes : it was their chief Sirb-
haSatru ('Enemy of the lions'). He saluted the
commander-in-chief who asked him : ' How is the wife
of my brother ? And your two sons, Sfimbara (deer)
and SSranga (antilope) are they in good health ? '
Rumanvat ordered to hand over to Siihhasatru a bale
of stuffs dyed with indigo, curcuma, and safran, besides

a thousand jars of sesamum oil Then appeared before

us deer whose limbs flashed fire like diamond bubbles ;
in bands, they passed and repassed, as swift as the wind.

One asks the chief of the Pulindas: 'None of

us has ever seen such beasts ! If you know about them,
explain them to us ! ' ' Neither do I know about them/
said the Pulinda, ' but my father did. On a certain
occasion, he taught me something which I will tell you.
...He whose arrow, once let fly, makes a pradaksina around
these beasts to return forthwith to the quiver, know that
he is a cakravartin ' (transl, Lacote, 55). All the traits
here appear life-like. The Pulindas are compared to the
burnt trunks; the Natyafiastra, XXI, 89, in fact,
presents that the Pulinda should be represented witji a
black complexion. The chief is of the size of a dwarf
(nikharva) : s< The Pre-Dravidians differentiate them-
selves from the Dravidians by their short stature n
(Thurston, The Madras Presidency, p, 184). The sons
of the chief have the names of animals; totemism is
still spread amongst the savage tribes of the plateaus,
history of the arrows which return to the quiver



t>RE-ARYAN AND PRE-DRAVIDIAN IN INDIA 91

and the resplendent deer appear to come directly from
the folklore of the Mundas the Santals.

The name of the Pulindas is interpreted in Tibetan
(Mahavyut., 188, 15) by gyun po "out-east" and in
Chinese by lu-kia l " the race which kills the beasts for
their food." Ptolemy, VII, 1, 64-, also gives to the
Poulindai the epithet of agriophagoi y a rare term, which
appears to be invented to translate an Indian original ;

agrio signifies "wild," phagos "eater of ; 5> one

hesitates to make a choice between "who live on wild
fruits " and " who live on raw meat." He assigns to
them a well-defined place in the interior behind LarikS,
the country of the Lafca, which has for its principal
towns amongst others; Barygaza (Bharukaccha), Ozene
(Ujjayinl) and Nasika (Nasika), from Malva to the source
of the Godavarf. It is therefore in the hills of Satpura,
the Vindhya, and the Aravalli that they are located
by him. Still further away, in the interior, he says,
are the Khatriaoi of whom the cities are, partly to the
east and partly to the west of the Indus.

The Kulindas have not acquired the same celebrity
as the Pulindas. Their name is rarely met with after the
epic period. In the Maha Bharata, however, their rdle is
more considerable than that of the Pulindas. They live
"amidst the Himalaya in a country which abounds
in elephants and in horses, all mixed with the Kirlta,
the Tangana and also with the Pulinda (sic) in hundreds,
a country loved by the gods, full of innumerable
marvels, III, 140, 10866; their king SubSbu receives
with respect the Pamjavas when they set out to visit



*.



92 PRE-ARYAN AND PRE-DRAVIDIAN

the GandhamSdana ; when returning, they follow the
same way, pass through Clna, Tukhara, Darada " and
then they found the countries of the Kulinda that have
so many jewels, and after crossing the Himalayan region
there where walking is difficult, they saw the fortress
of the king Subahu," HI, 177, 12850, Arjuna setting
out to subdue the north begins with the conquest of
the Kulinda ; then he turns towards Anarta, Kalakuta,
Sakala, II, 26, 996, In the Rajasuya of Yudhisthira,
the Kulinda appear in the cortege of the northern peoples,
inhabitants of the banks of the river Sailoda: Khasa,
Pffrada, Tangana, etc., which bring as tribute the gold
of ant-hills, the chowries made of the tails of yaks and
the honey of the mountain, II, 52, 1859. Section 85
of the VIII book relates at length the struggle in which
the Kulindas were engaged, during the great battle,
mounted on their swift elephants, well-equipped, covered
with gold and born in the Himalaya. In the description
of the world, VI, 9, 370, the Kulindas are classed near
the P&rada and the Kunthaka; the country has its
Piemont, the Kulindopatyaka, VI, 9, 363, But once
again here, as we have already observed in other cases,
the surreptitious variants attest that the name is not
understood; the name of Kalinga, which is better
known, tends to supplant it ; VI, 9, 347 C has Kalinga ;
VII, 121, 4819 C has Kulinga and the translator of P. C.
Boy writes Kalinga. The Vyu P. 45, 116 writes
Kulinda in the list of the Udlcya "Northerners":
Gandhara Yavanas caiva, etc., the Matsya P., in the
corresponding verse 113, 41 substitutes Pulinda which is
better known, but wrongly placed here ; the Markandeya,
57, 37 substitutes Kalinga. We recall these mentions
of Kulinga which we have already seen alternating



PRE-ARYAN AND PRE-DRAVIDIAN IN INDIA 98

through confusion with the Bhulinga. Another variant
is worthy of notice. The Maha Bharata, XIII, 88,
2104, in the list of degraded tribes, already quoted,
writes : Kalindaa ca Pulindas ca ; P, G. Roy's translator
introduces here the Kalinga. But the sketch of Indian
geography given by VSgbhaba (Kavyanusamnain } and
copied by Hemacandra (Kavyanusasana, \i < ), enumerating
the northern mountains, beyond Prthudaka (Pehoa),

gives : Himalaya-Jaland/iara-Kalind-Endrakila

partatah', Hemacandra writes "Kalindrendra" probably
under the influence of the following syllable dra. It is
interesting to state that the form collected by Ptolemy
VII, 1, 42 Kulindri(ne)j presents the same alteration;
the analogy of the name of Indra, so popular and so
frequently used at the end of compounds, has not failed
to affect the final inda. As regards the vowel of the
initial syllable, the a can be original, as in Kalinga ; the
hypothesis is rendered probable by the sacred name of
the holy Yamuna, This is Kalindl (Amara, I, 10, 31)
"the daughter of the mountain Kalinda," as the Narmada
is " the daughter of the mountain Mekala " (Amara, ibid,
the same verse). Ptolemy VII, 1, 42 places the Kulin-
drine " below the sources of the Bibasis (Vipasa), the
Zaradros (Satadru), the Diamouna (Yamunft) and the
Ganges." The Yamuna is therefore exactly the daughter
of the mountains of the Kulinda country. The Greek
geography confirms the Indian evidences.

The BrhatsaihhitS of Varaha Mihira supplies another
variant of the same name, In the XIV chapter, the
editor H. Kern has twiee adopted the reading Kauyinda,
in preference to other readings Kaulinda and Kaulindra
furnished by equally good manuscripts. Undoubtedly
it refers to the Kulinda ; in verse 30, they appear in a



9* FEE-ARYAN AND PRE-DRAVIDIAN

catalogue of the peoples of the North-East, with Kirata
and Clna, two lines after AbhisSra-Darada-Tangaga-
Kuluta; in verse 33, the King Kauninda follows Avanta,
Anarta, Sindhu-Sauvira, H&rahaura and the sovereign
of Madra. In the same treatise, but in another chapter,
I V, 24, the printed text has Kaulinda, with the variant
Kaulindra, in a list which includes Traigarta, Malava,
Sibi, Ayodhyaka. The form Kuntnda is warranted, in
any case, by the legend of a large number of old coins,
found mostly in the districts of Saharanpur and Ambala
" the Piemont of the Kulinda." These pieces are either
silver or copper ; the work is quite varied, and they cover,
certainly, a fairly long period of time, beginning from
the second century before the Christian era. The design is
overloaded ; on the obverse a woman standing with her
left hand in her hip, offering with her right hand a fruit
to a stag (or a buffalo) standing and turned to its right,
bearing a symbol between its horns ; over its chine, a
kind of square railing crowned by a parasol; on the
reverse a caitya with three stories of arches surmounted
by the parasol ; on the right, a tree inside a railing ;
on the left, a Svastika and a symbol with triangular
head ; higher up, a nandipada ; below, a serpent. The
legend is most often written in two scripts, Brahmi and
Eharos(rl ; in Br&hml : Amagabhutisa Maharajasa raftah
*(rajna) Kunadasa ; the Kharostrl: Eana Kunidasa
Amaghabhatisa Maharajasa, The use of the two writings,
each of which characterises a world and a civilisation,
indicates well the importance of the region occupied by
the Kulinda-Kuninda ; it is also in their territory, or
very near them, in the valley of Kangra, that digraphic
inscriptions in Brahmi and Kharo^bl have been discovered
(Ep. I*d., VII, 116).



PRE-ARYAN AND PRE-DRAV1DIAN IN INDIA 95

Pulinda-Kulinda, Mekala-Utkala (with the group
U Kalindl and on the other



100 PRE-ARYAN AND PRE-DRAVlDIAtf

hand we have seen, more than once, the variant Kulinga
instead of Kalinga and the ethnique Bhulihga has also
an u. Experts in the Austro- Asiatic languages may
venture, risky as it often is in matters of local onomastic,
to propose etymologies for these names. Father Schmidt,
without thinking of it, has already suggested for Kalinga
an etymology which, if correct, would open a direction
for new researches. In the list of lexicographic corres-
pondences accompanying his article B. E. F. E, O., VII,
261, he writes under the rubric of words with initial
L (n. 151) : " Kalan [in Nicobarese], white-bellied sea-
eagle, Cuncuma lemogaster^ Khmer Khlen, Stieng Klin,
a kind of milan (Sanskrit Kalinga) : The Sanskrit
Kalinga can very well signify f the shrike with a forked
tail,' Am. 2,5,16; Hem., 1383. Bohtlingk and Roth
suggest (P. W. 1 , sub verbo) for this meaning, an analysis of
the word into Kalim+ga \ they had been, without know-
ing it, anticipated by Sarvananda, who comments, on
Am , 2, 5, 16 : kaliih gahanam gacchaltti Kalingah.
KsIrasvSmin, has also his own interpretation : ke lingam
cUdasya Iraliftgah ' it has on its head (to) its characteristic
sign, its tuft.'" These fancies are a useful reminder of the
shifting ground of etymology. If the correspondence indi-
cated by Father Schmidt be recognised as exact, one would be
led to suppose that the eponyms of regions that we have
studied were totems. But the word Kalihga cannot fail
to suggest a completely different connection with the
Tibetan word ^trt=Sk. dvipa. Tibetan, of course,
belongs neither to the Mumjfi family nor to the Austro-
Asiatic group; but it has so many features in common
with these languages that it oannot be completely kept
aside from them. The word glin, now-a-days pronounced
lib in the classical use of Lhasa, contains a prefix ifa;



AND PRE-DRAVIDIAftlN INDIA 101

Tibetan has transformed into sonants all the surd explo-
sives of the preformatives : g, d, b, for * (a), t (a), p ()>
and eliminated the vocalic element which sustained them.
The word glin fall back therefore on an older form ha-lib.
Its meaning is identical to Sanskrit dvlpa " island/ 1 with
all the secondary meanings deriving from it : " Isolated
territory, territorial division large or small/ 1 etc. The
simple word, without a preformative, can be met with
in Sikkim, amongst the Lepchas who are considered to
be the most ancient inhabitants of the country ; it is
lyan " the earth " in every sense of the word : " the
earth, territory," etc. In Tibetan, the word ff-lin is
combined with the affix -ka ; g-lin-ka means, according
to S. C. Das, "garden" ; "pleasure park"; the English-
Tibetan dictionary of Lama Dawasamdup Kazi renders,
in fact, the English word garden by ldum-ia\ glin-ka-,
cAod z'in. Jaschke, however, gives for glin-ka "a small
uncultivated river-island or low land/ 1 In Lepcha, too,
the word lyan is combined with the affix -ka to designate
" a space where there is no village/' It seems impossible
to isolate this word from the name of Lanka, which
designates in Sanskrit the island par excellence, the island
of the R&k*asas, where reigns the demon Ravarm, the
violent adversary of divine Rama. The vocabulary of
Yule and Burnell, Hobson-Jobson, *, v. Lunka adds "also
an island in general." The learned compilers have
probably borrowed this indication from the Telugu Dic-
tionary of Brown. They indicate still another meaning
of the word. " A kind of strong cheroot much prized
in the Madras Presidency and so called from being made
of tobacco grown in the ' islands f (the local term of which
is labka) of the Godavery delta." No reference at all ;
it is therefore from the real usage that they have gathered



102 PRE-ARYAN AND PftE-DRAVlDIAtf

this signification, which we had not the occasion to verify
on the place. The laiika cigar brings us back in an
unexpected manner to Kalinga with the islands of the
Godavari delta. And in fact the OazeUer of India, 1 s. v.
Godavari, fully confirms this evidence: "The land on
which tobacco is grown consists for the most part of
alluvial islands lying within the banks of the Godavari

river, called lankds, which are flooded every year

Tobacco seems to be grown on any part of the lankds

almost indifferently The tobacco of the lankds

would command a good price in European markets."
The word is therefore a current one. But some epigraphic
documents prove that this vocable was equally in use
more in the north, in the valley of the Mahanadi.
A deed of gift, which comes from the state of
Sonpur on the Mahanadi and published by B. C,
Mazumdar (Ep. Ind., XII, 237), is issued by a local
prince related by an unknown lien with the sovereign
of TrikaliAga, and which takes title of Pa'acimalankadhi-
pati; Mr. Mazumdar observes about this name that
" the peoples of Sonpur still know by tradition that the
state of Sonpur once bore some name as Paacimalanka."
Another document coming from the same state and
published by the same editor (Ep. Ind., XII, 2 18) is
granted Lankavarttakasamnidhau. The editor proposes
to identify Lanka varttaka with a high land to be found
in the bed of the Mahanadi and which is called Latikes-
varl. The two inscriptions are of very late epoch and
go back only three or four centuries earlier. One cannot
read in Pliny without surprise, the passage already cited,
VI, 18 : Insula in Gauge eat magna amplitudinis genlem
continent unam nomine Modogalingam. "There is an
island of large extent in the Ganges, which contains only



PRE-ARYAN AND PRE-DRAVIDIAN IN INDIA 108

one nation called Modogalinga." Whatever may be the
first element (cf. ante) the mention of the island in the
Ganges recalls inevitably glin (ga-ling) which signifies
island and these lanka which we find in the bed of
the Godavari and the Mahanadi up the Ganges,

The element lanka reappears in a certain number of
geographical names in the neighbourhood of the Malaya
peninsula, One can have no hesitation in recognising
it in the country of Kia-mo-lang-kia, Lang-kin, Lang-kia
slu, of the Chinese travellers and annalists, Lenkasuka
of Nagarakretagama, Ilangasogam of the Tamil inscription
of Rsjendracola I in Tanjore. Mr. Ferrand has collected
all the texts in an Appendix (III) of his article on Malaka
(J. A., 1918, II, 134, 145, and 153); he has discussed
there the proposed localisations and has located it with
much probability on the eastern coast of the Malaya
peninsula, right in the Isthmus of Ligor. The identity
Kia-mo>lang-kia=.Lang-kia-*hu, generally admitted, sup
poses that the word I auk a forms an organic element in
the whole name ; besides, Yi-tsing does not hesitate to use
alternatively, in course of the same passage, Lang-kia-*hu
and Lang-kia (Relig. Kmin., p. 57 and 100); the History
of the Leang dynasty also gives the two forms Lang-ya
and Lang-ya-sieou. The meaning of the final element;
su, suk, 8oga remains to be determined. As regards the
word Kia-mo-lang-kia, by which Hiuan-tsang designates
the same country (Mtm>, II, 82), Stanislas Julien has
constituted the Sanskrit original Kamalanka on the basis
of the type of Kamarupa in which also the first two
syllables are equally represented by the characters kia-Ko *



104 PRE-ARYAN AND PRE-DRAVIDIAN

and this restitution has been accepted without discussion.
However Mr. Ferrand (J. A., 1918. II, 145), who is
familiar with the Malayan languages, has presented a
sagacious observation of this name. " Kamalanka" he
says, " is a curious form of toponomastic. One cannot
help connecting the first two syllables with those of the
Sanskrit name of Assam, KSmarupa. It cannot be
doubted that in both cases the Sanskrit kawa ( love ' is
simply a pun recalling by assonance, more or less exactly,
the indigenous word." Judging by the native terms used
in Hiuan-tsang's transcription, it is probable that his
Indian informants had pronounced Kamalaiika. But
this name had certainly taken still another fonii in
Sanskrit, in the time of Hiuan-tsang. The Mafijusrlmu-
lakalpa, the Sanskrit original of which has been discovered
by the admirable Ganapati Sfistrl, and of which we have
also a Tibetan and an incomplete Chinese translation,
names the islands of Karmarafiga with the island of
cocoanuts and Varugaka (Baros, Sumatra) and the islands
of the naked (Nicobar), Bali and Java as the regions
where the language is indistinct, without clearness, rude,
and abounding in the letter r.

