Hanneder ToEdit
Hanneder ToEdit
Hanneder ToEdit
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
How to Edit
European Textual Criticism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
The Insistence on Stemmatics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
Author Variants and the Hand of the Copy Editor . . . . . . . . . . . . . . . . . . 75
Ks.emarāja, the Editor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
Premodern Indian Textual Criticism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
Textual Criticism in Early Indology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
Early Indological Editions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
Lachmann . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
Indological Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
The Pitfalls of Editing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
Some Text-Critical Principles in Practice . . . . . . . . . . . . . . . . . . . . . . . . . 165
Text-Critical Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
Résumé . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228
Further Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
Preface
1 Especially since French academic hours are, as I have learned, actually of sixty minutes duration,
whereas German “hours” in universities traditionally start fifteen minutes after the full hour
(cum tempore)—a remnant of the monastic hourly prayer. 2 As one American doctoral student
recently told me bluntly: “We are Americans, if we edit, we are not getting jobs.”
menial task of producing printed editions of texts, but as a very basic way to
communicate with the thought world of authors of texts.
As is well-known, most Old Indian Texts have never been treated in this
manner. Wilhelm Rau (1922–1999), who taught Indology at Marburg Univer-
sity, has in an inaugural lecture suggested to all scholars of that subject to edit
one text during their lifetime in order to ameliorate the situation,3 but in view
of the large number of badly edited important texts, as well as inedita that
may be important but remain unknown, this can only be a drop in the ocean.
My own entry into editorial philology started when the supervisor of my
master thesis, Claus Vogel (1933–2012), suggested to me that instead of merely
translating Ks.emarāja’s Pratyabhijñāhr.daya, I could also consider editing it.
Not much more information or supervision was given and I had to find my
own way into editing. Then I had the privilege to work subsequently with
three teachers who all gave practical examples of what it means to be a prac-
tising editor: Michael Hahn (1941–2014) devoted his whole energy to bringing
the forgotten authors of Buddhist poetry to light and was as invaluable an
inspiration as Alexis Sanderson, who has single-handedly reconstructed the
history of Śaivism through working exclusively from primary sources and
more often from manuscripts than from printed texts. And finally I could
receive a prolonged and intensive training in practical editing when I had the
privilege to work in the Moks.opāya Project directed and supervised by Walter
Slaje.
With the habit of working with new sources being formed it is difficult
to avoid looking at manuscripts whenever the opportunity arises. Therefore,
when I worked regularly in the Bodleian Library in Oxford, as well as in
the library of the School of Oriental and African Studies in London, both of
which hold interesting collections of Kashmirian manuscripts, I started to
edit some of the minor works of Sāhib Kaul4 and a few other works. The
wealth of completely unknown texts was fascinating, and it is obvious that
5 These are the following projects financed by the German Research Community (DFG). Stanislav
Jager: Kollation der Birkenrindenhandschrift Ś14 des Utpattiprakaran.a des Moks.opāya (HA
5698/1–1, 2008–2010), Anett Krause: Sāhib Rāms Arbeiten zur Geschichte Kaschmirs. Erst-
edition, Übersetzung und Analyse. (DFG, HA 5698/4–, 2011–2014), Stanislav Jager: Bhāskaras
Cittānubodhaśāstra: Kritische Edition der ersten drei Kapitel mit dem Kommentar des Autors
(DFG, HA 5698/7–1, since 2014). Anna Martin and Maximilian Mehner: Sāhib Rāms Adaption des
Ahlāq-i Muh.sinı̄ (DFG HA 5698/9–1, since 2016). Other projects will be mentioned in the course
ˇ
of the lectures.
In this way a small corpus of editiones principes of Kashmirian Sanskrit works
has been or is being created in Marburg, but since the projects are in different
stages there has never been a good time to give an overview on these activities
and to explain some of the results to a more general audience. On the following
pages such an attempt will be made.
When talking about the practical side of editing, we cannot pass over
the fact that the introduction of the B.A./M.A. system all over Europe has
made it almost impossible to train new editors. With the phases of academic
training being by default now only two (M.A.) or three (B.A., PhD) years, no
sustained effort over a longer period is possible. For training editors in a
complicated language like Sanskrit this will most probably not do. The task
is only seemingly addressed by new centres for philology, which in order to
receive funding have to follow latest academic fashion, as the Berlin based
Zukunftsphilologie,6 but do not seem to aim at actually producing editions of
unknown texts.
It seems that the prospects for editorial work are not very promising, and
this is especially unfortunate, since it must not be forgotten that Kashmirian
literature should claim special attention: after the enormous efforts by Kash-
mirian scholars in the last century, the events of 1990, the dramatic exodus
of Kashmiri Hindus, which by the way has nearly escaped public attention,7
has pushed especially the later Sanskrit literature of Kashmir further out of
sight. Under these circumstances, editors with a knowledge of local details
are no more to be expected. Editing these texts is therefore far from being the
leisurly activity in academic ivory towers it should be, it is the urgent task to
save this part of the world’s cultural heritage from oblivion.
And of course this is only one, historically very productive, but still fairly
small region. If we start to talk about Indian manuscripts in general, and
hear about modern estimates ranging from few millions to thirty million
(Pingree),8 we may still recall the words of William Jones: “wherever we direct
our attention to Hindu Literature, the notion of infinity presents itself.”9
6 See also Jürgen Hanneder: »Zukunftsphilologie oder die nächste M[eth]ode«. In: ZDMG 163.1
(2013), p. 159-172. 7 Compare Walter Slaje: “Kashmir Minimundus. India’s Sacred Geography
en miniature”. In: Roland Steiner (ed.): Highland Philology. Results of a Text-Related Kashmir
Panel at the 31st DOT, Marburg 2010. Halle 2012, p. 26–28 8 William M. Calder and Stephan
Heilen: “David E. Pingree: an unpublished autobiography”. In: Greek, Roman, and Byzantine
9
��
�
La M ancha to Śrı̄nagara – The Translation of the D
. ān
Kviks. ot. a
14 D. Francisco García Ayuso: Sakúntala. drama del poeta indio Kalidasa en siete actos.
Version directa del sanskrit. Madrid: Imprenta de la Biblioteca de Instruccion y Recreo 1875
(Biblioteca Sanskrita), first page. 15 Haug mentions his Spanish student in a letter to his former
teacher Ewald: “Unter meinen Zuhörern ist ein Spanier, der mehrere Jahre hierbleiben will, um
Arabisch und Sanskrit zu lernen. Er will das Studium des Sanskrit u. der Sprachwissenschaft
nach Spanien verpflanzen.” Briefe an Ewald. Ed. R. Fick. Göttingen [no date], p. 105f. 16 El
Nirvâna buddhista en sus relaciones con otros sistemas filosoficos. Madrid 1885. 17 La Filología
en su relacion con el Sanskrit. Madrid 1871. 18 Shyama Prasad Ganguly: »El Quijote in
India: Some Transcultural Considerations«, p. 57. 19 For the following, see Annemarie Jordan
Gschwend and Johannes Beltz: Elfenbeine aus Ceylon. Luxusgüter für Katharina von Habsburg
(1507-1578). Zürich: museum rietberg 2010.
chest carved with a variety of motifs, which exemplify the quality of cultural
interaction.
In 1514 the famous German artist Albrecht Dürer produced one copper en-
graving called “Dudelsackpfeifer” (bag piper), in which he depicted a person
playing an Irish bag pipe, an instrument which Dürer might have seen in 1513,
when his patron Maximilian I fought alongside English troops. Apparently the
Ceylonese Ambassador to the Portuguese court, Śrı̄ Rāmaraks.ā Pan.d.ita, took
some of these prints to his homeland, where they appear on art objects from
1543 onward. The ivory chest kept in Zurich does not only display Dürer’s bag
piper, which had thus travelled from Nürnberg to Śrı̄ Laṅkā, but also several
other European motifs.
And the exchange was not merely eastwards. Around the same time an
Indian Rhinoceros was sent to Portugal, where it attracted much attention
before it drowned by accident. Dürer, being an effective businessman, reacted
quickly and produced a drawing of the animal with a short explanation of its
history.20 Copies of this print, which is among those that have been described
as early instances of tabloid journalism,21 have circulated widely. What un-
derlines the speed and intensity of communication is that Dürer in the print
dates the arrival of the animal to the 1st of May 1513. In fact it arrived only in
1515, the year when Dürer published his print. Dürer had apparently received
the information through the channels of Nürnberg merchants much earlier.
Thus given this exchange, when Cervantes published his Don Quixote in
1605, it is even conceivable that he used Indian motifs, as has been diagnosed
in the prophesy of the ape in the second part of the Don Quixote, which is
told by one “Benengeli”, a name referring perhaps—as indicated by some
Cervantes specialists—to someone from Bengal.22
But this phase of exchange seems to have had no lasting influence. Ro-
manists in the last century had to start from afresh. When in the year 2005
20 The text on this often reproduced work reads: “Nach Christiegeburt / 1513 Jar AD[omin]i
1. May hat man dem großmechtigisten König Emanuel von Portugal gen Lysabona aus India
pracht ein solch lebendig Thier. das nennen sie Rhinocerus/ Das ist hie mit all seiner gestalt
Abconterfect.” 21 This comparison suggests itself when Dürer’s other similarly circulated prints
of spectacular incidents like siamese twins are taken into account. See Dürer. Kunst – Künstler –
Kontext. Ed. Jochen Sander. München: Prestel 2014, p. 295–306. 22 Jorge Flores: “Distant
Wonders: The Strange and the Marvelous between Mughal India and Habsburg Iberia in the
Early Seventeenth Century”. In: Comparative Studies in Society and History. 49.3 (2007), p. 565.
the 400th anniversary of the publication of the first edition of Cervantes’ El
Ingenioso Hidalgo don Quixote de la Mancha was celebrated, Indian Roman-
ists drew up a survey of the reception of the work in India.23 Although the
book was partly conceived more as an Indian response to Cervantes than a
strictly academic survey, it showed what was, in any case, to be expected: that
Cervantes’ work, to be exact, English translations of it—an exception was one
copy of the Spanish original in possession of William Jones—had spread to In-
dia in the nineteenth century, and that translations from these English versions
into Indian vernaculars were gradually being produced. These translations
involved, in accordance with the tastes of the time, literary domestications,
as was common also in other pre-modern traditions of translation. Much of
the foreign flavour of the text was thus lost. Perhaps most visibly, the names
of the characters were “indianized”: Still in the Bengali translation of 1931,
Don Quixote became Don Kusti,24 Dulcinea became Tilottamā. Such highly
adaptive translations were not unknown in India, rather the Indian Pandits
acted very much like the classical translators of Europe, who would render
the foreign with the well-known and edit out anything contrary to classical
standards. It is easy to ridicule such substitutions, as for instance through the
case of a medieval translation from Spanish to German, in which the aceitu-
nas, “olives”, certainly unknown to the German readers at the time, became
“Bratwürst”, that is, fried sausages. But quite apart from such spectacular
cases, the tendency to transform the realia and style of a text into something,
which the reader would recognize as poetically valid is quite understandable.
In this literary texts and religious scriptures may differ considerably: Tibetan
translations of the Buddhist canon from Sanskrit on the contrary were literal
and etymological to the extreme.
Translations into Sanskrit were no exception to the general trend. In
his translation of Shakespeare’s Midsummer Night’s Dream into Sanskrit as
Vāsantikāsvapnam R. Krishnamachariar writes:25
23 Quixotic Encounters. Indian Responses to the Knight from Spain. Ed. Shyama Prasad Gan-
guly. Delhi: Shipra Publications 2006. 24 Quixotic Encounters., p. 21. 25 Kumbhakonam
1892, p. ii.
places, without being detrimental to the general tenor of the pas-
sages in the original. The ideas are enlarged in some places, but
the enlargement is generally in keeping with the dominant feel-
ings. There are deviations in details with a view to keep up the
characteristics of the Sanskrit drama. Some few passages preg-
nant with such ideas as cannot be brought home to our Pandits
have been omitted, as also some passages which relate purely to
Western habits and customs.
26 “Ich glaube man ist auf dem Wege, die wahre poetische Übersetzungskunst zu erfinden; dieser
Ruhm war den Deutschen vorbehalten [. . .]” A. W. Schlegel: “Nachschrift des Uebersetzers
an Ludwig Tieck”. In: Athenaeum. Eine Zeitschrift von August Wilhelm Schlegel und Friedrich
Schlegel. Zweiten Bandes Zweites Stück. Berlin 1799, p. 281.
The utmost precision in imitating grammar and metrics should
be fused with the highest level of uninhibited vitality.27
Now two of the Jena Romantics happen to be the founders of Indology in Ger-
many: Friedrich and August Wilhelm Schlegel. The latter translated profusely
from English, Spanish, Italian, etc.—he is the acclaimed German translator of
Shakespeare—before turning to Sanskrit, and we may therefore wonder how
the Romantic program worked when applied to Sanskrit, a language that is
quite unlike the others he had encountered before.
Firstly we have to see that there was the practice to use Latin as a sort
of didactic link language, which enabled the learner to understand Sanskrit
sentences without much loss. In the first generation of Sanskrit Studies there
were hardly any printed books, and utterly insufficient study materials. One
therefore used interlinear Latin versions to learn and understand Sanskrit, as
shown in the following text from Bopp’s famous edition of one episode from
the Mahābhārata.28
As the example shows, the Latin version was used not so much as a translation
in its own right, but as a sort of meta-language to understand the Sanskrit
more easily: the word order is preserved, each Sanskrit word is translated
with only one Latin word, the compound is indicated by the inflected but
hyphenated pseudo-compound viraseni-filius.
Where this was the learning method, a closely literal translation of Sanskrit
into German must have been the obvious choice. And for the readers of the
writings of the Early Romantics the finishing touch of such a novel attempt
27“Die möglichste Strenge in der grammatischen und metrischen Nachbildung soll mit dem
höchsten möglichen Grade freier Lebendigkeit vereinigt werden.” A. W. Schlegel: Rezension
von “Vier Tragödien des Aeschylos”. Jenaer Allgemeine Literaturzeitung 1804 [reprint: August
Wilhelm von Schlegel’s Sämmtliche Werke. Ed. Eduard Böcking. Leipzig: Weidmann’sche
Buchhandlung 1846, vol. 12, p. 161.] 28 Francisco Bopp: Mahá-Bhárati Episodium. Secundae
emendatae editionis. Berolini 1830, p. 3.
at translation would be an imitation of the metre of the original. Such an
attempt was made very early by Kosegarten, but has been virtually forgotten.
The same line in his translation runs a follows:
Kosegarten has managed to preserve not only the sense, but has realized the
Sanskrit metrical structure based on the quantity of syllables, which German
can only imitate through a stress accent (ictus). Such experiments, of course,
could work only with simple texts and simple metres.
It is unfortunate for more than one reason that Kosegarten has been largely
forgotten, and for that reason a small excursus to the university town of Jena
is necessary. It has been also forgotten that Kosegarten was the third professor,
next to Schlegel and Bopp, to introduce Sanskrit in a German university
around 1818. Kosegarten was trained in theology and had a special interest
in Oriental languages, which originally meant those pertaining to Christian
theology, but soon included Arabic, Persian, and finally Sanskrit. Kosegarten
had learnt Sanskrit in Paris29 and started teaching in Jena in 1818. There he
soon became friends with the German national poet Goethe in Weimar, who
had known his father. Goethe even became godfather to his son. In a brief
note on Indische Dichtungen Goethe explicitly mentions Kosegarten as the
one who explained to him the extent to which the content and character of
these works were modified by English translations:
29On Kosegarten and Sacy, see Michel Espagne: “Silvestre de Sacy et les orientalistes alle-
mands”. In: Itinéraires orientalistes. Revue germanique internationale 7 (2008), p. 79–91.
clearly visible from a translation of some verses directly from the
Sanskrit, which I owe to Prof. Kosegarten.30
Kosegarten left Jena after only a few years to return to his home town Greif-
swald, and the reason for this was unknown. I literally stumbled upon the
explanation when I was looking at decrepit tombs in the old Jena cemetery.
One tombstone turned out to be that of Kosegarten’s first wife. The withered
epitaph gives her dates and it seems that Kosegarten soon after the death of
his wife transferred to his home town to remarry.31
Kosegarten later became the first critical editor of the Pañcatantra, and
it was only when Johannes Hertel found better manuscripts of the text in
Kashmir, and poured undeserved scorn on his pioneering predecessor, poor
Kosegarten was forgotten even in indological circles.
But wrongly so. Kosegarten also worked on Persian literature, and when
Goethe published his famous West-Eastern Divan, he reviewed it. We must add
that this work of Goethe is always referred to when the German appreciation
of foreign cultures and religions needs to be emphasized, and it is a standard
quotation when politicians need to establish a connection to Islamic cultures.
In Germany Goethe stands for the ideal poet and enlightened politician, and it
is just a small aspect of this polymath that he is also unique for his appreciation
of “the East”. This has even influenced post-colonial literature, where Goethe
is sometimes styled the only person beyond the colonial mindset of the day.
Early Sanskritists are never mentioned in this connection, they usually have
to play the role of the culprits who invented the idea of an Aryan people.
There have been backlashes against this Goethe cult. In 1999, just a decade
after German reunification, Weimar became European Capital of Culture, and
this was the time of endless new biographies, works on “The spirit of Weimar”,
on Schiller, Herder, but above all on Goethe. Some authors became weary of
the Goethe cult and tried to highlight the dark side of Goethe, of a politician
30 “Alle diese Gedichte sind uns durch Übersetzungen mitgeteilt, die sich mehr oder weniger vom
Original entfernen, so, daß wir nur ein allgemeines Bild ohne die begrenzte Eigentümlichkeit
des Originals gewahr werden. Der Unterschied ist freilich sehr groß, wie aus einer Übersetzung
mehrerer Verse unmittelbar aus dem Sanskrit, die ich Herrn Professor Kosegarten schuldig
geworden, aufs klarste in die Augen leuchtet.” Goethe’s Werke. Vollständige Ausgabe letzter Hand.
Stuttgart: Cotta 1833, vol. 49, p. 147. 31 Details will appear in my forthcoming monograph on A.
W. Schlegel.
who claimed almost the highest income in the state of Saxony-Weimar, one
who in his various functions did not prevent the public execution of a young
woman, who sent military troups against students’ prostests in Jena, who
vigorously exercised his right to academic and other censorship, and last but
not least, who had set up a net of spies.32 The latter was not without belligerent
implications in a town, where the era of GDR style spying activity had just
ended and where current and prospective state employees were under scrutiny
for having been recruited by the notorious Ministry for State Security.
Some allegations levelled against Goethe were no doubt exaggerated, and
meant mainly to counteract the prevailing idealization. One author diagnosed
a Goethe taboo,33 implying that there are things one cannot say about Goethe,
and the following would probably also fall under that category.
There is an epilogue by Goethe to his Divan, where he writes about Islam
and Indian religions and expresses his sympathy for Mahmud of Ghazni’s
iconoclasm during his Indian raids in rather clear terms. That Goethe detested
Indian art and especially the many-armed deities of India is well known, but
to agree to a destruction of such images would come as a shock to many
admirers of Goethe, and is therefore mostly omitted.34 At the time when it was
published, Goethe’s pronouncement provoked angry reactions by indologists.
The more so, since he did not merely attack sculpture, but also religion by
saying in the same text: “basically Indian religion is worth nothing”.
Even in monographical works on Goethe’s appreciation of the East, based
on Indian and Persian literature, there is usually not a word about this. And
thus Kosegarten, who had written a review on the work and had protested
against the overtly negative tone of Goethe, is not mentioned either. To be fair,
in a letter to Kosegarten Goethe agreed that he had been too harsh, but his
public opinion never changed.
