Open Access

Who’s Zoomin’ Who? Bhagavadgītā Recensions in India and Germany

International Journal of Dharma Studies20164:4

DOI: 10.1186/s40613-016-0026-8

Received: 4 October 2015

Accepted: 20 January 2016

Published: 13 April 2016


This article discusses the political and theological ends to which the thesis of different “recensions” of the Bhagavadgītā were put in light of recent work on the search for an “original” Gītā (Adluri, Vishwa and Joydeep Bagchee, 2014, The Nay Science: A History of German Indology; Adluri, Vishwa and Joydeep Bagchee, 2016a, Paradigm Lost: The Application of the Historical-Critical Method to the Bhagavad Gītā). F. Otto Schrader in 1930 argued that the “Kashmir recension” of the Bhagavadgītā represented an older and more authentic tradition of the Gītā than the vulgate text (1930, 8, 10). In reviews of Schrader’s work, Franklin Edgerton (Journal of the American Oriental Society 52: 68–75, 1932) and S. K. Belvalkar (New Indian Antiquary 2: 211–51, 1939a) both thought that the balance of probabilities was rather on the side of the vulgate. In a trenchant critique, Edgerton took up Schrader’s main arguments for the originality of the variant readings or extra verses of the Kashmir version (2.5, 11; 6.7; 1.7; 3.2; 5.21; 18.8; 6.16; 7.18; 11.40, 44; 13.4; 17.23; 18.50, 78) and dismissed them out of hand (1932, 75). Edgerton’s assessment was reinforced by Belvalkar, who included a survey of various other “versions” of the Gītā in existence, either by hearsay or imitation. Belvalkar was especially hostile to the possibility of a Kashmir recension, because, as he noted, “once Schrader’s thesis is accepted as proved, it raises the possibility of other recensions of the Poem being current at different times in different parts of India” (1939a, 212). Belvalkar was consequent in following this assessment in the Critical Edition of the Bhagavadgītā, and the text reprinted there is essentially that of Śaṅkara’s eighth-century commentary. Yet, the publication of the Critical Edition has not sufficed to end the controversy, as witnessed by the latest spate of works (Vedavyas, E, 1990, Ancient Bhagavad Gita: Original Text of 745 verses, with Critical Introduction) (Bhattacharjya, Sunil Kumar, 2013, The Original Bhagavadgītā. Complete with 745 Verses) that claim to have discovered the “original” Gītā. In light of these attempts, we raise the question: why is there such interest in identifying versions of the text at variance from the normative one? And why do Indologists, in the name of “critical” scholarship, continue to welcome the creation of apocryphal versions of Indian texts? We suggest that there is a historical link between German Indology’s emphasis on creating new texts and German Protestantism.

For of playing the master and the critic there is nowadays neither measure nor end.

—Martin Luther, Luthers Bibelübersetzung (1530)


In this paper, we reconsider the debate over “recensions” of the Bhagavadgītā in light of recent work on the search for an “original” Gītā (Adluri and Bagchee 2014; Adluri and Bagchee 2016a). F. Otto Schrader in 1930 argued that the “Kashmir recension” of the Bhagavadgītā—reconstructed by him on the basis of an incomplete manuscript in Śāradā characters1 and a single Gītā commentary, the Sarvatobhadra of Rājanaka Rāmakavi (a. k. a. Rāmakaṇṭha)2 —represented an older and more authentic tradition of the Gītā than the vulgate text (Schrader 1930, 8 and 10). In spite of criticisms of Schrader’s thesis by both Edgerton and Belvalkar and in spite of the completion of the Critical Edition of the Bhagavadgītā in 1947, confusion over the normative form of the Bhagavadgītā continues, as witnessed by the latest spate of works (Vedavyas 1990; Bhattacharjya 2013) that claim to have discovered the “original” Gītā.3 Scholars’ confidence in being able to identify a more “original” Gītā, whether an expanded or an abridged version of the one currently normative in India, seems not to have died. As all such efforts go back to the view that there were different recensions of the Gītā in existence, it seems worthwhile to revisit the question: is there a Kashmir recension of the Bhagavadgītā? And what might the existence of such a recension reveal about the textual history of the Bhagavadgītā?

Rumors of a Hidden Gītā

Reports of the discovery of a longer recension of the Bhagavadgītā, corresponding exactly to the length of 745 verses mentioned in the Gītāmāna (112*4), in 1917 stimulated great excitement over the possibility that there might be versions of the Bhagavadgītā that differed from the normative version of the text (that is, that found in Śaṅkara’s commentary, which, at least since the eighth century, has been the recognized version of the text). This suspicion (that is, that there might be versions of the Gītā other than the one most widely read and recognized in India) directly triggered F. Otto Schrader’s efforts at identifying a different version of the Gītā.5 In his “Neues über die Bhagavadgītā” of 1927 (Schrader 1927, 171–83), Schrader stated that the information of the Gītāmāna “simply cannot mean anything else than that the composer of these verses attributed to Vaiśampāyana had before him a different version of the Bhagavadgītā than that currently found in the Mahābhārata [and] that this other Gītā was older” (172; Schrader’s emphasis). He further claimed that “it now appears that that this Bhagavadgītā of 745 verses [that is, as recorded in the Gītāmāna] (or a reconstruction thereof?) is still available. At least, such a Gītā text is at the basis of the colossal Gītā-bhāṣya of the Haṃsayogin currently being published, which is being edited by Pandit K. T. Sreenivasachariar in Madras and of which so far two volumes have been published: the Upodghāta (310 pages) and the commentary on the first adhyāya (170 pages)” (172). A year later, Schrader also commented: “The reigning view since Schlegel that the Bhagavadgītā has been transmitted perfectly uniformly (“nulla codicum discrepantia”) is a misconception. Thus far, it is possible to demonstrate the existence of two versions of the Bhagavadgītā that differ significantly from the familiar one [that is, the text of the Śāṅkarabhāṣya]: the first, namely, the one commented on by the Haṃsayogin, due to its unique sequence of verses and the inclusion of other verses of the Mahābhārata; the other due to a large number of unique readings, besides several verses not attested anywhere else” (Schrader 1928, 97). Schrader added: “Referring to his article in the Garbe Festschrift [for the first], the speaker [that is, Schrader] presented a provisory description of the second on the basis of the Gītārthasaṃgraha of Abhinavagupta and via reference to a second Kashmiri Gītā commentary, that of Rāmakaṇṭha (MS in the India Office), which he hopes to examine in London” (97). Thereafter, Schrader carried out significant work on this text in numerous articles published between 1930 and 1935, claiming that the existence of this alternative recension revealed much about the textual history of the Bhagavadgītā and its dissemination and reception in northern India (Schrader 1930; 1931, 748–53; 1933, 40–51; 1934, 348–57; 1935a, 146–49).6 But before we look at these claims, let us first look at the Bhagavadgītā edition of the Haṁsayogin, as published by the Śuddha Dharma Maṇḍala.

A Gītā edited from secret manuscripts?

The edition that Schrader in his 1927 and 1928 articles placed so much trust in did not stand under a lucky star. In 1917, the Śuddha Dharma Maṇḍala of Mylapore, Madras, an organization founded in 19157 and associated with the Theosophical Society, published an edition of the Bhagavadgītā that was to be the first of several publications claiming to be based on old, obscure, and secret manuscripts. Titled Bhagavad-Geeta of Bhagavan Sri Kṛishna and the Geeṭarṭha-Sangraha of Maharshi Gôbila, the edition was edited by “Pandit K. T. Sreenivasachariar of Madras” and featured an English foreword by “Dr. Sir S. Subrahmanya Iyer, K. C. I. E., L. L. D.” Sir Subramania was one of the leading intellects of the Anglicized Indian community of Madras and one of only three Indian lawyers to make it to the High Court Bench. Born in Madurai district in 1842, he rose, by 1899, to be Acting Chief Justice of the Madras High Court and Vice-Chancellor of the Madras University in 1904. But, as Derrett notes, “during his tenure as a Judge it was known that his real interests lay in spiritual matters. Immediately after retirement, he joined the Theosophical Society as Vice-President. He was by that time an L. L. D., and had been gazetted K. C. I. E. Sir Subramania Iyer was one of the first figures in Madras and it was an honour for the Theosophical Society to have him so highly placed between 1907 and 1911. Dr. Annie Besant thought so highly of him that she continued to publish matter of his writing even when it ceased fully to agree with her own notions” (Derrett 1977, 178).

In his foreword to the text, Sir Subramania noted that the edition of the Gītā published by the Śuddha Dharma Maṇḍala was “in a very real and substantial sense, a new one.” It was based on “trac[ing] and duly incorporat[ing]” the “seventy and odd verses, which ought to form part of the Scripture” by “transpos[ing] them to their proper place in the Geeṭā” from its “parent work, the Mahâbhâraṭa,” “thereby making it what it was in the Bhâraṭa of twenty-four thousand ślôkas, the predecessor of the epic as we have it now” (Iyer 1917, i). However, this was not the extent of the edition’s innovation: rather, the Gītā had been rearranged into “twenty-six chapters adopted in accordance” with the commentary of one “Hamsa-yôgi,” as dictated in the latter’s commentary known as the “Khaṇda-rahasya.” “This commentary,” Sir Subramania claimed, “hitherto little known outside the ancient Organization called the Śuḍḍha-ḍharṃa-maṇdala is a veritable store-house of invaluable interpretations of the secret teachings, contained in some of the most important Āryan sacred books such as parts of the Vêḍa, a number of the Upanishaḍs, the Mahâbhârata, the Râmâyaṇa and a few of the Purâṇas” (ii). Immediately thereafter, however, he noted that “the name Hamsa-yôgi is not the name of a specific individual author but that of an aḍhikâra-purusha, who is a member of the occult hierarchy engaged in the spiritual government of our globe, and who is charged with the duty of furnishing to the world esoteric explanations of scriptural teachings from time to time” (ii). In the conclusion, and perhaps not without a certain contradiction with his preceding statement about the creation of the edition, he noted of the sources consulted for the edition:

The manuscript which the Editor had already in his possession, was a copy made many years ago from the manuscript of Svâmi Yôgânanḍa who is evincing much interest in the publication of this and other works hitherto known only to the members of the Śuḍḍha-ḍharma-maṇdala to which he belongs. The Editor’s said copy has been carefully compared with two others lent to him by Svâmis Śankarânanḍa and Bhavânanḍa, both of whom are also members of the said organization. Furthermore, all the manuscripts have been checked with reference to a palm-leaf hand-book, nearly a century old, containing the first few words of all the verses of the Geeṭa in the order in which they stand in the present edition. This handbook belonged to the Editor’s maternal grandfather, Ṭiruvêlunḍoor Bhâshyam Ṭiruvênkatâ Châriar, an erudite and versatile paṇdit, a friend and relation of Paravasţu Rangâ Châriar, the greatest Samskṛṭ scholar of his time in this Presidency and the author of a remarkable Samskṛṭ lexicon. (xxxv–xxxvii)

In contrast, “the manuscript of Gôbhila’s Kârikâ used by the editor was kindly lent to him by one of the Svâmis occupying a very high position in the Organization. It will thus be seen that much care has been taken in making the present edition accurate and reliable, and it is hoped that it will have a wide circulation” (xxxvii).

Sir Subramania’s comments about the origins of the edition stand in contrast with the editor’s, who notes that “the existence of this Geeṭa of the Śuḍḍhas was first discovered to me by my much revered grand-father; who, as has long been well-known, was a staunch adherent of Śuḍḍha Dharma. My grand-father was Sremaḍubhaya Bāshyam Thiruvenkatachariar of Tiruvinḍalur. For his own use, he had once prepared an index of the first quarters of the verses of the Bhagavaḍ-Geeṭa, under the guidance of which, and from the manuscripts of the Swāmi Yōgānanḍa, Śankarānanḍa and Bhavānanḍa, has this Geeṭa been edited” (Srinivasachariar 1917, 21), and in complete contradiction to the comments of R. Vasudeva Row, the translator of the edition, according to whom publisher and editor, Sir Subramania and Pandit Sreenivasachariar, had been “enabled to do so [that is, publish the edition] by some of the Elders of the Suddha Dharma Mandalam, who graciously secured for the purpose of the Manuscript of this Text, along with those of other philosophic works, in Their safe keeping, from the archives of Suddha Kosha in the Maha Guha [great cave] in Northern India, known to a few among us” (Row 1939, vii). These discrepancies, both between the different accounts of the origins of the text and also within individual accounts, suggest a less scrupulous ancestry for the edition, and, in fact, Derrett has speculated that the “‘Haṃsa Yogi’ was the Madras pandit himself” (Derrett 1977, 179). There are also problems with the way the text of the Bhagavadgītā has been compiled, which clearly presupposes later ideas and modes of scholarship, especially historicist and critical scholarship. But before we evaluate the bona fides of the edition, let us first take a look at the arrangement of the materials itself and its contents.

The theosophic Gītā of the Haṁsayogin

Although Sir Subramania and K. T. Sreenivasachariar both claimed that the text of the Bhagavadgītā in the edition was based on ancient manuscript(s), it is clear that the actual arrangement of the text deferred not to a manuscript containing the text but to a manuscript of the Haṁsayogin’s commentary. For, Sir Subramania explicitly declared that the “novelty and value [of the edition] rests chiefly on the division of the Scripture into twenty-six chapters adopted in accordance with Hamsa-yôgi’s commentary known as the Khaṇda-rahaysa” (Iyer 1917, ii). Even if this commentary turned out to be based on an actual manuscript of a traditional commentator—something we have every reason to doubt—it is clear that the arrangement of the edition itself was novel in the sense of being reconstructed to reflect and to bring out the principles of Śuddha Dharma philosophy. In fact, publisher and editor both acknowledged as much: Sir Subramania noted that “those, who have so long been accustomed to the division of the book into eighteen chapters, will naturally demand an explanation of the present division into twenty-four chapters, now for the first time made publicly known” (Iyer 1917, ii); and Sreenivasachariar noted: “concerning the twenty-four Geeṭas it contains, Gōbhila, speaking on the Āḍiparva of the Mahābhāraṭa, says ‘Pravriṭṭi Dharma or the Dharma of Forthgoing involves the Taṭvas; likewise, Taṭvas mark Nivriṭṭi Dharma or the Dharma of returning also. The ten senses, the five material elements, the ṭanmāṭras or the properties of the atoms of these elements, Mula-prakṛṭi or the element of the monad or Āṭman, Mahaṭ, the element of intuition, Manasa or the mind and Ahankāra or the I-concept—these are known as the Taṭvas. These twenty-four Taṭvas are also the theme of the Bhagavaḍ-Geeṭa” (Srinivasachariar 1917, 2). In fact, the entire undertaking was an attempt to create a new, authoritative text—new, because it had to address a contemporary audience and because Sir Subramania was interested in pursuing a program of national spiritual renewal—even though to gain that authority it had to embed itself in the tradition, indeed, claim that it was not new at all but represented an ancient and timeless tradition. Hence, Sreenivasachariar could claim:

Such is the Geeṭa which the Śuḍḍhas hold up and revere and it consists of twenty-six chapters; it is made up of as many geeṭas (twenty-four) as there are letters in the Gāyaṭri, each letter of which signifies one Taṭva or element. Reckoning up the verses, the Geeṭa contains seven hundred and forty-five stanzas. We say seven hundred and forty-five stanzas on the authority of the Mahābhārata itself which in the last chapter of the Geeṭa-parvan incorporated in the Bheeshma-parva, says, ‘The Lord, Kṛishṇa, spoke six hundred and twenty stanzas; Arjuna spoke fifty-seven; Sanjaya, sixty-seven; and Dhṛṭarāshtra, one. These together make the Geeṭa’. (2)

Figure 1 (on the following page) presents the actual arrangement of the chapters of the Śuddha Dharma Gītā.
Fig. 1

List of Chapters and their Titles in the Śuddha Dharma Maṇḍala Gītā (Row 1939, x)

Without entering into extensive detail about these chapters, let us here convey a sense of just how closely their arrangement is shaped by theosophic ideals. The best way to do so is again to quote from Rajagopala Iyengar’s translation of the pandit’s preface to the text. He writes:

The Geeṭa of the Śuḍḍas consists of the following chapters: On the coming or the genesis of the Geeṭa. On Nara and Nārāyaṇa; on avatars or incarnations; on heirarchs [sic]; on the instruction of hierarchs in the due arts; on the knowledge of causes; on beatitude; on the nature and characteristics of those engaged in the study and practice of the teachings of the group of the foregoing six chapters; on the securing by these men, of the three instruments or means of the method of execution; on Māyā, to be taught those that strive for the three instruments; on Mōksha or liberation, which gives Prāpṭi or the attainment of the higher for those who have learnt this Māya; on the aspect and nature of Brahman adored by those aspirants after liberation; on Nara and Nārāyaṇa who, with their sinless volition, have the power of assuming manifold forms; on the method of Prāṇāyāma, which is the emblem of the concentrated activity in those that study and practise the teachings of this second group of six chapters; on the Supreme Self, the object to be realised by those aspirants; on Akshara or the spirit, veiling itself in the Cosmos; on the study of this aspect of the spirit; on the office and function of the Paramahamsas; on Sanyāsa or renunciation; on the self, the causer of yōga, for the aspirants who study and practise the teaching of this third group of six chapters; on Prakriṭi or the Not-self, on related details and particulars; on knowledge and the method of knowledge; on the method of yōga which the study and practice of the teaching of these last chapters lead to; and finally, the hymn on the realisation of Brahman. (Srinivasachariar 1917, 15–16)

Pandit Sreenivasachariar’s Gītā was, however, not only theosophical in respect to its contents but also in respect to several ideas concerning history, the relation of texts to the community, and the appearance of a savior. But as these ideas are best examined when we discuss the peculiar historical and social conditions under which this and several other expanded Gītās were produced in the twentieth century, we shall discuss them there.

