American Philosophical Society

Linguistic Prehistory of India

Author(s): Murray B. Emeneau
Source: Proceedings of the American Philosophical Society, Vol. 98, No. 4 (Aug. 16, 1954), pp. 282
-292
Published by: American Philosophical Society
Stable URL: http://www.jstor.org/stable/3143883
Accessed: 21-10-2015 10:23 UTC

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

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

American Philosophical Society is collaborating with JSTOR to digitize, preserve and extend access to Proceedings of the
American Philosophical Society.

http://www.jstor.org

This content downloaded from 157.182.150.22 on Wed, 21 Oct 2015 10:23:28 UTC
All use subject to JSTOR Terms and Conditions
 LINGUISTIC PREHISTORY OF INDIA
MURRAY B. EMENEAU
Professorof Sanskrit and General Linguistics,Universityof California, Berkeley
(Read April 23, 1954)

At some time in the second millenniumB.C., one of its interpreters3 has called it-"drab" or
probablycomparativelyearly in the millennium, "dull"4 mightbe a better,because a less subjec-
a band or bands of speakersofan Indo-European tive, word.
language, later to be called Sanskrit, entered For our presentinterest,however,it is to be
India over the northwestpasses. This is our recalled that this culture was literate. The
linguisticdoctrinewhich has been held now for many short inscriptionson seals are writtenin
morethan a centuryand a half. There seems to a script that is non-identicalin its formswith
be no reason to distrustthe argumentsforit, in either the Egyptian hieroglyphsor the Meso-
spite of the traditionalHindu ignoranceof any potamian cuneiform,but that is somewhatsim-
such invasion, their doctrine that Sanskrit is ilar to the earlystages of both in that it probably
"the language of the gods," and the somewhat developed froma picture-writing, rebus-likein
chauvinisticclingingto the old traditioneven nature, and was apparently,like both, syllabic
today by some Indian scholars. Sanskrit,"the in its structure.5 The unknownscriptand sub-
language of the gods," I shall thereforeassume ject-matter,combinedwith the shortnessof the
to have been a language broughtfromthe Near inscriptions,have so far defied convincingde-
East or theWesternworldby thenomadicbands. cipherment.
These invaders did not penetrate into a lin- The most ambitiousattempt up to date, that
guisticvacuum. What did theyfindto the east of Father Heras,6 has posited that the language
of the passes? If the archaeologists'reconstruc- is an old memberof the Dravidian familywhich
tion and chronologyof the events are at all is now located chieflyin South and Central India,
well-founded,they foundover the whole of the but whichstill has an outlyingmember,Brahui,
Indus Valley a highculture(the Harappa culture in the highlandsof Baluchistan to the west of
of the archaeologists),which was a siblingor a the Indus Valley. There is nothing a priori
remote cousin of the high cultures of the Near against thisassumption. Speakers of Dravidian
East-Egypt, the Fertile Cresent, Mesopo- languagesnow numberabout 90 millionsor just
tamia.1 Here weregreat and small cities,whose underone quarterofthe populationofthe Indian
life was based on the riverineagricultureper- subcontinent (making the family the fifthor
mitted by a great river and its tributariesin a sixth largest in the world). The geographical
vast, irrigable,alluvial valley. The probabili- distribution,and the nature of the boundary in
ties are that theculturewas as rigidlyregimented Central India between Dravidian speakers and
and as firmlybased on a systemof serfdomor the speakers of the Indo-Aryanlanguages that
slavery as were the Near Eastern cultures.2 So descend fromthe invaderlanguage Sanskrit,are
far the excavations have not yielded, and it good evidence that Dravidian has been steadily
seems improbablethat they will in futureyield, retreatingbeforeIndo-Aryan. The Dravidian-
evidence of a material brilliancesuch as that of speakers of the farthest South, the Tamil-
Egypt and Mesopotamia. It may be that the
Harappa culturewas the "repellent" thingthat 3Piggott,op. cit. 201.
"Dull" is Wheeler's word in Tamil culture2: 20, 1953;
1 See, e.g., Mackay, Ernest, Early Indian civilizations, he declares, however, in The Indus civilization,63 f. that
London, Luzac and Co., 1948; Wheeler, R. E. M. (now the remains suggest "an aesthetic capacity more broadly
Sir Mortimer), in various places, esp. Ancient India 3: based than the recovered examples of it alone would
74-76, Jan. 1947, and The Cambridgehistoryof India, indicate," an aesthetic capacity perhaps expended in part
supp. vol. The Indus civilization,Cambridge, Univ. Press, on the perishable art of wood-carving.
1953; Piggott, Stuart, Prehistoric India to 1000 B.C., 6 Piggott, op. cit. 178-181. Kroeber's term "stimulus
Pelican Book, Harmondsworth, Middlesex, 1950, with diffusion"must be called to mind.
good select bibliographies. 6 Many articles in Indian journals have been devoted to
2 Wheeler, AncientIndia, loc. cit. this "decipherment." It seems unnecessary to list them.
PROCEEDINGS OF THE AMERICAN PHILOSOPHICAL SOCIETY, VOL.98, No. 4, AUGUST, 1954.
282

This content downloaded from 157.182.150.22 on Wed, 21 Oct 2015 10:23:28 UTC
All use subject to JSTOR Terms and Conditions
VOL. 98, NO. 4, 1954] LINGUISTIC PREHISTORY OF INDIA 283

