Paradigm Lost: The Application of the HistoricalCritical Method to the Bhagavad Gtå
Vishwa Adluri and Joydeep Bagchee
It may well prove to be the case that to study Western interpretations of
the Gtå is in fact tantamount to studying in microcosm Western reactions to something larger—Indian religion and culture in its entirety.
—Eric J. Sharpe, The Universal Gtå
The worst readers are those who go about it like marauding troops: they
remove what they can make use of, befoul and derange the rest, and
blaspheme the whole.
—Friedrich Nietzsche, Menschliches Allzumenschliches
This article makes a contribution to methods and theory in the study of
Indian traditions by asking a simple question: how useful is the historicalcritical method for studying Indian texts?1 We adopt the application of
this method to the Bhagavad Gtå as a test case to make wider claims
about the validity of the method. Specifically, we focus on the criteria2
that have been proposed for identifying “layers” in the text.3 The starting
point for this article is the view, widely shared by scholars (Stietencron
1996: 6–7; Hanneder 2001: 240; Malinar 2007: 174), that the historicalcritical method offers a superior—indeed, the sole scientific method (see
Hacker 1961: 489)—for the study of Indian texts. Yet how objective are the
results attained using this method? In continuation of the argument of The
Nay Science: A History of German Indology (Adluri and Bagchee 2014b),
in this article we look at six criteria proposed for identifying layers in the
International Journal of Hindu Studies 20, 2: 199–301
© 2016 Springer
DOI 10.1007/s11407-016-9187-4
200 / Vishwa Adluri and Joydeep Bagchee
Bhagavad Gtå: (1) “theism,” (2) names and epithets of K®‚~a, (3) repetition, (4) specific words or phrases, (5) formal markers, and (6) philosophical schools. We argue that none of these criteria lead to results that
are objective, noncircular, and capable of independent verification. In
conjunction with the materials presented in The Nay Science, where we
showed that the search for layers in the Bhagavad Gtå resulted from a
powerful myth regarding the Mahåbhårata’s origins in a heroic, war epic
(Adluri and Bagchee 2014b: 30–155), these arguments demonstrate the
limited usefulness of so-called historical-critical research into the Bhagavad
Gtå. In light of the problematic history of much of this research, not least
its supersessionism,5 we suggest scholars might more usefully look at the
Bhagavad Gtå anew, without the baggage of historicist dogmas.
The article makes a three-part argument. In the first part, we show that
the criteria proposed for identifying layers in the Bhagavad Gtå do not
work. In the second part, we show that the reason the problems with
scholars’ analyses of the Bhagavad Gtå were glossed over is because of
the myth of a heroic original epic. This myth supplied the missing historical “data” that filled in the gaps in the scholars’ analyses and made their
reconstructions appear plausible. However, the existence of this myth
cannot explain why these scholars insist that their interpretations are more
rigorous and more objective than those of the tradition. Rather, as we argue
in the third part of our argument, the latter is best explained in terms of
the millenarianism of the new science, for which overcoming native traditions of commentary was a necessary step in accomplishing the “education
of humanity.”6 It is to highlight the role of this millenarianism, deeply
embedded within the seemingly secular study of Indian traditions in
academia, that we now undertake a reconstruction of some recent Gtå
interpretations.
The Historical-Critical Method and the Bhagavad Gtå
The original application of the historical-critical method in biblical criticism was to make a distinction between two categories in the text: that
which “witnessed to salvation in Christ, and that which did not” (Rogerson
1985: 17). Taking their inspiration from this model of differential analysis,
in the latter half of the eighteenth century biblical critics such as W.M.L.
de Wette began to apply the method to work out a “history of Israelite
The Bhagavad Gtå and the Historical-Critical Method / 201
religion…radically at variance with the view implied in the Old Testament
itself” (Rogerson 1985: 29). The perceived advantages7 of this method
led scholars to extend it to the study of the “sacred texts” of other religions
as well, that is, to “writings of great age and complexity [that]…posed
the eternal philological problems of transmission, corruption, interpretation” (Turner 2014: 369). Among the texts considered natural candidates for such “philological” exploration were the Mahåbhårata and the
Bhagavad Gtå. Scholars considered these texts candidates not only
because of their antiquity and significance, but also because, as with the
Old Testament, they were key resources for working out a history of
Indian religion “radically at variance”8 with the view implied in the texts
themselves.9 As with biblical criticism, here also the application of the
method relied principally upon the perception of changes to the text, claims
regarding the relative authenticity or antiquity of different doctrines, and
the identification of alleged contradictions in the text (Levenson 1993: 2–
4).10 A spate of allegedly critical reconstructions of the text11 led in the
twentieth century to the view of the Bhagavad Gtå as a composite text
(see Figure 4 on pages 259–61). In spite of the subjective and a priori
nature of these reconstructions (see Adluri and Bagchee 2014b: 156–
313), the search for layers in the Bhagavad Gtå continued well into the
twentieth century.12 We even find a multiplication of the number of
potential layers in some of the latest accounts (Malinar 2007; see Adluri
2010: 105, providing a useful summary), though many of them were, in
fact, derivative (see, for example, Jei 2009b: 32, 32n3, acknowledging
the debt to Charpentier 1930). In light of these continuing efforts, and
especially given the fact that many of the accounts were contradictory
and even circular, it appears opportune to pass review on some of the
more significant recent attempts. We shall focus here on a tetrad of articles:
P.L. Bhargava’s 1977 and 1979 articles “Additions and Interpolations in
the Bhagavadgtå” and “Names and Epithets of K®‚~a in the Bhagavadgtå,” and Mislav Jei’s 1979 and 1986 articles “The First Yoga Layer
in the Bhagavadgtå” and “Textual Layers of the Bhagavadgtå as Traces
of Indian Cultural History.”13
Additions and Interpolations in the Bhagavad Gtå
P.L. Bhargava’s first article was motivated by two concurrent interests.
202 / Vishwa Adluri and Joydeep Bagchee
First, in opposition to the traditional view of the text (which he called the
“orthodox Hindu” view), he sought to show, on historical-critical grounds,
that K®‚~a was originally a historical character. In his words, “although
K®‚~a, son of Vasudeva and Devak, appears to have been a historical
person who was in all probability identical with K®‚~a, son of Devak,
mentioned in the Chåndogya Upani‚ad,” “no sober historian” would
“concede that the Bhagavadgtå contains the actual words spoken by
K®‚~a to Arjuna on the battlefield of Kuruk‚etra” (Bhargava 1977: 357).14
Second, he sought to show that the Bhagavad Gtå was “the composition of a poet who wanted to epitomize the teachings attributed to K®‚~a”
(357). Thus, he claimed that “after a patient and critical study of this
work,” he had arrived at “the conclusion that a considerable part of the
Bhagavadgtå in its present form consists of additions and interpolations,
since this part is inconsistent with the rest of the work which must have
been the original Bhagavadgtå” (357). In defense of this hypothesis, he
invoked the thesis that “each one of the great teachers of the world has
suffered deification at the hands of his followers” (357). He argued or,
perhaps better, declared that “K®‚~a too could not escape this fate. It is
certain that K®‚~a was originally a human teacher who was later deified”
(357).15 The Bhagavad Gtå, he averred, consisted “of two clear-cut parts,
one of which, barring a few sporadic verses which can be shown to be
interpolated, regards K®‚~a as a human teacher, while in the other part
K®‚~a claims to be the omnipotent, omniscient and omnipresent God in
almost every verse” (357). Notably, whereas “in one part Arjuna always
addresses K®‚~a by his name or one of his well-known epithets as Acyuta,
Keçava, Govinda, Madhus¨dana, Janårdana, Mådhava, Vår‚~eya, H®‚keça
and Keçini‚¨dana,” in the other part “he addresses him as Puru‚ottama,
Bh¨tabhåvana, Bh¨teça, Devadeva, Jagatpati, Parameçvara, Viçveçvara,
Ananta, Deveça, and Jagannivasa—all names of the Supreme Being” (357).
Bhargava, however, did not pursue this idea further in his first article;
rather, he focused on an a priori method of analysis closely related with
the German Indologist Richard von Garbe’s method for identifying (and
removing) so-called pantheistic interpolations from the text.16 Thus, he
argued that there were no interpolations in the first chapter, for, “since
the first chapter…describes merely the despondency of Arjuna, there is
neither the necessity nor the room for interpolation” (Bhargava 1977:
358).17 In contrast, he argued that there was one interpolation in the second
The Bhagavad Gtå and the Historical-Critical Method / 203
chapter: “The only verse in which K®‚~a appears in the role of God is
v. 61 and [hence] it is clearly interpolated” (358). His argument for this
verse being interpolated was as follows: “V. 61 describes the result of
restraining senses with K®‚~a as the supreme goal, which seems to be a
paraphrase of v. 68 which describes the same thing…, but without any
reference to K®‚~a’s divinity. Verse 61 is, therefore, not only out of place,
but also redundant” (358). Bhargava also identified two interpolations in
the third chapter. According to him, the first of these comprised verses
22–24, as “these verses in which K®‚~a is said to be engaged in action for
the good of the world” produced “a semblance of unity” with the preceding verse “wherein it is said that other men do whatever a great man
doeth” (358). However, he argued, that “the continuity of thought is in
no way disturbed” by omitting the three verses (358).18 The second interpolation comprised verses 30–32, for they were “clearly irrelevant” (358).
According to Bhargava, whereas verses 29 and 33 described the role of
nature (prak®ti) with respect to the ignorant and the knowledgeable, these
three verses interrupted the train of thought with a reference to K®‚~a.
From this, he concluded that they were “undoubtedly interpolated” (358).
In a similar manner, he also identified several interpolations in chapters 4
(1),19 5 (2),20 and 6 (3),21 while claiming that chapters 7 to 12 as a whole
were interpolated.22 Bhargava also identified three interpolations each in
chapters 1323 and 14. According to him, chapter 14 contained three interpolations from verses 2–4, 19, and 26–27. His argument was as follows:
whereas the first verse was “introductory,” the “subject matter” of the
chapter only began with verse 5 (359). Thus, he concluded, verses 2–4
must have been interpolated “only with the object of showing that K®‚~a
is greater even than the Brahma of the Upani‚ads” (359). In his view,
“these words could scarcely have been uttered by K®‚~a of the Chåndogya
Upani‚ad fame” (359). Likewise, he considered verse 19 to be an interpolation as it laid “emphasis on K®‚~a’s divinity” (359). In his view, this
made it “clearly an interpolation” (359). Finally, he argued that verses
22–25, in which K®‚~a describes the marks of those who have crossed
over the qualities (gu~a) and concludes with the words “ ‘he is said to
have crossed over the qualities’,” originally constituted the final verses of
the chapter (359). However, according to him, “someone…who wanted to
include devotion to K®‚~a also among the marks of a person who has
crossed over the three qualities has appended vv. 26 and 27” (359).
204 / Vishwa Adluri and Joydeep Bagchee
Bhargava also identified two interpolations in chapter 15,24 one each in
chapters 16 and 17, and three in chapter 18. His exact argument for considering verses 17–20 of chapter 16 an interpolation was as follows: according to him, “these four verses add[ed] nothing new to the subject matter
of this chapter”; they were “meant only to emphasize the divinity of
K®‚~a” (Bhargava 1977: 360). Further, he argued that verse 21 “clearly
seems to have been originally in continuation of v. 16” (360), and hence
the intervening verses must have been inserted later. Along similar lines,
verses 5–6 of chapter 17 must be interpolations. They were “clearly out
of context”; K®‚~a, in this chapter, intended to describe “the three types
of worship, the three types of food, the three types of sacrifice, the three
types of austerities and the three types of gifts” (360), but verses 5–6 did
not correspond to any of these topics. Hence, he concluded they were
“clearly interpolated” (360). The eighteenth chapter posed more complex
challenges.25 According to Bhargava, it contained four interpolations
(18.54–58, 18.64–71, and 18.77). Regarding the first of these, he claimed
that there was a direct contradiction between verses 53 and 54–58. Further,
verse 58 appeared to repeat an idea found in 53 (casting aside egoism),
from which he concluded that this verse had been inserted with the intent
of restoring the context with verse 59. Likewise, he claimed that there
was a direct contradiction between verses 61–62 and verses 64–71. Thus,
whereas in the former, K®‚~a speaks of God “in the third person” and does
not offer “the slightest indication that K®‚~a is identical with Him,” in the
latter, “he does not ask him to flee to God but using the first person he
exhorts him [Arjuna]” to worship him (360). In his view, this made the
two sets of verses “so inconsistent…that it is really amazing that the contradiction has so far escaped the notice of scolars [sic]” (360). Indeed, they
were “so contradictory to and inconsistent with each other that no comment
is called for” (361). Finally, he also argued that verse 65 was “a repetition
of v. 34 of chap. IX,” and from this he concluded “that one and the same
person…[was] responsible for all the interpolations in the Bhagavadgtå”
(361). Bhargava also identified two final interpolations in the chapter:
the verses after 71, which were “in the manner of a phalaçruti of the
Bhagavadgtå” and hence could not “but be interpolated,” and verse 77,
which had “again been interpolated to give the impression that Sañjaya
was familiar not merely with the conversation between K®‚~a and Arjuna
but also with the universal form of K®‚~a described in chap. XI” (361).26
The Bhagavad Gtå and the Historical-Critical Method / 205
An A Priori Scheme for Making Claims about the Bhagavad Gtå
However, in spite of Bhargava’s claim of having undertaken a critical
study of the poem, there was nothing critical about his approach. Indeed,
it was not so much a method as an a priori scheme for making claims
about the text. For instance, the basic assumption of his approach was
that the human, historical level had to be primary and K®‚~a’s status in
the poem (as superhuman figure, as object of reverence, as universal
being, and as God) could only be explained in terms of euhemerization.
This theory, in fact, had a long history.27 In the context of the Bhagavad
Gtå, it was first advanced by Edward Washburn Hopkins (1895: 389;
but Hopkins was almost definitely drawing on Holtzmann (1892 and
1893) and canonized by Richard von Garbe (1905; and see also Jacobi
1921: 717).28 Garbe had claimed that an “impartial historical observation
of the sources shows that K®‚~a in the most ancient period was a man
and later—in a continual development—became a demigod, god, and
All-Being” (1905: 31). He criticized the view of K®‚~a as an incarnation
of God as an “inversion of the real relationship” and a “myth of transformation” and argued instead that “in our case, euhemerism is the correct
view” (31). However, Garbe’s “impartial historical observation” was not
so very impartial after all, since he subscribed to a relatively orthodox
Protestant view of religion (see Malinar 2003: 121), and, furthermore,
was interested in making certain claims about the nature of the religious
corruption of Hinduism (see Adluri and Bagchee 2014b: 189–91, 198–
200, and 391n126). In this project of placing the historical, worldly subject
at the origin of religious developments, the pseudocritical method of
identifying a history of ideas in the text came to his rescue. He could now
eliminate the philosophical and ontological sections of the poem, while
insisting that he was interested only in reconstructing the original.29
If we now look more closely at Bhargava’s approach, we find that the
method does not actually tell us anything objective about the poem itself.
Rather, it proceeds according to a set of principles that we may list as
follows:
(1) Criterion of preference of the nontheistic formulation: If one verse
says something that another already says, but says it without reference to
K®‚~a’s divinity, then the former verse is to be preferred.
(2) Criterion of preference of the nontheistic sequence: If by removing
206 / Vishwa Adluri and Joydeep Bagchee
one or more verses that refer to K®‚~a’s divinity, the continuity of thought
is not disturbed, then those verses are an interpolation.
(3) Criterion of preference of the nontheistic level: If in a sequence of
verses, one verse refers to K®‚~a, then this is an interpolation because
reference to K®‚~a is a priori ruled out as part of the discussion or intent,
and hence any such reference represents a divergence from a level posited
as more fundamental for the poem.
(4) Attribution of intent: The scholar attributes specific motivations to
the hypothetical interpolator (for example, the desire to enhance K®‚~a’s
status), and this becomes a further criterion for him to identify interpolations.
(5) Attribution of consistency: The scholar makes assumptions about
K®‚~a’s character and argues that sections of the poem in which K®‚~a
displays traits contrary to the scholar’s assumptions about K®‚~a’s basic
character traits must be interpolations.
(6) Criterion of nontheistic reference: Between two possible references
of a verse, the reference to a nontheistic passage (that is, one making no
reference to K®‚~a’s divinity) is to be preferred.
(7) Criterion of guilt by association or guilt by position: This is a widening of application of the criteria; it will permit Bhargava (and others)
to make much cleaner excisions of verses to avoid the kind of grammatical
inconsistencies their method would otherwise leave in the text.
(8) Criterion of grammatical inflexibility: This is Garbe’s criterion;
K®‚~a cannot refer to himself in the third person.
(9) Criterion of semantic inflexibility: It is not possible to say similar
things in two ways to introduce modulations or nuances of meaning.
(10) Criterion of literary inflexibility: If in one place K®‚~a refers to
God in the third person and says Arjuna should flee to him and in another
speaks of himself in the first person and says he should be his devotee,
then this is a contradiction, for fleeing is not the same as being devoted
(it would be interesting to know what kinds of tolerances in the meanings
of words these critics permit).30
(11) Criterion of (postulated) epic qualities: Only those parts of the
Bhagavad Gtå that display epic qualities or themes are original; hence
elements such as the concluding phalaçruti are interpolations.
When spelled out in this way, it is clear that the principles are not in
fact rigorous criteria, but are simply part of an a priori scheme for making
The Bhagavad Gtå and the Historical-Critical Method / 207
claims about the Bhagavad Gtå. In effect, Bhargava has gone through
the text, identified all the verses that refer to K®‚~a as a God or confer the
attributes of divinity upon him, and declared these verses to be interpolations in keeping with his a priori theory that K®‚~a cannot be God. Indeed,
the principles tell us nothing objective about the text, but merely how
Bhargava will react to a given verse in the Gtå.
This can be confirmed by a simple experiment. If we group the verses
excised by their respective criterion, we see that the reason certain verses
are excluded is the same each time: all of the objectionable verses contain
first-person references by K®‚~a to himself (for example, words such as
matparaª, måm, and aham). Thus, verses 2.61, 5.29, 6.13–15, 6.30–31,
6.47, 13.3, 13.11, 13.19, 14.19, 14.26–27, 15.6–15, 15.18–19, 16.17–20,
and 17.5–6 contain one or more of the following words: matparaª, måm,
mat cittaª, and matsaμsthåm; verses 3.22–24, 4.1–15, 13.11, 13.19, 14.19,
14.26–27, 15.6–15, 15.18–19, 16.17–20, and 18.54–58 likewise contain
one or more of the following: me, mama, aham, mayå, and bhakto ’si me;
and so on. The following table presents a full list of Bhargava’s excisions,
along with their respective criterion and the reason for excision:
Criterion
Verses
Reason
1
2.61, 5.29, 6.13–15, 6.30–
31, 6.47, 13.3, 13.11,
13.19, 14.19, 14.26–27,
15.6–15, 15.18–19,
16.17–20, 17.5–6
matparaª, måm, mat cittaª,
matsaμsthåm
2
3.22–24, 4.1–15, 13.11,
13.19, 14.19, 14.26–27,
15.6–15, 15.18–19,
16.17–20, 18.54–58
me, mama, aham, mayå,
bhakto ’si me
3
3.30–32, 4.1–15, 13.11,
13.19, 14.19, 14.26–27,
15.6–15, 15.18–19,
16.17–20
mayi, me
4
14.2–4, 14.19, 14.26–27,
15.18–19, 16.17–20, 18.77
mama, aham
208 / Vishwa Adluri and Joydeep Bagchee
5
14.2–4
mama, aham
6
16.17–20, 18.54–58
måm, aham, mat bhaktim,
bhaktyå måm, mat
vyapåçrayaª, mat prasådåd,
mat paraª, mat cittaª
7
17.5–6
måm
8
18.64–66
me paramaμ vacaª, i‚†o ’si me,
manmanå bhava madbhakto
madyåj måμ namaskuru|
måm evai‚yasi satyaμ te
pratijåne priyo ’si me\
9
18.64–66
me paramaμ vacaª, i‚†o ’si me,
manmanå bhava madbhakto
madyåj måμ namaskuru|
måm evai‚yasi satyaμ te
pratijåne priyo ’si me\
10
18.65–66
manmanå bhava madbhakto
madyåj måμ namaskuru|
måm evai‚yasi yuktvaivam
åtmånaμ matparåya~aª\
11
18.67–71
måm, mayi, mat bhakte‚u, me,
aham
Further, this reconstruction does not bear out Bhargava’s thesis that the
Bhagavad Gtå was originally a poem referring to a human and historical
character called K®‚~a, for, even if we were to accept the classification
into two types of passages (that is, verses with first-person references by
K®‚~a and verses without first-person references by K®‚~a), this still does
not establish the priority of the latter. Indeed, no historical sequence is
implied at all, since there is no reason why K®‚~a or the poet of the Gtå
could not have shifted between first-person and second-person statements.
Even the individual arguments do not support such a conclusion. For
instance, although Bhargava excised verse 2.61 on the grounds that it was
“redundant” as it merely repeated something found in 2.68 with the minor
difference that, in contrast to the latter, it referred to K®‚~a’s divinity, his
The Bhagavad Gtå and the Historical-Critical Method / 209
argument does not hold. It is unclear how verse 2.61 can be “redundant,”
since—by his own admission—it expresses a different idea than verse
2.68. Verse 2.61 would only be redundant if the thought or the content
of the two verses were identical. Thus, what Bhargava really means is
not that 2.61 is formally redundant, but that it expresses a redundant
notion, namely, the notion of K®‚~a’s divinity. Since, from the perspective of his theory, this notion is redundant not only here but also everywhere else it occurs, his argument does not actually prove anything. It
is, formally speaking, not even an argument. It merely restates his initial
hypothesis. In a similar way, his argument for excising verses 3.22–24,
namely, that “if we omit these three verses the continuity of thought is
in no way disturbed” (Bhargava 1977: 358), is not a valid argument.
Theoretically, the Bhagavad Gtå could be reduced to any two wellconnected verses. The argument places unnecessary restrictions upon the
poet’s freedom to change themes or to introduce nuance; it also begs the
question. Even if there were a contrast between the two sets of verses, we
would still have to show that the former (that is, the nontheistic verses)
were original. This is something Bhargava does not address at all. Finally,
we also cannot assume that a shift in the discussion from the level of
impersonal prak®ti to personal K®‚~a in verses 3.30–32 implies an interpolation. This would only hold if we were to tacitly assume that any such
shift in level must be evidence of interpolation, but this would be, once
again, to beg the question.
We may now list the problems with Bhargava’s principles as follows:
(1) Criterion of preference of the nontheistic formulation: This is not a
critical canon because it tells us nothing about the poem, but says something about the scholar instead.
(2) Criterion of preference of the nontheistic sequence: This is also not
a critical canon because it does not tell us why this pair of well-connected
verses (that is, 3.21 and 3.25) should be preferred over the other pair (that
is, 3.22–24), as it is equally conceivable that 3.22–24 formed the original
Gtå (along with other verses) and that 3.21 and 3.25, and so on, interrupted this continuity.31
(3) Criterion of preference of the nontheistic level: The distinction is
not something we come across in the poem (for example, when we
notice that a writer is using two levels such as a historical and a fictional
level or a historical and a symbolic level and is moving back and forth
210 / Vishwa Adluri and Joydeep Bagchee
between the two). Rather, the distinction is posited in advance and posited
in a way that makes it absolute. It is, in fact, nothing but the original
thesis of the scholar (that is, that there are two levels or two layers in the
Gtå).
(4) Attribution of intent: From the fact that a passage x appears to
enhance K®‚~a’s status, we cannot conclude either that the passage was
inserted with this motive or that it was inserted by this person (that is,
someone wishing to enhance K®‚~a’s status). The original author may
have wished to enhance K®‚~a’s status in this segment, or he may not
have wished to do so and still composed a passage one of whose effects
was that it enhanced K®‚~a’s status. The fallacy is a form of the genetic
fallacy.
(5) Attribution of consistency: This criterion entails two fallacies. First,
it assumes that the K®‚~a of the Bhagavad Gtå is a real person and that
this person is identical with the K®‚~a of the Chåndogya Upani‚ad.
Second, it also assumes that the K®‚~a of the Chåndogya Upani‚ad is a
real person and, furthermore, that the K®‚~a of the Chåndogya Upani‚ad
always acted in a perfectly consistent way.
(6) Criterion of nontheistic reference: There is no justification for taking
a verse to refer to a nontheistic antecedent rather than a theistic antecedent;
it is not a critical criterion.
(7) Criterion of guilt by association or guilt by position: There is nothing
specifically theistic about 17.5, but Bhargava takes it along with 17.6
because it constitutes a unit with it; it is not a critical criterion.
(8) Criterion of grammatical inflexibility: This criterion places undue
restrictions on an author’s freedom to use language.
(9) Criterion of semantic inflexibility: This criterion places undue
restrictions on an author’s freedom to use language.
(10) Criterion of literary inflexibility: This criterion places undue
restrictions on an author’s freedom to use language.
(11) Criterion of (postulated) epic qualities: The unstated major premise
of the final argument is that no work of the nontheistic nature of the
Bhagavad Gtå can have a phalaçruti. Even if we accept the claim that
the Gtå is nontheistic, the major premise is not necessarily true. In fact,
the further premise supporting it is that the Bhagavad Gtå being an epic
poem, a phalaçruti would be foreign to its character. This is not an argument, but the hypothesis stated in a different form.
The Bhagavad Gtå and the Historical-Critical Method / 211
Names and Epithets of K®‚~a in the Bhagavad Gtå
In a second article, published two years later, Bhargava again attempted
to demonstrate that K®‚~a in the Bhagavad Gtå was a historical character
who later underwent a process of divinization. Here he took up an argument originally introduced in the 1977 article, namely, that the names
and epithets of K®‚~a could be a guide to the stage of composition of the
relevant portion of the text. Bhargava distinguished between two stages
of the text: a first stage, in which K®‚~a was only a mortal individual, a
warrior, or a chieftain; and a second stage, in which K®‚~a was elevated
to a God.
In his view, this transformation reflected a wider change in Indian
society from a heroic K‚atriya culture to a ritualistic Bråhma~ic culture.
As with the earlier article, Bhargava proceeded via a simple method of
analysis. He divided the names and epithets of K®‚~a in the Bhagavad
Gtå into two groups—heroic names, pointing to the earliest layer, and
divine names, pointing to a subsequent theistic revision of the original
poem—and argued that the type of names in evidence were an indication
of whether the verse belonged to the earlier, heroic text or the later, theistic
text. Each of these groups consisted of further subgroups. Thus, in the
first group, he included five types of names: (1) names,32 (2) patronymics,33 (3) names referring to K®‚~a’s qualities as a warrior,34 (4) names
referring to his moral qualities,35 and (5) names descriptive of K®‚~a’s
qualities rather than epithets.36 In the second group, Bhargava included
seven types of names: (1) compounds of puru‚a (spirit),37 (2) compounds
of para or parama (supreme),38 (3) compounds of deva (God),39 (4) compounds of bh¨ta (being),40 (5) compounds of jagat (world),41 (6) compounds of viçva (all),42 and (7) miscellaneous terms.43 From the distribution of these terms in the text, he concluded that the Bhagavad Gtå
must originally have consisted of only chapters 1–6 and 12–18.44 He
argued that this portion of the poem must have been “composed at a time
when K®‚~a was still regarded as a human teacher” (Bhargava 1979: 96).