Karrarahgakhyadvipenn Nadikerasamudbhave \ dvtpe

Tarusake caiva Nagna Balitamudbkave \ Yavadvipe va

sattveQH tadanyadvipasamudbhara \ Faca rakarabahula tu

vacd asphutataih gala \ avyaltta nitiura caiva sakrodka-

pretayonisu.

Ed. Ganapati ; II p. 322.

I shall not stop to consider here all the names which
are so important for the study of Indian archipelago ;
I have already made them the subject of a communication



PRE-ARYAN AND PRE-DRAVIDIAN IN INDIA 105

to the Societe Asiatique (J. A., 1921, 1, 332) and which
I propose to publish later on. The Chinese translation
omits in. this passage the name of Karmaranga. The
Tibetan translation (ed. of Pekin, p. 197a) gives as its
equivalent las chon " action-colour ;" it is the literal
translation of the two words karma (action) and raiiga
(colour), which had been believed to be discovered on the
analysis of the name of the country. The form of the
word Karmaranga, with its two r, attests the frequency of
the sound r in this group of languages. The sound r has
been substituted for the I of lanka and also been introduced
in the first element. The word, however, is not absolutely
isolated. Bana, in the Hargaoarita, twice mentions the
shields of Karmaranga in the course of the seventh
chapter (edit. Nirnayasagar, p. 232); the warriors
who surrounded Harga wore as ornaments the leathern
bucklers of Karmaranga which were round and variegated
(kirmira-karmarangacarmaman$alamandan ; ibid, p. 243) ;
among the presents sent by the king of KSmarupa, there
were the leathern bucklers of Karmaranga ornamented
with designs drawn by the lustre of gold and with
beautiful borders : ruMrakaftcanapattrabhangabhauguranam
atibandhuraparivesanam Karmarangacarmanam sambharan.
The Kashmir edition has the reading karmaraiiga ; the
Bombay texts followed by Cowell in his translation,
pp. 203 and 214, read kardarafaga-, the epithet kirmlra
intentionally selected for the alliteration, would suffice
to warrant the reading karmaranga, made still more
suitable by the use of the word carma which follows it :
kirmirakarmarangacarma. The commentator Samkara,
in his Samketa comments on the first passage : kardar-
ahgakani kardarangadisodbhavani bahutuvarnasutra-
racitani carma^i. spkotatah

u



106 PRE-ARYAN AND PRE-DRAVIDIAN

Icardarangacarmani : "the leather coming from the
country of Kardarauga, fabricated with many golden
threads; the skins of Kardaranga have a lump of flesh
of glistening colour/' on the second passage he repeats
that Kardaranga is the name of the country of origin
of these bucklers : KardarangadeMhavanaih sphetakanam,
It is therefore necessary to reject the explanation given
by Bohtlingk (P W*, sul verbo)\ " hochroth (bright red);
cf. fymiraga" Besides Karmaranga, the Mafijusri-
mulakalpa mentions also the name of Carmaranga twice
in chap. 0, p. 206, and in chap. 22, p. 233 ; in both the
passages Carmaranga is mentioned along with Kalasa-
varapura (KalasahvS, p. 206; Kalasamukhya, p. 233)
Samata(a and Vanga. The last two countries lead us
to the delta of the Ganges Kalasapura (or Kalasavara-
pura, etc,) is a city of Suvarnadvfpa according to the
evidence of the KathasaritsSgara, 54, 108 (the Manjarl,
in the corresponding tale, XV, 207 ff., does not give the
name of the city), In the collection of Nepalese minia-
tures studied by M. Foucher (Etude sur I' Iconographie
bonddkique de PInde), the representation of "Bhagavat
at Kalasavarapura" (MS. A. 15, Calcutta, n 13) imme-
diately follows that of "Dlpankara in Yavadvlpa " (ibid',
n. 12). M. Pelliot has collected (B. E. F. E.-O, IV, 360)
several Chinese texts which mention this city and from
which it is evident that Kalasapura was situated in the
north of To-ho-lo, itself located in the north of Fan-pan
which lies on the Malaya peninsula, in the same line
as Bandon or Ligor. Carmaranga therefore brings us
to the same regions as Karmaranga and is perhaps only
a variant of the same name. It should be observed that
the two chapters of Maiijusrlmulakalpa in which the
name of Carmaranga is met with are wanting not only



PRE-ARYAN AND PRE-DRA VIDUN IN INDIA 107

in the Chinese version but also in the Tibetan version
of the work. The Brhatsamhita, XIV, 9, in its catalogue
of the peoples of the South-East (dgneyi) combines F?a-
Nalikera-Carmadvlpah ; Kern has translated (/. B. A. S. }
n. s,, V, 83) this as " The Island of Bulls, of Cocoas, of
Tree-barks/' but the mention of Nalikera by the side of
carma clearly proves that Carmadvlpa corresponds here
to Carma- or Karma-rangadvlpa of the Mafijusrlmula-
kalpa. The Brhatsamhita, in the same chapter, verse
23, mentions a people of Carmaranga again amongst
the populations of the farthest North-West, pele-mele
with the Sulika (Sogdians), the Ekavilocana (Monoph-
thalmes), Dlrghagrlva (long-necked), etc. This refers,
without doubt, to the same people that the Maha Bharata,
VI, 9, 355 calls Carmaman^ala and the PurSnas (Mark.,
57, 36; Vay., 45, 115) call Carmakhamjika. Mr.
Pargiter has connected this last denomination with the
name of Samarcand. In fact they are enumerated
between the Pahlava and the Gandhara, and thus
would be located on the confines of the Indian and Iranian
world.

The reputation of the skins of Karmarafiga appears
to explain Ptolemy's note on the population of the
" Brigands " Xqorat' which he locates exactly in the
surroundings of Karmaranga, on the southern shores of
the great gulf, i.e., the Gulf of Siam (VII, 2, 6 and 21) :
" It is said that the natives of the country of Brigands
live like beasts, inhabit the caverns, and that they have
skin almost like that of hippopotami impenetrable by
arrows." The region had some centres of population
and even a port of commerce : " Samara(n)de, Pagrasa,
PithdnobastS which is a market, Akadra, Zabai which
is the city." It can be supposed that Samara(n)dS is



ld8 PRE-ARYAN AND PRE-DRAVIDlAtf

an alteration of the name which has finally taken in
Sanskrit the alternate forms Carmaranga and Karmaranga
(of. infra the forms Camariz, Camarix).

India received not only some bucklers of Karmara&ga
hide ; she also received from this country a fruit which
has been acclimatized in India and which continues to
bear even to-day, but slightly transformed, the name of
its land of origin. Karmaranga is the Sanskrit designa-
tion of the fruit which the Europeans call carambole ;
Lushington (List of Vernacular Trees... in ihe Madras
Presidency, n 365) mentions also the names of the
Coromandel Gooseberry; Sweet Climbing; Square Tamarind.
According to the same authority the Uriya name is
koromohga (by metathesis), in Telugu koromonga and
tamarta, in Tamil sagadam, sis am, sigam^ kandasadgam,
tamarattai ; in Malay alam Murappuli, kamaran'gam,
pulinsi, tamaratta, in Canarese darehuli, karmaranga,
kirinelliy darepuli ; Khory and Katrak (Materia Medica,
II, 152) adds for Bengali kamaranga and kamarak,
for Oujrati kamarak^ for Hindi kamaranga and kamrakh*
According to the Hindi Sabda Sagara, Uamarakh is the
name of the tree ; the name of the fruit is (in Hindi)
kamaranga and karmaraiiga. Thus in the designation of
this fruit we find again the fluctuation attested by
Kamaranga (Kia-mo-lang-kia) of Hiuan-tsang against
Karmaranga of the Manjudrlmulakalpa. The glossary of
Yule and Burnell has an excellent and copious article on
this fruit, s.v. carambola. The name has been registered
by Linne who has classified the tree under the title of
Averrhoa, Carambola. The original habitat of the tree is
the islands of Moluccas (Lushington, loc. cit.) ; the
Kannarafigain this case could represent a stage in the trans-
plantation of the carambol tree ; this country, therefore,



t> RE-ARYAN AND PRE-DRAVlDIAN IN INDIA 109

would from very early times, have been a market of
exchange between India and the farthest islands of the
archipelago. l Its situation on the eastern coast of the
isthmus of Ligor marked it out for this role. The
Chinese texts collected by M. Ferrand confirm the fact ;
the most expressive of these texts appear to me to be
the biography of Paramftrtha translated by Takakusu
(B.E.F.E.O., IV. 62). When the Indian monk, tired of
his sojourn in China, wished to return to his country
(he was a native of UjjayinI), his biographer only says :
" he thought of embarking for Lang-kia-w"

The name of Karmaranga is mentioned in the famous
inscription of Rajendracola I at Tanjore although no one
has yet recognised it. In the list of countries conquered
by the Indian conqueror, after IlatigaSogam come
Mapappfijam, Mevilimbangam, Vajaippanduru, Tajaittak-
kolam, MSdamalingam. I have already dealt with
Talaittakkolam elsewhere in connection of Takkolaof
Ptolemy and the Pali texts ; I shall have to return to the
last name directly. " Mevilimbangam and Vajaippanduru
do not afford any identification at present/ 1 says M. Coedee



1 The name of carmaranga is given by Rajanigbagtu, 3, 123 as
a synonym? of avartikd which is the name for the senna Cassia
acutifolia (Lushington, List, 965 : African senna, Kordojan ., Nubian .,
Officinal s., Senaar s. t Surat 8.). As these designations indicate, the
plant is a native of Nubia and Abyssinia. The name Carmaranga
does not appear to point to the place of origin. Tbis name has not
yet been met with in either learned or ordinary usage except for the
reference made by tbe R&janighantu compiled at a very late period,
probably in tbe thirteenth century. Among other names of senna
collected in this compilation appears also ran'galata in which rang a is
not accompanied by carma. Lushington *B list does not mention any
of these names for Cassia aoutifolia.



110 PRE-ARYAN AND PRE-DRAVLDtAN

in his fine article on the kingdom of Srlvijaya (B. E. F.
E. 0., XVIII, 6, 15). M. Ferrand has only reproduced
the name without adding anything to it in his detailed
review of this work (/. A., 1919, II, 172). The Malaya
name of the carambol tree is balimbing or belimbing ; in
India this name is used for a variety of carambol tree,
the Averrhoa Bilimli of Linne, which yields a sweeter
fruit than karmaranga ; in Telugu : bilibili, bilumbi t
gommarekn, pnlmukaya ; in Tamil : koBsittamarattai)
pilimbi pulissakay, pulima j in Malayalam : bilimpi,
barissaklta, vilumpi ; in Canarese : bilimbi, bimbuli
(Lushington, List, 866 ; cf. also Yule-Burnell, s. v. blimbee.
The note furnished by Yule and Burnell, s. v. Carambola,
may be conveniently reproduced here ; " Sir J. Hooker
observes that the fact that there is an acid and a sweet-
fruited variety (blimbee) of this plant indicates a very old
cultivation.") But the evidence of Garcia de Orta, amongst
others, shows clearly that for a competent connoisseur
karmaraiiga and bilimbi are equivalent terms : " These
carambolas are called in Canar and the Decan camariz
and in Malaya balimba" (Yule-Burnell, s. v Carambola.)
And Linschoten (ibid) : " The fruit which the Malabarie
and the Portuguese call Carambola is called Camarix in the
Deccan, Camarix and Carabeli in Canara and Bolumba in
Malay/' Mevilimbangam should, therefore, be analysed,
in the inscription of Tanjore, like Ma-Damalingam, Ma*
Nakkavaram, as Me-7ilimbangam ; it is clear that Vilim-
bangam is the Indian transcription of Malaya belimbing
which is the equivalent of Karmaranga. The Indian
name of the fruit derived from the name of the
country, has become in its turn the indication of the
country itself ; Karmaranga has become the country of
the carambol tree, and as can be seen in the list



PRE-ARYAN AND PRE-DRAVIDIAN IN INDIA 111

of the names borrowed from Lushington, the Malaya
name of the fruit has made its way in all parts
of South India along with the name given by
Sanskrit. But then the identity of Kamalanka-L8nkasuka
is to be abandoned, because Lenkasuka is clearly mention-
ed in the Inscription of Tanjore under its Tamilised form
Ilangafogam. The two countries are certainly very near
each other, but they cannot be confounded and Map-
papalam is probably to be located between them, as it is
mentioned between the two names.

By the side of Lang-kia (shu) and symmetrical with
Ka(r)mal(r)aiik(g)a comes the name of Tam(f)aUnga>
with the alterance k- t- of which Kalinga-Tilinga, Kosala-
Tosala have given us incontestable examples. Mr. Coedes
has recognised the name in the inscription of Tanjore
where it appears under the form Ma-Damalihgam ; he
has also discovered the same name, this time under the
form of Tambralinga, in an inscription coming from Jaiya
and now preserved in Bangkok. I have pointed out
elsewhere that it is to be found in the Maha-Niddesa
under the form Tambalinga (Piolemfo le Maha-Niddesa
et le Brhatkatha, B. E. F, E. 0., Jubilee collection). M.
Coedes has also recognised the identity of this name
with the country of Tan- ma-ling mentioned amongst the
vassals of San-fo-t'si in the Tchu fan tche of Tcbao. Ju
kua (transl. Hirth and Rockhill, p. 62) and described
in a special note (p. 67 ff .) immediately before Ling-ya-sse
[kia], Lankasuka. Mr. Rockhill published later in
(T'oung Pao, 1915, p. 123) another note on Tan-ma-ling
taken from the Tao yi tche liao ; Tan-ma-ling, according to
this text, is adjacent to Sha-li Fu-lai-ngan; but the
position of this last place is also uncertain (cf. Blagden,
J. R. A, S,, 1913 ; p. 166). M. Coedes, after considering



112 PRE-ARYAN AND PRE-DRAVIDIAN

all the previous opinions concludes with probability
that the country of Tamalinga certainly covered Jaiya
and most probably Nagor Sri Dharmaraj. As regards the
name TSmbralinga written in the inscription of Jaiya,
M. Coedes observes : " Tdmbra is a Prakritic form of
tamra, copper, still used in Singhalese. The meaning of
the expression Tamralinga is not very clear. Taking
linga in the sense of mark or character, Tfimralinga could
signify (the country) which has copper as its character-
istic, but I do not believe that copper has ever been reported
in the north of the Malaya peninsula. It can be supposed
also that the country derived its name from a ' linga of
copper' which had some celebrity." Probably copper has
no more to do with this name than with most of the
other place-names where Sanskrit has tamra; we shall
explain this later on in connection with the name of
TSmralipti.