speakers, early acquired a high culturewith all arbitrary meanings, that has convinced the
the features of the North Indian culture that interestedscholars that the method used has
was framed in Sanskrit. Even the grammer been an unconvincingone-that, to put it more
and the literarycriticismof the Sanskriticcul- crudely,by thismethodone can give the inscrip-
ture were among the early borrowingsand are tions any interpretation one wishes.
the subjects of the earliest extant Tamil text. In fact,promisingas it has seemed to assume
As is usual in India, no exact dating of this early Dravidian membershipfor the Harappa lan-
literatureis possible, but it is clearly no later guage, it is not the only possibility. Professor
than the earliest centuriesof the Christian era W. Norman Brown has pointedout (The United
and may well be pre-Christian. Tamil scriptis States and India and Pakistan, 131-132, Cam-
known from the Arikamedu finds of the first bridge, Harvard University Press, 1953) that
century A.D. And the presence of Asokan in- NorthwestIndia, i.e. the Indus Valley and ad-
scriptionsin Middle Indo-Aryandialects as far joining parts of India, has during most of its
south as Mysore State in the thirdcenturyB.C. historyhad Near Eastern elementsin its political
is witness to the spread of Sanskritic culture and cultural make-up at least as prominently
southwardspriorto this period-for who could as it had true Indian elementsof the Gangetic
have been addressed by Asoka in Indo-Aryan and Southerntypes. The passage is so impor-
dialects if not literatebrahmanswho had settled tant that it is quoted in full:
in the South in some numberson theirmissionary
occasions? More ominous yet was another consideration.
Partitionnow would reproducean ancient,recur-
Be that as it may, it has been claimed that the ring,and sinister
incompatibilitybetweentheNorth-
far southern,Tamil form of the brahmanical west and the restof the subcontinent, which,but
Sanskritic culture of India shows in its earliest fora fewbriefperiodsof uneasycohabitation,had
remains so many specifichigh featuresthat are kept thempoliticallyapart or hostileand had ren-
not North Indian, that we must posit a high deredthesubcontinent weak. Whenan
defensively
culturein South India priorto the spread of the intrusivepeoplecame throughthepassesand estab-
brahmanical culture there. Archaeologicalevi- lisheditselfthere,it was at firstspiritually closerto
dence to substantiate these claims, however,is the relativesit had leftbehindthan to any group
still lacking.7 Yet, taking all together, the alreadyin India. Not untilit had been separated
assumptionthat the languageofthe Indus Valley fromthoserelativesfora fairlylongperiodand had
succeededin pushingeastwardwould it loosenthe
documentswas Dravidian is clearlynot fantastic. externalties. In
periodafterperiodthisseemsto
In spite of this,the attemptsat interpretation have been true. In the thirdmillennium B.C. the
have not been convincing. Since the subject- Harappa culturein the Indus Valley was partly
matterof the documentsis unknown,any sub- similarto contemporary westernAsian civilizations
ject-matter,i.e. meaning,that is assignedto any and partlyto later historicIndian cultureof the
elementof the script,is arbitrary. A succession Gangesvalley. In the latterpartof the nextmil-
of arbitrarymeaningsthus assigned to the sym- lenniumthe earliestAryans,livingin the Punjab
bols (with,to be sure,a littleaid fromSumerian and composingthe hymnsof the Rig Veda, were
texts) may, in fact, be made to coincide with a apparentlymorelike theirlinguisticand religious
Dravidian-likesuccessionof phonemes,but only kinsmen, theIranians,thanliketheireasternIndian
contemporaries.In the middleof the nextmillen-
for short stretchesof the material. The Indus nium the Persian
Achaemeniansfor two centuries
Valley material is made up of a fairly large held the Northwestas satrapies. AfterAlexander
number of short inscriptions,but it is only by had invadedIndia (327/6-325B.C.) and Hellenism
positingnumerousarbitraryvariationsof sound had arisen,the Northwesttoo was Hellenized,and
and of meaningforthe same symbolsthat any- once morewas partlyIndian and partlywestern.
thinglike a coherentsystemcan be made to run And afterIslam enteredIndia,the Northwest again
was associatedwithPersia,Bokhara,CentralAsia,
through the total corpus. It is this necessity rather thanwithIndia,and considered itselfIslamic
to vary, or, in other words, this self-givenper- firstand Indiansecond.
missionto vary in the sound and meaningthat The periodsduringwhichthe Punjab has been
are attached to the smallishnumberof different culturallyassimilatedto the restof northern India
symbols, plus the impossibilityof checkingthe are fewif any at all. Periodsof politicalassimila-
tionare almostas few;perhapsa partof the fourth
IWheeler, Ancient India 4: 89, July 1947-Jan. 1948, and thirdcenturies
B.C. undertheMauryas;possibly
with some bibliography. a briefperiodunderthe Indo-GreekkingMenander

This content downloaded from 157.182.150.22 on Wed, 21 Oct 2015 10:23:28 UTC
All use subject to JSTOR Terms and Conditions
284 MURRAY B. EMENEAU
in the second centuryB.C.; anotherbriefperiod
underthe Kushanasin the firstand secondcentury
A.D.; an even brieferperiodunderthe Muslimking-
dom of Delhi in the last quarterof the twelfthcen-
turyA.D.; a longone underthegreatMughalsin the
sixteenthand seventeenth centuriesA.D.; a century
underthe British,1849-1947.
Though this refersto cultural and political
factors,it is a warningthat we must not leap
to linguisticconclusionstoo hastily. The early,
but probably centuries-longcondition in which
Sanskrit,a close ally of languages of Iran, was
restrictedto the northwest(though it was not
the only language there) and the rest of India
was not Sanskritic in speech, may well have
been mirroredearlier by a period when some
other language invader fromthe Near East-a
relativeofSumerianor of Elamiticor what not-
was spoken and written in the Indus Valley.
It is notruledout,ofcourse,that thisanonymous
Near Easterner might have been only one of
the languages spoken in the Indus Valley-
perhaps that of invadersand conquerors-while
the indigenous population spoke another lan-
guage-perhaps one of the Dravidian stock, or
perhaps one of the Munda stock, which is now
representedonly by a handfulof languages in
the backwoodsof Central India.
On leaving this highly speculative question,
we can move on to an examinationof the San-
skrit records, and we find in them linguistic
evidence of contacts between the Sanskrit-
speakinginvadersand the otherlinguisticgroups
withinIndia.
Whenevertwo language communitiescome in
contact and remain in contact for any appre-
ciably long period, the languages have some
effectupon each other's structure. Borrowing
of wordsin one or the otherdirectionor in both
is the most obvious effect. But there may also
be a shiftof sound systems,borrowingof deriva-
tional or inflectionalmorphemes,or borrowing
of syntacticaltraits.
Sanskrit scholarship in the West soon saw
that some of the non-Indo-Europeanfeaturesof
Sanskritwere Dravidian (or possiblyMunda) in
type. The retroflex(domal or cerebral) con-
sonants in Sanskrit may be explained for some
of their occurrences as being the reflexesof
Indo-European consonant clusters of certain
types. The fact, however, that the later in
Indo-Aryanlinguistichistorywe go, the greater
is the incidenceof retroflexconsonants,and the
further factthatmostofthe Dravidian languages
This content downloaded from 157.182.150.22 on Wed, 21 Oct 2015 10:23:28 UTC
All use subject to JSTOR Terms and Conditions
[PROC. AMER. PHIL. SOC.
and Proto-Dravidian itself have this type of
consonantin abundance (the case is not so clear
for Munda, but is in all probabilitysimilar),
can only lead to the conclusion that the later
Indo-Aryandevelopmentsare due to a borrow-
ing of indigenous speech habits throughbilin-
gualism,and to the well-groundedsuspicionthat
even the early developmentof retroflexesfrom
certainIndo-Europeanconsonantclustersresults
fromthe same historiccause.' The same argu-
ment applies also to the developmentof absolu-
tives (otherwisecalled gerunds)in Sanskrit; this
non-finiteverb formand its syntacticuse are so
closely parallel to a feature of Dravidian and
so unlike what is found in the other old Indo-
European languagesthat we mustcertainlyposit
Dravidian influencehere.9
Prima facie, it should have been easy to ex-
amine the Sanskrit dictionaryfor possible bor-
rowingsfromDravidian. There were,however,
severalblockingfactors. Firstand perhapsmost
importantwas the assumption,usually but not
always only implicitlymade and seldom argued
or supported by evidence, that the Sanskrit-
speaking invaders of Northwest India were
people of a high,or better,a virile,culture,who
foundin India only culturallyfeeblebarbarians,
and that consequentlythe borrowingsthat pat-
ently took place fromSanskrit and later Indo-
Aryanlanguagesinto Dravidian werenecessarily
the only borrowingsthat could have occurred.
Indian civilizationitself,with its enthronement
of Sanskrit at the expense of other languages,
taughtWesternscholarsto thinkthisway about
Sanskrit. Moreover, the early days of Indo-
European scholarship were without benefitof
the spectacular archaeological discoveries that
werelaterto be made in the Mediterraneanarea,
Mesopotamia, and the Indus Valley. It was
but natural to operate with the hidden,but
anachronistic, assumption that the earliest
speakers of Indo-European languages were like
8 This doctrineis held by, e.g., Bloch, Jules, Sanskrit et
dravidien,Bull. Soc. Ling. Paris. 25: 1-21, esp. 4-6, 1925;
Some problems of Indo-Aryan philology, Bull. Schoot
Orient.Stud. 5: 731-733, 1930; L'Indo-aryen du Veda aux
temps modernes,53 ff.,325, Paris, 1934; Katre, S. M.,
Some problemsof historicallinguisticsin Indo-Aryan, 135
if., Bombay, 1944; Prokosch, E., A comparativeGermanic
grammar, 39, Philadelphia, 1939; Wackernagel, Jakob,
AltindischeGrammatik1: 165, ?144(a) Anm., Gottingen,
1896, with earlier bibliography,Gundert in 1869, Ztschr.
deutsch. morgenldndischen Ges. 23: 517 ff., having appa-
rentlythe earliest suggestion.
9 Cf. Bloch, Bull. School Orient.Stud. 5: 733-735, 1930.
VOL. 98, NO. 4, 19541 LINGUISTIC PREHISTORY OF INDIA 285