In contrast, the remaining six chapters (that is, 7–13) and “a few sporadic
verses” must have been “added after his deification” (96). Bhargava also
placed the period of the composition of the “original Bhagavadgtå [that
is, the Bhagavad Gtå comprising chapters 1–6 and 12–18]” in the fifth
century BC, since (he argued) “K®‚~a was already deified in the fourth
212 / Vishwa Adluri and Joydeep Bagchee
century B.C. according to the testimony of Megasthenes” (96).
Since Bhargava’s argument for a two-stage composition of the Gtå
turns on the distribution of K®‚~a’s names in the text, let us first look at
the occurrences of the names identified by him. The following table lists
the names against the eighteen chapters of the Bhagagavad Gtå:
The Bhagavad Gtå and the Historical-Critical Method / 213
Contrary to Bhargava’s claims, however, the table does not bear out his
thesis of a distinction between heroic and theistic parts of the poem. Only
two of the central six chapters (that is, chapters 10 and 11) actually have
a preponderance of divine names or epithets for K®‚~a. The remaining
chapters attributed to the theistic layer (that is, chapters 7–9 and 12) do
not stand out in any way from the nontheistic parts. Chapters 7, 9, and 12
do not feature any terms for K®‚~a as a divinity, while chapter 8 has only
a single term, which could be explained either as a statistical anomaly45
or as the result of a later interpolation (8.1 being the interpolated verse).
Further, what little plausibility the argument has rests on a redefinition of
K®‚~a’s names to make them fit their proposed categories. For instance,
Bhargava defines the name K®‚~a as “the name of the propounder of the
philosophy of the Bhagavadgtå” (1979: 93), even though this is precisely
the point in question: does the name K®‚~a in the Bhagavad Gtå refer
to a historical character or to a God? Likewise, he defines Keçava and
Govinda as “appellations of K®‚~a [that] had practically assumed the
force of names” (93). But these terms are also widely used and recognized
as divine names of K®‚~a. The argument is not helped by redefining the
names under debate in question-beginning ways. In fact, a look at the
remaining names Bhargava lists as “heroic” names of K®‚~a shows that
he has not appreciated the theological significance of any of them. Thus,
Mådhava is defined as “after an ancestor…named Madhu”; Vår‚~eya as
“after…[an] ancestor named V®‚~i”; Våsudeva as “after his father Vasudeva”; Janårdana as “destroyer of evil persons”; Madhus¨dana as “slayer
of Madhu”; Keçini‚¨dana as “slayer of Keçin”; H®‚keça as “controller of
senses”; Acyuta as “not deviating from righteousness or not yielding to
pasions [sic]”; Yogin, Yogeçvara, and Mahåyogeçvara as “since K®‚~a
was an adept in yoga”; Aris¨dana as “slayer of enemies” (further glossed
as “like any other great warrior”); Mahåbåhu as “indicative of physical
excellence and strength”; and Prabhu as “Master or Lord,” “because of
his [Arjuna’s] reverence for K®‚~a as a great teacher” (93, 94). However,
many of these terms have well-defined theological meanings. For instance,
Madhus¨dana is most often understood to refer to K®‚~a’s slaying of the
demon Madhu (see Mahåbhårata 3.194.8–30 and 12.335.16–64), which
marks it out as a theological term. Bhargava, however, interprets the
reference to be to “a wicked person bearing the same name as one of the
ancestors of K®‚~a” (94). No source is cited. Further, the interpretation
214 / Vishwa Adluri and Joydeep Bagchee
rests on interpreting K®‚~a as the name of a historical person with a
human genealogy, even though this is the very point under contention.
Terms such as Yogeçvara and Mahåyogeçvara (Lord of Yoga and Great
Lord of Yoga) are reinterpreted as references to K®‚~a’s skill in yoga,
thus denuding them of their theological significance and reducing çvara
to a term of hyperbolic praise. Indeed, the very meaning of yoga, that is,
what it means to be a master of yoga, is left out of consideration here.
We could scarcely imagine a more thoroughgoing banalization of the text
than this.
In pursuit of his thesis of the divinization of K®‚~a and because he does
not wish to acknowledge the divinity of K®‚~a, Bhargava has in fact
finessed the data. Rather than providing an objective analysis of the names
of K®‚~a, he has ensured that even names with as straightforward a
pedigree as Prabhu (Lord), Acyuta (the Unfallen One), and Janårdana
(explained by Çaμkara as either “One who inflicts suffering on evil men”
or as “He to whom all devotees pray for worldly success and liberation”)
are now redefined so as to fall into the convenient categories readied for
them. Even if we grant Bhargava his contention regarding the other names
and move just these seven (that is, Yogeçvara, Mahåyogeçvara, Prabhu,
Acyuta, Janårdana, Madhus¨dana, and Keçini‚¨dana) into their respective columns, we obtain a radically different picture of the text, as is indicated by the corrected figures in parentheses. Chapter 1, supposedly the
oldest chapter of the heroic layer, now contains five theistic epithets. In
contrast, chapters 8 and 14, supposedly part of the theistic layer of the
poem, contain just two. Further, chapter 2 contains as many theistic epithets (2) as chapters 8 and 14, while chapter 18 contains more theistic
epithets (4) than any of the so-called theistic chapters except two (chapters
10 and 11). Chapters 3 and 6 at any rate contain more theistic epithets (1)
than chapters 7, 9, and 12.
Bhargava’s article thus cannot be cited as evidence of a human, historical K®‚~a or as evidence of a two-stage or multistage composition of
the Gtå. It is not a work of objective scholarship. He has not made the
case that the distribution of K®‚~a’s names in the poem reveals something
about the poem’s history. Indeed, even if his claims about the distribution
of K®‚~a’s names were true, this would still not prove that chapters 1–6
and 13–18 were original to the poem, whereas chapters 7–12 were added
later, for we could have a real distinction between the use of the names
The Bhagavad Gtå and the Historical-Critical Method / 215
and the central chapters could still be the oldest part of the poem. Without
evidence for the existence of the historical character and without evidence
of the divinization of this historical character, we simply have no reason
to posit that K®‚~a is anything other than what he appears as in the text: a
literary figure, who is presented as an advisor, a friend, and a guru and,
ultimately, as the Supreme Being.46 The only reason Bhargava thinks to
separate out a historical, human prototype from the divinity is because he
subscribes to a theory of the euhemerization of K®‚~a that first became
popular around the late nineteenth century.47 It is this theory rather than
any evidence that sustains his analysis. Both Bhargava’s articles are
more correctly located within a tradition of Christian exceptionalism that
finds it easier to accept disingenuous and self-serving arguments than to
acknowledge the claims of a rival God48 and whose continuing incarnation in German academic scholarship is Indology.49
Textual Layers of the Bhagavad Gtå as Traces of Indian
Cultural History
In 1979, in explicit continuation of the work of Wilhelm von Humboldt,
Adolf Holtzmann, Jr., Richard von Garbe, Hermann Oldenberg, and
Rudolf Otto (Jei 1986: 628–29), the Croatian scholar Mislav Jei
proposed that the presence of repetitions in the Bhagavad Gtå could be
used as a criterion to identify the different layers in the text. According to
Jei, when he started “reading the Bhagavadgtå” “without any intention
to search for the layers,…[he] became aware of repetitions in the text of
both meaning and expression” (629). He further noted that the repetitions
in the text appeared to fall into one of two classes. On the one hand, there
were the “continuity repetitions,” which “display[ed] definite relationships to one another: complementary or opposing relationships, or relationships such as exist between question and answer, for example, which
serve to develop a given subject” (629). On the other, there were the
“duplication repetitions,” which “repeat[ed] something said elsewhere,
while giving it a different connotation” (629). Arguing that this difference in character was significant for a reconstruction of the text, Jei
wrote that he “assumed that the continuity repetitions betrayed those parts
belonging to the same sequence of text, and that the duplication repetitions, on the other hand, indicated parts belonging to different sequences”
216 / Vishwa Adluri and Joydeep Bagchee
(629). On the basis of this assumption, he proposed a structure for the
Bhagavad Gtå:
Figure 1. Jei’s simplified provisionary scheme for layers in the
Bhagavad Gtå50
In the first phase of its life, Jei claimed, the Bhagavad Gtå existed
as a simple heroic dialogue between Arjuna and K®‚~a—the K®‚~årjunasaμvåda as he called it. This poetic part of the poem would have been a
part of the original heroic epic and would likely have corresponded to
verses 1.1–2.4, 2.9–10, and 2.31–37 of the present text. Its main function
would have been to convince Arjuna to return to battle. Accordingly, it
The Bhagavad Gtå and the Historical-Critical Method / 217
would have been composed in a simple, heroic style and contained only
injunctions relevant to the warrior’s code. Alongside and possibly in
competition with this heroic poem, however, there also existed a hymn
composed in tri‚†ubh meter. This hymn had as its explicit aim the glorification of K®‚~a. Jei proposed that the “hymn evidently presuppose[d]
the epic part of the poem, but since it contain[ed] a proper description of
Arjuna’s dilemma, and not only K®‚~a’s answer to it, it must have been
intended as a separate composition, and not as a layer added to the epic
episode” (1986: 633). At some stage, an unnamed composer added a
Såμkhya layer corresponding to 2.11–30 and 2.38 of the present text.
This Såμkhya layer, according to Jei, offered a different answer in
response to Arjuna’s dilemma: instead of appealing to the warrior code,
it introduced the theme of not grieving either for the living or the dead
(açocyån anvaçocas tvaμ [2.11]).
At a later stage still, different composers added four Yoga layers in
succession, turning the Epic-Såμkhya text into a “yogaçåstra.” Since the
later Yoga layers after the first also incorporated material from the Upani‚ads, the text simultaneously gained the appearance of an “upani‚adaª.”
This expanded text, however, existed separately both from the tri‚†ubh
hymn, which comprised verses 2.5–8 and 11.10–50 (and possibly 9.20–
21 and 15.15), and from the original epic episode, which, Jei conceded,
might have still continued to exist in its pure form alongside its revised,
Upani‚adic form.51 The author of the tri‚†ubh hymn, however, was seemingly less interested in the latter than the former for, Jei claimed, “he is
absorbed in giving the epic K®‚~a infinite might and dignity of the only
God paradoxically incarnated among the warriors, in subordinating all
the vedic gods to him and in declaring him to be ak‚aram paramam,
viçvasya paraμ nidhånam, çåçvatadharmagoptar, avyaya and puru‚aª
sanåtanaª (XI(18), 38)” (1986: 633). This tri‚†ubh hymn, Jei argued,
which subordinated the warrior Arjuna to K®‚~a, declaring him to be his
çi‚ya, or disciple, would probably have constituted the “Ur-Bhagavadgtå”
because “neither epic nor yogaçåstra includes anything that would justify
such a title” (634).52 However, Jei alleged that the composite text we
now know as the Bhagavad Gtå only arose once a “Våsudevabhakta”
undertook the “immense enterprise” of “synthesiz[ing] this hymn with
the upani‚adaª, connect[ing] their layers and ingeniously reinterpret[ing]
them in the sense of the bhakti, introduc[ing] new contents and reinte-
218 / Vishwa Adluri and Joydeep Bagchee
grat[ing] the whole in the epic Bhårata again by means of st. XVIII 72–
73, which, with the words kari‚ye vacanaμ tava, after so much instruction, skillfully connect[ed] the Gtå with the continuation [of the battle]
in the BhP. [Bh‚maparvan] XLI 1” (634).53 At this point, the Epic layer
comprising chapters 1 and (parts of) 2 along with its Såμkhya and Yoga
overlays comprising the remainder of chapter 2 and most of chapters 3–9
(or 10?) was combined with the tri‚†ubh hymn, most of which was placed
at the end of chapter 10 with the exception of a few verses that were
inserted into chapter 2 (as verses 5–8), chapter 9 (20–21), and chapter 15
(15).54 The tri‚†ubh hymn itself had to be enclosed in a few additional
verses (composed possibly by the Våsudevabhakta?) to bring it into line
with his overarching bhakti orientation.55 Thus, Jei argued, whereas
“the introductory çlokas [of chapter 11] present the marvellous r¨pam
aiçvaryam to us, the hymn [in the central part of the chapter] confronts us
with the ghorar¨pam” (632).56 Likewise, “the God’s assertion in the
hymn that nobody in the world of men except Arjuna can see him in
this form is most clearly reinterpreted in the concluding çlokas as the
instruction that there is no other means to see God in this form except
through bhakti” (632). Along with enclosing the tri‚†ubh hymn in new
verses, the Våsudevabhakta also seems to have been responsible for
making bhakti interpolations into the earlier layers with the intent of
transforming their character. Thus, he seems to have separated out a
section of the tri‚†ubh hymn (distinguished, Jei claimed, by its “pathetic
style” and by the fact that it puts “in Arjuna’s mouth a humble request
addressed to K®‚~a: yac chreyaª syån niçcitam br¨hi tan me, çi‚yas te
’ham çådhi måm tvåm prapannam” [630]) and inserted it into the Epic
layer. As a result this layer, which had earlier extended from 1.1–2.4 and
nine verses thereafter, now came to form verses 1.1–2.4, 2.9–10, and
2.31–37 of the Gtå.57 Likewise, he also made bhakti interpolations into
the first Yoga layer that earlier extended from 2.39–4.41 and 4.42, but
is now interrupted at 3.22 and 3.33 by two bhakti interpolations that
redefine the meaning of the word çre‚†ha (meaning “best” or “distinguished” or “superior”) to mean K®‚~a rather than Arjuna, as was originally the case. Finally, the Våsudevabhakta also seems to have been
responsible for creating the verbal resonances, not just between his layer
or his insertions and the existing layers in the text, but also the existing
layers themselves.
The Bhagavad Gtå and the Historical-Critical Method / 219
A Conveniently Flexible System for Classifying Repetitions
In Jei’s scheme, repetitions are a guide to the expansion of the text
because each repetition is one of two types: it can be either a continuity
repetition, in which case it develops a given subject, or it can be a duplication repetition, in which case it reinterprets what was previously said.
A continuity repetition indicates parts of the text belonging to the same,
original sequence. In contrast, (the presence of) a duplication repetition between layers permits us to identify parts belonging to different
sequences. Let us consider some examples.
According to Jei, “we meet with duplication repetitions in chapter II
for the first time in the description of Arjuna’s despondency” (1986: 630):
Epic
Hymnic
1.33
ye‚åm arthe kåk‚itaμ no
råjyaμ bhogåª sukhåni ca|
ta ime ’vasthitå yuddhe
prå~åμs tyaktvå dhanåni
ca\
2.6
na caitad vidmaª kataran
no garyo; yad vå jayema
yadi vå no jayeyuª|
yån eva hatvå na
jijvi‚åmas; te ’vasthitåª
pramukhe dhårtar傆råª\
1.35
etån na hantum icchåmi
ghnato ’pi madhus¨dana|
api trailokyaråjyasya hetoª
kiμ nu mahk®te\
2.8
na hi prapaçyåmi
mamåpanudyåd; yac
chokam uccho‚a~am
indriyå~åm| avåpya
bh¨måv asapatnam
®ddhaμ råjyaμ surå~åm
api cådhipatyam\
2.37
hato vå pråpsyasi svargaμ
jitvå vå bhok‚yase
mahm| tasmåd utti‚†ha
kaunteya yuddhåya
k®taniçcayaª\
11.33
tasmåt tvam utti‚†ha yaço
labhasva; jitvå çatr¨n
bhuk‚va råjyaμ
sam®ddham| mayaivaite
nihatåª p¨rvam eva;
nimittamåtraμ bhava
savyasåcin\
In this example, the words or phrases marked in bold represent duplication repetitions, whereas the words or phrases marked in italics represent
220 / Vishwa Adluri and Joydeep Bagchee
continuity repetitions. Thus, according to Jei, whereas avasthitå in the
Epic layer at 1.33 is picked up and repeated in the Hymnic layer at 2.6
(duplication repetition), mah in the Epic layer at 1.35 is picked up and
repeated within the same layer at 2.37 (continuity repetition). Likewise,
whereas the words jitvå and bhok‚yase and the expression tasmåd utti‚†ha
in the Epic layer at 2.37 are picked up and repeated (with slight changes
of inflection) in the Hymnic layer at 11.33 (duplication repetitions),
®dddhaμ råjyaμ in the Hymnic layer at 2.8 is picked up and repeated
within the same layer at 11.33 (continuity repetition). Accordingly, 2.6,
2.8, and 11.33 must constitute a separate layer or a separate work from
1.33, 1.35, and 2.37. The words used are similar, but they have acquired
a new meaning, which, in Jei’s opinion, suggests that someone intentionally intended to repeat the external form of the earlier layer to suggest
continuity, though of course, he or she was only duplicating the outward
form of the layer and not really continuing its thought.58 Hence, the
repetitions in question are duplication repetitions rather than continuity
repetitions. In contrast, when we look at the repetitions within a layer
(for example, between 1.35 and 2.37 or between 2.8 and 11.33), we see
that they continue the thought of that layer and hence mark those verses
as belonging to a single layer.
This analysis appears plausible, but when we look more closely at the
definition of continuity repetitions and duplication repetitions, we find
that the two types of repetitions are not criteria at all because they are not
defined in a way independent of their content. For instance, whether a
repetition should be considered a continuity repetition or a duplication
repetition is simply a function of whether it occurs within verses that Jei
thinks belong to the same layer or it occurs within verses that he posits as
belonging to distinct layers. For instance, if we ask why the repetition of
avasthitåª in 2.6 should be classified as a duplication repetition rather than
a continuity repetition, we find that the explanation is provided entirely
in terms of a perceived change in the text. In Jei’s words:
Stanzas II 5–8 evidently repeat the contents of some preceding stanzas,
but they differ from these, firstly, by having a more pathetic style, secondly, in metre, being composed in tri‚†ubhs and the stanzas repeated
in çlokas, and, thirdly, in sense: this is most obvious in st. 7 which is
unusual in that it anticipates st. 9, putting in Arjuna’s mouth a humble
The Bhagavad Gtå and the Historical-Critical Method / 221
request addressed to K®‚~a: yac chreyaª syån niçcitam br¨hi tan me,
çi‚yas te ’ham çådhi måm tvåm prapannam. In st. 9, on the contrary,
Arjuna simply decides na yotsye, without asking for anything. These
are all typical features of duplication repetitions (1986: 630; emphasis
added).
Contrary to Jei’s statement, however, these features cannot literally be
“typical features of duplication repetitions,” for the simple reason that
they are specific to the context. No other verse in the Bhagavad Gtå, for
instance, has the words yac chreyaª syån niçcitaμ br¨hi tan me, and so
on, contrasting with its succeeding verse, which has the words na yotsye.
Even if we accept the first two of Jei’s three criteria for identifying
duplication repetitions in contrast to continuity repetitions (that is, the
fact that they have “a more pathetic style” and furthermore “differ from
these [that is, the preceding stanzas belonging to the heroic layer]…in
metre”), we do not make much progress, since these criteria are also of
limited applicability. The criterion of “a more pathetic style,” for instance,
applies only to the transition from the Epic layer to the Hymnic layer,
and here also it is not true of the Hymnic layer as a whole but only of its
first few verses (2.5–8). Further, “pathetic style” is too imprecise a characterization to justify excision of these three verses.59 Likewise, the criterion of meter also applies only to the transition to the Hymnic layer, since,
with the exception of 2.20, 2.22, 2.29, and 2.70, Jei considers all tri‚†ubh
verses to belong to this layer. Even if we interpret Jei charitably to mean
duplication repetitions are those repetitions that give a new emphasis or
inflection to the original and thus contrast with the heroic outlook of the
original,60 the problem persists:
(1) There is no intrinsic, formal criterion for identifying a duplication
repetition.
(2) A duplication repetition is always identified when certain words are
repeated between two sections that the scholar has postulated as belonging
to different layers.
(3) Duplication repetitions cannot occur within a layer because by definition a repetition is classified as a duplication repetition whenever the
scholar notices a shift in layer.
Similar problems afflict Jei’s discussion of continuity repetitions. If
the problem with his definition of duplication repetitions was that it was
222 / Vishwa Adluri and Joydeep Bagchee
overly specific, being restricted in its application to a particular set of
verses, and, furthermore, not defined in a way independent of the contents
of those verses, the problem with his definition of continuity repetitions
is the obverse: his use of continuity repetitions is so broad that there is no
single definition of continuity repetition that can encompass all usages.61
Indeed, there cannot be a definition of continuity repetitions because:
(1) Jei is not using or not identifying formal features of these repetitions.
(2) Neither can he use content to define the continuity repetitions
because these repetitions occur in both types of layers (that is, original
and nonoriginal).
In his discussion of the continuity repetitions within the Epic layer (that
is, those between verses 1.35 and 2.37) and within the Hymnic layer (that
is, those between verses 2.8 and 11.33), Jei writes that “in both stanzas
we have cases of developing the respective argument from the previous
stanzas. The repetitions refer back to what is repeated as do answers to
questions” (1986: 631). Thus, far from using a formal criterion to identify
the continuity repetitions, he makes his perception of whether a verse
belongs together with another in the same layer the criterion for positing
a continuity repetition. For instance, he assumes that verse 2.37 provides
the appropriate, heroic answer to Arjuna’s questions in 1.33 and 1.35;
consequently, mahm in 2.37 must be a continuity repetition corresponding
to mah in 1.35. Likewise, the reason the repetition of ®ddhaμ råjyaμ from
2.8 in 11.33 is considered a continuity repetition, is because he sees it as
giving the appropriate, nonheroic, theistic answer to Arjuna’s questions
in 2.6 and 2.8.
In contrast, the reason the repetitions across the two layers (that is,
avasthitå in 1.33 corresponding to avasthitåª in 2.6 and jitvå, bhok‚yase,
and tasmåd utti‚†ha in 2.37 corresponding to jitvå, bhuk‚va, and tasmåt
utti‚†ha in 11.33) are considered duplication repetitions, is because the
two sets of answers embody different attitudes. In Jeic’s words:
It is remarkable that at the same time we meet again with duplication
repetitions. Accompanying the identical differences in style and metre
we have here, in addition, quite different answers. To Arjuna’s decision
not to fight comes the heroic answer [in 2.37]: If you fight, you will
win victory or heaven! If you desist, you will fall into dishonour. Is
The Bhagavad Gtå and the Historical-Critical Method / 223
there anything worse for a k‚atriya? So fight! In answer to Arjuna’s
prapatti to K®‚~a, on the contrary, the God in his revelation replies: I
have already killed your enemies, you should be only an instrument of
my work! The litteral [sic] cross repetitions: duplication of avasthita,
BHUJ, jitvå and tasmåd utti‚†ha, along with synonymity of trailokyaråjya with surå~åμ cådhipatyam versus continuity of mah, resp. råjyam
®ddham, would even be enough in themselves to indicate that we have
here to do with two complete and different descriptions of Arjuna’s
dilemma and K®‚~a’s reply to him: an epic one and a hymnic one, no
less (1986: 631).
The problem is obvious: the criteria are not being defined in a way independent of the layers. In fact, they play no role in the analysis at all. Jei
already has a definite idea of what the original layer would have been
and the so-called criterion of continuity repetition or duplication repetition is simply functioning as a cover. This becomes especially clear from
his example, which we reproduce again below along with arrows indicating the connections he posits:
Figure 2. Problems with the criteria of continuity repetitions and duplication
repetitions62
If we look at the same example again, we can ask why the repetitions
224 / Vishwa Adluri and Joydeep Bagchee
between 2.8 and 11.33, which Jei classifies as continuity repetitions in
order to draw together the two verses into one layer, could not just as well
be classified as duplication repetitions.
Figure 3. Determining when a repetition is a continuity repetition or a
duplication repetition
Further, since continuity repetitions are not a unique feature of the
original layer, there is no reason why the repetitions between 2.8 and
11.33, even if we wanted to call them continuity repetitions, should be
continuity repetitions of a different, Hymnic layer. They could be continuity repetitions on par with those found between 1.35 and 2.37 and hence
also features of the original, Epic layer. The criterion of continuity repetitions, then, is not in itself an argument for classifying certain verses as
part of a separate layer, for, as Jei defines these repetitions, they at
most demonstrate a certain continuity between verses without explaining
why the continuity between those verses (for example, between 1.35 and
2.37) is of a different order than that between yet others (for example,
between 2.8 and 11.33).
If continuity repetitions are not a unique feature of the original layer,
what of duplication repetitions? Do they at least have formal features (for
example, that they are restricted to a specific layer and hence permit us to
identify this layer with a degree of confidence)? Since duplication repetitions never occur within a layer, it seems, at least initially, that here we
must at last have a formal criterion for identifying layers. Yet, this only
appears to be the case since duplication repetitions, like continuity repeti-
The Bhagavad Gtå and the Historical-Critical Method / 225
tions, are also not defined in a way independent of the presumed layer to
which they belong. In other words, they, no less than their counterparts,
are actually superfluous to the search for or the identification of layers.
Since duplication repetitions are identified on the basis of the scholar
noticing a shift in layer (which he does on the grounds that there is a
change in tenor in the text or a new theme is introduced), the fact that
duplication repetitions are restricted to discrete layers does not actually
help us identify a layer. The scholar may at most refine his scheme of
layers by successively positing a series of duplication repetitions, but the
identification of these so-called duplication repetitions does not add to
the evidence for the existence of a layer. Let us consider Jei’s discussion of his Bhakti layer. He offers the following table as an example of
how noticing further duplication repetitions can lead us to identify new
layers in the text (in this case, the Bhakti layer). He prefaces his table with
the following comment: “if we now continue to follow it [the Hymnic
layer], we shall find abundant duplication repetitions in this chapter, where
the tri‚†ubhs of the hymn have been embedded in introductory and concluding çlokas” (Jei 1986: 631).
Tri‚†ubh Hymn
Bhakti Layer
11.22
rudrådityå vasavo ye ca
sådhyå; viçve ’çvinau
marutaç co‚mapåç ca|
gandharvayak‚åsurasiddhasaμghå; vk‚ante två
vismitåç caiva sarve\
11.6
paçyådityån vas¨n rudrån
açvinau marutas tathå|
bah¨ny ad®‚†ap¨rvå~i
paçyåçcaryå~i bhårata\
11.23
r¨paμ mahat te
bahuvaktranetraμ
mahåbåho
bahubåh¨rupådam|
bah¨daraμ
bahuda삆råkarålaμ
d®‚†vå lokåª pravyathitås
tathåham\
11.10
anekavaktranayanam
anekådbhutadarçanam|
anekadivyåbhara~aμ
divyånekodyatåyudham\
11.48
na vedayajñådhyayanair na
dånair; na ca kriyåbhir na
tapobhir ugraiª|
11.53–
54
nåhaμ vedair na tapaså
na dånena na cejyayå|
çakya evaμvidho
226 / Vishwa Adluri and Joydeep Bagchee
evaμr¨paª çakya ahaμ
n®loke; dra‚†uμ tvad
anyena kurupravra\
dra‚†uμ d®‚†avån asi måμ
yathå\ bhaktyå tv
ananyayå çakya aham
evaμvidho ’rjuna| jñåtuμ
dra‚†uμ ca tattvena
prave‚†uμ ca paraμtapa\
In this example, we have no instances of continuity repetitions between
the verses in one layer identifying them as members of the same layer.
However, we do have a noticeable number of repetitions between the verses
on the left-hand side of the table and those on the right. If we had placed
the two columns in sequence (that is, verses 11.22 and 11.23 occurring after
verses 11.6 and 11.10 and verses 11.53–54 occurring after verses 11.22
and 11.23), these repetitions would have appeared as continuity repetitions
linking the six verses together in a single, uniform layer. Yet why have
we chosen to classify the repetitions in these verses as duplication repetitions rather than continuity repetitions? Jei’s argument is as follows:
The style is no longer so different; but while the introductory çlokas
present the marvellous r¨pam aiçvaryam to us, the hymn confronts us
with the ghorar¨pam. The God’s assertion…that nobody in the world
of men except Arjuna can see him in this form is most clearly reinterpreted in the concluding çlokas as the instruction that there is no other
means to see God in this form except through bhakti. Çlokas are, at the
same time, advocating the view that even devas and dånavas cannot
see him in this form in which he contemplates himself (X 14–15, XI 3).
