On Recent Studies in Hindu Grammar 9781463221775

William Dwight Whitney reviews the work of Bruno Liebich and R. Otto Franke, two scholars whose work was foundational to

183 84 2MB

English Pages 27 [31] Year 2009

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

On Recent Studies in Hindu Grammar
 9781463221775

Citation preview

O n Recent Studies in H i n d u Grammar

A n a l e c t a Gorgiana

339 Series Editor George Anton Kiraz

Analecta Gorgiana is a collection of long essays and

short

monographs which are consistently cited by modern scholars but previously difficult to find because of their original appearance in obscure publications. Carefully selected by a team of scholars based on their relevance to modern scholarship, these essays can now be fully utili2ed by scholars and proudly owned by libraries.

On Recent Studies in Hindu Grammar

William Dwight Whitney

w

1 gorgia? press 2009

Gorgias Press LLC, 180 Centennial Ave., Piscataway, NJ, 08854, USA www.gorgiaspress.com Copyright © 2009 by Gorgias Press LLC Originally published in All rights reserved under International and Pan-American Copyright Conventions. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning or otherwise without the prior written permission of Gorgias Press LLC. 2009

1

ISBN 978-1-60724-593-3

ISSN 1935-6854

Extract from The ^American Journal of Philology 14 (1893)

Printed in the LTnited States of America

ON RECENT

STUDIES

IN

HINDU

GRAMMAR.

ig7

of doctor o f philosophy creditably with a dissertation on one o f K a l i d a s a ' s plays, and went to E n g l a n d for further study and for e m p l o y m e n t . H e was fortified, a m o n g other things, with a letter o f introduction to a S a n s k r i t scholar o f G e r m a n birth, then long resident in L o n d o n . T h i s scholar, on being consulted in regard to plans and pursuits, told him that all his hitherto acquired k n o w l e d g e h a d no real foundation, and was essentially w o r t h l e s s ; that, if h e wished to accomplish anything, h e must d r o p all besides and devote himself for two or three years exclusively to the study o f P a n i n i ; when that h a d b e e n done it would b e time to talk o f s o m e t h i n g else. J u s t how m u c h this rebuff had to do with turning m y friend's attention away to other studies I do not k n o w ; but, at a n y rate, until his death s o m e years after h e was not heard o f further in S a n s k r i t . S u c h was, doubtless in its most intense form, the spirit o f t h e devotees o f the native H i n d u g r a m m a r a generation ago. And, t h o u g h it has b e e n in s o m e m e a s u r e subdued since, it is b y no means extinct, when a man o f real learning and ability like D r . F r a n k e can still maintain (in his Casuslehre, etc., noticed a b o v e , p. 68, or p. 6 o f the reprint) that our profounder knowledge o f S a n s k r i t is to b e especially proportioned to our d e e p e r penetration into Panini's t e a c h i n g s — a g a i n s t which is to be set, as antidote, the s a m e author's e x p o s u r e o f Panini's failure in the article o f compounds. I t is, o f course, m u c h to the credit o f Panini that he exercises such a bewildering fascination over t h e minds o f those who involve themselves in the labyrinth of fyis r u l e s — t h o u g h t h e influence admits, I believe, of a natural explanation. I am fully persuaded that a n y one who should m a s t e r the H i n d u g r a m m a t ical science without losing his head, who should b e c o m e t h o r o u g h l y familiar with Panini and escape being Panini-bitten, would be able to m a k e e x p o s u r e s o f the weaknesses and s h o r t c o m i n g s and needless obscurities o f the g r a m m a r on a scale hitherto unknown. W .

D.

WHITNEY.

I I . — O N R E C E N T S T U D I E S IN H I N D U

GRAMMAR.

Nine years ago (in October, 1884) I published in this Journal a paper entitled " T h e study of Hindu grammar and the study of Sanskrit." It was intended to emphasize the difference between Sanskrit on the one side and Pänini with his successors on the other, and to point out the true place of the native grammar as an important division of Sanskrit science, requiring to be studied as such, and not as the foundation of our knowledge of the Sanskrit language. Since that time there have appeared a number of contributions to our knowledge of the Hindu grammar, from the pens of two younger scholars of decided ability, then unknown; and these contributions I propose to examine briefly, especially in order to see how they stand related to the question above set forth. The first of them appeared in 1885, and was entitled " T h e casesystem of the Hindu grammarians, compared with the use of the cases in the Aitareya-Brähmana"; it was a doctorate-dissertation by Bruno Liebich; the author is at present a privat-docent in the Breslau University. Its first part, printed in vol. X of Bezzenberger's Beiträge zur Kunde der indogermanischen Sprachen, was a digest of the system of rules laid down by Pänini for the use of the cases, and was very welcome, as must be every contribution to an easier understanding of the peculiarities and difficulties of the Hindu science. A few words as to the system may not be out of place here. Pänini does not take up the cases as forms of nouns, setting forth the various uses of each, after our manner; he adopts the vastly more difficult and dangerous method of establishing a theoretical list of modes of verb-modification by case, or of ideal case-relations (he calls them käraka, 'factor' or 'adjunct'), to which he then distributes the cases. Almost as a matter of course, however, his case-relations or käraka are not an independent product of his logical faculty, but simply a reflection of the case-forms; they are of the same number as the latter, and each corresponds to the general sphere of a case: they are kartar ( ' a c t o r ' = nominative), karman ('act' = accusative), sampradäna ('delivery' = dative), karana ('instru-

172

AMERICAN

JOURNAL

OF

PHILOLOGY.

m e n t ' = instrumental), adhikarana ('sphere' = locative), and apadana ('removal' = ablative). The genitive has no defined character, but is provided for by stating, when all the other caseuses have been rehearsed, that the remainder are those of the genitive. A s for the definitions of the case-relations, it may suffice to say that the karman is described as belonging, first, to that which the actor in his action especially desires to obtain or attain (as in " h e makes a mat" " h e goes to the village"*) ; or, second, to that which, though itself undesired or indifferent, is connected with the action in a similar manner. Anything more crude or unphilosophical than this could not well be imagined. There is not an identity between the use of a given case and the presence of its generally corresponding case-relation, because, for example, in a passive sentence, as "the mat is made by him," mat is still called karman or 'act,' though nominative, and him still kartar or 'actor,' though instrumental. Thus there is no recognition of the grammatical category of subject of a verb ; and this leads, as could not be helped, to numerous obscurities and difficulties. Then, in the second part of the paper (ibid., vol. X I , 1887), the author proceeds to classify under this scheme, in all its headings and sub-headings, its general rules and its exceptions, the facts of case-use in the Aitareya-Brahmana: a careful and creditable piece of work. The results of the comparison are precisely what we should expect to find them, knowing well, as we do, the relation of the language of the Brahmanas to Panini's Sanskrit: there is a good degree of general agreement—as there would have been found to exist even if the Rig-Veda instead of a Brahmana had been compared; since changes of syntactical construction, perhaps even more than changes of form, are of slow progress in every language, leaving the main body of older usages long untouched. Alongside of this agreement are met with just the differences that could not fail to appear: constructions in the Brahmana that are unnoticed in Panini, as they are wanting in classical Sanskrit; and especially a host of details in Panini of which the Brahmana exhibits no examples. There is absolutely nothing to show, or even to give reason to suspect, that any special relation exists between Panini and this Brahmana any more than any other of the same class of works, specimens alike of the Brahmana stage of development of ancient Indian language. The conclusion is that, whatever its defects of theory, Panini's case-syntax proves to be a fairly good practical scheme; and the

ON RECENT

STUDIES

IN

HINDU

GRAMMAR.