the classical Greeks or Romans-prosperous, The third blocking factor has been for long
urbanized bearersof a high civilizationdestined the general ignoranceof and indifference to the
in its laterphases to conquerall Europe and then Dravidian languages, even among professed
a great part of the earth-rather than to recog- Indological linguisticscholars. They must not,
nize them for what they doubtless were-no- ofcourse,be judged too harshly. The Dravidian
madic,barbarouslootersand cattle-reivers whose languages are not easy, most of them are lan-
fateit was throughthe centuriesto disruptolder guages spoken by backwoods "primitives"and
civilizationsbut to be civilized by them. This are badly reported,the four literarylanguages
was in all probabilitythe event in India as it have enormousliteratures;ars (etscientia)longa,
was in Greece or in the late Roman Empire.'0 vita brevis.
This assumptionled in the long run to another Finally,a fourthblockingfactorhas been the
block-the methodologicaltendencyof the end generalcaution of Indo-European scholarswhen
of the nineteenthand the beginningof the twen- confrontedwitha substratumsituation. In this
tieth centuryto attempt to findIndo-European they were justified,forapart fromBasque, the
etymologiesfor the greatest possible portionof substrata that have been operated with in the
the vocabulariesof the Indo-Europeanlanguages studyof the historyof European languageshave
(see also note 10), even thoughthe object could been languages that are fragmentarily recorded
only be achieved by flightsof phonologicaland and badly known(e.g. Gaulish Celtic), not really
semanticfancy. Latin perhapswas the greatest on record (e.g. Illyrian),or not understood(e.g.
suffererfromthis urge," but none of the lan- Etruscan). In the case ofSanskrit,however,the
guages was exempt,and Sanskritwas no excep- Dravidian substratumis easily accessible in its
tion. It was the less pardonablein dealingwith dozen or more living languages and in that a
a language spoken and writtenin India, where Proto-Dravidian can be worked out, given
even casual inspection of the Dravidian lan- enough scholarsinterestedin the matter.
guages would have suggestedsome borrowings, The end resultof the block,however,was that
at least, fromDravidian into Sanskrit."2 very fewscholarsattemptedto identifyborrow-
ings from Dravidian into Sanskrit; those who
10 Mingled with this essentiallyEurope-centeredattitude were interested worked unmethodically and
was that other strand in late eighteenthcenturythinking, without establishingcriteria for recognitionof
the "romantic" one stressed by Mayrhofer, Manfred,
Saeculum 2: 54 ff.,1951. It led to Sanskrit being regarded probable, possible, and unlikelyexamples, and
(to use Mayrhofer's phrase) as "die ur-ste aller Ur- their results were universally ignored. The
sprachen," as an exemplar of purityand freedomfromall Sanskrit etymologicaldictionaryof Uhlenbeck
non-Indo-European influence. The Indo-European sav-
ages, in short,were the noblest of all noble savages. But (1898-1899)3 and the Indo-European etymo-
this, of course, is ethnocentrismall over again. logical dictionaryof Walde and Pokorny (1930-
11Cf. Lane, George S., Lang. 25: 335-337, 1949. 1932)14 completelyignore the work of Gundert
12 The borrowings from the vernacular Middle Indo-
Aryan into Sanskrit, when the latter became, as it early
(1869),15 Kittel (1872, 1894), 16 and Caldwell
did, a hieraticand then a dead language in which speakers (1856,i 1875).17
of Middle Indo-Aryan languages composed freely,form
another portion of the Sanskrit lexicon which must be as Tedesco does. But the Sanskrit "garland" word may
identifiedand not subjected to fantastic Indo-European well be borrowedfromthe Dravidian group of words given
etymologizing. Paul Tedesco has made many notable in Appendix 1, etym. 13, which see fordetails.
contributionshere, but he also is willing to operate with 13 Uhlenbeck, C. C., Kurzgefasstesetymologisches
Werter-
a methodology in which the Dravidian languages do not buch der altindischen Sprache, Amsterdam, 1898-1899.
exist except as borrowersfromIndo-Aryan. One example He mentions non-Sanskritic languages, though without
of this is his treatmentof words for "belly, stomach" in any detail, s.vv. dravidIi,palli. Mayrhoferhas apparently
Sanskrit pitaka- "basket," Archaeologica orientalia in found several more such instances; Saeculum 2: 55, 1951:
memoriamErnst Herzfeld,218, Locust Valley, N. Y., 1952. "verzeichnetkaum ffinfW6rtereinheimischerAbkunft."
Whatever may be the etymologyof Br. pid "belly, stom- 14 Walde, Alois, and Julius Pokorny, Vergleichendes
ach" (so with Bray, rather than p[h]ij with Linguistic Worterbuchder indogermanischenSprachen, 3 v., Berlin,
Survey of India), the Go. pir id. is rather to be inter- 1930-1932.
preted as going with Te. pEgu, pregu "entrail, gut, bowel" 15 Gundert,H., Ztschr.deutsch.morgenldndischen Ges. 23:
than as borrowed from an Indo-Aryan *petta-. More 517-530, 1869.
striking is the Sanskrit word mdld- "garland, wreath," 16 Kittel, F., Indian Antiquary 1: 235-239, 1872; A
which can be provided with an Indo-Aryan etymology Kannada-English dictionary,xiv-xlv, Mangalore, 1894.
only with the utmost ingenuity (Jour. Amer. Orient.Soc. 17 Caldwell, Robert, Comparativegrammar of the Dra-

67: 85 ff.,1947). The "rope" meanings in modern Indic vidian languages,565-579, 2d ed., 1875. I have no access
vernaculars may belong to homonyms,to be etymologized to the 1st ed. of 1856.