After this excursus on Kosegarten, we have to return to the question of
how to translate Sanskrit. Early German indologists thought that while other
European languages, notably French and English, were unable to push the
32 Tilman Jens: Goethe und seine Opfer. Eine Schmähschrift. Düsseldorf: Patmos 1999. 33 The
book appeared exactly in time for the celebrations. W. Daniel Wilson: Das Goethe-Tabu: Protest
und Menschenrechte im klassischen Weimar. München: DTV 1999. 34 The passage is, however,
reproduced in Veena Kade-Luthra (Ed.): Sehnsucht nach Indien. Literarische Annäherungen
von Goethe bis Günther Grass. München: Beck 1991, p. 83ff.
boundaries in order to imitate Sanskrit, German was uniquely suited to the
literal translation approach. Statements as the above that Germans were des-
tined to revolutionize translation are often read by non-Sanskritists mainly
as a political statement, as anti-French or anti-British. But while the non-
Sanskritist readers may have caught the undertone right, they usually did not
understand the main sense.35 That is that, while in English word order and
absence of differentiated inflection forces the translator very early to deviate
from the syntax of the original, the German translator can hope to imitate
many more of the Sanskritic features, including even the compounds. The
German poet Friedrich Rückert was the most outspoken proponent of this
literal method. He held the opinion that—I try to paraphrase—we should
not destroy this rich Indian poetical vegetation by severing it into single
branches and flowers and thus destroy the real beauty in it. Through cutting a
long sentence into pieces one destroys its very life. And then he adds: “The
English and the French cannot avoid it, but we can, if we only use our an-
cient Reichskammergerichtsperiodenbau.”36 This word is itself an example of
what he means: it is an awkward sounding German compoud meaning the
construction (-bau) of sentences (-perioden-) as used in a Reichskammerg-
35 Even if one reads the early pronouncement of Schlegel quoted above, we find the following
complementary description of the enormous problems every translator into German is facing:
“Die Sprache der Römer konnte nur durch unsägliche Mühe und Gewalt für die Poesie urbar
gemacht werden, und so hat auch bey uns die Undankbarkeit des Bodens zu einer mühsameren
Cultur genöthigt. Unsre Sprache ist halsstarrig: wir sind desto biegsamer; sie ist hart und rauh:
wir thun alles für die Wahl milder gefälliger Töne; wir verstehen uns sogar im Nothfalle zu
Wortspielen, einer Sache, wozu die Deutsche Sprache am allerungeschicktesten ist, weil sie immer
nur arbeiten, niemals spielen will. Wo sind denn nun die gepriesenen Wundervorzüge, die
unsere Sprache an sich, zur einzig berufnen Dollmetscherin aller übrigen machen sollen? Ein
Wörterreichthum, der gar nicht so überschwenglich ist, daß er nicht beim Uebersetzen oft Armuth
sollte fühlen lassen; die Fähigkeit zusammenzusetzen, und hie und da neu abzuleiten; eine etwas
freyere Wortstellung, als in einigen modernen Sprachen gilt, und endlich metrische Bildsamkeit.”
A. W. Schlegel: “Nachschrift des Uebersetzers an Ludwig Tieck”. In: Athenaeum. Eine Zeitschrift
von August Wilhelm Schlegel und Friedrich Schlegel. Zweiten Bandes Zweites Stück. Berlin 1799,
p. 282f. 36 “Die ganze dichtverwobene Laubmasse einer solchen indischen Vegetation, nach
unserer Art in einzelne Ränkchen und Blüthen aufzulösen, zerstört den eigentlichen Zauber jener
Poesie; man kann einen solchen Satz nicht in Sätzchen zerschneiden, ohne ihm die Sonnen des
Lebens entzwei zu schneiden. Der Engländer und der Franzose können nicht anders, wir aber
können’s, wenn wir auch nur unseren ehemaligen Reichskammergerichtsperiodenbau zu Hülfe
rufen wollen; wir brauchen nichts, als seine Prosa in Poesie zu verwandeln.” (p.141)
ericht, which is a court of justice (-gericht-) in the Holy Roman Empire, thus
meaning a formal language characterized by complicated expressions. The
main problem with this approach is that while many would agree that it is an
ingeniuous imitation of the original, only few would agree that it is poetic.
Even in German, and despite Mark Twain,37 long compounds sound inelegant.
Therefore the German reader of Rückert’s translations occasionally wonders.
We can interpret part of the poem easily: The subject, who views an image
of his beloved, is so drawn to this image that he even disregards the original.
Then follows an arthāntaranyāsa, a comparison to explain the first two lines.
He is like a traveller, who ignores the “Lab-fluth”, which the reader will grasp
as nutrition. But why it is here “a flood of nutrition”, remains unclear. Only
the Sanskritist can know that words like “flood”, “flock”, etc. are used as
plural markers in compounds, but the real question is: Do we need this image
in the target language? What has flood to do with the food? Probably nothing,
because such plural markers do not always fit into the context with their
primary sense. The other hurdle is considerable. No reader of German will
easily understand “Gazellen-dürstung”. “Dürstung” is not a word commonly
used, and Rückert uses it elsewhere in the sense of staying thirsty for a certain
37 Mark Twain, in his well-known essay on The Awful German Language (published as an
appendix to his A Tramp Abroad, Hartford 1880), caricatured almost all grammatical phenomena
that invariably puzzle speakers of English: grammatical gender, separation of verbal prefixes
and of course compounds, on which he says: “Some German words are so long that they have a
perspective [. . .] These things are not words, they are alphabetical processions. And they are not
rare; one can open a German newspaper at any time and see them marching majestically across
the page,—and if he has any imagination he can see the banners and hear the music, too. They
impart a martial thrill to the meekest subject.”
time, as when camels drink only after five days of “Dürstung”. And only the
Sanskritist will recognize that this word translates mr.gatr.s.n.ikā, which means
fata morgana. What Rückert has done is to give us an interpretation of the
not so obvious etymology of the word. It belongs to those translations that,
as Fritze has said,38 one almost understands, as soon as one understands the
original.
Schlegel detested Rückerts translations and exaggerated it in the following
verse:39
Deine Sanskritpoesiemetriknachahmungen
Sind voll von goldfunkelnagelneuen Benamungen
Du überflügelst in wortschwallphrasendurchschlängeltmonostrophi-
schen Oden
Die Weilandheiligenrömischenreichsdeutschernationsperioden.
Deine mit Dank erkanntwerdenwollenden Bemühungen sind höchlich
zu rühmen:
So muß man die Himavatgangesvindhyaphilologiedornpfade beblümen.
So German Indology has or rather had a double legacy: the literary modest
approach of Schlegel and the bold stretching of the natural boundaries of
German as practised by Rückert. Several developments in German Indology—
among which the etymological method in Vedic studies needs to be especially
mentioned—contributed to a rise of the literal translation method. Especially
in the second half of the twentieth century, a ban was pronounced on free
translation, and a so-called concordant translation was held to be the only
valid way of translating from Sanskrit into German. The proponent of this
highly influential school was Paul Thieme, discordant voices, as the one of
Paul Hacker,43 were hardly heard. As a result we now have extremely literal
translations even in popular works not intended for Sanskritists, the most
problematic aspect of which is the wide-spread perception that the more
unidiomatic and incomprehensible a translation is, the more literal and philo-
logically accomplished it must be. Hence it is often more rewarding to read
English translations of Sanskrit works than German ones, the erstwhile ad-
41 “Was ich hier gebe, ist keine Uebersetzung, sondern eine freie Nachbildung. Alle dichterischen
Uebersetzungen sind nur unvollkommene Annäherungen. Die Annäherung kann durch die
Unnachahmlichkeit und Unerreichbarkeit des Originals in eine so weite Ferne verwiesen werden,
daß man dann wohl besser thut, die Sache gar nicht zu unternehmen. Die indische Sprache scheint
mir, ohne alle Rücksicht auf Gehalt und Form der Schriften, ein solches unerreichbares und
unnachahmliches Original zu sein.” 42 “Wörtliche Übersetzungen, mit genauer Nachahmung
der metrischen Form, mochten und mögen für einzelne Proben zweckmäßig sein, um den Lesern
einigermaßen eine Vorstellung von dem Tone des Originals zu geben, so wie man etwa ein
Faksimile von einer Handschrift in Kupfer stechen läßt. Für erzählende Gedichte von größerem
Umfange würde ich aber diese Verfahrensweise nicht empfehlen: ich besorge, die indische Poesie
möchte dabei allzu sehr in Nachtheil gesetzt werden.” 43 Paul Hacker: “Zur Methode der
philologischen Begriffsforschung”. In: ZDMG 115 (1965), p. 294–308.
vantage of the flexibility of German has turned out to be a severe defect as
regards the reception of German translations from Sanskrit.
With this background on the theory and practice of translating Sanskrit
into Western languages, we shall now return to translations from Western lan-
guages into Sanskrit, namely of Cervantes’ famous novel Don Quixote. In the
volume about India’s encounter with Quixote, most authors follow an agenda
based on Romance Studies, and the methods employed, the topics pursued,
and the academic standards are those connected to contemporary “Literatur-
wissenschaft” adapted to an Indian context. Whereas the European scholar
would perhaps use the theories of Auerbach to approach his text, the Indian
scholar might want to use Abhinavagupta’s theories of literature and drama.
But here the application of some of the theories is too haphazard to be mean-
ingful. Speaking on topics like “Cervantes and India” we find the rasa theory
juxtaposed with Vedāntic illusionism, or Bharata with Aristotle,44 a blend of
theories to which all ingredients could have rightly protested, but this is a theo-
retical universe of its own and out of tune with more historical or philological
flavours of research. To quote Abhinavagupta’s Abhinavabhāratı̄ as saying
“fiction should always talk about the archetypal”45 only suggests to us that we
better keep these approaches separate. This also applies to the post-colonial
approach with its all-pervading power axiom, according to which everything
necessarily has a political meaning or implication. In one publication on the
global impact of Don Quixote, the fact that William Jones occasionally read
Don Quixote with his wife46 is interpreted as an unwillingness to share it with
the “aborigines”.47 I shall stop with the review of secondary literature here,
44 Preeti Pant: “The Natyashashtra and the Quixote on the Understanding of Fiction”. In:
Quixotic Encounters, p. 36. 45 Ibid., p. 37. No further bibliographical reference is given.
46 Mentioned in the entry “India” in the Gran Enciclopedia Cervantina, Vol. 6, p. 6184. 47 “Das
bedeutet auch, dass die Hierarchien, die die Machtverhältnisse und das Prestige der verschiedenen
Kulturen im Verhältnis zueinander ausdrücken, sich in einem ständigen Wandel befinden und
durchaus widersprüchlich bleiben. In dieser Situation ist es nicht verwunderlich, dass einer der
ersten dokumentierten Leser des Quijote in Indien anscheinend nichts unternimmt, um seine
Lektüre mit den Einheimischen zu teilen.” See Henriette Partzsch: “Don Quijote globalisiert.
Übersetzung und Universalität eines Klassikers der Weltliteratur”. In: Quijotexte, Quijothemen,
Quijotheorien. Acht Annäherungen an Don Quijote. Ed. Marco Kunz. University of Bamberg
Press 2009, p. 135. Of course, the basis for this interpretation is Edward Said.
although many far-reaching interpretations could be quoted.48 Common to
all articles in the publication on Cervantes in India—with the exception of
two by Ganguly—is the fact that none deals with individual translations, nor
with the sources. The English translations actually used are never touched
upon. It seems no attempt is made to deal with petty philological details.
It is therefore no wonder that the translations of the Don Quixote into
Sanskrit and Kashmiri have remained almost unknown. These translations
were produced by the Kashmirian Pandit Nityananda Shastri and Jagadhar
Zadoo between 1935 and 1936, and the credit for making this known goes
to S. N. Pandita, who described these translations briefly in a chapter of
his book on Marc Aurel Stein.49 The background given there is roughly as
follows: When Stein was trying to acquire funds to continue his researches, the
Harvard Sanskritist Charles R. Lanman tried to find support for his Central
Asian tours from within the university. Thus, during the years 1928 and
1930, Stein was invited to give lectures on Central Asia in Harvard and the
University eventually funded his expeditions. It was in Massachusetts that he
met Carl Tilden Keller (1872–1955). After studying in Harvard, Keller worked
as an accountant in Boston, and was a renowned book collector interested
among other things in translations of Cervantes’ Don Quixote into various
languages of the world. The Houghton Library of Harvard University houses
his collection, which contains one section devoted to his exchanges with
Stein.50 In his article Pandita does not explicitly disclose where some of
his source materials, especially the exchange of letters between Stein and
Nityananda Shastri, are kept, but it is to be assumed that they remain in the
Nityananda Shastri Library Collection in New Delhi.
48 Regarding the Urdu translations one author comes to a wildly speculative result: “Für die
westliche Leserin ohne Urdu-Kenntnisse stellt sich hier die faszinierende (und ohne Antwort
bleibende) Frage, ob Sarshar den Quijote vielleicht nicht deshalb übersetzt hat, weil er sein
eigenes Schreiben beeinflusst hat, sondern weil er in ihm so etwas wie einen Nachklang der
indischen Wurzeln des europäischen Erzählens gefunden hatte – hier sei an die Bedeutung
von Calila e Dimna für die Entwicklung der spanischen Erzählformen erinnert.” Ibid., p. 138.
49 S. N. Pandita: Western Indologists and Sanskrit Savants of Kashmir. New Delhi: Siddharth
Publications 2002. See there Kashmiri and Sanskrit Translations of Don Quixote, p. 269–287.
50 Carl Tilden Keller Collection Concerning Sir Aurel Stein (MS Am 2532). Houghton Library,
Harvard University. There are also letters by Stein, correspondence of Keller with others about
Tibetan and Mongolian translations of Don Quixote, etc.
The drawback of Pandita’s description is that in order to provide a contin-
uous and edifying historical narrative he sometimes adds his own interpreta-
tions and makes it more difficult to ascertain the plain facts:51
In fact we know the actual motive for and history of the Sanskrit and Kashmiri
translations of Don Quixote from a letter of Stein to Nityananda (16.10.1935),
as reported by Pandita:
In March 1937 Stein acknowledges the receipt of the translation, but adds a
further request: “I also returned the Sanskrit translation to Pandit Zadoo
with the request to have it properly written in the old Kashmiri Devanagari
style. He acknowledged the detailed instruction but up to the present, the
desired copy has not yet come to hand. I should be grateful if you would kindly
remind Pandit Zadoo about this work [. . .] The fair copy is in Mr. Keller’s
hand, who would prefer to keep it. Please ascertain whether the printing can
be done.”53
In May 1937 he acknowledges the receipt of the manuscripts, but nothing
followed upon it. Perhaps all attempts to publish it came to naught during
the Second World War. What remained are the Harvard manuscripts of the
translations.54
The scribe, who produced the fair copy, was apparently not very conversant
with Nāgarı̄, since he commits a number of mistakes typical for transcriptions
from Śāradā. Mostly these have been emended by the second hand in red
ink, which is most probably that of Stein himself.55 Prefixed to the Sanskrit
translation is a note by Stein and by Zadoo:
Nityananda Shastri did fare well with this; he decided to retain what he could
not translate, but then added something familiar to clarify the context:
The Sanskrit reader will notice through this that the European prayers with
their unfamiliar names were not unlike mantras. They were used for purifica-
tion and to this end had to be accompanied with a statement of the application
(viniyoga). Another such clarification concerns the cross, which, as the trans-
lator adds, is used maṅgalārthe. There are many other interesting loan-words
like paim
. ta “pint” or raun.d.-t.ebal vı̄ra “knights of the round table”.
Here is a short passage from the English version of Jarvis with its Sanskrit
translation:
The passage is interesting in many respects.
1. The Kashmirian translator could not know that the “Sable Mountain”
is the Sierra Morena. He interprets “sable” as a proper name and tran-
scribes it. This is one of the instances, through which one can identify
the English sources of this translation, for many authors do not trans-
late Sierra Morena as “sable mountain”. After some reading our guess
was that the most likely candidate is the translation of Jarvis, which
was for two hundred years the most popular English translation of the
work.56 But Jarvis’ translation, first published in 1742, was re-edited
56 See Cervantes: Bulletin of the Cervantes Society of America. 30.2 (Fall, 2010): 207–209.
with modifications to the wording every few years. Until 1936 there were
nearly hundred re-editions. It is therefore no surprise that for a long
time none of the versions available to us—often in the form of scans
from Keller’s library—were an exact match. After a protracted search
Dragomir Dimitrov recently indentified as the most likely candidate the
edition in the Oxford World Classic Series (Nos. CXXX und CXXXI)
with a preface by James Fitzmaurice-Kelly. It was published first in
1907 with regular reprints, and Stein might have preferred to send this
inexpensive and readily available publication to Kashmir rather than
any older version of Jarvis.
The whole project of editing D.ān Kviks.ot.a began with leisurely afternoon read-
ing sessions, which were pursued for a few semesters by Dragomir Dimitrov,
Stanislav Jager, Maximilian Mehner, and myself, occasionally including other
scholars as Martin Straube or visiting scholars like Chandra Bhushan Jha
(Delhi) and Shrikant Bahulkar (Pune). They combined interests of the vari-
ous participants, in Spanish, in Modern Sanskrit, manuscript reading, etc.,
but in the end it yielded more interesting results than expected. Some of
these, which—it must be emphasized here—are the result of our collective
endeavour, have been summarized here. Further work on the edition and de-
tailed analysis of the Sanskrit translation now rests in the hands of Dragomir
Dimitrov.
From a modern Indian perspective a translation of Don Quixote into San-
skrit may seem negligible. The author of the first translation from Spanish
directly into Hindi, Vibha Maurya, says in an interview57 that “there are trans-
lations of El Quijote into Bengali, Marathi, Tamil, Kashmiri, and Hindi. But
always made from English.”58 The earlier Hindi translation by the Sanskritist
Chavinath Pandey (1964) she considers to be unsuitable for the modern reader,
being written in a highly sanskritized Hindi.59 So it seems that the Indian Cer-
vantes experts have failed to notice that this translation is a singular instance
of an unusually literal translation of a European classic into Sanskrit. Whether
this is the first translation into Sanskrit that is not in the classical adaptive
style, I am unable to say, since to my knowledge the history of translations of
European works into Sanskrit still needs to be written.
57 Interview conducted in November 2007 by Ma. Teresa Elizarrarás for Revista de Estudios
Cervantinos 4 (2007–2008). 58 “Se debe mencionar que existen traducciones de El Quijote en
bengali, marathi, tamil, casamiri y en hindi. Pero siempre hechas del inglés.” Ibid. 59 “En el
año 1964 la Academia de Letras de India publicó la primera traducción hecha del inglés por un
profesor de sánscrito, el Dr. Chhavinath Pandey. La traducción contiene errores de comprensión
del original, así como el lenguaje es arcaico, sanscritizado, por eso poco asequible.” Ibid.
“Surpassing B ān. a’s Style” — Two Kashm irian Authors of the
Nineteenth Century
Of our next author, Sāhib Rām, we do not know a date of birth, but he probably
died in 1872. He was employed at the court of the Kashmirian Mahārāja Ranbir
Singh,60 and for providing context for his works some historical facts may
be of help. Ranbir’s father, Gulabh Singh, had been officer in Lahore, in the
army of the famous ruler of the Sikhs, Ranjit Singh, whose realm at the time
included Kashmir. Ranjit Singh was apparently fond of the two Dogra princes
Gulabh Singh and his brother, and appointed them as local rulers to those
areas from which the family came from, that is, Jammu and Kashmir. When
Ranjit Singh died almost the whole of his family including potential heirs to
his throne were killed in the fights about succession. In this phase Gulabh
Singh, being as it were a Rāja without a functioning Mahārāja, could regain
control over Jammu and Kashmir, and when the British, taking advantage of
the confusion over the Lahore throne, defeated the Sikhs in 1846, he could
by means of skillful diplomacy negotiate some independence of Jammu and
Kashmir. To the British, who were outwitted by his skills, he became the
“Jammu Fox”,61 whereas from an Indian perspective he was a clever ruler, who
managed to retain his distance from the British. This phase in Anglo-Indian
history is of course well-known, even in public memory. The golden throne of
Ranjit Singh is on display in London in the Victoria and Albert Museum, and
Ranjit Singh’s most famous diamond, the Koh-i-Noor is now part of the British
Crown Jewels. A large number of travel accounts gives a lively impression of
this phase in Indian politics.62
Gulabh Singh’s third son—the first two fell prey to assaults during the
fights about succession—was Ranbir Singh, who succeeded his ailing father
60 See for instance Sukh Dev Singh Charak: Life and times of Maharaja Ranbir Singh, 1830–1885.
Jammu 1985. 61 Bawa Satinder Singh: The Jammu Fox. A Biography of Maharaja Gulab
Singh of Kashmir, 1792–1857. Carbondale 1974. Compare also Robert A. Huttenback: “Gulab
Singh and the Creation of the Dogra State of Jammu, Kashmir, and Ladakh.” In: The Journal
of Asian Studies 20. 4 (1961): 477–488, and Bawa Satinder Singh: “Raja Gulab Singh’s Role
in the First Anglo-Sikh War.” In: Modern Asian Studies 5.1 (1971): 35–59. 62 A particularly
entertaining description of this time is given by the Transsylvanian traveller Martin Honigberger.