The Kashmir Gītā of F. Otto Schrader

The Haṁsayogin’s Gītā is not read today and is widely considered to be a fake. But its historical significance ought not to be underestimated, chiefly in the way it inspired other scholars to also search for different recensions of the Bhagavadgītā. Here, we wish to focus on one such example: the Kashmir Gītā of F. Otto Schrader. As we have seen, in his search for an expanded version of the Bhagavadgītā, Schrader set out from this very edition as evidence for the fact that the Bhagavadgītā as found in the vulgate had not always been the normative version of the Bhagavadgītā. Thus he cited Edgerton’s assertion that “there is absolutely no documentary evidence that any other form of the Gītā than that which we have was ever known in India. This, of course, does not prove that none ever was known; but it leaves a strong burden of proof upon those who maintain such a theory.”8 Against Edgerton, he argued: “one does not need to search far for the proof demanded by Edgerton. It is found, namely, in the Mahābhārata itself in the chapter immediately following the Bhagavadgītā.” Thereafter, citing the Gītāmāna śloka, he stated: “What does this mean? It simply cannot mean anything else than that the composer of these verses attributed to Vaiśampāyana had before him a different version of the Bhagavadgītā than that currently found in the Mahābhārata. Further, it also suggests the hypothesis that this other Gītā was older and that therefore not only the round verse number 700 but also the chapter number 18 according with the number of books of the Mahābhārata of the Gītā found in the current Mahābhārata is the result of a revision” (Schrader 1927, 172; Schrader’s emphasis). In pursuit of this thesis, Schrader first turned to a reconstruction of the Haṁsayogin’s Gītā on the basis of the introduction and the commentary on the first adhyāya, which is surprising since, as we have seen, both these contain sufficient clues to suggest that the edition was a bogus work9; evidently Schrader was not sufficiently skeptical of the Haṁsayogin’s efforts or did not take the publisher’s and editor’s introductions with the seriousness they merited.10 Thereafter, in later works, he turned to a reconstruction of the version of the Bhagavadgītā he considered to have been current in Kashmir. As this second of our bogus Gītās is of more interest to us here, we shall resume the story with Schrader’s arguments for a “Kashmir recension” of the Bhagavadgītā.

In 1928, Schrader followed up his introduction of the Gītā of the Haṁsayogin with a presentation at the Deutscher Orientalistentag in Bonn where he gave a presentation on the transmission of the Bhagavadgītā. In this presentation, whose transcript is unfortunately not preserved to us but of which we have a small notice in the report of that meeting, he presented, as he put it, “a provisory description of the second on the basis of the Gītārthasaṃgraha of Abhinavagupta and via reference to a second Kashmiri Gītā commentary, that of Rāmakaṇṭha (MS in the India Office), which he hopes to examine in London” (Schrader 1928, xcvii). He further noted that “the use of this manuscript in Germany for the purpose of an extensive publication about the Kashmir recension of the Bhagavadgītā has, in the meantime, been generously granted him; it contains the text almost completely.” Further, “in the British Museum he [the speaker, that is, Schrader] found an incomplete birch bark manuscript, not as yet listed in the published catalogue, containing only the text (in Śāradā script), which appears to agree exactly with that of the commentary cited” (xcvii). These were to be the basis of his next work, a reconstruction of the “Kashmir recension” of the Bhagavadgītā, published in 1930.

Schrader’s reconstruction was based on the evidence of three sources: two commentaries and a single fragmentary manuscript containing only the text of the Gītā (up to chapter 8, verse 18). The commentaries in question were the Sarvatobhadra of Ramakaṇṭha, used from manuscript no. 3271 in the library of the India Office, London, while the Bhagavadgītārthasaṁgraha of Abhinavagupta was used from the edition of the Nirnaya Sagar Press of 1912. According to Schrader, “all three sources are to some extent vitiated by the influence of the vulgate.” “This is most patent in the MS of the Sarvatobhadra where often the text preceding the commentary is in conflict with the latter [that is, the commentary],” but it also seems to have affected the other two sources, for “in the edition of Abhinavagupta’s commentary the vulgate has here and there crept not only into the pratīkas but even into the commentary itself, though, fortunately, without creating any doubt as to A.’s reading. And in the birch-bark MS the influence of the vulgate is quite obvious, e.g., in its containing two verses of the vulgate (II, 66–67) missing in both commentaries, and in its giving twice stanza V, 19 (ihaiva tair jitaḥ sargo, etc.) of the vulgate, viz., in that place and, moreover, after VI, 9, i.e., the place where the commentaries have it” (Schrader 1930, 2). In constituting his “Kashmir recension,” Schrader provided a list of the Kashmiri variants where the text of his sources differed from that of the vulgate. In places where his Kashmiri sources were not unanimous, he printed both readings next to each other, without making any attempt to reconcile them or, indeed, reduce them to a common one. As a sample, we provide the first chapter11 from his text of the Bhagavadgītā (23–24):
  • 1.1a dharmakṣetre kurukṣetre sarvakṣattrasamāgame (samavetā yuyutsavaḥ)12

  • 1.7c nāyakān mama sainyasya saṁjñārthaṁ tān bravīmi te (nāyakā)

  • 1.8a bhavān bhīṣmaś ca karṇaś ca kṛpaś śalyo jayadrathaḥ (samitiṁjayaḥ)

  • 1.8c aśvatthāmā vikarṇaś ca saumadattiś ca vīryavān (saumadattis tathaiva ca)

  • 1.9c nānāśastrapraharaṇāḥ nānāyuddhaviśāradāḥ (sarve yuddhaviśāradāḥ)

  • 1.11a ayaneṣu tu sarveṣu yathābhāgam avasthitāḥ (ca)

  • 1.18a pāñcālas ca maheśvāso draupadeyāś ca pañcake (drupado draupadeyāś ca sarvaśaḥ pṛthivīpate)

  • 1.21c ubhayos senayor madhye rathaṁ sthāpaya me ’cyuta (senayor ubhayor madhye)

  • 1.24c ubhayos senayor madhye sthāpayitvā rathottamam (senayor ubhayor madhye)

  • 1.28a kṛpayā parayāviṣṭo sīdamano ’bravīd idam (viṣīdann idam abravīt)

  • 1.29a sīdanti sarvagātrāṇi mukhaṁ ca pariśuṣyati (mama gātrāṇi)

  • 1.30a sraṁsate gāṇḍīvaṁ hastāt tvak caiva paridahyate (gāṇḍīvaṁ sraṁsate hastāt)

  • 1.32a na kāṅkṣe vijayaṁ kṛṣṇa na rājyaṁ na sukhāni ca (na ca rājyaṁ sukhāni ca)

  • 1.33c ta eva me sthitā yoddhuṁ prāṇāṁs tyaktvā sudustyajān (ta ime ’vasthitā yuddhe prāṇāṁs tyaktvā dhanāni ca)

  • 1.35c api trailokyarājyasya hetoḥ kiṁ u mahīkṛte (nu)

  • 1.37a tasmān nārhā vayaṁ hantuṁ dhārtarāṣṭrān svabāndhavān (sabāndhavān)

  • 1.39c kulakṣayakṛtaṁ doṣaṁ saṁpaśyadbhir janārdana (prapaśyadbhir)

  • 1.47c utsṛjya saśaraṁ cāpaṁ śokasaṁvignamānasaḥ (visṛjya)

As can be seen, the differences are not very significant13 and Belvalkar, in fact, thought that they were the kind that could easily be explained through the normal processes of scribal error, gloss, and simplification of the text. According to him, the majority of the 282 Kashmiri variae lectiones could be classified into one of four types:
  1. 1.

    “The Kashmirian variants (with a sporadic exception or two) seek to remedy and regularize the grammatical defects of the current (or Śāṁkara) text” (for example, 2.50ab, 2.60a, 3.23a, 5.1 cd, 6.39a, 7.18b, 9.31c, 10.9d, 11.41b, 14.23d, 16.2c, 16.13b) (Belvalkar 1947a, lxxviii).

  2. 2.

    “The Kashmirian variants seek to simplify and normalize the syntax” (for example, 1.7c, 5.21b, 6.13c, 6.19d, 6.21a, 8.17b, 9.11d, 10.16b, 10.19b, 10.28c, 11.44d, 18.8a, 18.50ab) (lxxix).

  3. 3.

    “The Kashmirian variants generally tend to smoothen (not always successfully) the difficulties in interpretation that have proved troublesome” (for example, 1.33b, 2.43b, 2.61b, 6.7b, 6.28b, 13.4d, 13.16a, 17.13a, 17.23c) (lxxx).

  4. 4.

    Finally, of four variants to which Schrader attached particular significance (2.5c, 2.11ab, and 2.21d), Belvalkar noted that all four of them were lectiones faciliores, and that Schrader had misinterpreted the manuscript evidence (lxxxii).14


Belvalkar also argued with compelling evidence from the other Mahābhārata manuscripts collated for the Critical Edition that the variants in Schrader’s three sources did not justify the assumption of an independent Kashmiri tradition, significantly different from the text of Śaṅkara. According to him, the variation was within the limits of that found in the other manuscripts; further, Schrader’s sources were not even unanimous among themselves. “It is to be noted that in the 282 places where Schrader reports Kashmirian varietas lectionis it is not always the case that his three Kashmirian authorities agree; and in such cases the discrepancies are explained as due to the influence of the Vulgate reading” (Belvalkar 1939a, 213). Further, Schrader’s three sources did not find support even from the Kashmiri manuscripts collated for the Critical Edition of the Mahābhārata, as one might expect if there really had been an independent tradition of the Bhagavadgītā in Kashmir. “One expects that our Kashmirian Mss. would confirm Schrader’s findings, and in a few cases they no doubt do so. But is it not rather surprising that in as many as 122 places (Vide Appendix 1) the Kashmirian and allied-Kashmirian Mss. used for the Critical Edition should not support the variant readings listed by Schrader? None of the other Mss. also, even in a single one of these 122 cases, registers Schrader’s readings, while in 12 other cases, shown in Supplements to Appendix 1, the ‘Kashmirian’ readings find only sporadic support from solitary Mss.” (213).

The amount of variation, furthermore, was not remarkable: Belvalkar noted that “our Śāradā MS records over 130 cases of such individual variations unknown to Schrader’s sources or in fact practically to any other MS.” He concluded that “we would not be justified in attaching any exaggerated importance to these cases of solitary variants, individually or cumulatively, and raise them to an independent ‘Recension’” (213). Indeed, Belvalkar found that “even in the matter of the additional stanzas and half-stanzas (as also of the omissions),” which had been Schrader’s strongest reasons for insisting on the existence of an independent tradition, “the Kashmirian Recension is not by any means peculiar” (215). He showed that other manuscripts also frequently added stanzas or half-stanzas as well as omitting some, but there was no reason to doubt the authenticity of the vulgate recension. Further, in his assessment, Schrader had rather overstated the evidence for an independent “Kashmir recension.” “Schrader designates the consensus of his three sources by the symbol ‘K,’ but that is rather misleading because his MS Lb breaks off after viii.18, and because Abhinavagupta passes over many words of the text in silence, so that not infrequently ‘K’ denotes only one authority” (215). The readings of Schrader’s sources did not have widespread support from other sources from Kashmir, making it more likely that the variants were cases of scribal error and corruption in one, isolated source that had been handed down and become the basis for the two commentaries.15 Indeed, some of the readings were not unique to Schrader’s sources at all, but had the support of other manuscripts,16 so that the editor chose those readings over those of the vulgate for the Critical Edition; others, by contrast, were clearly due to errors that could have occurred independently.17 Summing up, Belvalkar noted: “a large number of his [Schrader’s] cases are merely solitary variations of individual Mss., while quite a few of the others are not peculiar to Kashmir, and have no probative value in establishing a ‘Kashmirian recension’. About a little over thirty per cent. of the cases adduced can be regarded as Kashmirian Pāṭhabhedas of the Gītā, but intrinsically they can almost all be proved to be secondary and posterior to the text of the poem as known to Śaṅkarācārya” (231). He therefore rejected Schrader’s claims regarding the “authenticity” and “priority” of the Kashmir recension (231).18

The difficulty with the Lectio Difficilior

Belvalkar’s careful sifting of the evidence was confirmed by Franklin Edgerton, who wrote in a review of Schrader’s book that, after a careful study of Schrader’s variants, he had been “forced to the conclusion that not one of them is conclusive” and “in a number of cases, on the contrary, the probabilities favor the vulgate reading, while in the remaining there seems hardly any reason to prefer” (Edgerton 1932, 70). Edgerton also disputed Schrader’s assertion that the stanzas of the vulgate omitted in his sources must be “secondary additions” because “they ‘cannot possibly have been omitted in K for want of interest or some such reason.’” He wrote that Schrader “attributes more care and system to ancient Hindu copyists and redactors than I should wish to, when he implies that the stanzas could not have been omitted in K by mere accident, or (which perhaps means the same thing) for some reason that escapes our ken” (74–75). In contrast, he claimed that the extra verses in the Kashmiri sources had to be seen as additions in them, rather than as conscious omissions in the vulgate: “Is there any reason to suppose that they [the three-line stanzas] would have appeared to an ancient Indian redactor as a blemish? I think that Schrader projects modern western ideas into the ancient Indian past. It is much more likely that the extra lines of K are secondary additions” (75). Edgerton also examined in detail a number of variant readings (in verses 2.5, 11; 6.7; 1.7; 3.2; 5.21; 18.8; 6.16; 7.18; 11.40, 44; 13.4; 17.23; 18.50, 78) that Schrader had asserted were more original than the vulgate readings; in all cases, he rejected Schrader’s readings as being more likely the lectiones faciliores. His conclusion was succinct and decisively stated:

In brief, I see no reason whatever for assuming the originality of any of the plus parts of K, nor of even a single one of K’s variant readings. I am obliged to conclude that the attempt to prove the superiority of K is a failure, and that, on grounds of general probability, we must continue to regard the vulgate text of the Gītā as the nearest approach we have to the original, especially since it seems to be supported by the genuine ‘Kashmirian’ version of the Mahābhārata as a whole. (75)

A complete evaluation of Schrader’s claims is only possible by re-examining all his sources and comparing these with the editor’s decisions in the Critical Edition.