This content downloaded from 157.182.150.22 on Wed, 21 Oct 2015 10:23:28 UTC
All use subject to JSTOR Terms and Conditions
286 MURRAY B. EMENEAU
More recentworkby JulesBloch (1925, 1930, wordsof much moregeneralsemanticrangealso
1934)18 attempted to salvage some items from occur, some (e.g. nira- "water," etym. 6) as
the early attempts,and in the 40's T. Burrow
in an importantseries of articles (1945, 1946,
1948)'9attemptedto set up methodologicalprin-
ciples (see Appendix 1) and suggestedDravidian
sources for some five hundred Sanskrit words.
Collaborative work by Burrowand myselfon a
Dravidian etymologicaldictionarywilladd more
items. The Sanskrit etymological dictionary
that ManfredMayrhoferhas just begun to pub-
lish in Germany (1953)2? takes account of this
recentwork.
It is clear that not all of Burrow's suggested
borrowingswill stand the test even of his own
principles.21Much labor will have to be ex-
pended by qualifiedscholarson methodsand on
the individual items, but it can be safely pre-
dicted that the work of these modern scholars
will yield an acceptable residuethat will at long
last be available for studies of Indic linguistic
prehistory.
It is possible already to take some tentative
steps.
As was to be expected,many of the borrowed
itemsare names of floraand fauna indigenousin
India and not elsewherein theold Indo-European
territory, e.g. the fragrantscrew-pine,Pandanus
odoratissimus(App. 1, etym. 1), the cardamom
(etym. 2), the house lizard,Lacerta gecko(etym.
3), perhaps the peacock (etym.4), certainlythe
white ant or termite(etym. 5). Some of these
words occur in Sanskritliteraturefromthe epic
on; others are hardly more than items in the
lexica (perhaps intendedas etyma fromwhich
the Dravidian words were to be derived). But
18 Bloch, Jules, Sanskrit et dravidien, Bull. Soc. Ling.
Paris 25: 1-21, 1925; Some problems of Indo-Aryan
philology: II. Indo-Aryan and Dravidian, Bull. School
Orient.Stud. 5: 730-744, 1930; L'Indo-aryen du Veda aux
tempsmodernes,322-328, Paris, 1934.
19Burrow,T., Some Dravidian words in Sanskrit, Trans.
Philol. Soc. 1945: 79-120; Loanwords in Sanskrit, ibid.
1946: 1-30; Dravidian studies VII: Further Dravidian
words in Sanskrit,Bull. School Orient.and AfricanStud. 12:
365-396, 1948.
20 Mayrhofer, Manfred, Kurzgefasstes etymologisches
Wirterbuchdes Altindischen,1 und 2. Lieferung,Heidel-
berg, Carl Winter, 1953, 1954.
21 One should certainlyrule out the Dravidian etymolo-
gies given forkuiaru- and vrfs-,whichoccur in one passage
each in the Rig-Veda and nowhere else in the whole of
Sanskrit literature and are of highly uncertain meaning
(Trans. Philol. Soc. 1946: 22-23). To provide these with
Dravidian etymologiesis as futileas to provide them with
Indo-European etymologies,which Uhlenbeck did for the
former(followingRoth, presumably).
This content downloaded from 157.182.150.22 on Wed, 21 Oct 2015 10:23:28 UTC
All use subject to JSTOR Terms and Conditions
[PROC.AMER. PHIL. SOC.
early as the epic. We can be sure of Dravidian
originfora fewpropernames; e.g. the late Vedic
and epic hero Nala seems to have a Dravidian
name "the good (or handsome) man," and the
southern Malaya mountains have a name de-
rived fromthe Dravidian *malay "mountain"
(etymologies7 and 8). Even the interjection
aye of the dramas probably has a Dravidian
origin (etym. 9). It is unexpectedto findthat
kal- "an art" (epic and later), which has not
been provided with an Indo-European etymol-
ogv, is probably derived from the Dravidian
verb kal- "learn," but the suggestioncannot be
rejectedout of hand and in fact is not unattrac-
tive when we rememberthat India had a high
culturebeforethe Sanskritspeakersarrived.
The greatestinterestattachesto the itemsthat
occur in the earliestSanskritrecorded,the Rig-
Veda. Burrowfindssome twentywords,a very
mixed lot includingthe "peacock" word men-
tioned before and one or two other labels for
specificallyIndian phenomena. Most of them,
however, are much more general; e.g. khala-
"threshingfloor" (etym. 10) and phala- "fruit"
(etym. 11). kAna-"one-eyed"is veryobviously
derived fromthe negativeadjective ("who does
not see") of the Dravidian verb *kan- "see. '2
Most strikingly,bala- "strength," which has
been one of the more certainexamples given in
the handbooks for the rare Indo-European
phoneme *b (though the meanings of the cog-
nates are not really identical), may be derived
from the very wide-spreadingand ramifying
Dravidian family of val- "be strong, strong,
strength" (etym. 12); one of the languages in
which Proto-Dravidian *v becomes b will have
to be involvedhere. Indo-Europeanistsmay be
inclinednot to accept this, unless they are al-
ready desirousof gettingrid of all examples of
the rare Indo-European *b.
If the Rig-Vedic examples, or any of them,
22 The word occurs in Rig-Veda 10.155.1, an Atharvanic
charm. It is common in later Sanskrit and has an abun-
dant progeny("blind" and "blind of one eye") in the Indo-
Aryan vernaculars; see Turner, Ralph Lilley, A compara-
tive and etymologicaldictionary of the Nepali language,
London, 1931, s.v. kdnu. Connection with an Indo-
European root (s)qer- "to cut" or with qel- "to stick" is
not obviously convincing. The Dravidian etymology is
in Burrow, Trans. Philol. Soc. 1946: 22; earlier Gundert,
Ztschr.deutsch.morgenldndischen Ges. 23: 521. Not to be
derived with Kuiper, F. B. J., Proto-Munda wordsin Sans-
krit,52, froma Proto-Munda *ga-da "defective."
VOL. 98, NO. 4, 1954] LINGUISTIC PREHISTORY OF INDIA
are accepted, this is evidence forthe presenceof
Dravidian speakersas fartowardsthe northwest
as the Panjab, i.e. the upper Indus Valley, in
the firstcenturies(it is uncertainhow many) of
the presence of Sanskrit-speakerson Indian
soil.23 It is not entirelyclear evidence for the
23 This thesis automatically rules out a recenthypothesis
formulated by C. von Ftirer-Haimendorfin Proceedings
of the37th Indian Science Congress,Poona, 1950, Part II:
Presidential Addresses, 175-189, esp. 176-180, and New
aspects of the Dravidian problem, Tamil culture2: 127-
135, 1953. He bases his arguments on the results of
excavation at Brahmagiri, in the Chitaldrug District of
Mysore State, by Wheeler in 1947 (Ancient India 4: 180-
310, July 1947-January 1948; for the chronology of the
site, esp. 200-202). The prior establishmentof a precise
chronologyat Arikamedu (near Pondicherry)by Wheeler
(Ancient India 2: 17-124, July 1946) allowed a cross-
dating for the Brahmagiri site and the establishment of
this tentative chronology: Brahmagiri Stone Ax culture
down to the beginning of the second century B.C., "con-
tinuing as a dwindling sub-culture through most of the
succeeding Megalith phase"; megalithic culture "after c.
200 B.C. to the middle of the firstcenturyA.D., overlapping
the Andhra culture"; Andhra culture fromabout middle
of the firstcentury A.D. to third century. The date for
the beginning of the megalithic culture is an extremely
tentative one. Wheeler pointed out, moreover, that a
specific feature of the megalithic culture at Brahmagiri
was the use of cists with "port-holes," that this site is
near the northernlimit of reportedsites with this feature
(Hyderabad is so far the farthestnorth) but that they are
found in considerable numbers south to the tip of the
peninsula, and that if it were not forthe "wide and formid-
able disparity in date between the Indian cists and their
Western analogues [Western Europe, 2500-1500 B.C.], a
significantinterrelationshipcould scarcely be questioned"
(183). He added a reference to an early, unverified
report of similar cists in the neighborhood of Karachi
(301), desiderated a furthersearch for megalithsin North
India, and proceeded to speculate (303-304) on a possible
Western origin for the Indian phenomena, with (so far
as I can see) a plain warning that Iron Age megaliths of
the Indian type are not forthcomingin the Western world
at the right period and in the right place, but with the
assumption of "the possibility of an integral connection
between the port-holedcists of India and those of western
Asia and Europe, in spite of the wide disparities of time
and place" and an unwillingnessto envisage independent
origin of the port-holedcists "in regions which, however
far apart, have long been interconnectedby sea."
Furer-Haimendorfin the two publications cited at the
beginningof this note, and with more certainty and detail
in the second than in the first,accepted with hardly a
reservationthe hypothesisthat the megalithicculture was
brought to South India from outside India towards the
middle of the firstmillenniumB.C. He findsit impossible
to conceive that the megalithic, iron-usingculture might
have been a local development in some area of South
India still unexplored archaeologically (as most of South
India still is), with some stimulus perhaps from outside;
I, not being an archaeologist or an anthropologist, am
unable to combat his point of view with any authority.
However, I am unable to accept his furtherhypothesis
This content downloaded from 157.182.150.22 on Wed, 21 Oct 2015 10:23:28 UTC
All use subject to JSTOR Terms and Conditions
287
Dravidian nature of the Harappd language or
of one of the Harappa languages; it does, how-
ever, lead towardsthat hypothesis.
If the Munda languages have been mentioned
only in passing, it is not because there were no
contacts between Munda speakers and those of
the othertwo language families. The reason is
that we know so littlethat is certainabout the
Munda family,eitherdescriptivelyor compara-
tively. Even the most ambitious attempts at
demonstratingborrowings from Munda into
Sanskrit have been based on fantasticallyun-
sound methods. Although certain Sanskrit
words with no Indo-European or Dravidian
etymologiesmightbe borrowingsfromMunda,
it seems sounder to ignorethe suggestionsthat
have been made, until clear criteriaforjudging
themare at hand (see furtherAppendix 2).
APPENDIX 1
SANSKRIT BORROWINGS FROM DRAVIDIAN
T. Burrow in Trans. Philol. Soc. 1946: 13-18
set forthcriteriaforidentifying
Dravidian words
borrowed into Sanskrit. I summarize them,
that the Dravidian languages, or their ancestor, were
introduced into India by the bearers of the megalithic
culture. He finds it impossible that the people with an
iron-using,megalithic culture and the neolithic peoples
overrun,could have spoken languages of the same family.
But this a priori position,that peoples of differentcultures
cannot be linguisticallyrelated, is contradictedagain and
again in the Indo-European and in other fields-.g. the
relations of Latin, Celtic, and Germanic speakers within
the Indo-European family in the time of Caesar and
Tacitus are notorious. The general linguistic position
taken by Ftirer-Haimendorfis untenable, and with it
falls his contention that North India could never have
been Dravidian-speaking. To rule out the negative does
not, to be sure, provide evidence that North India was
ever Dravidian-speaking. But if Dravidian loanwords
are found in the Rig-Veda, this is positive evidence which
fillsthis particularvacuum most satisfactorily.
It is regrettablethat Manfred Mayrhoferwas persuaded
(Die Substrattheorien und das Indische, Germanisch-
romanischeMonatsschrift34: 233-234 and fn. 9, 1953) by
this tenuous archaeological hypothesis to abandon the
position that he had taken earlier (Arische Landnahme
und indische Altbevolkerungim Spiegel der altindischen
Sprache, Saeculum 2: 54-64, 1951) on solid linguisticevi-
dence. This is not to say that one can agree with all
details of Mayrhofer'searlier construction (e.g. too much
weightwas put on less than satisfactoryMunda evidence),
but at least in his theory Dravidian speakers were among
the early peoples met by the invading Indo-European
speakers. It is not quite clear what position he takes in
the introductionto his Kurzgefasstesetymologisches Wdrter-
buch des Altindischen(1953), but certainly he holds that
there are Dravidian loanwords to be found in the Vedas
(10).
288 MURRAY B. EMENEAU [PROC. AMER. PHIL. SOC.