See Johann Martin Honigberger: Als Leibarzt am Hofe des »Löwen vom Panjab« Ranjit Singh.
Nachdruck der Reiseerlebnisse, Wien 1853, mit einem Nachwort von Jürgen Hanneder. Halle 2011.
in 1856. When a year later the Mutiny endangered British sovereignty in India,
Ranbir Singh agreed to send Kashmirian troops to assist the British in quelling
the uprise. Cleverly, he did not expect a reward, but it was understood that
his rule in Jammu and Kashmir would be accepted by the British.
The reign of Ranbir Singh was not without problems: in 1872 there was a
conflict between Muslim fractions who dominated the economically impor-
tant weaving of the Kashmirian shawls. When the European market for these
fabrics collapsed after the French-German war, there was an uprise of the
weavers. There were also several famines and earthquakes with terrible conse-
quences for large parts of the population. His reign also saw a kind of revival
of Hindu culture. He founded the Raghunatha Temple Library, which houses
a large collection of manuscripts copied often from Srı̄nagar archetypes. He
opened schools, employed teaching staff and encouraged the translation of
important works63 in order to account for the multilingual culture of Kashmir.
While it seemed to some authors within Iranian Studies that Sanskrit had at
the time become negligible as a source language,64 this does not hold true for
Ranbir Singh’s time.
Also the research on older Kashmirian literature received a great impe-
tus through the activities of Ranbir Singh. Many works that now belong to
the most interesting areas within Sanskrit literature were discovered when
scholars made research tours to Kashmir and were assisted by the Kashmir
government. The most well-known report was made by the German indolo-
gist Bühler.65 Bühler, equipped with letters of recommendation from a former
employee of Ranjit Singh, was received by Ranbir Singh himself, a meeting he
describes as follows:66
63 “A special feature of service in these institutions was the teacher’s ability to translate. A part of
the salary in each grade was fixed as remuneration for translation work [. . .]” Sukh Dev Singh
Charak: Life and times of Maharaja Ranbir Singh, 1830–1885, p. 255. 64 Compare Siegfried
Weber: Die persische Verwaltung Kaschmirs (1842–1892). Wien: Verl. der Österr. Akad. der Wiss.
2007. Vol. 1, p. 56f. 65 Detailed Report of a Tour in Search of Sanskrit Mss. made in Kásmír,
Rajputana, and Central India. Bombay/London 1877. 66 Op. cit., p. 3f. 67 The transliteration
is strangely wrong, it should read ran.avı̄ra.
it. He is also versed in the Śâstras, especially in Vedânta and
Dharma, on which latter he is said to have composed a treatise.
He received me very kindly, and gave orders that all Pandits whom
I might wish to see should be asked to visit me, and that every
assistance should be given me. He was also good enough to take
me to his Mudrissa, and to allow me to examine some of his
pupils in his presence. The active manner in which he took part
in the examination showed that he was well acquainted with the
subjects taught, and that he took a real interest in the work of
education. This Mudrissa, which is the chief educational institu-
tion in Kaśmîr, contains, besides a Sanskrit college where poetry,
poetics, grammar, and philosophy are studied, Persian classes
and a school of industry. Mathematics are taught, according to a
Dogra translation of the Lîlâvatî. Its head is Pandit Râmjîv, the
son of Pandit Râjkâk, who combines the office of Superintendent
of Education with that of the revenue officer [. . .] I examined
several classes in Sanskrit, Euclid, and algebra, and most of the
boys did very fairly.
The quotation confirms the multilinguality in his reign. The textbook for
mathematics is in Dogri, perhaps the most practical, down-to-earth language
used. But also Sanskrit as well as Persian were cultivated; English is not
mentioned. The parallel use of Sanskrit and Persian naturally lead to questions
about the underlying social realities, and Bühler, when he enquired about the
two spheres, was told that both were totally apart: “When I first inquired into
the relations between these several sections of the Kaśmîrian Brahmans, I
was told that the Sanskrit-studying and the Persian-studying Pandits did not
intermarry.”68 But that was apparently only the ideal, because he adds: “Later
my informants recollected cases of marriages between children of officials
and of the men of Śâstras, and they modified their statement accordingly.”69
The time of Ranbir Singh is interesting, because in this period the foun-
dations were laid for the various editorial activities that culminated in the
enormous canon of Kashmirian works available now. Without it we would not
know of the long and continuous tradition of Indian historiography, and an-
70 Hilko Wiardo Schomerus (1879–1945) was Professor for the History of Religion in Halle.
71 See now Luther Obrock (ed.): Marc Aurel Stein – Illustrated Rājataraṅgin.ı̄. Together with
Eugen Hultzsch’s Critical Notes and Stein’s Maps.. Halle 2013. 72 See Gerard Clauson:
“Catalogue of the Stein collection of Sanskrit Mss. from Kashmir.” In: JRAS 1912, p. 578–627.
73 Jürgen Hanneder: “Modernes Sanskrit. Eine vergessene Literatur.” In: Pāsādikadānam.
Festschrift für Bhikkhu Pāsādika. Hrsg. Martin Straube et al. Marburg 2009, p. 205–228.
point I merely want to say that in the cases I shall present here judgement was
passed before the accused had even been heard. These texts have not been
made the object of research before.
In contemporary Indology there is the wide-spread notion that modern
Sanskrit can in no way match the sophistication and perfection of classical
Sanskrit and that modern authors are at best second rank. For this reason
it is worth reading what Bühler had to say about a poet whom he visited in
Kashmir:
74 Bühler: Detailed Report, p. 26. 75 Judging from recent German examples, the focus is
usually on religion, often Buddhism is required, rarely Hinduism. Older standard specializations,
as Vedic studies, are no more in focus, specializations that would seem natural in other philologies
(as, for instance, literature) seem out of the question, although it is quite unclear, why this is so.
some of the unknown pieces in Stein’s collection of Kashmirian manuscripts.
The first result was the so-called Fifth Rājataraṅgin.ı̄ of Dāmodara edited by
my student Bidur Bhattarai in Marburg,76 a continuation of those of Kalhan.a,
Jonarāja, Śrı̄vara including the editions by Śuka or Prājyabhat.t.a. While the
work was very interesting, it did not display the sparkling style Bühler was
talking about. Stein’s copy was prepared directly from the autograph, and it
seems that Dāmodara passed away before finishing the work. Stein writes:
“At the time that these passages are passing through the press news reaches
me that this most learned and amiable of all Kashmirian scholars has fallen a
victim to the epidemic now raging in the Valley [. . .]”.77 The apograph seems
to reproduce the author’s copy in much detail: we find that in few places the
text is given in two columns, one contains a prose version of the text, the other
a versified one. Most probably Dāmodara formulated his ideas in prose first
and then put them into Ślokas, for the main part of the text itself is metrical.
This is one of the rare and interesting cases, where we can look into an author’s
workshop!
Then we have a so-called “letter-writer”78 attributed to Dāmodara, which
Stein describes as follows: “Specimens of letters, composed by the late Pandit
Dâmodar by order of Maharaja Ranbir Singh. Original copy, sold to me by
Pt. Mahânandajîva, son of Pandit Dâmodar, who attests that this letter-writer
was adapted from a Persian text. Srinagar Sept. 30, 1892 M. A. Stein.”79 If we
look at this text, we find letters addressed to various recipients, to the “son”
(apatya), to the head of a village, to the king (rājānam . prati vijñaptipatrikā
fol. 8), to a father (fol. 10ff.), uncle, guru (16ff), the rest is again devoted to an
apatya. Since there is more than one hand involved and the quality of the text
is in some parts problematic, it is doubtful whether this was really, as attested
by Pandit Mahānandajı̄va, the “original”.
Then there is another letter-writter attributed to his father, but partly
written by the son, which is called “Lekhaśiks.ā śrı̄pan.d.itasahibrāmakr.tih.”.
In this manuscript the name sahibrāma stands over a deleted dāmodara. An
explanation is given in Stein’s description: “This copy was prepared from the
76 Master thesis. Marburg 2010. 77 Preface to Marc Aurel Stein: Kalhan.a’s Rājataraṅgin.ı̄.
Vol. 1. Sanskrit Text with Critical Notes. Bombay/Leipzig 1892, p. xviii, note 1. 78 For the genre,
see Ingo Strauch: Die Lekhapaddhati-Lekhapañcāśikā: Briefe und Urkunden im mittelalter-
lichen Gujarat. Berlin: Reimer 2002. 79 Ms. Stein or. c. 10, last folio.
original MS. of the author (written partly by him partly by his son Pandit
Damodar) in September 1892. M. A. Stein.”80 From the copy we cannot
know where the juncture was. It is written in Devanāgarı̄ and has quite a few
corrections by a second hand, which concern usually omissions and misspelt
letters.
For possible candidates for the artistic style Bühler was praising, there
are furthermore the fragmentary historiographic materials written not by
Dāmodara, but by his father Sāhib Rām, which have been edited in another
Marburg project by Anett Krause.81 After perusing these materials, it is
possible to make sense of Bühler’s enthusiasm.
The second Lekhaśiks.ā, which is attributed to both authors, is written in a
highly artistic Praśasti-like style, which indeed reminds one of the classical
prose novel in Sanskrit. The assumption that such a work, in an environment
dominated by a persophone administration, was fed by the Persian genre of
inšā � must have been obvious to Kashmirians,82 for since Mogul times there
were munšı̄s trained with Persian handbooks.83 In this Persian epistolography,
however, one distinguished an Indian and a Persian style,84 and since such
scribes were also Hindus, the functions of the Munshi and the Kāyastha
could easily overlap. Whatever the source of Sāhib Rām’s and his son’s letter-
writer,85 both appear to be an attempt to re-establish an official Sanskrit
correspondence, which would accord with the other attempts at Ranbir Singh’s
court to support Sanskrit.
More efforts would have to be made to establish the sources for the
letter-writers, but this is—precisely because of the sophisticated style—not
a straightforward task, since the work is puzzling and at the same time in-
80 Ms. Stein. or. d. 34. 81 Sāhibrāms Arbeiten zur Geschichte Kaschmirs: Erstedition und
Analyse ausgewählter Textstellen. Ph.D. dissertation, Marburg 2016. 82 See Adrian Gully:
The culture of letter-writing in pre-modern Islamic societies. Edinburgh: Edinburgh University
Press 2008. And: Ishtiyaq Ahmad Zilli: “Development of Insha literature to the end of Akbar’s
reign”. In: The making of Indo-Persian culture: Indian and French studies (2000). 83 Tuhfat
al-hind, Nigārnāma-yi munšı̄. See Alam & Subrahmanyam: “The making of a Munshi”. In:
Comparative Studies of South Asia, Africa and the Middle East 24/2 (2004), p. 61–72. 84 See
Moin Mohinuddin: “Sabk-i-Hindi”, Indo-Iranica (Calcutta) 1959. 85 The genre seems to have
a certain continuity in Sanskrit, for which see Strauch: Die Lekhapaddhati-Lekhapañcāśikā.
The introduction in Pushpa Prasad: Lekhapaddhati. Documents of State and Everyday Life from
Ancient and Early Medieval Gujarat. New Delhi: Oxford University Press 2007 is also worth
reading.
teresting on many accounts.86 Even without being able to judge the Persian
literary tradition myself it was my impression that the work can be explained
in continuity of ornate Sanskrit prose.
As an example I give the beginning of one such polite address to an
unnamed recipient (Ms Stein or. d. 34, fol. 16v):
svasti śrı̄madanugun.agan.ānavadyasvavidyādyudbhūtaprabhū-
tasadyaśah.stomasomāvadātı̄kr.tāśes.āvakāśes.u madı̄yamano-
vihagārāmes.u śrı̄maddhı̄vānasāhibhopākhyaśrı̄śrı̄śrı̄matkr.pā-
rāmes.u madhusūdanadevadattaprasādadattā aśis.ām . tatayah.
samullasantatarām
Other letters of the same collection display an even more elaborate string of
compounded expression: one example, of which I quote only the beginning,
is to be found on folio 14:
91 For the details on the Persian source I am indebted to Christoph Werner and Anna Martin.
92 S. R. Sarma: “From Yāvanı̄ to Sam . skr.tam. Sanskrit Writings inspired by Persian Works.” In:
Studies in the History of Indian Thought 14 (2002) [Department of Indian Philosophy. Kyoto
University], p. 71–88. 93 He refers to Akbar Shah’s Śr.ṅgāramañjarı̄, which is a scholarly reaction
to Bhānumiśra’s Rasamañjarı̄. Ibid., p. 77. 94 Ibid, p. 79f.
translations into Sanskrit. First, there is Śrı̄vara’s Kathākautuka (1505), an
adaptation of the Persian classic Jusuf-u Zulaikha, which has received some
attention because of the pioneering studies by Richard Schmidt.95
The Delārāmākathāsāra
95 Richard Schmidt: Das Kathākāutukam des Śrı̄vara verglichen mit Dschāmı̄’s Jusuf und
Zuleikha nebst Textproben. Kiel: Haeseler 1893. And his: Śrı̄vara’s Kathākāutukam: Die Geschichte
von Joseph in persisch-indischem Gewande. Sanskrit und Deutsch. Kiel: Haeseler 1898. 96 Edited
as volume 77 of the Kāvyamālā Series. Bombay: Nirnaya Sagara Press 1902. 97 The title of
rājānaka was given to members of the families of Kashmirian ministers. 98 The term is used
here not in a modern sense implying a “literal translation”, but as encompassing a variety of
types of pre-modern adaptations or acculturations. 99 For the following I would like to thank
Ulrich Marzolph for providing me with the crucial materials, and Anna Martin for helping me
with many Persian related questions. 100 I am a bit puzzled by the word vācya here. It is not
very likely the author means to refer to the “blemishes” of the work.
The work is therefore unlikely to be much older than the time of Śrı̄vara, but
no further clue has surfaced up to now. S. R. Sarma, to whom the credit for
pointing out the importance of the work must go, says about it:101
The poem deals with the adventures of Ibrahı̄m and Murād, the
two sons of the Sult.ān Muh.ammad of Aleppo. The book is named
after a courtesan Dil-ārām who plays an important role in the
story. The fortunes of the princes depend on the possession of a
bird with a magic heart.
The work indeed contains the internationally wide-spread motif of the The
Magic Bird-Heart, which has been studied extensively by Aarne in his dis-
sertation102 and is classified as AaTh/ATU 567.103 Elements of the story are
known from Sanskrit sources, for instance, the boy who finds gold under
his pillow every day and becomes king is related in the Kathāsaritsāgara, a
fairly complete version is contained in the Kathāratnākara (to be discussed
below), and there are further Kashmirian non-literary versions.104 There is
a Persian version in Nahšabı̄’s T.ūt.ı̄-Nāma,105 but while it shares motifs, it
ˇ
is more straightforward and there is no exact match for many of the motifs
contained in the Delārāmākathāsāra.
Since there is neither translation nor analysis of the work, I shall give a
summary below, but it has to be noted that the text as presented by the only
edition is defective—there is a gap between chapters 2 and 3, which the editors
identify exactly.106 Occasional question marks by the editor referring to a
corrupt reading as well as emendations in brackets suggest that the editors
did not have access to a second manuscript, at least this is as much as we can
say in the total absence of any introduction.107
101 Sarma: “From Yāvanı̄ to Sam 102 Annti Aarne: Vergleichende Märchen-
. skr.tam”, p. 84.
forschungen. Helsingfors 1908, p. 143–200. 103 I am very grateful to Ulrich Marzolph for
identifying the story and providing me with his analysis. See Ulrich Marzolph: “Vogelherz:
Das wunderbare V.” In: Enzyklopädie des Märchens 14.1 (2011), p. 292–295. 104 Aarne: Ver-
gleichende Märchenforschungen, p. 168. 105 See The Cleveland Museum of Art’s Tūtı̄-Nāma.
. .
Tales of a Parrot. By Ziya’ u’d-din Nakhshabi. Translated and edited by Muhammed A. Simsar.
Cleveland/Graz 1978, p. 323–330. 106 Apparently the two last verses (24–25) of the second
chapter are missing and the first 12 and a half verses of chapter 3. 107 The reader should note
that, since this is a Kashmirian text, foreign words spelt with an “e” may well transcribe “ı̄”, so,
for instance, Merabhaktā might also have been pronounced Mı̄rabhaktā.
�
The Story of Delārāmā (Summary)
The work starts with an unusual invocation, in which the deity is simply called
khadaiva, a word which is not in the normal (Hindu) vocabulary,108 but seems
deliberately used to avoid giving the “Muslim” story a Hindu frame. The
second verse quoted above gives “Muslim works” as the source, states that the
text is a translation or adaptation and that its purpose is entertainment.
The work itself commences with a description of the famous country
Halābha (1.3–14)—already identified as Aleppo (H.alab) by Sarma—as a pros-
perous and peaceful realm by employing a figure of Sanskrit poetics, which
involves a type of hidden praise (vyājastuti), that is by describing a defect in
such a way that it turns out to be a virtue. For instance, it is said that this
country was not free from sorrow/concern (cintā), but it was only the concern
for embracing women, drinking alcohol and games; quarrel was only found
between clamouring birds, rage only between co-wives, and so forth.
With the images employed in the first description the author sets the
tone right from the beginning: The reader is to expect a tale involving erotic
episodes—most of the images point in this direction. Immediately the reader
becomes aware of the fact that this is no plain and simple narrative,109 but
that the change of metre for dramatic effect, the insertion of descriptions
of a country, or more often of a woman, and the alternating of such highly
poetic passages with others, in which the progress of the plot is narrated in
less ornate language reveals a conscious effort on the side of the author to
transform his sources into a piece of poetry.
In this country there is a Sultan110 Muhammad and his wife Merabhaktā.
They are introduced with the usual polite words, but in view of the prevailing
conventions of exuberant praise of the monarch this sounds rather lukewarm
(1.15–16). The “good” monarch, being too attached to love games with his
wife and other lovers,111 does not care for his kingdom—thereby providing a
completely different angle to the introductory praise of the lack of concern in
108 The only instance of a use of that word that I could find is in Śuka’s Rājataraṅgin.ı̄. 2.79
(Srikanth Kaul (ed.): Rājataraṅgin.ı̄ of Śrı̄vara and Śuka. Hoshiarpur 1966). 109 As the
parallel in the Kathāratnākara, for which see below. 110 Sanskritized as suratrān.a. 111 This
is explicitly emphasized in 1.17a.
his country! This goes on until his best ministers also become disinterested
(1.18).
Then one day a barber brings a mirror to his courtyard, and when he looks
at his own grey hair and beard he is shocked that he has not accomplished
anything worthy of a king, but wasted his best years (1.21). Frustrated he goes
on a hunting party, where “on the way his sharpened arrows strike down to
earth many living beings” (1.23). He strays from his party, reaches a cavern
and sees a deer, shining and fearless.
Here, until the end of the chapter (that is, 1.25–35), a different metre is
employed to emphasize the crucial encounter that follows: “When the king
saw her, he immediately drew the sharpened arrow112 to strike her. But the
deer paused and spoke in human voice to the startled king: “O Lord! Today
I have come for your sake, therefore relax your drawn bow.113 I shall tell you
something [. . .]” (1.25–26).