The problem, however, with arguing from the lectio difficilior is that it is often subject to challenge. In many cases, opinion will differ upon which reading will have appeared more obvious to a scribe. Except in cases where one of the forms is truly rare or one is orthographically easily mistaken for another, more complex form, determining the lectio difficilior often comes down to a matter of weighing the probabilities. A headstrong editor, then, will quite often be able to controvert his critics’ objections, as Schrader, in fact, did vis-à-vis Edgerton. (By contrast, he never responded to Belvalkar’s evaluation of the evidence, either after 1939 or 1947, perhaps because Belvalkar did not argue solely from the lectio difficilior, but also presented extensive additional evidence showing that Schrader’s alleged “Kashmir” readings were neither as unique nor as unusual as he thought.) Thus, Schrader noted that “fourteen whole verses and four half-verses of the Kashmir Recension are missing in the vulgate. Edgerton considers them all to be interpolated, apparently for no other reason than that the vulgate does not contain them. However, factual reasons appear to me to speak against not a single one of these verses; but a precise metrical examination of the Bhagavadgītā (also worthwhile for itself) could perhaps also identify interpolations here, perhaps also the Schallanalyse [sound analysis?]. [But] it must for now remain non liquet [that is, without an applicable law or a decision deriving therefrom]” (Schrader 1933, 43). Further, he rejected Edgerton’s arguments for preferring the vulgate reading aśocyān anvaśocas tvaṁ prajñāvādāṁś ca bhāṣase over the Kashmiri aśocyān anuśocaṁs tvaṁ prājñavan nābhibhāṣase (2.11), with the words: “I [that is, Schrader] see here in V a very old corruption (Speijer expressed this view already in 1902 in his article “Ein alter Fehler in der Überlieferung der Bhagavadgītā” in ZDMG vol. 56) and in K the original reading. The meaning according to Edgerton is: ‘. . . and you (presume to) utter speeches concerning wisdom! (altho you are so foolish as to mourn those who should not be mourned)’. This is not very different from the four of the native explanations I have listed, namely that of Śrīdhara’s (prajñāvatāṃ paṇḍitānāṃ vādān śabdān kathaṃ Bhīṣmam ahaṃ saṃkhye ityādīn kevalaṃ bhāṣase na tu paṇḍito ’si), but, for me, even less acceptable than it. For the ‘presume to’ that one must think in addition is a far too dubious makeshift” (43–44). He added: “I cannot concede the necessity of holding on to the reading of V, because it is supposedly the lectio difficilior. A lectio difficilior is often based on a mere corruption of the text: anuśocaṃs (K), through leaving out the anusvarā, became anuśocas and further anvaśocas; and, as a consequence, the impulse for changing it arose, since one now needed a ca or tu in the second pāda” (45).

Schrader also contested Edgerton’s reasons for preferring the vulgate reading hatvārthakāmāṁs tu gurūn ihaiva over the Kashmiri na tv arthakāmas tu gurun nihatya,19 on the grounds that “Edgerton translates: ‘but having slain my gurus as they seek to attain their ends, I should eat food smeared with blood right in this world (without waiting for such a punishment in a future life)’ and adds, by way of clarification, that Arjuna means ‘a guru’s wish should be law to him.’ But as the latter is self-evident to every Hindu, one must ask oneself how it is that this simplest of explanations of the verse did not occur to any of the commentators. This can, in my view, only be the case if arthakāma was not familiar to them in the sense stipulated by Edgerton (‘desiring their ends’), but only as a fixed composite with the two meanings that are alone attested of ‘avaricious’ and ‘obliging.’ I therefore hold on to the Kashmiri reading arthakāmas” (Schrader 1933, 45–46). Regarding verse 6.7, Edgerton had argued that the terms ātman and paramātam were often treated as synonymous and so there was no real reason for objecting to the vulgate’s jitātmanaḥ praśāntasya paramātmā samāhitaḥ.20 Schrader responded huffily: “regarding the Bhagavadgītā’s reluctance to treat God and the soul as two separate metaphysical principles, I have expressed myself often enough (e.g., in my introduction to Bhagavadgītā XIII in Religionsgeschichtliches Lesebuch, issue 14, p. 20); Edgerton’s lecture on this subject was unnecessary” (Schrader 1933, 46). He further claimed that “the reason for it [that is, paramātman] had now become clear to me: not because it was ‘precisely equivalent to ātman’ [as Edgerton had claimed], but rather, in order to express the contrast to the ātman in jitātmanaḥ (which was concealed in the words yenātmaivātmanā jitaḥ of the preceding verse) the author this time says paramātma. He namely had the idea of two ‘selves’: the lower, which is falsely called so and the true self, here paramātman, simply called ātman in the previous verse (just like the lower)” (46; emphasis in original). Nonetheless, he rejected Edgerton’s conclusion: “even so the verse remains suspicious in the vulgate form. For the word paramātman not only does not appear in the older Upaniṣads but also in the so-called middle Upaniṣads that immediately precede the Gītā and have been cited word for word by it (Īśa, Kaṭha, etc.) and in the Gītā it first appears, ignoring our verse [that is, 6.7], in the additions that begin (according to Oldenberg’s and Garbe’s and also my opinion) with adhyāya XIII” (46). Finally, Schrader also contested Edgerton’s interpretations of 3.2,21 6.16,22 7.18,23 9.40,24 11.44,25 13.4,26 17.23,27 and 18.78,28 granting only four as open to question (that is, 1.7,29 5.21 and 18.8,30 and 18.5031).

Stemmatic arguments for K?

The lectio difficilior is an established principle in textual criticism and, when used correctly, it can in fact lead to the restoration of better, that is, unusual and presumably more archaic readings, only its use can so often be controverted in practice by a determined opponent that it is not always as useful as it might be. Especially in the case of German scholars, who saw no role for themselves in Sanskrit Studies if they were not to controvert the opinions of their peers, the lectio difficilior often turned out to be a source of great problems. Since German Indologists were precisely interested in putting their little knowledge of Sanskrit on display, drawing them into a discussion of the merits of different readings—or, being drawn by them into a discussion of the merits of different readings—played inescapably into their hands.32 They could now put on display their philology, which, as Heyne remarked, consisted of nothing more than “the vanity of wanting to seem brilliant through emendations.”33 (Heyne, in contrast to the German Indologists, was extremely circumspect about replacing the accepted reading34; the Indologists might have taken a page out of his book and evolved a more mature, less shrill philology.) For this reason, we focus here on stemmatic arguments for evaluating proposed reconstructions. This emphasis arises not from a rejection of the lectio difficilior, which has been used to great effect in the Critical Edition of the Mahābhārata—in fact, there is very little disagreement about it when scholars have a shared commitment to the truth (Belvalkar’s and Edgerton’s assessment of the vulgate is an example of the concord possible when using the lectio difficilior as the criterion), but from an awareness that the Indologists’ Achilles’ heel is textual criticism. We shall focus here on Schrader’s arguments for a “Kashmir recension,” as he is especially vulnerable on stemmatic grounds.

Even assuming that the Kashmir recension represents an independent, parallel text of the Bhagavadgītā and even assuming that it was the older recension (as compared with the vulgate), how are we to imagine its transmission? Schrader discusses the question in three places: his 1930 study, his 1933 article, and his 1938 article. In the 1930 text, Schrader argued that there must have been several recensions of the Bhagavadgītā in circulation to judge by the testimony of Al-Biruni. Further, as “not one of the additional stanzas [in the Kashmir recension]” appeared to be “unmistakably an interpolation” (Schrader 1930, 10; emphasis in original) (though he did not rule out the possibility, indeed, he considered it “quite likely,” “that there are additions in K (such as V, 17 in the birch-bark MS.)” [10]), he concluded that “at least some of the additional matter of K, as compared with V, may have been in the original Gītā before the number of its stanzas was cut down to seven hundred” (10). In contrast, he argued that the verses missing in the Kashmir recension (2.66 and 2.67) must have been additions in the vulgate: “they were not in the original Gītā” (11), and hence, while both recensions had undergone changes, the Kashmir recension had to be seen as preserving more of the original. Although Edgerton, in his review of Schrader’s book, pointed out a major flaw in this argument (the excellent and old manuscript no. 2137 of the India Office Library, London contains the vulgate version of the Gītā creating, in Edgerton’s words, “a certain presumption in favor of the vulgate text” [Edgerton 1932, 70]), Schrader held on to his theory, though now offering a more complicated account of the relation of all three recensions (that is, the “original,” the vulgate, and the Kashmiri). (Note, however, that Schrader’s objection to Edgerton’s argument, “this [that is, that Edgerton considered none of the readings of K to be more original than the corresponding ones of V] would mean that in Kashmir the recension V developed at some point into K and was then later once again replaced by V” (Schrader 1933, 42) is nonsense, entailing as it does a petitio principii.) Thus, he now argued that “if we therefore assume that the text of K has undergone changes—and this is surely most probable—and accordingly eliminate the obvious corrections and the verses that appear to be added from it [that is, from K], we arrive at a text as the foundation of K, that can be considered a parallel text to the vulgate, i.e., as a text that, together with this [the vulgate], goes back to an original that has not been preserved or that has not been found as yet” (42–43).

This interpretation was tantalizing as it provided support for the German scholars’ thesis of a much shorter “original” Gītā, later revised by “Brahmanic” redactors (and, in fact, Schrader did argue, immediately after presenting this idea, that “for the contemporary discipline [Wissenschaft] there can only be the one ‘working hypothesis’ that the text of the Bhagavadgītā was not always fixed, but rather was first [fixed] from the period onwards in which the vulgate, through the spread of Viṣṇuism all over India, which took place relatively late, compelled its universal acceptance” [43; Schrader’s emphasis]), but it crucially failed to take into account the evidence of the Gītāmāna śloka—something we know, from Schrader’s other writings, that he was explicitly interested in. Thus, in his final article on the topic, Schrader took a different tack: he now argued that even if the original Gītā had been a much shorter work, it must have developed into the present recensions (that is, the vulgate and the Kashmiri) via a 745-verse Gītā, for, “however late that passage [that is, the Gītāmāna] may be, the Bhagavadgītā to which it refers must be older and may be even considerably older, and those ślokas must have been composed by one who had actually before him a Bhagavadgītā of that description” (Schrader 1938, 62). On the basis of this hypothesis, he now essayed the view that “at a time when the Bhagavadgītā was not yet as sacrosanct as it has been since more than a millennium this same Bhagavadgītā as we now have it” must have been “re-written with some additions and omissions” (62–63). The additions required to bring the number of verses up from an indeterminate, but definitely lower figure35 to 745 presented no problem, for, as Schrader argued, “we can well imagine that the revisor added bona fide to, or interpolated in, the speeches of Śrī Kṛṣṇa some more stanzas of the Lord’s known to him from other sources. We can also understand that he saw no harm in adding, where this appeared to be desirable to him for the sake of clearness or otherwise, some ślokas (borrowed or of his own making) to those spoken by Saṁjaya” (63; Schrader’s emphasis). However, the omissions, required to bring the number of verses spoken by Arjuna down from eighty-four in the present Gītā (that is, the vulgate) to the fifty-seven noted in the Gītāmāna śloka presented a greater problem. According to Schrader, it was inconceivable that, once present in and recognized as a genuine part of the Bhagavadgītā, these verses could have been removed by any author. “We can, however, not understand his omitting any ślokas of the present Gītā. For, what could have induced him to do so?” (63; Schrader’s emphasis). Hence, he concluded: “No other conclusion can be drawn from this consideration but that those 27 ślokas exceeding in the current Gītā the number of those spoken by Arjuna in the longer one were not known to the compiler of the latter, and thus the supposition that he had before him our present Gītā falls to the ground” (63; Schrader’s emphasis). In order to bring the number of verses spoken by Arjuna in the Bhagavadgītā in line with the Gītāmāna figures, he therefore proposed excising verses 1–2, 15–30, and 36–44 of chapter eleven, which gave him a “Gītā” of 673 ślokas, from which he again subtracted two ślokas (because a variant in a Kashmiri manuscript refers to the number of verses spoken by Arjuna as fifty-five rather than fifty-seven) to arrive at the figure of 671 ślokas. Once he arrived at this figure for the “original” Gītā (though not the most original Gītā, the product of the German scholars’ fantasy), he reversed course and argued,

we [now] get a glimpse of the gradual growth of the Gītā from an unknown stage or number of stages to 671 stanzas (i.e., the vulgate minus 29 stanzas of Arjuna’s) and, from that point, in two diverging lines viz., (1) to a Gītā of 698 stanzas, i.e., 671 increased by 27 of Arjuna’s and then, by means of two more stanzas attributed to him, to the 700 stanzas of the vulgate; and, on the other hand, (2) to 741 stanzas, i.e., the vulgate minus 29 stanzas of Arjuna’s and plus 70 stanzas, not contained in the vulgate, of Śrī Kṛṣṇa’s and Saṁjaya’s, and from here, viz. by adding from the complete vulgate 2 more stanzas of Arjuna’s and besides 2 more stanzas of Saṁjaya’s to the 745 stanzas of Vaiśampāyana’s Gītā. (66)

On the basis of this two-stage analysis (first stage: using the evidence of the Gītāmāna śloka; second stage: using the evidence of the Gītāmāna śloka in a Kashmiri variant), Schrader concluded: “if this is what has happened, then not 27 but 29 ślokas spoken by Arjuna have to be accounted for as missing in the original Gītā, and the two latest ones would have to be either XI, 1–2, or VIII, 1–2, or two of the group X, 12–16, say 12 and 13” (66–67).36

Schrader as a textual critic and a critic of the Gītā

The problem with Schrader’s claims, however, is that they are not really stemmatic, as becomes apparent once we draw up explicit stemmata in place of his implicit ones (see Fig. 2 below).
Fig. 2

Schrader’s Hypothetical Stemmata

In the first case, if the scribes of K and V made conscious changes in each of their exemplars while copying from their common ancestor O, there is no stemmatic way to judge the originality of the readings of either K or V. The standard assumption in textual criticism is that scribes copy everything in their source and more, that is, that while they may introduce unconscious errors and insert glosses or even entire lines, they never consciously omit anything (Reynolds and Wilson 2013). If, however, the scribes of K and V were to have both added and omitted materials, we would be unable to account for the original reading in all the cases where K and V differ. Further, if K itself were to contain a number of interpolations vis-à-vis O, there would be no way of judging which of its fourteen full and four half-verses is original: any of these could be the interpolated verses.37 In effect, Schrader’s first argument, though apparently “text-critical,” cuts off the branch he is sitting on and forces a return to a subjective Konjekturalkritik.38

Likewise, his second argument assumes that we can, in the absence of stemmatic means, identify which of these fourteen full and four half-verses is original. Belvalkar and Edgerton had both pointed out of his original argument that, in the case of a crux between K and V, the balance of probabilities was wholly on the side of V—that is, that the additional verses in K were more likely additions in this recension rather than omissions in the vulgate. Schrader, however, thought he could get around this objection by conceding that some of the verses in K might be unoriginal and still maintain the primacy of the “Kashmir recension.” Thus he now argued that K could have developed from K' through the addition of a certain number (y) of verses, but K' would still be more original than V if V had developed from O also through the addition of a certain number (y) of verses. In this way, while he would have taken the force out of Belvalkar’s and Edger-ton’s objections (by casting doubt on the authenticity of the verses of V itself), he would still have retained K as the superior recension, albeit not in the form of K itself as it was present to him in his two and one-half sources but in the form of a hypothetical ancestor of K called K' to be reconstructed through eliminating the excess verses in K.

The problem with this argument, however, is that in the absence of a manuscript for comparison, there is no way Schrader can get from K to K': at most he could arbitrarily eliminate a number of verses, but the resulting text would not necessarily be more original than K—he would only be assuming that it was so. Further, if, as he explicitly claimed, V attained its present extent of 700 verses through the addition of y verses, then the original O must have contained less than 700 verses.39 If K' contains 714 full and four half-verses less y verses (that is, the total number of verses in K less those verses Schrader acknowledged might be interpolations in K [signified here by the symbol y]), then the fourteen full and four half-verses less y verses not found in V could be features of O only if some of the other verses in K' were additions in it. Consequently, Schrader would again have to go through K' and identify which verses were original in it and which not. The regress from K to K' does not solve anything; it only shifts the difficulty up one level.40 Schrader thus finds himself once again in the situation of using subjective criteria to determine the original form of the Bhagavadgītā—an undertaking as precarious as it is pointless.

Finally, the third of Schrader’s arguments for a “K” recension is the worst. From the disparity in accounts between the Kashmiri version of the Gītāmāna śloka, reported by S. N. Tadpatrikar (Tadpatrikar 1937, 357, n.1),41 and the vulgate version42 of that śloka, he concluded that the Bhagavadgītā must have been expanded in two stages: (1) a first stage, in which were added twenty-seven additional ślokas spoken by Arjuna, giving rise to the Kashmiri version of the Gītāmāna śloka (which attributes fifty-five ślokas to Arjuna); and (2) a second stage, in which were added an additional two ślokas spoken by Arjuna, giving rise to the vulgate version of the Gītāmāna śloka (which attributes fifty-seven ślokas to Arjuna). Schrader is obviously presuming that the two scribes or the scribes of the two versions had access to the same manuscript or text, except that one gained access to it at a later date, after it had undergone a further stage of expansion. But this is not true! There is no evidence that the scribe of V is reporting on the same text, albeit in expanded form, as the scribe of K. Both might be reporting on independent traditions, that is, that the relation of the two instead of being:
$$ \begin{array}{c}\hfill \mathrm{K}\mathrm{s}\hfill \\ {}\hfill \Big|\hfill \\ {}\hfill \mathrm{V}\mathrm{s}\hfill \end{array} $$
might be:
$$ \begin{array}{ccc}\hfill \Big|\hfill & \hfill \hfill & \hfill \Big|\hfill \\ {}\hfill \mathrm{K}\mathrm{s}\hfill & \hfill \hfill & \hfill \mathrm{V}\mathrm{s}\hfill \end{array} $$

Or, the scribe of one may simply have erred in his count (or erred in writing seven for five).