with some comment,as follows: (1) the Sanskrit simus": Skt. ketaka-, ketaki- id. Burrow, T.,
word should have no Indo-European etymology Trans. Philol. Soc. 1946: 16: "the diphthongai
-it should rather be: no certain or obvious in the Tamil and Malayalam wordsis an indica-
Indo-European etymology; (2) there should be tion that the word is originallyDravidian"; so
a wide currencyof the etymonin the Dravidian also the suffix-ai in Ta.-Ma. and Tu.
languages and it should be a basic element in 2. Ta. elam "cardamom plant,Elettariacarda-
the vocabulary; (3) "a word is shown to be momum"; ela-v-arici "cardamom seed"; Ma.
Dravidian in originif it is clearlyto be derived elam "cardamoms"; elatt-ari"cardamom seed";
from some Dravidian root" (e.g. candana- Ka. el-akki,yal-akki,yalaki "large cardamoms";
"sandalwood tree, ointment" is Dravidian in Kod.. e-l-akki "cardamom seeds"; e la male,
originsince the correspondingDravidian nouns e lati male "cardamom plantation"; Tu. el-akki
cdntu, etc. are specialized derivativesof a verb "cardamoms"; Te. ela, elaki "cardamomplant";
meaning "to rub into a paste"); (4) the word elakulu "cardamom seeds": Skt. ela "carda-
should be of some antiquity in Dravidian (e.g. mom." Kittel, no. 85.
occurringin the earliestTamil texts); (5) com- 3. Ta. palli; Ma. palli; Ko. e-paj; To. pasy;
parative lateness of appearance of the word in Ka. palli; Kod. palli; Tu. palli; Te. palli,balli
Sanskrit (or, one may add, appearance only or "house lizard, Lacerta gecko": Skt. palli, pallika
firstin Sanskritlexica) increasesthe probability id. (lexical); Mar. palli, popular pal. Kittel,no.
of its beinga borrowing;(6) in each case possible 59; Emeneau, M. B., Univ.of Calif. Publ. Class.
phoneticcriteriashould be looked for; (7) like- Philol. 12: 261, fn. 31, 1943; Burrow,T., Trans.
wise, semantic developmentscan sometimesbe Philol. Soc. 1946: 10.
taken as criteria. There should be added, pos- 4. Ta. mayil, maninai;Ma. mayil; Ko. mil;
sibly as a corollaryto (2), that if the word de- To. mi s; Ka. mayla, maylu; Kod. mayl; Tu.
notes something peculiar to the Indian geo- mairu; Pa. manjil, maniil;011. manigil;Go. mal;
graphical or social scene, a Dravidian origin is Kui medu, melu; Kuwi mellu "peacock": Skt.
moreprobable than an Indo-European one. mayura- id. (in two compounds in Rigveda
Not all these criteriacan be broughtto bear 3.45.1, 8.1.25), mayfiru-"peahen" (Rigveda
in all cases. Comparativesimplicityand avoid- 1.191.14, an Atharvanic charm). Burrow, T.,
ance of the assumptionof tortuousphonological Bull. School Orient.and African Stud. 11: 608-
and semanticdevelopmentsshould also be aimed 610, 1945. He discusses Przyluski's Austro-
at, followingthegeneralpracticeofall disciplines Asiatic suggestion (Bull. Soc. Ling. Paris 26:
("Occam's razor"), and may well at times tip 99 f., 1925) and shows that Skt. mayura- is
the scales forborrowingfromDravidian rather closer in formto the Dravidian series than to
than foran Indo-European etymologythat has the posited Austro-Asiatic*marak or the like.
been suggested. He notes also that r in mayuira-instead of I is
The etymologies that follow have been re- not entirely unexpected in Rigvedic Sanskrit.
ferredto in the body of this paper. The sigilla Further discussion by Jules Bloch, Bull. Soc.
for the various languages and the bibliography Ling. Paris 25: 16, 1925. The suggestion
are those given in Lang. 21: 184, fn. 1, 1945, (Walde-Pokorny2: 243) of an Indo-European
and Bull. School Orient.and African Stud. 15: etymologyon the basis of connexionwith Skt.
98, fn. 1, 1953. The followingshort list gives mimati "bellow, bleat," mayuf-n. "bellowing,"
the Dravidian languages in the order (essen- Gk. IALAhl~c "neigh," etc. is not convincing.
tiallygeographic)in whichtheyare quoted, plus 5. Ta. purru, purram "white anthill"; Ma.
two other sigilla used in the etymologies:Ta. purru "ground thrownup by moles,rats, esp. a
= Tamil, Ma. = Malayalam, Ko. = Kota, To. whiteanthill"; Ka. puttu,putta "whiteanthill";
= Toda, Ka. = Kannada (Canarese), Kod. Kod. putti id.; Tu. pufica id., "snake's hole";
= Kod.agu (Coorg), Tu. = Tuju, Te. = Telugu, Te. putta "anthill, snake's hole, heap, lot,
Kol. Kolami, Nk. = Naiki, Pa. = Parji, 011. crowd"; Kol. (Kin.) putta "white anthill"; Nk.
= Ollari, Go. = Gondi, Kui, Kuwi, Kur. = putta id.; Pa. putkal, (NE) putkal id., (S) putta
Kurukh, Malt. = Malto, Br. = Brahui; Skt. "nest insideanthill"; 011.putkal "whiteanthill";
= Sanskrit,Mar. = Marathi. Go. putti id.; Kui pusi id.; Kuwi [F] puci "ant-
1. Ta. kaital, kaitai; Ma. kaita; Ka. kWdage, hill," punja "ants' nest (earthen)"; Kur. putta
kedige; Tu. kedai, kedayi, kWdayi;Te. geddgi "anthill"; putbelo "white ants' queen" (belI
"the fragrant screw-pine,Pandanus odoratis- id.); MVIalt.pute "anthill": Skt. puttika- "the