The deer says it has descended from heaven and predicts that the king
will lose all his belongings because of his bad karma. He is presented with
the option to either suffer it at once, or go back to his kingdom and wait until
it has ripened, implying that then the result will be much worse. The king
is undecided and wants to ask his wife first. The deer agrees, he returns to
his city, brings in Merabhaktā, dismisses all attendants and relates with tear-
choked voice the words of the deer (1.31). Merabhaktā is greatly troubled but
argues that it is better to face fate as long as there is enough power in his body
to endure such hardship (1.32). The king leaves in the morning and under the
pretext of going hunting, leaves his servants, ministers and soldiers behind.
With this the first chapter ends. Perhaps the chapter change is introduced for
dramatic effect, but it does not accord well with normal Kāvya conventions,
according to which a chapter is often in a single metre, whereas the last verse
is typically marked by a different metre. In our text there is such a change in
the last verse,114 but then the new metre continues into the second chapter.
The king meets the deer and tells her that he accepts his misfortune today
(2.1). The deer is astonished and disappears. Now the karma takes immediate
effect. First his horse suddenly dies. The king, now walking the wilderness
112 The Sanskrit expression is literally “to draw the arrow from (abl.) the string”. 113 Lit.:
“arrow and bow”. 114 The encounter with the deer is phrased in Vasantatilakas, the last verse of
the first chapter is in Mālinı̄.
with lowered head, when approaching his camp (kat.aka), sees a party of
armed hunters approach115 (2.3). Faced with this, all his companions116 take to
flight on their “excellent horses”, leaving the king behind on foot (2.4). When
the whole army is gone,117 the wild tribesmen118 see the poor pedestrian, and
bind him. They take his ornaments, beat him and leave him covered in blood
(2.6). Back in the city, his former comrades quickly entrust the kingdom to
someone else (2.7).
When the king eventually returns to his city, he sees his beloved wife
Merabhaktā crying and with sullied clothes roaming the streets in fear (2.9).
She tells her husband with tear-choked voice that the kingdom has been
very quickly given by the ministers to another king who came to town (2.12)
and that she has to flee—as her husband, who would surely be killed by the
enemies (2.13).
Thereupon the couple emigrates to a land that is neither virtuous nor
characterized by bravery, where not even poetry is held in esteem (2.17).
There the king, whose body is “soft”, that is, not used to hard work, has to
resort to a detested livelihood worthy only of the lowest castes. Every morning
at sunrise his wife gives him bread to eat on the way and he returns only in
the evening from his work in the forest to sell fire wood119 in the town market.
After some time his wife gives birth to two sons, who bear the marks of kings.
Of chapter 3 unfortunately the first twelve and a half verses are missing,
and the gap in the story can only be tentatively filled from what follows. The
third chapter as transmitted starts within the conversation of the queen with
a merchant, who pays her three hundred dı̄nāras for an egg-shaped jewel120
and promises to pay her the same amount for more similar jewels (3.16). Now
it is said, and this shows what must have been the content of the lost verses,
that on the next day she approaches the bird cage with great hope and finds
another incomparable shining egg (3.21). The beginning of the third chapter
must have contained the story how the queen acquired the magical bird. The
queen sells one jewel egg daily and acquires wealth so that she can tell her
115 One would better read vyādhasenājagāma for vyādhasenā jagāma. 116 One may take
āmātya here as different from the sacivas he is said to have left behind (1.35cd) in order to avoid
contradiction. 117 Here the story remains confused: if it is his army, then he did not come alone
as said before. 118 Called variously vyādha (3d), śabara (4b) and pulinda (5a). 119 Clarified
later, see 3.30: indhana. 120 man.im an.d.am.
husband that there is no more need to pursue his exhausting job as a wood
collector (3.28). The couple regains wealth and even employs servants.
But the queen meets the merchant regularly to sell the jewels and brings
food and alcohol from her house, and they start a sexual relationship. Blinded
by passion the queen does not even think of her two royal sons any more, and
the chapter ends with the narrator’s lament that the bad behaviour of women
has no end.
In the fourth chapter we hear that in the house of the merchant there was
an old woman, who was a witch.121 Seeing the coming and going day and night,
the old woman is amused and uses all arguments to convince the merchant
to tell her what she so eagerly wants to hear. The merchant cannot keep the
secret, excusing himself by the remark that the mind of women is unsteady as
an ape (4.6cd). He tells her that she is the wife of a disrobed king and that she
sells her jewels to him, but that now she has become more closely attached.
The old woman responds with a challenge: if she is really so attached he
should test her by asking the bird in her house to be slaughtered and brought
for food (4.11). The merchant is convinced she would do so, but the witch
smiles and says that he does not even know the truth about all this—here the
reader will rightly guess that the witch knows about the magical bird.
When the queen comes on the next day the merchant acts distant and
when the queen apologizes for any offence and proclaims herself his slave, he
explains that he does not like the food she brings, and asks her scoldingly122
not to hide a bird in her house, but to prepare it as the next meal.
She kills the bird without mercy (4.28), cooks it and hides the meal
guarded by one female servant (4.29). When her two sons, the princes who
had played outside, return to the house and ask the servant for something to
eat they end up accidentally eating the magical parts of the bird: the older
brother eats the head with “the bone” (asthi), the younger brother eats the
breast (vaks.as).
The servant relates the matter to the queen, she brings the remainder of
the meal to the merchant and the old women eats the rest of the bird, but
without the crucial magical parts. After a longer description of the happy
121 If that indeed is the intention of the word Śākinı̄. This magical element in the story is needed
to explain why she knew the secret of the bird. 122 Absurdly he calls her pum . ścalı̄, one “going
to another man.”
reunion of the queen and the merchant, the old woman comes to know that the
head has been missing and here the text supplies to the reader an important
detail,123 the prediction that whoever eats the head becomes king (6.21). More
specifically, as the old woman explains, it is the “bone” (asthi) in the head and
the breast of the bird (6.27). Since it cannot be digested by anyone after having
eaten it, it must be still in the body of the young princes. When she comes to
his house, the merchant again feigns anger, and she regrets having killed the
bird that had brought her 300 dı̄nāras a day. The merchant now demands that
she kills her sons (6.4), which she accepts. After the lament of the narrator
about the blinding desire of women, he emphasizes his point through another
description of love-making, at the end of which the queen goes home, firm in
her resolve to kill her sons and promises a servant freedom if he commits the
deed (6.17).
The sons convince the servant to let them go by promising him thou-
sand dı̄nāras—the elder son had in the meantime become rich through the
bird’s bone he had involuntarily swallowed—and provides him with a written
document setting him free. The servant and the two sons flee to another
country.
Now, in chapter 7, the focus of the story shifts to the princes, who reach a
town where the king had just died. There the ministers receive a prediction
in their dreams that Ibrāhı̄ma, the son of Sultan Muhammada would present
himself and should be entrusted with the kingdom (7.3f.). The two princes
wandering around enter the town and sit down in a mosque (masedā).124 They
decide to split up, the older would remain in the town, the younger would go
further afield, and they agree to meet in the mosque in the evening.
When the older prince reaches the palace and tells his story, the ministers
immediately make him king (7.13). Blinded by the affluence of his new position
he forgets about meeting his younger brother, who in vain waits in the mosque
in the evening. It is only here that we are told the name of the second brother
Morāda Bhaks.ah..
When the new king, after missing his brother in the mosque, searches for
him, he reaches a park, sits down under a tree and sees a beautiful woman
123 This might already have been stated in the lost part where the queen acquired the magical
bird. 124 The later Rājataraṅginı̄s transcribe masǧid as masjedā (Dāmodara) or masjyedā
.
(Śrı̄vara).
up in a mansion. It turns out that she is Delārāmā, a prostitute, (7.27) and
a bawd tells him that one night would cost him three hundred dı̄nāras. The
prince pays for the same day and when the bawd wants to throw him out the
next morning, he pays day after day, and the courtesan, wondering about the
source of money, decides to deceive him (7.44).
Delārāmā tricks the unsuspecting prince, in a moment when “his whole
mind was taken away by alcohol intoxication”,125 into telling her where his
money came from. When he is unconscious from drinking, she administers
an emetic and retrieves the bird’s “bone”. When he cannot pay for the next
day, he is thrown out by the servants.
Now he wanders in the woods without any belongings left, frustrated
about what the prostitute had done to him, without the means of livelihood
and wishing to die, when he sees three men arguing with each other, as it
turns out, about magical objects: a purse, which gives three hundred dı̄nāras
to the person holding it, and a parrot in a cage, which can magically transport
the owner to the place of his wish. The prince tricks the three men and uses
the bird to go to Delārāmā: In an instance, after pronouncing his wish, he
finds himself near her bed (8.16). Delārāmā is astonished and explains that
she was not responsible, but rather blames the bawd and the other servants
in her mansion for pushing her around (8.21) and for throwing him out. He
should not be mad or despise her, but “play” as before (8.23). Then, when she
is asleep, he has both magically transferred to an island in the middle of the
ocean (8.31). In the colophon this is identified as sindhudvı̄pa.
When she awakes the next morning to see not her house but just water
surrounding her, she is frightened, suspects some magical trick and laments
her fate. When the prince tells the parrot to bring them food and drink, she
is pacified, and this goes on for a while. One day she makes him drink too
much, and forgetting even Delārāmā’s previous deceit (9.31) he tells her the
whole story behind the bird, and when he under the influence of sex, food and
alcohol falls asleep (9.32), she uses the bird to be transported back into her
house.
In the morning the prince realizes what happened and is devastated. He
does not know what to eat and of course how to leave the island. Wandering
�
The other Sanskrit version of the story is contained in Hemavijaya’s Kathā-
ratnākara,127 which has long remained unnoticed, but contains the story in
an exemplary version.128 The text was available in print, but when Johannes
126 This punishment is emphasized: 12.33b, 12.34d, 13.5b. 127 Edition: [Hemavijaya: Kathā-
129“Der Text, der unserer Übersetzung zugrunde liegt, ist die uns vorliegende eigenhändige
Niederschrift des Verfassers selbst, wie im dritten Bande dargelegt werden wird. Die indische
Ausgabe ist eine gröbliche Entstellung desselben, wie wir gleichfalls im dritten Bande zeigen
werden.« Hertel: Kathāratnākara, p. xx. 130 Personal communication by Anett Krause.
main story? According to Aarne’s analysis this is an original feature,131 and we
may at least deduce that the source of the DRK did not name the bird.
Many details in the story could be discussed afresh by comparing these
two Sanskrit versions,132 for instance the discrepancy as regards the part of
the bird that has to be eaten to gain the magical effect. Some of the variation
stems from insecurities of translation,133 but both in the KR and the DRK the
magical part of the bird has to remain unchanged in the body of the prince,
for it has to be recovered through an emetic only later. In the KR it is the
cockscomb (cūd.ā) that is eaten, in the DRK it is explicitly mentioned that the
magical part is a “bone” (asthi) that cannot be digested.
One interesting question for anyone transforming the fairy tale into a
literary piece is: how can the characters know about the magical property of
the bird. In the KR with its Jaina background it is a Jaina monk who knows
and divulges the secret, whereas in the DRK it is a witch (Śākinı̄). What is
perhaps also noteworthy is that in the KR the son Gun.acandra is not tricked
by the courtesan Rūpasenā, but by her bawd. The parallel role played in the
DRK by Delārāmā differs in that she is ultimately responsible, not her bawd,
and therefore punished by being transformed into a donkey; in the KR it is
the bawd who has to suffer the blows by the stick.
The magical objects are acquired in the KR from the pupils of a Yogi.
When he travels back magically to the courtesan, he is tricked again by the
bawd. In this passage the similarity between the versions is remarkable. But
whereas in the KR the whole story is told from a Jaina perspective, although
not very obtrusively, the DRK seems fairly neutral. No religious background
comes to mind immediately.
Apart from the many differences in details, one gets the impression that,
on the level of the plot, the author of the DRK has woven three different
strands into one connected and coherent narrative:
131“Die bestimmung der art des vogels ist ohne zweifel ein späterer zusatz.” Aarne: Verglei-
chende Märchenforschungen, p. 174. 132 The Kathāratnākara has not yet been extensively
used, and the DRK has not at all been noticed. 133 For instance, what Hertel identifies as
“Schildknorpel” (kākalaka) also means “Kehlkopf ” and is therefore not to be distinguished from
Aarne’s “kropf ”.
1. A (quasi-)historical setting involving a Sultan Muhammad of Halābha
with his wife Merabhaktā and his two sons Ibrāhı̄m and Murād Bhaks.o
is prefixed to the whole story.
The first fusion is an intelligent device to remedy two “problems” in the plot.
Only in the DRK it is in the end the righteous heir who gains the throne, which
his father had lost. And the dramatized explanation of the poverty of the king
leads naturally into and provides a context to the bird-heart story. The reason
why the story of Delārāmā takes such prominence is not clear. But the first
and the last element explain why the author attributes the story to Muslim
(Persian) sources.
One peculiar feature of the DRK is its misogynist attitude, which has the
side-effect that some otherwise arbitrary elements, which other versions had
to leave unexplained, are not noticed. What is however noticed as a rather
unpleasant aesthetic break in the general tone of the work is the sadistic
punishment of Delārāmā when transformed into a donkey. But this is a trait
of the original story and appears in many versions.134
An interesting feature of the DRK without obvious parallels lies in the
appearence of Yoginı̄s. These are semi-divine, magical and positive135 figures,
even morally superior to the king! In the end it is the Yoginı̄ who admonishes
the king to stop his disproportionate and inhuman punishment of Delārāmā.
One crucial difference is of course that the DRK is the most elaborate
Indian version of the story discovered so far. The author exercises his freedom
to dwell on detailed descriptions to develop characters—especially that of
Delārāmā—in the manner of a Kāvya, and it gives more attention to their
responsibility or their reaction to moral dilemmas. This, together with a more
sophisticated language, makes it the most distinguished Sanskrit version of
the story of the Magical Bird-Heart. Whether the unknown source used by
because Yoginı̄s are also imagined as animal-headed dreadful and dangerous magical beings.
Bhat.t.a Āhlādaka contained this configuration of motifs is hard to say, but
the poetical presentation of the story is necessarily that of the author alone.
A re-edition and translation of this Kashmirian work is obviously highly
desirable.
The Vı̄raratnaśekharaśikhā
A text even lesser known than the previous one, and the last example
for a translation of a Persian text into Sanskrit, is Sāhib Rām’s transla-
tion of the Akhlāq-i Muh.sinı̄ into Sanskrit. It is worth noting that all
three examples of translations of non-technical works from Persian into
Sanskrit, the Kathākautuka, the Delārāmākathāsāra, and Sāhib Rām’s
Vı̄raratnaśekharaśikhā, are by Kashmirians. Since I am not competent to
talk on details as far as the Persian source is concerned, I shall merely say
a few words on Sāhib Rām’s translation and give an example for how the
Sanskrit Pandit tried to adapt the materials.136
Sāhib Rām composed his metrical translation probably in 1853/54,137 and
there are a few manuscripts of the text, the most important being a complete
one in the IGNCA, New Delhi,138 written during the lifetime of the author,
from which the following examples are taken.
136 An edition and analysis is presently the focus of a project in Marburg as a collaboration
between Iranian and Sanskrit Studies. 137 M. A. Stein: Catalogue of the Sanskrit manuscripts
in the Raghunatha temple library of His Highness the Maharaja of Jammu and Kashmir. Bom-
bay: Nirnaya-Sagar 1894, p. 104. 138 IGNCA Ms. No. 455. 139 Literally: “it is enunciated”
Here śukla is not used in a Sanskrit meaning, but for Persian šukr and it is
defined thus: śuklam. śuddham. karmetyarthah., śuklatvam. viśes.an.advāren.ā-
ha sarvasādhāran.am iti uttamavarn.ahı̄najātiparyantam . samānam iti [. . .]
141The interpretation in the auto-commentary is more complicated, but need not concern us
here.
“śuklam. means pure action. This purity (or: nature of śukr) is further ex-
plained as being common to all, that means equal for all, from the highest
estate (varn.a) to the lowest cast (jāti).”
Here the following question arises: why was this concept of purity im-
portant enough not to be transformed into a “Hindu” concept of purity, but
to occasion a note about its definition? At present we can only guess that
the author wanted to highlight that this is a specific concept, which has no
counterpart in the Hindu thought world, since it applies to all people.
The second chapter treats of sauśı̄lyam, which literally means “good con-
duct”, but is defined here as “desireless action” (nis.kāmam . karma), which
of course reminds one of the Bhagavadgı̄tā’s karmayoga. This introduction
spans twenty-four verses, only then follows the recast of the Persian narrative,
which is announced with the words: “On that matter there is an old tale, in
which the king Khalapū is described.”142
The next verse clarifies that this king is called Khalapū,143 which is not only
an unusual name, but also if analysed as normal Sanskrit would mean some-
thing like the “barn cleaner”. In a literature where names indicate character,
such an interpretation is unlikely. According to the auto-commentary this pe-
culiar name denotes someone versed in nı̄ti, who is purifying (pū) the rogues
(khala) through proper nı̄ti, here obviously in the sense of “jurisdiction”.144
Khalapū thus is a name for the king as supreme judge, a sense that no
doubt one would have been able to express more clearly in Sanskrit. But
the word is an attempt to capture the arabo-persian halı̄fa. The Sanskrit has
ˇ
neither a voiced h, nor an f, which is regularly transcribed in Indian languages
as ph, which in Kashmirian pronunciation loses its aspiration.
Now follows the story, in which the Kalif conducts a court case, but steps
down after becoming biased. The details are outside the scope of this sum-
mary, but what should have become clear is that the translator carefully
transforms some elements of the story into an Indian context, but preserves
others. Not only would a detailed study prove interesting, what I was hoping
142 atrāpy udāharantı̄mam itihāsam. purātanam / Khalapūr madhyadeśı̄yo rājā yatra hi varn.yate.
(2.25) 143 madhyadeśe nr.pālo ’bhūt Khalapūnāmadheyakah. (2.26ab) 144 khalapūr
satyam
. prācı̄naracitāh. stutayo ’syā mude satām
bhavanto ’pi prasādāya stutir asyā vidhı̄yatām
Sāhib Rām says that he complied with their request and here follows in the text
fragment a Śrı̄mahārājñı̄stotra in fifteen Śārdūlavikrı̄d.ita stanzas. The author
then almost apologizes for this adhoc composition written at the “request of
others” and compares it to a fire-fly in the face of the light of the sun.
There is surely more to be gained from studying inedita.
145As in the combination of an ethnological and a philological approach. 146 I quote them
from the edition by Anett Krause, “Textfragment C”.
How to Edit
�
European Textual Criticism
1 The following is but a very brief summary of Widu-Wolfgang Ehlers: “Antike und
klassisch-philologische Editionsverfahren”. In: Hans Gert Roloff (ed.): Geschichte der Edi-
tionsverfahren vom Altertum bis zur Gegenwart im Überblick. Berlin 2003, p. 17ff.
often we see that through Devanāgarı̄ manuscripts contamination spreads
more easily, whereas regional ones preserve a more stable version of texts.
In Kashmirian literature we see that Devanāgarı̄ versions were often not
prepared by Pandits but by non-expert scribes and that they abound in all
kinds of mistakes. Here to edit a text without the help of Śāradā manuscripts
is not advisable. But also the oblong layout in most Indian manuscripts is
noteworthy. Even when in palm-leaf manuscripts the text is organized in
columns the reader has to finish the first line in a column and only then jump
to the second line.2
Coming back to Europe, we see that before the age of modernity a sophis-
ticated art of criticism (ars critica)3 evolved, which treated the transmission
of texts not as a science, but as an art practised by experts. The most famous
of these scholars is probably Poliziano (1454–1494), who was in many respects
ahead of his time. There were two wide-spread methods of editing or rather
preparing texts. In case of problematic passages one corrected the text with
the help of further manuscripts (emendatio ope codicum), or through one’s
own corrections (emendatio ope ingenii). Only later the term emendatio
came to mean something different: the “correction” of a passage against all
manuscripts. The development of the corresponding Sanskrit term is com-
parable. The word śodhanam means purification, correction of a text rather
like emendatio in the older sense, but is used, especially in modernity in the
combination śodhayitvā prakāśitam for “editing”.