Schrader further compounds his error by crossing this first misconception (that is, entering the two texts “reconstructed” on the basis of the testimony43 of the Gītāmāna śloka one below the other) with a second misconception: although both versions of the śloka speak of an expanded Gītā of either 741 or 745 verses, he treats their information as applicable to the vulgate Gītā as well, and thus enters V alongside K44 into his stemma to yield a bipartite stemma with expansion of both branches in two stages. Exactly where Schrader is getting the idea that the Bhagavadgītā must have been expanded “from an unknown stage or number of stages to 671 stanzas (i.e., the vulgate minus 29 stanzas of Arjuna’s)” (Schrader 1938, 66; Schrader’s italics) is unclear since the Gītāmāna śloka does not speak of the vulgate having less verses. It is clear that Schrader is trying to square the testimony of this śloka with the evidence of the vulgate text, and since the only way he can square the fact that the Gītāmāna śloka’s figures are both simultaneously greater and lesser than the vulgate’s (greater in the total number of verses; lesser in the number of verses attributed to Arjuna), the only way he can do so is to treat the two as two separate operations performed on a common ancestor.45

Unfortunately, from the fact that a śloka refers to a Gītā with twenty-nine verses less spoken by Arjuna than in the version we possess, the only thing that we may conclude is that such a Gītā may have existed and not that our version is related to this Gītā by the addition of twenty-nine verses to a common core. For we do not know what the fifty-seven verses spoken by Arjuna in that other Gītā may have been: they may have been completely different verses or they may have included some of the ones we now try to identify as the “additions” in our Gītā, which is to say that Schrader’s calculations, carried out at such length and with every indication of profundity and respect for the learned views of his peers Richard Garbe, Moriz Winternitz, and Rudolf Otto, are worse than useless. Likewise, his projected bipartite stemma, though carried out with every intent of making a contribution to the “textual criticism” of the Bhagavadgītā, is pointless.46 Schrader has not understood the basics of how to work with manuscripts or how to reconstruct manuscript traditions and that he has intervened so massively in the Bhagavadgītā tradition, leading to confusions about the extent of the text that persist to this day,47 illustrates just how baleful the import of German “critical” methods into native scholarship has been for the textual transmission of Indian texts.48

Reconstructing the past using projected texts

The fragmentary and misleading nature of Schrader’s evidence49 makes a more comprehensive evaluation of his claims impossible at the present juncture,50 though it has already become clear that his desire to say something about the Bhagavadgītā has more in common with Sir Subramania’s efforts than we might like to concede. Thus, in the three remaining sections of this paper, we wish to draw attention to certain parallels between Schrader’s work and the efforts of Sir Subramania (or, as we like to call them, the Hamburg yogin and the Haṁsayogin [see Fig. 3]). These sections will also offer an interesting perspective on the sociology of the human sciences, especially “Indology.”
Fig. 3

The Haṁsayogin and the Hamburg yogin: Sir Subramania Iyer and Friedrich Otto Schrader

Schrader’s claims about the Bhagavadgītā were not restricted to merely identifying a slightly different text. Rather, this new text—even though, as we have just seen, the evidence for it was tenuous and the text itself unremarkable—became the basis for wide-ranging claims about Indian antiquity—claims that had direct consequences for indigenous intellectual and “religious” traditions. We first quote some of these claims, cited in order of their appearance.

Even before he had seen the London manuscripts (or any manuscripts whatsoever), Schrader had already staked out a position on the original extent of the Gītā, in conformity with German scholars’ views of the poem. Thus, in a 1910 article (Schrader 1910, 336–40), his earliest article on the Bhagavadgītā, he opined: “I am, completely in agreement with Garbe, of the view that the Bhagavadgītā has not always been the uniform and contradiction-free whole, as which it has since an indeterminate (and probably indeterminable) era been regarded universally in India; rather, [I am of the view] that it emerged through the expansion and possibly revision of an originally smaller work” (340). Even though he did not have any evidence on which to base these assertions, Schrader claimed: “here [in the Gītā] the later, viṣṇuistic version had been preceded by an earlier, unsectarian (“purāṇic”) version” (340). He also claimed that “the oldest Gītā as part of the pre-viṣṇuite Mahābhārata was already at an end with [verse] II, 38 (naivaṃ pāpam avāpsyasi)” (340). He followed this claim up in his 1927 essay with the assertion that “[the Haṁsayogin’s] fantastic statements at any rate show—something that was until now unknown and contested among us—that the idea of earlier and shorter versions of the Bhagavadgītā is present even in India” (Schrader 1927, 178; Schrader’s emphasis). In his 1930 publication, Schrader wrote, “the conclusion to be drawn from all this [namely, that his two commentaries did not refer to the vulgate text of the Bhagavadgītā] is that the vulgate of the Bhagavadgītā was still unknown in Kashmir by the end of the tenth century” (Schrader 1930, 8; Schrader’s emphasis). He added: “And hence it further follows that no commentary based on the vulgate can up to that time have been known in Kashmir, i.e., that the Gītā Bhāṣya attributed to Śankarācārya and for this reason believed to be the oldest Gītā commentary preserved to us was likewise unknown there [that is, in Kashmir] up to that time at least” (8; Schrader’s emphasis). Further, casting doubt on the authenticity of the Śāṅkarabhāṣya, he wrote, “but how is it that apparently neither the name of Śaṅkara nor any unmistakable reference to his system is found in Abhinavagupta’s works or those of his predecessors? Can the reason be that in Kashmir (as an ancient stronghold of Buddhism) Śaṅkara’s system was not looked at as something sufficiently original to require a special refutation?” (8–9).

The years 1931 to 1935 saw the publication of three further articles on the subject, with Schrader by now established as the reigning specialist on the Bhagavadgītā. Thus, in the first of these, “The Bhagavad-Gītā in Ancient Kashmir,” he wrote: “the existence of a Kashmirian recension of the Bhagavad-Gītā became first known to me when I found that in the great Abhinavagupta’s ‘Epitome of the Gītā’ (Gītārthasaṁgraha) some verses are explained which are missing in the vulgate. I then found in London the materials which enabled me to edit the text. Thus, every stanza besides those of the first book being repeated in the Sarvatobhadra, the complete text could be restored by me” (Schrader 1931, 748–49). In contrast to his earlier study, however, where he had argued for the originality of the “Kashmir recension,” he now disputed the authenticity of both versions: “now the discovery that the Gītā text commented on by Śaṅkarācārya, Rāmānuja, Madhva, etc., was preceded in Kashmir by a text diverging from it in about three hundred places naturally raises the question of which of the two texts is the original Gītā. The most likely answer I have to this question is: Neither. There are features in both versions which appear to be unoriginal. The inevitable conclusion, then, seems to be that we are so far not in possession of the original of the Bhagavad-Gītā, but only of two not very different recensions of it” (749).

Schrader’s next article focused on expanding these “historical” details. Against Edgerton (who had pointed out that, contrary to Schrader’s thesis of a “Kashmir recension,” the Kashmirian manuscripts examined by the editors for the Critical Edition, including the old and excellent manuscript no. 2137, did not contain the Kashmiri but the vulgate text), he argued that an examination of the same manuscript had shown that it had “traces of the Kashmirian recension such as na tv arthakāmas tu instead of hatvārthakāmāṁs tu (II, 5), omission of II, 67–68a-b, and many others.” He argued that “this is a confirmation of my claim that the suppression of the Kashmirian recension by the vulgate took place gradually: the traces mentioned are the last of a struggle [between the recensions] that is manifest not only in the separate manuscripts of the Gītā, in Gītā commentaries and quotations but also, naturally, affected the entire Mahābhārata tradition” (Schrader 1933, 41–42). He added: “With Śaṅkara’s triumph, the vulgate also triumphed in Kashmir and only it [that is, the vulgate] was copied except for in circles of the orthodox [altgläubig], which grew ever smaller” (42). Schrader appears to have been absolutely convinced of this thesis’s correctness, for he claimed: “the absence of the vulgate [text] in the Kashmiri Mahābhārata as well will be attested to by every manuscript from that period that may still come to light (to my knowledge, until now not a single birch bark manuscript of the Bhīṣmaparvan has been found). I therefore hold on firm [to the thesis] that in approximately the ninth to the eleventh centuries in Kashmir the sole recognized or—more probably—sole known text of the Bhagavadgītā is the one edited by me” (42).51 Schrader also criticized Edgerton’s thesis that the “Kashmir recension,” due to its greater extent and its more classical Sanskrit forms, was more likely a modern, corrected and expanded copy of the text. According to him, “this would then mean that in Kashmir the recension V once developed into K and later was then once again replaced by V. But so far we know nothing of an earlier existence of V in Kashmir [the same thing could be said, with equal justification of Schrader’s K; in fact, the argument is circular, because it presumes that K became the standard recension in Kashmir such that it would need to be “replaced” by V]; and by far the largest number of the 250 cases in which K diverges from V do not permit a decision regarding greater originality, whereas in over 30 cases, in my view, the probability of greater antiquity lies on the side of K” (42). In fact, as in his 1931 article, Schrader used doubts about the correct reading to cast doubt on the authenticity of both versions. Thus he averred that “If we therefore assume that the text of K has undergone changes—and this is surely most probable—and accordingly eliminate the obvious corrections and the verses that appear to be added from it [that is, from K], we arrive at a text as the foundation of K, that can be considered a parallel text to the vulgate, i.e., as a text that, together with this [the vulgate], goes back to an original that has not been preserved or that has not been found as yet. For the contemporary discipline [Wissenschaft] there can only be the one ‘working hypothesis’ that the text of the Bhagavadgītā was not always fixed, but rather was first [fixed] from the period onwards in which the vulgate, through the spread of Viṣṇuism all over India, which took place relatively late, compelled its universal acceptance” (42–43; Schrader’s emphasis).

Schrader’s final article in this series also returned to the theme of multiple recensions of the Bhagavadgītā, though now he linked it up with European searches for the “Ur-Gītā” since Wilhelm von Humboldt. According to Schrader, the figure cited in the Gītāmāna (745 verses, of which 620 are attributed to Kṛṣṇa, fifty-seven to Arjuna, sixty-seven to Saṁjaya, and one to Dhṛtarāṣṭra) could be explained through a combination of additional verses existing in an older source or sources (thus bringing up the total to 745, including forty-six additional verses for Kṛṣṇa and twenty-six for Saṁjaya) and interpolations in the current Gītā (responsible for bringing the total for Arjuna to twenty-seven in excess of the Gītāmāna figure). According to him, from the fact that no scribe or editor could have wished to omit verses once they were regarded as part of the Gītā, even those attributed to Arjuna, “it can mean nothing less than that those 27 ślokas, since they were missing but cannot have been omitted in Vaiśampāyana’s Gītā must be a later addition to the original Gītā, no matter whether the latter itself is only a later recension truer to the original than those known to us was the basis of the Gītā of 745 ślokas referred to in the summary” (1938, 63; Schrader’s italics). He continued: “that there are interpolations in the Bhagavadgītā was noticed as soon as it became known in Europe, viz., by the great Wilhelm von Humboldt who found it difficult to believe that the last seven adhyāyas up to XVIII 62 are an original part of the Gītā” (63; Schrader’s italics) and then presented a summary of European scholarship up to his day.52 In concord with a broad tradition of scholarship that has always had its problems with the eleventh chapter, the theophany of the viśvarūpa, Schrader argued that “it is easy to see that the bulk of the additions we are looking for must be contained in canto XI. For, this most admired canto of the Bhagavadgītā is on account of its highly imaginative character specially favourable to enlargement, and of the two cantos containing many ślokas spoken by Arjuna (viz., [chapter] I with 21 and XI with 33) it is the one where a fairly large number thereof can be spared. Not much is lost and the context is not disturbed if we omit stanzas 1 and 2, 15 to 30, and 36 to 44. This gives us exactly twenty-seven stanzas, it being remarkable that apparently not one more can be spared in the adhyāya in Arjuna’s speeches” (64; Schrader’s emphasis).53 At the end of these and still more complex operations, Schrader concluded that the Bhagavadgītā “riddle” (he means the fact that the Gītāmāna śloka refers to a text of 745 verses) “reveals by its very form the existence of two recensions of the Bhagavadgītā, to wit: one which was longer than the two known to us (the vulgate and the Kashmirian one) and one which was shorter and therefore, at least as regards the eleventh canto, more original than both of the two. There has been, as I said, since W. v. Humboldt’s days the persistent persuasion among Western scholars, shared also by some in India, that the original Bhagavadgītā must have been shorter than the current one” (Schrader 1938, 67; italics and emphasis Schrader’s).54

Sources of philological authority

We will not address further here Schrader’s claims for an “original” Gītā, as enough has been written on this subject elsewhere.55 Rather, we want to draw attention to the way in which the creation, indeed, just the mere suggestion, of a new text is enough to transfer philological and epistemic authority over the text from one set of guardians and interpreters to another. This is a phenomenon that can be observed very well in the case of Schrader’s “Kashmir recension.” Although his manuscript evidence was thin (in reality, he had done no more than found scribal variants in one manuscript, and possibly two commentaries that had been based on this text, that is, an ancestor of or a manuscript similar to this one), Schrader was able to leverage the ensuing uncertainty over the text to a position of authority. The editors of the journal The Aryan Path,56 for instance, introduced his essay with the words, “This paper embodies the substance of Dr. Schrader’s recent researches on the Gita together with some new ideas, and also, for the first time, a translation of the stanzas preserved in Kashmir only” (Schrader 1931, 748).57 Setting out from the hypothesis of a different version current in Kashmir (a hypothesis that was not even remotely proven by his sources), he developed this hypothesis into an expansive account of why the Bhagavadgītā, as read and transmitted in India for centuries, was neither the most ancient nor the most authentic. Not only was he able to challenge the normative form of the text, as it had been read and interpreted by Indian commentators for centuries but he was also able to rehabilitate western scholars’ theories of an “original” Gītā by arguing from the variation between the two recensions (a variation that can be quite satisfactorily explained in terms of the minute changes that are introduced each time a text is copied) that they must both have descended from an original different from the versions of all the commentators. Schrader did not explain how far back he placed this original or, indeed, if he identified this source with the text hypothesized by him as the “original” Gītā (that is, up to 2.38) or merely with a somewhat less expansive version of the text, but it is clear that, once this possibility was granted, there was no real reason to stop with the hypothetical source of the two recensions and any number of hypothetical Gītās (even the Ur-Gītā) could be admitted.58 Further, invoking a standard trope of the Indologists,59 Schrader argued that “the Kashmir recension was in ancient times the only Bhagavad-Gita existing in Kashmir, until, as late as about the eleventh century AD, it was ousted by the vulgate when the latter was introduced in that country together with the works of Śaṅkarācārya” (749; Schrader’s emphasis). Indeed, he developed this claim into a theory of a full-blown sectarian conflict in Kashmir, in which the traditional Gītā of the Pratyabhijñā school was replaced by the vulgate. According to him, the text of his sources itself revealed traces of “the last of a struggle [between the recensions] that is manifest not only in the separate manuscripts of the Gītā, in Gītā commentaries and quotations but also, naturally, affected the entire Mahābhārata tradition” (Schrader 1933, 41–42). Even though these traces can be much more easily and naturally explained through the normal means of contamination (for example, comparison of manuscripts, occurrence of variant readings in the margins or marginal glosses, absorption of these variants or glosses into the text), Schrader argued for a real historical event and a real consonance between the features he believed he could see in the text and this hypothetical event. Thus, he wrote: “with Śaṅkara’s triumph, the vulgate also triumphed in Kashmir and only it [that is, the vulgate] was copied except for in circles of the orthodox [altgläubig], which grew ever smaller” (60). In his 1935 article, he further expanded on this theory, arguing that “a combination of certain facts seems, indeed, to justify the assumption that the spread of the vulgate in Kashmir began at about the time of Yogarāja, i.e., in the twelfth century. Before this time there is not only no trace of the vulgate text of the Gītā but also, so far as I can see, not a single indubitable reference to Śaṅkara or his works. I am, therefore, inclined to believe that it was during the reign of the Kashmirian king Harṣa (1089–1101), whose love of Dākṣiṇātya fashion has been noticed by Kalhaṇa in his Rājataraṅgiṇī, that the works of Śaṅkara and his school, and so the vulgate of the Gītā on which Śaṅkara’s Gītā-bhāṣya is based, began to attract the attention of the Kashmirians” (Schrader 1935a, 148). Schrader conceded that “the lateness of this date—over three centuries after Śaṅkara—is embarrassing” (148). But he had an explanation ready to hand:

The dig-vijaya of Śaṅkara was very likely not nearly as complete as the traditional accounts would make us believe. The Kashmirian ācāryas, on the other hand, were so much engrossed in their own Śaiva darśana that they may have been practically inaccessible to any system which did not proclaim the Āgamas as its source and base. They were, moreover, acquainted with Buddhism, Kashmir having been its stronghold for a long time, and must have, therefore, regarded the Śāṅkara Vedānta, if they knew it, as what it really is: a Vedāntic adaptation from Buddhist philosophy. As such, i.e., unoriginal and heretical, as it appeared to them, it may have been ignored until the time when the flower of the Pratyabhijña school was over and a ruler of the country had taken a fancy to the south. (148–49)

Spurious archaeology, fraudulent presentation of scripture, and practical organization

Derrett has argued that the production of new, modern and yet supposedly timeless scriptures such as those created by the Śuddha Dharma Maṇḍala should be seen as a response to a unique intellectual climate, one in which Hindu scriptures, to be considered authoritative, both had to be ancient60 and yet could not call upon established systems of commentary and exegesis to offer a recontextualization (of the ancient wisdom), that is to say, the tradition of reception had been interrupted. Under these conditions, the only way Sir Subramania’s “yearnings for spirituality and for the spiritual regeneration of his country” could be met were through falsification of the documentation (Derrett 1977, 178).61 Derrett holds out the prospect of a renewal and a restoration of these systems (“The answers are bound to come with time. The pandits of that day will welcome a method of regulating society which will neither have a need for the participation of complacent foreigners, nor fear to build upon the flattened rubble of the latter’s ignorance, disdain, and neglect”; xv), but we, who still work under the conditions of this immense rupture and delegitimation of the tradition, must look for alternative ways to make sense of what has happened in history. Here we wish to offer a commentary by way of offering a parallel.