I73

demonstration of the fact is to be received with t h a n k s ; it is a valuable contribution to our appreciation of the great g r a m m a r i a n . W h e t h e r , h o w e v e r , the author views it in just this light is a little questionable ; for he a d d s as s e c o n d title to his essay " a contribution to the s y n t a x of the S a n s k r i t l a n g u a g e " — w h i c h it plainly is not. Is it, forsooth, the B r a h m a n a that he has been e x a m i n i n g , t o see w h e t h e r its case-constructions are such as t h e y o u g h t to be ? or is this part of its g r a m m a r now better understood than hitherto, or a r r a n g e d in a manner which w e shall be disposed to accept as preferable to, for e x a m p l e , D e l b r i i c k ' s ? N o t h i n g of all this ; it is simply that Panini has been tested b y a bit of real l a n g u a g e , and the test has turned out not to his disadvantage. T h e misapprehension that s o m e t h i n g done for Panini is done for the S a n s k r i t l a n g u a g e is precisely w h a t m y f o r m e r paper was especially intended to discourage. D r . L i e b i c h a d d s at the end his o w n estimate of the results of his w o r k : " 1 . T h e A i t a r e y a - B r a h m a n a is older than Panini." T h i s were better stated the other w a y : namely, that Panini is later than the B r a h m a n a ; since it is really the grammarian, a n d not this m e m b e r of the literature, that is under examination. As for the relation itself, it is not o n l y true, but a t r u i s m ; no o n e h a v i n g a n y k n o w l e d g e of the subject has or could h a v e any question about i t ; our author's paper is not a demonstration, but m e r e l y an illustration, out of o n e department of g r a m m a r , of a fact a l r e a d y incontrovertibly established on m a n y and sufficient grounds. T h e author a d d s as f o l l o w s : " I t [the B r a h m a n a ] b e l o n g s to the V e d i c period, but to the close of the latter, and stands fairly near to Panini ( u n d o u b t e d l y m u c h nearer than to the R i g - V e d a in the other direction)." H e r e again we h a v e truths, but, since there has been no comparison m a d e b e t w e e n B r a h m a n a and V e d a in the paper, t h e y are incorrectly put forward as its "results." F u r t h e r , " 2 . T h e doctrine of Panini reposes upon a careful a n d acute observation o f t h e actual l a n g u a g e . " H e r e it is a little doubtful w h e r e the stress of the assertion lies, and w h a t counter-proposition is intended to be gainsaid. N o one, certainly, would think o f d e n y i n g that Panini o b s e r v e d and described with r e m a r k a b l e acuteness and to the best of his ability. N o r , again, I s h o u l d think, that he described an actual l a n g u a g e — " a n " rather than " t h e , " for j u s t w h a t l a n g u a g e he was d e a l i n g with is one o f the d i s p u t e d points. T h e author's a d d e d r e m a r k s indicate that he thinks it a b o o k - l a n g u a g e ; if a n y t h i n g in the rules is not

174

AMERICAN

JOURNAL

OF

PHILOLOGY.

capable of being instanced, it is, he suggests, because so much of the literature has been lost. This seems an untenable view, and has doubtless been since abandoned by him. T h e question will come up again further on. Four years later (1890), in the same Journal ( X V I 1-2), a kindred topic is taken up by another scholar, Dr. R. Otto Franke, now a privat-docent in the Berlin University. T h e title of his paper is " T h e case-system of Panini compared with the use of the cases in Pali and in the A j o k a inscriptions." He builds upon the foundation laid by Liebich, adopting the latter's scheme of Paninean case-uses, and looking for correspondences to them in the dialects confessedly later than Panini, as the Brahmana was confessedly earlier. Here also he finds all the agreement that could reasonably have been expected; and, as the ground has been comparatively little worked over, his work is much more truly a contribution to the syntax of the dialects of India than is that of Liebich. He brings to light one very curious thing : that for a problematic rule of Panini's, declaring the future tense to be usable in describing something recently past, examples are quotable from the Pali, though they have never been discovered in Sanskrit. But his general views as to Panini and the Sanskrit seem rather strange. He calls Liebich's little work " a beginning toward the accomplishment of the very pressing task of determining by internal evidence Panini's position in the literature, and so, indirectly, that of the Sanskrit"—as if nobody, before the appearance of this doctorate-dissertation, had done anything worthy of mention in that direction; or as if the position of Panini's Sanskrit in the history of development of Indian language had not long been clear enough. And he points out that, in spite of the partial agreement between the case-uses in the AitareyaBrahmana and Panini's rules, we ought not to conclude that the Brahmana was the exclusive, or even the principal, foundation of the rules—as if it could ever enter into the mind of any reasonable person to draw such a conclusion. He then gives us the same warning in regard to the Pali, which is even, if possible, more superfluous. H e further admits it as possible, though on the whole less probable, that Panini may have "collected the phenomena of very diverse dialects, and fused them together into an integral whole"—than which nothing could well find less to be said in its favor. But to the question as to what the Sanskrit of Panini really is the same author returns in a special paper entitled " W h a t is

ON RECENT

STUDIES

IN

HINDU

GRAMMAR.

175

S a n s k r i t ? " , dated in N o v e m b e r , 1889 ( t h o u g h first published in vol. X V I I , 1891, of B e z z e n b e r g e r ' s B e i t r ä g e ) . R a t h e r more than half the p a p e r is o c c u p i e d with the m o r e specific inquiry as to w h a t Pänini means b y bhäsä, a w o r d that he uses only seven times, or too s e l d o m to set forth its significance with the desirable clearness. ' P o p u l a r s p e e c h ' is its natural s e n s e ; but the u s a g e s q u o t e d from it b y Pänini as o p p o s e d to his o w n a p p r o v e d lang u a g e show that it was no Prakritic dialect (that is the chief result of the author's investigation); and it is as evidently not one of the older V e d i c d i a l e c t s ; there seems to remain, then, only one possibility : it is essentially S a n s k r i t , only not w h a t Pänini accepts as g o o d S a n s k r i t ; it includes those w o r d s a n d phrases which, t h o u g h m o r e or less current, he d o e s not r e g a r d as w o r t h y to be p e r p e t u a t e d . T h i s conclusion a p p e a r s to be a reasonable a n d safe one. 1 T h e second half of the paper then deals more directly with the inquiry as to w h a t Pänini's S a n s k r i t really i s ; and the author's opinion is e x p r e s s e d in these terms (pp. 7 5 - 7 6 ) : " P ä n i n i ' s S a n s k r i t is a c c o r d i n g l y in the main bhäsä. A n d yet on the other hand, it is neither bhäsä nor a living l a n g u a g e . " T h i s is not particularly c l e a r ; nor is it m a d e v e r y m u c h clearer b y the reasonings, and the quotations of the views of others, that follow. It is to m e so strange as fairly to be called unaccountable that these authors t a k e no notice w h a t e v e r of the e v i d e n c e of the dramas u p o n the subject. In the latter w e see a condition of society in w h i c h educated p e o p l e talk Sanskrit, while the uneducated talk Prakrit, in dialects m o r e or less different from one another. S o far as I can perceive, there is not any reason to question that this state of things was real at the time w h e n those d r a m a s were p r o d u c e d which then set the rule for all future time. T h e s p e a k e r s all understand one a n o t h e r ; the difference b e t w e e n Sanskrit and P r ä k r i t is not y e t sufficient to prevent t h a t ; the P r a k r i t - s p e a k e r s can even, in an e m e r g e n c y , put in a phrase of S a n s k r i t ; and, on the other hand, w h e n K i n g Purflravas g o e s mad, he casts off the restraints of education, a n d talks in part Präkrit, like a woman. T h a t , now, is j u s t the present character of S a n s k r i t : an educated or learned dialect, k e p t in existence, nearly u n c h a n g e d , b y instruction, b y learned and literary use, a m o n g l a n g u a g e s now b e c o m e so d i v e r s e from it that its k n o w l e d g e is confined to a v e r y small c i r c l e ; such, too, has been its 1 It is, however, rejected by L i e b i c h , in his ' K ä g i k ä ' (p. xxv), to be described further on. L i e b i c h suggests no substitute.

176

AMERICAN

JOURNAL

OF

PHILOLOGY.

character for at least two thousand years, while the true vernaculars have been growing further and further away from it; and such must unquestionably have been its character at the outset, when their divergence, and its separate life, first began. That it was itself originally a vernacular seems to me a matter of course; nor do I see that any one has the right to say that Pánini's speech was not a living one, unless he then enters into a full explanation of what he means by a living language as distinguished from it. Sanskrit was the natural successor of the dialects of Veda, Brahmana, and Sütra, and as much "living" as any of these had been, when the literary and learned class took it in hand, and, with the aid of grammatical science, fortified it against the further effect of the changes that were bringing out of it the various Prakritic dialects (taking that word in its widest sense). There is no absolute line to be drawn between living and dead languages. If the Sanskrit has never failed of being kept up by a constant tradition from teacher to pupil, though in a limited class, there is a real sense in which it has never died, but is still a living tongue. In another and equally correct sense, no language is alive that is not an out-and-out vernacular, spoken by a whole community, and having no inferior dialect below it in the same community; in this sense, to be sure, the Sanskrit of the series of grammarians of whom Pánini was the chief and virtually the last was not a fully living tongue; it had Prakritic dialects under it. Moreover, as soon as it took on the character of a learned dialect, it began as a matter of course to be stiffened into something a little unnatural ; no dialect ever fell into the hands of grammarians without suffering from their pedantry. But I can find no reason whatever for supposing that it was not their own language, the language which they themselves spoke and which they thought alone worthy to be spoken by others, that they set themselves to describe. W h a t ever Pánini's special original part in the work may have been, we know that he left it still abounding in errors, both of omission and of commission ; the important additions and corrections of Katyayana and Patanjali, to say nothing of their numerous but more insignificant successors, amply prove this; and it is frankly conceded in many points by these latest students of the system, unlike the scholars of a generation or two ago. T h e task Pánini attempted was beyond the power of mortal man to accomplish, especially in the form adopted by him—which is one that no sensible man should ever have chosen, yet on account of which,

ON RECENT

STUDIES

IN

HINDU

GRAMMAR.