Poliziano already made a full collation of all manuscripts the basis of his
editions, a decisive step ahead and a demand for which now the nineteenth
century critic Lachmann is credited. Poliziano emphasized the importance of
very old manuscripts, excluded apographs from consideration and developed
a proto-stemmatology.
But this was only one exceptional scholar at the time. The common prac-
tice was quite different. When near the end of the fifteenth century the
invention of the printing press in Europe caused a large-scale production of
2 A new study of the layout of pre-1350 North Indian manuscripts has just been completed by
Bidur Bhattarai: Dividing Texts: Conventions of Visual Text-organization in North Indian and
Nepalese Manuscripts up to ca. CE 1350. Ph. D. dissertation. Hamburg 2015. 3 This is at the
same time the name of some works on the topic, as for instance the one by Johann de Clerq
(Joannis de Clerici: Ars Critica. Londini MDCXCVIII.) For the genre, see Klara Vanek: Ars
classical texts, these first editions were typeset by using mostly one recent
manuscript that was emended according to the subjective taste of the editor
and then printed. The manuscripts used often do not survive. We have a
similar phenomenon in nineteenth century India, where after the wide-scale4
introduction of printing technology many editions were first printed from one
or two manuscripts that are now no more available. Although this time saw
very learned and able editors, for instance in the famous Kāvyamālā series, in
general we do not know how much emending went into these first prints.
Through frequent reprints these first versions became the standard text
and were uncritically perceived as “the text” of an author. Here I am talking
about East and West alike. It seems to me that the cultural difference of a
pronounced book culture in Europe as against a manuscript culture with a
strong emphasis on oral transmission in India did not make much difference
in this respect. No culture can easily avoid the impression that a printed
book contains a valid text. In the West these first editions were called the
“received text” (textus receptus), since through frequent reprints they became
widely-known standard versions. Because of their low quality this term came
to mean as much as an uncritical and doubtful text, since it is based on unclear
sources and there is the suspicion of undocumented emendations. From these
editions, also called vulgate, one cannot know what was actually transmitted
in the manuscripts. This also applies to Indian vulgate editions.
In the following centuries we have exceptional textual critics like Joseph
Scaliger (1540–1609), or Richard Bentley (1662–1742), but no different method.
Heavy emendations were quite common and this so-called “conjectural cri-
tique” remained a standard term in the nineteenth century.
An important test case of textual criticism is of course the criticism of
the New Testament. The textus receptus of the NT was not different from
many other editions prepared at the time.5 While there were innumerable
manuscripts it was prepared by the famous Erasmus of Rotterdam by using
only few, and as it turned out not very good ones. Furthermore, this text
shared the fate of other texts: once printed, it became the standard version,
but with the distinction that here it was not perceived as the work of an
4 The prehistory starting from sixteenth-century Goa is described in Anant Kakba Priolkar:
The Printing Press in India. Bombay 1958. 5 This summary is of course quite superficial. For a
detailed treatment, see Jan Krans: Beyond what is written. Leiden: Brill 2006.
author, but the word of God, so that any attempt at suggesting better readings
was initially doomed to failure. Any change in the text meant a change in a
theology, on which the church had long settled, and so textual critics could
easily become what some modern American evangelicals still think of them:
heretics. The absurdity of this situation can be made clear through pointing
to the fact that while the textual-critics were charged with the allegation of
breaking the tradition, they were actually trying to restore the earliest literary
traditions on which Christianity was built. But when it comes to textual
criticism of revealed scriptures, philology and the politics of religion have an
uneasy relationship.
Similar phenomena are met with in other religions, but also in non-
theologian compartments of philology. For opposing critical editions and
changes to a vulgate version we do not need religious feelings as an expla-
nation. Common human inertia is usually sufficient. When the edition of
the oldest recension of the Yogavāsis.t.ha, the so-called Moks.opāya, started, it
was for some scholars quite difficult to accept the thought that they had for
many years worked from a vulgate edition that was in many places next to
meaningless. But usually in the field of religion things are more complicated
because we have the conflicting ideals of academic freedom of research as
against religious freedom. In European academia eventually the old consensus
that nothing could be researched that was against the doctrine of the Christian
church broke down, the West not only accepted that we live in a heliocentric
system, and that the world was not created as it is described in the Bible, in
the universities theology could no longer control what other, secular subjects
had to work on. As a result the study of religion is now ideally conducted in
Germany in two faculties: one insider perspective of Christianity in theology,
and a neutral perspective in the philosophical faculties. This is the rationale
behind the distinction between a confessional study of religion and a secular
“science of religion” (Religionswissenschaft). Here I am talking mainly about
Germany, in other countries boundaries differ. In any case, the history of
textual criticism is often intricately connected to our view of religious and
intellectual freedom.
6 Paul Maas: Textkritik. Leipzig 1960 (1 1927). 7 Martin L. West: Textual Criticism and
Editorial Technique. Stuttgart 1973. 8 L. D. Reynolds and N. G. Wilson: Scribes and Scholars.
Oxford 1991. 9 “Sanskrit Literature”. In: D. C. Greetham: Scholarly Editing. A Guide to
Research. New York 1995, p. 575–599. 10 Reynolds and Wilson: Scribes and Scholars, p. 207f.
complicated by the presence of two or more variant readings, each
with a claim to be the transmitted text. The whole of this second
stage is sometimes still given its traditional, though misleading,
name—emendatio.
There is firstly the recensio, which implies surveying all existing manu-
scripts of a text and establishing the genetical relationship between all these
manuscripts with the help of shared error. The underlying idea is that some
errors are of such a quality that they are unlikely to have been committed
independently by scribes. Their presence shows that all manuscripts display-
ing the same errors derive from a common source. Through such “binding”
errors manuscripts of a given text fall into groups or branches that can be
arranged into a genealogical tree (stemma) which assigns specific values to
groups of manuscripts. Let us assume we have five manuscripts of a text,
called A, B, C, D, and E. We start from the assumption that all manuscripts
contain the same (version of the) text, that they, in other words, are ultimately
derived from one and the same archetype, which they all transmit, but each
with different scribal errors. If this basic assumption turns out to be unlikely,
if the manuscripts rather appear to go back to related, but ultimately separate
versions or recensions, then the edition of one original text is a much more
speculative matter. The problem in this case is that it is unlikely that such a
reconstruction of an original version resembles a historically existing text.11
In such cases an edition of the various recensions of a text may be the more
convincing approach.12
If there is a single archetype, we need to establish its text and then find
out whether the resulting text is convincing, or whether time and scribes have
changed the wording beyond recognition, so that we need to mark certain
passages as incomprehensible, or suggest emendations. But first we have to
find a method how to identify the readings of the archetype. Here it is useful
11 Examples would be the critical edition of the Mahābhārata (Poona 1928ff.), where the claim
to reconstruct an “Ur-Text” was not really made, or Edgerton’s reconstructive edition of the
Pañcatantra (Franklin Edgerton: The Panchatantra reconstructed. New York 1924), where,
however, large parts of the text are marked as being insecure. 12 This has been done by
Edgerton in his edition of the widely differing recensions of the Vikramacarita (Franklin
Edgerton: A Hindu Book of Tale: the Vikramacarita. Baltimore 1912). Here the four distinct
versions are edited side by side and no “Urtext” is postulated.
to remind oneself that in many cases there are obvious choices: The original
reading may simply be the one that makes sense, is grammatically correct, or
can be identified by other common-sensical methods, for which no textual
criticism is needed. But as soon as we collate a greater number of manuscripts
for a given text, we are likely to find that in a given passage there may be more
than one viable reading. A close familiarity with the style of the author and
other background knowledge will enable the attentive editor to narrow down
the candidates, hopefully to a single convincing original reading, but only the
naïve or the over-confident will believe that they can always identify what the
author wrote.
13Another technique is of course to employ “neo-stemmatology”, for which see Yumi Ousaka
and Moriichi Yamazaki: “Genealogical Classification of Saddharmapun.d.arı̄ka Manuscripts
Based on Many-Variable Analysis”. In: Literary and Linguistic Computing 17. 2 (2002), p. 193–206.
But it is of not much use to formulate general rules, as the examples of
historical editions and the controversies surrounding them show. One might,
for instance, in the face of the different versions of the Mahābhārata, for which
one would not propose a common archetype, refrain from classical editing
and just print one version. But this is exactly what Sukthankar has argued
against when writing on his predecessor’s edition in his Prolegomena to the
critical edition: “Professor Sastri’s edition is an excellent demonstration of the
inadequacy of the underlying principle, which has been repeatedly advocated,
showing up its defects as nothing else could. What Prof. Sastri set out to do is
(to quote his own words): ‘to print the text as it is in the original palm-leaf,
liberty being taken only to correct scriptorial blunders, to weigh the different
readings in the additional manuscripts and choose the more important ones
[scil. readings] for being added to the text by way of footnotes.’ How difficult
it is to carry this out verbatim in practice and at the same time to present a
half-way readable text may be realized when we see how Sastri has had to
doctor his text.”14
Despite all these complications in normal scenarios the idea of a “best
manuscript” will be treacherous. It may be true that one manuscript has
statistically the largest share of good readings, but this cannot be an argument
for following it blindly, since it invariably will be wrong in some places. More
importantly, a fixation on one manuscript can lure the editor into not taking
other evidence seriously, in which case he may easily overlook corruptions.
Another “intuitive” method would be to count manuscripts and follow
the majority reading. Suppose we have five manuscripts, labelled A to E. If
in a given passage E had one interesting reading worth considering, whereas
all other manuscripts agreed on another, perfectly acceptable reading; as
every editor knows there is a psychological hesitation to follow the single
manuscript E against all others. We tend to assume that it is more likely that
one scribe made a mistake than all the others, so the well-attested reading
can easily command some kind of preference. Here stemmatology cautions
us from a logical flaw: Imagine that the constellation of error shows that in
our group of sources A, B, C, and D agree in significant error. The case would,
for instance, be clear if they all would have exactly the same gaps (lacunae).
15
this is not entirely correct: when the main division of a stemma is into two
branches,16 then the method may give us no more than two readings of equal
weight. Here the editor has to decide according to other criteria, an unpleasent
practical complication of stemmatic theory, and only hesitantly discussed in
some works on stemmatics, since according to some critics editors then
silently resort to the best manuscript method,17 or even twist the stemma in
order to allow them more choices.
Here, as usual, Reynolds and Wilson provide a lucid treatment of the
dilemma:18
It is important to note that the resulting text is that of the “archetype”, which
is the oldest manuscript from which all others ideally derive. This archetype
may still be far removed from the autograph, the author’s copy. If, for in-
stance, a text survived half a century after the death of the author in only one
manuscript, then the method will not be able to retrieve anything but the
text from which all other copies have descended. The distance between the
autograph and the archetype cannot be bridged by this method.
For this reason, the editor is to examine whether the text gained through
stemmatical considerations, the archetype, is likely to be that of the author, or
whether, for instance, considerations of style or content, force us to doubt the
authenticity of words, or passages. This examinatio leads to cases where we
can suggest a reading that solves the inconsistency (emendatio), other cases
can only be diagnosed as helpless and have to be marked as corrupt.
Fränkel: Einleitung zur kritischen Ausgabe der Argonautika des Appolonius. Göttingen 1964,
p. 131. 18 Scribes and Scholars, p. 221.
This is in perhaps undue brevity and simplification the stemmatical
method, which goes under the name of Karl Lachmann, not so much his-
torically,19 but as a convenient designation that generally implies the inclusion
of further developments. One usually refers to the succinct and brilliantly
dense booklet Textkritik by Paul Maas for its valid formulation.20
At this point it is common for the more practically minded handbooks
to add a cautionary remark: “The apparent simplicity and finality of the
stemmatic method as outlined above is deceptive.”21 For, the stemmatic
method works only under certain conditions. (1) It presupposes that there
was a single archetype. (2) Scribes are supposed to have produced only
errors, but not to have corrected anything. (3) And a scribe should not have
used more than one manuscript and selected readings from one or the other
(“contamination”). If a section of the transmission is contaminated, then
the selection of variants “is greatly hindered, if not made impossible”22 and
each variant becomes what Maas calls a “presumptive variant”, a potentially
viable reading. In other words, in the case of contamination, regardless of
the position of a manuscript in the stemma, or its overall quality, whether an
editor says in the preface “generally correct” or “unreliable”, all variants are
again equal.
In order to show that this is far from being mere theory, but for an editor
an everyday experience, I shall present one current example. In our project of
editing the Sanskrit lexikon of Maṅkha, the Maṅkhakośa with its T.ı̄kā, recently
19 “Many scholars contributed to the elaboration of the stemmatic theory of recension; this had
been formulated in all its essentials by the middle of the nineteenth century and, although his
own contribution is much slighter than had been supposed, it is still associated with the name
of Karl Lachmann.” Reynolds and Wilson: Scribes and Scholars, p. 209f. 20 Paul Maas:
Textkritik. Leipzig 4 1960, 1 1927. The truth is, by the way, more complicated, which may be
guessed from the fact that in Maas’ work the name of Lachmann does not even occur. More
practical is the introduction by Martin West (Textual Criticism and Editorial Technique. Stuttgart
1973). Another very readable overview is given in L. D. Reynolds and N. G. Wilson: Scribes and
Scholars. Oxford 1991, p. 207ff (“Textual Criticism”), and there are many more. 21 Reynolds
and Wilson: Scribes and Scholars, p. 214. 22 I quote the English translation of Maas’ work by
Barbara Flower (Oxford 1958), here p. 7. Reeve has rejected this translation (“I decline to quote
[. . .]” Michael D. Reeve: “Eliminatio Codicum Descriptorum: A Methodological Problem”.
In: Editing Greek and Latin Texts. New York 1989, p. 2, fn. 3) because it has rendered the German
word “Hilfe”, repeated in one passage, with “aid” in the first instance and “help” in the second. I
have some sympathy for the attempt to smoothen the extremely terse style of Maas.
Lata Deokar and myself agreed on a preliminary edition of one passage in
the Maṅkhakośat.ı̄kā. We had two manuscripts from the Stein collection in
Oxford, one being the transcript of the other, so in fact we had merely one
source.
In verse 38 the Maṅkhakośa defines the word vālikā in the sense of, among
others, “child” (śiśau). The commentary gives the following example for this
meaning from literature:
vālikāracitavastraputrikākrı̄d.anena sadr.śam
. tad arcanam
ityādau śiśau /
“This [type of] worship is like playing with a [nicely] clothed doll
made for a child”—here [vālikā] has the sense of “child”.
The passage has not been edited by Zachariae in the only edition of this text—
the details of this will be discussed later—23 and the quotation was therefore
never identified. It is quoted in the Mahārthamañjarı̄parimala24 as from the
Arcanātrim. śikā, which can be identified as the Paramārcanatrim. śikā of Nāga,
25
who is dated to the eleventh century by Sanderson.
There does not seem to be any textual problem here and we could lay the
matter at rest, but there is the following marginal note: krı̄d.anena krı̄d.itena
napum . sake bhāve ktah.. What is quite odd here, is that the pratı̄ka krı̄d.anena
is explained by krı̄d.itena, and that the grammatical explanation does not
pertain to the word to be explained, but to the synonym. If we look at the
wording in the printed text of the Mahārthamañjarı̄, we find that krı̄d.itena is
the reading there. Only this constellation explains the note. A scribe who was
also a scholar made a note on krı̄d.itena, but the explanation was entered into
the text and the note, which now seemed confusing became further garbled.
This is a case, where stemmatics cannot help us, it is only the totality of
readings, testimonia and notes, that suggests that krı̄d.itena would have to be
the preferred reading.
23 See below, p. 160. 24 Ed. T. Gan. apati Sāstrī (Trivandrum Sanskrit Series 66). Trivandrum
1919, p. 111. The reading is, however, bālikā-. 25 See Alexis Sanderson: “The Śaiva Exegesis
of Kashmir”. In: Tantric Studies in Memory of Hélène Brunner. Pondichéry 2007, p. 292 and 411.
To this the date of Maṅkha now furnishes a solid upper limit.
The day after we had decided on such a conjecture I could check the
manuscript of the Maṅkhakośat.ı̄kā in the Bhandarkar Oriental Research Insti-
tute, which Zachariae described as being often “very corrupt” and that “with
this MS. alone an edition of the commentary would have been impossible”
(p. 2). But this manuscript reads —as if to demonstrate the text-critical obser-
vation that a “bad” manuscript may have the right reading alone—krı̄d.itena.26
So we must remind ourselves of the important rule that if contamination
is present, and in most transmissions of Sanskrit texts it is, every variant, even
in a bad, or stemmatically irrelevant manuscript, becomes a viable choice.
Unfortunately some earlier manuals of textual criticism, some of which
are less schematic and therefore of more practical value than the one by Maas,
have been mostly forgotten. It also has to do with the language barrier, or
rather the increasing limitation of the anglophone world to publications in
English, that Maas, who has been translated into English, is read, whereas
others are not. The introduction by Kantorowitz,27 for instance, is still worth
reading. It teaches stemmatics briefly, but then devotes much space to “in-
ner” criteria. He talks about the literary-historical criterion, the criterion
of language history, the “history of objects (Sachgeschichte), the criterion
of the history of transmission, etc. He formulates stemmatical rules, which
do not mathematically lead—as in Maas’ description—to the archetype, but
merely allow us to formulate a hypothesis to be tested with the other criteria.
A whole chapter is devoted to psychological criteria. There are also many
helpful cautionary remarks, for instance, that the application of stemmatology
is likely to produce wrong results in a transmission with manuscripts that
contain glosses.28 In India such codices with marginalia are frequent.29
Another earlier manual is extremely practically minded. It describes
everything around editing, from meeting other colleagues to preparing the
26 To be fair, Zachariae says that, “not a few leaves are pretty correct”. Ibid. 27 Hermann
Kantorowitz: Einführung in die Textkritik. Leipzig 1921. 28 Op. cit., p. 49. 29 For a study
devoted to marginalia in Sanskrit manuscripts see Camillo Formigatti: Sanskrit Annotated
Manuscripts from Northern India and Nepal. Ph.D. dissertation. Hamburg 2013.
apparatus for the typesetter. When it comes to stemmatics it remains very
brief and highly pragmatic.30
30 “Noch bevor alle Handschriften kollationiert sind, sobald genügend Material für ihre Kenntnis
gesammelt ist, muß man sich bemühen, ihr gegenseitiges Verhältnis zu bestimmen. Ein Stemma
aufzustellen ist oft nicht möglich, umso weniger je größer die Zahl der Handschriften ist. Je öfter
ein Text abgeschrieben wurde, desto mehr beeinflußten sich die Abschriften untereinander, die
Beziehungen kreuzten sich und immer neue Mischtexte entstanden. Daher ist oft unter einer
großen Masse keine einzige direkte Abschrift einer noch vorhandenen Vorlage nachzuweisen.