The alternation of tha preformatives kam-tam, which
we believe to have recognised in the names of Kam-fang=
Kamalanka, Karmaranga, Kamaranga, etc., and of Tarn*
Ung = Tfimralioga, etc., appears to be reproduced on the
very soil of India, in the region where we have already
noticed some ethnic couples differentiated only by the
preformatives. Kamarupa and Tamralipli seem to furnish
a new example.

The word hamarupa, considered from the standpoint
of Sanskrit, is a regular compound, in current use, of
which the meaning is perfectly clear : kama } desire +rupa
form. The association of the two words is not at all strange
as the Buddhist cosmology distinguishes in the universe
the world of Icama, Itamadhatu and the world of rupa,
rupadkatu. The use of the term of kamarupa in literature
constantly mdfcates the faculty of metamorphosing



PRE-ARYAN AND PRE-DRAVIDIANIN INDIA U3

at will. As regards the country of K&marupa,
though it is not mentioned in the epics, which include
it in PrSgjyptisa, KSlidasa names it (Raghuv., IV, 84) ;
in the time of Haifa Slladitya, it was a first class state
in Indian politics, and had relation with China. Owing
specially to Tantrikism Kamarupa has enjoyed a durable
popularity; this is a pltha, a sacred place, where
one of the relics of the Devi was adored. The Yoginltantra,
patala XI (quoted in Ep. Ind. 9 XII, 68) traces the frontier
of KSmarupa thus :

Nepalasya Kaftcanadrirn, Brahmaputrasya wmgamam
Karaloyam samarabhya yavad Dikaravasimm
nttarasyam Kanjagirih Karatoya tu pastime
tlrthasrestka Dikavnadl purvasyaifa Girikanyake
dafaine Brahmaputrasya Lakfiayah satiigamavadhi
Kamarupa iti k/iyatah sarva'saxtresu niscitah

t( From the mountain Kancana in Nepal up to the
confluence of the Brahmaputra, from the Karatoya to
Dikkara-VSsinl, the northern limit is the mount Kaiija ;
in the west the Karatoya, in the east the Dikgu (Oh,
daughter of the mountains), in the south the confluence
of the Laksa with the Brahmaputra ; this is the
territory which all the treatises call by the name of
Kamarupa."

The official nomenclature still continues to designate
the western part of Assam by the name of Kamrup. But,
in the religious sense, Kamarupa also includes Bhutan,
Kuch Behar and Bangpur. The temple of Kam&khyfi,
near Gauhati, is considered to be the mystical centre. It
is a mountainous region where one finds all the dialects of
India : of Aryan, Mumjft, Tibeto-Burman, and Mon-
families. At the time of Hiuan-tsang's visit

15



114 PRE-ARYAN AND PRE-DRAVIDIAN

King BhSskaravarman, was " a descendant of the God
NSrSyana " -, he was " of the caste of the Brahmanas,"
and had the title of " Kumara. " t( Sinoe the possession of
the kingdom by his family up to his time, the succession
of princes covers .a space of a thousand generations 19
(Mem. II, 77.) The evidence of his contemporary Bftna
(Haraacartta, chap. VII) confirms almost all these details.
Finally we possess since a few years ago an inscription
of King BhSskaravarman (Nidhanpur plates, JEp.
Ind., XII, 65), which takes back the genealogy up to
King Bhagadatta, the famous adversary of the Pfindavas,
by a long list of ancestors. However, when he had
business with others than Indians, the same prince boasted
of another origin altogether. When the envoy of the
T'ang dynasty, Li Yi-piao, paid him a visit during the
course of his mission (643-646) the king in a private
conversation, told him : " the royal family has handed
down its power for 4,000 years. The first was a holy
spirit which came from China (Han-ti) flying through the
air " (8he-kiafang tche, ed. Tok. XXXV, 1, 94b, col. ult.).
As though he would show sympathy for China, he asked
the envoy to get him a portrait of Lao-tseu and a Sanskrit
translation of the Tao-to-king. The Emperor, on his part,
wished to respond to this desire and promulgated an
edict asking the master of the Law, Hiuan-tsang, to
prepare the translation in collaboration with Taoist
teachers (cf. on this episode the article of M. Pelliot :
Autour d'une induction Sanscrtte du Tao4o king in Toung-
pao, Vol. XIII, 1912, p. 851 ff.). After the dynasty of
BhSskaravarman, Kftmarnpa never ceased to be subject
to barbarian races who have been gradually Hinduized.
The best known period is that of the Ahoms, of the
Tai or Shan rape, which succeeded in creating a kind of



PRE-ARYAN AND PRE-DRAVIDIAN IN INDIA 115

original civilisation and in maintaining their power from
the 13th to the 19th century.

The Brahmins have naturally invented a legend to
explain the name of K&marupa : it is there that Aama,
Love, sent by the gods to put an end to Siva's mourning
after the death of his wife, and to awaken in him again
the tender passion, was burnt by an angry look of Siva
and then recovered his original form (rupa). By the
side of this childish explanation, it will suffice to state
that the easternmost province of the kingdom of Assam,
on the very confines of Burma, bore the name of Namrup ;
Namrup was on the other side of the Dikhu, which marks
the religious limit of Kamarupa on the East, midway
between this river and the Upper Chindwin, in a hardly
accessible region which has always served as a shelter
for the vanquished. The climate there is deadly ; an
Assamese proverb, quoted by Gait (History of Assam, p.
144) says that if a bird flies over the country, the' bats
get back their lives, and if steel enters into the soil,
it will change into wax. A Brahman would not have
felt any difficulty in interpreting the name Namrup as
Namarupa, a compound so natural and so familiar that
it would seem to suggest itself spontaneously : Nama-
rupa, name and form, are the essential categories into
which existence is resolved. But the Philosophy of the
Upani$ads is out of place in this wild corner and the
Metamorphoses still more so in KSmarupa, We have
here " barbarous " names in which the same element noted
as rup and Sanskritised as rupa, is associated with pre-
formatives kam and nam. It would not, therefore, appear
strange if we discover the same element with the pre-
formative tarn formed with the prefix ka regularly
increased by nasal. It is this combination which



116 PRE-ARYAN AND PRE-DRAVlDIAN

appears to me to be at the root of the name of
Tamralipti.

Tgmralipti was for centuries the greatest port of the
Bay of Bengal. It is there that the missions exchanged
between A6oka and the King of Ceylon embarked and dis-
embarked (Mahavafivta XI, 38 ; XIX, 6). Fa-hien embarked
there for Ceylon ; Yi-tsing disembarked there when coming
from China and it is there also that he embarked for
Srlvijaya in Sumatra. The city, on account of its
importance, is frequently mentioned in stories ; the Maha-
bhSrata also mentions, very often, the city, the kingdom,
and the king, e.g. I, 186, 6993 ; II, 29, 1098 ; 51,
1874- VI, 9, 364; VII, 70, 2436; 119, 4716... The
Jaina Prajrlapana calls the city the capital of Vanga
in the list of the ariyas of the first order, the khettariya
which we have already quoted many times : Eayagila
Magaha Campa Aihga taAa Tamalitli Famga ya (Ind*
St., XVI, 397). The Dasakumara (story of Mitragupta)
makes it a city of Suhma. It commanded the entrance
and the exit of the river Ganges. To-day, it is a market-
town with some thousands of souls, called Tamluk, on
the Bupnarayan, not far from its confluence with the
Hughli. It is strange that the name of such a famous
city has never taken a definite form. The dictionary of
Hemcandra quotes (V. 979) four forms of the name :
Tamalipta, Damalipta, Tumalipti, Tamalint, and besides
two surnames : Stambapur and Visnugrha ; the Trikamja-
6ea (2, 1, 11) adds Tamalika. The Sabdakalpadruma
adds even another Tamoliptl. To all these designations
we must add another which is met with the most often,
Tamralipta (5, e). In almost all the passages, the manus-
cripts hesitate and the two forms tamra and tama are
alternately given in the same text. The Chinese pilgrims



RE-ARYAN AND PRE-DRAVIDIAN IN INDIA nf

transcribe To~mo-li4i 1 (Fa-hien), or Tan-mo-li-ti* (Hiuan-
tsang) and Tan-mo-li-li 8 (Yi-tsing), Ptolemy (VII, 1, 76)
writes Tamalites. MacCrindle (hd. Ant., XIII, 864)
has compared Tamalites with the name of Taluctoe men-
tioned by Pliny, VI, 18, on this side of India. The
name of TamraliptI seems to have been transported to
Cambaye in the course of the middle ages : the Paitca-
darvdachattraprabanclha (ed. Weber, 3) begins one of its
tales with : Stamlhatirthe Tamaliptyam Jayakarna-
bhupagrhc "in Cam bay, at TfimraliptI, in the palace of
the King Jayakarna;" and the SimhasanadvStrimsika
(Ind. Stud., XV, 252) begins with the history of a king
Tamraliptarsi who lives in Gujrat, between the Sabarmati
and the Mahi (Ourjarimandale Sabhravatl-Mahilanadyor
antare vanam vidyate tatva raja Tamraliptarqih). Weber
says in his note on this passage (Paftcad^ p. 71) that
Cambaye bears also the name of Tamravatt, or Trambavatl
from Guzerati tram&uih = Sk. tamra "copper" because
according to the legend, the city bad in ancient times
a wall of copper. Weber does not give the reference ; he
probably got his information from Biibler, who is quoted
in the notice on Cambaye. It is probable that the name
of Stambapur, given as synonym of Tamralipti by
Hemacandra, applied in reality to Cambaye, the Sanskrit
name of which is Stamlha-tirtha Prakrit Khambha. Fancy
has been given a free play on these numerous variants.
The Imperial Gazetteer of Tndia, 1 *. v. Tamluk, writes



/f If



Pg
Ik



118 PRK-ARYAN AND PRE-DRAVIDIAN

" The very name of the city points to its ancient un-
orthodox character ; but it has been so cleverly mani-
pulated that it has at last become a title of honour.
The Grammarians derive the word from tamo (tamas)-
lipta, ' tainted by obscurity or sin.' But a legend relates
that Vinu, under the form of Kalki, having got heated
in destroying the demons, let his sweat fall on the earth,
and the place which was fortunate enough to receive
this sacred perspiration derived its name from that and
became a holy place " (tama " fatigue "+lipta). All these
pretended etymologies show once more the uncertainty and
the embarrassment of the interpreters. TamalinI, Tamalika
are hopeless efforts to find a meaning out of this
mysterious groups of syllables ; one has tried to discover
there the name of the tree tamala, Xantkochymu* pictortu*
(Lushington, List, n. 178). The Jaina BhagavatI in the
second sataka tells the story of a Moriyapntta of Tamalitti
who called himself Tamali aod seems to have been
the object of a local cult. Copper, tamra, appears to
have discouraged the exegetists, who, however, could have
given to Tamralipti also, as to TamravatI=Cambaye, the
walls " coated with copper." One sees now how futile
it would be to attempt a literal interpretation of so many
geographical names of India with lamja, as the first
element. The list is infinite ; the most famous case is the
name of Ceylon, Tamraparni, Tambraparnt or Taproiane of
the Greeks. If copper has sometimes its ration d'etre there,
in most cases it only reveals an ancient prefix tarn placed
at the beginning of old denominations ; such is the case 'with
Tamralinga, Tamalinga, T9m(ba)l9ng, which we have
considered above.

If the first element of Tam(r)alipti can be thus
explained, it may not be impossible that the second lip,



PRE-ARYAN AND PRE-DRAV1DIAN IN INDIA 119

corresponds to ruf, becoming in Sanskrit rupa, from Eamrup
= Kamarupa. The kingdom of TSmralipti almost bordered
on Kftmarupa : To pass from one to the other Hiuan-
tsang had only to traverse the small kingdom of
Samata$a. The river which waters Tamluk, the Rupa-
Narfiyana seems to preserve in the first element of the
compound the pre-Aryan word whieh appeared also in
the name of the country. The correspondence i(lip) =
u(rup) is possible ; the modern name Tamluk seems to
have preserved the timbre of the real vowel, altered in
the Sanskrit adaptation, (The final k of Tamluk, sub-
stituted for p of rup, is normal in the present domain of
Tibeto-Burmans ; at the time of writing this, I receive
the Report of the Superintendent, Archaeological Survey,
Burma, 1923 ; on page 28, 1 find that the saint Upagupta
of Sanskrit Buddhism is venerated in Burma under the
name of Upagok). The name of Srlvijaya, now put
again in full light, shows constantly an identical altera-
tion ; the Chinese transcriptions always render the
Sanskrit syllable vi by the character which desig-
nates Buddha and of which the pronunciation bud is not
doubtful. The timbre u(o) is still preserved in the Arabic
transcription Sribuza, Serboza. On the opposite border
of the vast domain of Indian civilisation, the name of
Kapi&a (Capita guam diruit Cyrus in Pliny, VI. 92),
becomes in the Tibetan translation of the Maha-M5yQrI :
Jka. bu. sa. The identity Kapifa=Kabufia leads to the
question whether it can extend to Eamboja also. It is
surprising, in fact, to observe that according to the Chinese
documents, the rdle of KapisS seems to have been of
the first importance for India in politics and religion,
while its name, so to say, is hardly ever met with in Indian
literature. The Mahfi Bhflrata which is so well acquainted



120 PBE-ABYAN AND PRE-DBAVIDIAN

with the North-West of the Indian world, does not
mention Kapi5 even once, while the name of Kamboja
appears there almost incessantly. Kamboja and KapiSa
seem to be two attempts to render the same foreign word
in a language which did not lend itself to the purpose :

^-. = hm ; of P- each has a labial followed by a palatal,
kam b j

unvoiced in the first case and sonant in the second ; the
middle term seems to have been in the two cases a spirant :
/ and z both of which are wanting in Sanskrit. The
spirant appears in Greek also, in the sibilant of the proper
name Kambyses=Kafn) bujiya, the son of Cyrus, of whom
the name probably recalled one of the conquests of his
father, the destructor of Kapi&. In any case, the two
spirants seem to be quite clear in the title of KadpJiizez
claimed by the Kuehans, when their power had spread from
the district of Kuei-shnang to that of Kao-fn = Kambu in
Chinese transcription. The title of Kad phizes is symmetri-
cal with the title of Taxiles, under which the king of
Taxila-Takgasila whose personal name was Ambhi, is famous
in the history of Alexander. Both are tadrdja, according
to the terminology of Panini, IV, I, 174 . Kamboja
has even the honour of a special sutra, IV, 1, 175,
kambojal Ink : there is no Vrddhi for designating the
king ; he is Kamboja (and not Kamboja). It is an excep-
tion which Katyayana and Pataftjali extends to a series
of princes : Coda, Kw'era, Kerala and which Candragomin,
II, 4, 104, completes with Saka, all "margraves," chiefs
established on the borders of India. I do not want to
push this complicated question further here. I propose
to take it up again elsewhere in detail with a view to
drawing certain conclusions that bear on the history of
the Kushans in particular, I will quote here only one