177

it is very likely, his contemporaries and successors especially admired him, and made him their supreme authority. Something like this, in my opinion, is what we have a right to say that we know about Pânini ; and the investigations of Liebich and Franke, while they bring nothing to light that contradicts it, merely illustrate here and there a point in it, and do not add notably to its amount, because they ignore it all, and assume that the most fundamental facts involved have still to be established. W h a t we really need further is added precision on a host of points -as to which we have as yet only general knowledge, and particularly a comprehension of how the grammatical system, in all its details, stands related to the language of the Sanskrit classical literature, which professes to be governed by it, and yet has evidently had a traditional life of its own, simply regulated by the grammar, and has by no means been produced under the latter's dictation. T o ask and answer, in all seriousness, such questions as whether a certain Brâhmana, or whether the Pâli, is Pânini's Sanskrit, or whether that language was a living one, appears to me the wrong way to arrive at any valuable result. In his conspectus of the views of various scholars as to the character of Sanskrit, given in the second part of his paper, Dr. Franke quotes with approval and acceptance an old expression of opinion by Weber, made at the very outset of his career, to the effect that " t h e development of Sanskrit and of the Prakrit dialects out of their common source, the Indo-Aryan mothertongue, went on with absolute contemporaneousness (yollstàndig gleichzeitig)." But I do not see why this is not an unscientific and untenable proposition. For example, pakkhitta and attâ or appâ are not contemporaneous with praksipta and âtmà in the historical development of language, any more than Ital. rotto and rotti with Lat. ruptus, -urn in their various case-forms ; and so hodu is preceded in point of time by bhavatu, being a later "corruption" of the latter, coming to take its place, as Fr. était of stabat, or fûtes of fuistis. And this is true of the great mass of Prakrit words, forms, and constructions ; they are developed later than, and come to be substituted for, the corresponding Sanskrit words, forms, and constructions. If there were anything to be found on Indian ground that is earlier than praksipta, and from which it and pakkhitta should have equally descended by a parallel process, then we might have a right to speak of their contemporaneity; but that is plainly not the case; it is the

AMERICAN

JOURNAL

OF

PHILOLOGY.

Sanskrit forms themselves, and not something older and more primitive than Sanskrit, that the Prakrit words presuppose; they have passed through the stage which the Sanskrit represents. That here and there exceptions are met with, altered items for which the original is not found in Sanskrit, or is found in Vedic Sanskrit, is without any force whatever as against the great mass of material of a contrary character ; such exceptions to the descent in toto of one dialect from another are the rule in all dialectic history, and might with equal justice be relied on to prove that Italian and French are in their development "absolutely contemporaneous" with Latin. A s the other half or side of the view already quoted, Dr. Franke adds: " T h a t the Sanskrit had become extinct when the Prakrit dialects first began to develop themselves is false." W h a t this means is quite unintelligible; it seems to go out of the way to deny a doctrine which no wellinformed student of language could by any possibility think of maintaining, and it accordingly has no claim to be criticized, but must be simply set aside as valueless. If, for example, atma had ever become extinct, whence should atta or appa have originated ? W h o would say that the e g g had been extinct when the chicken first began to develop itself? But, somehow or other, those whose ancestors had said atma began to say atta instead, the one pronunciation passing into the other, with no extinction intervening. It was, however, only a part of the community who did thus; a part, doubtless much the smaller one, continued to say atma; and the two forms went on in currency side by side, as educated and as popular speech, in the same way as in many cases elsewhere in the world; and atma was Sanskrit, and, with some help and some mishandling on the part of grammarians, has maintained itself in being to this day, in the literature which we call Sanskrit, and which, rather than the grammarians' treatment of it, is the true and proper object of the study of the Sanskrit scholar. Next was produced by Dr. Liebich, in 1891, a valuable collection of studies entitled " Panini: a contribution to the knowledge of Indian literature and g r a m m a r " ; it makes a small octavo volume of 164 pages. T h e first study, or chapter, deals with Panini's period; the author reviews briefly the opinions that have been held by different scholars respecting the matter, and, without attempting to bring any new evidence to bear upon it, comes to the moderate and sensible conclusion that only a certain degree

ON RECENT

STUDIES

IN

HINDU

GRAMMAR.

179

of probability can be arrived at: "after Buddha and before Christ" represents to him the measure of this probability. The second chapter treats of the principal later grammarians who have continued and modified Panini's work ; in regard to the earliest and most important of them the same chronological uncertainty prevails. The third is entitled "Panini and the remaining literature," and is an attempt to determine where in the succession of the ancient literature of India, from the earliest Veda down, Panini comes in. It takes as starting-point the wild views of Goldstiicker, with their refutation by Weber; it points out further the insufficiency of the evidences relied upon for the prevailing opinion that Y a s k a is earlier than Panini; and it then proceeds to its principal task, of applying to the general question a new, a numericalstatistical, method of solution. The author counts off, namely, a thousand personal verb-forms occurring in succession in each of four different monuments of the literature—the Aitarey a-Brahmana, the Brhad-Aranyaka, two Grhya-Sutras (Ayvalayana and Paraskara), and the Bhagavad-Gita: representatives respectively of the Brahmana, Upanishad or later Brahmana, Sutra, and epic stages of development of Indian speech—and then applies to them the rules of the grammar, to see how many and what forms unauthorized by Panini appear in the several texts. The examination is creditable to the industry and learning of its author, and its results are interesting; we can hardly go further than that and pronounce them important. For they are essentially illustrative o n l y ; they put in a numerical form peculiarities which were already familiarly known to characterize the different classes of works instanced. Not a new item, so far as I can see, is brought to light; nor is any made more certain than before. Thus, six of the seven classes of Brahmana divergencies drawn out on pages 23-4 have long been recognized as such ; and how many examples of each class may chance to occur in a given amount of text is a matter of indifference. As for the seventh, represented by a single case, the lengthened final of the imperative krdhi, it is an error ; such a protraction does not belong to the Brahmana language, as, indeed, it has no right of occurrence anywhere except in verse; where it appears here (ii 2. 21), it is simply copied from the R i g Veda verse (i 36. 14) on which the Brahmana is engaged in commenting, and of which it repeats a whole pada (including krdhi) with merely the substitution of the more regular carandya for carathaya in it; and the retention of the l is not improbably even

180

AMERICAN

JOURNAL

OF

PHILOLOGY.

a misreading, s u c h as this B r a h m a n a has in no v e r y small n u m b e r (it m a y b e a d d e d that the author, doubtless misled b y Panini, describes krdhi falsely as a present instead of an aorist imperative). A n d so also in each of the r e m a i n i n g cases. T h a t is to say, the m a t t e r is not one to which the numerical m e t h o d of investigation is well s u i t e d ; this w o u l d be m u c h better applied between, for e x a m p l e , different texts o f the same class, as different B r a h m a n a s , to see whether it w o u l d y i e l d any evidence as to their respective periods ; and p e r h a p s the part of the w h o l e investigation which is of most value is the c o m p a r i s o n which it m a k e s possible between A i t a r e y a - B r a h m a n a and B r h a d - A r a n y a k a , the latter b e i n g part of a B r a h m a n a also, but plainly later, as was a matter o f course for an U p a n i s h a d . Instead, again, of the B h a g a v a d G i t a , w h i c h no one doubts to be a c o m p a r a t i v e l y recent addition to the M a h a b h a r a t a , it were m u c h to be wished that the author had selected s o m e t h i n g out of those parts of the epic which are m o s t p r o b a b l y to be r e g a r d e d as its original nucleus, in o r d e r to cast m o r e light u p o n the really difficult and hitherto doubtful question how and how m u c h the epic differs from the classical or Paninean Sanskrit, and w h y . T h a t B r a h m a n a and U p a n i s h a d a n d S u t r a antedate Panini w e k n e w just as certainly before this investigation was m a d e as w e k n o w it n o w ; the posteriority of the B h a g a v a d - G i t a , again, could h a r d l y h a v e b e e n questioned, h o w e v e r the case m a y stand with the earliest epic. T h e criteria applied to the d i v e r g e n c e s of the G i t a from g r a m m a t i c a l strictness are of a less satisfactory and decisive character. T h e decided m a j o r i t y (21 against 16) of the irregularities c o n c e r n the voice of the v e r b ; but, t h o u g h the looseness of at least the later epic in this r e g a r d is certainly excessive, it is likely that Panini's rules limiting the e m p l o y m e n t of the voices are e x c e p t i o n a l l y artificial and discordant with g e n u i n e u s a g e ; our author himself so j u d g e s e x a m p l e s of them (e. g., p. 28) in connection with the Brahmana. A s for the causative perfects with asa (3 in number), Panini's failure to a u t h o r i z e t h e m must be either an oversight or a piece of p e d a n t r y . A n d fucas, since this aorist o c c u r s in V e d a and B r a h m a n a , m i g h t be d e e m e d a sign rather o f antiquity than o f m o d e r n date. T h e harvest of results from the chapter, then, must be confessed a rather scanty one. In the sixth and seventh chapters the author returns to the A i t a r e y a - B r a h m a n a and the B r h a d - A r a n y a k a , in order to see w h e t h e r any difference of period can be established a m o n g their