Even though obviously a forgery,62 Sir Subramania’s edition of the Gītā should not be seen as the sole example of a fabricated Gītā in this period. The early years of the twentieth century are rife with examples of apocryphal Gītās, from the “pantheistic Gītā” of Adolf Holtzmann Jr. to the “theistic Gītā” of Richard von Garbe, from the “epic Gītā” of Hermann Jacobi to the “Kṛṣṇa Gītā” of Hermann Oldenberg, and from the “trinitarian Gītā” of Rudolf Otto to the “Aryan Gītā” of Jakob Wilhelm Hauer.63 To these, we could also add the “soldier’s Gītā” of Theodor Springmann, the “Brahmanic Gītā” of Georg von Simson, and the “pseudo-German Gītā” of Mislav Ježić, not to mention the countless other Gītās in existence, either by hearsay or by imitation including the two most recent attempts (Vedavyas 1990 and Bhattacharjya 2013) to produce an “original” Gītā of 745 verses (Fig. 4 clarifies the relationship of these Gītās to each other). As we have argued in several articles, these texts must be seen as symptomatic of the peculiar conditions of the reception of Indian texts in modernity. Western and especially German academics had no interest in reading the texts as they were read in India and yet they desperately wanted to be a part of their unfolding history. As a text commentarial tradition, that is, a tradition in which the fundamental form of dialogue was the commentary written on the canonical works of the tradition, the Indian tradition was particularly vulnerable to contestation of its sources. Thus, the easiest way to disrupt it was to question the texts on which the tradition’s interpretations were premised, something that can be done most easily by raising the possibility of multiple and contradictory versions of the texts. Since the very system of debate nurtured in India depends upon a shared text (though interpretations can differ), by denying such a common basis western scholars could effectively paralyze the tradition. Further, if every text was already “sectarian” in origin (that is, either as a whole or in its parts, each of which was attributed to a distinct sectarian source), native scholars had every reason to mistrust the possibility of a philosophical exchange or dialogue. Thus, at the same time as they found themselves pushed into retreat on the texts and having to secure them philologically against western scholars’ criticisms, they also found themselves unable to restart ancient traditions of commentary. Modern scholarship had arrived in India: the pandits and śāstris’ dialogical partners would no longer be other native scholars with a commitment to shared principles of argumentation, evidence, and aim of scholarship, but the western scholars themselves.
Fig. 4

An Overview of Gītā Criticism, Illustrating the Influence of Schrader’s Higher Criticism on his Lower Criticism

Schrader’s researches on the Bhagavadgītā, occurring as they do at the cusp of this important period in Indian intellectual history, are revelatory of the processes at work in this transfer of epistemic authority from the native scholars to a new class of professional guardians of Indian texts: the western-trained historical critics. They illustrate how, by creating doubts about the normative form of the Bhagavadgītā, by highlighting the existence of multiple recensions of the text, by arguing in terms of the “religious” agendas and anxieties of those involved, and so on, western scholars were effectively able to interrupt the traditional transmission of texts and the accessory sciences required for their interpretation. Little wonder, then, that Schrader welcomed news of the Haṁsayogin’s Gītā with open arms.64 In place of the text of the tradition, western scholars advanced their own ideas of specialization, of critical research and of historical reconstruction. It is characteristic that none of the texts of the Bhagavadgītā produced in this period could get by without reference to western scholarship or western methods: the tradition had lost its ability to self-authenticate.65 Even the Haṁsayogin, that archetype of eastern “mystical” consciousness is a product of this new paradigm. His work gives testimony of the advent of historical consciousness in India: he writes that “the eventful Mahābhārata War” is a real event, having taken place “over one hundred and thirty centuries” ago (Row 1939, viii); the new text along with its commentary recommends itself over existing versions precisely in respect of “its utter catholicity and singular freedom from sectarian and class bias of any kind whatever” (vii–viii); and the arguments for its authenticity, peculiarly enough, are now no longer derived from the assent of a legitimating community or even from the simple expedient of claiming to have discovered ancient manuscripts in a secret cave, but rather, from text-critical or historical-critical reflections of the kind that would not have appeared (and, as a matter of fact, did not appear) unusual to Schrader.

The great men of old, high-souled protagonists of the several cults and mighty philosophers, did not, in polemical argument, advance the idea of the one pure and eternal Dharma, in which are the seeds of all faiths and creeds; since, it would not have served to set off the sectarian doctrines they might enlarge upon, to describe alongside the single, seminal ḍharma that is at the root of all sects and faiths and the knowledge of the oneness of everything. Further, they did not care to commit their most intimate doctrines, the subtlest Truths, to the mere treatises they wrote. It must now be abundantly clear that it is against all argument to say that the Geeṭa of the Śuḍḍhas cannot be authentic simply because great men of the past have not referred to it, in works mainly controversial, where they sought to confound their opponents by loudly setting forth their own doctrines. (Srinivasachariar 1917, 9–10)66

Now, for centuries, the Geeṭa has been known to consist of only eighteen chapters; as such indeed did the great men of yore who interpreted the Geeṭa, Śankara, and others, together with their followers, accept it. Had they been aware of this other Geeṭa prized by the Śuḍḍhas, would not they have made in their works at least a single reference to it, were it only to disagree with its drift and teachings, refute and condemn it, and thus strengthen the cause of their own doctrines they expounded? Therefore those that rely on authoritative sources should approve of such a work as this Geeṭa of the Śuḍḍhas. (9)67

The flow of knowledge between two cultures is always a two-way process, but, in this case, we think the credit for teaching the Indians techniques for producing fabricated texts rests squarely with the Germans.


We began this paper with a quote from Luther as a segue; let us now draw everything together. Even though Roman Catholic versions of the Bible are themselves not free of interpolations and glosses, perhaps no translator of the Bible was ever as wanton and unapologetic about his changes to the text as Luther, as recorded in his famous exchange with Emser in 1530.68 In fact, as Luther makes clear in this exchange, the aim of his translation is not to restore a more original reading, but, through introducing the possibility of differing interpretations of the same text, to contest the traditional authority of the church.69 But Luther was aware that once the possibility of contesting the meaning of the received text is admitted there is no end to the process, and that is why in the very next line following this line, that is, after the line “for of playing the master and the critic there is nowadays neither measure nor end,” he added: “And let every man be warned against other copies [that is, other than his own].” In other words, he saw that the question of the correct reading ultimately comes to an end only with a decision in favor of one text over another: it is a question of which text one finally recognizes as authoritative. The introduction of the possibility of different readings of the text, then, does not so much eliminate the question of authority as it serves, rather, to create a fluxing configuration out of which a new authority may arise.70 In a similar vein, when German Indologists from the nineteenth century onwards contested the normative form of the Bhagavadgītā, they did so not in order to arrive at a more stable, more accurate reading of the text (one clearly does not enhance confidence in a text by casting doubt on its readings) but precisely to create a situation of uncertainty, in which established authorities might be discarded for new ones. By massively destabilizing the transmission of the text (as Luther in his time before them had done with the Bible), Indologists created a different source of intellectual and institutional authority, and thus what they did was political through and through. At the latest when a reader reads the Bhagavadgītā she will collapse the many potential readings of the text to one.71 Thus, she will not trouble too much whether the reading is hatvārthakāmāṁs tu gurūn ihaiva or na tv arthakāmas tu gurun nihatya. As long as both yield an acceptable sense, she will select one, because she is ultimately interested in using the text for self-reflection and self-transformation, that is to say, in an ethical relation. It is only the Protestant who, thinking he is saved sola fide (by faith alone), can endure in the face of an infinity of readings. But this infinity is a bad infinity. It indicates the apotheosis of subjectivity, but a subjectivity dirempt of all humanity, even in the minimal form of a concern for the self. It is an empty placeholder, even though the academic and critic may pride himself, thinking he is superior to the “naïve” reader, who always makes a choice when reading the text. Hovering between damnation and faith, the Indologist, like Luther, cannot choose, or, rather, he chooses not to choose, thus keeping faith open as a possibility. Pasquali’s words, “with regard to recensio, philologia profana is still, without knowing it, a tributary of philologia sacra,”72 thus still hold true, even if perhaps in a sense different from the one the author originally intended.


No. 6763 D; British Museum, London.


No. 3271; India Office Library, London. Schrader also used a second Gītā commentary, the Bhagavadgītārthasaṁgraha of Abhinavagupta, but as he did not have access to the manuscripts for it (he cited it according to the Nirnaya Sagar Press edition of 1912), we have not counted it here.


Some of this is, of course, rehashed, as, for example, Brockington’s introduction (2015) to the recent Norton edition of the Bhagavadgītā (Flood 2015). But the question still remains: why has scholarship not penetrated beyond the original purposes for which it was created? (Brockington’s arguments against the originality of the Bhagavadgītā, discussed in Brockington 1998, are addressed in Adluri and Bagchee 2016a.)


ṣaṭ śatāni saviṁśāni ślokānāṁ prāha keśavaḥ |

arjunaḥ saptapañcāśat saptaṣaṣṭis tu saṁjayaḥ |

dhṛtarāṣṭraḥ ślokam ekaṁ gītāyā mānam ucyate | |


Born March 19, 1876 in Hamburg, Friedrich Otto Schrader studied Indology in Göttingen, Straßburg, and Kiel, graduating with a doctorate from Straßburg in 1902. Schrader was active in India from 1905 to 1916, where he served as the director of the Adyar Library in Madras; he returned to Germany in 1920. From 1921 until 1945, Schrader was the Ordinarius for Indology and comparative linguistics at the Christian-Albrechts-Universität in Kiel, where his researches focused mainly on Buddhism, the Dravidian languages of India, and on creating general surveys of Hinduism. Further biographic details can be found in Sprockhoff (1963a, 1–11); a bibliography of his publications in Sprockhoff (1963b, 12–32 and 1982b, xi–xxii).


All of Schrader’s publications on the Bhagavadgītā, with the exception of Schrader 1928, 1935a, 1935b, and 1958, are reprinted in Schrader 1982.


The date is only approximate: no historical records appear to exist for its establishment, but David and Nancy Reigle claim that “The existence of the Suddha Dharma Mandala (Śuddha Dharma Maṇḍalam) was announced by Sir S. Subramania Iyer in the article, ‘An Esoteric Organisation in India,’ published in 1915 in The Theosophist” (Reigle and Reigle 2007, 3). According to Derrett, this article was actually published in four parts in The Theosophist 36, no. 2 (1915): 407, 499, and 614 and 37, no. 1 (1915): 196–206 (Derrett 1977, 178, n. 91a). The original article is cited as Iyer 1915 in the bibliography.


The reference is to Edgerton (1925, 99).


Schrader’s comments on the Haṁsayogin’s Gītā and his attempts to date and locate his work using style and vocabulary make delightful reading and are one of the best examples of the pitfalls the western scholar falls into when he attempts to adopt a critical tone vis-à-vis the past. For this past can quite often be bogus, a deliberate deception on the part of those involved. The Haṁsayogin’s Gītā presents us with a case where we can clearly see this because we have actual historical information about its dates and manner of composition, but how many other such works must there be where the author adopted the tone of a different historical period? Historical-critical reconstructions, the stock in trade of the Indologist, depend upon the attribution of a tremendous naïveté and lack of dissimulation to Indians, at the same time as the Indologist asserts that Indian accounts of history are not to be trusted. The contradiction is unsustainable and lasts only a little longer than the Indologist’s hold on the public imagination.


Schrader dismisses Sir Subramania’s introduction with the words “the extensive, but unfortunately completely unscientific [unwissenschaftlich] introduction of the deceased Dr. Sir Subramania Iyer” (Schrader 1927, 173, n. 1), but if he had only seen that this was the only part of the edition that really deserved attention and that it was the closest to a genuine treatment of the conditions of the origin of this version. Obviously, the distinction between the “scientificity” of the German scholar and the “unscientificity” of the native scholar is too narrow to encompass all the possible permutations and combinations that might result. In this case, it is the Haṁsayogin’s Gītā, which Schrader treated all unawares as genuine material to be historically studied and determined and related (that is, to a western audience) by him, which is clearly and definitely unscientific (and not just in the sense in which for the Indologist all materials from India’s past are “unscientific” but in the sense of not being a suitable object for scientific research), whereas Sir Subramania’s introduction, dismissed as “completely unscientific” by Schrader, is actually the only available object of research here: it is this that Schrader should have studied had he wished to arrive at reliable historical conclusions.


This is perhaps not the ideal choice, as the first chapter is not representative (the first chapter is preserved only in the birch bark manuscript and there are no variant readings). A discussion of Schrader’s reconstruction of the text, however, would have exceeded the scope of this paper.


Italicized text indicates the changes in Schrader’s sources; terms in parentheses are the corresponding Critical Edition readings. The Critical Edition reading is the same as the vulgate reading in all the cases cited by Schrader, except for 1.28c where the Critical Edition reading (dṛṣṭvemān svajanān kṛṣṇa yuyutsūn samavasthitān) is exactly that given by Schrader as the reading of his Kashmiri source (the corresponding vulgate reading, as given by Schrader, is: dṛṣṭvemaṁ svajanaṁ kṛṣṇa yuyutsuṁ samupasthitam).


Note that this chapter is based on the evidence of a single witness, the Kashmiri birch bark manuscript no. 6763D (the two commentaries only make a few comments about the first chapter), so that the readings Schrader has listed are in fact lectiones singulares and would have been eliminated as such in the process of reconstruction.


Schrader’s reading of the evidence is rather different. He also groups the readings into four categories, but his categories differ; and are as follows:

  1. 1.

    “There are, first, in K a small number of readings which are apparently but corruptions of V” (for example, 1.28 cd, 2.5c, 2.43 cd, 2.51c, 3.27, 5.5, 14.24, 17.6, 2.60a).

  2. 2.

    “Secondly, there is a large number of readings in K which look like grammatical or stylistic emendations of V” (for example, 3.23a, 5.24a, 9.14, 10.16, 10.19, 10.24, 11.41, 11.48, 11.54, 18.8).

  3. 3.

    “Thirdly, there is the rather numerous class of readings which appear to be original readings of the Gītā preserved in K but corrupted in V” (for example, 2.5, 2.11, 6.7).