PRE-ARYAN AND PRE-DRAVIDIAN IN INDIA ll\

case which is very characteristic of the alternatipn Kapisft-
Eamboja. The Ramayana, in the Kashmirian recension
(Weber, Bamayana, p. ?5 note), reads in the verse IV,
44, 23, Arattam Kapisam Battim. Ksemendra, in his
Manjarl Civ. 252), writes Aratta Balhi Kamboja?. Between
the i of Kapisa and the o of Kamboja, the intermediary
appears to be a soft u, analogous to the French , which
Sanskrit does not possess and which is supposed by the
alternation of (Sri) vi (jay a) and of (She-li)-fo (**bit*)
(she) in the Indonesian domain. In precisely the same
domain, we find a name analogous to but not identical witfy
Kamboja ; it is Kam-vuja, the present Cambodge, which the
Singhalese and the Tibetans have equally made Kamboja,
but which never appears with an o in the epjgraphic texts
written on the spot. Kam-vuja seems to make a counter-
part of rl-vija(ya), and specially of the form transcribed
by 8an-fo-tfi (=* 8am-bu* -jay) in Chinese and noted as
semloja in Javanese (for different forms of this name
ef. Ferrand, J. A., 1919, II, 158). Schlegel (Toung-pao>
II, 176) had already supposed the relation between

V V

Kemboja and 8emboja t in which he found the Malay name
of* the Plwneria acutifolia ; but it is a plant of recent
importation, a native of central America and now spread
over India and Indonesia (Liuhington, List, .1825; Pagoda
tree : Jasmine spurge ; Spanish American Jasmine) ; the
names which it has received in Uriya (gosampigc), aqd
in Tulu (gosampige ; sampai) are manifestly related with
the Malay name. If the relation between Kamboja and
Samboja be exact, it will be necessary to go back still
further to explain it. It may be then questioned whether
the name at the bottom is not that of the Bhoja, which it
is true, has an aspirate initial ; but the analogy of
(he wprcjs bhojana, thoga ^nd others derived from the

16



12& PRE-ARYAN AND PRE*DRAVID1AN

root bhuj in Sanskrit was strong enough to introduce this
alteration. The Bhojas have been admitted to an elevated
rank in Brahmanical society, on account of their alliance
with K?sna ; the MahS BhSrata in which they often appear
generally associates them with the Vrsui and the Andhaka,
the two principal clans of the party of Krsna. But Asoka
classes them still amongst the frontier countries (Edict,
XIII) on the same footing with the Yavana and the
Kamboja. The geography of the Puranas (Mark., 57,
58; Payu, 45, 182; Matsya ; 118, 52) cesses them
amongst the populations of the Vinc^^s where
they are found with the Kosala, T nnnft Utkala,

He CODQ

etc. 1 In modern times the name Bhoj.^ v ^.cae more
familiar than that of Kamboja and reacted on the



1 By a singular chance, the name of Kamboja and that of
Cambodge present , independently and at a very long interval, the
same alteration of the initial. Ptolemy, VI, 11, 6 when describing
Bactriane, places to the south of the Oxus the Tambyzoi by the side of
the Tokharoi; it is not at all doubtful that Tambyzoi is here the
equivalent of Kamboja as Tokharoi of Tukh&ra.

On the other hand the name of Cambodge has undergone in China
a final transformation or rather deformation at the beginning of the
17th century; the name (of Kan-po-tche) becomes there Tong-pu-
'tch'ai (Pelliot, B.E.F.E.O., II, 126). Prof. Pelliot proposes to con-
sider the character which is read as tong as faulty and to substitute for

the character & whichi as Men. It is not perhaps

a mere chance ; the initial ft, becoming palatalised has in both
the cases changed into a dental.

To note further that one of the mouths of the Ganges, the western-
POit, bears in Ptolemy, VII, 1. 18 & 80 the name. of JTamZwon,



PRE-ARYAN AND PRE-DRAV1DIAN IN INDIA 123

latter. The last dynasty of Vijaynagar, in the 16th
century, regularly couples the two names : its princes pride
themselves of being honoured by KaThbhoja-Bhoja-Kalinga,
etc. See for example Up. Ind. XI, 884 : XII, 848 ;
XIII, 229, XVI, 318. The name of Kamboja is always
written there as Kaihbhoja in imitation of Bhoja. The
pandits of Vijayanagar could have produced in support of
their spelling an almost sacred authority ; the grammarian
Yaska, the predecessor of Panini, in the famous passage of
the Nirukta, II, 1, 4, where he mentions the use of the verb
sav in the sense of " going " amongst the Kamboja adds
a curions interpretation of the name : mvatir galikarma
Kambojegv eva bha$yate Kambojah Kambalabhojah kamanlya-
bhoja va kambalah kamaniyo bhavati , . . " The Kamboja, these
are " Bhojas with the mantle of wool (kambala)" or " the
Kamboja as kamanlya (desirable) " ; the desirable (for
them), is a mantle of wool (on account of the cold,
adds a gloss)." Yaska gives this etymology in order to
show clearly the position of the Kambojas in relation to
the Aryas, whose linguistic usage he opposes to that of
the Kambojas in the following sentence ; the Karft-boja,
as they are a branch of the Bhoja, do not form a part
of the Arya. Thus, at an ancient epoch close to what
is called the Vedic period, an erudite and sagacious
grammarian analyses the name of the Kambojas, as we
propose to do, into kam + bhoja and he frankly admits
his difficulty in explaining by Sanskrit the element Knm
placed at the head of the name.

After this long digression, my attempt to discover
in Tamalipti a correspondance with the name Kamarupa
may perhaps not seem so daring. Besides, I do not pretend
to put forward anything except a hypothesis for discussion.
Whatever may be the final issue, it would not affect the



PRE-ARYAN AND PRE-DRAVlDIAN

conclusions of our work. These pairs of ethnics, identical
in form, differentiated only by their initial, apply to
regions placed by couples in close juxtaposition. The
process of differentiation by an initial pref ormative is
foreign to the two groups of languages, Aryan and
Dravidian, which have created the civilisation of historic
India. It is characteristic of a family of languages,
which is even to-day spread over a vast domain, from
the Himalayas to the Island of Palks, and which still
maintain themselves in the hills in the interior of India.
The later representatives of this ancient race are the
unconscious inheritors of a civilisation which had its
greatness. It had created in India veritable political units
of a considerable extent, so strongly linked with the real
life of the country that they have continued to exist
across the millenniums up to our days. The parallel
existence of these twin ethnics, as it were, sometimes they
are triplets, throws a curious light on the political and
social constitution of this civilisation. Mr. James Hornell,
in a brilliant essay on the origin and the ethnological
signification of the Indian boats (Memoirs of the Asiatic
Society of Bengal, Vol. VII, n 8, 1920) has been led
by his researches of a technical character to conclusions
which he had to accept he says, though he was not
prepared for them, and these conclusions are not without
analogy with ours. He admits a strong Polynesian
influence on the Pre-Dravidian population of the Southern
coast of India : a wave of Malayan immigration must
have arrived later, after the entrance of the Dravidians
on the scene, and it was they who brought from the
Malaya Archipelago . the cultivation of the cocoa-tree.
I still recall, as a sort of excuse for my boldness, that
the erudite and inventive A. Weber wae hot afraid to



PRE-A&YAN AND PRE-DRAVIDIAN IN INDIA 125

point out the possible relation between an unexplained
word, which remains still inexplicable, t&buuam, which
he had met with in the Veda of Magic (Atharvav. ; V, 18)
in a formula against the poison of serpents and the
Australo-Polynesian word tabou (Sitzb. Ak. Wis8. Berlin,
1876, XXXI, p. 684) ; he thought of a late borrowing,
coming from India by way of the Indian colonies of
Indonesia, fiarth, when criticising this Memoir in the
Bulletin des Religions de VInde (QSuvres, II, p. 54),
reasonably observed : " It attaches a very great fortune
to three obscure syllables lost in a corner of the Atharva-
veda " ; he added nevertheless : " If there be a relation
between the two word?, it would be rather, as it seems to
me, in the inverse sense," There is occasion now to take
the question up again and to pursue a methodical research
.to disclose the influences which this past had exercised on
the development of Indian civilisation. Prof. Przyluaki has
already boldly opened the way for the vocabulary (Mew.
Soc. Ling., XXII, 205, Bull. Soc. Ling., XXIV, 118 ft.).
We must know whether the legends, the religion and
the philosophical thought of India do not owe anything
to this past. India has been too exclusively examined
from the Indo-European standpoint. It ought to be
remembered that India is a great maritime country, open
to a vast sea forming so exactly its Mediterranean, a
Mediterranean of proportionate dimensions which for a
long time was believed to be closed in the south.
The movement which carried the Indian colonisation
towards the Far East, probably about the beginning of the
Christian Era, was far from inaugurating a new route,
as Columbus did in navigating towards the West.
Adventurers, traffickers and missionaries, profited by the
technical progress of navigation and followed, under the best



1*6 PRE-ARYAN AND PRE-DRAVIDIAN

conditions of comfort and efficiency, the way traced
from time immemorial, by the mariners of another race,
whom the Aryan or Aryanised India despised as savages.
The daring and skill of these men she was unable to
appreciate before and she continued to ignore all that
she owed to them.



APPENDIX

TO
PART I.

FURTHER NOTES ON NON-ARYAN LOANS
IN INDO-ARYAN

BY

JEAN PRZYLUSKI.



I

Non-Aryan Loans in Indo-Aryan

SANSKRIT Matanga l

The elephant is capable of taking objects by its trunk.
This is why in some languages its name is derived from
the word signifying " hand." In Sanskrit haslin, &arin,
" elephant " means an animal which has a hand (hatta-,
Tiara-). In a part of the Indonesian group, liman is a
name of elephant derived from lima " hand/'

In the A ustro- Asiatic languages the words which
mean ' hand ' fall under two principal types : with or
without the final consonant.

(a) With the final consonant ; dialects of the Malay
peninsula : tong, tung, tak (cf. Khmer dang \iang} ; Stieng
tong " the wrist, handle ").

(6) Without the final consonant : Mon toa \iai] ;
Khmer day [tat] ; Annamite lay ; Palaung lai ; San tali
tiy etc.

It is probable that all these forms originated from
an ancient tan (Skeat and Blagden had already proposed
tail with reservation) as in the Austro- Asiatic languages
a final consonant often disappears and yields place
to a vocalic element i and y. It is thus that one



1 Jean Przylaski, Bulletin de la SociJte' de Linguistique,
(79), pp. 98-108.

17



180 PRE- ARYAN AND PRE-DRAVID1AN

may explain the forms like Malay and Batak tafia* ;
Malagashy tkfia* ; Cam tafii* ; Jarai tSfia*, all of which
mean " hand."

Some of the dialects of the Malay peninsula for
designating ' elephant ' has a word with the element
tan, ton : tanyal, maintonp.

If some of the details of these modern forms are very
obscure, the case is different with Indo-Aryan matanga
" elephant." One can discover there, at the first sight, the
element tail preceded by the prefix ma. Matanga-, which
is thus the name of an animal, is also used sometimes
to mean an aboriginal tribe of India, as it very often
happens (cf. Oldenberg, La religion du 7da, trans.
Henry, p. 71,72).

The existence of a prefix ma- in the Austro- Asiatic
languages is proved specially by the following example :
it is found in San tali maran which A. Campbell translates
as "great, large, big, huge, to become or cause to
become great, large, big, huge ; first-born, principal,
bead, chief " ; martin is derived from a root ran, Ian and
a prefix ma- as is indicated by the words which mean
" great " in languages of the same family : Cam, praun,

Jara\* profi, gloft. In modern Annamite Ion " great "
has only preserved the root. But the middle Annamite
mid* still preserved, in the 17th century, the trace of
the element ma.

In Khasi a particle of emphasis ma- is placed before
the pronouns. " Ma- prefixed emphasises the pronoun ;
nga la ong, ma - ngal said, even I " (Linguistic Survey
of India, II, p. 9). It even seems that this partible of
emphasis used in polite forms before the pronoun of the
second person exists in a contracted form along with jt.



ARYAN LOANS IN INDO-ARYAN idl

*

In face of Bahnar e, th, Kaeeng and Halang at "you,
thou," there is in Kon-tu and Sue mat, and in Ann ami te

may or way which have the same meaning, In Stieng
two different forms either with or without m, exist side
by side for specific use ; ei is used when speaking to a
woman and mei when addressing a man.

SANSKRIT mayura-, mayiika-, maruka-.
A name of peacock may&ra- already appears in the
Rgveda, Prof. Ju'es Bloch has recently compared it with
Dravidian, Munda and Indo-Chinese forms (B S,L. XXV,
p. 16). The principal Austro- Asiatic, forma are the
following :

Santali marak'

Savara mara

w

Cam amrak

Malay mera

w

Crau brak

Stieng Irak

Mon mra

Most of these words present a final guttural and the
ancient form seems to have been marak preserved almost
intact in Santali where V is the notation of implosive.
Tn marak the prefix ma can bo isolated from the root rak
which imitates the cry of this bird. In fact in the Santali
dictionary of A. Campbell the following word is found.

raV " to weep, to beseech, the call, cry or note of
a beast, bird or insect/'

Amongst the examples given under this word we have :
marak' rak' " peacock-crow which is earlier than cock-



182 PRfl-ARYAN AND PRE-DRAVIDIAN

The same syllable rak explains Malay sorak, Khmer
srek and Jarai Itraih, all of which mean " to cry.' 1 The
peacock, of which the shrill cry is heard before that of
the cock is the bird that cries rak. In Sanskrit a name

of peacock is maruka which is formed after the Austro-
Asiatic word, if one accepts the equivalence 0=w of which
I have already given examples in , a previous article
(B. 8. L. XXIV, p. 120).

v

Besides markka- another name of peacock in Sanskrit
is mayuka mentioned in the supplement of the
dictionary of Hemacandra, The change of the first into
the second of these forms can be easily explained in
the A ustro- Asiatic languages. It is thus that we have in

W V

Malay merah, Cam moriah, and Jarai miyah for the ad jective
" red."

Mayuka and mayura, which have the same meaning
are so much similar to each other that nobody is justified
in separating them. Moreover, one feels tempted to
think of their relationship because -fea and -ra are regular
suffixes in Indo- Aryan. But how has one passed
from the one form to the other ? The Vedic language,
learned and aristocratic, has transposed the popular word,
and images to a higher plane ; it is just the cause of its
obscurity and prestige. Mayuka- borrowed from the abori-
ginal tribes was of too humble an origin to be used in the
Vedic hymns for the most noble bird. Was it possible to
change the termination f Mayuka- had the appearance of
being formed from mayu + ta-.* In substituting -ka by -ra
an unexpected and consequently less popular aspect was

1 The Indo-AryanB were accustomed to .this kind of arbitrary
analysis. The word wo' god, 1 for instance, has been created in an
unjustified manner by isolating the initial a from asura.