ON RECENT

STUDIES

IN

HINDU

GRAMMAR.

l8l

constituent parts. H e r e again is, as in the particular noted above, a good and suitable application of the statistical method, and it leads to trustworthy and interesting conclusions. In the A r a n y a k a are discovered no notable indications of diversity of a g e ; but in the Brahmana the author finds good reason to believe, as had been inferred by others before him, that the concluding chapters are more modern than the rest. Between the parts of the volume devoted to the first and to the second examination of these two works intervene a couple of chapters, of which the former, the fourth, is headed " P a n i n i ' s relation to the language of India"—that is to say, the relation of Panini's Sanskrit to the other dialects. T h e chapter is chiefly composed of a succinct statement of the views of other scholars, to which the author then appends his own v i e w ; and this is simply a summary of what he has illustrated in the preceding chapters as the relation between Panini's dialect and the Brahmana and Sutra on the one side and the epic on the other. Then (p. 50) he appends as final result a wholly new and original classification of the entire body of dialects of India. T h e y are divided into three categories: pre-classical, classical, and post-classical. T o the classical division are referred, besides " t h e doctrine of Panini," the Brahmanas and Sutras also, which the author has himself previously recognized as pre-Paninean! this leaves as pre-classical only " t h e samhitas of the four V e d a s . " But the third division, the post-classical, is still more wonderfully constructed: besides the " i n d e p e n d e n t " epic, it contains the whole literature which we have been accustomed to know as Sanskrit, namely " K a l i d a s a , etc., originated under the influence of the g r a m m a r " ! W h a t is left to constitute the classical subdivision " b. Doctrine of P a n i n i " is very obscure; it can be only Panini's g r a m m a r itself (so that such sentences as idamo rhil, gatlkutadibhyo 'niiinhit are classical, as contrasted with Kalidasa's compositions), and in addition all the works that might, could, would, or should have been written in strict accordance with it, and not merely " u n d e r its influence," if there only were any such. Now I had myself, in m y former paper, laid stress on the difference between the purely hypothetical "grammarians' S a n s k r i t " and the Sanskrit of the literature; but I never went so far as to maintain, with D r . Liebich, that the two even belonged to different prime divisions of the whole history of language in India (thus, II. b. grammarians' S a n s k r i t ; I I I . b. Sanskrit of the literature).

182

AMERICAN

JOURNAL

OF

PHILOLOGY.

Just half our author's volume (pp. 8 2 - 1 6 1 ) is occupied by two studies which are reckoned as Appendix I and Appendix II. The one is a digest of the teachings of the native grammar (Panini, the Mahabhasya, and the K a f i k a ) respecting the voiceinflection of the verbal roots, as active or middle or both; the other is a similar digest for the formation of feminine declensionstems from the corresponding masculines. These two appendixes constitute, in my opinion, the substantially valuable part of the volume; they exemplify what needs to be done for all the various subjects included in Panini's treatise. The next step, now, should be to compare in detail the statements thus drawn out with the actual facts of the language as exhibited in the whole series of monuments of the literature, from Vedic down to classical and epic, in order to determine what is the relation between the two, and then what the former, the prescriptions of the grammar, are worth; until that is done, no contribution has yet been made to our knowledge of the language, but only to our knowledge of Panini. It casts a shade of unreality over the whole subject of voice-conjugation that the voices of the thousand or twelve hundred false roots are not less carefully defined by the dhatupatha than those of the eight or nine hundred genuine ones. There is left for our consideration only the fifth chapter, in which the author takes up and attempts to answer my own objections, given in my paper of nine years ago, to the confusing of the study of Panini with that of Sanskrit, and the thrusting of the grammarians' dialect into the place in our attention which the real language of the recorded literature ought to occupy. I propose to examine here this reply, and see how effective it is. Dr. Liebich's first point is, as was my own, the dhatupatha, or list of roots, which is given as part of the material of the grammar, and really even its foundation, since it is upon them that tlie rules of the grammar profess to go on and build up the structure of the language—and that not only grammatically but lexically, for the grammar includes the system of derivation, with definition of the modifications wrought in each root-sense and stem-sense by the added suffixes. On this point the author offers a criticism which he is obliged himself to withdraw in the next paragraph : he first accuses me of treating Panini rather unfairly, since the dhatupatha was the part of his work most likely to be deformed by later corruptions; but then allows that I was perhaps (as is indeed plainly the case) criticizing the whole system of the grammarians

ON RECENT

STUDIES

IN

HINDU

GRAMMAR.

183

as it lies b e f o r e us, of w h i c h the list of roots o b j e c t e d to forms u n d e n i a b l y an inseparable part. B o h t l i n g k gives it in l e n g t h and breadth in his recent s e c o n d edition o f Panini, finding n o t h i n g else to put in its p l a c e ; and it must h a v e g o n e h a r d with him, w h o k n o w s w h a t in S a n s k r i t is real and w h a t is sham better than almost any other living scholar, and w h o has in the P e t e r s b u r g lexicons d o n e m o r e than a n y one else to m a k e plain their distinction, to introduce into his w o r k s u c h a m a s s o f worthless r u b b i s h ; I h a r d l y c o m p r e h e n d h o w he should h a v e prevailed on himself to d o this without e x e r c i s i n g his critical a c u m e n upon it, a n d s e p a r a t i n g in s o m e w a y the false from the true. O u r author talks of p r o b a b l e interpolations, and intimates that he d e e m s them posterior to the g r e a t trio of Panini, K a t y a y a n a , and Patanjali, a c k n o w l e d g i n g that m y criticisms m a y be " m o r e or l e s s " applicable to their successors. W e l l , I should think s o ; and m o r e rather than less. T h i s free and easy w a y of d i s p o s i n g of the subject is quite characteristic of the w h o l e g u i l d of partizans of the native g r a m m a r . It appears impossible to bring any one of t h e m to stand up and face fairly the question o f the dhatupaiha. T h e r e are not far f r o m nine h u n d r e d real authenticable roots in Sanskrit. W e c o u l d believe that the uncritical interpolations of later g r a m m a r i a n s m i g h t a d d to this n u m b e r a d o z e n , or a score, or fifty, or (to t a k e the e x t r e m e ) e v e n a h u n d r e d or t w o ; but it is the wildest o f nonsense ( o n l y s t r o n g expressions suit the case) t o hold that t h e y could swell t h e n u m b e r to o v e r t w o t h o u s a n d ! S u c h increase is thus far w h o l l y unexplained, perhaps forever unexplainable, and certainly most u n p a r d o n a b l e ; and until it is in s o m e w a y accounted for the admirers of t h e H i n d u science o f g r a m m a r o u g h t to talk in v e r y h u m b l e tones. I f these roots are not the ones r e c o g n i z e d by the w o n d r o u s three, w h e n a n d under w h a t circumstances and b y w h o s e influence were the additional t w e l v e h u n d r e d foisted in, to the a b a n d o n m e n t and loss of the old g e n u i n e list? T h e difficulty of explaining this s e e m s not less g r e a t than that of s u p p o s i n g the w h o l e t w o thousand as old as Panini h i m s e l f ; both are h a r d e n o u g h ; and, in either event, the taint of falsity attaches to the H i n d u s y s t e m as w e k n o w it and are e x p e c t e d to use it. A s concerns the three points of the m i d d l e periphrastic perfect, the m i d d l e precative, and the s e c o n d a r y passive forms, nothing that the author s a y s tends to c h a n g e at all the aspect of the case as stated b y m e : n a m e l y , that these are formations which, t h o u g h

184

AMERICAN

JOURNAL

OF

PHILOLOGY.