  4. 4.

    “The three readings I have so far dwelt on are, I believe, the most interesting ones of all; the rest can be dealt with in a much shorter way” (1.7, 3.2, 5.21, 18.8, 8.17, 6.16, 7.18, 11.8, 11.40, 11.44, 13.4, 17.23, 18.50, 18.78, 1.7, 2.5b, 2.6d, 2.10, 2.11, 2.21, 2.35, 3.2, 3.23c, 3.31d, 3.38, 4.18d, 6.21a, 6.21d, 6.28c, 6.37a, 10.42, 11.43, 16.3b, 16.8d, 16.19, 17.13, 17.26) (Schrader 1930, 11–18; all emphasis Schrader’s).


“A Provincial Recension of the Bhagavadgītā such as Schrader claims for Kashmir should imply that all or nearly all Mss. hailing from that Province through direct or indirect line of scribal transmission exhibit a sufficient number of varietas lectionis which (a) are generally common to the group and (b) are not to be found, except sporadically, in other groups of Mss. belonging to other Provinces. We have now seven ‘Kashmirian’ sources to deal with: 1. The London Śāradā MS used by Schrader (Lb); 2. The Commentary of Abhinavagupta (Ca); 3. The Commentary of Rāmakavi (Cr); and 4–7 of our Mss. which provisionally are designated Ś, K, H, and A” (Belvalkar 1939a, 214).


“[T]he so-called Kashmirian Recension is not peculiar to [the] Kashmirian text-tradition, but is more widely distributed, so much so that in a few cases I have adopted it for the Critical Edition without even a wavy line underneath, and in others with the wavy line” (215).


“[W]hile in quite a number of other cases, although neither of the above procedures was adopted, the rejected Kashmirian reading received support from such diverse sources as to place it beyond ‘Provincialism’ and in a few cases even demanded a wavy line below the adopted Vulgate reading” (215).


We have not cited here Belvalkar’s distinction between a “version” and a “recension” (a “version” “embody[ing] modifications happening during the course of scribal transmission from a common codex”; a “recension” “connot[ing] more deliberate and far-reaching alterations in the text”; 214, n. 1), both because we do not think the distinction is sustainable and because Belvalkar himself later abandoned the claim (see Belvalkar 1941, 25–26).


“Schrader (with some previous interpreters) feels that an api is needed after arthakāmām; but none is at all called for, and those translators who supply api misunderstand the meaning. The Kashmirian text reads arthakāmas, a nominative agreeing with Arjuna; but its version of the line is otherwise obviously corrupt and secondary, as Schrader himself admits (na tv arthakāmas tu gurūn nihatya; note double tu!), and there is not the slightest reason to abandon arthakāmān” (Edgerton 1932, 72).


“With Boehtlingk and Garbe, paramātma is here to be taken as precisely equivalent to atman; Schrader’s objections to this are purely subjective and sufficiently disproved by the parallels which Garbe quotes. The individual self is repeatedly and in all possible contexts called īśvara, paramātman, and all other epithets which apply to the Supreme One; the plain fact being that in early Sanskrit texts these terms mean both at the same time, and it is rarely if ever possible to draw a sharp line between the two concepts. The Kashmiri version, parātmasu samā matiḥ, is in itself harmless, but there is not the slightest reason to suppose that it was more original” (72).


“I continue to find the doubled iva suspicious” (Schrader 1933, 47).


“Non liquet” (47).


Matan admittedly cannot be thought of as an improvement but certainly as a corruption such as have often defaced such a completely clear text” (47).


“There are also other metrical errors to be found in K and might in fact be an indication of its greater age. But Edgerton rightly points out that samāpnoṣi here (in contrast to the commentators) can be understood as ‘attain, win’” (47).


“One has also earlier attempted to help oneself out with irregular sandhi. But I consider it most improbable that the Gītā would already have applied the image of a lover and his beloved to God and his devotee” (47).


“I consider it out of the question that with brahmasūtrāṇi the same thing is meant here as with vedānteṣu in Mbh. XII, 8971. I have heard Upaniṣads, even the ones composed in prose, ‘sing’, but never sūtras” (47).


“Even if Edgerton could cite a text according to which the Brahmans are created from the syllable om (that this is true of the Vedas is well known), the difficulty would remain that the following verses by no means fulfill the parallel between om, tat, sat and Brahmans, Vedas, and sacrifices he supposes, but rather, only connect om, tat, sat and asat in a vague manner with sacrifices, donations and asceticism, but nowhere connect either om nor tat nor sat specially with the Brahmans” (47).


“As before, I am of the same opinion as Schlegel (an opinion also shared by Böhtlingk): ‘At, vel sexcentis codicibus repugnantibus, nihilominus contenderem, hunc versum emendatione egere. Sunt frequentissimae versuum clausulae: iti me matiḥ vel iti matir mama, sed particular iti nullo modo ab iis abesse potest’. The original text [Urtext] either had dhruvā iti (K) or the reading dhruvāṇīti [sic] found by Schlegel in two of his manuscripts and supported by Rāmāyaṇa II, 20, 29” (47).


“Edgerton correct: nāyakā (V) more original than nāyakān (K)” (47).


“A decision is not easy, but Edgerton may be right” (47).


“Along with Edgerton (Garbe) I now consider the lectio difficilior to be acceptable, [and] K to be secondary here” (47).


A splendid example of this is offered by Speyer (1902, 123–25). After noting that Kṛṣṇa in Bhagavadgītā 2.11 praises Arjuna’s “wisdom,” he asks, “which wisdom [welche Weisheit]?” (124) and argues that there is no wisdom in the preceding passages that might justify this interpretation. Native commentators, in fact, had to come up with an addendum (“you speak words of wisdom, but you are not wise”) to explain this discrepancy (that is, they interpreted Kṛṣṇa’s words ironically). He continues: “there is also, in addition, a philological obstacle, which, however, the theological-dogmatic exegesis of the native commentators did not consider very significant,” and then he discourses: “However, it [the philological obstacle] is there and, when correctly recognized, a significant factor for doubting the correctness of the vulgate [text]. I doubt very much if a composite like prajñāvāda, ‘words of wisdom’ = wise words, wise speech, can exist in Sanskrit at all. Probably one could say prājñavādaḥ = prājñāṃ (or prājñasya) vādaḥ but prajñāyā vādaḥ would only be permissible if the prajñā were thought of as a person. Only an interpretation such as λόγοι τῆς Σοφίας, ‘words spoken by Prajñā’ is excluded in our passage. The explanation prajñāvādaḥ = prajñāvatāṃ vādaḥ is, considered purely philologically, awkward and untenable. It corresponds much rather to the language of the Bible and to Latin—verba sapientiae = verba sapientis or sapientium—than to Sanskrit, where the use of the abstract term to characterize the bearers of the term is unknown in this dimension” (124–25). After concluding, “in short, prajña is corrupted,” Speyer instead offered the emendation prajā. He argued that this term better brings out the contrast between the general opinion and the view of the wise and that its use in Kṛṣṇa’s speech, where it “has the meaning of pṛthagjana, Greek ἰδιώτης” “indicates the sensitivity of understanding [Feinheit der Auffassung] of the talented composer of the Gītā.” “He will probably have placed this prajā Kṛṣṇa in his mouth in order to express himself as gently as possible vis-à-vis Kṛṣṇa there where he has to reprimand his friend [that is, Arjuna]; how easily he [that is, the poet] could have said bāla-, ajña-, mūrkha-vādān bhāṣase” (125).


As reported by Sauppe (1872, 89).


See Sauppe (1872, 89): “Not only does he [Heyne] repeatedly describe it as a kind of vanity of wanting to seem brilliant through emendations, [a kind of vanity] from which he has freed himself, but the insignificant variants of such ancient witnesses of the transmission as we have them in Virgil in the manuscripts of the fourth, fifth and sixth centuries, which he knows and records, also move him only rarely to make a change to something once it is present in the editions or to a discussion of the differences in the transmission.”


In his earliest article on the Bhagavadgītā and the earliest on the topic, Schrader averred that the “original” Gītā would have only comprised chapter 1 and chapter 2, verses 1–38, that is, a total of eighty-five verses. See Schrader (1910, 336–40). It is not clear whether Schrader had this figure in mind when he spoke of the “original” in the 1938 article.


Schrader does not say why exactly these verses must be eliminated; probably he is deferring to Garbe’s views here, which he cited a little earlier.


Schrader would regard the fourteen full and four half-verses as being, tentatively, original, remove an unspecified number of verses (y) from these (which ones, he did not specify), and reconstruct the archetype. In contrast, he would regard all of the additional verses in V (+2) as being interpolations, as they are not present in K. But this would not be a stemmatic reconstruction: his hypothetical archetype would be based on a single witness (K) corrected ope ingenii and using completely arbitrary criteria. Even if we were to get around this problem (for example, by admitting that both K and V had been produced through mechanical transcription of a common source), where K and V give different readings, we would have a crux. We would have to choose the readings of one based on intrinsic grounds, and, since the standard assumption in textual criticism is that scribes never consciously omit anything (but may add verses of their own), the presumption must be that the additional verses in K are additions in this recension rather than omissions in the rival recension—a fact both Belvalkar and Edgerton noted.


For further examples that this is what the Indologists were really doing most of the time, when they claimed to be engaged in “textual critcism,” see Adluri and Bagchee (2014) for examples from Richard Garbe, Hermann Oldenberg, and Georg von Simson; Adluri and Bagchee (2016a) for examples from Jarl Charpentier, Mislav Ježić, and Przemysław Szczurek; and Adluri and Bagchee (2016b [In press]) for examples from Andreas Bigger and Michael Witzel.


The simplest way for the fourteen full and four half-verses less y verses not found in V but found in K' (and of which Schrader claimed that all of them are features of their common ancestor O) to have descended to K' from O is if O contained 714 full and four half-verses less y verses. In that case, V would have inherited 700 of those verses but only K' would have inherited the full complement of verses lost in V. If O contains only 700 verses and we are to assume with Schrader that the fourteen full and four half-verses less y verses unique to K' are all original, then an equal number of verses (that is, fourteen full and four half-verses less y verses) must be eliminated from both V and K', that is, not all of the 700 verses in V can be original and likewise a corresponding number of the 700 remaining verses in K' must be additions in K'. We might possibly be able to identify these verses if K' and V differed in respect of these verses, but since they are in fact identical we must assume that the same verses (and not just the same number of verses) as were added to V were added to K'. Eliminating them, then, will become a matter of subjective iudicium and we will be returned to the same position as the nineteenth-century critics of the Bhagavadgītā such as Richard Garbe and Hermann Oldenberg, who gaily went through the text and eliminated verses that seemed unoriginal to them. If, however, O has less than 700 verses, as Schrader’s argument suggests, then we would not even know how many verses in V and K' are to be eliminated as additions from them: O might have had any length, even as little as eighty-five verses (the number Schrader opted for in his 1910 article as being the length of the “original” Gītā). The only thing that is given here—given not because it is proven by codicological or philological investigations but because Schrader insists upon it—is that the fourteen full and four half-verses less y verses of K' are features of the original O. We do not have much reason for confidence in Schrader’s abilities either as a textual critic or as a critic of the Bhagavadgītā.


Something similar applies to Van Buitenen’s attempt to revive Schrader’s argument for a “K” recension (Van Buitenen 1965, 99–109). Van Buitenen treats Schrader’s K as a descendant of Bhāskara’s text (denoted by the siglum Bh. in his—explicit—stemma), Śaṅkara’s and Bhāskara’s texts (the former denoted by Ś) as apographs of a common source (O), and argues that Bh. must have descended from a second source (besides the vulgate, that is, the text of the Śāṅkarabhāṣya) that he had before him, “which must be considered a Vorlage of what now survives as the Kashmir version” and “is consistently superior to [the] version of the Gītā now known from Kashmir sources” (104 and 105). (Note that this makes Van Buitenen’s stemma contradictory to what he says of it; but this issue demands a separate article and we cannot get into it now.) Unfortunately, Van Buitenen has no way of getting from Bhāskara’s text to the hypothetical ancestor of Śaṅkara’s and Bhāskara’s texts (which he designates with O), the more so as he considers Bhāskara to have had access to and combined readings from two sources (Śaṅkara’s text and the hypothetical ancestor O). Bhāskara’s text Bh. is playing the same role in his stemma as Schrader’s hypothetical ancestor of K, K', and the fact that he has a manuscript source for Bh. only seemingly overcomes the problem of moving from K to Bh. (or K' as the case may be) because his procedure for establishing the relationship of Bh. to Ś is faulty: Van Buitenen treats Bh. and Ś as apographs of a common source, even though he has only examined the first chapter of Bhāskara’s text against the vulgate and he does not have a first chapter for Śaṅkara’s text to compare against Bhāskara’s (the first chapter is not preserved in Śaṅkara’s commentary). In other words, he enters what are essentially two different works into the same stemma as though he has been able to establish their relation and this is because he has treated Śaṅkara’s text as being essentially equivalent to the vulgate even though he criticizes the editor (that is, Belvalkar) for doing so. It is clear that the confusion is in Van Buitenen’s own mind: the constituted text is based on the evidence of the manuscripts and those commentators who did, in fact, comment on this part of the Gītā. Belvalkar cannot be accused of consituting the first chapter on the basis of Śaṅkara’s commentary because he explicitly notes that “Cś [the commentary of Śaṅkara] begins from st. 11 of Gītā adhy. 2” (Belvalkar 1947b, 114). The relationship of Bhāskara’s text to Śaṅkara’s must first be established using those parts for which both are preserved; only then can Van Buitenen’s inferences regarding their presumed ancestor be legitimate.


Tadpatrikar does not mention which these manuscripts are, other than to note that one is in Śāradā characters and the other in Kashmiri Nāgarī. Belvalkar refers to them as K1 and Ś (Belvalkar 1939b, 335).


We are calling the version found in the manuscripts Ś1, K0–2, B2–4, Dn, and D4.8 (*112 in the Critical Edition) the “vulgate” version of the Gītāmāna śloka. Perhaps “northern” version might have been more accurate, except for the circumstance that the two Kashmiri manuscripts mentioned by Tadpatrikar are also northern sources.


By “testimony” should here be understood not what scholars generally call “manuscript evidence” for Schrader has not in mind the evidence of its readings but the fact that a verse either gives or appears to give information about some source (or, possibly, about the scribe’s own exemplar). It is thus also not what scholars have termed the “codicological evidence” (Timpanaro calls it the prove materiali or the “physical evidence”) but a form of hearsay contained in the tradition itself. What exact weight should be assigned it (and whether it should be assigned a weight at all) is very much in doubt. Certainly, the two other attempts to account for it, S. K. Belvalkar’s and S. N. Tadpatrikar’s (1939b and 1937), are not very convincing. Vishnu S. Sukthankar as a matter of fact thought that no weight should be assigned to such figures (his main example was the Parvasaṁgraha figures of the Mahābhārata but the principle applies to all other such calculations) because of the ease with which “the data of the Parvasaṁgraha can be manipulated”—“far more easily [at any rate] than those of the manuscripts of the text” (Sukthankar 1933, xxxiii).


Note that these sigla differ from those used earlier and refer to the vulgate and Kashmirian recensions of the Gītā (the earlier ones referred to the vulgate and Kashmirian versions of the Gītāmāna śloka, that is, not to manuscripts or manuscript versions but to variants of a specific verse and, still more accurately, to the differing accounts contained in those verses).


That is, that first the vulgate Gītā was developed from the common archetype of 671 verses by the addition of twenty-seven verses, then the Vaiśaṁpāyana Gītā was developed from the same ancestor by the addition of seventy verses (though not the twenty-seven attributed to Arjuna added in the vulgate), and then another two.


Recall that textual criticism only functions on the basis of the assumption of mechanical transcription of a single source; if the scribe made conscious changes to his exemplar, both adding, dropping, and composing new verses, we would have no reason to treat it as an apograph of the same source. We might attempt to derive some relationship through noting the common passages between it and another exemplar such as the vulgate but note that this would be an exercise in higher criticism. It would furthermore be subject to a basic objection that arises in relation to all studies in Quellenkritik: rather than being an independent method alongside Textkritik, Quellenkritik is actually a specific subset of the latter—indeed, an erroneous application or a malformation of the latter. Quellenkritik attempts to identify the common source of two texts based on their shared passages but this runs into the fundamental objection that only shared innovations prove manuscripts related. Two manuscripts that have the same passages may have both gotten them from a third or from different sources altogether or one from the other. But without significant errors to connect them and without an appreciable length of common text containing a significant number of such errors (this is why textual criticism does not work for very short passages such as might be quotations of one text in another) we would be unable to establish their relationship.