NON-ARYAN LOANS IN INDO-ARYAN 133

given to the word. Perhaps the suffix -ra was also more
respectable than -ka as it opposed its popular equivalent -la.

Mayura-, once admitted into the religious literature,
had evolved like other Indo-Aryan words. The existence
of a Prakrit form mo? a explains the name of the Maurya
dynasty. This word which the Chinese translators render
by " the family of the Peacock " id to be classed with
Matanga amongst the names of tribes and royal clans
related to animal or vegetable of which the list will grow
with the progress in our research.

Thus the prefix na- y which we have discovered in
Santali maran and in several A ustro- Asiatic pronouns is
found again afc the beginning of the names of "elephant "
and ".peacock." The elephant, the lord of the forest,
and the peacock, the king of birds, occupy a high place
in the order of beings; maran qualifies that which is
either physically or morally great; in Khasi the particle
ma- is emphatic. To what extent is the element ma a
real prefix ? Is it not rather an element of vocabulary,
a word having an augmentative value ? In a large
number of Indo-Chinese and Indonesian languages mas
or mah is a name of gold, the precious metal par excellence.
In Javanese, the same word mas which means ' gold ' is
also a title of noblese. In Stieng mah means 4 great '
and in Bahnar there is mah " gold " and ma " straight,
the right side."

The problem only can be indicated here as it presents
itself almost in the same manner as other prefixes.
Thus the prefix lam which we have found out in Indo-
Aryan tambulam betel" (B*. S. L. XXIV, p. 256),
normally precedes the names of trees in Stieng and
Bahnar, is used in the languages of the same family
to form the name of ' tree ' : tam4on, torn-Ion, tom-chi



IB* P&fi-ARYAtf AND PRE-DRAVlt>IAN

etc and is in Khmer and in Stieng, under the forms

dom [torn], torn, a real name which means " trunk of a
tree, principle, origin."

SANSKKIT makula-, muknta.

As soon as we admit that in certain words the exis-
tence of prefix ma- had probably the augmentative value
and was susceptible to take . the form mu- we can try
to analyse in the same way other Indo-Aryan words
suspected to have been borrowed from the Austro-Asiatic
languages. It will be seen later on that makuta-, mukuta-
" diadem " contain the same prefix ma-, mu-. Murala,
the name of a fish is also the name of a people (Katha-
saritsagara, 19, 96), which has been assimilated to Kerala
(Diet, of Saint- Petersburg under murala) ; besides Muraia
or Murandala is the name of a river in the country of
Kerala. We have here a play of suffixes and infixes which
denotes an Austro-Asiatic origin. From kerala- one
passes to murala by a substitution of the prefix mu- for
ke- and murandala is derived from murala by the infixation
of tid, a double infix frequent in the Austro-Asiatic lan-
guages. In supposing that ma-, mu- has here .too the
augmentative value it can be understood why this prefix
has been used for designating a river, a people or an im-
portant tribe and the fish eponymous with this tribe. One
may be even tempted to explain in the same way Vedic
marut, name of wind, of the wind-gods and of the tempest,
of which no satisfactory explanation has as yet been given.
If the peacock *marulc(a) is " the bird that cries," could
not the wind and the Maru'ts be called "those who cry"
or the " roarers " ? In one of the Sakai dialects (Malay
peninsula) parug is the word for wind with which one
may compare, in the same region, ru " roar " (like a tiger)



NON-ARYAN LOANS IN INDO-ABYAN 185

v

and M6n paru " roaring, sound." It is true that we
have in Sanskrit marnt and not *maruk, but there is
in Indo- Aryan a root rud meaning precisely " to cry,
to lament, to roar " and the nouns like rntam " cry "
and Budra a the roarer," name of the great god, the
chief of the Maruts. It might be conceived that by
analogy with rud, rutam, the name of wind became marnt.
But here is merely a conjecture. The study of the
proper names in Indo- Aryan cannot yield in many cases
definite results so long as we ignore the importance of
Austro-Asiatic contributions to the religion and civilisa-
tion of ancient India. 1



1 Before publishing such a comprehensive study I have begun to
show in two monographs tbe influence of the Austro- Asiatic civilisation
on the religious and social institutions of ancient India : Etudes
Asiatiques, La princesse a Vodeur de poisson et la nagi dans lea tradi-
tions de VAsie orientate and Journal Asiatique, 1924, II, p. 101 ff. Le
prologue-cadre det Mill* et une nuits et le theme du Svayafavara,



II

Names of Indian Towns in the
Geography of Ptolemy T

Pura is one of the words which mean " town " in
Sanskrit and names of towns ending in pura were numerous
in ancient India. It is, therefore, quite strange that
we meet only two such names in the Tables of Ptolemy :
Selampoura and Mapoura. 2 We find besides a town
which is called in different manuscripts either Fentapolis
or Mentapolis and this word may be a hellenised form
of an Indian name ending in pura. The prologue of the
Suka-saptati mentions a town called Paficapura and places
there a family of Br&hmanas. 3 Pentapolis represents
exactly Paficapura. Pentapolis therefore is comparable
to Pentagramma (Ptol. I, 57), which seems to siand
exactly for Paficagranna. The federation of five villages
(paiicagramt) are known to the law texts (Yajnavalkya, 2,
272). In Pentagramma only the first part of the name
is Greek, but in Pentapolis the whole name is such.
The hellenisation of Rome toponcmies should not astonish
us in a book which enumerates countries like Argyrft

L Jean Przyluski, Bulletin de la SoctW da Linguistique, 1926 (83),
pp. 218-229.

* I refer once for all to the edition and index of Mr. Renou, La
gtographie de Ptoltmte, L'Inde (FIJ, 1-4), Paris, Champion, 1925.

8 The same town is also mentioned ip the 40th section of the
work*



INDIAN TOWNS IN PTOLEMY'S GEOGRAPHY 137

and Khryse and islands like Heptanesia and Trio&ia.
We know on the oth er hand that some Indiab towns were
formed by agglomerations * and this fact justifies names
like Paficapura, Pancagrftma,

By the side of names in -pura we find names of towns
of the type Hippokoura, Barakoura. Koura was no
doubt an Indian word meaning "town." The Mahftbhftiiata>
mentions on several occasions the town of Dantakura and
associates it with the name of the Kalinga people. It
does not seem to be doubtful at all that Dantakura was
"a geographical name either analogous to or identical
with Dantapura," the town of the Kaliogas (3* Lvi :
Notes Indiennes in Journ. As., 1925,1, 4&-^5, infra, App. to
Part III). Dantapura may mean " the city of elephant's
tusks." The country of Kalinga was famous for its
elephants (Le Parinirvdna et let Funerailles du Buddha, pp,
117-8). Dantakura being a designation analogous to
Dantapura, it is tempting to suggest the equivalence :
jDra = /cwm=town, It will be seen just now that this
hypothesis explains a certain number of facts.

In names like Hippokoura, Barakoura, the intervocalic
occlusive k can easily change into a sonant g. This is why
we find in the Tables of Ptolemy another type in
-gflura by the side of that in -koura', Nagagoura,
Spuannagoura, Astagoura. Naga no doubt stands for
naga " dragon " ; Souanna for snvarna (of. Pali .savanna)
" gold " ; asta for asta " eight/' Nagagoura was probably
the city of naga (cf. modern Nagpur) ; Sounnagoura,
"the city of gold" and Astagoura a name with a
numerical element at the beginning comparable to



1 On the different agglomerations comprised under the sole name
of Vaisali, cf. Uvasagadatao^ edit. R. Hoernle, foU. 1, n. 8.

18



188 PRE-ARYAN AND PRE-DRAVIDIAN

Pentapolis, Pentagramma. 1 The manuscripts of Ptolemy
mentions still another town called Gammogoura or
Brammokoura, The last reading conveys an acceptable
meaning. Brammokoura was doubtless the city of Brahma
(ef. in Ptolemy the towns Bramme and Bramma and
in Burma the famous Prome). The hypothesis pura=*
fefittt thus becomes more probable.

: I have shown elsewhere (Journ. A*., 1926, pp. 25*9)
that in some Indian names the initial frequently disappears.
This phenomenon explains the passage otpura or feira
to ura. In fact the names of towns ending in -oura are
numerous in Ptolemy : Poloperoura, Koreoura, Karoura, etc.
Numerous are also the names with the termination -our.
These names stand in the same relation to those in -oura
a* the modern names in pur to those in pura. Nagour
for example is comparable to Nagpur and Nagaoura.

Besides the names in the Tables of Ptolemy ending
in 0ifra and -our are not without equivalents in modern
onomafiticon. Cannanore or Kannanur or Kannur, for
example, is the city (ur) of Krsna (cf. Hofaon-Jobson,
8.7.) ; and to this Dravidian word ur " city " or " village >f
corresponds in Canarese uru and in Tamil ura ( Ling. 8urv.
Ind. IV, pp. 325 & 679). It is also a priori possible that
the names of cities in -oura and -our are derived from
pur& as well as kfira. The quantity of u in the Dravidian
won) fifa, nru, ur makes us more inclined to admit the
second alternative. 1

. To sum up, besides the Indo-Aryan words like p*ra,
gtfma, etc., the names of towns in the Tables of Ptolemy

1 For the name of towni with " eight " & its firit element, of.
Atthakanagara at Sanchi loser., No. 20i, and in A*guttttro t V. 842.

> We have pir in Vedic bat pur a hat always .



INDIAN TOWNS IN PTOLEMY'S GEOGRAPHY lad

have for their second element, a series of forms * konra,
goura, oura, our which are interconnected with each other
and can all be explained by the single word kura meaning
"city."

As it often happens, the results which we have just
arrived at. raise new problems. If pura is Indo-Aryao,
to which languages do kura and other words of the series
properly belong ? What is the origin of the word kura ?

We have just seen that the word Dantal&ra seems
to be .modelled after Dantapura. There are still two
other forms. Dandagula in Pliny VI, 78, corresponds
to Dantakura and Professor Sylvain Lvi has precisely
shown that Paloura is another term for Dandagula (Notes
Indiennes, p. 53). We have thus four forms corres-
ponding with each other : Dantapura, Dantakura,
Dandagula^ Paloura. If we suggest the equivalence :
pu ra**kura=gula i= oura= city then it necessarily follows
that :

danta=danda='})al=zkvi8k of elephant.

Now it remains to be answered in which languages pal
can mean "elephant's tusk." We have in the Austro-
Asiatio family :

Malay Peninsula bala', bala, baP

Khmer phltik

Stieng bfok

Kaseng blok

Sue lott

Haling milb

Sedang bota

Jarai, Bahnar tola

Cam bala

Kon-tu palb



PR&ABYAN AND PRE-DKAVIDIAN

All these words mean "elephant's task" and are
connected with a root bal which signifies " horn " in the
Austro-Asiatic languages and hence, means of defence for

V V

the elephant. The example of Khmer : phluka fish '* " fish " ha (common Austro-Asiatio).

1 Deliteoh, Glower.

/bid, p. 128, and cf. B.S.L., XXV, pp. 56-59.



148 f RE-ARVAN AND PRE-DRAVlDIAN

Bomerian. Anstro- Asiatic.

Jcu " (ktia) "

riff " skin covered with " hairs," sok (Mon)

"hair," sok (Stieng)

"wool" suk (Old Khmer)

800}

mk \ (Malay Peninsula)
salt)

gblu " man" " man " kol (common Munda)

kur (Malay Peninsula)

mnlu mono (Halang)

ur benu (Lave)

ur-an (Cam)
or-an (Malay)
burn" fruit" " fruit peK

pele (Malay Peninsula)
pie

plei (Bahnar)
phU (Khmer)

gnla" great" " great " karu

gal kadu (Nikobarese)

' karu (Chowra)

kadni (Malay Peninsula)

mah " great " Augmentative prefix ma

lie demonstrative Demonstrative ni, ne (common

Austro- Asiatic).

These analogies, of which the lift is cot exhausted,
can be explained as loan-words. It is neither impossible
that Sumerian should be related to the Austric languages.
But it would be premature to try to choose between the
two alternatives.



Ill

Non-Aryan Loans in Indo-Aryan

Kodumbara, Odumbara 1

In a series of articles published in the Mimotres
and the Bulletin de la SocictS de Lingimtique I have begun
to show, since 1920, what the Indo-Aryan vocabulary
owes to the Austro-Asiatic 2 languages. After having
prepared the way, I have marked out in these languages
a certain number of words of which the initial has either



1 Taken from Prof. J. Przyluski's article : Un ancien peuple
du Penjab : Let Udumbara, J. As,, 1926, pp. 25-36.

8 While speaking of " Austro-Asiatic " languages I attribute
to this word a meaning of comprehensive nature not in use till now.
I use it for designating a linguistic family which singularly surpasses
the limits of Austria Asia. It is difficult to find a geographical term
which will be sufficiently general to include all the languages spoken,
not only in Eastern Asia but also in a large part of Oceania, Africa
(Madagascar) and no doubt to the north of the Pacific in ancient
times. "Austro-Asiatic " is suggested by the hypothesis that Austria
Asia is the region whence the people speaking these languages seem
to have dispersed. Such was the opinion of H. Eern who admitted
the continental origin of the Malay o- Polynesian peoples. This
hypothesis has not yet been confirmed by any fact. P. Schmidt
who was the first to speak of " Auatro- Asiatic " languages reserve
this word for the languages of the continent and makes the " Austro-
Asiatic " language a subdivision of the Austria language. This
last term has the disadvantage of apparently confining the language,
still largely spread to the north of the equator, only to the Austria
hemisphere. Probably it would be better to reserve the word
"Austria" for designating the languages spoken in Oceania before
the arrival of the Malayo-Polynesians.



150 PKE-ARYAN AND PRK-DRAVIDIAN

softened or completely disappeared. It is thus that for
" son, child " we have the following forms :

kon kkon han 1

In han the guttural initial is reduced to an aspirate.
On the other hand in Khmer amlas " cotton " comes
from an ancient *kambas of which the guttural initial has
completely disappeared (Bulletin de la Socitte &e Linguit-
tique, 1924, p. 70, and supra, p. 23). The comparison
of the words for man and woman in the Munda languages
reveals an analogous phenomenon :

" man " koro 9 Aar, horol, hara, hor

" woman M kurl, kori, era

Here for the same root we have different condition?
of the initial, k, h and zero. It would be easy to show
that the same phenomenon has occurred in the case of
other initials in that linguistic family. Thus for u salt "
we have :

tampoying, empoya, ambang 2

For the " breast/' Cam has two forms : barann and
araun. In Santali the Hibiscus sabdariffa is either
bambara or ambaro. To brang * black" in Bahnar (<.*berang)
corresponds heram in the speeches of the Malay Peninsula,
hireng in Rawi and ireng in Javanese. It will be easy to
multiply examples. When the principle once comes to
light the difficulties begin in our attempt to draw con-
clusions from it.

1 Memoir es de la Societf de Linguistique, 1921, p. 209 (c/. supra,
. p. 29). The form han is common to several Munda languages :
Mundari, Santali, etc.

* Cf. Skeat and Blagden : Pagan Races o/ the Malay Peninsula,
II, p. 702.