It again contradicts XI 22 (q.v.). In the hymn Arjuna asks Våsudeva to
see his caturbhuja form again (17 and 46), but çloka 51 interprets his
saumyavapus (50) simply as his human form. This implies that the hymn
has been embedded in a bhakti layer of close infinity [?], but the seams
are still recognizable (1986: 632).
As with the previous example, where Jei defined the repetitions between
1.33 and 2.6 and between 2.37 and 11.33 as duplication repetitions on the
grounds that “we have here…quite different answers” (631), here also
the real reason why the repetitions are taken to be duplication rather than
continuity repetitions is the apparent difference in content between the
The Bhagavad Gtå and the Historical-Critical Method / 227
two groups of verses. That means, however, that the repetitions are really
playing no role in the identification of the layers. At best they are a feint,
since, if Jei wanted to assert the existence of a layer, he could have done
so directly on the basis of the perceived differences in content rather than
invoking so-called duplication repetitions. Indeed, though he claims that
the characteristic feature of duplication repetitions is that they outwardly
reproduce the form of the text (though actually giving it a new meaning
or inserting new contents into the text), thus evincing the author’s interest
in working his own doctrines into the text in a way that will not draw
attention to his alterations, his own analysis shows that this cannot be the
case. Consider, for example, the so-called duplication repetitions between
the Epic and the Hymnic layers:
Epic
Hymnic
1.33
ye‚åm arthe kåk‚itaμ no
råjyaμ bhogåª sukhåni ca|
ta ime ’vasthitå yuddhe
prå~åμs tyaktvå dhanåni
ca\
2.6
na caitad vidmaª kataran
no garyo; yad vå jayema
yadi vå no jayeyuª| yån
eva hatvå na jijvi‚åmas; te
’vasthitåª pramukhe
dhårtar傆råª\
1.35
etån na hantum icchåmi
ghnato ’pi madhus¨dana|
api trailokyaråjyasya hetoª
kiμ nu mahk®te\
2.8
na hi prapaçyåmi
mamåpanudyåd; yac
chokam uccho‚a~am
indriyå~åm| avåpya
bh¨måv asapatnam
®ddhaμ råjyaμ surå~åm
api cådhipatyam\
2.37
hato vå pråpsyasi svargaμ
jitvå vå bhok‚yase
mahm| tasmåd utti‚†ha
kaunteya yuddhåya
k®taniçcayaª\
11.33
tasmåt tvam utti‚†ha yaço
labhasva; jitvå çatr¨n
bhuk‚va råjyaμ
sam®ddham| mayaivaite
nihatåª p¨rvam eva;
nimittamåtraμ bhava
savyasåcin\
Jei does not explicitly clarify who was responsible for these verbal
resonances. But it seems clear that it cannot have been the original composer of that section. Commenting on the verbal repetitions between the
Epic and Hymnic layers, Jei writes that “the litteral [sic] cross repeti-
228 / Vishwa Adluri and Joydeep Bagchee
tions: duplication of avasthita, BHUJ, jitvå and tasmåd utti‚†ha, along
with synonymity of trailokyaråjya with surå~åμ cådhipatyam versus
continuity of mah, resp. råjyam ®ddham, would even be enough in themselves to indicate that we have here to do with two complete and different
descriptions of Arjuna’s dilemma and K®‚~a’s reply to him: an epic one
and a hymnic one, no less” (1986: 631). This suggests that the creator of
the tri‚†ubh hymn, although in reality providing a “different description…
of Arjuna’s dilemma and K®‚~a’s reply to him” took pains to adapt his
language (though not his meter) to the text of the original (631). But elsewhere Jei also claims that the tri‚†ubh hymn was never intended to be
part of the original epic episode. In his words, “this hymn evidently presuppose[d] the epic part of the poem, but since it contain[ed] a proper
description of Arjuna’s dilemma, and not only K®‚~a’s answer to it, it
must have been intended as a separate composition, and not as a layer
added to the epic episode” (633). In that case, the composer of the tri‚†ubh
hymn, not knowing that his composition would later be incorporated into
the Bhagavad Gtå, could not have authored its duplication repetitions.63
Further, if duplication repetitions are defined as repetitions that undertake
a superficial assimilation of new text to the original (though introducing
new ideas, ones possibly at odds with the view of the original), the duplication repetitions in the tri‚†ubh hymn must have been authored by the
person responsible for combining the tri‚†ubh hymn with the epic poem,
that is, by the Våsudevabhakta. And since this applies not only to the
tri‚†ubh hymn but also to all the parts of the Gtå that underwent a later
revision, this means that we have no evidence for duplication repetitions
being expressions of the intentions of the author of the respective section.
At best, they would be a guide to the actions or intentions of the Våsudevabhakta.64 Since the duplication repetitions are actually posterior to the texts
that were combined to produce the poem, they are of no use in determining
the form or extent of these texts. In order to know and to be able to identify
what changes the Våsudevabhakta made, we would already have to know
what texts he had before him.65 That means, however, that the duplication
repetitions are actually not criteria for identifying layers. On the contrary,
Jei is using the thesis disingenuously to justify the assumption of layers.
This would not be a problem if Jei could point to some objective feature that distinguishes duplication repetitions from continuity repetitions.
Yet, as we have seen, a duplication repetition is identified precisely on
The Bhagavad Gtå and the Historical-Critical Method / 229
the basis of an assumed shift in layer. In other words, whether a particular
repetition is to be classified as a continuity repetition or a duplication
repetition is entirely a function of its place in the scholar’s a priori scheme
of the Bhagavad Gtå’s layers, and that means that the duplication repetition cannot actually be used, in turn, to corroborate that scheme. The
terms continuity repetition and duplication repetition are therefore misleading: the repetitions neither intrinsically continue anything, nor do
they intrinsically duplicate anything. It is the scholar who determines in
each case whether a repetition continues something or duplicates something. The criterion of duplication repetitions falls for a third time out of
consideration as a guide to the Bhagavad Gtå’s layers.66
Similar problems arise concerning the second of Jei’s criteria for
identifying layers, namely, his concept of beginning and ending markers.
After introducing the concept of continuity repetitions, Jei writes:
“Especially valuable instances of continuity repetitions were the beginning and the ending markers of particular sequences, which sometimes
enable one to detect the layers and then to ascertain in greater detail how
they could have been intermingled” (1986: 629). As an example of beginning and ending markers, he presents the following sequence of verses:
Epic
2.10
2.37
tam uvåca h®‚keçaª
prahasann iva bhårata|
senayor ubhayor madhye
vi‚dantam idaμ vacaª\
hato vå pråpsyasi svargaμ
jitvå vå bhok‚yase mahm|
tasmåd utti‚†ha kaunteya
yuddhåya k®taniçcayaª\
Såμkhya
2.11
açocyån anvaçocas tvaμ
prajñåvådåμç ca bhå‚ase|
gatås¨n agatås¨μç ca
nånuçocanti pa~itåª\
2.15
yaμ hi na vyathayanty ete
puru‚aμ puru‚ar‚abha|
samaduªkhasukhaμ dhraμ
so ’m®tatvåya kalpate\
2.30
deh nityam avadhyo ’yaμ
dehe sarvasya bhårata|
tasmåt sarvå~i bh¨tåni na
tvaμ çocitum arhasi\
2.38
sukhaduªkhe same k®två
låbhålåbhau jayåjayau|
tato yuddhåya yujyasva
naivaμ påpam avåpsyasi\
230 / Vishwa Adluri and Joydeep Bagchee
As in the previous examples, the italicized text indicates continuity
repetitions (in this case, all between the verses of the so-called Såμkhya
layer), while the bold text indicates duplication repetitions. Additionally,
the beginning and ending markers of the two sections are highlighted in
gray. Jei comments:
II 11 begins with a complete change of tenor and introduces the new
theme: açocyån anvaçocas tvam. Coming to st. 30, we recognize the
corresponding conclusion introduced by tasmåt. Such a relationship of
continuity repetitions I have called that of beginning and ending markers.
The cross repetitions in II 38 referring to, e.g. II 14,67 on the one hand,
which 38ab continues, and to II 37 (q.v.) which 38cd duplicates and
reinterprets, reveal stanzas 11–30, 38 to be a layer ingeniously embedded in the epic. St. II 39 christens it såμkhya. This then proves definitely what the epic episode really was [!] (1986: 632).
Contrary to Jei’s claim, however, neither the beginning marker açocyån
anvaçocas tvaμ, and so on (2.11) nor the ending marker tasmåt (2.30)
permit us to identify the layer with any degree of confidence. Even though
the idea that those interpolating the text would have neatly enclosed their
interpolations with clear, formal indications (analogous to the Sanskrit
word iti, which indicates a quotation or the end of a phrase) is appealing,
in practice the criterion does not hold up to examination. This becomes
especially clear in the case of the so-called beginning markers.
As we have seen, Jei posits the existence of a beginning marker at
2.11 because he thinks it introduces a different theme from the Epic layer,
which it interrupts after 2.10. Thus, whereas the former contains a heroic
response to Arjuna’s dilemma (namely, the injunction to fight and obtain
either heaven or earth in 2.37),68 the Såμkhya layer provides a different
response to this dilemma. It responds in terms of the necessity not to
grieve, as is appropriate for the wise man. Thus, from Jei’s perspective,
the words açocyån anvaçocas tvaμ function as a beginning marker for
this new layer, which, as he says, “begins with a complete change of
tenor and introduces the new theme” (1986: 632). As this only means
that he thinks there is a change in the text at this point, nothing is gained
for his argument for the existence of a layer to call açocyån anvaçocas
tvaμ a beginning marker: he could simply have asserted his sense that a
The Bhagavad Gtå and the Historical-Critical Method / 231
new layer begins here. If there is no intrinsic, formal characteristic for a
beginning marker and the marker is always identified on the basis of the
scholar’s perception of a new layer,69 it cannot be used, in turn, as confirmation of the existence of that layer. Similar problems afflict Jei’s identification of the beginning markers of his next two layers, the Yoga layer
and the Bhakti layer. Concerning the first, Jei writes: “II 39 is again a
beginning marker; in referring to what has been said, it resumes the såμkhya
layer and not the epic one.…[It simultaneously] introduces…[a] new
theme: yoga” (633). The Yoga layer, according to him, is clearly marked
out as a distinct layer, not only by the presence of continuity repetitions
throughout its extent, but also by the occurrence of the ending marker
tasmåt at its conclusion in verse 4.42. Let us first look at his illustration
before we consider his analysis:
Epic
2.10
2.37
tam uvåca h®‚keçaª
prahasann iva bhårata|
senayor ubhayor madhye
vi‚dantam idaμ vacaª\
hato vå pråpsyasi svargaμ
jitvå vå bhok‚yase mahm|
tasmåd utti‚†ha kaunteya
yuddhåya k®taniçcayaª\
Yoga
2.39
e‚å te ’bhihitå såμkhye
buddhir yoge tv imåμ
ç®~u| buddhyå yukto yayå
pårtha karmabandhaμ
prahåsyasi\
4.41
yogasaμnyastakarmå~aμ
jñånasaμchinnasaμçayam|
åtmavantaμ na karmå~i
nibadhnanti dhanaμjaya\
4.42
tasmåd ajñånasaμbh¨taμ
h®tsthaμ jñånåsinåtmanaª|
chittvainaμ saμçayaμ
yogam åti‚†hotti‚†ha
bhårata\
As before, the continuity repetitions (within the Yoga layer only) are
indicated by means of italics; the duplication repetitions (between the
Epic layer and the Yoga layer) are indicated by means of bold letters.
The beginning marker (såμkhye at 2.39) and the ending markers (tasmåt
at 2.37 and 4.42) are indicated by means of a gray highlight. Jei
comments on this layer as follows:
232 / Vishwa Adluri and Joydeep Bagchee
The ending marker of this yoga layer we can recognize by continuity
repetitions in IV 41 (see illustration 6 [reproduced above]).…There are
more repetitions than illustration 6 shows because the jñåna too was, in
the meantime, explained as såμkhya (III 3). In the next stanza IV 42
we recognize the cross repetitions: yogam åti‚†ha, the conclusion of the
layer, is connected with the duplication repetition of the epic layer
tasmåd…utti‚†ha (v. II 37), thus concluding the whole poem for the
third time within its epic frame of Bh‚maparvan XXIV 10 and XLI 1
(1986: 633).
However, here also, the beginning marker is not identified on the basis of
a formal characteristic. In fact, there is something quite strange about the
so-called beginning marker e‚å te ’bhihitå såμkhye: although supposedly
the beginning marker of the Yoga layer, it features the word såμkhya and
thus, as Jei says, seems to “christen” (1986: 632) the previous layer.
The only feature that seems to justify us in considering it the beginning
marker of this (that is, the Yoga) layer is that it features the word yoga
and thus, as Jei says, “introduces the new theme: yoga” (633). But if
the beginning marker is being identified on the basis of the supposed
beginning of a new layer (which, in turn, is being done on the basis of
Jei’s suspicion that a new theme is introduced here), then it can hardly
be used in turn to confirm the existence of that layer.70 Jei could just as
well have asserted that a new layer begins here, without the search for a
“beginning marker” that marks nothing except the beginning of the layer
that he thinks it marks. Unless the term beginning marker is defined in a
formal sense, the thesis that there are markers that permit us to identify
the presence of new layers where the only characteristic these markers
possess is that they stand at the beginning of the layers we think exist in
the text is no better than to assert that we think there are layers in the text.71
The thesis of special markers that indicate these beginnings proves nothing,
and the markers themselves are superfluous.
If there is no formal definition of a beginning marker, what of the socalled ending markers? Are they any more reliable as a criterion for the
existence of layers in the Bhagavad Gtå? Initially, at least, it seems that
Jei must be on stronger grounds here, since he defines the ending marker
in a formal sense. The ending markers consist of the word tasmåt. Thus,
we can now reliably use this criterion to identify the end of different
The Bhagavad Gtå and the Historical-Critical Method / 233
layers without having to rely on our subjective perceptions of changes in
the text. For instance, tasmåt in 2.37 would mark the end of the Epic layer.
Likewise, tasmåt in 2.30 would mark the end of the Såμkhya layer. The
following table gives the scheme of layers as identified by Jei using the
criterion of the ending marker tasmåt:
2.37
hato vå pråpsyasi svargaμ jitvå vå
bhok‚yase mahm| tasmåd utti‚†ha
kaunteya yuddhåya k®taniçcayaª\
End of the Epic
layer beginning
1.1–47, 2.1–4,
2.9–10, 2.37
2.30
deh nityam avadhyo ’yaμ dehe sarvasya
bhårata| tasmåt sarvå~i bh¨tåni na tvaμ
çocitum arhasi\
End of the
Såμkhya layer
beginning 2.11–30,
2.38
4.42
tasmåd ajñånasaμbh¨taμ h®tsthaμ
jñånåsinåtmanaª| chittvainaμ saμçayaμ
yogam åti‚†hotti‚†ha bhårata\
End of the Yoga
layer beginning
2.39–4.41, 4.42
Unfortunately, the ending marker tasmåt is no more successful than the
beginning markers in permitting us to reliably identify the layers in the
text because the term also occurs at 1.37, 2.18, 2.25, 2.27, 2.50, 2.68, 3.15,
3.19, 3.41, 4.15, 5.19, 6.46, 8.7, 8.20, 8.27, 11.33, 11.44, 16.21, 16.24,
17.24, and 18.69. If Jei were serious about using the marker tasmåt in
a formal sense, the Epic layer would have ended at 1.37. Likewise, the
Såμkhya layer should have ended at 2.18, with a new layer at 2.19–25
and a further layer at 2.26–27; the Yoga layer at 2.50, with new layers at
2.50–68, 2.68–3.15, 3.15–19, 3.19–41, and 3.41–4.15; the Upani‚adic
layer at 5.19, with a new layer at 5.19–6.46; and the Bhakti layer at 8.7,
with new layers at 8.7–20 and 8.20–27. Along similar lines, the Tri‚†ubh
layer should have ended at 11.33, with a new layer at 11.33–44, and the
(second) Upani‚adic layer should have ended at 16.21, with new layers at
16.21–24, 16.24–17.24, 17.24–18.69a, and 18.69b–d. As this list demonstrates, even though Jei claims to introduce tasmåt as a formal marker
for the end of a layer in the text, it is not being used in any objective sense.
Of twenty-five occurrences of the term in the text, Jei only considers
three. Of these three, only one (2.37) occurs where his model of the
Bhagavad Gtå’s layers predicts it will occur. The remaining two either
234 / Vishwa Adluri and Joydeep Bagchee
occur slightly before their predicted place in the text (for example, at
2.30 even though the Såμkhya layer actually ends at 2.38) or slightly
after (at 4.42 even though the Yoga layer actually ends at 4.41). Between
these three, there are numerous other occurrences that would require us
to break the layer and insert new ones, were we really relying on tasmåt
as a formal criterion.
Since both the beginning and the ending markers fail as objective, noncircular criteria for the identification of layers, this means that we are
returned to Jei’s initial hypothesis, namely, that there are continuity
and duplication repetitions in the text that permit us to identify and reconstruct the layers of the Bhagavad Gtå. Earlier, we noted that the criterion
of continuity and duplication repetition was insufficient to permit us to
objectively identify the existence of layers in the text, since these categories were not defined in a way independent of the contents of the text and
that means, ultimately, in a way independent of our perception of changes
to the text. Having seen that there is really no intrinsic distinction between
the presumed layers in the text (the supposed interpolators were neither
consistent enough to use tasmåt only in those instances where they wished
to end their interpolation, nor clear-sighted enough to combine the beginning of their interpolation with some kind of formal indication for the
beginning of a new layer rather than leaving us to conjecture at the beginning on the basis of the change in theme), let us consider a final example
where we see that even the basic requirement for a duplication repetition,
namely, that it be a repetition, does not hold.
Yoga
Bhakti
3.21
yad yad åcarati çre‚†has
tat tad evetaro janaª| sa
yat pramå~aμ kurute
lokas tad anuvartate\
3.22
na me pårthåsti kartavyaμ
tri‚u loke‚u kiμ cana|
nånavåptam avåptavyaμ
varta eva ca karma~i\
3.25
saktåª karma~y
avidvåμso yathå kurvanti
bhårata| kuryåd vidvåμs
tathåsaktaç cikr‚ur
lokasaμgraham\
3.23
yadi hy ahaμ na
varteyaμ jåtu karma~y
atandritaª| mama
vartmånuvartante
manu‚yåª pårtha
sarvaçaª\
The Bhagavad Gtå and the Historical-Critical Method / 235
Jei introduces this example as an illustration of the changes the
Våsudevabhakta would have made to the existing components of the
Bhagavad Gtå to bring it in line with his bhakti philosophy. According
to him, this example constitutes “the first clear interpolation of the bhakti
layer inside the previous layers of the poem” (Jei 1986: 634). As before,
we might justifiably ask why the repetitions on the right-hand side of the
table should be considered duplication repetitions rather than continuity
repetitions (in which case, we would have simply listed verses 3.22 and
3.23 under verse 3.21 and before 3.25). Jei’s answer is as follows: “The
verbal correspondences are here less important than the semantic ones:
åcarati–varte, lokaª–manu‚yåª, pramå~a–vartman, all implying that the
çre‚†ha is the God Våsudeva and not Arjuna as was first intended by his
friend, charioteer and teacher K®‚~a, according to his role in this yoga
layer” (634).
However, there are a number of problems with this thesis. First, the
criterion of repetition has been diluted to the point that it is no longer
even necessary that we have a repetition in the strict sense (that is, as a
verbal repetition). Indeed, me and aham are not repetitions of çre‚†ha in
any sense. They are merely taken to be repetitions because Jei thinks
that they interpret çre‚†ha in a new way, relating it to K®‚~a rather than
Arjuna (as would have been the case—he claims—in the original, heroic
poem). Second, we also note that Jei, in explaining why he considers
the words vartate, manu‚yåª, and vartman to be repetitions (of åcarati,
lokaª, and pramå~a, respectively), has already assumed the existence of
a Yoga layer and, moreover, what stance this layer would have taken concerning the relationship of Arjuna to K®‚~a, even though this is something that was supposed to be established first in the analysis. In other
words, once again the repetition has been identified as a duplication repetition on the basis of the assumption that it imparts a new interpretation
to the preceding stanzas. There is no intrinsic, formal criterion for a duplication repetition. In this case, Jei is even willing to set aside the criterion
of repetition to make his case; or rather, he reintroduces the criterion only
after he has identified his layers so as to give them a veneer of scientific
legitimacy. It now really seems as though, “without any intention to search
for the layers” and merely on “reading the Bhagavadgtå” (Jei 1986:
629), he has come across the layers, which would suggest they are an
objective feature of the text that any unbiased observer might find.72
236 / Vishwa Adluri and Joydeep Bagchee
The First Yoga Layer in the Bhagavad Gtå
In a second article, published the same year as this first one was delivered
(at Weimar in 1979), Jei attempted to extend these principles to study
the changes within specific layers. As his example, he chose the Yoga
layer—according to him, “the oldest” (Jei 1979: 555) of the Gtå’s
layers.73 In the first section of his article, Jei provided an overview of
his methodology and a brief summary of the results he had arrived at so
far using these methods. In the second, he then took up the question of
the mechanics of the addition of the Yoga layer. As with the earlier
examples, he attempted to show both that the Yoga layer was a coherent
layer (as evinced by the existence of continuity repetitions between the
verses of the layer) and that it was a discrete layer (as evinced by duplication repetitions between it and the other layers). Let us consider his
first example.
First Yoga
2.39
e‚å te ’bhihitå såμkhye
buddhir yoge tv imåμ
ç®~u| buddhyå yukto yayå
pårtha karmabandhaμ
prahåsyasi\
3.2
vyåmiçre~aiva våkyena
buddhiμ mohayasva me
tad ekaμ vada niçcitya
yena çreyo ’ham åpnuyåm
4.41
yogasaμnyastakarmå~aμ
jñånasaμchinnasaμçayam|
åtmavantaμ na karmå~i
nibadhnanti dhanaμjaya\
4.42
tasmåd ajñånasaμbh¨taμ
h®tsthaμ jñånåsinåtmanaª|
chittvainaμ saμçayaμ
yogam åti‚†hotti‚†ha
bhårata\
Second Yoga
5.1
saμnyåsaμ karma~åμ
k®‚~a punar yogaμ ca
çaμsasi| yac chreya
etayor ekaμ tan me br¨hi
suniçcitam\
As we have already noted the arguments for the Yoga layer being a
The Bhagavad Gtå and the Historical-Critical Method / 237
coherent layer, we need only briefly recapitulate them here. As before,
Jei took 2.39 to be the common ending verse of both the Epic and
Såμkhya layers, while simultaneously also being the beginning verse of
the Yoga layer. The word såμkhya is allegedly the ending marker of the
previous layer, while also serving, in conjunction with yoga, as the beginning marker of the new layer. In contrast, the ending marker of the Yoga
layer tasmåd utti‚†ha recalls, at least in form if not content, the tasmåd
utti‚†ha of the Epic layer at 2.37. (It is, therefore, a duplication repetition
and not a continuity repetition.)
The more interesting argument, however, concerns Jei’s reasons for
positing a new layer beginning after 4.42. He writes that “in order to prove
conclusively that the layer ends here, we should, on the one hand, analyse
the whole of the subsequent text and check whether it could be in entirety
of a different and later origin” (Jei 1979: 548). But, as this task is “not
feasible in this paper” and “we must take it for granted [!] that this analysis
has already been done” (548), Jei restricts himself to just one example.
This is the verse 5.1, allegedly the beginning of the second Yoga layer.
According to him, this verse repeats some of the words of verse 4.41 from
the previous layer (that is, saμnyastakarmå~aμ from 4.41 and çreyo from
3.2).74 However, “it evidently introduces a new text sequence,” for it does
not continue with the theme of yoga from the previous layer but introduces
a “doctrinal” theme, namely, “the difference between the yoga and the
saμnyåsa” (548). This, in Jei’s view, justifies us in the assumption that
a new layer was begun here. Why does 5.1 represent the beginning of a
new layer and not just the introduction of a new theme? Jei’s answer is
as follows: first, there was no mention of saμnyåsa until now; second,
karmayoga had been explicated until now as “an activity free from passions and selfish motives” (but without reference to actual renunciation);
and third, the new layer “seems to have arisen, partially, out of a misunderstanding of the preceding sequence, and, in particular, of the expression
‘yogasaμnyastakarman’ in st. IV.41” (548). According to Jei, this
expression, “possibly puzzled the author of the following sequence, and
consequently it had to puzzle Arjuna” (548). Finally, the reason given for
the verses following 5.1 being a new layer is that the individual or individuals behind this interpolation “wish[ed] to elaborate the text in the
light of a new conception of the saμnyåsa” (548). This led them to create
what Jei terms “the second yoga layer” (548).
238 / Vishwa Adluri and Joydeep Bagchee
Nonetheless, even if this establishes the beginning and end points of
the first Yoga layer (that is, 2.39 and 4.42), this does not mean that the
Yoga layer still exists, as a whole, in the form in which it was first interpolated into the poem. For the Yoga layer has itself been subject to interpolation. As evidence, Jei focused on verse 2.54, of which he noted:
“It is remarkable how clearly II.54 differs from the rest” (1979: 549).
According to him, whereas “the other questions are existential; although
less personal than the epic stanza II.4, [and] they still express Arjuna’s
abhorrence of the ‘karma ghoram’ or the ‘påpam,’ and so they remain
attuned to the epic,” “II.54, on the contrary, is an astonishingly scholastic
question” (549). “It appears without any convincing relation to the preceding text, [it is] compositionally without function, and it introduces
quite a new terminology, e.g. ‘prajñå’ instead of ‘buddhi’ etc.” (549).
This justifies us, Jei claimed, in the assumption that it is an interpolation. Further, he argued that the insertion of this verse “could have been
motivated by the expression ‘sthåsyati…samådhau (…buddhiª)’ in the
preceding stanza” (549). The following table represents the scheme of
repetitions:
First Yoga
2.53
3.1
çrutivipratipannå te yadå
sthåsyati niçcalå|
samådhåv acalå buddhis
tadå yogam avåpsyasi\
jyåyas cet karma~as te
matå buddhir janårdana|
tat kiμ karma~i ghore
måμ niyojayasi keçava\
Prajñå Interpolation
2.54
sthitaprajñasya kå bhå‚å
samådhisthasya keçava|
sthitadhª kiμ prabhå‚eta
kim åsta vrajeta kim\
2.55
2.56
2.57
2.58
2.59
2.6175
2.67
2.68
2.69
2.70
sthitaprajñas tadocyate
sthitadhr munir ucyate
tasya prajñå prati‚†hitå
tasya prajñå prati‚†hitå
(continues 2.58)
tasya prajñå prati‚†hitå
tad asya harati prajñåμ
tasya prajñå prati‚†hitå
(continues 2.68)
(“attracted” by 2.68)
The Bhagavad Gtå and the Historical-Critical Method / 239
3.36
atha kena prayukto ’yaμ
påpaμ carati p¨ru‚aª|
anicchann api vår‚~eya
balåd iva niyojitaª\
The first Yoga layer is interrupted at 2.53, and an interpolation consisting of duplication repetitions that pick up the form but not the content of
this verse begins at 2.54. This interpolation, according to Jei, displays
all the features of being an independent layer: not only does it feature
duplication repetitions intended to make the reader think it is an organic
part of the older layer, but it also features continuity repetitions within
the verses of the layer. Thus, sthitaprajña at 2.54 is repeated at 2.55 and
recalled in sthitadh at 2.56, whereas the term prajña runs through all of
the verses 2.57–58, 2.61, and 2.67–68.76 Notably, this term differs from
the buddhi of the preceding stanzas. Verses 2.59 and 2.69, to be sure, do
not have any continuity repetitions that mark them out as belonging to
this layer, but they “seem rather to continue st. 58, resp. 68, than to
belong to the original first yoga layer” (Jei 1979: 550). Finally, there is
no reason for 2.70 to be included as part of this interpolation, but verse
2.68 might “by its proverbial character” have “attract[ed] the only
tri‚†ubh stanza in our sequence, II.70” (550). The original sequence of
the Yoga layer resumes at 3.1, which manifests a clear relationship with
3.36, as evinced by the existence of continuity repetitions between them.