See the editions by Vedavyas (1990) and Bhattacharjya (2013). And see also Shastri 1941, discussed and already rejected as an apocryphal edition in Belvalkar 1943.


A similar argument is made by Patrick Olivelle (1998, 173–87).


The complete extent of his explanations to his reconstruction in his 1930 article is as follows: “The arrangement of the edition attempted in the following pages requires just a few remarks. The ‘vulgate’ (V) which I have used as the basis of my comparison is the text of the Gītā as known to Śaṅkara and his school. My critical apparatus is naturally but small. There is none at all for the first adhyāya, because for it the birch-bark MS (B) only was available. For adhyāyas II to VIII (exclusive of VIII, 18 ff.) I had the two commentaries (R, A) and the birch-bark MS, for the rest only the two commentaries, or in the many cases in which A is silent, only R in its twofold form: the text (Rt, occasionally vitiated by V) and the comment (R) following each verse or group of verses. For the establishment of the text the commentaries had, of course, to be decisive: when a reading was found explained (not merely mentioned) in one or both of them, it was necessarily preferred to a different reading in B or Rt; and in such cases the latter if agreeing with V, i.e., obviously secondary, was as a rule not mentioned at all. In many cases where no apparatus is given this means that the reading concerned is the only one and sufficiently guaranteed. Where R or A or both of them are not mentioned, though expected (as, e.g., ad VI, 40), they do not repeat the word or words concerned nor throw any light on them. Differences between R and A are always noticed; between these (or either) and B, however, as a rule only when B has not the vulgate reading. Rt is generally not given except where it disagrees with R (or A, B) otherwise than by agreement with V. The extracts from the commentaries contain the complete comments on the stanzas or half-stanzas missing in the vulgate, so far as these are explained at all, and, moreover, all that is essential in the explanation of the readings peculiar to K. Obvious clerical errors, inclusive of wrong punctuation, I have silently corrected” (Schrader 1930, 18–19; emphasis in original).


We have examined P. C. Divanji’s critical apparatus of section B (appendix II) in the Critical Word Index to the Bhagavadgītā (Divanji 1946). Unfortunately, as those readings are based on Schrader’s 1930 edition, as well as there being other problems with his procedure (1946, xv and 193), this does not help very far. Divanji is altogether too sympathetic toward Schrader’s theories.


Schrader’s assertion was proved wrong: all of the Kashmiri manuscripts of the Mahābhārata (specifically of the Bhīṣmaparvan) found turned out to contain the vulgate version of the Gītā.


“More than a century has elapsed since (1826), and during it many scholars have expressed their opinion on the problem with the result that in the West the almost general opinion is at present that the Gītā cannot have been from the beginning what it is now. Farthest of all went the late Professor Winternitz who, not satisfied with Garbe’s deletion of 170 stanzas, refused to recognize as original parts of the Gītā the whole of the famous eleventh canto as well as the whole of the last six cantos with the sole exception of XVIII 55–56. The last one who wrote on the problem is the late Professor Rudolf Otto of Marburg University. He endeavoured to show that the Gītā had much the same evolution as the Mokṣadharma: as the dying Bhīṣma’s ‘few reconciling and consolatory speeches,’ which alone could have been part of the epic proper, became the ‘nest’ of numerous treatises, so the ‘primitive Gītā’ (Urgītā) of 156 verses gradually grew, first by the intrusion of eight ‘didactic treatises’ (Lehrtraktate) and then by many ‘glosses,’ the same (with some exceptions) as, and a few more than, those pointed out by Garbe. His view of canto XI was identical with Humboldt’s, viz., that it is the very acme of the Gītā, and thus diametrically opposed to that of Winternitz” (63–64).


After making such massive and unwarranted incursions into the text, Schrader rather brazenly accuses the hypothetical “redactor” of the “original” Gītā of a “complete absence of a sense of propriety (let alone historical sense)” (65, n. 1). We might say the same thing of him. For more examples of Schrader’s wanton, irresponsible, and partisan approach to identifying “interpolations” and proposing conjectural “original” readings, see Schrader 1929a and 1929b.


On the German scholars’ need to claim a more illustrious predecessor for their method of “criticism” in Wilhelm von Humboldt, see Adluri and Bagchee (2016a, 285, n. 98).


See Adluri and Bagchee (2014, chapter 3), and see also Adluri and Bagchee (2016a).


An English-language journal founded in Bombay in 1930 and published by the Theosophy Company. Its editor and board were drawn from Bombay’s elite anglicized Parsi community.


And see also Schrader (1936, 107–18) where Schrader bitterly fought out the battle for authority over the text with the “scholar of religion” Rudolf Otto. In his words, “it is extremely disconcerting that this ‘Ur-Gītā’ [the reference is to Otto’s reconstructed text, bearing in mind that not only Schrader himself but also Richard Garbe, Hermann Jacobi, and Hermann Oldenberg had themselves offered reconstructions of the “original” Gītā; apparently, these posed no problem for Schrader, since they were “indologically” legitimated], even before critical scholarship [Kritik] could say a word about it, has been placed before a wide audience. For, first of all, [Otto’s edition] is still lacking in the foundation on the basis of which such constructions can attain security: the philological investigation, oriented toward [a knowledge of] grammar, vocabulary, style, and meter” (112).


This was, in fact, how Schrader was understood by Belkvalkar. Belvalkar wrote: “The traditional extent of the Bhagavadgītā as reported by Śaṅkarācārya is just seven hundred ślokas or stanzas, and the orthodox Indian Commentators have attempted to make these seven hundred stanzas (neither more nor less) yield a self-consistent system of Ethics and Metaphysics. The late Professor R. Garbe and his pupil, the late Professor Rudolf Otto, essayed to prove the inherent impossibility of such an attempt by drawing attention to the composite nature of the present Bhagavadgītā. Garbe postulated two disparate strata in the Poem: Otto was not content with anything less than ten or eight of them; but neither has, in my opinion, succeeded in proving that the Gītā in its present form is incapable of being understood as a whole which may allow for the original divergent thought-phrases (when established as such), and yet transcend them all in a higher philosophical synthesis. Another German savant, Professor F. Otto Schrader of Kiel, has attempted to attack the authenticity of the traditional extent of the Bhagavadgītā from a somewhat different point of view. Schrader tries to show that the text of the Poem to which the Gītābhāṣya of Śaṅkarācārya gave currency (and consequently the Gītābhāṣya itself) was completely unknown in Kashmir upto about 1,000 AD, some two hundred years after the time of the great Bhāṣyakāra. To persons brought up in the belief that the text of the Bhagavadgītā has remained, like the Vedas, almost immune from varietas lectionis this would come as a great shock, particularly if it is claimed, as Schrader in fact claims, that the earlier and hence the authentic Gītā is that preserved by the Kashmir Recension, and not the one on which Śaṅkara wrote his Bhāṣya. It is of course not claimed that this ‘Kashmirian’ Bhagavadgītā, from the purely philosophical point of view, differs vitally from the accepted text; but once Schrader’s thesis is accepted as proved, it raises the possibility of other recensions of the Poem being current at different times in different parts of India. All this, once admitted, would go to discountenance the view that the Bhagavadgītā had a definite philosophical import and so had once constituted an important landmark in the evolution of Indian Philosophy. Schrader’s thesis is, it follows, of more far-reaching consequence than would appear at first sight.” (Belvalkar 1939a, 211–13). Schrader was not unaware of the opposition to the fragmentation of the text in India (see Schrader 1929a, 174: “Still the idea of interpolation in the Gītā must not, as is often done in India, be ridiculed as the caprice of hypercritical minds”).


For further examples, see Adluri and Bagchee (2014).


“Why was it necessary to suppose that ancient works were brought to light, while at the same time circumstances made it evident that the compositions were recent, indeed contemporary? What mental process allowed this self-deception to occur? Firstly as the Sanātana Dharma Dipikā shows, and evidence of the conditions of thought in the 1770’s confirms, Hindu scholarship is wedded to the notion that all authoritative scripture is part legislation and part prophecy and is located, actually or ideally, in the first years of the Kali-yuga, except for works of even greater prestige which must be located earlier than all yugas. The conception that living persons can lay down law, even in committees of public gatherings, is foreign to Hindu thought. The converse is obviously correct also. No one who proposes to persuade others can hope for success unless he shows that what he teaches is not his own but the common heritage of the race” (Derrett 1977, 180).


It did not harm that, as Derrett puts it, “at the time Sir Subramania was learning from Mrs. Besant spurious archaeology, fraudulent presentation of scripture, and practical organization” (178). Thus the problem is a western one in a twofold sense: the impact of western colonization both created the conditions under which a recourse to traditional systems became impossible and proffered a solution.


Derrett offers this summary of the conditions of its production: “The prime movers of the Śuddha Dharma Maṇḍala (which survived until as late as 1952) were principally Sir Subramania, now well over 70 years of age, and a Pandit K. T. Sreenivāsāchāriar of Madras. It is evident that the former lent his name and did most of the English writing (which often has a legal flavour) for the Maṇḍala’s series of publications. In more than one publication, a photograph of Sir Subramania appears, dressed as a sanyāsī, with his name and titles added, in order to give prestige to this bogus organization. The Sanātana Dharma Dipikā purports to be a purāṇa. It is full of mock puranic material. In Part I, chapter 1, sec. 181 we are told that the deity said ‘I will re-establish the organization named Śuddha dharma-maṇḍala: it is as old as time, excellent and makes its appearance with each kalpa (era).’ At sec. 289 we are told that all people will have one faith, one caste, in the Kali-yuga (the present age); at sec. 292–94 that all dharmas will be unified. At ch. 2, sec. 76–78 we are told that women will choose their own husbands. Social reform is evidently within the scope of the spurious literature which is now put forth. In 1918 the Yoga Dipikā of Bhagavān Nārāyaṇa was published in the same series, purporting to be a work of the deity, with a commentary by Haṃsa Yogi. In 1922 the Bhagavad Gītā was published also with the commentary of Haṃsa Yogi, but we are told at p. 7 that Haṃsa Yogi is not anyone’s name, but the title of an office. Evidently ‘Haṃsa Yogi’ was the Madras pandit himself. By this time the public is invited to supply money to subsidize the publication of commentaries on the entire Veda in the sense of the Ś. D. Maṇḍala, and so the enterprise which had started as an intellectual aberration takes on the form of a criminal conspiracy. The works published were bogus, but presented as if they were genuine. The author is also non-existent. Religious and social regeneration is the aim, and no doubt the conspirators were pioneers in some respects, hiding however under the garb of restorers of a lost literature” (Derrett 1977, 179).


For an account of these Gītās, see Adluri and Bagchee (2014, chapter 3). Georg von Simsons “Brahmanic Gītā” is also covered in this chapter; for Ježić’s “pseudo-German Gītā” (see next line of main text), see Adluri and Bagchee (2016a).


Could Schrader have known that the Haṁsayogin’s Gītā was a forgery? The evidence of the various forewords and introductions is unambiguous, but, as we have seen, Schrader chose to neglect these. At any rate, we know that other German scholars, reading the works of the Śuddha Dharma Maṇḍala around the same time, came to radically different conclusions: Wilhelm Printz, for instance, reviewed the publications of the Śuddha Dharma Maṇḍala and declared, “what must be understood under S. Dh. M. is adequately clear from [publication] nr. 5 [An Esoteric Organization in India], from the texts nr. 2 and 4 [Yoga Deepika of Bhagavan Narayana, and the Commentary of the Hamsa Yogi and Sanatana Dharma Deepika of Bhagavan Hamsa Yogi] as well as the English introductions: an esoteric organization with a transcendental hierarchy. The latter has its seat in the northern Himālaya (or, north of the Himālaya) in Badarī-vana” (Printz 1929, 257). Printz also noted that the founder of the organization was Sir Subramania Iyer, “K. C. I. E., L. L. D., sometime Acting Chief Justice of the High Court of Judicature, Madras, known as a long-term member of the Theosophical Society and [that] his book (nr. 5) is shot through with comparative references to it [that is, the Society]” (259). Nor was there any doubt about the identity of the Haṁsayogin: “About the author and commentator Haṃsayōgin as well we are sufficiently informed. Nr. 4 begins in Purāṇa style with Yōgis seeking the Ṛṣi Haṃsayōgin in Badarī-vana and requesting him to present the S. Dh. M. [to the world]. In the foreword to nr. 7 (and similarly in other places) S. Subrahmanya Iyer clarifies (7) that ‘The name Hamsa Yogi is not the proper name of any specific writer but the title of an office in the organization, filled, from time to time, by one or more of the most learned among the members thereof; these office-holders being charged with the duty of elucidating esoteric teachings contained in the chief Hindu Sacred books’” (259–60). Printz adds: “Finally, regarding the Bhagavadgītā we learn from this protagonist [that is, Sir Subramania] in the foreword to nr. 3 right at the beginning: ‘It is needless to say that none of the verses, in question, are of modern invention and all of them are to be found in the parent work, the Mahâbhârata, but in parts and contexts where they should not be. All that has been done is merely to transpose them to their proper place in this Geeta, thereby making it what it was in the Bhârata of 24,000 ślôkas, the predecessor of the epic as we have it now’” (260). Printz was not in any doubt about the mechanisms of textual production of the Dharma Maṇḍala; he notes that “one does not need to add a long criticism to the state of affairs described above. It is at any rate interesting to see how the Theosophic doctrine, which, under Annie Besant and C. W. Leadbeater, has taken over extensive elements from Tantric teachings of a south Indian nature, has had a reverse effect on a Vaiṣṇava circle, which, of course, has not had any great success’” (260–61). We can only speculate as to why Schrader so lightly overlooked this evidence. Perhaps the attractions of announcing a “new” Gītā edition and a previously unknown “classical” commentator to European audiences were too compelling, or perhaps he really did take the Haṁsayogin’s historical claims for statements of fact. Then, again, which we consider most probable, he may have been looking for evidence to make his own claims about having discovered a “Kashmir recension” of the Bhagavadgītā and that the Bhagavadgītā as we have it is not the “original” Gītā more probable. Curiously, in his last article on the Gītā (Schrader 1938, 62–68), he seems to have realized the deception (“We have, then, here an attempt to fabricate a Bhagavadgītā conforming to our Mahābhārata passage and consequently later than it”; 62), but he offers no retraction or, indeed, mention of his earlier studies, where he had taken the Haṁsayogin’s evidence seriously.


This is the essence of Biardeau’s objection to critical editions of the epics and purāṇas (Biardeau 1968, 115–23); a point that V. M. Bedekar in his response (Bedekar 1969, 210–28) unfortunately failed to grasp.


Compare Schrader’s “Both the commentators, Rāmakaṇṭha as well as Abhinavagupta, must have been completely ignorant of what is now the vulgate text of the Bhagavadgītā. They could not otherwise (not, at least, Rām, who mentions even variants such as anantaram for nirantarā XII, 12) have failed to adopt or at least mention some of the better readings of the vulgate, as, e.g., samaduḥkhasukhaḥ svasthaḥ in XIV, 24 instead of the nonsensical samaduḥkhasukhasvapnaḥ of the Kashmir text. Nor could they possibly have not even once mentioned the omission in the vulgate of one of those stanzas which are found only in the Kashmir text. In this respect it is worthy of note that Abhinavagupta at least is not afraid to state interpolations; also that Rāmakaṇṭha says (ad XI, 27) that he, on the strength of his having consulted many manuscripts of the Bhagavadgītā, does not agree with ‘somebody’ who metri causa omits the half-śloka beginning with tattejasā nihatā nūman” (Schrader 1930, 7–8; Schrader’s emphasis). And see also Schrader (1933, 42): “It can be considered out of the question that, had the Kashmirian Mahābhārata in Abhinavagupta’s time contained the Bhagavadgītā in the form of the vulgate, that he [that is, Abhinavagupta] would not have used this as the basis for his commentary or at least would have spoken out against it [that is, the vulgate text].”