So hat z. B. P. de Lagarde die Ansicht ausgesprochen (die sich freilich kaum in vollem Maß
bestätigen wird), daß von den unzähligen Handschriften der Septuaginta keine einzige ganz
wertlos, aber auch keine von schweren Korruptelen ganz frei sei. In solchen Fällen kann man kein
Stemma aufstellen; dagegen ist es möglich, gewisse Gruppen zusammenzunehmen, die durch gle-
iche Verderbnisse (besonders wichtig sind Zusätze und Lücken) ihre gemeinsame Abstammung
verraten [. . .] Ist der Wert der einzelnen Handschriften auf Grund ihres Verwandtschaftsverhält-
nisses festgestellt, so werden die Handschriften (am besten nach der Reihenfolge ihres Wertes)
verglichen; die Haupthandschrift aber sollte womöglich noch einmal eingesehen werden, nach-
dem die Varianten der übrigen gebucht sind.” Otto Stählin: Editionstechnik. Ratschläge für
die Anlage textkritischer Ausgaben. Leipzig 1914, p. 36f. 31 Reynolds and Wilson: Scribes
and Scholars, p. 211. 32 I am deliberately phrasing this carefully, because—as Schmidt has
33 In classical studies this seem to be the consensus: “In den letzten Jahrzehnten sind die
Überlieferungsverhältnisse vieler Texte durch arbeitsintensive Untersuchungen geklärt worden,
wodurch immer deutlicher wird, daß Kontamination verschiedener Stränge fast überall die Regel
ist. Trotz dieser Schwierigkeiten kann bei vielen Texten ein glaubwürdiges Stemma konstruiert
werden”. Joseph Delz: “Textkritik und Editionstechnik”. In: Fritz Graf (ed.): Einleitung in die
lateinische Philologie. Stuttgart 1997, p. 58. 34 Salemans (see below) removes the pejorative
term “common error” from the Lachmann method in favour of “common non-original change”
(p. 19). 35 West: Textual Criticism and Editorial Technique, p. 39. 36 L. D. Reynolds (ed.):
Texts and Transmission. Oxford: Clarendon 1983, p. 201. 37 Ibid. 38 West: Textual Criticism
and Editorial Technique, p. 47. 39 See, for instance, on the successful attempt to establish a
known stemma from published data: Peter Robinson and Robert J. O’Hara: “Report on the
Textual Criticism Challenge 1991”. In: Bryn Mawr Classical Review 3(4) (1992): p. 331–337.
has produced fairly sophisticated guidelines as to which variants should be
used and the type of erroneous stemmas that are produced, if these rules are
not adhered to.40 He says: “Many text genealogists assume incorrectly that all
the textual differences between text versions, sometimes with the exception of
small differences in spelling, can be used to draw text genealogical trees. From
the four types of parallelism presented we can see that many non-spelling
variants can offer false information about the kinship of text versions. It
could be argued that parallelisms do not frequently occur in text versions
and, therefore, could be filtered out by statistical or mathematical analysis,
like cluster analysis. For text genealogy this is a dubious approach. A single,
trustworthy variant can provide better information about the shape of a text-
genealogical tree than a thousand other untrustworthy variants.”41
On the other hand Philipp Maas has been successful in testing the cladistic
approach for a Tibetan text by using all variants as the basis for a stemma.42
He warns against what we might call too scientistic expectations, and proposes
to judge connective variants according to philological principles and thereby
avoid problems, since “phylogenetic software—like the human mind—can
easily get confused by contamination and parallelism.”43 In another article he
suggests combining the cladistic analysis, the “quantitative approach” with a
“philological discussion of selected variants (i.e. a qualitative approach)” in
order to produce a stemma for a contaminated transmission.44
But we must ask again. Suppose the refined techniques of computer-
aided neo-stemmatology are able to provide us with a stemma of a highly
contaminated transmission of a text. Would we select readings according to
stemmatic rules? Or would we declare contamination and thus bend these
rules whenever the supposed original reading occurs on the wrong side of
the stemma? If so, the stemma would be useful for understanding the course
40 Benedictus Johannes Paulus Salemans: Building Stemmas with the Computer in a Cladis-
tic, Neo-Lachmannian, Way. The Case of Fourteen Text Versions of Lanseloet van Denemerken.
Nijmegen 2000, p. 300–302. 41 Salemans, op. cit., p. 71. 42 Philipp Maas: “A Phyloge-
2015. 47 “Hence, in the present study the role of cladistics ends here.” Op. cit., p. 88. 48 Op.
cit., p. 114. 49 See the case of the reading karan.atvena on p. 173. 50 Adherents of computerized
neo-stemmatology will not necessarily agree to this.
text was now seen as an ahistoric mixture. In the beginning of the twentieth
century new methods marked a complete turning away from “Lachmann”.51
In modern philologies, where we have authorized prints rather than
manuscripts, the problems faced by an editor are entirely different. If an
author has, for instance, worked again on his text and has published more
than one version of it, then stemmatics must fail. Lachmann had in these
cases conflated authorized prints and has been criticized for this. In these
cases, we can only print one version or the other, or produce a genetic edition
that shows how the development of different versions of a text by the same
author can be analysed in an edition.52 Then we must not forget that editors
of philosophical literature have claimed that their method differs53 from those
applied for literature.
In Sanskrit textual criticism almost all these problems occur too. There
are transmissions like those of classical texts, where we may successfully apply
stemmatics. Then there are texts, where we have a multitude of recensions, but
no archetype.54 Apart from this we have the tradition of recitation, the orality
of Indian texts, and for some areas a living theological, philosophical and
literary tradition, which means that texts copied were in use by experts, and
thus were surely not just copied, but corrected and edited. A given manuscript,
especially when accompanied by a commentary, can therefore be quite unlike
the source envisaged by Maas: it could well be a pre-modern edition.55 For this
51 See Hans-Gert Roloff: “Karl Lachmann, seine Methode und die Folgen”. In: Hans Gert
Roloff (ed.): Geschichte der Editionsverfahren vom Altertum bis zur Gegenwart im Überblick.
Berlin 2003, p. 76f. 52 Hans Zeller: “Die Entwicklung der textgenetischen Edition im 20.
Jahrhundert”. In: Hans Gert Roloff (ed.): Geschichte der Editionsverfahren vom Altertum bis
zur Gegenwart im Überblick. Berlin 2003, p. 143–207. 53 Walter Jaeschke: “Editorische
Verfahren und Leistungen philosophischer Ausgaben.” In: Hans Gert Roloff (ed.): Geschichte
der Editionsverfahren vom Altertum bis zur Gegenwart im Überblick. Berlin 2003, p. 287. 54 For
some such cases von Hinüber has diagnosed a complete failure of normal textual criticism. Oskar
von Hinüber: “Der vernachlässigte Wortlaut. Die Problematik der Herausgabe buddhistischer
Sanskrit-Texte.” In: Kurt Gärtner (ed.): Zur Überlieferung, Kritik und Edition alter und neuerer
Texte. Mainz 2000, p. 17–36. 55 For some scholars this is so obvious that they mention it only
in passing: “The genealogical lines of Śāradā manuscripts are usually blurred by the tendency of
the Kashmiri pandits, who themselves copied the manuscripts, to make something like a critical
edition ante litteram.” Rafaelle Torella: “Pratyabhijñā and Philology”. In: JAOS 133.4 (2013),
p. 707. Kantorowitz speaks of “old emendation” and reminds the reader of the fact that there have
been philologists before Lachmann (op. cit., p. 8).
reason editors of Sanskrit text have regularly rejected stemmatics in favour of
other approaches.
But the whole question has been reopened by Michael Witzel in a recent article,
in which he reviews “Textual criticism in Indology and in European philology
during the 19th and 20th centuries”56 and diagnoses a lack of critical editions:
“Surprisingly, Indian (or South Asian texts in general) texts have hardly seen
any critical editions in the strict Lachmannian sense. What goes under this
name usually are editions that merely include a selection of variants. It is
remarkable that over the past 200 years or so only about a dozen truly critical
editions, with stemma, of Sanskrit texts have been prepared.”57
Witzel not only describes recent practices, but demands an indological
best practice by classifying editions in the following way: “Class A: crit. ed.
with reasoned pedigree (stemma) of MSS. Class B: edition that makes use of
SOME MSS that are NOT described well and only occasionally / inconsistently
mentioned in the footnotes, the so-called critical apparatus. Class C: edition,
made as the editor sees fit, with no MSS used / mentioned; maybe with one or
two MSS (never explained) used and very sporadically quoted, or not at all.
Most Skt. editions are of type C and B, only a dozen or so are of type A, so far.”
This is the elaborate version, found in an internet forum, which makes his
point quite clear. Since it is difficult to get an overview on all Sanskrit editions,
one might add a few editions with stemmas, so there are actually more than
a dozen, but the number of editions featuring a stemma is indeed smaller
than one might expect. For Witzel this means that editors have fallen short
of complying with obvious standards, or in his words that “the 19th century
indological “tradition” of presenting “semi-critical” editions without proper
stemma continues unabated: even in Olivelle’s Manu only the N. and S. Indian
branches are distinguished [. . .]”58
Before continuing I would like to mention two editions with stemmas that
are not mentioned by Witzel. The first is Rau’s edition of the Vākyapadı̄ya,59
which has a stemma, a lucid introduction, but does not report many readings.
56 In: Electronic Journal of Vedic Studies 21.3 (2014), p. 9–90. 57 Op. cit., p. 47. 58 Op. cit.,
p. 48. 59 See below.
The other edition is an attempt to inculturate stemmatics into Sanskrit
terminology by V. L. Joshi.60 In a long introduction, where he even translates
some passages from the often quoted article by Housman into Sanskrit,61
Joshi has not only tried to explain the details of editing in Sanskrit, he has also
coined Sanskrit terms for the Latin technical terms of editing, as for instance,
recension (upaśākhā) etc. It is unfortunate that this attempt has not caught
on, and sparked neither a discussion nor even any serious reaction.62
The main thrust of Witzel’s proposal is to follow “the stemmatics of the
Lachmann school of 1810”,63 which implies “strictly adhering to the principle
of establishing a family tree of manuscripts (stemma). The method has been
summed up, after more than a hundred years of trial and error, by P. Maas
and M. L. West and for India by S. M. Katre.”64 Witzel also holds that the
recensio will produce a text that in a second step has to be emended: “Different
than Bentley’s ideas about criticism, textual criticism does not immediately
extract a meaning from a text. Such questions are better postponed until its
proper wording has been established. Even then, both the establishment of
a text and its preliminary editorial interpretation naturally go hand in hand.
After these initial steps of editing a text, higher textual criticism (emendatio)
comes in. Based on our knowledge of the grammar, style, parallel passages or
typical expressions of the author concerned—repeating here the Alexandrian
model—we can scrutinize the archetype MS and propose to make certain
corrections to that text.”65
Naturally one is interested to see what solution Witzel proposes for the
problem of contamination, which he admits is “always present”,66 but he is
Scholarly Editing. A Guide to Research. New York 1995, p. 575–599. 63 Witzel: Textual criticism
in Indology, p. 18. 64 Ibid. 65 Op. cit., p. 21. Also on p. 72: “having a stemma (or we may
add, an early version of a Bardic text) is only the starting point for the processes of emendatio
and producing a critical edition.” 66 Op. cit., p. 26.
It is well known that contamination is the rule in the (edition) of
the Sanskrit epics, which makes a true critical edition impossible
(apart from the problems of oral bardic transmission), as stressed
by the Mbh. editor Sukthankar.
However, the exact nature of contamination can usually be deter-
mined fairly easily. It has not been taken into account that the
influence of the many Indian scripts and the diversity of local
pronunciation allows to trace various strands of transmission and
to detect ‘aberrations’ from the individual local ‘norm.’67
Witzel, while being himself quite positive about stemmatics in the face of
contamination, I think, underrates the indological consensus against it. He
mentions a few examples of editors who consciously avoided constructing a
stemma on what they considered weak evidence,68 but remains ignorant of
such prominent examples as Coulson in his edition of the Mālatı̄mādhava,
who says under “ms. relationships”:69
As causes for this Coulson gives examples of manuscripts that were copied
from one recension, but where someone had added readings from another
recension. There is, by the way, ample evidence for this in the transmission
of many texts, which shows that such avenues of conflation were common.70
67 Op. cit., p. 27. 68 “By now, some voices have been raised with regard to contamination
and the “impossibility” to establish a stemma for Indian texts, for example: J. Hanneder, ed. of
Abhinavagupta’s Mālinı̄ślokavārttika 1/1-399, Groningen 1998, p. 40-45; R. Adriaensen et al., 190
ed. of the Skandapurān.a, vol. I, p. 39.” Op. cit., p. 41. 69 Michael Coulson: A Critical Edition
of the Mālatı̄mādhava. Delhi: Oxford University Press 1989, p. xviii. 70 See for instance Jürgen
Hanneder: “The Yogavāsis.t.ha and its Kashmirian recension, the Moks.opāya. Notes on their
Textual Quality”. In: WZKS 44 (2000), p. 183–210.
Often only parts of a text were revised with the help of further manuscripts,
and as a result, one may find in one’s critical apparatus that in one chapter
there seems to be a fairly consistent grouping of manuscripts (let us say AB
against CD), but that this changes completely in the next section (AC against
BD) and again further on. Similar observations are in fact quite commonly
voiced in text-critical introductions, and then usually the editor refrains from
providing a stemma on inconclusive evidence. In recent publications I was
under the impression that not only a sort of consensus had formed about this,
but also that in such cases the method described by Srinivasan (see below)
was very worth considering. This might have been premature, since Witzel
does not even mention this book in his treatment of textual criticism and
contamination.
It is therefore time to reconsider some of the facts and arguments in more
detail. The first task would be to investigate the nature of contamination in
pre-modern Indian transmissions of Sanskrit texts, which has some bearing
on whether stemmatology can be usefully applied. The other point concerns
the history of indological editing and its relation to the method of Lachmann.
It has become clear from the preceding description of classical, that is, stem-
matic editing that it works under certain preconditions only. One precon-
dition we have not yet dealt with sufficiently concerns the so-called author
variant. It means that when an author has transmitted to posteriority more
than one version of a passage or even a work, we have not one but more
archetypes, or at least variants that are quite unlike the ones envisaged in
Maas’ handbook, since they are not scribal errors, but valid authorized read-
ings. In a manuscript culture such second versions could manifest themselves
in marginal annotations in the autograph or in more than one copy being
prepared under the supervision of the author etc. An editor who uses stem-
matics may not want to accept that these things exist, many classical scholars
expect a good writer to work like Horace told his pupil in his Ars Poetica, to
publish only when the work has come to perfection, show it to no one before
and never change one’s mind afterwards. Common sense, the fact that this
admonition had to be given in the first place, and examples from modern
philologies show that this may be an honoured rule, but not necessarily a
wide-spread practice. Some authors may have worked without leaving any
trace of the production of texts, but it would be quite naïve to assume that all
or even most of them did.
There are only a few proven71 cases of medieval Sanskrit autographs. One
incontrovertible case is Ratnakan.t.ha’s commentary on the Haravijaya, written
in 1603 according to the colophon by the author himself. This statement could
have been copied by a scribe, but the peculiar cursive hand of the manuscript
allows us to identify its writer as the author himself.72
Another instance that is not well-known concerns the text of the Kathā-
ratnākara (early seventeenth century). Here Hertel has claimed in his transla-
tion of the work that the author’s copy has survived.73 But since the manuscript
is now lost, we cannot verify the case.74 What is even more interesting is Her-
tel’s statement that in this autograph, the corrections in faded ink are by the
first hand, that is by the author himself.75 Here we have one case that is like
those in modern philologies, where it is not untypical to have more than one
authorized version.
It is surely the case that we cannot prove that author variants are among the
variants we find in the transmission of a text, but the phenomenon has been
described sufficiently for European classical philology, in a work that is well
worth perusing,76 but has never been translated or reprinted. Its bearing on
stemmatics is briefly summarized in English by Reynolds and Wilson: “One
final complicating factor is the possibility that the ancient author himself made
corrections or alterations to his original text after publication. Sometimes
71 Claims for autographs are sometimes based on the observation that a manuscript is ex-
tremely correct. While such cases may carry a certain probability, I cannot see why this should
alone be considered as a proof. 72 For the author, see Jürgen Hanneder, Stanislav
Müller 1920, p. xx. 74 In his handwritten notes Hertel provides some of the arguments: “Das
(sic) B die Handschrift des Autors selbst ist, ergibt sich z. B. aus S. 84.” (Universitätsarchiv Leipzig,
NA Hertel 12b, p. 485). 75 “Der Korr. mit blasser Tinte ist der Schreiber selbst, der also spätere
Nachträge eingetragen hat.” (Universitätsarchiv Leipzig, NA Hertel 12b, p. 486). 76 Hilarius
Emonds: Zweite Auflage im Altertum. Leipzig 1941.
these would be extensive enough to justify us in speaking of a second edition.
Under the conditions of the ancient publishing trade a second edition was
much less likely to supplant its predecessor than in the modern world. Cicero’s
attempts to revise or eliminate errors in his works did not affect all the copies
from which our archetypes descended. The two versions circulated side by
side throughout antiquity with horizontal transmission taking place.”77
In European editing already the Humanists surmised that variants might
have been not merely caused by scribal and other errors, but also by the inter-
ference of the author himself. The wealth of examples provided by Emonds
shows that this is very likely, but also very difficult to prove, when it comes
to practical editing. The question whether a given variant, which fits every-
thing we know about the author, is an author variant, or rather an intelligent
conjecture, cannot often be solved convincingly.
But for the sake of our investigation it is not important to prove specific
cases, but to assess the general probability: Is this scenario just theoretically
possible, is it rare, or is it always to be expected in Indian textual criticism.
If our answer tends to the latter, there is the practical problem with this
observation that its application remains difficult. If we suspect an author
variant any time we encounter a second convincing variant, we cannot edit a
text any more. However, the answer has some bearing on one of our leading
questions, that is, the applicability of stemmatics.
There are examples that show that proof of author variants remains diffi-
cult and tends to complicate the editing process. Even in well-researched cases,
like the famous Middle-English poem Piers Plowman, which is transmitted in
around fifty manuscripts and at least three distinguishable versions, on which
much effort has been spent, we are left with the expected conflicting opinions:
some imagine a single versatile poet, others a multitude of authors, and it
depends on the temperament of individual scholars to commit to a side.
Then there is a further unsolved problem: How did authors publish their
works? Were they copied only privately by those interested. Did authors give
their works to a publishing, that is, to a copying house, was there a copy editor?
Were copies produced only after completion of the work? Was there a second
edition? Sanskritists might reject all these deliberations as inapplicable: for
78 See Walter Slaje: Bacchanal im Himmel und andere Proben aus Maṅkha. Wiesbaden 2015.
The verse can be interpreted in a variety of ways, and I have tried to give a
neutral rendering. The meaning given by Jonarāja in his succinct, but excellent
commentary is much more specific. He says that sūktibhejas.am . means the
remedy for a Kāvya, in the present case for Maṅkha’s Śrı̄kan.t.hacarita, and
that it consists of the removal of errors through the kind experts present at
its first recitation: yaih. sadbhih. sūkteh. kāvyasya bhes.ajam
. dos.anivāran.am
.
saujanyena hetunā svavaidus.yād vitanyate. If we then regard the context,
in which the participants of the literary circle, who are about to hear the
work of Maṅkha, are thus described, it would mean that these experts—the
Ālam. kārika Ruyyaka was among the listeners—were known or even expected
to give hints and corrections to the author. But if so, then the manuscript
mentioned in the text to which these corrections were applied and the last
Sarga added, would not have looked like an autograph, but like an exemplar
that was corrected. Would all scribes have known how to apply the changes
and ignore the first version?
If this seems unlikely, far-fetched or even shocking, I would like to point
out one verse by Somendra, who reports in a postscriptum to his father’s
Avadānakalpalatā that he had given the work to one “ācāryah.”:
Now śuddha, when it comes to language and texts, means “correct”, often in
the sense of grammatically correct. What Sūryaśrı̄ was credited for was not to
produce a nicely written copy, but to purify the text of errors, in other words
he acknowledges, as we would do in a book, the help of an editor.
It seems, we could open the door behind some texts a little and could get
the impression that on the other side there are some hitherto unknown char-
acters silently involved in the production of literature. My argument is that
Sanskritists, frustrated by the paucity of sources that could illuminate this
background of particular texts, failed to notice it, even when it was staring
into their face. One such failure is connected with the famous Śivastotrāvalı̄
by Utpaladeva, which has been edited79 and also translated a few times.80
The Śivastotrāvalı̄ is a collection of Stotras attributed to the author Ut-
paladeva, who lived in Kashmir two generations before Abhinavagupta, per-
haps around the middle of the tenth century. Kashmirian libraries house a
large number of manuscripts of this work attesting to its popularity. Often
it is accompanied by Ks.emarāja’s commentary, who is the third in a line of
religious transmission from the author.