Jei also identified six other interpolations in the first Yoga layer:
(1) the second Yoga layer interpolation (2.72)77; (2) the first yajña interpolation (3.9–16); (3) the second yajña interpolation (4.23–34); (4) the
svadharma interpolation (3.35)78; (5) the åtman interpolations (3.17–18,
3.42–43); and (6) the Bhakti layer (2.61, 3.22–24, 3.30–32, 4.1–15, 4.35).
Disregarding the two single-verse interpolations (1 and 4), let us consider Jei’s arguments for the first and second yajña interpolations, the
åtman interpolations, and the Bhakti layer. Regarding the first yajña interpolations, Jei noted that “there are two segments of our sequence which
treat of…yajña” (1979: 550). The first of these was 3.9–16; the second
4.23–34. The former, Jei noted, began with a verse “the second half of
which seems to duplicate III.9 while introducing the rest of the segment
[that is, by featuring the theme of yajña]” (550). Let us look at the verses
in relation.
240 / Vishwa Adluri and Joydeep Bagchee
First Yoga
3.8
niyataμ kuru karma tvaμ
karma jyåyo hy akarma~aª|
çarrayåtråpi ca te na
prasidhyed akarma~aª\
3.19
tasmåd asaktaª satataμ
kåryaμ karma samåcara|
asakto hy åcaran karma
param åpnoti p¨ru‚aª\
First Yajña Interpolation
3.9
yajñårthåt karma~o ’nyatra
loko ’yaμ
karmabandhanaª|
tadarthaμ karma
kaunteya muktasagaª
samåcara\
3.10
3.1179
3.12
3.13
3.14
3.15
3.1680
sahayajñåª prajåª
(continues 3.10?)
dåsyante yajñabhåvitåª
yajñaçi‚†åçinaª santo
yajñåd bhavati…yajñaª
nityaμ yajñe prati‚†hitam
(continues 3.15?)
In this example, 3.8 and 3.9 have continuity repetitions (the word
karma) that are also in keeping with the basic theme of the first Yoga
layer. However, a new theme (that is, yajña) is introduced at 3.9. This
verse is the first of an eight-verse interpolation (3.9–16); only at the end
of this sequence (ignoring verses 17–18 for the moment) does the poem
return once again to the original theme of karma (in verse 19). However,
in order to make the break less conspicuous, the anonymous composer of
3.9–16 has emulated the outward form of the verse he displaced from its
original position after 3.8 to 3.19 and made it seem as though verse 3.9
(and the seven verses that follow it) is a genuine continuation of verse 3.8.
There are two problems with this thesis: the underlying assumption is
both erroneous and self-defeating. An interpolator would only have needed
to emulate the form of 3.19 to suggest that his verse 3.9 originally stood
in that place (where 3.19 originally stood) if his readers had been aware
that 3.19 was the successor to 3.8. In that case, it makes sense to assert
that, by using similar language, he deceived readers into thinking they
were reading the same text (that is, the text as they knew it), when, in fact,
following the bridge verse 3.9, they were actually being led to a com-
The Bhagavad Gtå and the Historical-Critical Method / 241
pletely different theme and completely different contents. This assumption is erroneous because if a composer’s aim is to suggest continuity, he
would not emulate the form of the verse he is displacing, but that of the
verse he is continuing—in other words, of 3.8. Further, if the purpose of
emulating the outward form of 3.19 is to suggest that 3.9 is the original,
organic continuation of 3.8 and there has been no substitution, why would
the composer of the interpolation leave the original verse (that is, 3.19)
in the text? (The argument is like suggesting that a jewel thief, having
replaced a valuable necklace with a paste imitation, then left the original
lying conveniently nearby for comparison.) The argument is thus also
self-defeating: if 3.9 was created to emulate 3.19, then the continued
presence of its model in the same text defeats the purposes for which 3.9
was originally created.
To be sure, Jei acknowledges that “the case is not quite certain” (1979:
551). However, his solution to the problem is incoherent: to make plausible the link between 3.9 and 3.19 he avers that “if there are some reasons…
to consider III.17–18 an interpolation too, then stanza 19 would follow, if
9–16 is an interpolation, at an earlier stage of the text history immediately
after 8, and the suspicion that the correspondences between 9 and 19,
each in its turn following after 8, could possibly present us with a case of
duplication, would be formally corroborated” (551). Actually it would not,
because the connection between 3.8 and 3.19 remains the same whether
or not we assume that they were consecutive, that is, whether or not we
assume that they were separated by the two verses 3.17 and 3.18. In any
case, if 3.9 is one of the two verses (the other being 3.19) originally “in
its turn following after 8,” then we would expect that the composer made
efforts to draw a connection between 3.8 and 3.9, but this is something
Jei has not shown at all (551). We would have the following four options:
Original
Situation
(without Åtman
Interpolation)
Yajña
Interpolation
(without Åtman
Interpolation)
Yajña
Interpolation
(with Åtman
Interpolation)
Original
Situation
(with Åtman
Interpolation)
3.8
3.8
3.8
3.8
3.19
3.9–16
3.9–16
3.17–18
3.19
3.17–18
23.19
3.19
242 / Vishwa Adluri and Joydeep Bagchee
In the first situation, 3.19 follows directly on 3.8. Thereafter, someone
inserted a yajña interpolation, duplicating (in 3.9) the outward form of
3.19, which he displaces to verse 16 in the chapter (the åtman interpolations do not exist as yet). Later still, someone inserted verses 3.17–18,
displacing verse 3.19 from sixteenth place to nineteenth place in the text.
Alternatively, 3.17 and 3.18 always existed in the text (or they were
inserted before 3.9–16), and 3.19 followed two verses after 3.8. Thereafter,
someone inserted 3.9–16 after 3.8, displacing 3.17–18 and 3.19 to their
present positions in the text. The implied connection between 3.9 and
3.19 is not strengthened in any way by assuming that someone inserted
3.17–18 between them. It only creates a convenient fiction of how 3.19
might have been functionally similar to 3.9, that is, that at one time it
occupied the same place after 3.8, without in any way being able to show
that it ever did.81
Let us come to the third of Jei’s examples. The second yajña interpolation is unusual in that it does not make reference (via duplication repetitions) to the original text or to its surrounding context, as we might
expect given Jei’s discussion of the rationale for duplication repetitions
thus far. Rather, it makes reference to the first yajña interpolation! The
following table illustrates the duplication repetitions between verses 3.9
and 4.23:
First Yajña Interpolation
3.9
yajñårthåt karma~o ’nyatra
loko ’yaμ
karmabandhanaª|
tadarthaμ karma kaunteya
muktasagaª samåcara\
Second Yajña Interpolation
4.23
gatasagasya muktasya
jñånåvasthitacetasaª|
yajñåyåcarataª karma
samagraμ pravilyate\
4.24
(continues 4.23?)
4.25
daivam evåpare yajñaμ…
yajñaμ yajñenaivopajuhvati
4.26
(continues 4.25?)
4.27
(continues 4.25?)
4.28
dravyayajñås tapoyajñå
yogayajñås tathåpare|
The Bhagavad Gtå and the Historical-Critical Method / 243
svådhyåyajñånayajñåç ca
yatayaª saμçitavratåª\
4.29
(continues 4.28?)
4.30
yajñavido
yajñak‚apitakalma‚åª
4.31
yajñaçi‚†åm®tabhujo…
ayajñasya kuto
4.32
evaμ bahuvidhå yajñå
4.33
yajñåj jñånayajñaª
paraμtapa
4.3482
(continues 4.33?)
Jei’s rationale for considering verses 4.23–34 a separate layer is that
“while the first segment [that is, 3.9–16] explains succinctly the theory of
the sacrifice in the original sense of the word, the second enumerates the
types of the sacrifice, subsuming under this conception a number of ascetic
practices and distinguishing the material sacrifice from the sacrifice by
knowledge alone” (1979: 551). Nonetheless, it is not clear why this second
interpolation would wish to suggest an organic continuity with the first:
should it not, first and foremost, wish to suggest an organic connection
with its surrounding text? And if the aim was to suggest an organic continuity with the first yajña interpolation, why was it not inserted there, that
is, immediately after 3.16? Further, according to the logic unfolded by
Jei, when a composer inserts a passage, he duplicates the form of the
verse he displaced. Thus, the composer of 3.9, in displacing 3.19 from its
position after 3.8, duplicated the form of 3.19. Should not then the author
of the second yajña interpolation, presuming he originally inserted or
wished to insert his passage as a continuation of 3.16, have also duplicated
the form of 3.19? According to our logic, which states that a composer
would be more likely to duplicate the form of the verse immediately preceding his insertion than the verse he has displaced after his insertion, the
author of 3.9–16 should have duplicated the form of 3.8 and the author of
4.23–34 should have duplicated the form of 3.16, but in either case there
is absolutely no rationale for duplicating the form of 3.9.
Jei’s fourth example is the most complex and, in many ways, the
most interesting. Let us first look at the verses.
244 / Vishwa Adluri and Joydeep Bagchee
Prajña Interpolation
2.55
prajahåti yadå kåmån
sarvån pårtha
manogatån| åtmany
evåtmanå tu‚†aª
sthitaprajñas
tadocyate\
Ka.Up.
1.3.1083
indriyebhyaª parå hy
arthå arthebhyaç ca
paraμ manaª| manasas
tu parå buddhir buddher
åtmå mahån paraª\
Ka.Up.
1.3.11
mahataª param
avyaktam avyaktåt
puru‚aª paraª| puru‚ån
na paraμ kiμcit så
k傆hå så parå gatiª\
3.41
tasmåt tvam indriyå~y
ådau niyamya
bharatar‚abha|
påpmånaμ prajahihy
enaμ
jñånavijñånanåçanam\
Åtman Interpolations
3.17
yas tv åtmaratir eva syåd
åtmat®ptaç ca månavaª|
åtmany eva ca saμtu‚†as
tasya kåryaμ na vidyate\
3.18
naiva tasya k®tenårtho
nåk®teneha kaç cana| na
cåsya sarvabh¨te‚u kaç cid
arthavyapåçrayaª\
3.42
indriyå~i parå~y åhur
indriyebhyaª paraμ manaª|
manasas tu parå buddhir yo
buddheª paratas tu saª\
3.43
evaμ buddheª paraμ
buddhvå
saμstabhyåtmånam
åtmanå| jahi çatruμ
mahåbåho kåmar¨paμ
duråsadam\
In order to understand this final example, recall that Jei (1986: 633)
originally posited a Yoga layer extending from 2.39 to 3.42. In the meantime, we have learned that the Yoga layer is interrupted at 2.54–60, 2.62–
70 (the prajña interpolation), 2.61 (the bhakti interpolation), 2.72 (the
second Yoga layer interpolation), 3.9–16 (the first yajña interpolation),
3.17–18 (the åtman interpolation), 3.22–24, 3.30–32 (the bhakti interpolation), and 3.35 (the svadharma interpolation). The first Yoga layer
resumes as a continuous layer from 3.36. However, according to Jei, it
only does so for five verses, for, in his view, we encounter a further set
of interpolations after 3.41. In his words, “st. 41 neatly continues the
The Bhagavad Gtå and the Historical-Critical Method / 245
preceding text [that is, 3.36–40], but 43 seems to duplicate 41” (Jei
1979: 552). Hence, we must posit a new interpolation here. The distinguishing mark of this new layer is its use of the terminology of åtman.
Thus, whereas “our yoga layer…[was] in its entirety consecrated to the
exposition of the role of the buddhi in the spiritual discipline of a warrior”
(552–53), this new layer identifies åtman as the central principle. Jei
notes that “ ‘åtman’ appears probably as a philosophical term in our
sequence only in stanzas II.55, III.17–18 and here. Elsewhere it is most
probably a pronoun” (553), but he rejects the first of these occurrences
on the grounds that verse 2.55 was already assigned to the prajña interpolation. This leaves us with four verses: 3.17–18 and 3.42–43. According
to Jei, the use of åtman terminology in these verses was likely “precipitated by II.55” (553). Evidently, a composer, noticing that the word åtman
had earlier appeared in a prajña interpolation in a Yoga layer embedded
in an epic poem after a Såμkhya passage that was shifting the buddhi
language of this layer, decided to take up åtman and make it the centerpiece of his philosophy. Indeed, Jei claimed that “the partial repetitions
(åtmany eva…†u‚†aª, åtmani/-am…åtmanå [that is, between 2.55 and 3.17
and 3.42]) betray[s]” that 2.55 was indeed the source of his terminology
(553). Jei acknowledged that “none of these hints is by itself a sufficient
reason to consider these stanzas as younger,” but he argued that “all of
them in conjunction with the repetition ‘(pra-) jahi påpmånam/çatrum…’
can hardly be accidental” (553). Furthermore, this was “all again mutually
corroborated by the considerations…concerning stanzas III.17–18” (553),
namely, that these verses might also, with profit, be considered interpolations as this had the advantage of bringing verses 3.9 and 3.19 closer
together.
Note that 3.18 does not actually have continuity repetitions that link it
with 3.17; Jei seems to have included it as part of the åtman interpolation merely because it continues the thought of the latter.84 Likewise,
3.42 has neither continuity repetitions with 3.17 and 3.18, nor duplication repetitions with 2.55. (It may be part of the first Yoga layer, since it
features the language of buddhi characteristic of that layer, or, again, it
might be a buddhi interpolation inserted by a composer inspired by the
appearance of this term in the Epic layer at 1.23 or, alternatively, in the
Såμkhya layer at 2.39.) It has a one-word duplication repetition with
3.41, but to make the thesis of duplication repetitions really plausible,
246 / Vishwa Adluri and Joydeep Bagchee
Jei has to go to the Ka†ha Upani‚ad to identify a suitable parallel.85
Thus, he now has sufficient continuity and duplication repetitions (at
least between 3.41 and 3.42 if not between 3.17 and 3.18) to make the
hypothesis of a single åtman interpolation at least plausible. The real
crux of Jei’s method here, however, is not the continuity or duplication
repetitions. As will have become clear from the discussion so far, these
criteria are only seldom and selectively applied. Further, in addition to all
the problems with the application of these criteria that we noted in the
preceding section, the attempt to identify interpolations within layers
using the criteria of repetition, specific words or phrases, and formal
markers generates entire new classes of problems:
(1) Interpolations, even when supposedly modeled on or in continuation of other interpolations, are not found where they were supposed to
be. This was most clear in the case of the second yajña interpolation,
which was found to be separated by nearly thirty verses from the first
yajña interpolation. Doubtless, a creative scholar will always be able to
fill in the discrepancy with any number of further “interpolations” that
successively displaced a given interpolation to its present position.
(2) The choice of which verse a given interpolation looks back to in
fashioning its duplication repetition is quite random. Thus, Jei asserts
that the repetition of çreyaª in verse 5.1 looks back to 3.2 of the previous
layer, but does not justify why 1.31c, 2.5b, 2.7c, 2.31d, 3.11d, 3.35a,
3.35d, 4.33a, 12.12a, 16.22d, and 18.47a could not just as well have been
the source of this verse’s use.
(3) Related with this point, there is the problem of arbitrariness when it
comes to specifying the logic of the selection. For instance, Jei claims
that the language of prajña in verses 2.54–68 was inspired by the use of
buddhi in 2.53, which it replaced. But why should it not just as well have
been inspired by buddhi in 2.39 (the Såμkhya layer); or 5.20 or 5.28 (the
second Yoga layer); or 6.9, 6.21, or 6.43 (the Upani‚adic layer); or 7.4,
7.10, 8.7, 10.4, 10.10, 10.12, 12.8, or 12.14 (the Bhakti layer)? Depending
on the verse we specify as the source, the logic for the composer replacing
this word with another will change and so will our account of what he
did and why.86
(4) Further, there is a peculiar flexibility in putting together duplication
repetitions. Thus, the çreyaª in verse 5.1 is supposedly borrowed from
3.2, but the saμnyåsaμ karma~åμ of the same verse is supposedly bor-
The Bhagavad Gtå and the Historical-Critical Method / 247
rowed from 4.41. Now if the point of fashioning a duplication repetition
is to suggest continuity with a previous layer or previous thought, then
how does it make sense for the composer of the interpolation to cherrypick elements of his repetition randomly from throughout the layer?
Surely, if the idea was to fashion an analogue to another verse, the composer would take as many elements as possible from this one verse. By
scattering his sources, he reduces the impact of the duplication. He also
makes the nonorganic, composite nature of the verses more evident.
(5) Jei groups all the verses between 2.54 and 2.70 that either feature
the word prajña or are related to it through the word stitha (that is, 2.54,
2.55, 2.56, 2.57, 2.58, 2.61, 2.67, 2.68) into a prajña interpolation (as
well as 2.59, 2.69, and 2.70, which either continue or are “attracted” by
other verses) but leaves behind other verses (2.63, 2.65, 2.66, and also
2.64; was the latter also “attracted” by some other verse?) that feature the
word buddhi as constituents of the original Yoga layer. But why should
these not just as well be part of the prajña interpolation and the use of
buddhi here a conscious effort to conform to the language of the first
Yoga layer? Or if Jei wishes to insist that the composer of the interpolation was absolutely committed to the language of prajña (intending it,
for instance, as a rival term to the concept of buddhi of some other school),
why should verses 2.63–66 not be interpolations into the prajña interpolation?87 Further, the verse need not be the ultimate semantic unit of
analysis. It is possible that only the word prajña was interpolated or that
after the prajña verses were inserted, further insertions were made.
(6) What is the evidence that the second yajña interpolation is reinterpreting the theory of the sacrifice of the first yajña interpolation, which
was sacrifice “in the original sense of the word” as Jei informs us (1979:
551)? Why should 4.23–34 not represent an organic development of the
idea of the original? Indeed, in a text whose overarching logic is to reinterpret all human action and, maximally, the entire universe sacrificially
(compare the vision of the viçvar¨pa in chapter 11), verses 4.23–34 represent a necessary link in the chain of expansion of the meaning of sacrifice. Likewise, there is no reason to dissociate 3.42–43 from 3.41 and
call them “åtman interpolations” (555). The two continue, gramatically,
logically, and semantically, the thought of 3.41 (as remarked in the preceding section, if Jei had wished to consider them part of the same
layer, he would have relabeled all their duplication repetitions as conti-
248 / Vishwa Adluri and Joydeep Bagchee
nuity repetitions and moved them into the same column under 3.41). The
only reason for doing so is the offending term åtman, which Jei thinks
is a sign of a new author who subscribes to a different school introducing
the terminology of his school into the text. As he states, “here is, may be
[sic], the proper place to sum up the main terminological implications of
these analyses. In the first yoga layer the word ‘åtman’ seems to be used
only as a pronoun. The terms ‘brahman’ and ‘yajña’ are not used at all.
Both appear only in the yajña insertions. The word ‘puru‚a’ means simply ‘man.’ The whole psychical sphere is mostly designated by the term
‘buddhi’ ” (553).
As the last two examples especially will have made clear, the grouping
of verses into interpolations and layers is not being carried out on the
basis of the criteria of continuity repetitions and duplication repetitions.
Already, in the last section, we noted that these criteria were a feint and
that the choice of verses deferred to other, less honest imperatives. It has
now become clear that the real basis for classification is actually a rather
banal understanding of “philosophical schools,” identified in terms of
their presumed terminology. In other words, Jei has gone through the
Bhagavad Gtå, removed all the verses that have words or concepts that
he thinks cannot have been part of the philosophical terminology of the
Yoga school (prajña, åtman, yajña, brahman, and puru‚a in a sense
other than “man”), and, in this way, he has tried to reconstruct what a
“pure” Yoga doctrine (as he imagines it) might have looked like. “What
remains after we have extrapolated [extracted? excised?] younger segments of our sequence, is more or less the original first yoga layer, as
preserved in the actual text of the Bhagavadgtå” (Jei 1979: 554). He
has not actually understood what the teaching of yoga in the Bhagavad
Gtå is (whether taken in its “original” form or any of its “extrapolated”
forms), nor does he possess an accurate idea of what the historical Yoga
school might have taught, nor has he made an effort to understand how a
yoga teacher might, employing the lexical resources of Sanskrit vocabulary (for example, prajña, åtman, yajña, brahman, puru‚a, and so on),
have tried to convey this. His entire procedure is reductive and circular.
Little wonder then that, at the end of this exercise, he states that “the first
yoga layer, as it revealed itself, teaches the buddhiyoga. It is a karmayoga which presupposes the knowledge of some elementary Såμkhya
conceptions” (554).
The Bhagavad Gtå and the Historical-Critical Method / 249
There are two problems with such an approach. First, not only is the
assignment of terms to a school and the identification of a school with
certain terms being carried out reciprocally,88 but there is also no argument for why the adherents of a school could not have extended themselves beyond their scholastic terminology to draw on language or
concepts in general use. Second, the attribution of verses to layers must
be consistent: either all references to a term (for example, buddhi) are
ruled axiomatically to belong to a school (for example, Yoga) or one
gives up the idea of using philosophical schools to identify layers in the
Bhagavad Gtå. The criterion cannot work as a criterion if one keeps
changing the rules. Finally, it is a sign of boorishness to reduce complex
texts to our neat little categories so that we can conveniently file away
the fragments in our mental filing cabinets. Have we really understood
the text when we classify a verse as either Såμkhya, proto-Såμkhya,
Epic Såμkhya, or classical Såμkhya, based on a priori and circular
conceptions of what these phases looked like? What is gained by these
classifications, given that they are usually made up as the scholar goes
along? Every school, when divested of the terms, concepts, and intellectual structures it works with and reduced to a one-point teaching, will
end up resembling that teaching. But, as Friedrich Nietzsche says, “if
someone hides something behind a bush, looks for it in the same place
and then finds it there, his seeking and finding is nothing much to boast
about” (1999: 147).
The Criteria of the Historical-Critical Method as Articles of Faith
Let us now review the results of the forgoing analysis of the criteria for
identifying layers in the Bhagavad Gtå before turning to a discussion of
the usefulness of the historical-critical method in the study of Indian traditions. In this article we focused on six criteria that have guided the use of
this method: (1) “theism,” (2) names and epithets of K®‚~a, (3) repetition,
(4) specific words or phrases, (5) formal markers, and (6) philosophical
schools. The results were uniformly negative. The first criterion relied
on the removal of all parts felt to be theistic in the sense of referring to
K®‚~a’s divinity based on an a priori thesis of the divinization of a folk
hero. It did not actually lead to insights into the textual history of the
poem. Neither did it lead to the devoutly wished-for confirmation of
250 / Vishwa Adluri and Joydeep Bagchee
K®‚~a’s progressive divinization. It merely produced a text conforming
to Bhargava’s expectations of it. The second criterion also relied on an a
priori thesis of the divinization of a folk hero, but it had to, additionally,
manipulate the evidence to make K®‚~a’s names fit the thesis. The third
criterion likewise suffered from circular reasoning: the classification of a
repetition (that is, as a continuity repetition or a duplication repetition)
invariably depended on the role a verse was assigned within an overarching scheme. The fourth and fifth criteria also turned out to be subjective
and subject to circular reasoning. Finally, the criterion of philosophical
schools turned out to rely on a priori ideas of what specific philosophical
schools could and could not have contributed to the text to determine the
terminology they could and could not have used. It does not offer a viable
means for identifying layers.
If the layers are not being identified on the basis of rigorous epistemological criteria, what then is driving these scholars’ adoption of the articles
of faith of higher criticism? Why is there a commitment to producing
scholarship that is reductive, circular, logically flawed, and yet makes a
great pretense of employing a scientific technical method? Why the constant insistence on the rigor of so-called philological methodology?89
(We say “so-called” because these methods have little to do with what is
ordinarily understood and practiced as philology in departments of
Classical Studies.90) This is where the problems raised in this article take
on a wider significance. What has been and continues to be presented as
“scientific” in contrast to traditional interpretations, as a universal method
in contrast to the particularity of traditional approaches, and as being a
rational and disinterested contemplation turns out to be otherwise when
subjected to closer scrutiny. Scientificity, objectivity, precision, and a
willingness to read from a variety of perspectives are laudable ideals. But
surely there is something wrong with a discipline that survives only
through the repeated citation of a small circle of authors conforming to
the dominant ideology91 and that, moreover, has to exclude all those
skeptical of its founding myths (for example, the Åryan epic, which, as
shown in The Nay Science [Adluri and Bagchee 2014b], remains the
basis of all German Gtå scholarship to the present day) to uphold its
hegemony.
It would, of course, be premature to draw conclusions regarding the
method based on the analyses of a handful of scholars. That is why, in
The Bhagavad Gtå and the Historical-Critical Method / 251
order to gain a proper understanding of how these articles of faith arose
and became canonical among German scholars, in The Nay Science we
traced the genealogy of Gtå scholarship back to both its textual context
in the Mahåbhårata and its historical context in the emergence of
Indology in Germany.92 As we demonstrated, the search for layers only
became germane to scholarship after the early nineteenth century. Prior
to this period, no reader of the Bhagavad Gtå had thought to look for
layers for the simple reason that the conditions that would make such an
interpretation plausible (assumption of multiple authorship, extended
composition, assumption of an Åryan epic, attribution of K®‚~a’s divinity
to a K®‚~a-worshiping sect, and so on) did not exist. It was only in the
wake of the pseudohistorical researches of Christian Lassen (1837 and
1847) and Adolf Holtzmann, Jr. (1892 and 1893) that it became meaningful to look for the presumed layers in the text. Even though there was no
reason to assume that there was an earlier, brief composition in which the
charioteer K®‚~a advised the hero Arjuna (in one version of Holtzmann’s
argument, they are replaced by Dro~a and Duryodhana93) to fight as
demanded by the knightly code of the Indo-Germanic warrior, the idea of
a hypermasculine epic narrative devoid of ethical dilemmas and complex
philosophical doctrines had tremendous appeal in post-Reformation
Germany. German scholars blamed the loss of German autonomy on the
sophistry and scheming of Catholic priests and regarded Martin Luther as
the emancipator and savior of the free-thinking and heroic Germans (see
Adluri and Bagchee 2014b: 112–21). Even though the project of identifying the so-called Epic layer of the Bhagavad Gtå thus had relevance
only to those invested in the Reformation narrative, scholars from around
the globe joined in the project due to the institutional dominance of
German Indology.94 The appropriation of the term critical to characterize
their researches, as opposed to the tradition, now universally repudiated
as “precritical” and “unenlightened,” played a major role in why scholars
took up this wild-goose chase. Who, after all, would wish to be labeled
“precritical”? It was better to keep faith with the Germans.