Additionally, Schrader’s comments (see preceding note) bear rich comparison with the work of David and Nancy Reigle of the Eastern Tradition Research Institute. Commenting on the historical priority of the texts of the Śuddha Dharma Maṇḍala, they note, “The great sage Gauḍapāda near the beginning of his Māṇḍūkya-kārikā reviews the various theories of ‘creation,’ or more accurately, manifestation. He concludes by giving his own view that the svabhāva of the deva, i.e., brahman/ātman, is the cause of manifestation, in full agreement with the Wisdom Tradition. Yet this is not the doctrine of the Advaita Vedānta school known today, which was established by Śaṅkarācārya, his disciple’s disciple. The Śaṅkarācārya known today teaches that the cause of the world is omniscient, omnipotent brahman as God (īśvara). He goes on to specifically reject svabhāva, ‘inherent nature,’ as the cause of the world. So virtually all Vedāntins today accept God, īśvara, rather than svabhāva as the cause of the world. But the Praṇava-vāda, in agreement with Gauḍapāda, says that manifestation is the svabhāva or inherent nature of brahman, and that svabhāva is declared everywhere to be the cause of the world . . . . These few examples are sufficient, I think, to show that Praṇava-vāda, and presumably other texts brought ought by the Suddha Dharma Mandala, are in agreement with the hitherto secret teachings brought out in The Secret Doctrine. Moreover, the last example indicates that they represent a more ancient tradition than the one currently found in India, as verified by a still extant source, Gauḍapāda” (Reigle and Reigle 2007, 9–10). Evidently, the contrast between a “scientific domain [wissenschaftlichen Bereich]” and “esoterica” that some have made (see Hanneder 2008) is insufficient to distinguish these two approaches. German Indology is no less “esoteric” than Theosophy: both traditions seek to identify more “original,” “secret,” or “lost” texts; both traditions believe in an esoteric tradition of knowledge, lost or hidden from the Indians; and both traditions place emphasis on a lineage of teachers, experts, or initiates, thought to have a mysterious ability to uncannily restore these lost texts. It is merely a matter of which tradition one professes allegiance to. That is why in the section above we referred to Otto Schrader and Sir Subramania as the Hamburg yogin and the Haṁsayogin.


See Martin Luther, Luthers Sendschreiben vom Dolmetschen (1530).


Luther’s writings on translation offer eloquent testimony for his views. Under attack by Catholic theologians for his translation of Romans 3: 28, a passage in which he interpolated the word “alleine” (alone) even though the Greek does not contain it, he typically did not debate his rivals on the merits, but turned the debate into a question of authority. He says: “If your papist makes much useless fuss about the word sola, allein, tell him at once: Doctor Martin Luther will have it so, and says: Papist and donkey are the same thing; sic volo, sic jubeo, sit pro ratione voluntas. For we do not want to be pupils and followers of the Papists, but their masters and judges; for once, we also want to show off and spar with the donkey heads; and as Paul boasts before his mad saints, so I will boast before these donkeys. They are doctors? So am I. They are learned? So am I. They are preachers? So am I. They are theologians? So am I. They are disputators? So am I. They are philosophers? So am I. They are are dialecticians? So am I. They are lecturers? So am I. They write books? So do I. And I shall further boast: I can expound Psalms and Prophets; which they cannot. I can translate; which they cannot. If these donkeys demand further responses to their useless blubbering over sola, then say only so much: Luther wishes it be so and says, he is a doctor above all doctors in all of the papacy. Therefore the word allein shall remain in my New Testament, and though all pope-donkeys [Papstesel] should get furious and foolish, they shall not turn it out.”


This is something Malinar overlooks, when she claims that, by setting aside the native commentarial tradition, she avoids the problem of authority. See Angelika Malinar (1996 and 2007). For, the problem of authority is only sublated and not overcome: not only does she in her work defer to the authority of the German scholars Richard von Garbe, Hermann Jacobi, Rudolf Otto, Jakob Wilhelm Hauer, and Heinrich von Stietencron, and so on, but she also seeks to constitute herself as a new authority. Her work can at most be taken to be a rejection of traditional authority not a rejection of the principle of authority tout court.


See V. S. Sukthankar’s comments in Sukthankar 1933: “We in the present century are apt to get nervous and irritable over misprints and variae lectiones. But an ancient Indian scribe, redactor or even commentator, not to speak of the common reciter (pāṭhaka)—if I read aright Indian literary history—was not perturbed in the least by a little difference in wording or in sequence, especially if the variant did not give an appreciably better, or appreciably worse sense” (lv). We must remember that the fetishism of the variant is a characteristically modern event, introduced by the forgetting of the urgent existential task.


Pasquali (1952, 8), cited and translated in Timpanaro (2005, 58).



Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (, which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.

Authors’ Affiliations

Freie Universität Berlin
Hunter College


  1. Adluri, Vishwa, and Joydeep Bagchee. 2014. The Nay Science: A History of German Indology. New York: Oxford University Press.View ArticleGoogle Scholar
  2. Adluri, Vishwa and Joydeep Bagchee. 2016a. Paradigm Lost: The Application of the Historical-Critical Method to the Bhagavadgītā. International Journal of Hindu Studies 20(2).
  3. Adluri, Vishwa and Joydeep Bagchee. 2016b. Philology and Criticism: A Guide to Mahābhārata Textual Criticism. London: Anthem Publishing (In press).
  4. Bedekar, V.M. 1969. Principles of Mahābhārata Textual Criticism: The Need for a Restatement. Purāṇa 11(2): 210–28.Google Scholar
  5. Belvalkar, S.K. 1939a. The So-Called Kashmir Recension of the Bhagavadgītā. New Indian Antiquary 2(4): 211–51.
  6. Belvalkar, S.K. 1939b. The Bhagavadgītā ‘Riddle’ Unriddled. Annals of the Bhandarkar Oriental Research Institute 19(4): 335–48.
  7. Belvalkar, S.K. 1941. Śrīmad-Bhagavad-Gītā with the “Jñānakarmasamuccaya” Commentary of Ānanda[vardhana]. Poona: Bilvakuñja Publishing House.Google Scholar
  8. Belvalkar, S.K. 1943. A Fake (?) ‘Bhagavadgītā’ MS?. Journal of the Ganganath Jha Research Institute 1(1): 21–31.Google Scholar
  9. Belvalkar, S.K. 1947a. Introduction. In The Bhīṣmaparvan for the First Time Critically Edited, ed. S.K. Belvalkar. Poona: Bhandarkar Oriental Research Institute.
  10. Belvalkar, S.K. (ed.). 1947b. The Bhīṣmaparvan for the First Time Critically Edited. Poona: Bhandarkar Oriental Research Institute.
  11. Bhattacharjya, Sunil Kumar. 2013. The Original Bhagavadgītā. Complete with 745 Verses. Delhi: Parimal Publishers.Google Scholar
  12. Biardeau, Madeleine. 1968. Some More Considerations about Textual Criticism. Purāṇa 10(2): 115–23.Google Scholar
  13. Brockington, John L. 2015. The Bhagavadgītā: Text and Context. In The Bhagavad Gita: A New Translation, ed. Gavin Flood. New York: W. W. Norton & Company. Originally published in The Fruits of Our Desiring: An Inquiry into the Ethics of the Bhagavadgītā for Our Times, ed. Julius Lipner, 28–47. Calgary: Bayeux, 1996.
  14. Brockington, John L. 1998. The Sanskrit Epics. Leiden: Brill.Google Scholar
  15. Derrett, J. Duncan M. 1977. Essays in Classical and Modern Hindu Law, vol. 2: Consequences of the Intellectual Exchange with the Foreign Power. Leiden: Brill.
  16. Divanji, P.C. 1946. Critical Word Index to the Bhagavadgītā. Bombay: New Book Company.Google Scholar
  17. Edgerton, Franklin. 1925. The Bhagavad Gītā, or Song of the Blessed One: India’s Favorite Bible. Chicago and London: Open Court.Google Scholar
  18. Edgerton, Franklin. 1932. Review of The Kashmir Recension of the Bhagavad Gītā, by F. Otto Schrader. Journal of the American Oriental Society 52(1): 68–75.View ArticleGoogle Scholar
  19. Flood, Gavin (ed.). 2015. The Bhagavad Gita: A New Translation. New York: W. W. Norton & Company.Google Scholar
  20. Hanneder, Jürgen. 2008. Review of Bhagavad Gita. Der Gesang des Erhabenen, trans. by Michael von Brück. Marburger Forum 9(4).
  21. Iyer, Subrahmanya. 1917. Foreword. In Bhagavad-Geeta of Bhagavan Sri Kṛishna and the Geeṭarṭha-Sangraha of Maharshi Gôbila, ed. Pandit K.T. Sreenivasachariar. Madras: Shuddha Dharma Mahamandala.Google Scholar
  22. Iyer, Subramania. 1915. An Esoteric Organisation in India. Theosophist 36(2): 407, 499, 614, and 37(1): 196–206.
  23. Malinar, Angelika. 1996. Rājavidyā: Das königliche Wissen um Herrschaft und Verzicht. Studien zur Bhagavadgītā. Harrassowitz: Wiesbaden.Google Scholar
  24. Malinar, Angelika. 2007. The Bhagavadgītā: Doctrines and Contexts. Cambridge: Cambridge University Press.View ArticleGoogle Scholar
  25. Olivelle, Patrick. 1998. Unfaithful Transmitters: Philological Criticism and Critical Editions of the Upaniṣads. Journal of Indian Philosophy 26: 173–87.View ArticleGoogle Scholar
  26. Pasquali, Giorgio. 1952. Storia della tradizione e critica del testo. Florence: Felice le Monnier.Google Scholar
  27. Printz, Wilhelm. 1929. Eine sonderbare Bhagavadgītā-Ausgabe. Zeitschrift der Deutschen Morgenländischen Gesellschaft 83: 256–61.Google Scholar
  28. Reigle, David, and Nancy Reigle. 2007. Publications of the Suddha Dharma Mandala. Cotopaxi, Colorado: Eastern Tradition Research Institute.Google Scholar
  29. Reynolds, L.D., and N.G. Wilson. 2013. Scribes and Scholars: A Guide to the Transmission of Greek and Latin Literature, 4th ed. Oxford: Oxford University Press.Google Scholar
  30. Row, Vasudeva R. 1939. Foreword. In Srimad Bhagavad Gītā of Bhagavan Sri Krishna (26 Chapters & 745 Slokas), translated into English by R. Vasudeva Row. Mylapore, Madras: The Suddha Dharma Office.Google Scholar
  31. Sauppe, Hermann. 1872. Johann Matthias Gesner und Christian Gottlob Heyne. In Göttinger Professoren. Ein Beitrag zur deutschen Cultur und Literaturgeschichte in acht Vorträgen, 59–98. Gotha: Friedrich Andreas Porthes.
  32. Schrader, Friedrich Otto. 1910. Über Bhagavadgītā II, 46. Zeitschrift der Deutschen Morgenländischen Gesellschaft 64: 336–40.Google Scholar
  33. Schrader, Friedrich Otto. 1927. Neues über die Bhagavadgītā. In Aus Indiens Kultur. Festgabe zum 70. Geburtstag von Richard von Garbe, ed. Julius von Negelein, 171–83. Erlangen: Palm und Enke.Google Scholar
  34. Schrader, Friedrich Otto. 1928. Zur Überlieferung der Bhagavadgītā. Zeitschrift der Deutschen Morgenländischen Gesellschaft 82: 97.Google Scholar
  35. Schrader, Friedrich Otto. 1929a. The Sacrificial Wheel taught in the Bhagavadgītā. Indian Historical Quarterly 5(2): 173–81.
  36. Schrader, Friedrich Otto. 1929b. Concerning Bhagavadgītā III, 5. Indian Historical Quarterly 5(4): 790–91.
  37. Schrader, Friedrich Otto. 1930. The Kashmir Recension of the Bhagavadgītā. Beiträge zur indischen Sprachwissenschaft und Religionsgeschichte 3. Stuttgart: W. Kohlhammer.Google Scholar
  38. Schrader, Friedrich Otto. 1931. The Bhagavad-Gita in Ancient Kashmir. The Aryan Path 2(11):758–53.
  39. Schrader, Friedrich Otto. 1933. Rezensionen der Bhagavadgītā. In Festschrift Moriz Winternitz, ed. O. Stein and W. Gampert, 40–51. Leipzig: Otto Harrassowitz.Google Scholar
  40. Schrader, Friedrich Otto. 1934. Ancient Gītā Commentaries. Indian Historical Quarterly 10(2): 348–57.Google Scholar
  41. Schrader, Friedrich Otto. 1935a. On the Form of the Bhagavadgītā Contained in the Kashmirian Mahābhārata. Journal of the Royal Asiatic Society of Great Britain and Ireland 1: 146–49.
  42. Schrader, Friedrich Otto. 1935b. Review of Śrīmad Bhagavadgītā Edited with Numerous Variants, an Exhaustive Introduction, and Critical Notes, by S. N. Tadpatrikar. Orientalische Literaturzeitung 38(6): columns 386–88.
  43. Schrader, Friedrich Otto. 1936. Review of Die Urgestalt der Bhagavad-Gītā, Die Lehrtraktate der Bhagavad-Gītā, and Der Sang des Hehr-Erhabenen, by Rudolf Otto. Orientalistische Literaturzeitung 39(2): columns 107–17.
  44. Schrader, Friedrich Otto. 1938. An Implication of the Bhagavadgītā Riddle. New Indian Antiquary 1(1): 62–68.Google Scholar
  45. Schrader, Friedrich Otto. 1958. Review of The Bhagavadgītā. A Fresh Approach with Special Reference to Śaṅkarācārya’s Bhāṣya. Text with Śāṅkara Bhāṣya and an Introduction and Notes, by P. M. Modi. Zeitschrift der Deutschen Morgenländischen Gesellschaft 108: 222–23.Google Scholar
  46. Schrader, Friedrich Otto. 1982. Kleine Schriften. Mit Ergänzungen aus dem Nachlass, ed. Joachim Friedrich Sprockhoff. Wiesbaden: Fritz Steiner.
  47. Shastri, Jivaram Kalidas. 1941. Shri Bhagavad Gita (745 Shlokas, Bhojpatra MS). Gondal, Kathiawar: The Rasashala Aushadhashram.Google Scholar
  48. Speyer, J.S. 1902. Ein alter Fehler in der Überlieferung der Bhagavadgītā. Zeitschrift der Deutschen Morgenländischen Gesellschaft 56: 123–25.Google Scholar
  49. Sprockhoff, Joachim Friedrich. 1963a. Friedrich Otto Schrader zum Gedächtnis. Zeitschrift der Deutschen Morgenländischen Gesellschaft 113: 1–11.Google Scholar
  50. Sprockhoff, Joachim Friedrich. 1963b. Verzeichnis der Schriften von Friedrich Otto Schrader. Zeitschrift der Deutschen Morgenländischen Gesellschaft 113: 12–23.Google Scholar
  51. Sprockhoff, Joachim Friedrich. 1982a. Zum Geleit. In Kleine Schriften. Mit Ergänzungen aus dem Nachlass, ed. Joachim Friedrich Sprockhoff, v–ix. Wiesbaden: Fritz Steiner.
  52. Sprockhoff, Joachim Friedrich. 1982b. Bibliographie der Schriften von Friedrich Otto Schrader. Zugleich Inhaltsverzeichnis der vorliegenden Kleinen Schriften. In Kleine Schriften. Mit Ergänzungen aus dem Nachlass, ed. Joachim Friedrich Sprockhoff, xi–xxii. Wiesbaden: Fritz Steiner.
  53. Srinivasachariar, Pandit K.T. 1917. Bhagavaḍ-Geeṭa: translation of the preface of Pandit K. T. Srinivasachariar by T. B. Rajagopala Iyengar, B. A., (Hon.). In Bhagavad-Geeta of Bhagavan Sri Kṛishna and the Geeṭarṭha-Sangraha of Maharshi Gôbila, ed. Pandit K.T. Sreenivasachariar. Madras: Shuddha Dharma Mahamandala.Google Scholar
  54. Sukthankar, V.S. 1933. Prolegomena. In The Ādiparvan for the First Time Critically Edited, i–cx. Poona: Bhandarkar Oriental Research Institute.Google Scholar
  55. Tadpatrikar, S.N. 1937. The Computation of the Bhagavadgītā. Annals of the Bhandarkar Oriental Research Institute 18: 357–60.Google Scholar
  56. Timpanaro, Sebastiano. 2005. The Genesis of Lachmann’s Method, ed. and trans. Glenn W. Most. Chicago and London: University of Chicago Press.
  57. van Buitenen, J.A.B. 1965. A Contribution to the Critical Edition of the Bhagavadgītā. Journal of the American Oriental Society 85(1): 99–109.View ArticleGoogle Scholar
  58. Vedavyas, E. 1990. Ancient Bhagavad Gita: Original Text of 745 verses, with Critical Introduction. Hyderabad: United Social Cultural and Educational Foundation of India.Google Scholar


© Bagchee and Adluri. 2016