taught b y Panini, are wanting in the traditional literary l a n g u a g e — a s m u c h so as verb-forms from the thousand and m o r e false r o o t s ; they b e l o n g to the g r a m m a r i a n s ' S a n s k r i t alone. Just h o w m u c h or how little e x c u s e Panini m a y h a v e had for setting t h e m up, that is a different and a minor question, to be d e c i d e d finally b y the general result of our examination of Panini's w a y of w o r k i n g , of selecting what he will a d o p t a n d what he will reject. T o m e they seem artificial and pedantic structures, reared on an obsolete and insufficient or m i s a p p r e h e n d e d basis. T h e author's well-intended correction of m y estimate oiprayoktase in T S . ii 6. 23 as 1st sing. I d o not find m y s e l f able to accept. T h e sentence is not, perhaps, absolutely c l e a r ; but the p r e s e n c e in it of a te ' f o r t h e e ' is to m e a tolerably certain indication that the v e r b is not 2d sing. ( ' I will e m p l o y to-morrow for thee at the sacrifice,' or ' a t t h y s a c r i f i c e ' ) ; no such possessive would be called for (or admissible, I t h i n k ) if the person were second. A n d -tase is obviously the true m i d d l e a n a l o g u e to active -tdsmi, as fase to fasmi and the l i k e ; while -take, as g i v e n b y the g r a m marians, is absolutely anomalous, being unsupported, so far as I k n o w , b y a single other phonetic fact of the l a n g u a g e . T h a t it occurs once (but only o n c e ) in the literature, in that v e r y late V e d i c document the Taitt. A r a n y a k a , w h o s e text is in m a n y parts e x t r e m e l y faulty, is b e y o n d q u e s t i o n ; but I would put forward the s u g g e s t i o n , as b y no means an impossible one, that the form is corrupt, and that the 1st sing, -tahe of the g r a m m a r i a n s is f o u n d e d solely on it. T h a t the native c o m m e n t a r y , it m a y be a d d e d , explains prayoktase in T S . as 2d sing, is not of the smallest particle o f i m p o r t a n c e ; an expositor schooled in Panini would of course do that, a n d is capable of d o i n g it against the most incontrovertible e v i d e n c e to the contrary. A n o t h e r matter w h i c h the author u n d e r t a k e s to defend against m y objections is Panini's determination of the cases w h e r e dh and w h e r e dh is to be used in the 2d plur. e n d i n g s dhvarn and dhve. H e is so far successful that he is able to show the g r a m m a r i a n s ' rules to admit in part a different interpretation from that put u p o n them by the later H i n d u authorities, and r e p o r t e d b y the E u r o p e a n g r a m m a r s which follow these rather than the l a n g u a g e itself. I was careful to allow for this possibility in so flagrant a case, putting in the caveat " i f the H i n d u g r a m m a r i a n s are r e p o r t e d rightly b y their E u r o p e a n pupils (which in this instance is hard to b e l i e v e ) " ; it now appears that a part of the reproach is capable of b e i n g

ON RECENT

STUDIES

IN

HINDU

GRAMMAR.

185

shifted from the shoulders of Panini to those of his later interpreters. But only a part. Panini uses in the first of his two rules one of his customary algebra-like signs, in, which is ambiguous, signifying either simply the i- and z^-vowels, or these together with the r- and /-vowels, the diphthongs, the semivowels, and h. But such an ambiguity is itself a palpable blot upon a system that claims to be so precise, and Panini's successors are little to blame, comparatively, if they have chosen the wrong meaning. Then, further, it is and must be equally a matter of uncertainty whether this same in is or is not to be carried over by implication from the first to the second rule; and this, again, is a characteristic and a pervading difficulty, running through Panini's entire work, and, as I said in my former paper, involving " a condemnation of the whole mode of presentation of the system as a failure." W h a t are the boasted terseness and exactness of the rules really worth, when in innumerable cases you cannot tell what they mean without first knowing what they ought to mean?—that is to say, when an acquaintance with the facts of the traditional language is necessary in order to the right interpretation of the grammar's diciurn respecting them ? T h e present is, at the best, a case where the interpreters have been too careless of the facts and the reasons of the facts. But, whatever improved explanation we may apply to them, there is plenty left to object to in Panini's rules. T h e 2d pi. precative middle is plainly declared to end in sidhvam or in sidhvam according to what letters precede the s (which might also be s) ; and this is senseless. If the ending is -§idhvam, it is so because the form is originally -si-s-dhvam, with the special precative sibilant between mode-sign and personal ending, as in 2d and 3d sing., -§i-§-thas and -si-s-ta ; if it is, on the other hand, -sidhvam, this is because, as in 1st persons and 3d plur., no such sibilant is present, and the ending is originally -§l-dhvam ; and no one can speak with certainty upon the point, because, as I have pointed out, not a single example of the form has been brought to light out of the literature, earlier or later (the probabilities are altogether in favor of si-s-dhvam, and so -sidhvam); but it is perfectly obvious that what precedes the -si- has nothing to do with determining the matter, any more than with determining the presence or absence of the precative sibilant in the 2d and 3d singular. It is equally plain that in the indicative of the z§-aorist we must always have dhvatn (which the known texts also always

186

AMERICAN

JOURNAL

OF

PHILOLOGY.

g i v e ) , because -idhvam necessarily results from the combination -is-dhvam, without any reference whatever to what m a y p r e c e d e the -is-; and the interpreters must r e g u l a t e t h e m s e l v e s accordi n g l y , if they wish to save Panini's credit. T h e author thinks he catches me in an error in s a y i n g , as concerns this point, that " a l l the q u o t a b l e e x a m p l e s . . . are o p p o s e d to their rule," and brings u p against me astodhvam etc. out of my g r a m m a r . B u t this o n l y s h o w s how carelessly or how unintelhgently he has read m y p a p e r ; for it is distinctly allowed there that the rule as g i v e n applies correctly to the s-aorist, and there is q u o t e d the e x a m p l e anedhvam (from anes-dhvam; by the w a y , this e x a m p l e and its like seem to show that in in the rule requires to be taken in its wider s e n s e ) : one of the striking things about the matter was that a prescription suiting well the one aorist had been wantonly e x t e n d e d to include the other, with which it h a d nothing to do, its application g i v i n g in e v e r y instance a different form from the theoretically correct one found o c c u r r i n g in the literature. B u t Panini undeniably takes the perfect also into his rule, m a k i n g its 2d plur. ending to be dime or dhve under the s a m e conditions as those laid d o w n for the aorists. T h e impropriety o f the combination and identical treatment of the t w o tenses is clear. T h e aorist has a l w a y s at the end of the stem a lingual sibilant—anes-, apavi§

to exercise its euphonic influence u p o n

the dh of the e n d i n g , while in the perfect there is none such. T h a t is to s a y , none unless the endings dhve and dhvarn are really b y origin sdhve and sdhvam; and this is a doctrine w h i c h has found, and perhaps still possesses, some adherents. B u t it has no foundation w h a t e v e r in the actual phenomena of S a n s k r i t , but solely in these b l u n d e r i n g rules of the native g r a m m a r . E x a m p l e s of the 2d plur. perfect, indeed, are of e x c e e d i n g rarity ; I a m able at present to point to only a single one (dadhidhve, o c c u r r i n g twice in R V . ) in the older l a n g u a g e . But, if we are to r e c o g n i z e sdhve in the perfect, w e plainly o u g h t to r e c o g n i z e sdhve and sdhvam also in the present (indie., impv., and opt.) and i m p e r f e c t ; and then we should not m e e t with forms like studhvam, janldhvam, bhavedhvam, akrnudhvam, but with studhvam and so on. It appears, then, that the only w a y to s a v e Panini's reputation in t h e matter is to s t r i k e the syllable lit ( m e a n i n g ' p e r f e c t ' ) out of his rule, as u n g e n u i n e ; and I w o u l d s u g g e s t that it was p e r h a p s intruded b y the same c u n n i n g h a n d that thrust into the dhatupatha m o r e than a thousand false roots without b e i n g

ON RECENT

STUDIES

IN

HINDU

GRAMMAR.