A study of the manuscript material of this text has been made by Con-
stantina Rhodes-Bailly.81 She comes to the conclusion that “there were no
major variants in any of the manuscripts that I studied, and that the textual
tradition of the Śivastotrāvalı̄ remained intact, without varying recensions.”82
The actual variants, which include synonyms as for instance śarı̄ra for svarūpa
are not reported by the editor and the text of the first edition is made the
basis. This is somewhat astonishing, since the edition of 1964 lists quite a few
variants, also in the verses itself, and moreover the commentator Ks.emarāja
himself mentions and comments upon variants readings.83
Rhodes-Bailly understands Utpaladeva’s verses as a “spiritual diary”, and
believes that we, the readers, are “accompanying Utpala on the wanderings
on a marvelous pilgrimage.”84 In this context, the opening verse is interpreted
as marking the “outset of the journey”,85 the initial understanding. In other
words, the interpretation of the work is biographical and it is at least implicitly
suggested that the journey ends, when the accomplished devotee has become
a siddha,86 and this is at the very end of the work.
While I have no objections to such an interpretation in general, I am quite
surprised that the presupposition that the Stotras are auto-biographical and
chronological is taken for granted. This is all the more astonishing, since no
reader of the Sanskrit text can avoid being told by the commentator Ks.emarāja
79 In the following I refer to the text as edited by Rājānaka Laks. man. a: The Śivastotrāvalı̄ of
Utpaladevācārya with the Sanskrit Commentary of Ks.emarāja. Varanasi 1964. 80 Kotru 1985,
Rhodes-Bailly 1987, Bonnet 1989. 81 Shaiva devotional songs of Kashmir: a translation and
study of Utpaladeva’s Shivastotravali. Albany 1987. 82 Op. cit., p. 3. 83 For instance, ad 18.7
and 19.4. The reader will notice that the statement that there are no (major, real, original etc.)
variants is a recurrent theme. 84 Op. cit., p. 2. 85 Ibid. 86 Op. cit., p. 23
in clear terms that Utpaladeva is not really responsible for the form, in which
his text appears:87
Two persons took care of the literary bequest of Utpaladeva, and they found
his verses in disarray, at least not as ready-made Stotras. So these verses were
placed separately into Stotras. In other words, the mixed verses were arranged
by the executors of the literary bequest of Utpaladeva, and it appears that
87 For the interpretation of this passage, see also Alexis Sanderson: “The Śaiva Exegesis
of Kashmir”, p. 399f. 88 Stotra 13 is called Samgrahastotra and Ksemarāja gives a separate
. .
introduction for this. 89 The fourteenth Stotra in the Śivastotrāvalı̄ is one such, since every
line begins with the word jaya. 90 The fifteenth is called bhaktistotra. 91 Sanderson takes
the last two together: “also a number of single-verse poetic hymns for his daily devotions.”
92 tathātvena “being thus”.
Ks.emarāja, despite living only few generations after the author, and in the
same lineage, had no way of cleaning up the transmission. The arrangement
of the verses is not one conceived of by the author, but by later redactors.
If it reflects the author’s spiritual biography, then the credit must go to the
medieval editors, who arranged the materials.
And finally the same applies to the names of these Stotras, as Ks.emarāja
further informs us:
śrı̄viśvāvarttas tu vim
. śatyā stotraih. svātmotpreks.itanāmabhir
vyavasthāpitavān iti kila śrūyate /
The editorial report by Ks.emarāja shows that no less than four persons were
involved in the redaction of the so-called Śivastotrāvalı̄: Rāma and Ādityarāja
ordered the literary bequest into twenty groups, Viśvāvartta named the result-
ing Stotras and Ks.emarāja made sense of the collection by commenting on
the Stotras in their sequence. Neither the name of the text itself nor most of
the names of the Stotras are original.
But Ks.emarāja is, apart from the parts he considers authentic—as for
instance the Saṅgrahastotra—, highly critical of the presentation of the trans-
mitted text. Already in the second verse he stumbles upon an incongruity,
which he blames on the redactor:
Ks.emarāja says he has been sparse with his criticism, but what we infer from
his statements is this: he regards the status of the edition of his predecessors,
which really is a new composition of fragments, as problematic. The verses
were often not intended to be part of Stotras, and to treat them as if they were
does not do justice to the author.
But as we know from more recent examples, such cautionary remarks
never work. A printed text—or here one commented upon—almost invariably
creates its own history. It seems that Ks.emarāja mentions the history of the
text in such unusual detail to alert the reader to the nature of the text, to
caution him that the author was not responsible for the arrangement. This
would be what we would expect from modern editors as well, but while we
find such text-critical awareness a millennium ago in Kashmir, it is much
harder to find nowadays, as some translations of this text suggest.
94 See above, p. 18. 95 Of course only if he applied corrections very cautiously. If he invented
his own “better” readings—not unlike the conjectural school of the nineteenth century—, then
it would be difficult to establish what happened. Unless we could actually trace the corrections
in the manuscript in which they occurred first, we would simply notice that there are passages,
where we find among the variants collected from a variety of manuscripts not just one correct
reading and many misreadings, but more than one convincing reading; convincing in the sense
that it is consistent with all additional pieces of information which a historical-critical editor
needs to take into consideration: known style of the author, his vocabulary, his background,
and so forth. 96 Exceptions are, for climatic reasons, regions on the northern outskirts of
A German speaker would have to read the text slowly, but would not have
problems to understand it. One might have to think longer on single words
thus disfigured, but in an intelligible context they pose no real problem. The
speed of this mental auto-correction shown by this odd example suggests
that smaller errors are quite unlikely to even distract a reader, he will correct
them before he knows. Here proficiency in the language is the all important
97 See Bastian Sick: Der Dativ ist dem Genitiv sein Tod. Köln 2004, p. 73.
criterion, a few years of school German may not do. Sanskrit is not different,
but the question of the mastery of language is more complex, since it has
been, for the time concerned, not a first language to anyone. An educated
scholar may reconstruct a garbled text correctly, whereas a scribe with a poor
knowledge of Sanskrit, may produce a text that looks more like our German
example above.
Of course here the specifics of Sanskrit come to play, the fact that many
combinations of letters may yield a meaning, but if one does not stretch the
boundaries too far, the same principle should work, as in the translation of
the text above:
There are further aspects of scribal psychology and typical errors that have
to do with the character of Sanskrit manuscripts: Most manuscripts consist
of unbound leaves, which may get scattered and can be reassembled wrongly.
Then the oblong format of most Indian manuscripts, which is derived from
the form of a palm leaf and produces extremely long lines, can lead to errors,
since the eye when jumping back to the left margin may easily miss the correct
line.98
Many errors and their correction must have seemed no doubt a simple
matter for a learned expert scholar. Like a reader of the German text above,
he would have silently corrected errors. This is what the tenth-century com-
mentator Bhat.t.a Rāmakan.t.ha states:99
98 This is why in European typography we learn that when the length of the lines (typically
between 60 and 80 character) and the leading (vertical distance between lines) is harmonized,
reading is less exhausting. 99 See Dominic Goodall: Bhat.t.a Rāmakan.t.ha’s Commentary on
the Kiran.atantra. vol. I: chapter 1–6. Critical edition and annotated translation. Pondichéry 1998,
p. cxviii. I have modified Goodall’s translation to emphasize the points I want to make.
Variant readings in this work (atra), whether good or bad, are not
[derived] from the original (mūla), but from the error of readers.
Therefore to point them out [only displays one’s] ignorance.
But this lament is not limited to any specific time or country or culture. Take
the following quotation: “The state of the holy books is really deplorable,
if their authority depends on unlearned copyists (as they mostly are) or
intoxicated typesetters”. This is by the famous Dutch editor Erasmus with
reference to the Christian world.101
100 Transl. Sachau, vol. 1, p. 18. 101 “Misera vero conditio sacrorum voluminum, si horum
auctoritas pendet ab indoctis, ut fere sunt, librariis, aut temulentis typographis”. Quoted from
Jan Krans: Beyond what is written. Leiden: Brill 2006, p. 7.
The verses frequently given by scribes to apologize for any errors and
thereby asking for a general excuse for the bad state of the text did not help.
These verses are, as if to prove the point, sometimes written in faulty Sanskrit.
Here I shall quote a more elaborate and sophisticated one:
More often the statement is to the effect that the copyist has just reproduced
his source without any change. For any errors the manuscript source is to be
blamed. This, by the way, would be the ideal scribe for stemmatology to work.
But the negative view of scribes is one-sided. Tripathi rightly emphasizes
that manuscripts were first written, corrected, read and preserved by scholars,
not by scribes.103 On the other hand, pre-modern Sanskrit scholars are casti-
gated for doing exactly the opposite: emending and even rewriting texts that
are transmitted in bad shape. The most famous case is reported by Bühler,
who talks of the practice of “cooking Sanskrit books”,104 which means that
“lacunae and defects in the original are filled in according to the fancy of the
Pandit who corrects them”. One example he specifically mentions concerns
one of the authors mentioned above, Sāhib Rām. According to Bühler he had
been ordered by Ranbir Singh to produce a copy of the Nı̄lamatapurān.a, and
102 The text is quoted by Colas from a source which is not accessible to me at present. He
prints paris.odhanı̄yam, which cannot be correct, but Colas’ translation indicated that he
meant pariśodhanı̄yam. See Gérard Colas: “Relecture et techniques de correction dans les
manuscrits indiens”. In: Christian Jacob: Lieux de Savoir 2. Les mains de l’intellect. Paris
2011, p. 501f. 103 granthānām . racanākārāh., tes.ām
. prathamalekhakāh., pān.d.ulipı̄nām
. śodhakāh.,
tāsām. pāt.hakāh., tāsām
. sam . raks.an.akartāraś ca pan.d.itā evāsan, na tu kāyasthāh.. Gaya Caran
Tripathi: “sam . .skr tapān d
.. ulipı̄nām. sam
. raks.an.e pan.d.itānām avadānam”. In: The Pandit. Tradi-
tional Scholarship in India. Delhi 2001, p. 204. 104 Georg Bühler: Detailed Report of a Tour
in Search of Sanskrit Mss. made in Kásmír, Rajputana, and Central India. Bombay/London 1877,
p. 33.
since the beginning of this text was lost, he “restored” it with the help of what
he knew about the mythology and history of Kashmir. This restoration was
used for the edition of the text by Kanjilal and Zadoo (the above-mentioned
co-translator of Don Quixote), whereas in de Vreese’s critical edition105 this
longer version has been rejected for critical purposes. But it is interesting what
de Vreese writes on Sāhib Rām’s augmented version: “As, however, the Pan.d.it
possessed an intimate knowledge of ancient Kashmirian history, the insertions
and additions by him deserve due consideration as a commentary.”106
In order to put Bühler’s remarks, which conjure a fairly negative image
of Indian editing, into perspective I would like to point out further examples,
where editors have added to texts what they thought was appropriate. The
first example is that of the famous editor Erasmus of Rotterdam (died 1536),
who produced during his lifetime a large number of editions of religious texts,
but also of secular texts of classical Latin and Greek authors. His editions were
usually based on merely one or two manuscripts and he does not often refer
to readings, but was prone to interpolation, a practice unfortunately typical
for the humanists. When critical editing started in Europe, these Humanist
editions were therefore completely rejected in favour of older manuscripts.
But the extent of such interventions was considerable. Erasmus in one
case even added a tract of his own to a collection, a fact that was recognized
only centuries later,107 and here the border between conjectural editing and
forgery is difficult to draw.
What makes this case even more spectacular is that the Greek and Latin
text of the Bible used throughout Europe and even viewed some time after
Erasmus as the unalterable word of God itself is actually such a haphazard
edition by Erasmus. Its deficiencies were obvious to scholars and theologians,
who had looked at the manuscripts, but the status of this edition as the divine
word was so strong that it was only in the nineteenth century that critics dared
to read with the superior variants. Before this it was possible to add these
variants in the footnote apparatus, but for religious reasons one could not
change the received text (textus receptus).
105 Konraad de Vreese: Nı̄lamata-purān.a or teachings of Nı̄la. Leiden 1936. 106 Op. cit.,
p. xi. 107 H.T.M. van Vliet: “Altphilologie und Editionswissenschaft in den Niederlanden”.
In: Geschichte der Editionsverfahren vom Altertum bis zur Gegenwart im Überblick, p. 42.
Later times were not exempt from the problem of additions and inter-
polations. One example108 would be the lost beginning of Aśvaghos.a’s Bud-
dhacarita.109 The text survives in one very old Nepalese manuscript (perhaps
1300), where the beginning is missing. Then we have what must be copies of
this manuscript, which however do contain the beginning. The first editions
print this beginning as if it was part of the text, but it was actually written by
the Nepalese Pandit Amr.tānanda in 1830, and as shown by Okano110 this be-
ginning has its origin in the Mahāsam . vartanı̄kathā. This augmented version
was the source of the later manuscripts. Here the rule that such interpolations
have to be dismissed is all-important in order to avoid the impression that this
part of the text is actually by Aśvaghos.a. If we print such additions within the
text, then there is the danger that even in academic circles, who should know
better, it may gradually be seen as part of the text. I am saying this rather
self-critically, as our next example will show, where a text written by an editor
has acquired the status of an original.
111 112
The Vim . śatikā, or preferably Vim. śikā of Vasubandhu was, like so
many other Buddhist texts, not transmitted any more in India, but survives
only in a single palm-leaf manuscript in the Durbar Library now kept in the
National Archives, Kathmandu. In view of the great importance of this text for
Buddhist idealist or illusionist thought this is all the more astonishing. Under
these circumstances the task of an editor may seem straightforward, because
he could be content with reproducing the text of the single manuscript with
occasional emendations, if these should prove necessary. But fortunately or
unfortunately, the parallel transmission of this text in Tibetan and Chinese
versions provides ample grounds for text-critical interventions.
108 For the following I am grateful to Roland Steiner. 109 See Roland Steiner: “Truth under
the Guise of Poetry. Aśvaghos.a’s “Life of the Buddha”.” In: Lives Lived. Lives Imagined. Biography
in the Buddhist Traditions. Ed. Linda Covill et. al. Boston 2010, p. 89–121. 110 Kiyoshi Okano:
Sarvaraks.itas Mahāsam 111 For this and
. vartanı̄kāthā. Ph. D. dissertation. Marburg 1997.
further materials, see Kano Kazuo: “Two Short Glosses on Yogācāra Texts by Vairocanaraks.ita:
Vim . śikāt.ı̄kāvivr.ti and Dharmadharmatāvibhāgavivr.ti.” In: Francesco Sferra: Sanskrit Texts
from Giuseppe Tucci’s Collection. Part I. Roma 2008, p. 350 and 353. 112 The following is a
brief summary in English of Jürgen Hanneder: “Vasubandhus Vim . śatikā 1–2 anhand der
Sanskrit- und tibetischen Fassungen”. In: Festschrift für Michael Hahn zum 65. Geburtstag von
Freunden und Schülern überreicht. Ed. by Konrad Klaus and Jens-Uwe Hartmann. Wien
2007, p. 207–214.
An edition of the Vim . śatikā was published as early as 1925 by Sylvain
113
Lévi from the codex unicus, and with the help of the Tibetan and Chinese
versions. This has remained the basis of further prints and studies. Much
later, in 1989, this manuscript was made accessible in a facsimile edition.114
On this basis a new edition with a Polish translation and a commentary was
published in 1999,115 in which an attempt is made to examine and improve
Lévi’s edition with the help of the manuscript.
When I read the text of the Vim . śatikā some time ago in a seminar held at
the indological Institute in Halle, it was natural to make the latest edition by
Balcerowicz the basis of the reading class, since there some misreadings
from the first edition had been corrected. We started reading the text and,
since none of us could read Polish, we altogether missed the fact that the first
two verses—although given in the edition in Sanskrit—are not actually trans-
mitted in Sanskrit. As is so often the case, the beginning of the manuscript
was missing.
When we retraced the sources, we saw that in Lévi’s edition the first two
verses are in smaller print and we could read that they had been translated
back from Tibetan and Chinese into Sanskrit by Lévi himself. Despite Lévi’s
care, the habit to print a complete text produced a natural tendency to view
verses one and two as part of the text, and this gradually led to a full integration
of the verses by Lévi in some cases.116 I was led into the same trap.
The example is mainly given to show what critical philologists know well:
if one makes the effort to investigate the sources for oneself, the outcome may
be quite unexpected. So when I tried to atone for my pramāda and conducted
a reinvestigation of the sources of this text, there was a surprise. The facsimile
edition contained not only the manuscript used by Lévi, but two more. One
was a direct copy of the old manuscript, the other transmits only the Kārikās
without the auto-commentary. To my surprise the latter contained all Kārikās,
that means also verses one and two.
Sanskrit Manuscript. The Trisvabhāvanirdeśa, the Vim . śatikā with its Vr.tti, and the Trim
. śikā with
Sthiramati’s Commentary. Ed. by Katsumi Mimaki, Musashi Tachikawa, and Akira Yuyama,
Tokyo 1989. 115 Piotr Balcerowicz and Monika Nowakowska. In: Studia Indologiczne 6
(1999), p. 5–44. 116 T. R. Sharma: Vijñaptimātratāsiddhi. With Introduction, Translation and
Commentary. Delhi 1993.
So let us first look at the published version, for it is the one that has been
printed and translated:
The variants against Lévi are idam against etat, yadvat against yathā, which
I think is blameless, taimiraka for taimirika, a type of common variation
in manuscripts, and most importantly, keśon.d.ukādi for keśacandrādi. We
might deliberate on the cause for the reading in the Tibetan, but the Sanskrit
version seems convincing, and one should not forget a small but decisive
detail: keśacandrādi is not a transmitted text.
Examples in this lecture often illustrate a text-critical principle, sometimes
merely a more anecdotal morale, like the famous rule of Kosambi that the
value of a manuscript is “inversely proportional to the fuzz made in lending
it”.117 One other such rule would be that once we unravel the genesis of an
edition, things may become much more complicated than we could have ever
imagined. In the previous example we could easily solve one textual problem
and dismiss a text that was written by a scholar rather than by the author, a
cautionary tale about so-called re-translations from Tibetan into Sanskrit. But
a second look at the manuscripts showed a more fundamental problem.
The text is contained twice in the Tibetan Canon, as no. 4061, which
contains only the Kārikās, and 4062 which contains the Vr.tti with the Kārikās
in the peculiar way interspersed with the auto-commentary as we know it
from the Sanskrit version.118 In no. 4061 we find the Tibetan of verse 1 as
printed above, but 4062, the version with the Vr.tti, does not transmit this
verse. La Vallee Poussin briefly alludes to this in a footnote to the translation,
but the implications of this have not been made clear and were not thought
through. In the commentated version the content of the first verse is expressed
in prose, nothing is missing from the argumentative structure of the treatise.
In other words, there were obviously two versions of the Vim . śatikā, one
Kārikā version and one Savr.tti version, both complete as far as the contents
are concerned, but differing in one verse. One likely explanation would be that
the commented version is the original, but when one wanted to transmit the
Kārikās separately, the main philosophical proposition, which was formulated
in prose, would have been missing and so this was made into an additional
verse, our verse 1. Only in Lévi’s edition this verse, which fits only in the Kārikā
version, was added to the commented version, thus producing a Vim . śatikā
with 22 verses and an ahistorical conflated text. Here the Tibetan translators
who preserved both versions separately were text-critically more far-sighted.
117D. D. Kosambi: Epigrams attributed to Bhartr.hari. Bombay 1948, p. 10. 118 The Tibetan
Tripitaka. Taipei Edition. Vol. XL. Nr. 4061 starts on p. 348, folio 5, line 4.
Coming back to our starting point, which was “cooking manuscripts”, it
appears that there are many methods to alter transmissions, and even with
the best of intentions.
The crucial difference is, however, that in modern editions an introduction,
the footnotes or other paratexts acquaint us with the history of the text and
other details. Pre-modern editors have on the contrary often worked without
leaving traces and changed a text silently. Such manuscripts, cleverly and
silently emended, may stand out in a group of manuscripts easily as the “best
manuscript” among more menial apographs. Absurdly, following such a best
manuscript is not always a bad idea, since the editor could have well known
more about the text, its author and the cultural background than a modern
editor, thus his chances of restoring a reading correctly might have been much
higher. But we cannot be sure, it may well be just an application of conjectures,
or an attempt to “improve” the text.