A Role among the Indians Similar to Martin Luther’s
Translation of the Bible?95
Historical criticism functions as an instrument of social critique in Indology
252 / Vishwa Adluri and Joydeep Bagchee
by positing desirable pristine origins and enjoining the search for these
origins upon scholars (Figueira 2002). This search is to be undertaken
through “purifying” whatever does not belong to the fantasy original (that
is, whatever is deemed non-Åryan, non-Vedic, nonheroic, Bråhma~ic,
religious, and/or philosophical). The application of this method to the
Bhagavad Gtå did not lead to a better understanding of the text, nor even
of the historical processes that produced it (ignoring, of course, the many
fabricated histories that have been posited for it, for example, a svadharma interpolator who adds the word svadharma, a prajña interpolator
with a similar attachment to the word prajña, and so on).96 But the larger
question here does not concern the validity of this method in Gtå Studies
(indeed, if the analyses generated in the name of historical criticism have
taught us anything it is how not to approach the text), but its usefulness
in general. As we have indicated, the method only appears worthy of
emulation as long as one does not look more closely at the scholarship it
generates and does not place this scholarship within its proper historical
hermeneutic horizon. Thus, we now have scholars who seek an interpolator for each new term and a “science” that proceeds entirely by means
of metaphors (verses that are “attracted” or “precipitated” by others) but
asserts that it represents the claims of objectivity tout court. What might
some of the alternatives be? As a way of engaging with the defenders of
the method and, also, as a way of considering the issue from all sides, we
consider ten potential objections to our conclusion:97
(1) The method is right; particular hypotheses may have failed, but
better hypotheses might still emerge. While we cannot rule out in principle
that better hypotheses might emerge, we cannot ignore the fact that the
hypotheses that emerged in the past century did not emerge as the result
of “stumbling” through the text. Rather, they were the result of an intellectual and institutional context that demanded a certain type of scholarship, conformed to a certain type of ideology, and answered to specific
scholarly prerogatives. Better interpretations will emerge only to the extent
that we interrogate how and why the current hypotheses arose, so as to
become aware of why we, as scholars, do what we do.
(2) The method is right; Jei’s application of the hypothesis is flawed,
but the idea is intriguing and other scholars may yet discover objective
features about repetitions. The problems with Jei’s hypothesis were
not limited to his application of it. Rather, the very idea that repetitions
The Bhagavad Gtå and the Historical-Critical Method / 253
could be distinguished depending on whether they “continue” or “duplicate” a layer turned out to be wrong. Repetitions can be distinguished
only if we already know which layer they continue or duplicate and the
manner in which they do so. The criterion is hence superfluous. It is, in
theory, always possible that the method might be refined. However, current attempts such as those by Przemyslaw Szczurek (2002, 2003, 2005–
2006, 2007, 2008) and Ivan Andrijani (2013) are not promising. Further,
since the hypothesis, besides its methodological shortcomings, also relies
on the dogmas of Christian Lassen, the two Holtzmanns, Richard von
Garbe, Hermann Jacobi, Hermann Oldenberg, Rudolf Otto, and others, it
is hard to see how any future attempt could be more convincing.
(3) The method is right; particular hypotheses may have failed, but the
idea of layers and interpolations should not be dismissed altogether. There
is no reason to assume the existence of layers in the Gtå. The idea of
layers is a fiction, whose historical origins can be traced back fairly accurately to the work of Holtzmann, Jr. (1893), specifically to his thesis that
the Gtå originally existed as a heroic, epic poem describing only the war
situation.98 As the common background of all their work,99 this prejudice
was the real reason that the Indologists found the idea of layers in each
other’s work so compelling.100 In contrast, we do not reject the use of
interpolation as an interpretive category. Indeed, native commentators
were aware of the category and made occasional use of it.101 However,
they were aware of the fact that interpolation is an interpretive category
as well as of the fact that it should be resorted to only as a last resort (for
example, when all other attempts to make sense of a verse have failed). It
should not be used as the primary category, and above all, it cannot and
should not replace the interpretive task. Finally, interpolations need to be
identified on the basis of manuscript evidence, by conducting a recensio
as V.S. Sukthankar did in the case of the Mahåbhårata critical edition. The
category “interpolation” should not be abused to justify a narrow, dogmatic view of the text—whether Lutheran solafideanism or Holtzmannian
pan-Germanism—by excluding other voices and themes.102
(4) The method is right, but it is intended to generate only plausibility
not certitude. Absolute certitude is an impossible epistemic standard in
the humanities, and neither are we asking for it. But when the practitioners of a method claim greater objectivity and greater scientificity than
native commentators (Stietencron 1996: 6–7; Hanneder 2001: 240;
254 / Vishwa Adluri and Joydeep Bagchee
Malinar 2007: 17; see also Jei 1986: 628–29 and Szczurek 2002: 56,
acknowledging the scientific primacy of the German school), they should
expect to be held to that standard. Further, even taking the criterion of
plausibility as our standard, historical-critical research has failed to meet
the mark. What is more plausible—that an author had a complex philosophical argument and used several terms to explore the issue from different
sides, or that every new word in the poem is evidence of an interpolator
who campaigned for just that term?103
(5) The method is right; progress has been made and some historical
insights achieved, even though no absolutely convincing scheme or set of
criteria that convinces all rational observers has been proposed. If one
examines the figure of Gtå interpretations on pages 259–61 of this article,
it will be clear that, in fact, no progress has been made. The application
of the historical-critical method to the Bhagavad Gtå has led neither to
the emergence of a scientific consensus, nor to a body of cumulative
evidence. Indeed, the only reason these theories appeared plausible to
scholars is because they had all accepted the thesis of a heroic epic poem
riddled with bhakti and Bråhma~ic interpolations.104 Even considered as
a response to the documentary impulse (Sheehan 2005: 102), historicalcritical scholarship on the Gtå has not enhanced our knowledge about
the history of the text. The so-called historical insights have rested on a
priori accounts,105 proposed on the basis of circular and self-referential
criteria.106
(6) The method is right, but it needs to be supplemented and balanced
by context-sensitive analysis of the Bhagavad Gtå. The aims of the
historical-critical method are antithetical to those of a philosophical
reading.107 Historical criticism emerged out of debates concerning the
interpretation of scripture in Protestantism, and it remains beholden to
the principles of the latter.108 Its intent is to, in the name of respecting
the historical context, make it impossible to read the text as a coherent
sequence, to read verses or doctrines or themes as being part of a literary
context, indeed, as even belonging to the same religion or theology
(Levenson 1993: 4–5, 100). We are forced to reject the method not only
in the interests of retaining an access to the text now only available to us
through the native commentarial tradition, but also because we recognize
that there is no access to the text that would not be conditioned by the
implicit theological and religious commitments of the scholar.
The Bhagavad Gtå and the Historical-Critical Method / 255
(7) Historical criticism is a valid form of scholarship, but it must in
future contend with a pluralistic tradition of commentary. Until now, the
practitioners of the historical-critical method in Indology have refused to
acknowledge that they too have only a partial grasp of the truth, that they
too are the products of specific historical conditions (Hanneder 2011:
124–25). Indeed, they have opposed the alleged universality and objectivity of their method to the—once again, alleged—particularity and
subjectivity of the Indian tradition (Hanneder 2001, 2010). Under the
pretense of excluding religious and theological perspectives from discourse, they have closed themselves to the hermeneutic encounter.109
Every interpreter of the Gtå in the past has been aware of the history of
interpretation. It is only the Indologist who in bad faith claims that, as his
work unearths a rigorously original meaning, he does not belong to the
history of interpretation; indeed, he is its end.
(8) Historical criticism is more widely applied than in Indology alone
(for example, in biblical criticism). Historical criticism has also come
under attack in biblical criticism, its original field of application, where
similar sorts of criticisms have been raised of it as those we have raised
here. As noted earlier (see note 108 on page 288), there are many who
reject historical criticism altogether as being a form of Protestant religious
experience. Historical criticism supports Protestantism by making the
literal meaning of the text decisive, thus undermining allegorical, typological, moral, and anagogical approaches to the text.110 There is now an
entire movement within biblical criticism that calls itself “postcritical
biblical interpretation.” “Postcritical biblical interpretation does not subscribe to the view, characteristic of much historical criticism, that ‘the
most primitive meaning of the text is its only valid meaning, and the
historical-critical method is the only key which can unlock it’ ” (Soulen
and Soulen 2011: 157, citing Steinmetz 2011: 4).111 The historical-critical
method has also come under attack in postmodern biblical interpretation. The latter criticizes modern biblical criticism for what it regards as
its “problematic or unsustainable premises.…These premises…include
(1) the view that biblical texts are artifacts that have a single, stable
meaning; (2) that a text’s meaning, though initially hidden from the
modern interpreter by temporal and cultural distances, can be recovered
by historical reconstruction; and (3) that the benefits of critical methods
accrue over time as methods become more sophisticated and data increase”
256 / Vishwa Adluri and Joydeep Bagchee
(Soulen and Soulen 2011: 158). All three have relevant parallels in Indology. Postmodern biblical interpretation also forces us to recognize that
“no single…community can claim to enjoy privileged knowledge of the
text” (159). “It challenges what it regards as modernity’s unwarranted
‘hegemony of the expert’ that has served to discount equally valid but
differently formulated interpretations” (159). By inducting marginalized
and underrepresented groups, it exposed that what presents itself as a
universal and all-encompassing view formulated in the name of a common
reason is in fact a partial and by no means disinterested perspective and,
simultaneously, made it possible for a number of other perspectives to be
articulated (feminist biblical interpretation, Afrocentric biblical interpretation, Asian biblical interpretation, and so on). Finally, in postcolonial
biblical criticism, scholars have brought to light how “Christian missionaries, official representatives of the colonizing powers, and colonized
peoples themselves used the Bible in ways that legitimated colonialism”
(156). They also challenged “modern western biblical criticism, whose
pretensions to universal validity are viewed as complicit with imperial
expansion and colonial rule” (156; emphasis in original). Those who
argue for the validity of the historical-critical method in Indology by
citing the example of its use in biblical criticism should be aware that this
field too has undergone a revolution in the meantime and that they speak
of a biblical criticism that has ceased to exist. If historical criticism has
been challenged even in biblical criticism from all these perspectives—
for its Protestantism, its problematic epistemological premises, its exclusion of other voices, its claim to speak in the name of all when in fact it
advocated the interests of a powerful few, its support of colonialism—
how then can we avoid asking the same questions of its use in Indology?
Can we really maintain that these issues are less relevant in Indology?
(9) Historical criticism as a method needs to be separated from historical criticism as an ideology. As a growing body of literature recognizes, the scientific pretensions of German Indology need to be located in
a broader discourse of power relations that denigrated the authority of
“native” commentarial traditions as hopelessly sectarian and benighted
(Bagchi 1996, 2003; Figueira 2002; McGetchin 2009; Yelle 2013; see
also Adluri and Bagchee 2014a).112 Unfortunately, the problem is not as
simple as separating the effects of historical criticism (as ideology, as
critique, as orientalism, as anti-Semitism, and as supersessionism) from
The Bhagavad Gtå and the Historical-Critical Method / 257
the method. As we showed in The Nay Science, the denigration of the
tradition is a sine qua non for the emergence of Indology. Indeed, because
German Indologists were not able to make a case for themselves aside
from the fact that they were not the tradition, “concrete interpretation of
Indian texts was indissolubly linked to the task of a critique of tradition”
(Adluri and Bagchee 2014b: 343). “Indology devoid of its interventionist,
aufklärerische, and simultaneously restorative and reformatory concerns
would be dissolved into the discipline of history” (388). Further, as we
demonstrated in the conclusion, historical criticism, “allegedly theologically neutral, turned out to have a major role to play in the delegitimization of all alternative sources of intellectual authority—Jewish, Catholic,
Greek or Indian” (434–35). Thus, we reject, with good reason, the suggestion that historical criticism can be practiced or, indeed, has a function
without its antitraditional, antiphilosophic biases.113 Our reconstructions
of Bhargava and Jei in this article demonstrated that the only function
of these flawed and circular analyses was to deny the literary context of
the Bhagavad Gtå—at the level of the book, the chapter, and even the
verse. It is time to recognize that historical criticism corresponds not only
to a community of interpreters, as Jon Levenson argues (1993: 123), but
also to a community of believers. The problem of what to do when we do
not believe in their “sacred drama of salvation being carried out by a
chosen Indo-European people” (Benes 2008: 110) is unavoidable.114
(10) The article dismantles erroneous theories of the origins of the
Bhagavad Gtå, but does not present a positive vision for future scholarship. This article’s main concern was to clear away misperceptions of the
text, especially a powerful myth regarding the Mahåbhårata’s origins in
a heroic, war epic. As we showed, it is this myth rather than any scientifically reliable analyses that explains the Gtå critics’ analysis. Indeed,
we found the continuing hold of this myth on the scholarly imagination
to be so strong that Bhargava and Jei preferred to finesse their criteria
(or the application of these criteria) rather than produce results that contravened this myth. The reason we did not articulate a positive vision for
future scholarship is twofold: (1) it is too early to do so until the problematic ground of existing theories of the Bhagavad Gtå is cleared up;
and (2) we are not interested in claiming that any particular approach is
the one true one. Our main interest in this article was to open up the field
of interpretations, which has been narrowed to the disadvantage of phi-
258 / Vishwa Adluri and Joydeep Bagchee
losophical interpretations.115 As we showed, as the dominant ideology
that shaped much of the study of the epic over the past two centuries,
historical criticism narrowed down the field of interpretation and led
scholarship down the blind alley of Protestant literalism. It was in order
to undo this narrowing of the field and to show that the constraints on an
independent and conscientious reading of the text today arise not from
Bråhma~ism but from the Neo-Bråhma~ism of the “experts” that we
undertook this critique. (In this respect, we are the more authentic inheritors of Martin Luther’s Reformation than the Indologists themselves, who
have activated the narrative of liberation of the mind from dogma only
in order to consolidate authority over the texts in their own hands.116)
P.L. Bhargava’s and Mislav Jei’s work is testimony to the fact that,
even today, human statutes count for more than the Word. As the work
of Madeleine Biardeau, Don Handelman, Alf Hiltebeitel, David Dean
Shulman, Bruce M. Sullivan, and others shows, historical concerns can
and must be combined with engaged and intelligent reading of the texts.117
The Bhagavad Gtå and the Historical-Critical Method / 259
260 / Vishwa Adluri and Joydeep Bagchee
The Bhagavad Gtå and the Historical-Critical Method / 261
262 / Vishwa Adluri and Joydeep Bagchee
Notes
1. The historical-critical method encompasses a variety of approaches,
and perhaps the narrower “text-historical method” (textgeschichtliche
Methode) would have been preferable here. However, this term poses
problems of its own: although it accurately describes the narrowest part
of what Indologists were doing, the term itself was not used in Indology
until Hacker’s 1961 article (for a discussion of Hacker and his apologetic
concerns as they relate to the study of India, see Bagchee and Adluri 2014).
The term is also misleading, because it suggests a parallel to what is
understood as “textual history” (Textgeschichte) in textual criticism, even
though it does not refer to an objective history of the text (constructed, for
instance, on the basis of what we know of the production and dissemination of texts, the practices of scribes, and so on), but to an a priori history
constructed out of the text. For the same reason, we avoid the term “text
critical” (textkritisch) when speaking of the Indologists’ work: the latter
rarely understood or applied the principles of textual criticism, even
though they adorned their work with frequent references to Textkritik,
or textkritischer Befund, or textkritische Perspektive (also textkritische
Untersuchung) (see von Simson 1969 and Malinar 1996; Bigger 1998 is
another example of the inflation in the use of this term). All translations
are ours unless otherwise noted.
2. Historically, ten criteria have been proposed: (1) epic elements or epic
style, (2) epic situation, (3) “pantheism,” (4) the numinous, (5) “theism,”
(6) names and epithets of K®‚~a, (7) repetition, (8) specific words or
phrases, (9) formal markers, and (10) philosophical schools. In this article
we focus on criteria 5 to 10; criteria 1 to 4 have already been covered in
The Nay Science: A History of German Indology (Adluri and Bagchee
2014b). One further criterion—namely, that of statistics, applied mainly
in the work of Smith (1968) and Yardi (1977–78)—may be worth examining, but at the time of writing the former article was unavailable to us.
Note, however, that statistical variation cannot actually show that different
sections have been composed at different moments in time because an
author is capable of modulating his style, especially to fit the contents of
different sections or to make different points. Yardi does not find significant differences in style between the adhyåyas of the Bhagavad Gtå
(see 1977–78: 1051: he rejects the thesis of multiple authorship), whereas
The Bhagavad Gtå and the Historical-Critical Method / 263
the claim that the Bhagavad Gtå was originally not part of the Mahåbhårata requires further study (1052).
3. This article uses inverted commas around “layers” to express its skepticism regarding both the existence of any such objective feature and the
heuristic value of any such distinction. Hereafter, the inverted commas
will not be repeated, but are to be understood.
4. Stietencron: “historical and philological methodology”; Hanneder:
“philological method”; and Malinar: “text-historical” (all three belabor
the contrast with traditional commentary).
5. Supersessionism (also known as “replacement theology”) is the view
that salvation in Christ replaces or supersedes God’s covenant with the
Jewish people and was a significant factor in the rise of historical biblical
criticism (see, for example, the work of Johann Salomo Semler or Ferdinand Christian Baur, where it is explicitly acknowledged as a motivation;
the new historical criticism is both the finest fruit of supersessionism and
a contributor to it). In the present context, we mean both its use against
the Hebrew tradition and against the Indian commentarial tradition. In the
latter case, it primarily took the form of a methodological supersessionism
(though evangelical overtones were not absent) that was aided by the
narrative that self-reflexivity and critical consciousness were both discoveries of the Enlightenment (see Schmitt 2012).
6. See Weber 1850a and 1855, important source texts for this view (the
relevant passage is cited in note 95 below); and Barnes 2003: 347–48,
discussing the Protestant origins of the idea of world history as a process
of education of humanity (the phrase, of course, is a play on Gotthold
Ephraim Lessing’s Die Erziehung des Menschengeschlechts, or The
Education of the Human Race).
7. Rogerson lists these as “the freedom to investigate questions of authorship of books, unity of books, and sources underlying books, without the
restraints imposed by traditional opinions on these matters deriving from
narrow views of the nature of inspiration” (1985: 27).
8. This history, as discussed elsewhere, had two components: it was,
first and foremost, an anthropological history, based on the “biracial”
theory of Indian history (Arvidsson 2006: 45), but it was also an eschatological history, that is, a history “embodying a concrete vision of what
India could be if it but followed the German scholar’s missives” (Adluri
and Bagchee 2014b: 392).
264 / Vishwa Adluri and Joydeep Bagchee
9. The Mahåbhårata, for instance, views history cyclically; it structures
time according to the four yugas (cosmic ages), where the characteristic
feature of these yugas is not temporal succession but increase of entropy of
dharma. Creation itself is seen only as a mimetic event, so that “history”
ultimately as a category is not real; the cosmos is the divine play, or kra
of Nåråya~a (see Mahåbhårata 12.336.56 and 12.339.20).
10. In Indology, the method was more often referred to as the “texthistorical method” (textgeschichtliche Methode), following Hacker’s
coinage and his valorization of this approach (see note 1 above). In his
view, scholars were to use “changes” in the text, such as “inversions of
the text, expansions, interpolations, and even individual word variants,”
to identify the historical evolution of the text and, mutatis mutandis, of
religious and intellectual traditions in India (“from such changes…one can
at times practically read off intellectual-historical [geistesgeschichtliche]
processes” [Hacker 1961: 489]). Hacker evidently did not see the problem
with using assumed changes to support assumed developments and even
more so to create a fabricated history of India, likely because the method
tied in so well with his fundamentalist evangelical concerns in reading
texts (see Bagchee and Adluri 2014). What is less understandable is why
scholars who were, at least nominally, secular also adopted Hacker’s
method.
11. After Holtzmann, Jr. (1893), who first claimed that the Bhagavad
Gtå originally consisted of a brief exchange before the battle (153–54)
but was expanded to include sections on pantheism and on yoga (154), a
number of interpretations and/or reconstructions followed, all of them
implicitly drawing on the former’s work in some way. Thus, Garbe (1905:
16) proposed that the original Gtå, contrary to Holtzmann’s views, would
have consisted of a brief theistic poem. Schrader (1910: 340) concurred
with Garbe’s view of a theistic revision of a pantheistic original but proposed that one could go even further in identifying the Bhagavad Gtå’s
earliest layers. He proposed that originally the poem would have ended
with verse 2.38, but probably a number of verses composed in a similar
style were added to this poem before the “Bhågavatas” took it over and
used it as a foundation for their composition. This suggestion was later
taken up by Jacobi (1918: 325–26) to produce an epic Gtå comprising
verses 1.1–47, 2.1–6, 2.9–12, 2.18, 2.25–27, 2.30–37, and 18.73. Oldenberg (1919: 334) expressed approval of Schrader’s view but thought that
The Bhagavad Gtå and the Historical-Critical Method / 265
the term Bhagavad Gtå should be reserved for the expanded “K®‚~aite”
poem that extended from 1.1–2.37 and 2.39–72 (except 2.26–27 which
he though were possibly interpolated).
12. Examples include Jei 1979, 1986, 2002, 2009a, 2009b, 2010;
Malinar 1996, 2007; and Szczurek 2002, 2003, 2005–2006, 2007, 2008.
Brockington (1997) also supports the “layering” of the Gtå but does not
offer an independent set of arguments for it (his article summarizes and
restates the views of Hermann Jacobi, Rudolf Otto, R. Morton Smith,
Mislav Jei, and Georg von Simson).
13. Jei published his first paper in 1979; but the paper published in
1986 is actually older. As Jei 1979 refers to and relies on the 1986
paper’s conclusions, we take up the latter first.
14. The emphasis on a historical K®‚~a draws heavily on this identification, even though Hiltebeitel has noted that “one cannot, of course, be
sure that this Upani‚adic passage refers to the same K®‚~a Devakputra”
(1979: 80n37; see also Olivelle 1996: 410). The argument is problematic
inasmuch as the “historical” K®‚~a Devakputra of the Chåndogya Upani‚ad is cited as evidence not only for the human origins of K®‚~a in the
Bhagavad Gtå, but also for his warrior origins (see Charpentier 1930:
124: “the Chåndogya Upani‚ad tells us about a certain K®‚~a Devakputra—and there is to me not the slightest doubt that he is identical with
the K®‚~a of the Great Epic—who was no doubt a k‚attriya [sic] and who
was the pupil of Ghora Ågirasa”). This is problematic because Chåndogya Upani‚ad 3.17.6, like the ¸g Vedic references (1.116.23, 1.117.7,
8.74.3; K®‚~a is the name of a ®‚i who is devoted to sacrificing soma to
the Nåsatyas), refers to a sacrificial context closely associated with the
soma. (Olivelle glosses the verse as: “here various aspects of living are
equated with the central elements of a Soma sacrifice: a man is consecrated [dkså] prior to undertaking a sacrifice; various preparatory rites
[upasad] are performed daily between the day of consecration and the
day of Soma pressing” [1996: 338]). There seems to be no support for a
heroic warrior K®‚~a if we look to the Chåndogya Upani‚ad.
15. Bhargava fails to notice that the argument entails a petitio principii:
K®‚~a’s status is precisely what is in question here. He also resorts to an
argument ad baculum: “It is certain that K®‚~a was originally a human
teacher who was later deified” (Bhargava 1977: 357; emphasis added).
As evidence for K®‚~a’s deification, he cited the Chåndogya Upani‚ad,
266 / Vishwa Adluri and Joydeep Bagchee
which “mentions K®‚~a, son of Devak, as a disciple of the sage Ghora
Ågirasa without any claim to divinity” (357), without noticing, once
again, that he had been guilty of a petitio principii. Incidentally, the
teaching ascribed to Ågirasa contains a metaphor that uses the yajña as
an organizing principle, and the salvation promised is not ‘going to heaven’
but the final soteriological goal of ‘entering into the sun.’ This works
against Charpentier’s claim that the K®‚~a Devakputra of Chåndogya
Upani‚ad is the same person as K®‚~a in the Bhagavad Gtå, since he
claims of the latter that he espouses (in verse 2.37) “the ideals of a chivalrous class and period, in a way strikingly like those of the Scandinavian
Viking time when the brave man did either win power and riches or go,
sword in hand, to the very material paradise of Valhall” (1930: 80).
16. Garbe’s method, if it can be called that, consisted in the removal of
verses of the Bhagavad Gtå he found at odds with his cherished thesis of
a “theistic” original with “pantheistic” interpolations. He adopted as his
guide any language indicative of an identity between God and the creation
or between K®‚~a and Brahman (Being), both of which aroused his suspicion of “pantheism,” and, in one of the most free-wheeling intellectual
enterprises to date, argued that the removal of all such passages proved
that the Gtå was originally theistic.
17. The argument is not quite clear. Perhaps Bhargava means that there
was no necessity for interpolation, since the first chapter merely set the
stage for the philosophical discussions that were to follow, but why he
thinks there should be no room for interpolation is unclear: surely philosophic interpolations could have been made here just as well as in the
second chapter, where they also begin amidst Arjuna’s despondency.
18. See Jei 1986: 634; Jei shares Bhargava’s suspicion that these
verses have been interpolated because they refer to K®‚~a as the standard.
19. Verses 1–15, rejected on the grounds that they “have nothing to do
with the subject-matter of this chapter [namely, the yoga of wisdom] and
have been prefixed to it only to establish the divinity of K®‚~a. They are
indubitably interpolated” (Bhargava 1977: 358).
20. Verse 29, rejected because it “affirm[s] the divinity of K®‚~a” and
because “the way to liberation has already been explained in the preceding
verse” so that this verse, “which emphasizes the knowledge of K®‚~a’s
divinity as the way for attaining Peace,” is “clearly interpolated” (Bhargava
1977: 358).
The Bhagavad Gtå and the Historical-Critical Method / 267
21. Verses 13–15, 30–31, and 47. Bhargava’s arguments for considering
them interpolations are as follows: whereas verse 12 discusses the topic
of yoga without reference to K®‚~a, “the following three verses say the
same thing with emphasis on the divinity of K®‚~a. They cannot therefore
but be interpolated” (1977: 358). Likewise, verse 29 defines the yogin
without reference to K®‚~a as does verse 32, but verses 30–31 “speak of
the yogin as one who sees K®‚~a everywhere and worships him. These
verses are therefore clearly out of context” (358). Finally, he also argued
that verse 46 “was undoubtedly the last verse of this chapter at one time
because K®‚~a here sums up his arguments why Arjuna should become a
yogin,” and hence there was no need for a further verse (that is, 47),
“which lays emphasis on faith in K®‚~a” (358–59).
22. Bhargava’s argument for considering these six chapters interpolations was as follows: “This chapter [that is, chapter 13] was undoubtedly
at one time the seventh chapter of the Bhagavadgtå before the present
chap. VII to XII were put as a wedge between it and the sixth chapter”
(1977: 359); see also the next section of the present article for further
arguments for these chapters being interpolated.
23. Verses 3, 11, and 19, rejected on the grounds that they were inserted
with the express intent of emphasizing K®‚~a’s status as a divinity. As
the arguments make for amusing reading, they are quoted here in full:
“Since the Knower of the Field has already been defined in the previous
verse [that is, 13.2], his identification with K®‚~a in v. 3 is the thought of
a later writer who wanted K®‚~a to be regarded as the omnipresent God”;
“the second interpolated verse is v. 11. K®‚~a has recounted the various
qualities which constitute wisdom in vv. 8–10 and someone who thought
that devotion to K®‚~a should also be included among these qualities has
deliberately interpolated this verse here”; “the last interpolated verse is
v. 19 which is clearly redundant and has been inserted here only to emphasize the importance of devotion to K®‚~a” (Bhargava 1977: 359).
24. Verses 6–15 and 18–19. According to Bhargava, chapter 15 was
“originally a very short chapter intended to explain the difference between
the soul that suffers constant changes,…the soul that attains the highest
goal,…and the Supreme Soul” (1977: 359). It led from verse 5 (description of the highest goal) to verse 16 (distinction between those who have
attained the highest goal from those who have not) to verse 17 (description of the highest Self). This, in his opinion, made it “clear that the inter-
268 / Vishwa Adluri and Joydeep Bagchee
vening vv. 6–15…[were] interpolated” (359). Likewise, he considered
that after the elaboration of the Paramåtman in verse 17, “no further elucidation was necessary” (360). But some “person who wanted to establish
the divine nature of K®‚~a added two more verses (18 and 19) in which
K®‚~a claims that he himself is the Highest Self mentioned in v. 17” (360).
25. “In none of the chapters is the contradiction between the genuine and
interpolated verses so glaring as in chap. XVIII” (Bhargava 1977: 360).
26. The following is a complete list of Bhargava’s excisions. The excisions for each chapter are listed individually with the criterion applied in
parentheses next to it (an asterisk indicates that the criterion is introduced
for the first time): chapter 1 (no excisions); chapter 2, verse 61 (criterion
1*); chapter 3, verses 22–24 (criterion 2*) and 30–32 (criterion 3*); chapter
4, verses 1–15 (criteria 2–3); chapter 5, verse 29 (criterion 1); chapter 6,
verses 13–15, 30–31, and 47 (criterion 1); chapter 13, verses 3 (criterion
1), and 11 and 19 (criteria 1–3); chapter 14, verses 2–4 (criteria 4* and
5*), 19 (criteria 1–4), and 26–27 (criteria 1–4); chapter 15, verses 6–15
(criteria 1–3) and 18–19 (criteria 1–4); chapter 16, verses 17–20 (criteria
1–4 and 6*); chapter 17, verses 5–6 (criteria 1 and 7*); chapter 18, verses
54–58 (criteria 2 and 6), 64 (criteria 8* and 9*), 65 (*10), 66 (criteria 8–
10), 67–71 (criterion 11*), and 77 (criterion 4). Chapters 7–12 are excluded
in toto as being “theistic,” so no specific criteria apply.
27. See Hiltebeitel (1979: 80n87) for a list of citations. Hiltebeitel’s
earliest reference is to Max Müller (1879), but the theory is also found in
Weber (1850b: 190), who, however, ascribes it to an unnamed text by
Henry Thomas Colebrooke. Bagchee (2011c) discusses further examples.
28. Looking forward, the idea is also absorbed by Jei: “On the other
hand, the well-known passage of Chåndogya-upani‚ad 3.17 mentions a
K®‚~a, son of Devak, as a pupil of Ghora Ågirasa, and confronts us
with the long-debated question whether it is the same person as K®‚~a
who teaches Arjuna in the Gtå.…In spite of differences between ChU
3.17 and the BhG, it is hard to reject or neglect the precious testimony of
the ChU as a historic parallel in reconstructing the genesis of the character
of K®‚~a as teacher in the BhG. The burden of proof lies with those who
reject the identity of the ‘two K®‚~as’ ” (2009a: 221–22).
29. The divinity of K®‚~a has posed a problem for almost every German
Indologist since Adolf Holtzmann, Jr. As supreme God, absolute Being,
an incarnation of God on earth, and God in a different and higher sense
The Bhagavad Gtå and the Historical-Critical Method / 269
than the nature gods of the earlier Vedic period, his depiction in the epic
was a source of theological anxiety for many scholars, who, while sharing
an exclusivist vision of Christian faith, “would never have admitted that
their ‘literary’ studies of the epic were laden with religious and theological
presuppositions” (Hiltebeitel 1979: 65–66). Nonetheless, it bears repeating
that “about the cosmic character or Çr K®‚~a, the epic itself is not in any
doubt. He is an Ûçvara. He is the Puru‚a of the Såmkhyas; the Brahman,
the Åtman or the Paramåtman of the Vedantins. The man of knowledge
affirms that Våsudeva is All (Våsudevaª sarvam iti, Gtå 7. 19).…There
is…not a single passage in the Mahåbhårata which does not presuppose
the divinity or the cosmic character of Çr K®‚~a” (Sukthankar 1957: 63;
emphasis in original). The attempt to remove K®‚~a from the epic is as
pointless as it is disingenuous.
30. Actually, Bhargava (1977: 360–61) overstates the contrast. The
expression he translates as “flee” is çara~aμ (gaccha) (18.62), and the
expression he translates as “be my devotee” is (mad)bhakto (18.65). Both
çara~am and bhakti (the former better translated as “taking shelter” or
“taking refuge”) are standard words of bhakti terminology (indeed, in the
later Tamil Vai‚~ava tradition as well as Sanskrit texts such as Lak‚m
Tantra, one of the standard terms to describe this way of approaching God
is çara~ågati or prapatti). Bhargava also does not notice that in verse 66,
one of the verses supposedly in contradiction with 18.61–62, the term
çara~am repeats. Only a Jeiian theory of a çara~am interpolator responsible for the çara~am verses (see later) could have saved him here.
31. This is the same criterion as Richard von Garbe’s and Moriz Winternitz’s “no gap may arise” (see Adluri and Bagchee 2014b: 182n104). It
suffers from the same fault: ativyåptido‚a, or being overly generic. Theoretically, the Bhagavad Gtå could be reduced to any two well-connected
verses.
32. K®‚~a, Keçava, and Govinda.
33. Mådhava, Vår‚~eya, and Våsudeva.
34. Janårdana, Madhus¨dana, and Keçini‚¨dana.
35. H®‚keça and Acyuta.
36. Yogin, Yogeçvara, Mahåyogeçvara, Aris¨dana, Mahåbåhu, and
Prabhu.
37. Puru‚ottama, Çåçvata Divya Puru‚a, Sanåtana Puru‚a, and Purå~a
Puru‚a.
270 / Vishwa Adluri and Joydeep Bagchee
38. Para Brahma, Para Dhåma, Parama Pavitra, and Parameçvara.
39. Deva, Ådi Deva, Devadeva, Deveça, and Devavara.
40. Bh¨tabhåvana and Bh¨teça.
41. Jagatpati and Jagannivåsa.
42. Viçveçvara and Viçvar¨pa.
43. Ananta, Anantar¨pa, Aja, Vibhu, and Vi‚~u.
44. Bhargava’s scheme compares favorably with Jarl Charpentier’s.
Before Bhargava, Charpentier had already claimed that “there is in the
allocutions of Arjuna no hint of the supremely divine nature of K®‚~a
until we arrive at the puru‚ottama in viii, 1. And it is only in the cantos x
and xi that Arjuna raises himself to a language of the purest bhakti by
using epithets like devadeva, viçveçvara, Vi‚~o, etc.” (1930: 105). From
this he concluded: “There is not the slightest doubt that this rising scale
has been conscientiously aimed at by the author of the earlier Bhagavadgtå” (105). The reason for preferring Bhargava’s scheme was to make
the point that it was not just European scholars who followed in the wake
of a German tradition of pseudocritical scholarship. Indian scholars also
joined in the grand game for the “original” Gtå, though typically with
even less understanding of what was at stake for western scholars in
disproving K®‚~a’s divinity. Charpentier, however, provides us with a
particularly perverse example of how scholars, determined to dismantle
K®‚~a’s divinity, resorted to falsification of the text. According to him,
“the reading me’cyuta [at 1.21c], whatever be the text of manuscripts
available at the present date, must be false; for this verse belongs to the
original epic text, and to its authors K®‚~a is not Acyuta, the Supreme God
Vi‚~u” (104). As though anticipating Jei’s thesis of how an interpolator,
inspired by the occurrence of a word in one verse, might add the word to
another or compose a verse around it, he claimed that “it is easily understandable that from here [that is, verse 18.73, where the word acyuta also
occurs] the final redactor of the Gtå transposed it to i, 21; for then Arjuna
would be made to address the Supreme Being as Acyuta the first and the
last time that he speaks to him within the Gtå” (104). Regarding Charpentier’s claim that “the original text [that is, the text after his emendation]
may have run somewhat like this: s. u. m. ratham sthåpaya keçava (or
mådhava)” (104n63), we can only comment, as Sukthankar did of Ruben’s
(1930: 251) suggestion that the reading of the archetype might have been
sahastramekaμ çlokånåμ sapta çlokaçatåni ca: “the line is metrically,
The Bhagavad Gtå and the Historical-Critical Method / 271
grammatically and stylistically impeccable, and does him credit. As we
do not, however, want to rewrite the Mahåbhårata, such manufactured
lines have no value for us” (Sukthankar 1930: 279).
45. It is impossible to keep all theistic vocabulary out of the original
nontheistic sections as the example of chapter 14 shows.
46. Further, in a text that purports to unfold an ontology, it makes sense
that the theme of K®‚~a’s divinity is introduced gradually. The text does
not dogmatically declare K®‚~a to be God, but leads Arjuna (and, by extension, the reader) on a pedagogic journey. Since the concept of divinity
germane to the text is that of God as the indweller (dehin), the soul or the
self (åtman), and the witness (såk‚in) of all beings, the thesis of K®‚~a’s
divinity is not contradicted by his appearing in mortal form. In fact, the text
has an entire theory of manifestation (compare s®jåmy aham [Bhagavad
Gtå 4.7]) to account for God’s appearance in a mortal form. Thus, we find
that it is only in the tenth chapter, when K®‚~a has completed his introductory descriptions of the techniques of self-control, purification, asceticism,
and meditation through which the philosophical adept gains a vision of
God, that Arjuna expresses his wish to hear of his divine glories. The stage
is set for an increase in “theistic” terminology. Likewise, it is only in the
eleventh chapter, when Arjuna expresses his assent to K®‚~a’s divine glories and asks to view his supernal form, that K®‚~a unfolds the spectacle
of his universal form (viçvar¨pa). Neither the structure of the text nor the
fact that the theophany unfolds only now is accidental. The text is fully
capable of accounting for the increase in theistic terms in some chapters.
47. This theory was first articulated by Holtzmann, Jr. (1892), who
claimed that K®‚~a was originally either a scheming politician of uncertain
Indo-Germanic ancestry or a lowly cowherd chieftain (Holtzmann is not
consistent) and only later became a deity once the Bråhma~as identified
him with their God Vi‚~u-Nåråya~a. The theory was taken up by Hopkins
(1895) and Garbe (1905, 1909), who used it to articulate their theories of
a new religion, which they referred to as “K®‚~aism.” In their view, K®‚~a
was originally a prophet and founder of a religion. After his death, the
Bhågavatas, the adherents of his religion, began to treat him as a deity,
calling him bhagavåna. Eventually, this religion became too powerful
for the Bråhma~as to ignore, and they were forced to absorb it into their
religion (variously called “Bråhma~ism” or “Hinduism”). Bhargava’s
participation in these scholars’ project of demythologization and historici-
272 / Vishwa Adluri and Joydeep Bagchee
zation indicates just how much of the Protestant narrative he has absorbed
(see also Ruben 1943 for another example of how this narrative can be
absorbed within a Marxist perspective).
48. See Hacker 1960, and also Schneider 1982. The essays in Stietencron 2005 continue these researches.
49. The deeply Protestant Albrecht Weber (see McGetchin 2009: 126)
made some of the most strenuous attempts to prove the Christian origins
of the worship of K®‚~a (see Weber 1850c: 399–400, 423, 1868: 339; the
passages are translated and discussed in Adluri 2015; see also note 95
below).
50. Reproduced from Jei 1986: 636. Jei’s scheme might be generously compared with the attempt by Whitman (1958) to identify the symmetrical composition of the Iliad. Classical philology was long a source
of inspiration for the Indologists, who attempted to replicate their colleagues’ work (see Oldenberg 1906; more recently, Witzel 2014 has made
the same claim). However, Whitman’s model has long been abandoned
in the classics; paradoxically, the Parry-Lord hypothesis put speculation
about the Iliad’s composition to rest in Homeric Studies. We say “paradoxically” because Jei claims that “the originally oral character of the
text would, possibly, have left it open to reciters and performers to add,
drop or change a word, a påda, a stanza or more. On the other hand, this
oral origin partly bestowed its syntax of epic formulae on the diction”
(1986: 629). However, if the Bhagavad Gtå’s origins are oral, then repetition cannot be considered a sign of interpolation, since, according to the
Parry-Lord hypothesis, bards made use of techniques such as formulaic
and ring composition. Thus, although the bardic hypothesis opens the
floodgates of speculation in Mahåbhårata Studies, it simultaneously works
against some of the claims about layers made in the name of oral origins.
51. “The possibility that along with the upani‚adic form of our poem its
pure epic form, included in the Bhårata, was still separately preserved,
cannot be excluded” (Jei 1986: 633).
52. The term “Ur-Bhagavadgtå” is possibly modeled on Otto’s “UrGtå,” by which the Marburg scholar meant the original Bhagavad Gtå
containing the core of the numinous experience (the mysterium tremendum)
prior to its expansion by eight doctrinal treatises (Otto 1934 and 1935).
Otto thought that this Ur-Gtå would have contained the same experience
of the holy as he found to be contained in the Old and New Testaments
The Bhagavad Gtå and the Historical-Critical Method / 273
and, above all, in Martin Luther’s De Servo Arbitrio (The Bondage of the
Will). He claimed that it would have been the original and most authentic
core of the text. In Jei’s case, it is not clear why this part of the text
should be more deserving of the term Bhagavad Gtå or Ur-Bhagavad
Gtå than any other. From his statement that “neither epic nor yogaçåstra
includes anything that would justify such a title,” it seems that he is interpreting the Bhagavad Gtå along similar lines as Otto: it refers primarily
to religious revelation, and hence the term Ur-Bhagavad Gtå would be
inappropriate for any other parts of the text such as the so-called epic poem.
53. The term “Våsudevabhakta” appears to be Jei’s own. It is not
found in any of the literature on the Bhagavad Gtå before this period,
though it is possible that it is borrowed from Dandekar’s (1975–76) discussion of the compound våsudevaka. Jei does not explain what significance he attaches to this term; perhaps he simply means “a devotee of
Våsudeva.” In later articles, this is replaced by “Bhågavata” and “ProtoBhågavata,” the terms preferred by the scholars he most cites.
54. Chapter 15 is possibly part of the bhakti stretch of the poem (Jei
is not explicit about where the Yoga layer ends and the bhakti poem
begins); in that case, the Våsudevabhakta would have inserted the verse
from the tri‚†ubh hymn into his own poem at the time of its creation.
55. That bhakti is not simply a doctrine or a movement or a sectarian
understanding, available ready-to-hand for “interpolation” into various
texts in accordance with the need of individuals to enhance their own
prestige does not occur to Jei because he fundamentally does not see
bhakti as an intellectual phenomenon. From his perspective, religious
developments can be explained without remainder by positing the representational needs of certain sects or individuals. The fact that positing a
“Våsudevabhakta” as the source of the “Våsudeva bhakti” does not take
us very far in understanding bhakti (in fact, the argument relies only on
the relative homonymy of the names to sound plausible; actually, Jei
has said nothing at all) eludes him. Yet Jei is himself only responding
to the demand for “religion-historical” (religionsgeschichtliche) explanations of bhakti that emerges from German Indology and its antediluvian
concerns (see Adluri and Bagchee 2016).
56. Jei seems to be borrowing the thesis from Rudolf Otto. Otto had
opined that “for this reason [that is, because it shows “ÇVARA, the
VIOLENT”] chapter 11, the great theophany, belongs first of all in our
274 / Vishwa Adluri and Joydeep Bagchee
primordial context (Urzusammenhang). It shows the Almighty All-Being:
in his ghorar¨pa and his viçvar¨pa, and the former is more important
than the latter” (1934: 9). He also followed this up with the observation
that, in the Calcutta edition of the Mahåbhårata, the description of K®‚~a’s
miracle in the Kaurava court from the Udyogaparvan lacks any reference
to K®‚~a as possessing manifold limbs, and so on. Otto claimed that the
first half of the theophany of the fifth book “actually does not speak at all
of a viçvar¨pa, but of the måyå of an all-powerful måyin, and portrays
him in his ghorar¨pa” (25). From this he concluded: “it was only later
that people tried to make out of this the viçvar¨pa of the All-God” (25),
and he extended this argument to the Gtå. Jei may be contemplating
something similar: in that case, the central sections (which depict K®‚~a
in his ghora form) would be primary, the work of the composer of the
tri‚†ubh hymn; only later, as the hypothetical Våsudevabhakta undertook
a revision with the aim of depicting K®‚~a as the God of bhakti, were
these horrific descriptions enclosed in a description of the viçvar¨pa, or
the universal form. Jei does not explain why it was essential that the
Våsudevabhakta make use of the central sections of the tri‚†ubh hymn.
57. Jei does not clarify why the Våsudevabhakta should have inserted
these verses precisely here. His generic explanation is that the Våsudevabhakta felt a need to bring the text in line with his bhakti philosophy.
However, this does not explain the presence of these particular verses in
this precise place. Logically, the Våsudevabhakta could have extended the
bhakti interpolation further, creating a Bhakti layer from 2.8 to chapter 11.
58. This point is debatable, but we have chosen to follow Jei’s argumentation here. In his opinion, 2.6, 2.8, and 11.33 express sentiments
different from (what he takes to be) characteristics of the heroic outlook.
Exactly why a hero could not have said “And we do not know what is
better for us: / That we defeat them or they defeat us; / Dh®tar傆ra’s men
are positioned before us, / After killing whom we have nothing to live
for” (2.6) or “There is nothing I see that might dispel / This sorrow that
dessicates my senses, / If on earth I were to obtain without rivals / A
kingdom, nay even the reign of the Gods!” (2.8; van Buitenen translation) he does not explain. Perhaps, like Holtzmann, Jr., he finds such
sentiments more characteristic of the “elegiac wisdom, the resignation,
being tired of life, of later Indian literature” than the “raw warrior-like air
of the old Germanic North” (1892: 45) that Holtzmann thought he
The Bhagavad Gtå and the Historical-Critical Method / 275
detected in the oldest parts of the epic. See also the next note.
59. The idea that a warrior cannot cry out in doubt or sorrow is a specifically modern and Romantic one. It has no support from either Indian
or Greek antiquity. In the Iliad, Achilles, the paradigmatic hero of the
poem, cries out to his mother when he is denied his honor; the passage
reads: “Then Achilles, in tears, / Withdrew from his friends and sat down
far away / On the foaming white seashore, staring out / At the endless
sea. Stretching out his hands, / He prayed over and over to his beloved
mother: / ‘Mother, since you bore me for a short life only, / Olympian
Zeus was supposed to grant me honor. / Well, he hasn’t given me any
at all. Agamemnon / Has taken away my prize and dishonored me.’ /
His voice, choked with tears, was heard by his mother / As she sat in the
sea-depths beside her old father. / She rose up from the white-capped sea
like a mist, / And settling herself beside her weeping child / She stroked
him with her hand and talked to him” (360–375; Lombardo translation).
Achilles’s desire for honor does not preclude him from weeping pathetically to his mother; he further asks his mother to intervene with Zeus on
his behalf to turn the tides of battle against the Greeks (420–30). Contrary
to the expectations of the Indologists, what sides a hero takes in battle is
thus secondary to heroism itself. This brings into question not only their
assumption of what a hero is, but also the thesis of an “inversion” in the
Mahåbhårata (Holtzmann 1892). Indeed, Achilles’s fit of “pathetic”
weeping, which specifically looks back to the opening lines of the Iliad
(“Rage: / Sing, Goddess, Achilles’ rage, / Black and murderous, that cost
the Greeks / Incalculable pain, pitched countless souls / Of heroes into
Hades’ dark, / And left their bodies to rot as feasts / For dogs and birds, as
Zeus’ will was done” [1–6]), is central to the purpose of the epic: without
this “pathetic” supplication, the central drama of the Iliad cannot unfold.
It seems that in the Iliad at least the distinction between a “heroic” and a
“pathetic” style is unwarranted. Incidentally, we might also see, in the
Greek epic, a caution against using repetition as a criterion to identify
“interpolations.” Commenting on these verses (362–92), Kirk notes: “The
earlier part, over half, of Akhilleus’ reply to Thetis’ enquiry is a long
summary, without the all-important speeches, of the events dramatically
described so far, and which have led to Akhilleus’ present distress. It is
surprising to find such a summary so close to the beginning of the whole
poem and so soon after the extremely full description of arguments and
276 / Vishwa Adluri and Joydeep Bagchee
events. Thetis does not need it…, nor does the singer or his audience at
this point—although sometimes such a résumé can be helpful to both”
(1985: 91). Classical philologists are much more circumspect about inferring an interpolation, on the basis of seeming repetition or seeming confusion.
60. The argument is, furthermore, circular, as it presumes the existence
of something that is supposed to be demonstrated by the analysis. This is
a major objection to the idea of continuity repetitions: the concept already
presumes that there are such things as layers in the Bhagavad Gtå, which
make it meaningful to group repetitions according to whether they continue the thought of the original or they merely replicate the form of the
original while changing its thought. Whatever Jei might say about
having accidentally come across repetitions in the text (1986: 629), his
idea of repetition being a guide to interpolation actually assumes composite origins of the Bhagavad Gtå.
61. Jei does define continuity repetitions as repetitions that “display
definite relationships to one another: complementary or opposing relationships, or relationships such as exist between question and answer, for
example, which serve to develop a given subject” (1986: 629). But this
definition specifically extends to parts that have already been identified
as contiguous in terms of their content, and hence the criterion of repetition is superfluous to the attempt to identify the “layers” in the poem.
We are specifically concerned here with a definition of continuity repetition that is: (1) not circular; and (2) not based on content. A formal definition of continuity repetitions (for example, that they are the repetitions
characteristic of the primary parts of the poem), however, is excluded
because Jei also extends the concept to describe the existence of (probable) continuities within secondary layers.
62. As earlier, italicized text indicates continuity repetitions; bold text
indicates duplication repetitions.
63. That is, unless we want to assume that he composed the duplication
repetitions in anticipation of what would later happen to his composition.
64. If this is so, textual layers would then no longer be traces of Indian
cultural history but only of a specific stage of that history, vitiating the
broader argument of Jei’s article, which is that the Bhagavad Gtå’s
layers might be a guide to Indian cultural history.
65. This does not pose a problem for Jei because he considers it self-
The Bhagavad Gtå and the Historical-Critical Method / 277
evident who the various “composers” and “redactors” of the Bhagavad
Gtå were and what motives they followed in composing and/or interpolating their various texts. In his opinion, all that the critic has to do is to
bring these “composers” and “redactors” into the probable order of their
activity, and he will soon have a reconstruction of the Gtå’s history.
What Jei does not realize thereby is that this order only appears probable
to him because of his acceptance of a priori theories about the origins of
the Gtå in a small, heroic poem and its later “interpolation” by individuals or sects following different ideologies such as “theism,” “K®‚~aism,”
“Vi‚~uism,” “bhakti,” and so on. For a criticism of this view of the Mahåbhårata as a text created through constantly being redacted to meet the
representational needs of various traditions, see Adluri and Bagchee (2016).
66. Even if we were to accept Jei’s thesis (namely, that the presence
of duplication repetitions is a guide to the “interpolations” in the Bhagavad
Gtå), nothing about duplication repetitions permits us to deduce the order
of the interpolations in the text. Even if duplication repetitions let us
identify verses 2.11–30, 2.37, and 3.22–23 as interpolations, nothing about
the repetitions tells us that verses 2.11–30, 2.37 must have been interpolated first. Since the order of the text is no longer normative (that is, that
verses 2.11–30, 2.37 must have preceded verses 3.22–23 in the text; the
former could have acquired their current position even if inserted after the
latter), the interpolation could have occurred in any order. The duplication repetitions can at most identify a, b, c, d, and so on, as distinct layers
without telling us that their sequence is a @ b @ c @ d, and so on. That
means, however, that the textual layers of the Bhagavad Gtå cannot function as a guide to Indian cultural history, as Jei claims, because all that
we have is a series of interpolations, without being able to say in any way
what the probable order of their sequence was.
67. This must be a typographical error; there are no common words
between 2.14 and 2.38. Jei likely means 2.15, which does correspond
to his analysis of the common features of 2.14 and 2.38.
68. Incidentally, the idea that everything following the injunction to
fight, especially the long philosophical discussions, is unnecessary in a
poem whose main problem is how to get a central combatant to return to
battle is a false idea. It is based on a prejudice that K®‚~a’s sole concern
in the Bhagavad Gtå is to get Arjuna to fight, which is not true: K®‚~a’s
real concern is to get Arjuna to fight for the right reasons, and that means
278 / Vishwa Adluri and Joydeep Bagchee
to distinguish between the agent of action and a higher self that is unaffected by action (see Bagchee 2011a, 2011b). These subtle philosophical
points escape the grasp of readers who proceed like Nietzsche’s “plundering troops” (1967: 72).
69. This accords with what we have found about duplication repetitions: the transition to a new layer is always identified on the basis of the
scholar’s perception of a new theme and not an objective criterion.
70. Jei does not explain why he thinks that yoga could not be an
organic continuation of the original text. In a work describing methods of
self-control (essential for the warrior) and methods of salvation (essential
for the philosophical adept; Arjuna is both), we would expect some discussion of both. In fact, the turn to yoga (which literally means “yoking,”
but here probably has the sense of “application,” “endeavor,” and also
“contemplation,” “meditation,” and “self-concentration”) is already indicated by verse 2.14, where Arjuna is told to endure the contradictory pulls
of heat and cold (and, in general, of the pairs of opposites). It is only on
the assumption that K®‚~a means to recount no philosophy, that his sole
goal (as is appropriate for the context of war) is to get Arjuna to return to
battle that the turn to yoga can appear as an “interpolation.” All German
Indologists from Adolf Holtzmann, Jr. through F. Otto Schrader, Hermann
Jacobi, and Hermann Oldenberg were unanimous that philosophical doctrines were out of place in—what they assumed was—a battlefield instruction; Oldenberg even wrote, “How strange, these theosophic teachings at
such a moment, this sudden silencing of the noise of battle in the face of
a mystic otherworldly stillness” (1922: 71; emphasis in original; compare
Jacobi’s “What epic poet would completely neglect to take into consideration the epic situation described by him in order to place a philosophical
conversation spanning over six hundred and fifty verses in the mouths of
two of his heroes at a moment when the opposed armies are about to begin
their attack?” [1918: 323]). The earliest historical source of Jei’s theory
of a Yoga layer, however, is Holtzmann, Jr., who in his 1893 reconstruction opined that, in addition to the two original elements of the text (nullity
of the fear of death and pantheism), “in the second chapter…in which
with verse 6.26.11…the actual didactic poem begins, K®‚~a’s answer
addresses both these points, but it also places a third way of looking at
things, one from the perspective of immersion (yoga), alongside” (1893:
154).
The Bhagavad Gtå and the Historical-Critical Method / 279
71. Jei’s confusion between what the so-called markers can actually
achieve and what he thinks they can achieve is easily explained by the
contradictory imperatives he is following. On the one hand, he is not
interested in the markers telling him anything about the poem’s articulation (in fact, where they do, as in the case of tasmåt, they pose a challenge to his thesis). Rather, he only wishes them to confirm an a priori
scheme he has of the Gtå’s “layers.” On the other hand, he requires the
markers as a means to offer “verification” of the existing scholarly prejudice about the Bhagavad Gtå. Hence the paradox that the markers are
being used to mark the divisions that Jei already thinks exist in the text.