NON-ARYAN LOANS IN INDO-ARYAN 151

KusinarS is the Sanskrit name of a town which
became famous by the parinirrana of Buddha ; the
Mahabtorata (VIII, 5, v. 187 ; XII, 101, v. 3736) speaks
of a people called USinara. These names form a pair of
which the elements can be nearly superposed ; the most
notable difference is due to the loss of the initial in
the second, a phenomenon frequent in the Austro-
Asiatic languages. Can we deduce from it that the
pair has been borrowed from that family of languages ?
It will be however imprudent, The apheresis is observed
in different linguistic families. Sumerian (Autran,
Langnes du Monde, p. 278) furnishes some such examples
and some Sanskrit words are precisely suspected to be
loans from that language.

Even supposing that Sumerian has nothing to do
here and that amongst the languages which were in
contact with Indo- Aryan in ancient times those which
belonged to the Austro-Asiatic family were the sole to
present cases of aphereses, it would be still imprudent
to say that the words which have lost their initial in
Sanskrit would necessarily be of Austro- Asiatic origin.
The softening of the initial lt>h appears to be dtfe to
the fact that, in the dialects in which it is produced
the pronunciation of the occlusive necessitated a strong
breath. The same pronunciation has been able to main-
tain itself and produce the same effects even after the
non- Aryan populations of India had learnt to speak the
Indo-Aryan languages and, since then, some words of
Indo-European origin could have lost their initial.

In short, the loss of initial in some words of Indian
languages would not prove their A UP tro- Asia tic origin. I
cannot therefore admit without reservation the following
formula in which Prof. Sylvain Lgvi recently summarised



152 PRE-ARYAN AND PRE-DRAVIDIAN

the results of his researches on some Indian geographical
names : " Pulinda-Kulinda, Mekala-Utkala (with the
group Um. The prefix with guttural initial in



1 The cerebral of *tumba which has disappeared in Sanskrit from
the beginning of the word tamba is preserved in the interior of the word
gofumba. Similarly in some Austro- Asiatic languages, for example
in Annamese, the old initial t has become t while it is retained at
the end of the word. It should be remarked that in Sanskrit there
are vqry few words with cerebral initial.



NON-AEYAN LOANS IN INDO-ARYAN 157

godumba is found again in ka4i-m*n, and with apheresis
in antemon (< *kantemov). The initial a of Sanskrit
alabu is probably also the rest of an ancient prefix.

In Javanese walu is the form which corresponds to
Malay labu; walu, without doubt, comes from *luwa
(<*iumba) by metathesis. Several other Indo-Chinese
words for gourd can be connected with the latter :
Bahnar puol\ Rongao puol ; Kha par and similarly
Annamese bd'u " Lagenaria mlgaris, gourd."

We now see how many different forms the same root
has taken in the vast domain where it has spread. It is
little probable that these variations are solely due to the
normal display of phonetic laws. We know that the
words having religious import are subject to systematic
deformation ; the word being tabooed under its normal
form, was much changed in use. The myth of gourd-
mother sufficiently explains the respectful fear, inspired
by the name of this fruit and the modifications the word
has undergone.

On the whole, it seems that an Austro-Asiatic root
like *tumba, either accompanied by affixes or not, has been
used to mean the cucurbit, i.e., the fruit which contains
a large number of seeds in it. To the same family
belongs the Sanskrit loan words : tumba, tumbi, tumbuka,
godumba, labu, alabu, alabu. The same root explains
still a few other Indo-Aryan words.

Udumbara is the Sanskrit word for Fieus glomerata,
a tree which is native of Burma and found in India
principally in the Sub- Himalayan zone (Watt,
Dictionary of conomic Products of India, s. v. Ficu*
ghmerata). The fruit of this tree resembles the European
fig except that it is smaller and of oval shape. Its
shape is rightly comparable to that of some small



158 PRE-ARYAN AND PRE-DRAVlDiAN

coloquintidas and the large number of seeds which it
contains adds to the resemblance. 1 Therefore we need
not be surprised to find an analogy between the names of the
coloquintida : Sanskrit tumba, and that of Ficus glomcrata :
Sanskrit u (lumbar a. It may be remarked however that
the cerebral of *tumba reappears in the classical Sanskrit
form udumbara. Such rapprochements seem to be more
justified when we consider the very words for udumbara
in the modern languages of India : Santali Iowa ; Chote
Nagpur dume ; Kherwarien dumer ; Oriya dimert Nepal
dumri. [Bengali dumur P. C. B.], Between labu, the
Malay word for different kinds of gourds (the Sanskrit
equivalent Idbu, the corresponding word in Javanese
walu) and Iowa the word for Ficu* glomerata in Santali,
the resemblance is as exact as possible and can be very
well explained by semantic laws.

What does then the initial u in Sanskrit udumbara
represent f This cannot but be the rest of an ancient
prefix which has now lost its initial and the analogy of
ketimun, godumba suggests here an ancient guttural.
Udumbara \ udumbara then would come from an ancient
*kutumbara \ *kutumbara and we are thus brought back
in a roundabout way to the hypothesis already formulated
above. 8

It is probably in the same family, from which Sans-
krit tumba, udumbara, etc., come, that we have to search
for the origin of names of some musical instruments
and an Indian caste. Labuki, the word for a sort of



1 In the eyee of the Austro-Asiatice the abundance of seeds was
certainly the characteristic feature of these fruits, as the myth of the'
gourd-mother giving birth to numerous children shows.

* Of the two kindred words for fine cotton stuff in Pali the one
has a cerebral : kotumbara while the other has not got it : kodumbara.



NON.ARYAN LOANS IN INDO-ABYAN 159

late is certainly derived from Skr. labu "gourd." l
Itemaru is the Sanskrit and the Bengali name of a small
drum which plays an important role in Indian iconography
as the attribute of several divinities (cf. Curt Sachs,
Die MusitMrumente Indien* und Indonesiens, 2nd ed.,
p. 75). The instrument called in Marathi ddmru, in Hindi
4amru t etc., resembles a gourd, with two swellings, cut in
such a way as to have only the two hemispheric ends. The
analogy of the names of this drum with those of the
udumbara fruit in the Indian languages can be therefore
explained by their common ressemblance to some cucurbits*
Tambura is the Marathi word for a kind of vina. It does
not appear doubtful at all that the instrument owes its
name to that of the two hollow appendices which are
suspended from the tube like the gourds from their stalk. 8

1 Cf. also the name of a kind of Bengali violin : alabu sarangl
(Curt Sacha, p. 131).

* This cithara must not be confounded with a lute which has almost
the same name : Hindi \ambura (Cart Sachs, ibid, p. 129, fig. 90), bnt
which resembles a mandolin of which the body is like a big calabash.
Mr. C. Sachs is of opinion that the name of this instrument is the
same as Persian lanbur, which is derived by metathesis from pandur,
name of a late in the Near East. India would have received the
tanbur from her western neighbours and modified the name by analogy
to that of Tumburu, the king of celestial musicians. In fact it seems
that Hindi tambura designates an instrument imported to India from
the west bat it is difficult to decide if the Hindi form tamburS is
due to a contamination with the name of Tumburu or that of the
cithara called tambura in Marathi. This last instrument seemi to be
indigenous in India and its name can be explained by the two
calabashes hanging from it. It is not however impossible that
Tumburu owes its name originally to the musical instrument which
has the shape of a gourd or a calabash ; according to this hypothesis
the name of the gandharva king would also belong to the family ;
a, urftimfcara, etc,



160 PRE-ARYAN AND PRE-DRAVIDIAN

Last of all, c^omba, in Sanskrit, is the name of a low caste
earning its livelihood by chanting and singing. The
word appears under the form Domva (Dova) in the list
of the Mlecchas taken by Weber from the Jaina texts
(Indische studien, XVI, p. 832). In the modern languages
of India, dome, dhombe, dombar, dombari, dumbar, dumbaru
designate the degraded aborigines scattered all over
India (Hobson-Jolson, s, v. dombe). It seems that the
same non-Aryan word or its variants have been used
for the musical instruments made of gourd or calabash,
as well as for the native musicians who played on
that instrument and the caste from which these poor
musicians were recruited.

On the whole, udumbara, the Sanskrit word for ficus
glomerata, belongs to the long series of words
borrowed by the Indo-Aryan languages from the Austro-
Asiatic ones. Udumbara also designates a country of
the Pen jab and its inhabitants. It has been observed
that in the neighbouring countries in the South-Sea, the
names of peoples or countries are frequently borrowed
from the vegetable world. If the name udumbara is of
Austro- Asiatic origin, there is nothing surprising that the
name of a people and a country has been derived from this
tree. On the other hand, the economic and linguistic
history forbids us to separate from it the word kodumbara,
which is used for a sort of tissue as well as for the people
who used to make it. It should therefore be admitted that
Udumbara, Odumbara, Kodumbara are the variants of the
same name of an Austro- Asiatic people of Northern India.



APPENDIX

TO

PART III



I

Paloura-Dantapura i

( SYLVAIN LK'VI )

On the eastern coast of India Ptolemy mentions a
locality called Paloura (VII, *, 16) which he chooses
as one of the bases for the preparation of his map.
He places Paloura at 136 40' East x 11 20' North at
the entrance of the Gangetic gulf and at 20' North of the
apheterium where the vessels bound for the " Peninsula of
Gold " (KhrusS Khersonesos) ceased to follow the littoral
and entered the high sea. The map of Ptolemy places
the apheterium at the southern extremity of an imaginary
peninsula which inclines south-eastwards starting from a
point which approximately agrees with the Calimere point,
right to the north of Ceylon and passes straight to
the north for extending at last in the west-east direction
towards the Gangetic delta. In his first book (I, 13, 5-7)
Ptolemy had already discussed in detail the position
attributed to Paloura by his predecessor, the navigator
of Tyr, and corrected in his own way the distance
indicated by this navigator between Paloura and the port
of Sada, on the opposite coast.

One is surprised to find that a locality placed in such
an exceptional situation on the route of maritime commerce
between India and the Far- East, is not mentioned else-
where The name belongs to the Dravidian type ; it is one

1 Cf. J. As., 1926, pp. 46-57 (Hotel Indiennes) already referred to
by Prof. Preyhnki ; cf. supn, p. 186.



164 PRE-ARYAN AND PRE-DRAVIDIAN

of the long series of names ending in *our and *ovra,
an element which has long been recognised as the
Dravidian word ur "city." Caldwell (Comp. Grammar
of the Dravidian Languages, In trod., p. 104) has explained
the name Paloura by Tamil pdl-tir " the city of milk."
But another explanation can also be offered. In Tamil
for f tooth' we have the' word pallu (Telugu : pallu ;
Canarese : liallu ; Malay alam : pallu ; Oond : pal, etc. ;
cf . Linguistic Survey, Vol. 1 V, 650-652, No. 37) ; Paloura
can be the "city of tooth." The Indian tradition is
acquainted and since long, with a "city of tooth,"
Dantapura, in the country of Kalinga, just in the same
region in which Paloura is located. It is only in the
Buddhist tradition that Dantapura has become famous ;
the name of the town is associated there with a famous
relic, the tooth of Buddha, which is worshipped till to-day
at Eandy in Ceylon. The common tradition on the
division of the relics just after the Parinirvana relates
that one of the teeth of the Master passed to the king-
dom of Kalinga (Dlgha, II, p. 167 ; Buddhavarfisa, Chap.
XXVIII ; Dnlva in Rockhill, Life, p. 147). A late poem
in Pali, the UMavajhsa (ed. J. P. T. 8. 1884) by
Dhammakitti, relates the pretended history of this relic ;.
it was brought to Dantapura by the Muni Ehema, under
the reign of Kalinga Brahmadatta ; it remained there as
the object of a cult till the reign of Guhaslva who for
saving it from defilement, entrusted it to his son-in-law,
Dantekumftra of IJjjayinI ; it was taken by him to
Tamralipti, and thence on board a vessel to Ceylon where
it was piously received by the successor of MahSeena,
Kitti-siri-megha (middle of the 4th century A.D.), the
same prince who had sent an envoy to Samudragnpta
for the Mahabodhi monastery. The two facts are related



PALOURA-DANTAPURA 165

to each other ; Kitti-siri-megha seems to have imposed
himself as a champion of Buddhist interests at a time
when India was passing through a period of reaction
against that religion.

For the Buddhist tradition, Dantapura is one of the
most ancient cities of India ; it appears as the first in
the list of six cities built by Maha Govinda, in the time
of King Renu :

Dantapuraifa Kalinganath Assak&nan ca Potanarh
Mahissatl Avantlnaih Sovlranan ca Rorukath
Mithilft ca Videhanam Campa Angesu mfipitft
Bar&nasl ca K&slnam ete Govindamffpitft.

This versified list has been inserted in the MahS
Govinda Sutta of the Dlgha NikSya, XIX, 36 ; it is also
found in the corresponding Sutra of the Dlrgh&gama,
extant in two Chinese versions, the one (Tok ; XII,
9, 28", 5) incorporated in the Chang A-han the long
Agama and the other translated separately (XII, 10,
50", 3) by She-hu (D&napala). It also occurs in the
Mahavastu, III, 08. Several stories of the Mahftvastn
have for their scene Dantapura in the country of the
Kalinga, III, 361, 364 (similarly cf. alto the Pftli
Jfitaka : Kurudhamma II, 67 ; Culla Kalinga, HI,
8; Kumbhaltara, II, 876; Kalifyabodhi IV, 230);
they likewise speak of events which took place "at
the time when Kalinga was reigning at Dantapura in
Kalifiga" (Kalingaratthe Dantapuranagare Kalvhge
rajjam karentc). In the KumbhaJcara (jataka) the king
who reigns at Dantapura in Kalifiga is the famous
Itarandn, whose name is associated with those of Naggaji
of Gandh&ra, Dummnkha of Uttara-Pancala y and Nimi
of Videha who abdicated for living like ascetics. Their



166 PRE-ARYAN AND PRIUDRAVIDIAN

glory is not less in the Jaina literature where Karantfu is
transformed into Karakaniju, who was also king of Kalinga
at Dantapnra ; the magnificent Jaina encyclopaedia!
now in course of publication, the Abhidh&na-Rajendra,
relates at length the biography of Karakautju (s.v.) and
refers to a series of texts ; it will suffice here to retain
the account of the Uttaradkyayana sutra, XVIII, 45-46,
with the commentary of Devendra. Dantapura of Kalinga is
also famous amongst the Jainas as the capital of King
Dantavakra (vakkaj, " the best of the Kgatriyas " accord-
ing to the SutraManffa I, 6, 22, specially known for
having involuntarily roused an emulation of devotion
and heroism between two friends, Dhanamitra and Dpjha-
mitra, the Indian parallel of Orestes and Fylades or of
Damon and Pythias, etc. ; the word danta " tooth " or
" tusk " has suggested the beginning of the story : the
wife of King Dantavakra, who was enceinte had a desire ;
she wanted a palace entirely built of ivory ; the king
therefore orders to keep all available ivory reserved for
his use. Unfortunately, the wife of the merchant
Dhanamitra, who was also enceinte, was seized by the same
desire ; for satisfying her the merchant and his friend do
not obey the order of the king ; then each of them puts
forth his claim to be punished ; the king is moved and
pardons them (cf. Abhidhfina-Rftjendra s.v. pacchitta>
* Vol. V, p, 186, and for the references s.v. Dantavakka).