187

d e t e c t e d or d e t e r r e d ; this latter trick was e v i d e n t l y b y far the h a r d e r to e x e c u t e . B u t D r . L i e b i c h finds two other defenses to m a k e (both on p. 58). F o r one thing, w e are not justified in a s k i n g for a reason w h y dhvam should in certain cases b e c o n v e r t e d into dhvam. " A s if," he exclaims, " w e w e r e able in any l a n g u a g e w h a t e v e r to trace e v e r y w h e r e the connection of cause and e f f e c t ! " Begging his p a r d o n , I assert that, on the contrary, in the combinations of stem a n d e n d i n g in S a n s k r i t e u p h o n y , w e d o not meet with any effect of w h i c h w e m a y not l o o k for a cause with g o o d expectation of finding it. If w e c a m e a n y w h e r e u p o n a dhvam without a discoverable reason, w e s h o u l d question its correctness, and hold it probable that s o m e one had blundered, that the text-tradition was corrupt, or the like. O n the other hand, if, as is actually the case, w e h a v e no dhvam for w h i c h w e cannot show a perfectly g o o d reason (few as, unfortunately, the instances are), and no dhve at all, and can put against this only the assertion of Panini a n d his successors and interpreters that such forms o u g h t to occur without any reason, I submit that the sole acceptable conclusion must be that these g r a m m a r i a n s , like grammarians e v e r y w h e r e else, h a v e blundered, a n d n e e d to be corrected. O u r author's remaining p l e a is one that, it must be confessed, g i v e s a tinge of the c o m i c to the w h o l e discussion. T h e difference, he points out, between dh a n d dh is v e r y slight, and it m i g h t be unfair to e x p e c t Panini in e v e r y case to distinguish the one correctly from the other ! T h a t is to s a y , if Panini prescribes a (¡,h w h e r e there is no g r o u n d for one, it m a y be simply the fault of his ear, w h i c h c a u g h t the s o u n d w r o n g . N o w I h a v e been accused, b y the author and others, of insinuating d e p r e c i a t o r y things about Panini, but I certainly never went so far as this. If the g r e a t g r a m m a r i a n h a d too dull an ear to distinguish a lingual mute accurately from a dental (like the typical, or mythical, G e r m a n , w h o cannot tell t and d apart), w h a t are all his teachings worth that i n v o l v e phonetic distinctions? T h e staff is b r o k e n o v e r Panini, a n d b y one o f his o w n partizans. T o c o n c l u d e (after passing without notice the other points m a d e b y m e ; the most important was the grammarians' derivation of the reduplicated aorist from the causative stem instead of from the root directly), D r . L i e b i c h t a k e s up m y criticism of the Paninean classification of c o m p o u n d s , d e f e n d i n g and extolling this classification; and he returns to the same subject, elaborating

i88

AMERICAN

JOURNAL

OF

PHILOLOGY.

his view still further, in the introduction to another later publication, " T w o chapters of the K a f i k a . " 1 According to him, the true scientific principle of arrangement of compounds, which must be regarded as underlying Panini's scheme, is furnished by syntactical subordination, after the following fashion: i. In the copulative compounds, as devamanusyas 'gods-and-men,' neither element is subordinated to the other, but both are coordinate ; 2. in the determinatives, the former element is subordinated to the latter, either as a case dependent on it or as an adjective (or its equivalent) qualifying it: examples are housetop, redbird ; 3. in the possessives, both are subordinated together to a word outside the compound, which they jointly qualify in the manner of an adjective: for example, redhead, i. e. redheaded, or possessing a red head; then, 4. there remains only one other possibility, namely that the second element should be subordinated to the first, as in aiimatram 'beyond measure': we might give as English parallel aboveboard or overboard (also, for the other Hindu variety, consisting of a participle governing a following noun, the English spendthrift or hategood; of this variety our author makes no account, because it is Vedic, and unnoticed by Panini). If, then, we are told, the subordinated element be represented by a minus-sign, and the other by a plus, we get thus the four combinations + + , — 1 - , , - ) — ; and these evidently exhaust all the possibilities of the case. Now this is in the real Paninean style, and proves Dr. Liebich to possess a double portion of Panini's spirit, if he be not the great grammarian himself in the latter's «th metempsychosis. Panini would have been proud to adopt it into one of his chapters, together with its algebraic notation, so akin with his own. But our author has to confess that it is not Panini's own scheme; it is only brought out full}' and distinctly by a much later successor. Moreover, that Panini's fourth class, the so-called avyayibhava compounds, is by no means limited to examples of the formula plus-minus, but includes a number of quite heterogeneous formations. Dr. Liebich is nevertheless confident that he recognized the unique value of the scheme, and had it plainly in mind ; only he sacrificed it, " perhaps with a heavy heart" ( K a f i k a , p. ix), on the altar of—brevity! This brings to our notice, and in a strikingly illustrative manner, another of Panini's leading characteristics and at the same time greatest weaknesses. The prime object aimed at by him (as in 1

Z w e i K a p i t e l d e r K a p i k a , B r e s l a u , 1S92, 8vo, p p . x l , 80.

ON RECENT

STUDIES

IN

HINDU

GRAMMAR.

189

no small measure in the i«/r«-style everywhere) is brevity, brevity at the cost of every other desirable thing—of theoretic truth, of connection, and, most of all, of intelligibility. The quality may be one that recommended his work to those who had to learn it by rote (though in its degree we have the right to question even that), but it is very much the opposite of a recommendation to us, and cannot but detract very seriously from our approval and admiration. And this especially when we see how capriciously the principle is applied—how many rules are squandered on details of the most trifling consequence, far below others that are omitted; on the quotation of other grammarians (the best way to confute whom was to leave them unnoticed); on the excerption {in more than 200 rules) of scattered particulars out of the Vedic language, which are valueless because they are merely specimens, making no pretense to completeness, while the motive of their selection is in many cases beyond the reach even of conjecture— and so on. If the grammar were sharply examined with reference only to this its leading motive, it would unquestionably be found to teem with matter for unfavorable criticism. But there is another and more fundamental difficulty lying behind Panini's oversight, or possible sacrifice, in not recognizing the fourth, the plus-minus, class of compounds in its true character, and thus rounding out a perfect scheme of classification, namely this: there is no such class; Dr. Liebich and his authorities, the later Hindu grammarians, are deceiving themselves with a false determination and notation; the avyaylbhava class, however composed, is not plus-minus, but minus-minus. By this is not meant that the component parts of such compounds do not stand in a plus-minus relation to one another; but so also do those of the ordinary possessives stand in a minus-plus relation; and if the possessive is nevertheless really a minus-minus compound, so is, for the same reason, the avyaylbhava. The copulative compound, composed of two (or more) nouns or adjectives, is itself noun or adjective accordingly, and is properly reckoned as plusplus ; the determinative is a noun or adjective with preceding limiting word, and it also is noun or adjective accordingly, and rightly minus-plus. It is different with the possessive, because, though this is not less a noun with a preceding limiting word, it has passed through a transformation making of it an adjective, which is to qualify something outside : mahabahu when it means ' a great arm' is determinative or minus-plu's; but when it means

I90

AMERICAN

JOURNAL

OF

PHILOLOGY.

' having a great arm' it is changed to minus-minus. If we represent the adjectivizing influence by a, we shall get the equation (minus-plus)0, = minus-minus, which is good linguistic mathematics ; at any rate, it is only in such a way that the possessive comes to be a minus-minus compound. But precisely the same is true of the avyaytbhava. Taking, for example, the participial compound dbharad-vasu 'bringing wealth,' we find it made up of a governing word and its object-noun; but it is not therefore a noun; it has been transformed to an adjective; its accus. sing, and nom. plur. are not abkarantam-vasu and abharanto-vasu, but abharad-vasum and abharad-vasavas ; it has undergone a similar transformation to that of mahabahu, and it is tninus-minus ; for its formula is again (plus-minus)" = minus-minus. But the proper avyayibhava is not an adjective, but an adverb; the phrase aii matram 'beyond measure' becomes as a compound atimatram 'excessively.' Here is plainly involved a similar fusion and transfer to that already described; and, if we represent the adverbmaking force by b, the proper formula for atimatram, is (plusminus)* = minus-mi?ius. But in real truth atimatram is still further from being a plus-minus compound; for to any one who considers the class historically it must be obvious that any such adverb is simply the neuter accusative of an adjective used adverbially, as neuter accusatives, among simple words and compounds of every kind, are wont to be used. For example, the first step from ati matram is the common adjective atimatra 'excessive,' of which the formula is (plus-minus)a \ then from this comes by another transfer the adverb, with the formula ((plus-minus)")'', or, more briefly, (plus-minus)**; and, as the adjective was minus-minus, the adverb is doubly so. Whether this double transfer be accepted or not (of course the acceptance does not imply that some of the adverbs have not been made directly, by analogy with the others of more regular development), the asserted plus-minus class is irretrievably lost, and with it the mathematically exhaustive and regular classification of Sanskrit compounds. It has, indeed, never been found that the facts of language could be reasoned on mathematically; and, whenever the attempt so to treat them is made, we have the right to expect to detect a misapprehension, as in the present case. W e may now decline to be touched by the spectacle of Panini's "heavy heart," and hold, on the contrary, that Dr. Liebich has probably done him for a second time signal injustice, in believing him

ON RECENT

STUDIES

IN

HINDU

GRAMMAR.