This content downloaded from 157.182.150.22 on Wed, 21 Oct 2015 10:23:28 UTC
All use subject to JSTOR Terms and Conditions
VOL. 98, NO. 4, 1954] LINGUISTIC PREHISTORY OF INDIA 289

white ant or termite"; pipilika-puta-"anthill" food and prosperity";na pal "teeth that grow
(pipilika- "ant"). Burrow, T., Trans. Philol. straightand regular"; To. nas, nasaO"beauty",
Soc. 1945: 111. Skt. puttikd-has been taken naso-n n. pr. of a man; Ka. nal, nalme "good-
to be a meaning development from puttikd- ness"; nalla "a good man; goodness,excellence,
"doll, puppet" by Boehtlingk-Rothand Uhlen- beauty"; KocJ. nalle "good"; Tu. nalu, nala
beck ("das puppendhnlichetier") and Monier- "good, cheap"; nalme "goodness, friendship";
Williams ("so called fromits doll-like form"); Te. naluvu "beauty, ability; beautiful"; Go.
this is clearlyad hoc. Proto-Dravidian *rrsur- (M) nelh "good": Skt. Nala- n. pr. of a man.
vives in Ta.-Ma. but develops in two different Emeneau, M. B., Univ. of Calif. Publ. Class.
ways (tt, q) in differentlanguages. Sanskrit Philol. 12: 255-262, 1943; this still seems to me
has borrowedfromtwo different Dravidian lan- to be a good etymology. Manfred Mayrhofer
guages, giving puta- and *putta-,both meaning agrees,SymbolaeHrozn' 5: 371, 1950.
"anthill" or more specifically"white anthill," 8. Ta. malai "hill, mountain"; Ma. mala
fromthe latter of which puttika- is derived by "mountain"; Ko. mal im "buffaloesof the Nilgiri
a Sanskritformativesuffix. tribes (i.e. mountainbuffaloes)"; mal aV "high
6. Ta. nir "water, sea, juice, liquor, urine, downs on westernhalf of Nilgiri plateau"; To.
dampness, moisture"; (nirpp-, nirtt-) "become mas o r id.; Ka. male "mountain,forest"; Kod.
thin or watery (as liquid food in cooking), be male "thick jungle land, cardamom plantation
wet, moist"; nirmai"propertyof water,as cold- in jungle on mountain"; Tu. male "forest,hill
ness"; ir "moisture,wetness"; iram "wet, mois- overgrownwith forest"; Te. mala "mountain";
ture,freshness, coolness"; irali (iralipp-,iralitt-), Kol. ma le "hill"; Pa. mala?7g"forest"; Br. mash
ri (iripp-,iritt-)"become moist,damp"; irippu "hill, mountain": Skt. malaya- "the mountains
"dampness"; iriya "damp, wet, cold"; Ma. nir whicb borderMalabar on the east (i.e. Western
"water, juice, moisture"; iram "moisture,dirt"; Ghats)" (epic +); "a celestial grove (Nandana-
irikka"grow damp"; irmam,irman,iran "damp vana)" (lexical); "a garden" (lexical); mdla-
cloth"; Ko. ni r "water"; To. ni-r"water"; i-rm "forest near a village; field" (lexical). Kittel,
"dampness"; Ka. nir "water"; ira "moisture, no. 164; also Gundert and Caldwell. The San-
dampness, wetness"; Kod. ni ri "water"; Tu. skritlexicographersderivemalaya- froma hypo-
niru "water"; Te. niru "water" (literary);nillu thetical root mal "hold, possess," since the
"water"; imiri"moisture"; Kol.-Nk. ijr"water"; mountain range "contains sandalwood"! The
Pa.-Oll. nir "water"; Kui niru "juice, sap, commentatorsare wildlyat variance about the
essence"; Br. dir"water,flood-water, juice, sap": meaningof mala- in Meghaduita16, some taking
Skt. nira-"water" (epic +); "juice, liquor" (lexi- it as a propername ofa field,othersas a common
cal); nivara-"water,mud" (lexical). Kittel,no. noun; certaintyseems impossible,but it is in
157; Bloch, J., Bull. SchoolOrient.Stud. 5: 739, all probabilitythe wordthat we are dealingwith,
1930; Burrow,T., Trans. Philol. Soc. 1946: 9. whatever its specific denotation in Kalidasa's
An old Indo-European etymology,that always verse.
required faith both in its phonologyand in its 9. Ta. aiya excl. ofwonder,of pityor concern;
semantics,has recentlybeen revivedwithall the aiyako excl. of pityor sorrow;ai "wonder,aston-
apparatus of the laryngealhypothesisby Louis ishment"; aiy-enal "uttering ai expressive of
H. Gray, Lang. 25: 376-377, 1949; it is hardly wonder, of distress or mental suffering,of as-
convincingvis-a-visthe general Dravidian word sent"; aiyaiyo excl. of pity or grief; aiyo excl.
for"water." of wonder,of pity or concern,of poignantgrief;
7. Ta. nal, nar-,nalla "good"; nala (nalapp-, Ma. ayya interj. of derision; ayyo, ayyayy6
nalant-) "resultin good, take a favorableturn"; interj.of griefor pain; Ko. aya- excl. of surprise
nalappu "goodness, benefit, success"; nalam or grief;aya- ava- excl. of grief;To. eya- excl. of
"goodness, virtue, beauty, profit,fame, pros- surprise;Ka. ayyo, ayyayyo,ayyayye interj. of
perity"; nalavu "goodness"; narku "good"; griefor astonishmentor compassion; Tu. ayyo,
narpu, narram, nanpu, nanmai, nanri, nannar ayyayyointerj.of grief,annoyance,or pain; Te.
"goodness" (some of thesealso mean "beauty"); ayyo, ayyo, ayyayo, ayyayyo, ayayo interj. of
Ma. nal "good, fine"; nalam, nallam "goodness, sorrow,lamentation,pity, pain, etc.; Kui aige,
beauty"; nalpu, nalma, nanni, nanma "good- aigo, aigona, aike, aiko, aikona interj.of annoy-
ness"; nalla "good, fine, handsome"; Ko. nal ance, impatience,or disgust; Kuwi (S) ijalijo,
va-yn"one whose mouthsmellsgood, who enjoys ijalesa (j = y) "alas"; Kur. ayo, ayo ge excl. of