The Mahft-Bhfirata mentions a prince named Danta-
vakra (pass. ; cf. Sorensen, Index, s.v.) but he is the
king of Kftruga, the country situated between Cedi and
Magadha, to the south of K&H and Vatsa. Dantavakra,
the Kftraga, is mentioned quite frequently in the Hari-
vaxhla, almost always in the company of Kaliftga; he
ia tW terrible adversary of K|-Da who at last kills him*



PALOURA-DANTAPURA 167

Tkat is doubtless the origin of the reading adopted by
the Southern manuscripts, in the passage of the Mahft-
Bhftrata quoted later on (p. 168) : " He (K^na) ie
destroyed the Kalinga [and] Dantavaktra."

I have not been able to discover any reference to
Dantapura in the Brahmanical literature ; however * late
epigraphio evidence proves that the name of this loeality
was in use for a long time. King Indravarman
issues a donation from his residence at Daniapunt
(Dantapuravasakat ; Ep. I*d., XIV, 881 : Purle Plates
of Indravarman, the [Gafiga] year 149) while the
kings of this dynasty generally issue their charts from
Kalifiganagara. Indravarman makes to a Brahmin
the donation of the village of Bhukkukfrra, in the Kfiraka-
rtyfara (modern Bhukkur in the PSlakonga Taluk), the
place where the inscription was found. The editor of
the inscription, G. Ramadas, observes : " On the road from
Chicacole to SiddhSntam [the name of a village which
Mr. Ramadas wants to reconstruct as Siddhftrthaka-
grama c the village of the Buddhists ' ! ] and very near
this last place a large piece of land is shown as the
site of the fort of Dantavaktra." Mr. Ramadas, if he
had known the Jaina legends, would not have missed
to notice the really surprising persistence of the memory
of King Dantavaktra or Dantavakra (the forms alternate
and are confounded with each other in Prakrit Danta-
vakka) about whom the legend connected with the name of
Dantapura has been just now mentioned.

Though the MahE-Bhftrata does not contain any
mention of Dantapura yet it refers, on several occasions in
connection with Kalinga, to aname of which the first element
is Aanta. In the 5th canto (adhy. 25, v. 708), Yudbijjhira,
wtrile recalling the exploits of his brothers speaks out,



168 PRE-ARYAN AND PRE-DRAVIDIAN

"The son of MSdrl, Sahadeva, has vanquished the
Kalifigas assembled at Dantakfira by throwing lance right
and left!"

MSdrlputrah Sahadevah Kalingan samfigatftn ajayad
Dantakure | VSmenisyan dakinenaiva yo vai mah&baladi
kaooid enarfi smaranti II

Further on in the same canto when Sarfrjaya reports
the words of Arjuna glorifying Krgna (adhy. 47, v. 1883) :
" It is he who has vanquished the Pftndya at Kav5$a and
destroyed the Kalin gas at Dantakura."

ayarii Kavfi(e nijagbana PSndyaifa
tathft Kalingfin Dantakure mamarda |

Against this passage the translation of P. C. Roy
has : " It was he who slew King Pftndya by striking his
breast against his and mowed down the Kalingas in
battle/ 9 The translator adds in a note " some texts read
kapatt nijaghana meaning ' slew in the city of Kapfi(a.' "
For his translation he has followed the reading of the
Calcutta edition : kapatena jaghana. The meaning, as
is evident, is quite different. The translator has followed
the commentary of Nllakantha, who adopts kapatena
jaghana and interprets Jcapaia as " thorax, the breast as
large as the door-piece ; " Nllakanfcha, in the second
pada, arbitrarily attributes the meaning of "the battle
in which one shows the teeth " to the word dantakura.

The southern edition reads (adhy. 48, v. 76) tavate
nijaghana and dantavaktraih mamarda. The gloss inter-
prets kavate as nagarabheda " a particular town ; " it is
silent about the rest. It is however curious to see the
King Dantavaktra, persistently associated with Kalinga,
reappearing here, even in defiance of the syntax which
does not allow two accusatives juxtaposed (tatha kalib-
gan dantovaktram mamarda).



PALOURA-DANTAPURA 169

The word dantafara is again mentioned in the Mahi-
Bharata, VII, 70, 7 at least in the Southern edition.
The poet recalls the exploits of Paradurftma in his fierce
fight against the Kfatriyas. " There, fourteen thousand
enemies of the Brahmanas and others still, he stopped
them and killed them at Dantakura."

Brahmadvisarti catha tasmtn sahasrani caturdafa
punaranyan nijagraha Dantakure jaghana ha.

The commentator mentions another reading, danta-
kruram ; t( in this case, he says, the word designates the
king of the country/' To say in other words, when
it is not a place-name formed with *kura, it is the name
of a man formed with *krura " cruel ; " and one should
take it in the accusative. The translation of P. C. Roy
follows the reading Dantakure and translates the passage
thus : " In that slaughter were included fourteen thousand
Br&hmana-hating Kshattriyas of the Dantakura country."
The Calcutta edition has preferred to read Dantakrfirarh
jaghana ha; it is also the reading followed by Nllakan(ha
whose gloss (taddcsa-dhipati) has been reproduced by
the annotator of the Southern edition. The authors of
the Petersburg Dictionary have (s. v. dantakruram) treated
the word as an adverb and translated it as " auf eine
grauenvolle Weise mit den Zahnen " = " in a ferocious
way with the teeth ; " they have referred to this passage
only. But later on, in the abridged edition, Bohtlingk
substitutes the nominal stem dantakrura for the adverb
dantakruram and interprets it thus : " Name of place
(according to Nllakan(ha) ; it should be read as dantakure
instead of dantakrftram"

The choice between Dantakura and DantakrVra, left
uncertain by the tradition of the MahS-Bhftrata and the

22



170* PRE-ARYAN AND PRE-DR&VIDIAN

meaning of the word too, also uncertain, are definitely
fixtd'bjjr tb* evidence of Pliny. In Book VI, IK, he
proposes to give the measurement of the coast up to the
Indtw, as be finds it, by indicating the distances, though
there is no agreement amongst the different itineraries.
He indicates as the first step : ab ostio Gangis ad promvn-
torinm CalingonetoppidumDandagula DQXXV M.passuum :
" From the mouth of the Ganges to the promontory of
Calinga and the fortified place of Dandagula, 625,000
steps." The promontory of Calinga, which constitutes
so clear a starting point for measuring the coast, is
evidently, without any doubt, the point where Ptolemy
locates the aphelerium of navigation towards the peninsula
of Gold and which marks according to him a sudden
change in the direction of the littoral. The nearest
oppidum cannot be anything except Paloura of Ptolemy,
otherwise called Dantapura and in the spelling of Pliny :
DandagulA; it is easy to recognise in it the name of
Dantakura. The distance of 625,000 steps, retained by
Pliny from the confused data of the itineraries, correspond
to -3, 84 5 stadia. Ptolemy calculates 500 stadia for one
degree to the equator and consequently on each meridian ;
the distance from the Ganges to Dandagula would there-
fore correspond to about 6 36' in Ptolemy. Between
Rtkrara and the westernmost mouth of the Ganges
Ptolemy indicates a distance of 7 50' in longitude (186
4*^-144 W) and 6' 55' in lattitude (11 20' 1 8 15').
Ii appears therefore that Ptolemy worked on data which
weft* very near the approximate calculation of Pliny:
bottfar the conception of the whole, which forced a
systematic deformation, he would have arrived at a suffi-
ciently faithful localisation in this part of his map;
th* ddfe of the Ganges is placed towards 22 P degree in



PALOttRA-DANTAPURA 111

the north; the region in which Dantapura can >be conve-
niently searched for and where the local tradition still locates
the fort (oppidum) of Dantavaktra is in the neighbourhood
of Chicacole and Kalfogapatam, "the city of the
Kalingas," a little to the north of 18 degree ; the distanqe
between these two points is of 5 to 6 degrees along the
littoral.

I have shown elsewhere * that the apfieterion en
Khrnsen of Ptolemy is identical with Caritrapura of Hiuan
tsang and some Sanskrit texts. 1 would point out another
feature more common both to the Greek and the Chinese.
Ptolemy, as we have seen, places the aphetcrium at .A
promontory where the coast turns suddenly from the
direction of W. N. W.-E. S. E,, to that of S.-N., inclined
at last towards the East separating the Argarique gulf
from the Gangetic. The Caritrapura of Hiuan tsaog is
situated on the South-Eastern border of Orissa (M4m,,
III, 90), and to the North-East of the kingdom of
Malakuta (Md, 124), Orissa, towards the South-East
borders the kingdom of Kong-yu-fo, Kongoda which
formed a mancjala of Daksiuakoala and corresponds
to the present district of Ganjam. " The frontiers of
this kingdom/' writes Hiuan tsang, " embrace several tens
of small towns which touch the hills and are situated
at the confluence of two seas." Such is at least the
translation of Julien ; Waiters (Notes, II, 197), criticise*
it and says, " there is no word for two in the text and the
term hat kiao here means the meeting of sea and Jand."
The pilgrim wished his readers to understand that the
towns at one extremity 'continued the hills '



1 C1. Etudes Asiatiques published by 1'Ecole
d'Bxtrtme Orient I.



172 PKE-ARYAN AND PRE-DRAVIDUN

ling) and at the other were on the sea shore. I would
not enter into this Sinological discussion ; I must however
point out that the word kiao regularly means " meeting
point ; exchange, mixture/' and the expression " situated
at the meeting-point of seas " very well renders the
geographical conception which Ptolemy had adopted for
apheterium.

To find a well defined turning point along the eastern
coast it is necessary to go up to the Palmyras Point
which brings one to the mouths of the Ganges, at
20 44' 40" N. x 87 V E. to the north of the mouth of the
Mahfinadl. But Ptolemy places the apheterium right to
the south of this river which he calls Manadas, midway
between its mouth and the mouth of the Maisolos which
represents at the same time the Godavarl and the Kistna.
Besides, the current which, during the South- Western mon-
soon, passes from the Indian coast to that of Burma, avoids
the Indian coast in the region of Chicacole at Kalinga-
patam at about 18 degree. I refer once again to the
text of Valentrjn (1727) which Yule has pointed out
(Proceed. Roy. Qeogr. Soc. 1882): "In the beginning

of February, a small vessel sailed towards Pegu with

a cargo which she took at Masulipatam From

this point she followed the coast up to the 18 degree
north and thence she reached the high sea for attaining
the opposite shore at 16 degree/' The Maps of India, even
in the 17th century, for example, that of William Blaca,
continue to mark a sudden turn and inclination of the
coast, in conformity with the indications of Ptolemy, between
the ports of Masulipatam and Bimlipatam (to the north
of Vizagapatam at 17 53' 15" N. and 83 29' 50" E,).

Moreover, it would be rather naive to wish to interprets
the data of Ptolemy too literally for all this region;



PALOURA-DANTAPURA 173

the precision in bis orders should not deceive anybody
on the real worth of his knowledge. He locates Paloura
a little to the north of the apheterium : Pliny, on the
other hand, starting from the mouths of the Ganges,
mentions first "the promontory of the Calingae" and
then, consequently to its south, " the fortified place of
Dandagnla ; " Ptolemy ignores the name of Kalinga,
either territorial or littoral ; perhaps one should recognise
this name, still famous, in the city of Kalliga which
Ptolemy LII. '. 93 enumerates amongst " the inland
cities " of Maisoloi where Pitundra also is met with. I
had already the occasion of pointing out the strange
transposition by which he has taken Tosall (Tdsalei
metropolis) of Orissa right into the territory of Pegu ;
it seems that the whole of Orissa and a part of the
neighbouring countries have undergone a similar trans-
position, probably through the confusion of the two
itineraries, one inland from the South to the North and
the other maritime from the West to the East.

Now that the name Dantakura is well established as
a geographical term, analogous to or identical with Danta-
pura, one will be surprised to recognise the enigmatic
element kura in the name of the kingdom Kurakarfc^ra j
the village Bhukkukura, given by King Indravarman while
he was at Dantapura, formed a part of the latter kingdom.
The editor of the chart, Mr. Ramadas, is astonished to
find here the term ra$tra "kingdom " while the provinces
of Kalinga are elsewhere designated as vigaya. The
Kuraka or Kura kingdom was perhaps an ancient expres-
sion, retained in a long use, for designating the territory
near the capital Dantapura.

This singular element fara which appears to be alter*
nating with Sanskrit pura for designating in combination



PBK-ABYAN AND PBB^DEAVIDIAN

with JDanta , the capital of Kalinga, recalls, at least through
analogy, the last syllable of the uame of the city which
Ptolemy writes Hippokoura (VII, 1,83). Hippokoura is
situated in the southern part of Ariake to the south of
Paithana (Paithaq, in the upper valley of the Godavari,
in the dominions of Nizam, district of Aurangabad) and
Tagara (Ter, Ther, in the dominion of Nizam, district
Naldrug), and to the north of Banaouasei (Banavasi in
Mysore, district Shimoga). Hippokoura like Dantakura,
is a royal capital ; it is basileion Baleokourou^ " the royal
residence of Baleokouros," The name of the sovereign also
seems to reveal an element kura. Baleokouros is, without
any doubt, an approximate transcription of the mysterious
Vijiv&yakura, which is joined with the name of kings
SStakarni Vasi$(hlputra and Satakarni Gautamlputra,
on a special type of coin, different from the type
generally used by these two princes and confined only
to the southern part of the Mahra$ha country, more
precisely to the district of Kolhapur, which
agrees well by its situation to the point which Ptolemy
indicates for Hippokoura. On the title of Vijivfiyakura
I cannot but repeat on my part what Prof. Bapson
has said in his excellent Catalogs of the Coins of He
Andhra Dynasty, 1908, to which I also refer for the descrip-
tion of the coins : " No satisfactory explanation has as yet
beet given of the forms Vilivayakura and Sivalakura "
(p. Ixxxvii). "Sivalakura" which is coupled in the
same way with .the name of King Mfitjharlputra, contains
also the element kura. The name of Hippokoura
reappears in the Tables of Ptolemy (VII, 2, 6) as the
name of a port situated in the immediate neighbourhood
and a little to the south of Simulla (Cemula, Chaul, 23
miks. south of Bombay); the Pariplue does not mention



PALOURA-DANTAPURA 175

it. Since this latter Hippokoura is located by Ptolemy on
the littoral of AriakS, which is distinguished as AriakS
Sadindn, it may be asked if it is not the same locality,
carried to the sea-shore through a wrong interpretation of
the itineraries,

One may feel inclined to compare the final toura in
Hippokoura with kourai which appears to be the plural
form in the name of Sosikourai (VII, 1, 10)
which certainly corresponds to Tuticorin ; here kourai
is doubtless the equivalent of the Tamil word kudi
" place of habitation, town " (cf. the texts, s. v. Tuticorin
in Hobson-Jobson by Yule and Bnruell). But the
identity kura-konrai is very doubtful.

Whatever might be the interpretation of the term
ASra, the identity Paloura=Dantapura seems to be now
definitely established. The geography of ancient India
thus finds in Pliny and Ptolemy new data for localising
the site of a large city of the past. The alternance
Paloura-Dantapura, besides, shows that in the time of
Ptolemy the Draviclian language was disputing the 1
territory of Ealinga with the Aryan dialect. Even to
our day, Chicacole and Kalingapatam and the tftkik
of Palkonda are in the Telegn country ; the limit between
the Aryan and the Dravidian apparently is more to
the north, almost midway between Chicacole and Oanjatn
(cf. Linguistic Survey, IV, 577).