I9I

c a p a b l e of b e i n g deceived b y an alluring t h o u g h false t h e o r y . T h e a d j e c t i v e c o m p o u n d s with g o v e r n i n g prior m e m b e r , w h e t h e r this be preposition or participle, are sub-classes, with the possessives, of the g r e a t class of s e c o n d a r y adjective c o m p o u n d s , as I h a v e located and described t h e m in m y g r a m m a r ; and the avyayibhavas are no class of c o m p o u n d s at all, but only a g r o u p in the l o n g list of adjective neuter accusatives used a d v e r b i a l l y . It m a y be further mentioned, as a curiously characteristic point, that our a u t h o r objects ( K a g i k a , p. xi, note 2) to the name " p o s s e s s i v e " as applied b y B o p p and his successors to the " m u c h - r i c e " (bahuvrihi) c o m p o u n d s , because s o m e o f them admit of b e i n g fairly rendered otherwise than b y ' h a v i n g ' or ' p o s s e s s i n g , ' and because the S a n s k r i t has no v e r b ' h a v e , ' and therefore Panini w o u l d not h a v e cast the sense into this form. T h e n also, it m a y be inferred, we are w r o n g to s p e a k of the " p o s s e s s i v e " suffixes in and vant, and to r e n d e r balin and balavant b y ' h a v i n g strength,' or to call madiya ' m y ' a " p o s s e s s i v e " pronominal a d j e c t i v e or tasya ' h i s ' a " p o s s e s s i v e " genitive. It m a y be p l e a d e d in reply that, since we n a m e them in our own l a n g u a g e and not in Sanskrit, w e h a v e e v e r y right to cast their real and undeniable sense into the form of nomenclature that best suits our e x p r e s s i o n ; and that the H i n d u s t h e m s e l v e s put the idea of possession as well as t h e y can into the definitions of these c o m p o u n d s b y their familiar f o r m u l a y a s y a .. . sa tathoktah, : they s a y , for e x a m p l e , " w h o s e arms are g r e a t " in place of our " h a v i n g great a r m s " : and it really seems to a m o u n t to the same thing. A t the close of his chapter, D r . L i e b i c h , conceiving himself to h a v e b r o k e n the force of all m y objections to setting Panini a b o v e the S a n s k r i t literature, and his g r a m m a t i c a l science a b o v e ours, regrets that I h a v e not b r o u g h t forward a happier selection of them. I, on t h e other hand, think m y s e l f justified in maintaining that, as t h e y all still stand in full vigor, they are a sufficient illustration and s u p p o r t of m y contrary estimate o f the native g r a m mar. B u t I am willing to add another point, which he indeed almost forces upon m y attention. A t the v e r y end, n a m e l y (p. 6 1 ) , he lifts up hands of horror at me (as did Speijer, in his Sanskrit S y n t a x , p. 189, note) for d a r i n g to stigmatize as a barbarism s o m e t h i n g w h i c h Panini e x p r e s s l y teaches (his alarm m a k e s him s e e it as double, or worse than double, and he puts it in the plural, as a thing h a p p e n i n g " o c c a s i o n a l l y " ) . H e ought fairly to h a v e q u o t e d the case, instead of merely referring to the

192

AMERICAN

JOURNAL

OF

PHILOLOGY.

rule about it. It is this: Panini teaches that a c o m p a r a t i v e and superlative adverbial ending m a y be a d d e d to a personal v e r b : thus, dadati ' h e gives,' daddiiiardm ' h e g i v e s more, dadatitamam, ' he g i v e s most.' T h i s is precisely as if o n e were directed to s a y in G r e e k Sidwrnrtpov (in this case, e v e n the suffix is identical) and didaxriTarov. N o w I maintain, and without any fear of successful contradiction, that such formations, no matter w h o authorizes them, are horrible barbarisms, offenses against the proprieties of universal I n d o - E u r o p e a n speech. T h e total absence of a n y t h i n g like them, or of a n y t h i n g s u g g e s t i n g even remotely the possibility of f o r m i n g them, in the pre-Paninean l a n g u a g e (one might just as successfully s e e k for s u g g e s t i o n s of diSmairepov in H o m e r or Plato), and their rarity later (no e x a m p l e of -tamdm is ever met with), a m o n g writers to w h o m a rule of Panini is as the oracle of a g o d , is e n o u g h to show that t h e y never formed any p r o p e r part of the language. P r o b a b l y t h e y were j o c o s e or s l a n g y m o d e s of e x p r e s sion (essentially bhasa, but far below the level of decent bhdsa), which some strange freak, p e r h a p s of amusement at their o d d i t y ( a n d Panini was entitled to s o m e compensation for the " h e a v y h e a r t " w h i c h his s u b s e r v i e n c y to brevity often cost him), led him to s a n c t i o n — i f indeed the rule permitting them be not another interpolation b y that mischief-maker w h o spoiled the list of roots. D r . L i e b i c h complains of the ( p r e s u m a b l y disrespectful) references to " t h e native g r a m m a r i a n s " which he finds too frequent in m y Sanskrit g r a m m a r , and k i n d l y advises m e to cast them all out. B u t this is in the h i g h e s t d e g r e e unreasonable. C o n s i d e r i n g the place which those g r a m m a r i a n s h a v e l o n g o c c u p i e d in the s t u d y of the l a n g u a g e , and the influence allowed them b y their E u r o p e a n successors, and that their w a y s of v i e w i n g and presenting t h i n g s h a v e determined in l a r g e measure the form of universal Sanskrit g r a m m a r , it is s i m p l y impossible to leave them out of account and unmentioned. I am sure I h a v e been as respectful to t h e m as I possibly could, and p r o b a b l y in the majority of cases quite s u c c e s s f u l l y — a t least h y p o t h e t i c a l l y respectful, stating their t e a c h i n g for what it m a y be worth, and l e a v i n g to the future the final determination of its value. It was h a r d l y respectful for him, o n his part, to pronounce (in his closing sentence) all m y references to them " e x t r e m e l y superficial and often inaccurate," without quoting a single instance to show that t h e y really bear that character. P e r h a p s , if he had done so, he w o u l d h a v e m a d e as signal a failure of it as he has of the attempt to refute the v i e w s and reasonings of m y former paper.

ON RECENT

STUDIES

IN

HINDU

GRAMMAR.

I93

A n e x t e n d e d review of L i e b i c h ' s Panini, b y D r . F r a n k e , is found in the G o t t . Gelehrte A n z e i g e n bearing date of D e c . 1, 1891 ( p p . 9 5 1 - 8 3 ) . It is, h o w e v e r , less a detailed examination and criticism of the former's views than an independent discussion of s o m e of the points involved, carried on with m u c h learning and acuteness. M a n y p a g e s are e x p e n d e d u p o n Panini's classification of the c o m p o u n d s ; and here D r . F r a n k e is far from s u p p o r t i n g L i e b i c h ' s answer to m y criticisms; on the contrary, he t a k e s m y side, setting forth the r e m a r k a b l e superficialities and incongruities of Pánini's w o r k in this department, especially as r e g a r d s the asserted class of avyaylbhdvas ; he m a k e s m a n y points of detail which I h a v e passed without notice in the a b o v e discussion of the theoretic g r o u n d w o r k of the classification. T h o u g h d a t e d in the following year, L i e b i c h ' s K a g i k a a n d its introduction were d o u b t less written before the a p p e a r a n c e of this review ; he would h a r d l y h a v e ventured to repeat his views, or w o u l d have cast them into a v e r y different form, if he had had before his e y e s their c o n d e m nation b y a fellow-partizan of Panini. In other points, F r a n k e ' s notice of L i e b i c h ' s w o r k is mainly l a u d a t o r y . T h u s , he " t h o r o u g h l y a p p r o v e s , " as " v e r y s u c c e s s f u l " (p. 962), the latter's futile pleadings as to the ending dhvam (including, I suppose, the s u g g e s t i o n of Panini's dullness of ear), a d d i n g , as his own contribution to the c o n t r o v e r s y , that a dh not seldom takes the place of dh in Prakrit, and that Prakritic c h a n g e s h a v e been k n o w n to w o r k their w a y into Sanskrit. B u t what has that to d o with Panini's definite prescription of dh in certain conditions which d e m o n s t r a b l y h a v e nothing to d o with the m a t t e r ? S o in Prakrit, in obedience to the s a m e general lingualizing t e n d e n c y , n in the majority of cases b e c o m e s 11; but that w o u l d be far from supporti n g a H i n d u grammarian w h o should teach that a r altered the next following n to 11 o n l y p r o v i d e d it were itself p r e c e d e d b y the sounds included in the designation in. A s for the great question of the 1200 false roots, D r . F r a n k e slips s m o o t h l y over it, m e r e l y e c h o i n g the other's r e m a r k , that it was an " unfortunate p r o c e e d i n g " on m y part to c o m m e n c e from that quarter m y attack upon the native g r a m m a r . Unfortunate, i n d e e d ; but evidently unfortunate only for the g r a m m a r : w h o could help starting from that most flagrant, wanton, a n d inexcusable of all its m a n y w e a k sides ? It is hardly worth while to say m u c h more than has been already said with r e g a r d to Liebich's K a ? i k a . It is a laborious and useful contribution to the s t u d y of Panini himself and of one

194

AMERICAN

JOURNAL

OF

PHILOLOGY.

of the m o s t noted c o m m e n t s u p o n his w o r k , s m o o t h i n g a little the w a y to their c o m p r e h e n s i o n f o r t h o s e w h o shall a p p r o a c h it hereafter. T h e author's m e t h o d is a n a r r o w l y restricted one ; the rule of Panini is g i v e n , not translated, and then follows a bald r e n d e r i n g of the K a g u k a ' s exposition, with h e r e a n d there brief notes a d d e d on one and another point in the l a t t e r ; f r o m a n y attempt at an i n d e p e n d e n t explanation, a n d y e t m o r e f r o m a n y criticism, the a u t h o r c a r e f u l l y refrains. T h u s , of the rule which introduces the w h o l e subject, samarthali padavidhih,, the K a f i k a g i v e s two entirely d i s c o r d a n t interpretations, illustrating, h o w e v e r , o n l y the latter of t h e m — w h i c h is a v e r y s t r o n g indication that the c o m m e n t a t o r s w e r e t h e m s e l v e s uncertain as to w h a t m e a n i n g r e a l l y l a y h i d d e n in its o b s c u r i t y ; and the translator p a s s e s the matter without a w o r d of r e m a r k , nor d o e s it occur to him to state whether in his opinion w e o u g h t to understand ' a w o r d - r u l e is c o m p e t e n t , ' or to f o r c e into the t e x t with e x t r e m e violence the sense ' a w o r d in the following rules is to be taken in connection with its s e n s e ' : it is o n l y an illustration of the o r d i n a r y principle that y o u must first find out what a rule of Panini o u g h t to s i g n i f y , and must then, at w h a t e v e r cost, interpret that signification into it. A n d the continuation is of a piece with the b e g i n n i n g . No one can well a v o i d being m o v e d to r e p u g n a n c e b y the fantastic obscurity with which the subject is p r e s e n t e d ; and w e k n o w a l r e a d y that the u n d e r l y i n g t h e o r y , the s c h e m e of distinctions and of classification, is a v e r y d e f e c t i v e one. T o claim, then, that it must be all l a b o r e d t h r o u g h b y the g e n e r a l b o d y of students of S a n s k r i t , in o r d e r that t h e y m a y d u l y understand the subject of S a n s k r i t c o m p o u n d s , is o b v i o u s l y unreasonable, not to s a y a b s u r d . Panini a n d his chief c o m m e n t a t o r s m u s t be w o r k e d over b y a small class of specialists, a n d not s i m p l y t r a n s l a t e d — t h a t is a m e r e b e g i n n i n g of the t a s k — b u t b r o u g h t into s u c h a form as to be r e a d i l y u n d e r s t o o d and assimilated b y the m a s s of s c h o l a r s . T h e s t u d y is e x c e s s i v e l y difficult, and on m a n y of the points i n v o l v e d in it certainty s e e m s unattainable. D r . L i e b i c h confesses (p. i) that h e found the r e n d e r i n g of these two little c h a p t e r s so h a r d that he could s c a r c e l y k e e p his c o u r a g e u p to c o m p l e t e the task. S p e i j e r has been a faithful student of the native g r a m m a r ; but of the discussions a n d criticisms of points in it on which h e occasionally v e n t u r e s in his S a n s k r i t S y n t a x , B o h t l i n g k (in a r e v i e w of the w o r k in Z . D . M . G . X L I 1 7 9 ff.) claims to refute nearly e v e r y o n e ; and now L i e b i c h ( K a g i k a , p. iv) declares

ON RECENT

STUDIES

IN

HINDU

GRAMMAR.

IÇ5

Bohtlingk, in spite of his life-long familiarity with the subject and his immense erudition, to have translated Panini sometimes incorrectly. Rather discouraging that for a student who is ambitious to get his knowledge of Sanskrit directly from native sources! I would be far from saying anything to discourage the study of Panini; it is highly important and extremely interesting, and might fairly absorb much more of the labor of the present generation than has been given to it. But I would have it followed in a different spirit and for a different purpose and in a different method. It should be thoroughly dissociated from the study of Sanskrit, though never without recognition of what it ma}' finally contribute to our knowledge of Sanskrit in addition to what we derive from the literature. As to what the literature contains, we need no help from the native g r a m m a r ; it is the residue of peculiar material that we shall value, and that we should strive to separate from the mass. And the study should be made a truly progressive one, part after part of the native system being worked out to the last possible degree and the results recorded, so that each generation be not compelled to begin anew the tedious and unrewarding task. At the beginning of the introduction to his Kafika, it is true, Liebich makes the claim that all Sanskrit students need to master Panini, if for no other reason, because the native commentaries cannot be otherwise completely understood, it being known that they abound more or less in references to the grammar and demonstrations founded upon it. There would be more in this consideration if the grammatical discussions were not precisely the most worthless part of the comments, which can be in all cases neglected with least fear of loss. W h a t the words mean, what allusions they contain, what is to be supplied to complete the sense, which of possible constructions is the right one—these are matters in regard to which the aid of the commentator is more or less (in proportion, namely, to the artificiality of the composition) welcome, sometimes even indispensable; but for the grammatical forms, the derivations, and everything else that Panini can be quoted for, the case is different. As for Sayana and his kind, even those who make the strongest claims in his favor will hardly venture to deny that the whole grammatical part of his exposition might be expunged from his text without loss of a jot or tittle of its value. It may be added that Dr. Franke also, in the first paragraphs of his review of Liebich briefly examined above, shows the same

I96

AMERICAN

JOURNAL

OF

PHILOLOGY.

disposition to e x a g g e r a t e and misrepresent the claims o f Panini to attention. H e quotes once more, as L i e b i c h had done before him, L a s s e n ' s unworthy insinuation that B o p p ' s g r o w i n g indep e n d e n c e of Panini was o w i n g to his i g n o r a n c e of h i m ! A s if B o p p did not k n o w Panini, both at firtt h a n d and in his E u r o p e a n representatives, sufficiently to j u d g e with full c o m p e t e n c e what his system was worth, and h o w far it required to be followed ! T h e r e is quite too m u c h of Panini left still in B o p p ' s g r a m m a r ; y e t to B o p p b e l o n g s the h i g h credit of m a k i n g the r e c o r d e d facts of the l a n g u a g e for the first time the basis of their orderly presentation, and of bringing the principles of E u r o p e a n grammatical science, a n d those of a new and d e v e l o p i n g c o m p a r a t i v e g r a m m a r , to bear u p o n Sanskrit. It is o w i n g to this that he became the real S a n s k r i t teacher t o E u r o p e , in a manner and d e g r e e far b e y o n d the reach of L a s s e n . D r . F r a n k e then g o e s on to vindicate for Panini various things to which he has not the s h a d o w of a j u s t title: as, 1. that not only for Sanskrit, but also for other A r y a n dialects and writings, Panini is of indispensable i m p o r t a n c e — w h i c h apparently means nothing more than that s o m e o f the p h e n o m e n a of dialects later than Sanskrit are to be found noted in his g r a m m a r ; 2. that the s t u d y of his rules has a formally e d u c a t i n g i n f l u e n c e — w h i c h is, I think, just the opposite of the truth, since their m e t h o d is p u r e l y mechanical, sacrificing e v e r y t h i n g else to brevity, ignoring connection and proportion, l a c k i n g all recognition o f the historical element, and therefore necessarily destitute of p h i l o s o p h y (we h a v e seen a b o v e that too m u c h Panini has led D r . L i e b i c h to d o u b t the relation of cause and effect in S a n s k r i t e u p h o n y ) ; 3. that it is Panini w h o has t a u g h t us to r e g a r d e v e r y word, e v e r y ending, even e v e r y letter as i m p o r t a n t — which is an accusation laid without a n y reason w h a t e v e r against western grammatical s c i e n c e ; and 4. that Panini is g o i n g to aid literary c h r o n o l o g y in a w a y that is hitherto for the most part only a matter of conjecture and of future h o p e — a n d w h i c h therefore, we m a y answer, it is as y e t too early to s a y a n y t h i n g a b o u t ; but, if there are such treasures hid in Panini, w h y d o not his partizans d e v o t e themselves to b r i n g i n g t h e m forth, instead of dwelling u p o n subjects w h i c h are far better understood out of the literature itself? Just forty y e a r s a g o , a G e r m a n student of more than o r d i n a r y ability, in c o m p a n y with w h o m I had w o r k e d for a season under a professor of the highest eminence in G e r m a n y , t o o k the d e g r e e