72. In reality, of course, Jei is confusing two things. On the one hand,
there are the repetitions, which he may well have come across serendipitously. On the other hand, there is the decision to treat the repetitions as
evidence of layers, which not only cannot be accidental but also shows
that Jei’s reading of the text is not as free of external influences as he
claims. There is nothing intrinsic about repetitions that says that they
must be due to layers in the text. As we note below, prior to the nineteenth
century no one had thought of or thought up the Bhagavad Gtå’s layers.
It is only after the development of so-called critical methods of analysis
in biblical criticism and their importation into Indology that scholars first
thought of identifying layers in the text. When Jei now uses terms such
as “layer” and “interpolation” (as well as the terminology for the acceptable layers such as “Epic,” “Hymnic,” “Såμkhya,” “Yoga,” “Theistic,”
“K®‚~aite,” and “Bhakti”), he reveals his reliance upon this tradition.
German prejudices about an original epic constitute the background of
his conscious reading.
73. Apparently, there was some confusion about this, because, according
to Jei’s 1986 scheme, the Yoga layer should have been the second
layer after the Såμkhya layer. In the 1979 article as well, the Yoga layer
is called the “youngest” of the Såμkhya, Epic, and Yoga layers (547).
Perhaps he means that the Yoga layer was added first to the Epic layer
and then the Såμkhya layer was, as he says, “embedded” between them.
But this still would not explain the Yoga layer being called younger than
the Såμkhya. Perhaps he simply is not clear about his own scheme.
74. The choice of this verse is quite random and illustrates the abuse to
which his method is susceptible. There is no reason why the çreya in
verse 5.1 should replicate the one in 3.2d; the term also occurs at 1.31c,
280 / Vishwa Adluri and Joydeep Bagchee
2.5b, 2.7c, 2.31d, 3.11d, 3.35a, 3.35d, 4.33a, 12.12a, 16.22d, and 18.47a.
Any one of these could have been the source of the poet’s choice, or he
might simply have looked up the word in a lexicon. That Jei alights on
this one occurrence shows that, once again, the desired results are guiding
the method.
75. Jei notes that this verse “could belong to the later bhakti layer”
(1979: 550) as it contains the expression matparaª; compare Bhargava’s
(1977: 358) argument for considering 2.61 a “theistic” interpolation (the
objections are always the same, the only thing that changes is the name
given each time to the particular interpolation). Perhaps the correct justification is neither “theistic” nor “bhakti” nor “K®‚~aite,” but simply ‘at
odds with the Christian view of possessing the one true God.’
76. In this example, in the interests of space, the full verses have not
been quoted.
77. Jei claims of this verse that it is “out of context in our layer” and
that it “clearly anticipates the next yoga layer in chapter V [this would be
the second Yoga layer]” (1979: 550). From this he concludes that its function was “to connect a later addition [the second Yoga layer?] with the
preceding text” (550).
78. Jei claims of this verse that it is “even more out of context than
II.72” and that it has “a similar role, namely, to anchor a late addition in
XVIII.41–48 in the preceding text [in the first Yoga layer?]” (1979: 552).
79. See next note.
80. Jei does not clarify why these verses are considered part of the
interpolation, other than to note that “stanzas 10–16 follow one another
with such a syntactical and semantical cohesion that it is obvious that
they represent, together with st. 9, a complete entity within the rest of the
text” (1979: 551). Apparently, it suffices to have continuity repetitions
only within some verses of an interpolation to identify all the members of
that interpolation as being interpolated. This makes the criterion of continuity repetition somewhat redundant, but Jei does not see that.
81. As a rule, hypotheticals are not made more plausible by adding
more hypotheticals, but that is how Jei’s approach (and, in general, the
historical-critical method) functions.
82. Jei writes that he is unsure whether this verse “belongs to this
segment [that is, the second yajña interpolation], to the primordial yoga
layer or to the bhakti interpolation in st. 35 [at times also called the sva-
The Bhagavad Gtå and the Historical-Critical Method / 281
dharma interpolation], recognizable by the affirmation ‘bh¨tåny açe‚e~a
drak‚yasi…mayi’ ” (1979: 551). However, he considers the second option
(that is, that it belongs to the first Yoga layer) “the least probable” (551).
83. Jei does not mark the words or passages duplicated in the Bhagavad
Gtå. Perhaps he means that the Bhagavad Gtå cites the Ka†ha Upani‚ad
without making any effort to duplicate it, for what would be the point of
trying to “duplicate” (in the special terminological sense in which Jei
means it) an external source?
84. This once again shows, if any confirmation was needed, that the
criterion of continuity repetitions is actually superfluous.
85. This is perhaps the only place in the entire article where Jei really
has a convincing parallel, but that the Ka†ha Upani‚ad is a source for
some of the Gtå’s verses was already known to most commentators.
86. The problem is more serious than we have made it here, for there
are other terms within the semantic field of cognition/intellect in the
Bhagavad Gtå such as manas and cit. What is the evidence that the composer of the “prajña interpolation” looked only to the buddhi verses in
creating his verses? Why could he not just as well have been inspired by
other verses (that is, those containing terms like manas or cit, presuming,
of course, that is how composers write, that is, that they look at verses
with certain words in a text and then model other verses on them, though
replacing one or two terms by the terms current in their “school”)? Indeed,
what is the evidence that buddhi came first, and it was not buddhi that
was modeled on prajña? As with the so-called continuity and duplication
repetitions, the chronological sequence of the interpolations has been
determined first and the analyses only constructed to provide a semblance
of objectivity.
87. This could have occurred in two ways: either the adherents of the
Yoga school, noticing that their text had been overlaid with prajña interpolations, struck back with a number of buddhi verses, or yet other rivals
of the prajña philosophers chose to advance the claims of buddhi terminology once more.
88. For instance, that Yoga is the school that uses the term buddhi and
buddhi is the term characteristic of the Yoga school.
89. Some startling examples are provided by Jei: “A careful philological analysis can discover in the ingeniously unified Bhagavadgtå a
complex structure of layers” (1986: 634); “It goes without saying that
282 / Vishwa Adluri and Joydeep Bagchee
such types of repetitions should be formally demonstrated in every case,
and the philological conclusions corroborated by a sufficiently close
network of clearly characterised repetitions”; “The actual task of philological research is to draw conclusions about a text only so far and so definitely as the material permits” (1979: 546); “On the basis of this distinction between ‘continuity repetitions’ and ‘duplication repetitions,’ I hope
to have discovered an objective criterion which enabled me to detect
different layers of text of the BhG in my previous articles…” (2009b:
31); “According to my analyses in Jei 2009[b], the cited passage in
Ka†ha is younger than at least the first 4 text layers of the Bhagavadgtå
and the Tri‚†ubh Hymn within it (largely in ch. 11), and therefore cannot
represent the oldest evidence of the concept. This can serve as an example
of the utility of the philological analytical approach to texts” (2010: 89n8).
90. The tenuous claim of Indology to being a science worthy of the
modern research university is nowhere more apparent than in the many
calls to recognize Indology as a form of philology. For examples, see
Adluri and Bagchee (2014b: 356–403, especially the sections “Steps
toward a Scientific Indology,” “Steps toward a Positivist Philology,” and
“Construing the (Natural) Scientific Character of Philology”). Since the
publication of this book, more examples have appeared of Indologists
wishing to legitimate themselves by laying claim to the legacy of philology
(Witzel 2014; Witzel’s claims to philology are addressed in Adluri and
Bagchee forthcoming).
91. The following is a partial list of those who cite Jei as authoritative: Witzel 2014: 34n158, 2006: 486n113; Brockington 1997: 29, 43,
1998: 269–70, 275 (both refer to Jei offering a “more sophisticated
analysis (than others to date)”; Malinar 1996: 115–386; Szczurek 2002,
2003, 2005–2006, 2007, 2008; and Andrijani 2012: 15n26, 2013: 21–22,
22n2, 25–30. Johnson (2007: 657) cites Jei (1986), Brockington (1998),
and Szczurek (2002) and argues (states?) that their work is “respected.”
Jei’s work is also recommended on the Indology mailing list as constituting “basic Indological research on the Gita”; “the last detailed critical
& historical discussion of the Gita” (the reference is to Jei’s 1979 article,
though erroneously, as the author clearly has in mind the 1986 article);
and a means of ensuring that discussion remains restricted to “Indological
Studies, not for broadcasting religious or personal philosophical persuasions [?]”—Michael Witzel to Indology list, March 26, 2001, “Is the Gita
The Bhagavad Gtå and the Historical-Critical Method / 283
Dishonest?” Available at: list.indology.info/pipermail/indology_list. indology.info/2001-March/025970.html (accessed December 1, 2015).
92. See also Sharpe 1985 and Herling 2006. Neither scholar, however,
traces German Gtå interpretations right up to the present. They also do
not focus specifically on Indology, but on a wider tradition of reception.
93. See Holtzmann 1893: 153.
94. Compare Jei’s claim: “by applying a very different methodology,
I hope to have attained more completely argumented and more precise
results than the former researchers. My observations, which are not as
yet concluded, imply, however, without presupposing it, that the poetic
parts of the Gtå are relatively more ancient than the didactic parts attached
to them, in accordance with what Otto and Oldenberg supposed, and that
the såμkhya and yoga layers precede the vedåntic elements, as Garbe and
Renou, in part conjectured, but that the bhakti layer comes last, contrary
to Garbe and Otto, but in accordance with Holtzmann’s and Hopkins’
intuition…” (1986: 629). And see also Szczurek, who writes that “for
over 170 years, papers taking positions on the problem of the redaction
of the text [the Bhagavad Gtå] have been published…especially in
efforts by a series of philologists who have tried to discover the genesis
of the Bhagavadgtå by proposing reconstructions of its original form, or
a close approximation” (2002: 56). The figures he accords to this critical
tradition are Wilhelm von Humboldt, F. Otto Schrader, Hermann Jacobi,
Hermann Oldenberg, Jarl Charpentier, and Mislav Jei.
95. The title is an excerpt from part of a letter addressed to Karl Otto
von Raumer by Albrecht Weber. The full passage reads: “The entire
weight of the religious and cultural structure of contemporary India appears
to rest on the Vedas. As soon as they are unveiled from the mysterious
darkness surrounding them till now…, and made accessible to all, all the
untruths shall be automatically revealed, and this shall, in time, put an
end to the sorry plight of religious decadence…of India. The critical
analysis and publication of Vedic texts shall assume a role among the
Indians similar to Luther’s translation of the Bible” (Weber 1855, cited
and translated in Sengupta 2004: 279).
96. See Hiltebeitel (1979), commenting on earlier attempts to supply
K®‚~a with “a historical, as opposed to a legendary, identity”: “Hopkins’
criteria for determining history are not only subjective but inane” (78).
The comment could almost be an epitaph for this tradition of scholarship.
284 / Vishwa Adluri and Joydeep Bagchee
97. The following section owes much to the input of four anonymous
readers of the article. In many places, their questions or comments were
absorbed directly into the text, without being able to quote them of course.
Nonetheless, we acknowledge the debt here: their careful reading was
invaluable to improving this article.
98. Several sources (Khair 1969: 12; Jei 1986: 628, 2009b: 31; Szczurek 2002: 56n3) confer this honor upon Humboldt (1826), perhaps
with the intent of claiming a more illustrious ancestor for their method.
However, von Humboldt never advocated the kind of historical reconstruction characteristic of later scholarship. His sole comment on the
form of the text was: “the first eleven chapters enclose the doctrine completely…If the conclusion appended to the end of the eighteenth chapter
(from çloka 63 onwards) were to follow after the final verse of the eleventh
chapter, I believe that the poem would hardly appear deficient” (Humboldt
1906: 327), but we find no reference to the fact that this text might either
have existed or have preceded the current Bhagavad Gtå. Other statements such as von Humboldt’s rejection of the theory of the Gtå as a
composite work (333) or of the theory that inflections in tone or content
in the poem might permit us to identify earlier parts (334, echoing Oldenberg 1922: 71 and Jacobi 1918: 323; both passages are cited in note 70
above) militate against such a hypothesis.
99. There appears to be some concern among western scholars to show
that the search for layers in the Bhagavad Gtå is not merely a feature
of their work. Brockington writes: “the tendency to dissect is not, however, confined to western scholars, as is shown by the work, originally in
Marå†h, of G S Khair, in which he concludes that the Bhagavadgtå was
composed by three ‘philosopher-poets’ at three different periods” (1997:
29); he also notes that “another such attempt is that by Purushottam Lal
Bhargava, 1977” (46n6; and see also Schrader 1927, arguing from the
Çuddha Dharma Ma~ala’s apocryphal Gtå that the “fantastic statements
[of the Haμsayogin]…are not completely worthless inasmuch as they
show…that the idea of earlier and shorter versions of the Bhagavad Gtå
is present even in India” [178; emphasis in original]). Similar comments
were made about Indian scholars to the authors of this paper by Mislav
Jei at the Seventh Dubrovnik International Conference on the Sanskrit
Epics and Purå~as (referring specifically to G.S. Khair and to K.T. Telang).
The need to claim predecessors in Khair and Telang is understandable, but
The Bhagavad Gtå and the Historical-Critical Method / 285
both Brockington and Jei miss the point: the real question is whether
they can demonstrate that the identification of layers was a concern in
Indian scholarship before the nineteenth century. Both Khair and Telang
are aware of the work of the German scholars. Khair (1969: 12–14)
discusses how “critical study” of the Gtå began with the posing of the
question of its unity (in the work of von Humboldt) and thereafter cites
the work of Albrecht Weber, Adolf Holtzmann, Jr., Edward Washburn
Hopkins, Richard von Garbe, Rudolf Otto, and Moriz Winternitz as
evidence of the scholarly consensus that the Bhagavad Gtå is a composite
work. Telang (1882: 1–36) begins by giving assent to Friedrich Max
Müller’s judgment of Indian thought as lacking a historical sense; he
thereafter cites the work of Albrecht Weber, August Wilhelm von Schlegel,
Christian Lassen, and Georg Bühler, and especially refers to Theodor
Goldstücker’s view of the transformation of the Mahåbhårata from a text
for the K‚atriya warriors to a work of Bråhma~ic ritual and philosophy.
Khair and Telang cannot be cited as representatives of an authentic Indian
tradition, while the Haμsayogin’s thoroughly modern creation suggests
just how much native traditions of scholarship have been contaminated
with western ideas and representations after the nineteenth century (for a
discussion of the relationship between Schrader’s work and the Haμsayogin’s, see Bagchee and Adluri 2015).
100. Some amusing examples are provided by Jei: “For my part, I
was struck by the fact that, with three minor modifications, Szczurek
identified exactly the same 22 bhakti interpolations throughout the text as
I did in my first attempt to survey them thirty years ago in 1979. I have
never published that part of my research, and it could not have influenced
Szczurek. In my opinion, it is a strong indication of the great degree of
certainty with which such text-critical research can be pursued with our
methodology” (2009b: 32n2; see also Jei 2009a: 262n77, where the
assessment is repeated in almost the exact same words), and by Szczurek:
“Jei’s research method seems to be an extremely important and accurate
approach in analysing the drafting of the poem’s text. The objectivity of
its philological procedure, including clear formal evidence for the conclusions, is the reason why the author of this paper [that is, Szczurek]
accepts the results of his analysis” (2002: 57). This assessment is also
repeated in Szczurek 2008: “Among over 20 works which have been
trying to present (sometimes quite contradictory) views on the history
286 / Vishwa Adluri and Joydeep Bagchee
of the poem’s text, Mislav Jei’s research has in my eyes been of the
greatest importance” (179). Szczurek notices neither that this research is
an amalgam of existing German views, nor that it is no less contradictory
than in the original sources.
101. Commentators use both the term prak‚ipta (interpolations) and
apapå†ha (corruptions). Occasionally, there may also be a phrase like
“some do not read [na pa†hanti] this verse.”
102. For a discussion of interpolation in the Mahåbhårata tradition, see
our work Philology and Criticism: A Guidebook to Mahåbhårata Textual
Criticism (Adluri and Bagchee forthcoming).
103. Brockington argues that “either it [that is, the Bhagavad Gtå] is
an integral part of the Mahåbhårata and directed pragmatically to Arjuna’s
situation, or it is a later insertion (which includes the possibility of a later
expansion of a brief original) developing a philosophically and theologically significant message from its Mahåbhårata context” (1997: 30), but
this dichotomy only holds if one assumes that the aims of the epic are contradictory to those of the Bhagavad Gtå. It is on the basis of this assumption and this assumption only that the philosophical discourse will appear
as an interruption of the text. Brockington explicitly makes this assumption, for he says: “among the reasons for thinking that the Bhagavadgtå
was not originally part of the Mahåbhårata is precisely the incongruity
of such a sermon at such a point in the narrative. Would both armies really
have waited while Krishna answers Arjuna’s doubts at such length, especially when the battle has actually begun, as BhG 1.20c indicates (prav®tte
çastrasaμpåte)?” (30), but perhaps he would now reconsider his views in
light of the evidence.
104. See, for instance, Jei’s comment that his analyses presuppose
“the flawless argumentation of Georg von Simson (1969)” (2009b: 32) or
Szczurek’s comment “Von Simson’s analysis of the text seems unassailable and it appears reasonable to proceed from his conclusion that the
Bhagavadgtå was not originally part of the epic” (2002: 55). Brockington
thinks that von Simson has identified the “mechanisms” by which the
Bhagavad Gtå was included within the Mahåbhårata (1997: 31, see also
1998: 147), without, however, having presented conclusive evidence for
its insertion (although he thinks there is “clear textual evidence” for its
insertion [1997: 31, 1998: 146]). Fitzgerald makes this dependence on
anti-Bråhma~ism as a hermeneutic principle explicit. After expressing
The Bhagavad Gtå and the Historical-Critical Method / 287
his basic agreement with von Simson’s thesis (“I sketched some of these
ideas [about the Mahåbhårata depicting the ‘restoration of proper, bråhma~ya kingship, which undertakes to use violence for the protection and
support of brahmins’ against the Mauryan empire] in ‘The Great Epic of
India as Religious Rhetoric’ in 1983, and Georg von Simson wrote in
‘The Mythic Background of the Mahåbhårata [1984],’ ‘I believe that the
main redaction of the MBh is to be understood as a reaction of orthodox
Brahminical circles against the religious policies of the Mauryas (223)’ ”
[Fitzgerald 2004: 121n173, 121]), he writes: “By Bhagavad Gtå I mean
basically the argument of the first twelve chapters of K®‚~a’s sermon and
demonstration to Arjuna. The BhG has an interesting and complicated
history. Some important studies over the past three decades by Georg
von Simson (‘Die Einschaltung der Bhagavad Gtå im Bh‚ma parvan
des Mahåbhårata’); Gajanan Shripat Khair (Quest for the Original Gtå),
and Mislav Jei (‘The First Yoga Layer in the Bhagavadgtå’ and ‘Textual
Layers in the Bhagavadgtå as Traces of Indian Cultural History’) have
proposed valuable arguments bearing upon the history of the BhG. Von
Simson’s work is particularly important, for it convicingly shows how
the BhG was inserted into its specific textual context” (Fitzgerald 2004:
140n240). For a criticism of von Simson’s so-called text-critical arguments, see Adluri and Bagchee (2014b: 277–96).
105. To be fair, the reader was not really proposing that historical
insights had been achieved in the case of the Gtå. The reference in the
original was to early Christian texts. Nonetheless, we thought we should
address this comparison as well.
106. For a classic example, see Malinar (2007), where we read that
“seen from the perspective of the debates on war and peace that pervade
the UdP [Udyogaparvan], the BhG can be regarded as a continuation or
even commentary on some of the issues raised in this book. This connection of the BhG to the UdP may have been one reason for including the
BhG in this part of the epic.…The following analysis is based on the
extant, critically constituted text of the UdP and thus deals with those
dimensions of meaning that were established by the time of the final
redaction of the epic. At this stage, the text of the UdP testified to a
cultural-historical situation in which different notions of asceticism and
heroism have already been developed. Therefore the UdP also includes
texts in which K®‚~a is presented as the highest god…[,] thus pointing to
288 / Vishwa Adluri and Joydeep Bagchee
the influence of the theology of the BhG in some parts of the UdP that
are then regarded as being later than the BhG” (35). How exactly someone
can use the Udyogaparvan to date the Bhagavad Gtå, but then use the
Bhagavad Gtå to date (and excise) parts of the Udyogaparvan is beyond
us.
107. This is so not only because historical critics insist that the only
context worthy of respect is the historical situation at the time of composition of the text under consideration (whereas there are, of course, other
kinds of contexts such as the literary context, the context of the work
within a wider set of works, within the tradition, and so on, that are equally
if not more important for understanding the meaning of a verse; see
Levenson 1993: 4), but also because historical critics have been peculiarly
resistent to the suggestion that the text undergoes a recontextualization
within the tradition and, indeed, each time a reader reads it. For an
example of such dogmatic blindness see Hanneder (2008).
108. Like Kugel, we recognize that “from its inception, this scholarly
discipline was fundamentally a Protestant undertaking, one might even
say, a form of Protestant piety” and that it has “in ways great and small,
still retained much of its particularly Protestant character” (1986: 22).
But we also recognize that, in the context of Indian Studies, the adoption
of historical criticism went along with a refusal to recognize Indian readers
as preservers and authentic interpreters of the text. Narratives of racial
and religious decline were invented to explain why authority over the
texts had to be transferred to the western critics (Garbe 1889; Oldenberg
1886, 1922; Witzel 2014). Because we reject the triumphalism of that
“underhanded theology” (Nietzsche 2005: 9) known as German Indology,
we see no scope for integrating its solafidean theology with our exegetic
concerns.
109. Sharpe puts it nicely: “one of the strangest things about the Western tradition of Gita hermeneutics has been its almost total lack of interest in the Hindu world’s own estimate of its own Gita.…The West has
approached the Gita from one of two angles: either as a piece of archaic
literature, to be dissected, analysed and placed in an essentially remote
religio-historical context, or as an exotic insight into the ultimate mystery
of the universe—a scripture which is Hindu only incidentally” (1985:
xiii).
110. Indeed, historical critics were often explicitly aware of their work
The Bhagavad Gtå and the Historical-Critical Method / 289
as contributing to Protestantism. Soulen and Soulen note: “historical critics
have often thought of themselves as continuing the aims of the Protestant
Reformation, above all, with respect to historical criticism’s commitment
to the exclusive validity of the literal (i.e., original) sense of the text, and
the necessity of interpreting the Bible free from the influence of ecclesiastical tradition and control.…The theologian Gerhard Ebeling, following
Rudolf Bultmann, has argued that there is a ‘deep, inner connection’
between the historical-critical reading of scripture and the Reformers’
doctrine of justification by grace through faith in that both function to
remove all false security” (2011: 89–90).
111. Postcritical biblical interpretation offers an explicit parallel to the
problem we have diagnosed as being symptomatic of German approaches
to the Gtå: “unlike methodologies such as source and form criticism,
which disintegrate the text into its antecedent kernels, postcritical biblical
interpretation assumes that the canonical form of the text was designed to
convey a message, and that finally the Bible itself is a text in its own right
in which all discernible units large and small take on new hues and connotations within the whole” (Soulen and Soulen 2011: 157).
112. For examples of comments by Indologists, see Garbe 1889: 90;
Roth 1855: v; Oldenberg 1906: 5–6; Charpentier 1930: 46; the comments
are echoed in Hanneder 2001, 2005, and 2010.
113. See Schechter (1915), who believes that the primary function of
historical criticism is to perpetuate—and institutionalize in “scientific”
form—the reaction against Hebrew tradition. “Higher anti-Semitism is
partly…contemporaneous with the genesis of the so-called Higher criticism of the Bible. Wellhausen’s Prolegomena and History are teeming
with aperçus full of venom against Judaism, and you cannot wonder that
he was rewarded by one of the highest orders which the Prussian Government had to bestow. Afterwards Harnack entered the arena with his
‘Wesen des Christenthums,’ in which he showed not so much his hatred
as his ignorance of Judaism. But this Higher anti-Semitism has now
reached its climax when every discovery of recent years is called to bear
witness against us and to accuse us of spiritual larceny” (37). (To this
list, Schechter might have added Christian Lassen and Albrecht Weber,
both holders of the Orden pour le mérite.) According to Schechter, the
characteristic feature of this new form of anti-Semitism (for which he,
combining the two terms, coined the memorable expression “higher anti-
290 / Vishwa Adluri and Joydeep Bagchee
Semitism”) is that it “burns the soul though it leaves the body unhurt”
(36). “The [Hebrew] Bible is our sole raison d’être, and it is just this
which the Higher anti-Semitism is seeking to destroy, denying all our
claims for the past, and leaving us without hope for the future” (37).
114. Benes uses the phrase in the context of describing the role of
comparative philologists in orientalizing European Jews. We obviously
are using it in a wider sense, yet in continuity with Benes’s sense. The
orientalization of “nonmodern” modes of study (that is, the Jewish but
also the Bråhma~ic tradition) is the central impulse for Indology.
115. But see Framarin (2009) for how scholars working outside of the
historicist paradigm have been able to make productive use of the text.
There has been a tradition of philosophical interpretation of the Gtå
outside of Indology and its fundamentalist concerns, but this would be a
separate article.
116. Malinar, Harder, and Oberlies (2011) claim to be sensitive to the
problem of discrimination but are still to articulate a vision of how they
see Indology developing in future. We are troubled by the fact that the
authors still have not offered a justification for the systematic exclusion
of native commentators as “religious” in the wake of the evidence presented in The Nay Science of German scholars’ Christian apologetic
concerns.
117. As Indologists have invoked the parallel with biblical criticism
several times, they might also look to developments in that field for inspiration. Here the name of Brevard S. Childs (Protestant but superbly
informed about the issues informing historical criticism) comes to mind.
His student Alan Cooper offers the following suggestions: “in my own
subsequent work, I suggested that the full history of interpretation served
as an indispensable bridge between the modern reader and the ancient
text, and that modern commentators ought to adopt three hermeneutical
principles of traditional interpretation: ‘the assumption that the text is
meaningful; the demand that interpretation be answerable to the text; and
the principle that all interpretations merely realize the text’s possibilities:
“new” interpretations, if they adhere to the first two principles, then add
to the repository of ideas that is the history of interpretation’ ” (2002: 26–
27; emphasis in original). There need not be a conflict between historical
criticism and other forms of commentary, provided Indology becomes
self-aware of its history and that what it too ultimately does is offer no
The Bhagavad Gtå and the Historical-Critical Method / 291
more than an interpretation. But there can be no space for narratives of
German scientific supremacy nor for stereotypes of Indians (Hanneder
2001, 2005, 2010).
118. This table is reproduced from Adluri and Bagchee (2014b: 309–
12), though with Charpentier’s scheme and also a number of corrections
that could not be incorporated into the final proofs. It lists the eighteen
chapters of the Bhagavad Gtå along with the deletions or changes made
by each author. Strikethrough signifies deletion—where the author excises
one or more verses from the chapter, the verse numbers are struck through;
where he excises the entire chapter, the complete verse range for the
chapter is listed and struck through. A question mark indicates uncertainty.
Transposition of verses and verses retained as part of a secondary Gtå
are indicated within parentheses.
119. Wherever there is some doubt about whether the author includes
the verses, we have chosen to be charitable and assume that he does. Thus,
the percentage figures at the end of the table represent the upper margin.
120. Hauer comments that “in…[his] opinion the original poem was
comprised of a brief conversation in chapter 2…and of the vision in chapter 11, which likely followed this conversation” (1937: 72n7). However,
he also includes translations of a number of verses in his outline of an
Indo-Åryan philosophy or metaphysics, which fall outside his “original”
Gtå. As these choices are especially revelatory of his concerns, we have
featured two columns. The first indicates the extent of Hauer’s deletions;
the second gives the reader a quick overview of the verses he translates
(and hence apparently found to resonate with his ideas of an Åryan
outlook).