II

Note on Tosala and Dhauli

(Of. p. 71)i
Text ' of Gamiavyuha quoted on p. 69.

Gaocba kulaputrehaiva Dak^inftpathe Amitatosale
janapade Tosalaih nama nagaraih tatra Sarvag&ml

pari vrSjako prati vasati Yenam i tatosalo janapadas

tenopajagftmopetya Toealain nagaraih parimargan pari-
gave^am&DO 'nupurveoa Tosalaifa nagaram anupr&ptah
surySstarngamanakfile sa Tosalarh nSma nagaram
anupravidya madhye nagarasrngfi^akasya sthitvS vlthl-
mukhena vlthlmukham catvarena catvaraih rath) ay a
rathySm Sarvagraminani adrakslt ratryam prasantayftifa
Tosalasya nagarasyottare digbhage Surabharh nama
parvataib tasya dikhare vividhatruagulma-ausadhivanSrS-
maracite mahivabhasaprftptam bhSskaram ivoditaih tasya

tarn avabhSsaiii d^tva (MSS. 33, 36, 4] of the

Bibliotheque Nationale de Paris).

This text of the Gandavyuka locates the mountain

" Surabha to the north of the city of Tosala ; the Chinese

translation of Buddhabhadra (398-421 A. D.) follows the

text faithfully. But the translation of Sikg&nanda and

Pr&jfia differs and locates it to the east of the city. They

1 I am glad to be able to add this note from the account furnished
by my friend Mr. Nirmal Kumar Bose, M.A., who hag travelled in thia
tract on many ooeationa in connection with anthropological studies.
P. C. B.

1 Thii text ia omitted on p. 69 through a mistake of mine, P, C, B,



NOTE ON TOSALA AND DHAULI 177

translated the official MS. of the King of Orissa which
should be considered as more authoritative in this case.

If Tosala is identified with Dhauli or its immediate
neighbourhood then the mountain Surabha is to be
identified with the Dhauli hill (also called Dhavalagiri)
as it is the only hill in that tract. Dhauli is situated
to the south or south-east of Bhu vanes war at a distance
of about 5 miles. The usual route to be followed is the
metalled road from Bhuvaneswar to Puri, It passes by
an ancient site known as Sisupalgarh and traverses the
DaySnadi which is to be crossed by ferry. The Dhauli
hill is situated to the west of thin road, near the river.
To the south-east of the Dhauli hill there is a large pool
oNwater called Kausulla-gang, which is still fed by the
Dayanadi during the rainy season when the water flows
into it by a stretch of lowland to the north of the Dhauli
hill. It was in all probability a dam, made /from an old
river-bed, in which water was preserved for the time of
scarcity. There is still the mark of an embankment
which stretched from the foot of the hill southwards
along the Kausulla-gang. Admitting that the pool of
water was an artificial dam, it must have been used to
supply water to a neighbouring town which has now
disappeared, and was situated by the side of the embank-
ment. The boundaries of that town were probably the
river on the west, the Kausullfi-gang on the east and the
Dhauli hill on the north or north-east. It is difficult
to determine if the river has changed its course in recent
times. But a study of the map shows that the main
current now flows by the river Bargovi.

It should be pointed out that Asoka's inscription is
found on the Eastern side of the Dhauli hill (facing the
highroad from Bhuvaneswar to Puri). If this road,



178 PRE-ARYAN AND PRE-DRAVIDIAN

in ancient times, passed by the town we are trying to
locate, it had probably taken a more easterly direction
and passed by the side of the hill and the Kausulla-gang.

If ancient Tosali is located at the site indicated above
then the Dhauli hill stood on its north or north-east, as
the accounts of the Gandaryuha want us to believe, and
the Kausulla-gang was to its east. The word gang is
generally taken in the sense of river. Kamulla-gang
therefore seems to have been originally a branch of the
river. It may be suggested that Kausulla is most pro-
bably based on the word Kosala.



ADDITIONS AND CORRECTIONS



page line for

85 18 is

36 6-7 the sources of in-

formation available
21 the movement
28 (J.A*., 1923, 1, p. 135)

37 24 a family

38 18-20 "If we ... northern

side"



22 certainly has been
very recently ac-
quired

26 though it is

39 1 ' in a general way ' . . .

2-3 Insert "it has hap-
pened"
last line cerebrla

41 23-24 Omit not only " and

80 Aika ...

42 very recent

44 3 (from bottom) : " which

is to be"



read

seems to be

the present state of

our knowledge
movements
(but see /, As.,

1923, I, p. 186)
families
" we must look into

northern Dravi-

dian, if we wish to

understand the

history of southern

languages "
is concerned only

with their modern

state
though, generally

speaking, it is
to be omitted
before "through"

cerebral

" but goes farther

back."
aika

'comparatively
recent 1

"which has been"



180



ADDITIONS AND CORRECTIONS



45



12
14
19



for

"p. 199"
"for the "

"The only thing left
is to consider "



23-25 "To illustrate the
course of a parallel
evolution nothing
more is necessary "

25 Omit " than to mention
that"

46 11 "facts of a substra-

tum"

47 13 Kudirei ...
Kora ...
16 Kruta ...
21 km

2 (from bottom} : MSL

XIX, p. 89

last line fouud ...

17 KM ...
50 5 Add: " Further, in the

absence of other
proofs no borrowing
from Dravidian will
be admitted for
Indo-Aryan words
with aspirates."

21 'kalte ...

23 Add : (cf. Jules Block
Le nom du Biz :
Xftude* Asiattqucs)



read

" p, 179 "

for an "

" So what remains
to be compared
are only "

te Nothing more is
necessary to ex-
plain why in
course of a parallel
evolution ..."



" facts due to a sub-

stratum "
kudirei



truta

kui

Ind.Ant. XLVIII,

p. 195
found
Mi



Mte.



ADDITIONS AND CORRECTIONS



181



page


!* /or


read


51


2 MA5 ...


varkar





8 JZruott* ...


Ernout


,n


. 1, 5 Tibetan Loan Words ...


Loan Words in






Tibetan


52


22 melasu ...


menatu


58


18 vigu


vlgu


))


t?tPW ... ...


vtvu


99


3 (from bottom) : malt


namali


54


20 panjna ...


panjka




Add after fruit " ...


Oon. pandna, Kur.






panjna "to ripen."


55


3 " ordinary article ' ' ...


"article of ordinary






use




21-22 which the Afghan has


which is found only




borrowed


in Afghan


56


4 (from bottom) : oth


oth


57


4 "for it"


" for words with






aspirates "


99


12 mill


mm


JJ


13 migei


migei





16 Przylaski


Frzyluski


99


17 cramuh


wamnh


n. 1, 6 "facade in the back"


" the courtyard in






the backside of a






house "


58


9 (from bottom) :


"dialects of Dravi-




(< dialects, Dravi-


dian w




dian"






4 (from bottom) : Vljana


vljana


i


8 :Telegu


Telugu


59


17 "must be 11


"maybe 11


64


24 Tosala-Otkala


Tosal-Otkala


65


1 Avafiyaka ; -mijjutti


Avasyaka-nijjatti



182



ADDITIONS AND CORRECTIONS



page

66

66


67

68 n
69
78
82
90
91

94
97



103
107



120



line for

12 while presenting



... and presented



12 Athena gounon

7 Nondon
1,2 MihTra ...

22 Subhakaradeva

8 Forchhamer

1 MekalS...

8 (from bottom) : presents

2 the Munrjas the San-



Athenagouron

Nardon

Mihira

Subh

Forchhammer

MekalSh

prescribes

The Mundas and the

Santals

Amagha

Mekala

VairSta

KaudambT

up to the Ganges
c dt>ipah
Sulika
extend

Kamboja (and not
Kamboja)



122 12 of n. 1, Insert before "the charac-

' ter 1 ' and "is read"

after "which."

12S 13 (from bottom) : Karat- Kamboja
boja

124 9 the Island of Palks ... Easter Island

125 last line the best , ... better.



9 Amagabhutisa
5 Mekal ...

7 (from bottom): Vairata
9 Kau-

sSmbi

5 up the Ganges
4 Carmadvlpah
9 Sulika ...

8 (from bottom): extends

9 (from bottom) : 'Kam*

boja (and not Kam-
boja)



INDEX

OF

SANSKRIT AND OTHER INDO-ARYAN WORDS TRACED
TO THE AUSTRO-ASIATIC ORIGIN.

(Words in italics belong to modem vernacular*.)



Accba 97
alabu 157

alaba 157
AAga 72
Auijra 84
bftla(v&la) 6
baga 19
bfirui 18
baru(jivin) 18
barai 18
baroj 18
Carmarafiga 106
Dfimalipta 116
damaru 159
dtmeri 158
dhombe 160
(Jomba 160
fomva 160
dome 160
dombar 160
dombari 160
dambar 160
dnmbarn 160
dwnri 158
dumur 158



gara 147
gcxfomba 153
gulika 18n



gulma 18n
jim.nd 25
Kabuda 119
Eadali 4
kakkola 98
kaliftga 74
kalabgala 18 n-1
kambala 6
Kamboja 119
Kimbhoja 123
EamaraAga 108
Eftmala&kA 103112
Eamarupa 103, 112
kandali 5
Eapi^a 119
kara 147
Earmaraftga 104
karpftsa 23
karpata 24
Kerala 134
Kosala 63
Eoocba 97
kodambara 149
kuri 26
ku& 26
kuri 26
ku^ya 140
kuta 145
knra 147



184



INDEX



Kodnmbara 158

Iftbu 156

labuki 168

Iagu4a 12

laknta 12

laftgala 8

ISfigula 8

lafigala 11

laftgula. 11

Iftfigula 11

laftka 101

(Laftka)

liftga 8

iTAga 15

makuta 134

mfttafiga 129

maruka 131

marut 135

mayfira 181
mayuka 131

Mekala 81
134
86

Murala 131
Mara^dalft 134
Ojlra 34
Odombara 149
p&lak 6n-l
pa^a zvi
pa^a 24
Paufijra 85
Pai?(Ja 85
Paitfra 85
Polinda 88
galmala 8



dftlmall 8
atobala 7
simbala 7
dimbala 6
Bimball 7
tabavam 125
tambulam 16
tambul (Pili) 16
tambulam (Pkt) 16
tamburft 159 n-2
tambolam 16
tambol! 16
TamraliAga 111
TambraliAga 111
Tftmbalifiga 111
Tflmralipti 116
Tamalipti 116
Tamalinl 116
Tamali 118
T&malitti 118
Tamrapar^I 118
Tamluh 119
Takkola 98
Tosala 63
Triliftga 74
tumbS 153

tumbi 153
tambuka 153
Udnmbara 153
Ufra 84
Utkala 81
Uihija 84
Ui?(?a 84
Sra t 147




MAP OF DHAULI (Surv. Ind-n 73)
Scale I m. to an inch



SOME PUBLICATIONS ON LANGUAGE AND

LITERATUEE BY THE CALCUTTA

UNIVERSITY.

The Origin and Development of the Bengali
Language, by Sunitikumar Chatterji, M.A.,
D.Lit., with a foreword by Sir George Abraham
Grierson, K.C.I.E., I.C.S. (Retd.), Director
of the Linguistic Survey of India. In 2 vols.,
F'cap 4to, cloth-bound, uncut edges. Rs. 20.

Vol. I Introduction and Phonology, pp. i-xci,
1-648.

Vol. II Morphology, Additions and Corrections,
and Index of Bengali words, pp.
049-1179.

" This admirable work," says Sir George Grierson in
his Foreword, " . . . . is a fine, example of wide knowledge
and of scholarly research, " An invaluable contribu-
tion to scientific study of the Modern Indo- Aryan languages
as a whole.

" it will honour the University fliid Indian Scholarship very

much. It is the first book of that amplitude and depth devoted by an
Indian to an Indian language "Prof. Jules Bloch.

44 It is a work of extremely high importance and value,

and serving as a model for future researches in other languages of
India." Prof. L. D. Barnett.

44 it would be impossible for any European Scbolar to bring

out anything so full of information your penetration of the subject

i* admirable, and you have ina.slcred Western methoda to perfection."
Prof. Sten Konow.

14 The author is thoroughly familiar with the Western methods

(/ philological-historical investigation and at the same time has a
knowledge of linguistic facts which no European Scholar could ever
hope to acquire." Prof. M. Winternitz.

44 1 know of no work of this kind where the matter has been

so .fully mastered, so thoroughly searched, so clearly exposed." Prof.
Sylvain Le*vi.



(2)

History of Bengali Language, by Bijaychandra
Majumdar, B.L., Demy 8vo. pp. 323.
Bs. 7-0.

The book gives a sketch of the origin of the Bengali
Language and the various influences linguistic, ethnic,
social that shaped and moulded its earlier history.

History of Bengali Language and Literature, by

Eai Bahadur Dineschandra Sen, B.A., D.Litt.,
Demy 8vo. pp. 1067. Ks. 16-12.

A comprehensive view of the development of the
Bengali Language and Literature from the earliest times
down to 1850.

The Yaishnava Literature of Mediaeval Bengal,

by the same author, Demy 12mo. pp. 312.
Ee. 1-6.

The book contains a connected history of the influence
of Vaishnava Literature of the Mediaeval Age on the
development of Bengali Language.

Bengali Prose Style, by the same author, Demy 8vo.
pp. 184. Ks. 4-4.

The book throws light on the linguistic features of the
earliest period of the Modern Bengali Prose. It is a history
of the evolution of Modern Bengali Prose.

Early Bengali Prose, by S. K. Mitra, Demy 8vo.
pp. 184. Es. 3-0.

The book contains a few typical specimens of old Bengali
Prose which was written before the advent of British rule
and the establishment of the Printing Press in Bengal.

Bengali Literature in the Nineteenth Century,

by S. K. De, M.A., D.Litt., Demy 8vo.
pp. 530. Es. 8-6.

A historical review of the course of Bengali literature
with a background of social and political history. The



materials have been collected from sources hitherto in-
accessible to many.

The Origin of Bengali Script, by B. D. Banerjee,
M.A., Demy 8vo. pp. 122. Bs. 3-0.

A history of the development of the Bengali Alphabet
and a valuable contribution to Indian Paleography.

Origin of the Indian Alphabet, by D. B.

Bhandarkar, M.A., Ph.D., Royal 8vo. pp. 24.
Re. 1-2.

The Origin and Character of the Purana Litera-
ture, by B. C. Majtmidar, B.A., B.L., Royal
8vo. pp. 24. Be. 1-2.

Introduction of the Alphabet into Tibet, by

Mahamahopadhyaya Satishcbandra Vidya-
bhushana, M.A., Ph.D., Royal 8vo. pp. 0.
As. 12.

A Note on Sanskrit Compounds, by 1. J. 3.

Taraporewala, B.A., Ph.D., Royal 8vo. pp. 6.
As. 12.

Contamination in Language, by the same author,
Boyal 8vo. pp. 13. As. 12.

Pali, Prakrit and Sanskrit in Buddhist Litera-
ture, by Sailendranath Mitra, M.A., Boyal 8vo.
pp. 19. Be. 1-2.

No comments: