Sacred Texts  Hinduism  Index  Previous  Next 

p. 1


IT has become quite a literary commonplace, that--to borrow the words of Professor Max Müller in one of his recent lectures--history, in the ordinary sense of the word, is almost unknown in Indian literature 1. And it is certainly a remarkable irony of fate, that we should be obliged to make this remark on the very threshold of an introduction to the Bhagavadgîtâ; for according to the eminent French philosopher, Cousin 2, this great deficiency in Sanskrit literature is due, in no inconsiderable measure, to the doctrines propounded in the Bhagavadgîtâ itself. But however that may be, this much is certain, that the student of the Bhagavadgîtâ must, for the present, go without that reliable historical information touching the author of the work, the time at which it was composed, and even the place it occupies in literature, which one naturally desires, when entering upon the study of any work. More especially in an attempt like the present, intended as it mainly is for students of the history of religion, I should have been better pleased, if I could, in this Introduction, have concentrated to a focus, as it were, only those well ascertained historical results, on which there is something like a consensus of opinion among persons qualified to judge. But there is no exaggeration in saying, that it is almost impossible to lay down even a single proposition respecting any important matter connected with the Bhagavadgîtâ, about which any such consensus can be said to exist. The conclusions arrived at in this Introduction must, therefore, be distinctly

p. 2

understood to embody individual opinions only, and must be taken accordingly for what they are worth.

The full name of the work is Bhagavadgîtâ. In common parlance, we often abbreviate the name into Gîtâ, and in Sanskrit literature the name occurs in both forms. In the works of Sankarâkârya, quotations from the Gîtâ are introduced, sometimes with the words 'In the Gîtâ,' or 'In the Bhagavadgîtâ,' and sometimes with words which may be rendered 'In the Gîtâs,' the plural form being used 1. In the colophons to the MSS. of the work, the form current, apparently throughout India, is, 'In the Upanishads sung (Gîtâs) by the Deity.' Sankarâkârya, indeed, sometimes calls it the Îsvara Gîtâ 2, which, I believe, is the specific title of a different work altogether. The signification, however, of the two names is identical, namely, the song sung by the Deity, or, as Wilkins translates it, the Divine Lay.

This Divine Lay forms part of the Bhîshma Parvan of the Mahâbhârata--one of the two well-known national epics of India. The Gîtâ gives its name to a subdivision of the Bhîshma Parvan, which is called the Bhagavadgîtâ Parvan, and which includes, in addition to the eighteen chapters of which the Gîtâ consists, twelve other chapters. Upon this the question has naturally arisen, Is the Gîtâ a genuine portion of the Mahâbhârata, or is it a later addition? The question is one of considerable difficulty. But I cannot help saying, that the manner in which it has been generally dealt with is not altogether satisfactory to my mind. Before going any further into that question, however, it is desirable to state some of the facts on which the decision must be based. It appears, then, that the royal family of Hastinâpura was divided into two branches; the one called the Kauravas, and the other the Pândavas. The former wished to keep the latter out of the share of the kingdom claimed by them; and so, after many attempts at

p. 3

an amicable arrangement had proved fruitless, it was determined to decide the differences between the two parties by the arbitrament of arms. Each party accordingly collected its adherents, and the hostile armies met on the 'holy field of Kurukshetra,' I mentioned in the opening lines of our poem. At this juncture, Krishna Dvaipâyana, alias Vyâsa, a relative of both parties and endowed with more than human powers, presents himself before Dhritarâshtra, the father of the Kauravas, who is stated to be altogether blind. Vyâsa asks Dhritarâshtra whether it is his wish to look with his own eyes on the course of the battle; and on Dhritarâshtra's expressing his reluctance, Vyâsa deputes one Sañgaya to relate to Dhritarâshtra all the events of the battle, giving to Sañgaya, by means of his own superhuman powers, all necessary aids for performing the duty. Then the battle begins, and after a ten days' struggle, the first great general of the Kauravas, namely Bhîshma, falls 1. At this point Sañgaya comes up to Dhritarâshtra, and announces to him the sad result, which is of course a great blow to his party. Dhritarâshtra then makes numerous enquiries of Sañgaya regarding the course of the conflict, all of which Sañgaya duly answers. And among his earliest answers is the account of the conversation between Krishna and Arguna at the commencement of the battle, which constitutes the Bhagavadgîtâ. After relating to Dhritarâshtra that 'wonderful and holy dialogue,' and after giving an account of what occurred in the intervals of the conversation, Sañgaya proceeds to narrate the actual events of the battle.

With this rough outline. of the framework of the story before us, we are now in a, position to consider the opposing arguments on the point above noted. Mr. Talboys Wheeler writes on that point as follows 2. 'But there remains one other anomalous characteristic of the history of the great war, as it is recorded in the Mahâbhârata, which cannot

p. 4

be passed over in silence; and that is the extraordinary abruptness and infelicity with which Brahmanical discourses, such as essays on law, on morals, sermons on divine things, and even instruction in the so-called sciences are recklessly grafted upon the main narrative.... Krishna and Arguna on the morning of the first day of the war, when both armies are drawn out in battle-array, and hostilities are about to begin, enter into a long and philosophical dialogue respecting the various forms of devotion which lead to the emancipation of the soul; and it cannot be denied that, however incongruous and irrelevant such a dialogue must appear on the eve of battle, the discourse of Krishna, whilst acting as the charioteer of Arguna, contains the essence of the most spiritual phases of Brahmanical teaching, and is expressed in language of such depth and sublimity, that it has become deservedly known as the Bhagavad-gîtâ or Divine Song. . . . Indeed no effort has been spared by the Brahmanical compilers to convert the history of the great war into a vehicle for Brahmanical teaching; and so skilfully are many of these interpolations interwoven with the story, that it is frequently impossible to narrate the one, without referring to the other, however irrelevant the matter may be to the main subject in hand.' It appears to me, I own, very difficult to accept that as a satisfactory argument, amounting, as it does, to no more than this--that 'interpolations,' which must needs be referred to in narrating the main story even to make it intelligible, are nevertheless to be regarded 'as evidently the product of a Brahmanical age 1,' and presumably also a later age, because, forsooth, they are irrelevant and incongruous according to the 'tastes and ideas 1'--not of the time, be it remembered, when the 'main story' is supposed to have been written, but--of this enlightened nineteenth century. The support, too, which may be supposed to be derived by this argument from the allegation that there has been an attempt to Brahmanize, so to say, the

p. 5

history of the great war, appears to me to be extremely weak, so far as the Gîtâ is concerned. But that is a point which will have to be considered more at large in the sequel 1.

While, however, I am not prepared to admit the cogency of Mr. Wheeler's arguments, I am not, on the other hand, to be understood as holding that the Gîtâ must be accepted as a genuine part of the original Mahâbhârata. I own that my feeling on the subject is something akin to that of the great historian of Greece regarding the Homeric question, a feeling of painful diffidence regarding the soundness of any conclusion whatever. While it is impossible not to feel serious doubts about the critical condition of the Mahâbhârata generally; while, indeed, we may be almost certain that the work has been tampered with from time to time 2; it is difficult to come to a satisfactory conclusion regarding any particular given section of it. And it must be remembered, also, that the alternatives for us to choose from in these cases are not only these two, that the section in question may be a genuine part of the work, or that it may be a later interpolation: but also this, as suggested recently, though not for the first time, by Mr. Freeman 3 with reference to the Homeric question, that the section may have been in existence at the date of the original epos, and may have been worked by the author of the epos into his own production. For that absence of dread, 'either of the law or sentiment of copyright,' which Mr. Freeman relies upon with regard to a primitive Greek poet, was by no means confined to the Greek people, but may be traced amongst us also. The commentator Madhusûdana Sarasvatî likens the Gîtâ to those dialogues which occur in sundry Vedic works, particularly the Upanishads 4. Possibly--I will not use a stronger word--possibly the Gîtâ may

p. 6

have existed as such a dialogue before the Mahâbhârata, and may have been appropriated by the author of the Mahâbhârata to his own purposes 1. But yet, upon the whole, having regard to the fact that those ideas of unity on which Mr. Wheeler and others set so much store are scarcely appropriate to our old literature; to the fact that the Gîtâ fits pretty well into the setting given to it in the Bhîshma Parvan; to the fact that the feeling of Arguna, which gives occasion to it, is not at all inconsistent, but is most consonant, with poetical justice; to the fact that there is not in the Gîtâ, in my judgment, any trace of a sectarian or 'Brahmanizing' spirit 2, such as Mr. Wheeler and also the late Professor Goldstücker 3 hold to have animated the arrangers of the Mahâbhârata; having regard, I say, to all these facts, I am prepared to adhere, I will not say without diffidence, to the theory of the genuineness of the Bhagavadgîtâ as a portion of the original Mahâbhârata.

The next point to consider is as to the authorship of the Gîtâ. The popular notion on this subject is pretty well known. The whole of the Mahâbhârata is, by our traditions, attributed to Vyâsa, whom we have already noticed as a relative of the Kauravas and Pândavas; and therefore the Bhagavadgîtâ, also, is naturally affiliated to the same author. The earliest written testimony to this authorship, that I can trace, is to be found in Sankarâkârya's commentary on the Gîtâ 4 itself and on the Brihadâranyakopanishad 5. To a certain extent, the mention of Vyâsa in the body of the Gîtâ would, from a historic standpoint, seem to militate against this tradition. But I have not seen in any of the commentaries to which I have had access, any consideration of this point, as there is of the mention in some

p. 7

[paragraph continues] Smritis and Sûtras of the names of those to whom those Smritis and Sûtras are respectively ascribed 1.

We must now leave these preliminary questions, unluckily in a state far from satisfactory, and proceed to that most important topic--the date when the Gîtâ was composed, and the position it occupies in Sanskrit literature. We have here to consider the external evidence bearing on these points, which is tantalizingly meagre; and the internal evidence, which is, perhaps, somewhat more full. And taking first the internal evidence, the various items falling under that head may be marshalled into four groups. Firstly, we have to consider the general character of the Gîtâ with reference to its mode of handling its subject. Secondly, there is the character of its style and language. Thirdly, we have to consider the nature of the versification of the Gîtâ. And fourthly and lastly, we must take note of sundry points of detail, such as the attitude of the Gîtâ towards the Vedas and towards caste, its allusions to other systems of speculation, and other matters of the like nature. On each of these groups, in the order here stated, we now proceed to make a few observations.

And first about the manner in which the Gîtâ deals with its subject. It appears to me, that the work bears on the face of it very plain marks indicating that it belongs to an age prior to the system-making age of Sanskrit philosophy In 1875, I wrote as follows upon this point: 'My view is, that in the Gîtâ and the Upanishads, the philosophical part has not been consistently and fully worked out. We have there the results of free thought, exercised on different subjects of great moment, unfettered by the exigencies of any foregone conclusions, or of any fully developed theory. It is afterwards, it is at a later stage of philosophical progress, that system-making arises. In that stage some thinkers interpret whole works by the light of some particular doctrines or expressions. And the result is the development of a whole multitude of philosophical sects, following the lead of those thinkers, and all professing to draw their

p. 8

doctrine from the Gîtâ or the Upanishads, yet each differing remarkably from the other 1.' Since this was written, Professor Max Müller's Hibbert Lectures have been published. And I am happy to find, that as regards the Upanishads, his view coincides exactly with that which I have expressed in the words now quoted. Professor Max Müller says: 'There is not what may be called a philosophical system in these Upanishads. They are in the true sense of the word guesses at truth, frequently contradicting each other, yet all tending in one direction 2.' Further corroboration for the same view is also forthcoming. Professor Fitz-Edward Hall, in a passage which I had not noticed before, says 3: 'In the Upanishads, the Bhagavadgîtâ, and other ancient Hindu books, we encounter, in combination, the doctrines which, after having been subjected to modifications that rendered them as wholes irreconcileable, were distinguished, at an uncertain period, into what have for many ages been styled the Sânkhya and the Vedânta.' We have thus very weighty authority for adhering to the view already expressed on this important topic. But as Professor Weber appears to have expressed an opinion 4 intended perhaps to throw some doubt on the correctness of that view, it is desirable to go a little more into detail to fortify it by actual reference to the contents of the Gîtâ, the more especially as we can thus elucidate the true character of that work. Before doing so, however, it may be pointed out, that the proposition we have laid down is one, the test of which lies more in a comprehensive review of the whole of the Gîtâ, than in the investigation of small details on which there is necessarily much room for difference of opinion.

And first, let us compare that indisputably systematized work, the current Yoga-sûtras 5, with the Bhagavadgîtâ on one

p. 9

or two topics, where they both travel over common ground. In the Gîtâ, chapter VI, stanzas 33, 34 (p. 71), we have Arguna putting what is, in substance, a question to Krishna, as to how the mind, which is admittedly 'fickle, boisterous, strong, and obstinate,' is to be brought under control--such control having been declared by Krishna to be necessary for attaining devotion (yoga)? Krishna answers by saying that the mind may be restrained by 'practice (abhyâsa) and indifference to worldly objects (vairâgya).' He then goes on to say, that devotion cannot be attained without self-restraint, but that one who has self-restraint, and works to achieve devotion, may succeed in acquiring it. Here the subject drops. There is no further explanation of 'practice' or 'indifference to worldly objects,' no exposition of the mode in which they work, and so forth. Contrast now the Yoga-sûtras. The topic is there discussed at the very outset of the work. As usual the author begins with 'Now therefore the Yoga is to be taught.' He then explains Yoga by the well-known definition 'Yoga is the restraint of the movements of the mind.' And then after pointing out what the movements of the mind are, he proceeds: 'Their restraint is by means of practice and indifference to worldly objects 1,'--the very terms, be it remarked in passing, which are used in the Bhagavadgîtâ. But having come thus far, the author of the Sûtras does not drop the subject as the author of the Gîtâ does. He goes on in this wise: Practice is the effort for keeping it steady.' 'And that becomes firmly grounded when resorted to for a long time, without interruption, and with correct conduct.' So far we have a discussion of the first requisite specified, namely, practice. Patañgali then goes on to his second requisite for mental restraint. 'Indifference to worldly objects is the consciousness of having subdued desires &c. (Vasikâra sañgñâ) which belongs to one having no longing for objects visible and those which are heard of' (from Sâstras &c., such as heaven and so forth).

p. 10

[paragraph continues] He next proceeds to distinguish another and higher species of 'indifference,' and then he goes on to point out the results of that self-restraint which is to be acquired in the mode he has expounded. That is one instance. Now take another. In chapter VI, stanza 10 and following stanzas, the Gîtâ sets forth elaborately the mode of practically achieving. the mental abstraction called Yoga. It need not be reproduced here. The reader can readily find out how sundry directions are there given for the purpose specified, but without any attempt at systematizing. Contrast the Yoga-sûtras. In the Sâdhanapâda, the section treating of the acquisition of Yoga, Patañgali states in the twenty-ninth aphorism the well-known eight elements of Yoga. Then he subdivides these elements, and expatiates on each of them distinctly, defining them, indicating the mode of acquiring them, and hinting at the results which flow from them. 'That inordinate love of subdivision,' which Dr. F. E. Hall 1 has somewhere attributed to the Hindus, appears plainly in these aphorisms, while there is not a trace of it in the corresponding passage in the Bhagavadgîtâ. In my opinion, therefore, these comparisons strongly corroborate the proposition we have laid down regarding the unsystematic, or rather non-systematic, character of the work. In the one we have definition, classification, division, and subdivision. In the other we have a set of practical directions, without any attempt to arrange them in any very scientific order. In the one you have a set of technical terms with specific significations. In the other no such precision is yet manifest. In one word, you have in the Gîtâ the germs, and noteworthy germs too, of a system 2, and you have most of the raw material of a system, but you have no system ready-made.

Let us look at the matter now from a slightly different point of view. There are sundry words used in the Bhagavadgîtâ, the significations of which are not quite identical

p. 11

throughout the work. Take, for instance, the word 'yoga,' which we have rendered 'devotion.' At Gîtâ, chapter II, stanza 48 (p. 49), a definition is given of that word. In chapter VI, the signification it bears is entirely different. And again in chapter IX, stanza 5, there is still another sense in which the word is used 1. The word 'Brahman' too occurs in widely varying significations. And one of its meanings, indeed, is quite singular, namely, 'Nature' (see chapter XIV, stanza 3). Similar observations, to a greater or less extent, apply to the words Buddhi, Âtman, and Svabhâva 2. Now these are words which stand for ideas not unimportant in the philosophy of the Bhagavadgîtâ. And the absence of scientific precision about their use appears to me to be some indication of that non-systematic character of which we have already spoken.

There is one other line of argument, which leads, I think, to the same conclusion. There are several passages in the Gîtâ which it is not very easy to reconcile with one another; and no attempt is made to harmonise them. Thus, for example, in stanza 16 of chapter VII, Krishna divides his devotees into four classes, one of which consists of 'men of knowledge,' whom, Krishna says, he considers 'as his own self.' It would probably be difficult to imagine any expression which could indicate higher esteem. Yet in stanza 46 of chapter VI, we have it laid down, that the devotee is superior not only to the mere performer of penances, but even to the men of knowledge. The commentators betray their gnostic bias by interpreting 'men of knowledge' in this latter passage to mean those who have acquired erudition in the Sâstras and their significations. This is not an interpretation to be necessarily rejected. But there is in it a certain twisting of words, which, under the circumstances here, I am not inclined to accept. And on the other hand, it must not be forgotten, that the implication fairly derivable from chapter IV, stanza 38 (pp. 62, 63), would seem to

p. 12

be rather that knowledge is superior to devotion--is the higher stage to be reached by means of devotion as the stepping-stone. In another passage again at Gîtâ, chapter XII, stanza 12, concentration is preferred to knowledge, which also seems to me to be irreconcileable with chapter VII, stanza 16. Take still another instance. At Gîtâ, chapter V, stanza 15, it is said, that 'the Lord receives the sin or merit of none.' Yet at chapter V, stanza 29, and again at chapter IX, stanza 24, Krishna calls himself 'the Lord and enjoyer' of all sacrifices and penances. How, it may well be asked, can the Supreme Being 'enjoy' that which he does not even receive?' Once more, at chapter X, stanza 29, Krishna declares that 'none is hateful to me, none dear.' And yet the remarkable verses at the close of chapter XII seem to stand in point-blank contradiction to that declaration. There through a most elaborate series of stanzas, the burden of Krishna's eloquent sermon is 'such a one is dear to me.' And again in those fine verses, where Krishna winds up his Divine Lay, he similarly tells Arguna, that he, Arguna, is 'dear' to Krishna. And Krishna also speaks of that devotee as 'dear' to him, who may publish the Mystery of the Gîtâ among those who reverence the Supreme Being 1. And yet again, how are we to reconcile the same passage about none being 'hateful or dear' to Krishna, with his own words at chapter XVI, stanza 18 and following stanzas? The language used in describing the 'demoniac' people there mentioned is not remarkable for sweetness towards them, while Krishna says positively, 'I hurl down such people into demoniac wombs, whereby they go down into misery and the vilest condition.' These persons are scarcely characterised with accuracy 'as neither hateful nor dear' to Krishna. It seems to me, that all these are real inconsistencies in the Gîtâ, not such, perhaps, as might not be explained away, but such, I think, as indicate a mind making guesses at truth., as Professor Max Müller puts it, rather than a mind elaborating a complete

p. 13

and organised system of philosophy. There is not even a trace of consciousness on the part of the author that these inconsistencies exist. And the contexts of the various passages indicate, in my judgment, that a half-truth is struck out here, and another half-truth there, with special reference to the special subject then under discussion; but no attempt is made to organise the various half-truths, which are apparently incompatible, into a symmetrical whole, where the apparent inconsistencies might possibly vanish altogether in the higher synthesis. And having regard to these various points, and to the further point, that the sequence of ideas throughout the verses of the Gîtâ is not always easily followed, we are, I think, safe in adhering to the opinion expressed above, that the Gîtâ is a nonsystematic work, and in that respect belongs to the same class as the older Upanishads.

We next come to the consideration of the style and language of the Bhagavadgîtâ. And that, I think, furnishes a strong argument for the proposition, that it belongs to an age considerably prior to the epoch of the artificial department of Sanskrit literature--the epoch, namely, of the dramas and poems. In its general character, the style impresses me as quite archaic in its simplicity. Compounds, properly so called, are not numerous; such as there are, are not long ones, and very rarely, if ever, present any puzzle in analysing. The contrast there presented with what is called the classical literature, as represented by Bâna or Dandin, or even Kâlidâsa, is not a little striking. In Kâlidâsa, doubtless, the love for compounds is pretty well subdued, though I think his works have a perceptibly larger proportion of them than the Gîtâ. But after Kâlidâsa the love for compounds goes through a remarkable development, till in later writings it may be said almost to have gone mad. Even in Bâna and Dandin, Subandhu and Bhavabhûti, the plethora of compounds is often wearisome. And the same remark applies to many of the copperplate and other inscriptions which have been recently deciphered, and some of which date from the early centuries

p. 14

of the Christian era. Take again the exuberance of figures and tropes which is so marked in the classical style. There is little or nothing of that in the Gîtâ, where you have a plain and direct style of natural simplicity, and yet a style not by any means devoid of æsthetic merit like the style of the Sûtra literature. There is also an almost complete absence of involved syntactical constructions; no attempt to secure that jingle of like sounds, which 'seems to have proved a temptation too strong even for Kâlidâsa's muse entirely to resist. But on the contrary, we have those repetitions of words and phrases, which are characteristic, and not only in Sanskrit, of the style of an archaic period 1. Adverting specially to the language as distinguished from the style of the Gîtâ, we find such words as Anta, Bhâshâ, Brahman, some of which are collected in the Sanskrit Index in this volume, which have gone out of use in the classical literature in the significations they respectively bear in the Gîtâ. The word 'ha,' which occurs once, is worthy of special note. It is the equivalent of 'gha,' which occurs in the Vedic Samhitâs. In the form 'ha' it occurs in the Brâhmanas. But it never occurs, I think, in what is properly called the classical literature. It is, indeed, found in the Purânas. But that is a class of works which occupies a very unique position. There is a good deal in the Purânas that, I think, must be admitted to be very ancient 2; while undoubtedly also there is a great deal in them that is very modern. It is, therefore, impossible to treat the use of 'ha' in that class of works as negativing an inference of the antiquity of any book where the word occurs; while its use in Vedic works and its total absence from modern works indicate such antiquity pretty strongly. We may, therefore, embody the result of this part of the discussion in the proposition, that

p. 15

the Gîtâ is removed by a considerable linguistic and chronological distance from classical Sanskrit literature. And so far as it goes, this proposition agrees with the result of our investigation of the first branch of internal evidence.

The next branch of that evidence brings us to the character of the versification of the Gîtâ. Here, again, a survey of Sanskrit verse generally, and the verse of the Gîtâ in particular, leads us to a conclusion regarding the position of the Gîtâ in Sanskrit literature, which is in strict accord with the conclusions we have already drawn. In the verse of the Vedic Samhitâs, there is almost nothing like a rigidly fixed scheme of versification, no particular collocation of long and short syllables is absolutely necessary. If we attempt to chant them in the mode in which classical Sanskrit verse is chanted, we invariably come across lines where the chanting cannot be smooth. If we come next to the versification of the Upanishads, we observe some progress made towards such fixity of scheme as we have alluded to above. Though there are still numerous lines, which cannot be smoothly chanted, there are, on the other hand, a not altogether inconsiderable number which can be smoothly chanted. In the Bhagavadgîtâ a still further advance, though a slight one, may, I think, be marked. A visibly larger proportion of the stanzas in the Gîtâ conform to the metrical schemes as laid down by the writers on prosody, though there are still sundry verses which do not so conform, and cannot, accordingly, be chanted in the regular way. Lastly, we come to the Kâvyas and Nâtakas--the classical literature. And here in practice we find everywhere a most inflexible rigidity of scheme, while the theory is laid down in a rule which says, that 'even mâsha may be changed to masha, but a break of metre should be avoided.' This survey of Sanskrit verse may, I think, be fairly treated as showing, that adhesion to the metrical schemes is one test of the chronological position of a work--the later the work, the more undeviating is such adhesion. I need not stay here to point out, how this view receives corroboration from the rules given on this subject in the standard work

p. 16

of Pingala on the Khandas Sâstra. I will only conclude this point by saying, that the argument from the versification of the Gîtâ, so far as it goes, indicates its position as being prior to the classical literature, and nearly contemporaneous with the Upanishad literature.

We now proceed to investigate the last-group of facts falling under the head of internal evidence, as mentioned above. And first as regards the attitude of the Gîtâ towards the Vedas. If we examine all the passages in the Gîtâ, in which reference is made to the Vedas, the aggregate result appears to be, that the author of the Gîtâ does not throw the Vedas entirely overboard. He feels and expresses reverence for them, only that reverence is of a somewhat special character. He says in effect, that the precepts of the Vedas are suitable to a certain class of people, of a certain intellectual and spiritual status, so to say. So far their authority is unimpeached. But if the unwise sticklers for the authority of the Vedas claim anything more for them than this, then the author of the Gîtâ holds them to be wrong. He contends, on the contrary, that acting upon the ordinances of the Vedas is an obstacle to the attainment of the summum bonum 1. Compare this with the doctrine of the Upanishads. The coincidence appears to me to be most noteworthy. In one of his recent lectures, Professor Max Müller uses the following eloquent language regarding the Upanishads 2: 'Lastly come the Upanishads; and what is their object? To show the utter uselessness, nay, the mischievousness of all ritual performances (compare our Gîtâ, pp. 47, 48, 84 3); to condemn every sacrificial act which has for its motive a desire or hope of reward (comp. Gîtâ, p. 119 4); to deny, if not the existence, at least the exceptional and exalted character of the Devas (comp. Gîtâ, pp. 76-84 5); and to teach that there is no hope of salvation and deliverance except by the individual self recognising the true and universal self, and finding rest there, where alone rest can be found 6' (comp. our Gîtâ Translation, pp. 78-83).

p. 17

The passages to which I have given references in brackets will show, that Professor Max Müller's words might all be used with strict accuracy regarding the essential teaching of the Bhagavadgîtâ. We have here, therefore, another strong circumstance in favour of grouping the Gîtâ with the Upanishads. One more point is worthy of note. Wherever the Gîtâ refers to the Vedas in the somewhat disparaging manner I have noted, no distinction is taken between the portion relates to the ritual and the portion which relates to that higher science, viz. the science of the soul, which Sanatkumâri speaks of in his famous dialogue with Nârada 1. At Gîtâ, chapter II, stanza 45, Arguna is told that the Vedas relate only to the effects of the three qualities, which effects Arguna is instructed to overcome. At Gîtâ, chapter VI, stanza 44, Arguna is told that he who has acquired some little devotion, and then exerts himself for further progress, rises above the Divine word--the Vedas. And there are also one or two other passages of the like nature. They all treat the Vedas as concerned with ritual alone. They make no reference to any portion of the Vedas dealing with the higher knowledge. If the word Vedânta, at Gîtâ, chapter XV, stanza 15 (p. 113), signifies, as it seems to signify, this latter portion of the Vedas, then that is the only allusion to it. But, from all the passages in the Gîtâ which refer to the Vedas, I am inclined to draw the inference, that the Upanishads of the Vedas, were composed at a time not far removed from the time of the composition of the Gîtâ, and that at that period the Upanishads had not yet risen to the position of high importance which they afterwards commanded. In the passage referred to at chapter XV, the word Vedântas probably signifies the Âranyakas, which may be regarded as marking the beginning of the epoch, which the composition of the Upanishads brought to its close. And it is to the close of this epoch, that I would assign the birth of the Gîtâ, which is

p. 18

probably one of the youngest members of the group to which it belongs.

It appears to me, that this conclusion is corroborated by the fact that a few stanzas in the Gîtâ are identical with some stanzas in some of the Upanishads. With regard to the epic age of Greece, Mr. E. A. Freeman has said that, in carrying ourselves back to that age, 'we must cast aside all the notions with which we are familiar in our own age about property legal or moral in literary compositions. It is plain that there were phrases, epithets, whole lines, which were the common property of the whole epic school of poetry 1.' It appears to me that we must accept this proposition as equally applicable to the early days of Sanskrit literature, having regard to the common passages which we meet with in sundry of the Vedic works, and also sometimes, I believe, in the different Purânas. If this view is correct, then the fact that the Gîtâ contains some stanzas in the very words which we meet with in some of the Upanishads, indicates, to my mind, that the conclusion already drawn from other data about the position of the Gîtâ with regard to the Upanishads, is not by any means unwarranted, but one to which the facts before us rather seem to point.

And here we may proceed to draw attention to another fact connected with the relation of the Gîtâ to the Vedas. In stanza 17 of the ninth chapter of the Gîtâ, only Rik, Sâman, and Yagus are mentioned. The Atharva-veda is not referred to at all. This omission does certainly seem a very noteworthy one. For it is in a passage where the Supreme Being is identifying himself with everything, and where, therefore, the fourth Veda might fairly be expected to be mentioned. I may add that in commenting on Sankarâkârya's remarks on this passage, Ânandagiri (and Madhusûdana Sarasvatî also)seems evidently to have been conscious of the possible force of this omission of the Atharva-veda. He accordingly says that by force of the word 'and' in the verse in question, the Atharvângirasas, or Atharva-veda must

p. 19

also be included. Are we at liberty to infer from this, that the Atharva-veda did not exist in the days when the Gîtâ was composed? The explanation ordinarily given for the omission of that Veda, where such omission occurs, namely, that it is not of any use in ordinary sacrificial matters, is one which can scarcely have any force in the present instance; though it is adequate, perhaps, to explain the words 'those who know the three branches of knowledge,' which occur only a few lines after the verse now under consideration. The commentators render no further help than has been already stated. Upon the whole, however, while I am not yet quite prepared to say, that the priority of the Gîtâ, even to the recognition of the Atharva-veda as a real Veda, may be fairly inferred from the passage in question, I think that the passage is noteworthy as pointing in that direction. But further data in explanation of the omission referred to must be awaited.

If the conclusions here indicated about the relative positions of the Gîtâ and certain Vedic works are correct, we can fairly take the second century B. C. as a terminus before which the Gîtâ must have been composed. For the Upanishads are mentioned in the Mahâbhâshya of Patañgali, which we are probably safe in assigning to the middle of that century. The epoch of the older Upanishads, therefore, to which reference has been so frequently made here, may well be placed at some period prior to the beginning of the second century B. C. The Atharva-veda is likewise mentioned by Patañgali 1, and as 'ninefold,' too, be it remembered; so that if we are entitled to draw the conclusion which has been mentioned above from chapter IX, stanza 17, we come to the same period for the date of the Gîtâ. Another point to note in this connexion is the reference to the Sâma-veda as the best of the Vedas (see p. 88). That is a fact which seems to be capable of yielding some chronological information. For the estimation in which that Veda has been held appears to have varied at different times. Thus, in the Aitareya-brâhman2, the glory

p. 20

of the Sâman is declared to be higher than that of the Rik, In the Khândogya-upanishad 1 the Sâman is said to be the essence of the Rik, which Sankara interprets by saying that the Sâman is more weighty, In the Prasna-upanishad 2, too, the implication of the passage V, 5 (in which the Sâman is stated as the guide to the Brahmaloka, while the Yagus is said to guide to the lunar world, and the Rik to the, human world) is to the same effect. And we may also mention as on the same side the Nrisimha Tâpinî-upanishad and the Vedic passage cited in the commentary of, Sankara on the closing sentence of the first khanda of that Upanishad 3. On the other side, we have the statement in Manu that the sound of the Sâma-veda is unholy; and the consequent direction that where the sound of it is heard, the Rik and Yagus should not be recited 4. We have also the passages from some of the Purânas noted by Dr. Muir in his excellent work, Original Sanskrit Texts, which point in the same direction 5. And we have further the direction in the Âpastamba Dharma-sûtra, that the Sâman hymns should not be recited where the other Vedas are being recited 6, as well as the grouping of the sound of the Sâman with various classes of objectionable and unholy noises, such as those of dogs and asses. It is pretty evident that the view of Âpastamba is based on the same theory as that of Manu. Now in looking at the two classes of authorities thus marshalled, it is plain that the Gîtâ ranges itself with those which are unquestionably the more ancient. And among the less ancient works, prior to which we may place the Gîtâ on account of the facts now under consideration, are Manu and Âpastamba. Now Manu's date is not ascertained, though, I believe, he is now generally considered to belong to about the second or third century B. C. 7 But . I

p. 21

[paragraph continues] Dr. Bühler, in the Preface to his Âpastamba in the present series, has adduced good reasons for holding that Âpastamba is prior to the third century B. C. 1, and we therefore obtain that as a point of time prior to which the Gîtâ must have been composed.

The next important item of internal evidence which we have to note, is the view taken of caste in the Bhagavadgîtâ. Here, again, a comparison of the doctrine of the Gîtâ with the conception of caste in Manu and Âpastamba is interesting and instructive. The view of Manu has been already contrasted by me with the Gîtâ in another place 2. I do not propose to dwell on that point here, as the date of Manu is far from being satisfactorily ascertained. I prefer now to take up Âpastamba only, whose date, as just now stated, is fairly well fixed by Dr. Bühler. The division of castes, then, is twice referred to in the Bhagavadgîtâ. In the first passage (p. 59) it is stated, that the division rests on differences of qualities and duties; in the second (pp. 126, 127) the various duties are distinctly stated according to the differences of qualities. Now in the first place, noting as we pass along, that there is nothing in the Gîtâ to indicate whether caste was hereditary, according to its view, whereas Âpastamba distinctly states it to be such, let us compare the second passage of the Gîtâ with the Sûtras of Âpastamba bearing on the point. The view enunciated in the Gîtâ appears to me plainly to belong to an earlier age--to an age of considerably less advancement in social and religious development. In the Gîtâ, for instance, the duties of a Brâhmana are said to be tranquillity, self-restraint, and so forth. In Âpastamba, they are the famous six duties, namely, study, imparting instruction, sacrificing, officiating at others' sacrifices, making gifts, and receiving gifts; and three others, namely, inheritance, occupancy, and gleaning ears of corn, which, it may be remarked en passant, are not stated in Manu. The former seem to my mind to point

p. 22

to the age when the qualities which in early times gave the Brâhmanas their pre-eminence in Hindu society were still a living reality 1. It will be noted, too, that there is nothing in that list of duties which has any necessary or natural connexion with any privilege as belonging to the caste. The Law lays down these duties, in the true sense of the word. In Âpastamba, on the contrary, we see an advance towards the later view on both points. You have no reference to moral and religious qualities now. You have to do with ceremonies and acts. You have under the head 'duties not mere obligations, but rights. For the duty of receiving gifts is a right, and so is the duty of teaching others and officiating at others' sacrifices; as we know not merely from the subsequent course of events, but also from a comparison of the duties of Brâhmanas on the one hand, and Kshatriyas, Vaisyas, and Sûdras on the other, as laid down by Manu and Âpastamba themselves. Âpastamba's rules, therefore, appear to belong to the time when the Brâhmanas had long been an established power, and were assuming to themselves those valuable privileges which they have always claimed in later times. The rules of the Gîtâ, on the other hand, point to a time considerably prior to this--to a time when the Brâhmanas were by their moral and intellectual qualities laying the foundation of that preeminence in Hindu society which afterwards enabled them to lord it over all castes. These observations mutatis mutandis apply to the rules regarding. the other castes also. Here again, while the Gîtâ still insists on the inner qualities, which properly constitute the military profession, for instance, the rules of Âpastamba indicate the powerful influence of the Brâhmanas 2. For, as stated before, officiating at others' sacrifices, instructing others, and receiving presents, are here expressly prohibited to Kshatriyas as also to Vaisyas. The result of that is, that the Brâhmanas become indispensable to the Kshatriyas and Vaisyas, for

p. 23

upon both the duty of study, of offering sacrifices, and making gifts and presents is inculcated. In his outline of the History of Ancient Religions, Professor Tiele, speaking of the 'increasing influence of the Brâhmans,' writes as follows: 'Subject at first to the princes and nobles, and dependent on them, they began by insinuating themselves into their favour, and representing it as a religious duty to show protection and liberality towards them. Meanwhile they endeavoured to make themselves indispensable to them, gradually acquired the sole right to conduct public worship, and made themselves masters of instruction 1'. And after pointing out the high position thus achieved by the Brâhmans, and the low position of the Kândâlas and others of the inferior castes, he adds: 'Such a position could not long be endured; and this serves to explain not only the rise of Buddhism, but also its rapid diffusion, and the radical revolution which it brought about 2.' To proceed, however, with our comparison of the Gîtâ and Âpastamba. The superiority distinctly claimed by the latter for the Brâhmana is not quite clearly brought out in the Gîtâ. 'Holy Brâhmanas and devoted royal saints' are bracketed together at p. 86; while the Kshatriyas are declared to have been the channel of communication between the Deity and mankind as regards the great doctrine of devotion propounded by the Bhagavadgîtâ. That indicates a position for the Kshatriyas much more like what the Upanishads disclose 3, than even that which Âpastamba assigns to them. The fact is further noteworthy, that in the Gîtâ each caste has its own entirely distinct set of duties. There is no overlapping, so to say. And that is a circumstance indicating a very early stage in the development of the institution 4. Besides, as already indicated,

p. 24

the duties laid down by Âpastamba and Manu as common to Kshatriyas and Vaisyas are the very duties which make those castes dependent to a very great extent on the Brâhmanas. Lastly, it is not altogether unworthy of note, that in the elaborate specification of the best of every species which we find in chapter X, the Brâhmana is not mentioned as the best of the castes, there is nothing to indicate the notion contained in the well-known later verse, 'The Brâhmana is the head of the castes.' On the contrary, the ruler of men is specified as the highest among men 1, indicating, perhaps, a state of society such as that described at the beginning of the extract from Professor Tiele's work quoted above.

We come now to another point. What is the position of the Gîtâ in regard to the great reform of Sâkya Muni, The question is one of much interest, having regard particularly to the remarkable coincidences between Buddhistic doctrines and the doctrines of the Gîtâ to which we have drawn attention in the foot-notes to our translation. But the materials for deciding the question are unhappily not forthcoming. Professor Wilson, indeed, thought that there was an allusion to Buddhism in the Gîtâ 2. But his idea was based on a confusion between the Buddhists and the Kârvâkas or materialists 3. Failing that allusion, we have nothing very tangible but the unsatisfactory 'negative argument' based on mere non-mention of Buddhism in the Gîtâ. That argument is not quite satisfactory to my own mind, although, as I have elsewhere pointed out 4, some of the ground occupied by the Gîtâ is common to it with Buddhism, and although various previous thinkers Are alluded to directly or indirectly in the Gîtâ. There is, however, one view of the facts of this question, which appears to me to corroborate the conclusion deducible by means of the negative argument here referred to. The

p. 25

main points on which Buddha's protest against Brahmanism rests seem to be the true authority of the Vedas and the true view of the differences of caste. On most points of doctrinal speculation, Buddhism is still but one aspect of the older Brahmanism 1. The various coincidences to which we have drawn attention show that, if there is need to show it. Well now, on both these points, the Gîtâ, while it does not go the whole length which Buddha goes, itself embodies a protest against the views current about the time of its composition. The Gîtâ does not, like Buddhism, absolutely reject the Vedas, but it shelves them. The Gîtâ does not totally root out caste. It places caste on a less untenable basis. One of two hypotheses therefore presents itself as a rational theory of these facts. Either the Gîtâ and Buddhism were alike the outward manifestation of one and the same spiritual upheaval which shook to its centre the current religion, the Gîtâ being the earlier and less thorough-going form of it; or Buddhism having already begun to tell on Brahmanism, the Gîtâ was an attempt to bolster it up, so to say, at its least weak points, the weaker ones being altogether abandoned. I do not accept the latter alternative, because I cannot see any indication in the Gîtâ of an attempt to compromise with a powerful attack on the old Hindu system; while the fact that, though strictly orthodox, the author of the Gîtâ still undermines the authority, as unwisely venerated, of the Vedic revelation; and the further fact, that in doing this, he is doing what others also had done before him or about his time; go, in my opinion, a considerable way towards fortifying the results of the negative argument already set forth. To me Buddhism is perfectly intelligible as one outcome of that play of thought on high spiritual topics, which in its other, and as we may say, less thorough-going manifestations, we see in the Upanishads and the Gîtâ 2. But assume that Buddhism was

p. 26

a protest against Brahmanism prior to its purification and elevation by the theosophy of the Upanishads, and those remarkable productions of ancient Indian thought become difficult to account for. Let us compare our small modem events with those grand old occurrences. Suppose our ancestors to have been attached to the ceremonial law of the Vedas, as we are now attached to a lifeless ritualism, the Upanishads and the Gîtâ might be, in a way, comparable to movements like that of the late Raja Rammohun Roy. Standing, as far as possible, on the antique ways, they attempt, as Raja Rammohun attempted in these latter days, to bring into prominence and to elaborate the higher and nobler aspects of the old beliefs. Buddhism would be comparable to the further departure from old traditions which was led by Babu Keshub Chander Sen. The points of dissent in the olden times were pretty nearly the same as the points of dissent now. The ultimate motive power also was in both cases identical--a sense of dissatisfaction in its integrity with what had come down from old times encrusted with the corruptions of years. In this view the old system, the philosophy of the Upanishads and the Gîtâ, and the philosophy of Buddha, constitute a regular intelligible progression. But suppose the turn events took was different, as is supposed by the alternative theory indicated above. Suppose Babu Keshub's movement was chronologically prior, and had begun to tell on orthodox, society. Is it likely, that then one of the orthodox party would take up the position which Rammohun Roy took? Would he still rely on old authorities, but with sundry qualifications, and yet earnestly assail the current forms of orthodoxy? I do not think so. I think the true view to be, as already stated, very different. The Upanishads, with the Gîtâ, and the precepts of Buddha

p. 27

appear to me to be the successive 1 embodiments of the spiritual thought of the age, as it became more and more dissatisfied with the system of mere ceremonial then dominant.

There are several other points of much interest in the Bhagavadgîtâ, such as the reference to the Sânkhya, and Yoga; the place assigned to the Mârgasîrsha month; the allusion to the doctrines of materialism; the nearly entire coincidence between a stanza of the Gîtâ and one in the Manu Smriti. But in the present state of our knowledge, I do not. think that we can extract any historical results from any of them. Without dwelling on them any further 2, therefore, I will only state it as my opinion that the Sânkhya, and Yoga of the Gîtâ are not identical with the systems known to us under those names, and that the Manu Smriti has probably borrowed from the Gîtâ the stanza common to the two works.

We now proceed to a discussion of some of the external evidence touching the age of the Bhagavadgîtâ. It is, of course, unnecessary to consider any evidence of a date later than the eighth century A. C., that being the date generally received, though not on very strong grounds, as the date of Sankarâkârya, the celebrated commentator of the Gîtâ 3 For the period prior to that limit, the first testimony to consider is that of Bânabhatta, the author of the Kâdambarî. The date of Bâna is now fairly well settled as the middle of the seventh century A. C. The doubt which the late Dr. Bhâu Dâjî had cast upon its correctness 4, by impugning the received date of king Harshavardhana, appears to me to have been satisfactorily disposed of by the paper of

p. 28

my friend Professor R. G. Bhândârkar on the Kâlukya dates 1. In the Kâdambarî, then, we have testimony to the existence of the Bhagavadgîtâ in the middle of the seventh century A. C. For in that work, which, as is well known, abounds with equivoques, we have a passage which compares the royal palace to the Mahâbhârata, both being 'Anantagitâkarnanânanditanaram 2,' which, as applied to the royal palace, means 'in which the people were delighted by hearing innumerable songs;' and as applied to the Mahâbhârata means 'in which Arguna was delighted at hearing the Anantagîtâ.' Anantagîtâ is evidently only another name here for Bhagavadgîtâ. The conclusion deducible from this fact is not merely that the Gîtâ existed, but that it existed as a recognised portion of the Bhârata, in the seventh century A. C. Now the Kâdambarî shows, in numerous passages, in what high esteem the Mahâbhârata was held in its days. The queen Vilâsavatî used to attend at those readings and expositions of the Mahâbhârata, which have continued down to our own times; and it was even then regarded as a sacred work of extremely high authority, in the same way as it is now. It follows., therefore, that the Gîtâ must have been several centuries old ill the time of Bânabhatta.

Prior in time to Bâna is the Indian Shakespeare, Kâlidâsa, as he is referred to in Bânabhatta's Harshakarita 3, and also in a copperplate inscription of the early part of the seventh century, as a poet who had then already acquired a high reputation 4. Unfortunately, it is not yet possible to fix exactly the date at which Kâlidâsa flourished. Still, I think, we have pretty satisfactory evidence to show that the middle of the fifth century A. C. is the very latest date to which he can be referred. In a small tract (written by me in 1873), discussing Professor Weber's theory about the Râmâyana, I have pointed out 5 that the Pañkatantra

p. 29

quotes from Kâlidâsa a passage which there is good reason to believe formed part of the Pañkatantra when it was translated for king Nushirvan of Persia about the beginning of the sixth century A. C. 1 Allowing for the time required to raise Kâlidâsa to the position of being cited as an authority, and for the time required for the spread of the fame of an Indian work to Persia in those early days, I think, that the middle of the fifth century is a date to which Kâlidâsa cannot well have been subsequent. Now in the works of Kâlidâsa we have some very remarkable allusions to the Bhagavadgîtâ. It is not necessary to go through all these allusions. I will only mention the most remarkable, one from the Raghuvamsa, and one from the Kumârasambhava. In Raghu, canto X, stanza 67, the gods addressing Vishnu say, 'There is nothing for you to acquire which has not been acquired. The one motive in your birth and work is the good of the worlds.' The first sentence here reminds one at once of Gîtâ, chapter III, stanza 22, the coincidence with which in sense as well as expression is very striking. The second sentence contains the words 'birth and work,' the precise words employed at Gîtâ IV, 9; and the idea of 'good of the worlds' is identical with the idea expressed in Gîtâ III, 20-24, the words only in which it is clothed being different. Couple this passage with the one from Kumârasambhava, canto VI, 67, where the seven Rishis say to the Himâlaya mountain, 'Well hast thou been called Vishnu in a firmly-fixed form.' The allusion there to the Gîtâ, chapter X, stanza 25 (p. 89), is, I venture to think, unmistakable. The word 'firmly-fixed' is identical in both passages; the idea is identical, and Mallinâtha refers to the passage in the Gîtâ as the authority which Kâlidâsa had in view. It follows, therefore, that the Gîtâ must be prior to Kâlidâsa's time. It may be added, that Kâlidâsa in his Raghu XV, 67, cites Manu as an authority

p. 30

for the proposition that a king must protect all castes and all orders or âsramas. Manu, therefore, must have lived considerably earlier than Kâlidâsa, and the Gîtâ, as we have already argued, must be considerably earlier, not only than Manu, but also than his predecessor Âpastamba. The Gîtâ, may, therefore, be safely said to belong to a period several centuries prior to the fifth century A. C.

The next piece of external evidence is furnished by the Vedânta-sûtras of Bâdarâyana. In several of those Sûtras, references are made to certain Smritis as authorities for the propositions laid down. Take, for instance, I, 2, 6, or I, 3, 23, and many others. Now three of these sûtras are very useful for our present purpose. The first we have to consider is Sûtra II, 3, 45. The commentators Sankarâkârya, Râmânuga, Madhva, and Vallabha 1 are unanimous in understanding the passage in Gîtâ, chapter XV, stanza 7 (p. 112), to be the one there referred to by the words of the Sûtra, which are, 'And it is said in a Smriti.' Now a glance at the context of the Sûtra will, I think, satisfy us that the commentators, who are unanimous though representing different and even conflicting schools of thought, are also quite right. Sûtra 43, in the elliptical language characteristic of that branch of our literature, says, 'A part, from the statement of difference, and the reverse also; some lay down that it is a fisherman or a cheat.' Sûtra 44 runs thus, 'And also from the words of the Mantra.' And then comes Sûtra 45 as set out above. It is plain, that the Sûtra No. 45 indicates an authority for something not specified, being regarded as part of some other thing also not specified. Now the discussion in previous Sûtras has been about the soul; so we can have little difficulty in accepting the unanimous interpretation of the commentators, that the proposition here sought to be made out is that the individual soul is part of the Supreme Soul, which is the proposition laid down in the Gîtâ in the passage referred to. The

p. 31

next Sûtra to refer to is IV, 1, 10. I shall not set forth the other relevant Sûtras here as in the preceding case. I only state that the three commentators, Sankara, Râmânuga, and Madhva, agree that the Gîtâ is here referred to, namely, chapter VI, stanza ii seq. Vallabha, however, I am bound to add, does not agree with this, as he interprets the Sûtra in question and those which precede and follow as referring to an entirely different matter. If I may be permitted to say so, however, I consider his interpretation not so satisfactory as that of the three other and older commentators. Lastly, we come to Sûtra IV, 2-19. On this, again, all the four commentators are unanimous, and they say that Gîtâ, chapter VIII, stanza 24 seq. (p. 80), is the authority referred to. And I think there can be very little doubt that they are right. These various pieces of evidence render it, I think, historically certain, that the Gîtâ must be considerably prior to the Vedânta-sûtras; and that the word Brahma-sûtras, which occurs at Gîtâ, chapter XIII, stanza 4 (p. 102), is correctly interpreted by the commentators as not referring to the Vedânta-sûtras, which are also called Brahma-sûtras, but to a different subject altogether 1. When were the Vedânta-sûtras composed? The question must at once be admitted to be a difficult one; but I think the following considerations will show that the date of those Sutras must, at the latest, be considerably earlier than the period which we have already reached in this part of our investigation. We may take it as fairly well settled, that Bhatta Kumârila, the celebrated commentator of the Pûrva Mîmâmsâ school, flourished not later than the end of the seventh century A. C. 2 A considerable time prior to him must be placed the great commentator on the Mîmâmsâ-sûtras, namely, Sabarasyâmin. If we may judge from the style of his great commentary, he cannot have flourished much later than Patañgali, who may now be taken as historically proved to

p. 32

have flourished about 140 B. C. 1 Now a considerable time must have intervened between Sabarasyâmin and another commentator on the Pûrva Mîmâmsâ, whom Sabara quotes with the highly honorific title Bhagavân, the Venerable, namely, Upavarsha. Upavarsha appears from Sankara's statement to have commented on the Vedânta-sûtras 2. We have thus a long catena of works from the seventh century A. C., indicating a pretty high antiquity for the Vedanta-sûtras, and therefore a higher one for the Bhagavadgîtâ. The antiquity of the Vedânta-sûtras follows also from the circumstance, which we have on the testimony of Râmânuga, repeated by Mâdhavâkârya, that a commentary on the Sûtras was written by Baudhâyanâkârya 3, which commentary Râmânuga says he followed. Baudhâyanâ's date is not accurately settled. But he appears to be older than Âpastamba, whose date, as suggested by Dr. Bühler, has already been mentioned 4. The Vedânta-sûtras, then, would appear to be at least as old as the fourth century B. C.; if the information we have from Râmânuga may be trusted. A third argument may be mentioned, bearing on the date of the Vedânta-sûtras. In Sûtra 110 of the third Pâda of the fourth Adhyâya of Pânini's Sûtras, a Pârâsarya is mentioned as the author of a Bhikshu-sûtra. Who is this Pârâsarya, and what the Bhikshu-sûtra? Unluckily Patañgali gives us no information on this head, nor does the Kâsikâ Vritti. But a note of Professor Târânâtha Tarkavâkaspati, of Calcutta, says that Pârâsarya is Vyâsa, and the Bhikshu-sûtra is the Vedânta-sûtra 5. If this is correct, the Vedânta-sûtras go very far indeed into antiquity. For Pânini can certainly not be assigned to a later date than the fourth century B. C., while that learned

p. 33

scholar, Professor Goldstücker, on grounds of considerable strength, assigned him to a much earlier date 1. The question thus comes to this, Is the remark of Professor Târânâtha, above set out, correct? I find then, from enquiries made of my venerable and erudite friend Yagñesvar Sâstrin, the author of the Âryavidyâsudhâkara, that the note of Târânâtha is based on the works of Bhattogî Dîkshita, Nâgogî Bhatta, and Gñânendra Sarasvatî, who all give the same interpretation of the Sûtra in question. It is certainly unfortunate that we have no older authority on this point than Bhattogî. The interpretation is in itself not improbable. Vyâsa is certainly by the current tradition 2 called the author of the Vedânta-sûtras, and also the son of Parâsara. Nor is Bhikshu-sûtra a name too far removed in sense from Vedânta-sûtra, though doubtless the former name is not now in use, at all events as applied to the Sûtras attributed to Bâdarâyana, and though, it must also be stated, a Bhikshu-sûtra Bhâshya Vârtika is mentioned eo nomine by Professor Weber as actually in existence at the present day 3. Taking all things together, therefore, we may provisionally understand the Bhikshu-sûtra mentioned by Pânini to be identical with the Vedânta-sûtras. But even apart from that identification, the other testimonies we have adduced prove, I think, the high antiquity of those Sûtras, and consequently of the Bhagavadgîtâ.

We have thus examined, at what, considering the importance and difficulty of the subject, will not; I trust, be regarded as unreasonable length, some of the principal pieces of internal and external evidence touching the age of the Bhagavadgîtâ and its position in Sanskrit literature. Although, as stated at the very outset, the conclusions we have deduced in the course of that examination are not all such as at once to secure acceptance, I venture to think that we have now adequate grounds for saying, that the various .and independent lines of investigation, which we have pursued, converge to this point, that the Gîtâ, on numerous and

p. 34

essential topics, ranges itself as a member of the Upanishad group, so to say; in Sanskrit literature. Its philosophy, its mode of treating its subject, its style, its language, its versification, its opinions on sundry subjects of the highest importance, all point to that one conclusion. We may also, I think, lay it down as more than probable, that the latest date at which the Gîtâ can have been composed, must be earlier than the third century B. C., though it is altogether impossible to say at present how much earlier. This proposition, too, is supported by the cumulative strength of several independent lines of testimony.

Before closing this Introduction, it is desirable to add a word concerning the text of the Bhagavadgîtâ. The religious care with which that text has been preserved is very worthy of note. Schlegel and Lassen 1 have both declared it as their opinion, that we have the text now almost exactly in the condition in which it was when it left the hands of the author. There are very few real various readings, and some of the very few that exist are noted by the commentators. Considering that the Mahâbhârata must have been tampered with on numerous occasions, this preservation of the Gîtâ is most interesting. It doubtless indicates that high veneration for it which is still felt, and has for long been felt, by the Hindus, and which is embodied in the expression used in the colophons of the MSS. describing the Gîtâ as the 'Upanishad sung by God 2.' In view of the facts and deductions set forth in this essay that expression existing as, I believe, it does, almost universally in Indian MSS. of the Gîtâ, is not altogether devoid of historical value.

Schlegel draws attention to one other circumstance regarding the text of the Gîtâ, which is also highly interesting, namely, that the number of the stanzas is exactly 700.

p. 35

[paragraph continues] Schlegel concludes that the author must have fixed on that number deliberately, in order to prevent, as far as be could, all subsequent interpolations 1. This is certainly not unlikely; and if the aim of the author was such as Schlegel suggests, it has assuredly been thoroughly successful. In the chapter of the Mahâbhârata immediately succeeding the eighteenth chapter of the Gîtâ, the extent of the work in slokas is distinctly stated. The verses in which this is stated do not exist in the Gauda or Bengal recension, and are doubtless not genuine. But, nevertheless, they are interesting, and I shall reproduce them here. 'Kesava spoke 620 slokas, Arguna fifty-seven, Sañgaya sixty-seven, and Dhritarâshtra one sloka; such is the extent of the Gîtâ.' It is very difficult to account for these figures. According to them, the total number of verses in the Gîtâ would be 745, whereas the number in the current MSS., and even in the Mahâbhârata itself, is, as already stated, only 700 2. I cannot suggest any explanation whatever of this discrepancy.

In conclusion, a few words may be added regarding the general principles followed in the translation contained in this volume. My aim has been to make that translation as close and literal a rendering as possible of the Gîtâ, as interpreted by the commentators Sankarâkârya, Srîdharasvâmin, and Madhusûdana Sarasvatî. Reference has also been frequently made to the commentary of Râmânugâ-kârya, and also to that of Nîlakantha, which latter forms part of the author's general commentary on the Mahâbhârata. In some places these commentators differ among themselves, and then I have made my own choice. The foot-notes are mainly intended to make clear that which necessarily remains obscure in a literal translation. Some of the notes, however, also point out the parallelisms existing between the Gîtâ and other works, principally the Upanishads and the Buddhistic Dhammapada and Sutta Nipâta. Of the latter

p. 36

[paragraph continues] I have not been able to procure the original Pâlî; I have only used Sir M. C. Swamy's translation. But I may here note, that there are some verses, especially in the Salla Sutta (see pp. 124-127 of Sir M. C. Swamy's book), the similarity of which, in doctrine and expression, to some of the verses of the Gîtâ is particularly striking. The analogies between the Gîtâ and the Upanishads have been made the basis of certain conclusions in this Introduction. Those between the Gîtâ and these Buddhistic works are at present, to my mind, only interesting; I am unable yet to say whether they may legitimately be made the premises for any historical deductions.

There are two indexes: the first a general index of matters, the second containing the principal words in the Gîtâ which may prove useful or interesting for philological, historical, or other kindred purposes.


1:1 Hibbert Lectures, p. 131.

1:2 Lectures on the History of Modem Philosophy (translated by O. W. Wight), vol. i, pp. 49, 50. At p. 433 seq. of the second volume, M. Cousin gives a general view of the doctrine of the Gîtâ. See also Mr. Maurice's and Ritter's Histories of Philosophy.

2:1 Ex. gr. Sârîraka Bhâshya, vol. ii, p. 840. It is also often cited as a Smriti, ibid. vol. i, p. 152.

2:2 See inter alia Sârîraka Bhâshya, vol. i, p. 455, vol. ii, p. 687, and Colebrooke's Essays, vol. i, p. 355 (Madras); Lassen's edition of the Gîtâ, XXXV.

3:1 The whole story is given in brief by the late Professor Goldstücker in the Westminster Review, April 1868, p. 392 seq. See now his Literary Remains, II, 104 seq.

3:2 History of India, vol. i, p. 293.

4:1 History of India, vol. i, p. 288,; and compare generally upon this point the remarks in Gladstone's Homer, especially vol. i, p. 70 seq.

5:1 Infra, p. 21 seq.

5:2 Compare the late Professor Goldstücker's remarks in the Westminster Review for April 1868, p. 389.

5:3 Contemporary Review (February 1879).

5:4 Madhusûdana mentions the dialogue between Ganaka and Yâgñavalkya as a specific parallel.

6:1 See to this effect M. Fauriel, quoted in Grote's Greece, II, 195 (Cabinet ed.).

6:2 Compare also Weber's History of Indian Literature (English translation), p. 187. The instruction, however, as to 'the reverence clue to the priesthood' from 'the military caste,' which is there spoken of appears to me to be entirely absent from the Gîtâ; see p. 21 seq. infra.