121. Oldenberg is typically vague about the extent of the original Gtå.
He often contradicts himself, tries to ingratiate himself with other authors,
and in general tries to articulate some kind of consensus view (actually, a
kind of mean average of all existing views). Hence there are two figures
for him, though we might easily have listed three or four, so confused is
he.
122. In a second attempt at determining the “ ‘original’ Gtå” (see Garbe
1914: 232n1), Garbe removed an additional 175 verses from the Bhagavad
Gtå, as advised by Winternitz (1907: 197).
123. If we included the deletions outside the Bhagavad Gtå, the figure
for von Simson (1969) would be negative: -991 or -141.57 percent.
292 / Vishwa Adluri and Joydeep Bagchee
References Cited
Adluri, Vishwa. 2010. Review of The Bhagavadgtå: Doctrines and Contexts, by Angelika Malinar. History of Religions 50, 1: 102–7.
Adluri, Vishwa. 2015. “Philosophical Aspects of Bhakti in the Nåråya~ya.” In Simon Brodbeck, Alf Hiltebeitel, Adam Bowles, eds., The
Churning of the Epics and Purå~as at the 15th World Sanskrit Conference, 127–54. New Delhi: Rashtriya Sanskrit Sansthan and D. K.
Printworld.
Adluri, Vishwa P. and Joydeep Bagchee. 2014a. “The Redemption of the
Brahman: Garbe and German Interpreters of the Bhagavadgtå.” In
Joanne Miyang Cho, Eric Kurlander, and Douglas T. McGetchin, eds.,
Transcultural Encounters between Germany and India: Kindred Spirits
in the Nineteenth and Twentieth Centuries, 68–83. New York: Routledge.
Adluri, Vishwa and Joydeep Bagchee. 2014b. The Nay Science: A History
of German Indology. New York: Oxford University Press.
Adluri, Vishwa and Joydeep Bagchee. 2016. “Bloß Glaube? Understanding
Academic Constructions of Bhakti in the Past Century.” In Emmanuel
Francis and Charlotte Schmid, eds., The Archaeology of Bhakti II:
Royal Bhakti, Local Bhakti, 79–126. Pondicherry: Institut Français de
Pondichéry and École Française d’Extrême Orient.
Adluri, Vishwa and Joydeep Bagchee. Forthcoming. Philology and Criticism: A Guide to Mahåbhårata Textual Criticism. London: Anthem.
Andrijani, Ivan. 2012. Filozofija vedånte: Filozofska tumaenja upani‚adi
Bijele Yajurvede. Zagreb: FF Press.
Andrijani, Ivan. 2013. “Historical Analysis of Textual Layers in Ancient
Indian Literature and Indian Cultural History.” In Danuta Stasik and
Anna Trynkowska, eds., CEENIS Current Research Series, 1: 21–43.
Warsaw: Dom Wydawniczy ELIPSA.
Arvidsson, Stefan. 2006 [2000]. Aryan Idols: Indo-European Mythology
as Ideology and Science (trans. Sonia Wichmann). Chicago: University
of Chicago Press.
Bagchee, Joydeep. 2011a. “The Bhagavadgtå: Philosophy versus Historicism.” Feature Review Article of Desire and Motivation in Indian
Philosophy, by Christopher G. Framarin. Philosophy East and West 61,
4: 707–17.
Bagchee, Joydeep. 2011b. “A Response to Christopher Framarin.” Philos-
The Bhagavad Gtå and the Historical-Critical Method / 293
ophy East and West 61, 4: 720–22.
Bagchee, Joydeep. 2011c. “Inversion, K®‚~afication, Brahmanization: The
Explanatory Force of Some Extraordinary Figures of Speech.” Journal
of Vaishnava Studies 19, 2: 127–41.
Bagchee, Joydeep and Vishwa Adluri. 2014. “The Passion of Paul Hacker:
Indology, Orientalism, and Evangelism.” In Joanne Miyang Cho, Eric
Kurlander, and Douglas T. McGetchin, eds., Transcultural Encounters
between Germany and India: Kindred Spirits in the Nineteenth and
Twentieth Centuries, 215–29. New York: Routledge.
Bagchee, Joydeep and Vishwa P. Adluri. 2015. “Recensions of the Bhagavad Gtå? Apocryphal Gtås in Germany and India in the Early Twentieth
Century.” Paper presented at the 16th World Sanskrit Conference, Bangkok, Thailand, June 28–July 2, 2015.
Bagchi, Kaushik. 1996. “Orientalism without Colonialism? Three Nineteenth-Century German Indologists and India.” Ph.D. dissertation. Ohio
State University, Columbus.
Bagchi, Kaushik. 2003. “An Orientalist in the Orient: Richard Garbe’s
Indian Journey, 1885–1886.” Journal of World History 14, 3: 281–325.
Barnes, Robin. 2003. “Images of Hope and Despair: Western Apocalypticism ca. 1500–1800.” In Bernard McGinn, John J. Collins, Stephen J.
Stein, eds., The Continuum History of Apocalypticism, 323–53. London:
Continuum.
Benes, Tuska. 2010. In Babel’s Shadow: Language, Philology, and the
Nation in Nineteenth-Century Germany. Detroit: Wayne State University Press.
Bhagavad Gtå. 1981. The Bhagavadgtå in the Mahåbhårata: Text and
Translation (ed. and trans. J.A.B. van Buitenen). Chicago: University
of Chicago Press.
Bhargava, P.L. 1977. “Additions and Interpolations in the Bhagavadgtå.”
East and West 27, 1–4: 357–61.
Bhargava, P.L. 1979. “Names and Epithets of K®‚~a in the Bhagavadgtå.”
Indologica Taurinensia 7: 93–96.
Bigger, Andreas. 1998. Balaråma im Mahåbhårata: Seine Darstellung im
Rahmen des Textes und seiner Entwicklung. Wiesbaden: Harrassowitz.
Brockington, John L. 1997. “The Bhagavadgtå: Text and Context.” In
Julius Lipner, ed., The Fruits of Our Desiring: An Inquiry into the Ethics
of the Bhagavadgtå for Our Times, 28–47. Calgary: Bayeux.
294 / Vishwa Adluri and Joydeep Bagchee
Brockington, John L. 1998. The Sanskrit Epics. Leiden: E.J. Brill.
Charpentier, Jarl. 1930. “Some Remarks on the Bhagavadgtå.” The Indian
Antiquary 59: 46–50, 77–80, 101–5, 121–26.
Cooper, Alan. 2002. “Biblical Studies and Jewish Studies.” In Martin
Goodman, ed., The Oxford Handbook of Jewish Studies, 14–35. New
York: Oxford University Press.
Dandekar, R.N. 1975–76. “The Beginnings of Vai‚~avism.” Indologica
Taurinensia 3–4: 169–86.
Figueira, Dorothy M. 2002. Aryans, Jews, Brahmins: Theorizing Authority
through Myths of Identity. Albany: State University of New York Press.
Fitzgerald, James L. 1983. “The Great Epic of India as Religious Rhetoric:
A Fresh Look at the Mahåbhårata.” Journal of the American Academy
of Religion 51, 4: 611–30.
Fitzgerald, James L., ed. and trans. 2004. The Mahåbhårata: Book 11, The
Book of the Women; Book 12, The Book of Peace, Part 1. Chicago:
University of Chicago Press.
Framarin, Christopher G. 2009. Desire and Motivation in Indian Philosophy. New York: Routledge.
Garbe, Richard von. 1889. Indische Reiseskizzen. Berlin: Verlag von
Gebrüder Paetel.
Garbe, Richard von. 1905. Die Bhagavadgîtâ: Aus dem Sanskrit übersetzt,
mit einer Einleitung über ihre ursprüngliche Gestalt, ihre Lehren und
ihr Alter. Leipzig: H. Haessel.
Garbe, Richard von. 1909. “Bhagavad-Gtå.” In James Hastings, ed., Encyclopaedia of Religion and Ethics, 2: 535–38. Edinburgh: T. & T. Clark.
Garbe, Richard von. 1914. Indien und das Christentum: Eine Untersuchung
der religionsgeschichtlichen Zusammenhänge. Tübingen: J.C.B. Mohr.
Hacker, Paul. 1960. “Zur Entwicklung der Avatåralehre.” Wiener Zeitschrift
für die Kunde Süd- und Ostasiens und Archiv für Indische Philosophie
4: 47–70.
Hacker, Paul. 1961. “Zur Methode der geschichtlichen Erforschung der
anonymen Sanskritliteratur des Hinduismus.” Zeitschrift der Deutschen
Morgenländischen Gesellschaft 111, 2: 483–92.
Hanneder, Jürgen. 2001. Review of Råjavidyå: Das königliche Wissen
um Herrschaft und Verzicht. Studien zur Bhagavadgtå, by Angelika
Malinar. Zeitschrift der Deutschen Morgenländischen Gesellschaft 151,
1: 239–41.
The Bhagavad Gtå and the Historical-Critical Method / 295
Hanneder, Jürgen. 2005. Review of The Pandit: Traditional Scholarship
in India, edited by Michael Axels. Zeitschrift der Deutschen Morgenländischen Gesellschaft 155, 2: 671–72.
Hanneder, Jürgen. 2008. Review of Bhagavad Gtå. Der Gesang des
Erhabenen, translated and edited by Michael von Brück. Marburger
Forum: Beiträge zur geistigen Situation der Gegenwart 9, 4. Available
at: www.philosophia-online.de/mafo/heft2008-4/Han_Bhag.pdf (accessed
July 30, 2008; site discontinued).
Hanneder, Jürgen. 2010. Marburger Indologie im Umbruch: Zur Geschichte
des Faches 1845–1945. Munich: P. Kircheim.
Hanneder, Jürgen. 2011. “Pretence and Prejudice.” Indologica Taurinensia
37: 123–37.
Hauer, Jakob Wilhelm. 1937 [1934]. Eine indo-arische Metaphysik des
Kampfes und der Tat: Die Bhagavadgita in neuer Sicht. Stuttgart:
Kolhammer.
Herling, Bradley L. 2006. The German Gtå: Hermeneutics and Discipline
in the German Reception of Indian Thought, 1778–1831. New York:
Routledge.
Hiltebeitel, Alf. 1979. “K®‚~a and the Mahåbhårata (A Bibliographical
Essay).” Annals of the Bhandarkar Oriental Research Institute 60, 1:
65–107.
Holtzmann, Jr., Adolf. 1892. Das Mahåbhårata und Seine Theile. Volume
1: Zur Geschichte und Kritik des Mahåbhårata. Kiel: C.F. Haessler.
Holtzmann, Jr., Adolf. 1893. Das Mahåbhårata und Seine Theile. Volume
2: Die neunzehn Bücher des Mahåbhårata. Kiel: C.F. Haessler.
Hopkins, Edward Washburn. 1895. The Religions of India. Boston: Ginn.
Humboldt, Wilhelm von. 1826. Über die unter dem Namen BhagavadGítá bekannte Episode des Mahá-Bhárata. Berlin: Druckerei der Königlichen Akademie der Wissenschaften.
Humboldt, Wilhelm von. 1906. Wilhelm von Humboldts Gsammelte
Schriften. Volume 5: Werke, 1823–1826 (ed. Albert Leitzmann). Berlin:
B. Behr’s.
Iliad. 1997. Iliad [by] Homer (trans. Stanley Lombardo). Indianapolis:
Hackett Publishing.
Jacobi, Hermann. 1918. “Über die Einfügung der Bhagavadgtå im Mahåbhårata.” Zeitschrift der Deutschen Morgenländischen Gesellschaft 72,
4: 323–27.
296 / Vishwa Adluri and Joydeep Bagchee
Jacobi, Hermann von. 1921. “Die Bhagavadgtå,” review of Die Bhagavadgîtâ: Aus dem Sanskrit übersetzt, mit einer Einleitung über ihre ursprüngliche Gestalt, ihre Lehren und ihr Alter, by Richard von Garbe. Deutsche
Literaturzeitung 42, 50–51: 715–24.
Jei, Mislav. 1979. “The First Yoga Layer in the Bhagavadgtå.” In,
Ludwik Sternbach Felicitation Volume, 1: 545–57. Lucknow: Akhila
Bharatiya Sanskrit Parishad.
Jei, Mislav. 1986. “Textual Layers of the Bhagavadgtå as Traces of
Indian Cultural History.” In Wolfgang Morgenroth, ed., Sanskrit and
World Culture: Proceedings of the Fourth World Sanskrit Conference,
628–38. Berlin: Akademie Verlag.
Jei, Mislav. 2002. “The Bå‚kalamantra-Upani‚ad and the Bhagavadgtå.” In Mary Brockington, ed., Stages and Transitions: Temporal and
Historical Frameworks in Epic and Purå~ic Literature. Proceedings of
the Second Dubrovnik International Conference on the Sanskrit Epics
and Purå~as, August 1999, 35–54. Zagreb: Croatian Academy of Sciences and Arts.
Jei, Mislav. 2009a. “The Relationship between the Bhagavadgtå and
the Vedic Upani‚ads: Parallels and Relative Chronology.” In Robert P.
Goldman and Muneo Tokunaga, eds., Epic Undertakings: Papers of the
12th World Sanskrit Conference, 2: 215–82. Delhi: Motilal Banarsidass.
Jei, Mislav. 2009b. “The Tri‚†ubh Hymn in the Bhagavadgtå.” In Petteri
Koskikallio, ed., Parallels and Comparisons: Proceedings of the Fourth
Dubrovnik International Conference on the Sanskrit Epics and Purå~as,
September 2005, 31–66. Zagreb: Croatian Academy of Sciences and
Arts.
Jei, Mislav. 2010. “To Be Liberated while Still Alive or to Die in Order
to be Liberated—in the Jñåna, Karma and Bhakti Yoga of the Bhagavadgtå—According to Different Text Layers.” In Andreas Bigger, Rita
Krajnc, Annemarie Mertens, Markus Schüpbach, and Heinz Werner
Wessler, eds., Release from Life, Release in Life: Indian Perspectives
on Individual Liberation, 87–110. Bern: Peter Lang.
Johnson, Kathryn Ann. 2007. “The Social Construction of Emotions in
the Bhagavad Gtå: Locating Ethics in a Redacted Text.” Journal of
Religious Ethics 35, 4: 655–79.
Khair, G.S. 1969. Quest for the Original Gtå. Bombay: Somaiya Publications.
The Bhagavad Gtå and the Historical-Critical Method / 297
Kirk, G.S. 1985. The Iliad: A Commentary. Volume 1: Books 1–4. Cambridge: Cambridge University Press.
Kugel, James L. 1986. “Biblical Studies and Jewish Studies.” AJS Newsletter 36: 22–24.
Lassen, Christian. 1837. “Beiträge zur Kunde des indischen Altertums
aus dem Mahâbhârata. I: Allgemeines über das Mahâbhârata.” Zeitschrift für die Kunde des Morgenlandes 1: 61–86.
Lassen, Christian. 1847. Indische Alterthumskunde. Volume 1: Geographie
und die älteste Geschichte. Leipzig: Verlag von L.A. Kittler.
Levenson, Jon D. 1993. The Hebrew Bible, the Old Testament, and Historical Criticism. Louisville: Westminster/John Knox.
Malinar, Angelika. 1996. Råjavidyå: Das königliche Wissen um Herrschaft und Verzicht. Studien zur Bhagavadgtå. Wiesbaden: Otto
Harrassowitz.
Malinar, Angelika. 2003. “ ‘K‚atriya-Glaube’ und ‘Opferwesen’: Richard
Garbe und die indischen Religionen.” In Heidrun Brückner, Klaus
Butzenberger, Angelika Malinar, and Gabriele Zeller, eds., Indienforschung im Zeitenwandel: Analysen und Dokumente zur Indologie
und Religionswissenschaft in Tübingen, 121–43. Tübingen: Attempto.
Malinar, Angelika. 2007. The Bhagavad Gtå: Doctrines and Contexts.
Cambridge: Cambridge University Press.
Malinar, Angelika, Hans Harder, and Thomas Oberlies. 2011. “Editor’s
Introduction.” Zeitschrift für Indologie und Südasienstudien 28: 1–2.
Max Müller, Friedrich, trans. 1879. The Upanishads. Part I. Oxford:
Clarendon Press.
McGetchin, Douglas T. 2009. Indology, Indomania, and Orientalism:
Ancient India’s Rebirth in Modern Germany. Madison: Fairleigh Dickinson University Press.
Nietzsche, Friedrich. 1967. Menschliches, Allzumenschliches. Volume II:
Nachgelassene Fragmente Frühling 1878 bis November 1879. Berlin:
Walter de Gruyter.
Nietzsche, Friedrich. 1999. “On Truth and Lying in a Non-Moral Sense
[1873].” In Friedrich Nietzsche, The Birth of Tragedy and Other Writings
(eds. Raymond Geuss and Ronald Speirs; trans. Ronald Speirs), 139–
53. Cambridge: Cambridge University Press.
Nietzsche, Friedrich. 2005. The Anti-Christ, Ecce Homo, Twilight of the
Idols, and Other Writings (eds. Aaron Ridley and Judith Norman; trans.
298 / Vishwa Adluri and Joydeep Bagchee
Judith Norman). Cambridge: Cambridge University Press.
Oldenberg, Hermann. 1886. “Ueber Sanskritforschung.” Deutsche Rundschau 47: 386–409.
Oldenberg, Hermann. 1906. “Indische und klassische Philologie.” Neue
Jahrbücher für das klassische Altertum, Geschichte und deutsche
Literatur und für Pädagogik 17: 1–9.
Oldenberg, Hermann. 1919. “Bemerkungen zur Bhagavadgtå.” Nachrichten
von der Königlichen Gesellschaft der Wissenschaften zu Göttingen, Philologisch-historische Klasse, 321–38.
Oldenberg, Hermann. 1922. Das Mahåbhårata: Seine Entstehung, seine
Inhalt, seine Form. Göttingen: Vandenhoeck and Ruprecht.
Olivelle, Patrick. 1996. Upani‚ads: A New Translation. New York: Oxford
University Press.
Otto, Rudolf. 1934. Die Urgestalt der Bhagavad-Gtå. Tübingen: J.C.B.
Mohr (Paul Siebeck).
Otto, Rudolf. 1935. Der Sang des Hehr-Erhabenen. Die Bhagavad-Gtå
übertragen und erläutert. Stuttgart: Kohlhammer.
Rogerson, John W. 1985. Old Testament Criticism in the Nineteenth
Century: England and Germany. Minneapolis: Fortress Press.
Roth, Rudolf von. 1855. “Vorwort.” In Otto Böhtlingk and Rudolf Roth,
eds., Sanskrit-Wörterbuch, 1: iii–iv. St. Petersburg: Kaiserliche Akademie
der Wissenschaften.
Ruben, Walter. 1930. “Schwierigkeiten der Textkritik des Mahåbhårata.”
Acta Orientalia 8: 240–56.
Ruben, Walter. 1943 [1941]. Krishna: Konkordanz und Kommentar der
Motive seines Heldenlebens. Vienna: Adolf Holzhausens Nfg.
Schechter, Solomon. 1915. “Higher Criticism—Higher Anti-Semitism.”
In Solomon Schechter, Seminary Address and Other Papers, 35–39.
Cincinnati: Ark Publishing.
Schmitt, Arbogast. 2012 [2008]. Modernity and Plato: Two Paradigms of
Rationality (trans. Vishwa Adluri). Rochester: Camden House.
Schneider, Ulrich. 1982. “K®‚~as postumer Aufstieg: Zur Frühgeschichte
der Bhakti-Bewegung.” Saeculum 33, 1: 38–49.
Schrader, F. Otto. 1910. “Über Bhagavadgtå II, 46.” Zeitschrift der
Deutschen Morgenländischen Gesellschaft 64, 2: 336–40.
Schrader, F. Otto. 1927. “Neues über die Bhagavadgtå.” In Julius von
Negelein, ed., Aus Indiens Kultur: Festgabe Richard von Garbe dem
The Bhagavad Gtå and the Historical-Critical Method / 299
Forscher und Lehrer zu seinem 70. Geburtstag dargebracht von seinen
Freunden, Verehrern und Schülern, 171–83. Erlangen: Palm und Enke.
Sengupta, Indra. 2004. “State, University, and Indology: The Politics of
the Chair of Indology at German Universities in the Nineteenth Century.”
In Douglas T. McGetchin, Peter K.J. Park, and D.R. SarDesai, eds.,
Sanskrit and “Orientalism”: Indology and Comparative Linguistics in
Germany, 1750–1958, 271–305. Delhi: Manohar.
Sharpe, Eric J. 1985. The Universal Gtå: Western Images of the Bhagavad
Gtå. A Bicentenary Survey. La Salle: Open Court.
Sheehan, Jonathan. 2005. The Enlightenment Bible: Translation, Scholarship, Culture. Princeton: Princeton University Press.
Simson, Georg von. 1969. “Die Einschaltung der Bhagavadgtå im Bh‚maparvan des Mahåbhårata.” Indo-Iranian Journal 11, 3: 159–74.
Simson, Georg von. 1984. “The Mythic Background of the Mahåbhårata.”
Indologica Taurinensia 12: 191–223.
Smith, Morton R. 1968. “Statistics of the Bhagavadgtå.” Journal of the
Ganganath Jha Research Institute 24: 39–46.
Soulen, Richard N. and R. Kendall Soulen. 2011 [1976]. Handbook of
Biblical Criticism. Louisville: Westminster John Knox Press.
Steinmetz, David C. 2011. Taking the Long View: Christian Theology in
Historical Perspective. New York: Oxford University Press.
Stietencron, Heinrich von. 1996. “Editor’s Introduction” (trans. Srilata
Müller). In Angelika Malinar, Råjavidyå: Das königliche Wissen um
Herrschaft und Verzicht. Studien zur Bhagavadgtå, 6–11. Wiesbaden:
Otto Harrassowitz.
Stietencron, Heinrich von. 2005. Hindu Myth, Hindu History: Religion,
Art, and Politics. Delhi: Permanent Black.
Sukthankar, V.S. 1930. “Epic Studies: III, Dr. Ruben on the Critical
Edition of the Mahåbhårata.” Annals of the Bhandarkar Oriental Institute 11, 3: 259–83.
Sukthankar, V.S. 1957. On the Meaning of the Mahåbhårata. Bombay:
Asiatic Society.
Szczurek, Przemyslaw. 2002. “Some Remarks on the So-called Epic Layer
of the Bhagavadgtå.” In Mary Brockington, ed., Stages and Transitions: Temporal and Historical Frameworks in Epic and Purå~ic Literature. Proceedings of the Second Dubrovnik International Conference on
the Sanskrit Epics and Purå~as, August 1999, 55–72. Zagreb: Croatian
300 / Vishwa Adluri and Joydeep Bagchee
Academy of Sciences and Arts.
Szczurek, Przemyslaw. 2003. “Brahmanirvå~a Versus Nirvå~a (Nibbåna):
Some Remarks on the Polemics with Buddhism in the Bhagavadgtå.”
In Renata Czekalska and Halina Marlewicz, eds., Proceedings of the
2nd International Conference on Indian Studies, 547–75. Cracow:
Institute of Oriental Philology.
Szczurek, Przemyslaw. 2005. “Bhakti Interpolations and Additions in the
Bhagavadgtå.” In Petteri Koskikallio, ed., Epics, Khilas and Purå~as:
Continuities and Ruptures. Proceedings of the Third Dubrovnik International Conference on the Sanskrit Epics and Purå~as, September
2002, 183–220. Zagreb: Croatian Academy of Sciences and Arts.
Szczurek, Przemyslaw. 2005–2006. “Jednoç w rónorodnoçci. Uwagi o
warstwach tekstowych Bhagawadgity.” Studia Indologiczne 12–13:
94–126.
Szczurek, Przemyslaw. 2007. “Juggling with ‘Åtman’: Remarks on the
‘Bhagavad-gtå’ 6.5–6.” Rocznik Orientalistyczny 60, 2: 212–38.
Szczurek, Przemyslaw. 2008. “Prajñåvådåμç ca bhå‚ase: Polemics with
Buddhism in the Early Parts of the Bhagavadgtå.” In Richard Gombrich
and Cristina Scherrer-Schaub, eds., Buddhist Studies: Papers of the 12th
World Sanskrit Conference, 8: 175–231. Delhi: Motilal Banarsidass.
Telang, Kâshinâth Trimbak. 1882. The Bhagavadgîtâ with the Sanatsugâtîya and the Anugîtâ. Oxford: Clarendon Press.
Turner, James. 2014. Philology: The Forgotten Origins of the Modern
Humanities. Princeton: Princeton University Press.
Weber, Albrecht. 1850a. “Ueber die Literatur des Sâmaveda.” In Albrecht
Weber, ed., Indische Studien: Zeitschrift für die Kunde des indische
Alterthums, 1: 25–67. Berlin: Ferd. Dümmler’s Buchhandlung.
Weber, Albrecht. 1850b. “Zwei Sagen aus dem Çatapatha-Brâhmana über
Einwanderung und Verbreitung der Ârier in Indien, nebst einer geographisch-geschichtlichen Skizze aus dem weissen Yajus.” In Albrecht Weber,
ed., Indische Studien: Zeitschrift für die Kunde des indische Alterthums,
1: 161–232. Berlin: Ferd. Dümmler’s Buchhandlung.
Weber, Albrecht. 1850c. “Analyse der in Anquetil du Perron’s Uebersetzung enthaltenen Upanishad: Fortsetzung.” In Albrecht Weber, ed.,
Indische Studien: Zeitschrift für die Kunde des indische Alterthums, 1:
380–456. Berlin: Ferd. Dümmler’s Buchhandlung.
Weber, Albrecht. 1855. “Letter to Karl Otto von Raumer, 12.10.1855.”
The Bhagavad Gtå and the Historical-Critical Method / 301
Humboldt University Archives, P.F. 1433.
Weber, Albrecht. 1868. Über die K®ish~ajanmâsh†amî (K®ish~a’s Geburtsfest). Berlin: Akademie der Wissenschaften.
Whitman, Cedric H. 1958. Homer and the Homeric Tradition. Cambridge:
Harvard University Press.
Winternitz, Moriz. 1907. Review of Vier philosophische Texte des Mahábháratam: Sanatsujâta-Parvan, Bhagavadgîtâ, Mokshadharma, Anugîtâ,
by Paul Deussen. Wiener Zeitschrift für die Kunde des Morgenlandes
21: 194–202.
Witzel, Michael. 2006. “Brahmanical Reactions to Foreign Influence and
to Social and Religious Change.” In Patrick Olivelle, ed., Between the
Empires: Society in India 300 BCE to 400 CE, 457–99. New York:
Oxford University Press.
Witzel, Michael. 2014. “Textual Criticism in Indology and in European
Philology during the 19th and 20th Centuries.” Electronic Journal of
Vedic Studies 21, 3: 9–91.
Yardi, M.R. 1977–78. “Theories of Multiple Authorship of the Bhagavadgtå.” Annals of the Bhandarkar Oriental Research Institute 58–59:
1049–54.
Yelle, Robert A. 2013. The Language of Disenchantment: Protestant
Literalism and Colonial Discourse in British India. New York: Oxford
University Press.
VISHWA ADLURI is Adjunct Associate Professor at Hunter College,
New York.
vadluri@hunter.cuny.edu
JOYDEEP BAGCHEE is Postdoctoral Fellow at Freie Universität
Berlin, Germany.
jbagchee@gmail.com