This content downloaded from 157.182.150.22 on Wed, 21 Oct 2015 10:23:28 UTC
All use subject to JSTOR Terms and Conditions
290 MURRAY B. EMENEAUT [PROC. AMER. PHIL. SOC.

pain or surprise; Malt. aya, ayyi, ayyu excl. of "ripen"; pum "fruit"; Ka. pan "be produced
wonder, woe, or joy; ay (y)oke, ay (y)okaboke (ripe fruit)"; pan, pannu "ripe fruit,ripeness";
"alas": Skt. aye excl. of surprise,recollection, Kod. panni "fruit"; Tu. palkuni, palkuni "be
or fear (esp. in dramas). In the Dravidian very soft (as an overripefruit),be pliant, flex-
languages these are either vocatives of words ible"; parnduni "be ripe, mature, (hair) turns
for "father" (all the southern languages) or gray"; parndu "ripeness, ripe fruit,ripe plan-
"mother" (Kui- Kuwi, Kur., Malt.), or have in tains, ripe, gray"; palu "ripening (as of fruit),
a secondaryway been equated with or assimi- half-ripe"; Te. paniuu "ripen, mature"; "fruit,
lated to such vocatives; note Ko. aya- ava- berry"; "ripe, mature"; panta "produce, crop,
"father! mother!" In Sanskrit there is no fruit,ripening"; Kol. panj- (pandt-) "become
such connection. An Indo-Europeanetymology ripe"; pandud "ripe fruit"; (Kin.) pan "fruit";
might be suggested with such interjectionsas Nk. pandl-"become ripe"; panjle "ripe fruit";
Gk. al, al, atal, Lith. ai, ai, German ei, and Pa. panl- "(plant) matures"; parfi- "ripen";
Skt. e, ai (both lexical) (cf.Walde-Pokorny1: 1), pal "ripe fruit, pus"; 011. parry(g)-"become
but the Dravidian formsare so much more cur- ripe"; Kuwi (S) pand.u"ripe fruit";Kur. panna
rent than the suggested Indo-European etyma (pafija) "ripen, (boil) festers,have a yellowish
that a borrowingfrom Dravidian seems more appearance (as after a prolonged illness)";
plausible. pafijkd "fruits"; Malt. palne "ripen"; panjek,
10. Ta. kalam, kalan "place, open space, panjeke "ripe": Skt. phala- "fruit"(Rigveda ?);
threshingfloor,battlefield";Ma. kalam "thresh- phalati "bear or produce fruit,ripen" (epic +).
ing floor,level space for spreading grains for Gundert, H., Ztschr. deutsch.morgenldndischen
drying,battlefield";Ko. kaim "place forthresh- Ges. 23: 519, 1869; also Caldwell and Kittel;
ing or dancing"; To. koln "threshingfloor"; Ka. Burrow,T., Trans.Philol. Soc. 1946: 10; Ammer,
kala, kana "threshingfloor,battlefield"; Kod. Karl, Ztschr.deutsch.morgenldndischen Ges. 51:
kala "threshingfloor"; Tu. kala "a square, bed 128, 1948; Mayrhofer,Manfred,Anthropos47:
of flowers,etc., place where pariahs assemble"; 664, 1952; with some uncertaintyBloch, Jules,
Te. kallamu "threshing floor"; kalanu "war, Bull. Soc. Ling. Paris 25: 17, 1925; Bull. School
battle,combat, (B) threshingfloor"; Kol. kalave Orient. Stud. 5: 740, 1930; Kuiper, F. B. J.,
"workshedin field,(Kin.) threshingfloor"; Nk. Acta Orientalia16: 305, 1938; the possibilityis
kalave "threshingfloor"; Pa. kali id.; 011. kalin not even mentionedby Lfiders,Heinrich,Kuhn's
id.; Go. kdar "sacred enclosure,threshingfloor"; Zeitschrift42: 198 ff.,1909. This hypothesisof
Kui klai "threshingfloor"; Kuwi (F) kranfi, borrowingis much more obvious than the sug-
kalomi id.; Kur. khall "field,piece of land suit- gested derivations from Indo-European roots
able fortillage"; Malt. qalu "fieldon the hills": meaning"swell" or "burst" or than Sturtevant's
Skt. khila- "threshingfloor,granary" (Rigveda connectionwith words for "leaf" (Lang. 17: 6,
+); "place, site" (lexical). Burrow, T., Bull. 1941).
Soc. Orient.and African Stud. 11: 133, 1943; 12. Ta. val "strong,forceful,skilful"; valamn,
Trans. Philol. Soc. 1946: 9. No Indo-European vallam "strength, power, right side, victory,
etymologyhas been plausibly suggested. authority"; vallavan, valavan, vallan, valldlan
11. Ta. paru (parupp-, parutt-) "ripen (as "strong man, capable man"; vallu (valli) "be
fruits,grain), grow mature,arriveat perfection, able"; vali "strength,power"; (valiv-, valint-;
become old, come to a head (boil), change color valipp-, valitt-) "be strong, compel"; valimai
by age, become pale or yellowish (as the body "strength,skill,hardness"; valiya "strong,big";
by disease), become flexible, pliant"; paru, valivu "strength";valu "strength,skill,ability";
paxruppu"ripeness, yellowness (of fruits),leaf (valupp-,valutt-) "be strongor hard"; valuppu,
turned yellow with age"; parunu (paruni-), vanpu, vanmai, "strength,firmness";valumai
pa.runu (paruni-) "grow ripe, become mellow, "strength,force,violence"; vala kkai, valaA kai
mature,be fullor perfect"; pa.ram"ripe fruit"; "right hand"; Ma. val, valu, valiya "strong,
Ma. parukka "grow ripe,become well-tempered, powerful,great"; vallu "be able, strong"; valip-
suppurate, decay"; "a fruit put to ripen"; pam, valima "greatness,bigness"; valiye "for-
paxruppu"ripeningof fruit"; paruppikka "ripen ciblv, suddenly"; valam "the rightside"; valah
artificially";param "ripe fruit,ripe plantains"; kai "right hand"; Ko. val "powerful, very,
Ko. parv- (pard-) "(fruits) become ripe, (boil, right"; val kay "righthand"; val(n) "man who
sore) opens"; pan "fruit"; To. posf- (post-) is clever at cheating"; valc- " (man) becomes