6:3 Westminster Review, April 1868, p. 388 seq.; and Remains, I, 104, 105.

6:4 P. 6 (Calcutta ed., Samvat, 1927).

6:5 p. 841 (Bibl. Indic. ed.); also Svetâsvatara, p. 278.

7:1 See, as to this, Colebrooke's Essays, vol. i, p. 328 (Madras).

8:1 See the Introductory Essay to my Bhagavadgîtâ, translated into English blank verse, p. lxvii. See also Goldstücker's Remains, I, 48, 77; II, 10.

8:2 p. 317; cf. also p. 338.

8:3 Preface to Sânkhya Sâra, p. 7 (Bibl. Indic. ed.)

8:4 History of Indian Literature, p. 28.

8:5 Are we to infer from the circumstance mentioned in Weber's History of p. 9 Indian Literature (p. 223, note, 235), that the author of these Sûtras was older than Buddha?

9:1 Sûtra, 12, Abhyâsa-vairâgyâbhyâm tannirodhah.

10:1 In the Preface to his Sânkhya Sâra, I think.

10:2 This is all that we can infer from the few cases of division and classification which we do meet with in the Gîtâ. A subject like that treated of in this work could not well he discussed without some classifications &c.

11:1 In chapter X the word occurs in two different senses in the same stanza (st. 7).

11:2 Compare the various passages, references to which are collected in the Sanskrit Index at the end of this volume.

12:1 And see, too, chapter VII, stanza 17, where the man of knowledge is declared to he 'dear' to Krishna.

14:1 Compare Muir, Sanskrit Texts, vol. i, p. 5. See, too, Goldstücker's Remains, I, 177.

14:2 This opinion, which I had expressed as long ago as 1874 in the Introduction to my edition of Bhartrihari's Satakas, is, I find, also held by Dr. Bühler; see his Introduction to Âpastamba in this series, p. xx seq., note. Purânas are mentioned in the Sutta Nipâta (p. 115), as to the date of which, see inter alia Swamy's Introduction, p. xvii.

16:1 Compare the passages collected under the word Vedas in our Index.

16:2 Hibbert Lectures, p. 340 seq.

16:3 II, 42-45; IX, 20, 21.

16:4 XVII, 12.

16:5 VII, 21-23; IX, 23-24.

16:6 VIII, 14-16; IX, 39-33.

17:1 See Khândogya-upanishad, p. 473, or rather I ought to have referred to the Mundaka-upanishad, where the superiority and inferiority is more distinctly stated in words, pp. 266, 267.

18:1 Contemporary Review, February 1879.

19:1 See also Sutta Nipâta, p. 115.

19:2 Haug's edition, p. 68.

20:1 Bibl. Ind. ed. 12

20:2 Bibl. Ind. ed., p. 221 seq.

20:3 Bibl. Ind. ed.: p. 11.

20:4 Chapter IV, stanzas 123, 124.

20:5 Vol. iii (2nd ed.), p. 11 seq. Cf. Goldstücker's Remains, I, 4, 28, 366; II, 67.

20:6 Âpastamba (Bühler's ed.) I, 317, 18 (pp. 38, 39 in this series); see further on this point Mr. Burnell's Devatâdhyâya-brâhmana, Introd., pp. viii, ix, and notes.

20:7 Professor Tiele (History of Ancient Religions, p. 127) considers the 'main features' of Manu to be 'pre-Buddhistic.'

21:1 P. xxxv.

21:2 See the introductory Essay to my Bhagavadgîtâ in English verse, published in 1875, p. cxii.

22:1 The remarks in the text will show how little there is in the Gîtâ of that 'Brahmanizing' which has been shortly noticed on a previous page.

22:2 As to the Kshatriyas the contrast with Manu's rules is even stronger than with Âpastamba's. See our Introduction to the Gîtâ in English verse, p. cxiii.

23:1 P. 120.

23:2 Pp. 129, 130.

23:3 See p. 58 intra; and compare with this Weber's remarks on one of the classes into which he divides the whole body of Upanishads, History of Indian Literature, p. 165. See also Muir, Sanskrit Texts, vol. i, p. 508; Max Müller, Upanishads, vol. i, p. lxxv.

23:4 Cf. Sutta Nipâta, p. 32; and also Mr. Davids' note on that passage in his Buddhism, p. 131.

24:1 P. 89 infra.

24:2 Essays on Sanskrit Literature, vol. iii, p. 150.

24:3 See our remarks on, this point in the Introductory Essay to our Gîtâ in verse, p. ii seq.

24:4 Introduction to Gîtâ in English verse, p. v seq.

25:1 Cf. Max Müller's Hibbert Lectures, p. 137; Weber's Indian Literature, pp. 288, 289; and Mr. Rhys Davids' excellent little volume on Buddhism, p. 151; and see also p. 83 of Mr. Davids' book.

25:2 Cf. Weber's History of Indian Literature, p. 285. in Mr. Davids' Buddhism, p. 94, we have a noteworthy extract from a standard Buddhistic work, touching p. 26 the existence of the soul. Compare that with the corresponding doctrine in the Gîtâ. It will be found that the two are at one in rejecting the identity of the soul with the senses &c. The Gîtâ then goes an to admit a soul separate from these. Buddhism rejects that also, and sees nothing but the senses.

27:1 The word Brahma-nirvâna, which occurs so often at the close of chapter V and also at chapter II, 72, seems to me to indicate that nirvana had not yet become technically pinned down, so to say, to the meaning which Buddhism subsequently gave to it, as the name of what it deemed the summum bonum. Nirvâna by itself occurs at VI, 15.

27:2 See some further remarks on these points in my Introduction to the Gîtâ in verse.

27:3 Professor Tiele (History of Ancient Religions, p. 140) says Sankara was born in 798 A.D.; on. the authority, I presume, of the Âryavidyâsudhâkara, p. 226.

27:4 Journal of the Bombay Branch of the Royal Asiatic Society, vol. viii, p. 250; and see, too, Indian Antiquary, vol. vi, p. 61. Dr. Bühler).

28:1 Journal of the Bombay Branch of the Royal Asiatic Society, vol. xiv, p. 16 seq.

28:2 P. 182 (Târânâtha's ed.)

28:3 See F. E. Hall's Vâsavadattâ, p. 14 note.

28:4 See Indian Antiquary, vol. v, p. 70.

28:5 'Was the Râmâyana copied from Homer?' See pp. 36-59.

29:1 Cf. Colebrooke's Essays, vol. ii. p. 166, seq. It may be remarked that this argument is not affected by the attempt to distinguish the Kâlidâsa of the Sakuntalâ from the Kâlidâsa of the Raghuvamsa. Because the work cited in the Pañkatantra is the Kumârasambhava, which indisputably belongs to the same author as the Raghuvamsa.

30:1 I am indebted to Professor M. M. Kunte for a loan of Vallabhâkârya's commentary on the Sûtras noted in the text. I had not seen it in 1875, when I last discussed this question.

31:1 Cf. Weber's Indian Literature, p. 241. See also Lassen's Preface to his edition of Schlegers Gîtâ, XXXV. Râmânuga takes the other view.

31:2 See Burnell's Sâmavidhâna-brâhmana, Introduction, p. vi note.

32:1 The authorities are collected in our edition of Bhartrihâri (Bombay Series of Sanskrit Classics), Introd. p. xi note. See also Bühler's Âpastamba in this series, Introd. p. xxviii.

32:2 See Colebrooke's Essays, vol. i, p. 332. An Upavarsha is mentioned in the Kathâsaritsâgara as living in the time of king Nanda, and having Pânini, Kâtyâyana, and Vyâdi for his pupils.

32:3 See the Râmânuga Bhâshya; and the Râmânuga Darsana in Sarvadarsana-sangraha.

32:4 Âpastamba, p. xvi.

32:5 See Siddhânta Kaumudî, vol. i, p. 592.

33:1 See his Pânini; and see also Bühler's Âpastamba in this series, Introd. p. xxxii note.

33:2 The correctness of this tradition is very doubtful.

33:3 Indische Studien I. 470.

34:1 See the latter's edition of the Gîtâ, Preface, p. xxvii.

34:2 In the edition of the Gîtâ published in Bombay in Saka 1782, there is a stanza which says that the Upanishads are the cows, Krishna the milkman, Arguna the calf, and the milk is the nectar-like Gîtâ, which indicates the traditional view of the Gîtâ--a view in consonance with that which we have been led to by the facts and arguments contained in this Introduction.

35:1 p. xl (Lassen's ed.)

35:2 Sankara's commentary states in so many words that the Gîtâ he used contained only 700 slokas.

Next: Chapter I