This content downloaded from 157.182.150.22 on Wed, 21 Oct 2015 10:23:28 UTC
All use subject to JSTOR Terms and Conditions
VOL. 98, NO. 4, 19541 LINGUISTIC PREHISTORY OF INDIA 291

stout, (heart) becomes bold, (grain) becomes "dewlap" may possiblyalso be taken as evidence
solid lump when boiled"; To. pas "right"; pas formalai as a native Dravidian word. For non-
koy "righthand"; paly- (pals-) "(child) becomes Dravidianists it should be noted that the Dra-
strong"; Ka. bal, bali "grow strong or firm, vidian formsare those that a borrowingfrom
increase"; bal, balu, bolu "strength,firmness, Sanskrit would have but are also those proper
bigness,abundance"; bala "right"; bala key/gey to a native Dravidian group of etyma.
"right hand"; balume, balme, baluhu, balpu
"strength";Kod. bala "power,strength";balate APPENDIX 2
"right (hand)"; ballye "great"; Tu. bala THE MUNDA LANGUAGES
"strength"; balatu "the rightside"; balata kai
"right hand"; balapini "gain strength,recover Identificationof borrowingsfromMunda into
health"; balike "prowess, strength,hardness"; Indo-Aryanand into Dravidian engaged several
balime, balume, balme "strength,might"; balu scholars in the 20's and the 30's, notably J.
"very large, great, severe, violent"; Te. valdti Przyluski and Suniti Kumar Chatterji. This
"clever person,expert"; vala "right"; vala cEyi was continued by the publication in 1948 of
"right hand"; valadu "much"; valanu "skill, F. B. J. Kuiper's Proto-Mundawordsin Sanskrit
excellence, possibility"; "right, possible, con- (Verh. d. K. Nederlandsche Ak. v. Weten-
venient"; valan instrumentalpostposition;val- schappen,Afd. Letterkunde,N. R. 51.3), a very
amu "largeness, stoutness"; valfida "stout, ambitious work. In spite of a few tempting
large"; valla "possible"; valladi "violence, op- borrowings among those suggested by these
pression"; Pa. vela key "righthand"; Go. wdllE various scholars (cf. Burrow,Trans. Philol. Soc.
"much, very"; Kuwi (F) braiyui,(S) blajugatti 1946: 3-6), there are several factors involved
(j = y) "strong"; Kur. bale, baleti "with the that give a conservativescholarpause.
help of": Skt. bala- "power, strength,might" First, the Munda languages are very poorly
(Rigveda +); balena, balatas "by means of." known to scholarship. Ambitious grammars
The v-languagesare primary;v > b in Ka. be- and dictionarieshave been publishedof Santali,
ginningin the 9th cent. A.D. Sanskritmay then Mundari, and So-ra- (Sabara or Savara), but,
have borrowedfroma Dravidian language with with all deferenceto the extended effortsand
secondary b. Kittel, no. 398; also Caldwell; the devotion of the missionariesand the educa-
Burrow,T., Trans.Philol. Soc. 1946: 19. Should tionistswho have worked here, we must judge,
we perhaps add also the name of the Rigvedic I think, that the intrinsiccomplexitiesof the
demon vala-, i.e. "the strongone"? It is just languages of this familyare not a suitable field
as good a suggestionas the derivations given for amateur effort. Extensive prefixing,infix-
from IE *wel- "drehen, winden, walzen" or ing, and suffixingmake both description and
*wer-"verschliessen,bedecken." comparisonof the languages very difficult,
and
13. Ta. malai "garland,wreath,necklace,any- until we have competentdescriptionsof several
thingstrungtogether";malai (malaiv-,malaint-) of them,comparisonsconcernedwith borrowing
"wear, put on" (Burrow,"put on as a garland," or with genetic relationshipswithin the family
based on the old text Purananuiru12, etc.); Ma. and outside it will be hazardous.
mala "garland, wreath,necklace, dewlap"; Ka. The second dubious factor concerningIndo-
male "wreath, garland, necklace"; Kod. madle Aryan borrowingfrom Munda is that so far
"necklace, dewlap, jungle cock's ruffor neck those who have tried to demonstratesuch bor-
feathers";Tu. male "garland,wreath,necklace"; rowings have been dealing not primarilywith
Te. mala id. Burrow,Bull. Soc. Orient.and Afri- Munda but with Father W. Schmidt's Austro-
can Stud. 12: 390, 1948. He notes that Ta. has Asiatic or with his Austric. Acceptance of the
malai in theoldestliterature(Purananuru60, 76) hypothesis that Munda, Khasi, Mon-Khmer,
and that the verb malai witha shortvowel is just and various other languages of Southeast Asia
as old and relatedas a derivativefromthe noun. form a language family to be called Austro-
The Ka. wordis classed by the Ka. grammarians Asiatic can so far only be on a differentplane
as a tatsama, i.e. "a term not borrowed from fromthe scholar'sattitudewith respectto Indo-
Sanskrit,but existingin Kannada as well as in European, Finno-Ugric,Turkic, Bantu, Malayo-
Sanskrit" (Kittel's definition);this native anal- Polynesian, Algonquian, or Athabaskan. In
ysis always may be taken as speaking in favor contrast with these, the material with which
of Dravidian origin. The Ma.-Kod. meaning Schmidtworkedis badly recordedand analyzed,

This content downloaded from 157.182.150.22 on Wed, 21 Oct 2015 10:23:28 UTC
All use subject to JSTOR Terms and Conditions
292 MURRAY B. EMENEAU [PROC. AMER. PHIL. SOC.

and in spite of some temptingsimilarities,it is often widely divergentforms; e.g. Skr. tunda-
clear that littleprogresshas been made beyond and caficu- [both 'beak'] from da-da, vent-,
themin the search forphoneticcorrespondences. kabari- and Xpida- [the firsttwo 'a braid of
The fieldis one in which accidental similarities hair,' the third 'a chaplet tied on the crown of
are still being found rather than etymologies. the head'] from wa-da-, . . . From the point
This scepticismbecomes morepronouncedwhen of view of methodologythis is, I think,the chief
Austronesian(= Malayo-Polynesian)is brought difficulty of these investigations,as by a purely
in to forman Austric family. Archaeologists, mechanicalapplicationof the 'sound-laws'nearly
historians and prehistorians,and ethnologists everythingcan be demonstrated." If these are
concerned with South Asia have erected lofty sound-laws (with or without quotation marks),
hypotheses on Schmidt's hypothesis, but the we have got back to the days of Voltaire. No
linguisticspecialist should know full well what wonder Kuiper astonishes himselfby the fact
a weak foundationSchmidt has provided and that "nearly everythingcan be demonstrated."
should refrainfromerectinghis buildingson it. Unfortunately, altogethertoo many of the Indo-
In identifyingIndo-Aryan borrowingsfrom Aryan borrowingsfromMunda astonish in this
Munda, Austric material has been used to the way. In addition, much derivation in the
fullestextent-groups of Austricetyma, includ- Austric languages depends on prefixationand
ing Munda words if they can be found in the infixation-thoughit is to be noted that littleof
badly recorded material, are provided as the this apparatus is so far known by satisfactory
bases for the borrowing. How hazardously comparative methods. If words may be split
speculative this procedurecan be is clear from up and prefixesand infixesdiscarded somewhat
a glance at Kuiper's introductionin which are arbitrarilyin the attempt to establishgroupsof
foundthese statementsof his method: (6) "the etyma,this derivationalmethodand the phono-
Austro-Asiaticconsonantal system had a rela- logical method sketchedabove are togetherun-
tively small number of phonemes with a wide likelyto establish much credence. But even so
range of possible realizations, the following thereare withoutdoubt a fewsane etymologies
sounds, forinstance, to be salvaged from the work already done,
thoughI do not know how to identifythem.
dit ' dh/th ' r jjc, s This slight critique of Munda and Austric
ift - dh/th r *z/*s studies makes clear, I hope, that the prime
desideratumin thisfieldis accurate and intensive
originallyhavingconstitutedone phoneme"; (7) recordingand analysis of the languages con-
"another Proto-Munda phenomenon inherited cerned. It is only after this is accomplished
fromprim. Austric is the nasalization and pre- that the way will be open forvalid comparative
nasalization of the consonants of a root"; (7) work and for the establishmentof a linguistic
"the vowelsare largelyinterchangeable. . . the prehistoryof the kind that is already possible
derivatives fromone and the same root have forIndo-Aryanand Dravidian.

This content downloaded from 157.182.150.22 on Wed, 21 Oct 2015 10:23:28 UTC
All use subject to JSTOR Terms and Conditions