So how can we find out about the principles employed by pre-modern
Indian editors? There are a few normative texts that incidentally deal with our
topic. One is on the donation of manuscripts,119 where we find descriptions
of the qualities of a scribe, rules for writing manuscripts, how to draw the
lines, which script to use, how to prepare the pen, the ink etc. Technical
terms are given there, as the manuscript’s dust jacket (malapr.s.t.ha) etc. Most
important for our purpose are the rules for correcting the text.120 The scribe
was supposed to correct by taking into account metre, context, surrounding
words, proper grammar, and as Dutta notes, this was to be done without prior
collation of other manuscripts: “Obviously it is one of the causes that brought
forth the widespread corruption of manuscripts in this country.”121 According
to the European terminology this would amount to an emendatio ope ingenii.
But these Purān.ic sources were fortunately not the last word on editing.
We may assume that for scholars it was clear that in order to edit a text a
survey of sources had to be made. But what to do with the evidence? Is it, as
119
The passage from Hemādri’s Caturvargacintāman.i and others are discussed in Kali Kumar
Dutta: “The Ritual of Manuscripts”. In: Our Heritage 19.1 (1971), p. 15–44. I am very grateful to
Manu Francis for alerting me to this work. 120 Op. cit., p. 31ff. 121 Op. cit., p. 32.
Rāmakan.t.ha tells us, unnecessary, or even a confession of one’s own insecurity
to talk about readings? No, says Abhinavagupta:122
122 Manavalli Ramakrishna Kavi (ed.): Nāt.yaśāstra with the commentary of Abhinavagupta.
Baroda 1926, vol. 1, p. 1f. 123 atra kecid iti pat.hanti [. . .] ayam
. ślokah. atratya iva na laks.yate,
pustakes.u kena kāran.ena gata iti na vidmah.. Madhusudan Kaul Shastri: The Ghat.akarpara
Kāvya of Kālidāsa. With the Commentary of Abhinavagupta. Srinagar 1945, verse 20, p. 19.
124 kecit tu śakārathakārayor lipisārūpyamohāt prathama ity ūcuh. E. Hultzsch: Kalidasa’s
.
Meghaduta. London 1911, p. 3.
text-critical presuppositions. Vallabhadeva assumed that a scribe had con-
fused the letters, in other words he has given a paleographical argument for
the source of error and appeals to meteorological facts for the solution, what
Kantorowitz has later termed the criterion of “Sachgeschichte”.
Thus, by far the best way to see an Indian editor practise his art is in
commentaries, where from stray remarks on variant readings we can gather
his method. Admittedly the literature to be examined is extremely vast, but
some scholars have done their best to collect statements of commentators on
matters of textual criticism.125
From these collections it becomes abundantly clear that commentators
often, perhaps as a rule, compared manuscripts systematically. Even though
only few readings are reported, the frequency of references to variant read-
ings suggests that this meant more than an occasional glance at a second
manuscript, even if this is naturally difficult to prove.
If we review the information on commentators given so far, we find a
variety of pre-modern Indian approaches to editing. We have already seen
in the above quotation from Rāmakan.t.ha something like the idea of silent
emendation. His contemporary Abhinavagupta on the contrary favoured
a discussion of competing variants. The famous commentator Mallinātha
regularly in his commentaries produces a statement, which we might interpret
as an abstinence from conjecture, although his statement is too unspecific
to be sure.126 Editors also mention their sources—in the absence of library
catalogues we cannot expect more than general adjectives to describe these
codices—, or the fact that they have collected manuscripts from different
areas, as Nı̄lakan.t.ha, the commentator on the Mahābhārata.127
125 Especially noteworthy are: Gérard Colas: “The Criticism and Transmission of Texts in
Classical India”. In: Diogenes 47 (1999), p. 30–43. French version: “Critique et transmission
des textes dans la littérature sanskrite”. In: L. Chiard and Chr. Jacob: Des Alexandries I.
Du livre au texte. Paris 2001. Ram Shankar Bhattacharya: “Use of manuscripts in textual
criticism by our commentators.” In: V. V. Dvivedi, J. Pandeya (ed.) Sampādana ke Siddhānta
aur Upādāna. Samyak-Vāk Series 5 (1990), p. 200–221. 126 nāmūlam likhyate kimcit [. . .]
. .
(The Raghuvam . śa of Kālidāsa. Ed. H. D. Velankar Bombay 1949, p. 1) might also mean that he
held his commentary free from irrelevant discussions. 127 bahūn samāhr.tya vibhinnadeśyān
kośān viniścitya ca pāt.ham agryam / prācām . gurūn.ām anusr.tya vācām ārabhyate bhāratabhāva-
dı̄pah.. The Mahābhāratam. [. . .] nı̄lakan.t.haviracitabhāvadı̄pākhyat.ı̄kayā sametam. Delhi: Nag
Publishers 1988, introductory verse 6.
A further category is formed by statements about the validity of certain
readings, from which we can reconstruct criteria for the selection of variants.
This sometimes also involves an assessment of the status or quality of the
manuscript, from which the reading stems. Here age can be one criterion: it is
mentioned, for instance, whether a manuscript is ancient, old, or modern. It
may be from a certain area, or “another area” (deśāntara) implying probably
a different script. We read of good manuscripts (satpustaka), or of decrepit
ones (jı̄rn.a). The implication here is of course that “old” implies authority,
a fact we can gather from a polemical remark on the blind belief in old
manuscripts: Jayanta Bhat.t.a, when commenting on the validity of tradition
says that everyone could produce a decrepit manuscript and claim that it is an
(unknown but valid) Āgama.128
There is unsurprisingly no stemmatics, but there are proto-stemmatical
considerations as we find them in pre-modern European critics. Agreement
in manuscripts is often a criterion for the adoption of a reading, and we have
an argument remotely resembling the eliminatio singulorum lectionum, when
the kvacitkah. pāt.hah. is not followed.
Beyond considerations on manuscripts and on constellations of manu-
scripts, the value of singular readings was of course assessed. Here we find a
large repertoire of terms and concepts that await further study. Ram Shankar
Bhattacharya has presented an invaluable catalogue of such passages and an-
nounced a monographical volume on the topic,129 which unfortunately never
appeared. For the present purpose I merely give a few examples from the liter-
ature mentioned above, enriched with material assembled through a quick
and rather coarse method, namely a search for the word “(variant) reading”
(pāt.ha) in the Göttingen Register of Texts in Indian Languages (gretil). The
computer-based search in this archive yielded many pertinent passages, in
which commentators were arguing for or against a reading. Commentators
talk about reading words as a compound (samastapāt.ha) rather than sepa-
rating it, they talk of a split in readings (pāt.habheda), that is, the existence of
variants in a given passage, we find augmented or deficient readings (adhika-,
132 Dominic Goodall: “Bhūte ‘āha’ iti pramādāt: Firm Evidence for the Direction of Change
Where Certain Verses of the Raghuvam . śa are Variously Transmitted”. In: ZDMG 151 (2001),
p. 103–124. 133 Ed. T. Venkatacharya. Delhi 1980. 134 na ksunno ’yam parair mārgas
. .. .
tatra sam carato mama. (Sāhityakan takoddhāra 11a). 135 One highly technical, āyurvedic
. ..
example is given by P. K. Gode in his Textual Criticism in the Thirteenth Century (Woolner
Commemoration Volume. Lahore 1940), p. 107f. Discussing one passage in Hemādri’s commentary
on the As.t.āṅgahr.daya, Gode comes to the conclusion that this commentator has a “love of
elaborate defense (samarthanā) of a single reading”. 136 Colas: Criticism and Transmission,
p. 30.
criticism” did not aim at reconstructing text “in its original form as the author
conceived it” but merely a best version.
The question is of course just an opener and the material provided by
Colas shows that “Indian textual criticism” as well as “European textual criti-
cism” are collective terms for highly heterogeneous practises. Therefore the
conceptual opposition between “Western” and “Eastern” criticism can be
misleading.137 What we may meaningfully compare are specific practices, and
here there are significant overlaps or even parallel methods.
137 For instance, Colas’ idea that it is “only modern research that has shed light on these inten-
tional interventions in the manuscript transmission.” (p. 35) would have to be modified in view of
the frequent identification of praks.ipta verses by commentators. There was also clearly an idea of
the style of the author as the main criterium, as we can read in the above-mentioned Ks.emarāja’s
commentary on Śivastotrāvalı̄ 20.21, where he judges a stanza to be not by Utpaladeva because
it displayed a style unworthy of him (asadr.śaśailı̄). 138 H. T. M. van Vliet: “Altphilologie
und Editionswissenschaft in den Niederlanden”. In: Hans Gert Roloff (ed.): Geschichte der
Editionsverfahren vom Altertum bis zur Gegenwart im Überblick. Berlin 2003, p. 41f.
• Rejection of conjectures even when readings are clearly corrupt, the
rule: sthitasya gatiś cintanı̄yā.139
• pāt.hakramād arthakramo balı̄yān. The idea that emendation is neces-
sary and that it has to start from the contents. Or phrased with typical
Housman polemics: “There is a foolish sort of conjecture, which seems
to be commoner in the British Isles than anywhere else, though it is also
practised abroad, and of late years especially in Munich. The practice
is, if you have persuaded yourself that a text is corrupt, to alter a letter
or two and see what happens. If what happens is anything which the
warmest good-will can mistake for sense and grammar, you call it an
emendation; and you call this silly game the paleographical method
[. . .] Haupt, for example, used to warn his pupils against mistaking
this sort of thing for emendation. ‘The prime requisite of a good emen-
dation’, said he, ‘is that it should start from the thought: it is only
afterwards that other considerations, such as those of metre, or possi-
bilities, such as the interchange of letters, are taken into account.”140
In the criticism of the epics commentators have always been aware of the fact
that the text often is problematic and has suffered at the hands of redactors
or users. In an article on Textual Criticism in the Thirteenth Century P. K.
Gode141 quotes from the manuscript of Madhvācārya’s commentary on the
Mahābhārata (BORI 275 of 1892–95), who comes to this diagnosis:
139 The translation of this maxim is somewhat difficult: “It is necessary to consider the situation
of what is there”. (Colas) “Il faut réfléchir à la situation de ce qui est là.” (Colas). Much more to
the point is “Some way must be sought to understand the text as it stands” (Pollock, reported by
Colas). 140 “The Application of Thought to Textual Criticism”. In: The Classical Papers of A. E.
Housman. Collected and Edited by J. Diggle and F. R. D. Goodyear. Cambridge: University
Press 1972, p. 1064f. 141 Woolner Commemoration Volume. Lahore 1940, p. 106. 142 grantha
is here used in the sense of a 32-syllable unit.
remove them, elsewhere they exchange them by negligence or
elsewhere in other ways.
In other words, the medieval diagnosis from within the tradition was fairly
drastic. Gode himself speaks of “vandalic tendencies” and “malpractices”.143
All this is hardly new or unexpected. Already Jacobi in his work on the
Rāmāyan.a has shown that in the transmission of this text commentators
followed different maxims and principles and criticized each other for it. I
shall translate the interesting passage here in full:144
After the main recensions had been determined and then were
transmitted mostly through writing through learned tradition,
their fate was quite comparable to those of other such texts.
The commentators deleted verses and passages they termed as
praks.ipta, chose between various readings, or tried to improve
a corrupt passage through conjecture. From the Tilaka we can
get a clear picture of these processes and an impression of the
opinion of its author Rāmavarman on textual criticism. There
readings are regularly mentioned, rejected or justified, by being
labelled “old” (prācı̄na), “transmitted” (pāṅkta or sāmpradāyika),
“based on many manuscripts” (bahupustakasammata), but also
“not transmitted” (apāṅkta) or “conjecture of a recent author”
(ādhunikakalpitah. pāt.hah.).
[Footnote:] Maheśvaratı̄rtha seems to have been more bold in
conjecturing. Rāmavarman often mentions his conjectures with-
out approval [. . .]: Tı̄rthas tu ‘atra jı̄vitasaṅgamah.’ iti pāt.ham
kalpayāmāsa.
Thus one simply needs to take into account the obvious fact that there were
different methods and principles in pre-modern Indian textual criticism.
A wholesale comparison with modern “scientific” methods (which ones?)
is likely to give a distorted picture. What this means for the prospect of
editing texts through the method of Lachmann should be by now clear. If the
transmission of a text displays too many excellent readings, many marginal
143 Ibid., p. 107. 144 Hermann Jacobi: Das Rāmāyan.a. Darmstadt 1970 [Bonn 1 1893], p. 9f.
variants, too few scribal mistakes and does not fall neatly into stemmatic
branches, then we might have among our sources, or subarchetypes one
old-Indian edition. In this case the Lachmann method cannot work.
In other words, according to Witzel, for the editors of the Veda the process
of recensio, collecting and analyzing all sources, was almost dispensable,
since there was only one oral text, give or take a few writing mistakes. The
indological problem, according to Witzel, was that later generations continued
with this method, although their text transmissions were entirely different.
This explanation of matters may appear sound, but has its own flaws, visible
already in the fact that Müller does not adduce the orality of the R.gveda. If we
want to see how such an argumentation would look like, we only need to read
a book that appeared also in Oxford only few decades later, namely Grierson
and Barnett’s edition of the sayings of the Kashmirian Saint Lallā, which are
based on an oral text as well as on a differing written version. Grierson called
this record “in some respects more valuable than any written manuscript”.146
Most importantly the oral and the written version are edited separately!
Müller does not say that he is taking into account an oral tradition. He
explicitly says that he is following the method of (amongst others) Lachmann.
And even worse, he is probably right.
But let us start with the basics. Every Sanskritist knows and every student
hears about the R.gveda having been memorized by heart for thousands of
years by an elaborate method that safeguarded its textual integrity. As a
result the Vedic recitation up to modernity preserved the text of the oldest
Indian literary document without change. According to our textbooks, a large
146 Lallā-vākyāni. Ed. George Grierson and Lionel Barnett. London 1920, p. 5.
number of linguistic studies has proved this and it is taken as an established
fact in Indology. I was also brought up to believe that while we may use textual
criticism for all sorts of texts, the Veda is beyond that, simply because there
are no variants. We read as much in Jan Gonda’s standard handbook on Vedic
Literature in his History of Indian Literature: “[. . .] the text of the R.gveda
has for many centuries remained unaltered—there are no variants [. . .]”147
For this statement he refers to the preface to the edition of the R.gveda by Max
Müller and other scholars have since reiterated this. The latest, more popular
restatement of the same is perhaps the following: “The entire corpus was
preserved orally with razor-sharp precision for three millennia, as if it had
been engraved in the neurons of Brahman families committed to reciting and
preserving it.”148
Of course, for a non-Vedicist used to reading non-Vedic Indian manu-
scripts this is a priori hard to believe. Here I am refraining from wandering
into more theoretical, or even polemical avenues, and have therefore excluded
the uninformed or absurd,149 but one such remark may be permitted. The
claim that textual criticism is inappropriate or has no scope is usually applied
to texts with a divine status. From the insider’s point of view such texts are
often not transmitted but revealed and therefore there is at least an unease
when it comes to variants. For when the revealed text is conceived of as a
source of correct knowledge beyond the scope of others (āgama as a pramān.a
beyond the others), there is not much space for logical arguments about the
correct reading. Here East and West, again, do not differ.
As a result there is a marked tendency to claim a variant-free transmission
for religious texts. To give another example:150
147 Wiesbaden 1975, p. 18. Many other similar quotations could be adduced. 148 David Shul-
man: “At the Heart of Hinduism”. In: The New York Review of Books LXII.1. 149 A combination
of both can be found in the writings of Vishwa Adluri and Joydeep Bagchee, who think that tex-
tual criticism is “mechanical”, whereas the related “historical criticism” is “pseudo-critical” and
“anti-semitic”. The Real Threat to the Humanities Today: Andrew Nicholson, The Nay Science, and
the Future of Philology. Publication on academia.edu, p. 2. 150 Sheldon Pollock: “Literary
Culture and Manuscript Culture in Precolonial India”. In: Simon Eliot et al.: Literary Cultures
and the Material Book. The British Library 2007, p. 88.
fered by the Bangla-language Caitanya-caritamrta (Immortal
Deeds of Caitanya) of Krsnadasa Kaviraj, a poetic biography of
the religious reformer Caitanya (died c. 1533), composed around
1600, not in Bengal, but far to the west in Vrindavan, an impor-
tant sacred centre of Vaishnavism. This is one of the most often
reproduced texts in the history of Indian manuscript culture, now
existing in more than two thousand copies—virtually identical
copies. There is none of Eisenstein’s ‘textual drift’ here [. . .]
I find it hard to believe that firstly this is the case and secondly that anyone
could have possibly ascertained the fact through a quick check of all 2000
manuscripts. But the interesting question would be, how one could safeguard
a text in that way.
Let us pretend that the critical edition of the R.gveda is just a normal
critical edition. Its history is connected with two scholars with interesting,
but accidental biographical similarities. The first editor was Friedrich Rosen,
a young and gifted German scholar who studied in Berlin, moved to Paris
and then took up the chair for modern oriental languages in London. He
did not apply for the Boden Sanskrit chair in Oxford because of its religious
orientation. He must have been an interesting figure, but not much is known
about him. Rosen’s pioneering attempt to edit the R.gveda was published
posthumously in 1838. His premature death prevented not only the completion
of his endeavour, but also the publication of his notes on the text. What we
have is a fragmentary edition with Latin translation, a remarkable work at a
time when many scholars still thought that the Veda contained the original
monistic religion of India. Rosen had merely two manuscripts at his disposal.
The next editor, Max Müller, was also German, studied in Leipzig, moved
to Paris, and then to England. He did apply to the Oxford Sanskrit chair, was
according to all we know the best candidate, but did not get it for the same
religious reasons.151 When Max Müller made a new attempt at editing the
R.gveda, he described his method as follows:152
151See Richard Gombrich: On Being Sanskritic. A Plea for Civilized Study and the Study of
Civilization. Oxford: Clarendon Press 1978. 152 Rig-Veda-Samhita. The Sacred Hymns of the
Brahmans together with the Commentary of Sayanacharya. Ed. Max Müller. Vol. 1. London: W.
H. Allen & Co. 1849, p. ix.
My principal object in this present edition is therefore to give
a correct text of the Rig-veda, and to restore from the MSS. a
readable and authentic text of Sāyan.a’s commentary. The former
was by far the easier task. The MSS. of the Rig-veda have gen-
erally been written and corrected by Brāhmans with so much
care that there are no various readings in the proper sense of the
word, except those few which are found noticed as such in the
commentaries or in the Prātiśākhyas. Even these are generally of
small importance, and seldom affect the meaning of a sentence.
One needs to let this information sink in for some time. Müller concluded
from the lack of variants in his manuscripts that they had been corrected
and thus presented the correct text. After rereading his introduction a few
times, I had the impression that Müller’s textual criticism of the R.gveda is
a topsy-turvy world, not unlike the one Alice encounters in the novel of his
Oxford colleague Lewis Carrol. In this world the correct text is not something
to be established, it is established before textual criticism begins, not despite,
but because of the intervention of later generations of transmitters. In other
words, everything we have learned from textual criticism is valued differently
here. Whereas Lachmann could have excluded all of Müller’s manuscripts
as contaminated on suspicion that later scribes had corrected them, here
Brahmins as guardians of the text have a completely different role. Of course,
the R.gveda, when seen as oral literature may be a special case, but neither
does Müller argue that it is, nor are his sources based on recordings or direct
intervention of Pandits; he is still editing a text from manuscripts, and the
extent to which normal rules of textual criticism are cancelled, is remarkable.
If we view this world from our normal coordinates, we find potential
problems in Müller’s approach. Whereas Rosen used two manuscripts, Müller
used three (plus two for the Pada-text), surely not an impressive coverage
of potential variant bearers. Any further collation of sources he considers
superfluous and argues:
Of course, this could still be true, although one or two checks of materials that
may not share an ancestor would have been reassuring. We should also be
told, why we cannot apply the principle to other texts. If we select few enough
manuscripts of a text of Kālidāsa and think that we in any case know what he
meant, we can surely dispense with most of textual criticism. Phrased in this
way, it is hardly short of the kind of text-critical madness as diagnosed by his
Cambridge colleague Housman, which we shall discuss shortly.153
Müller also adduces as an additional argument for his approach the first
edition of Rosen, which only serves to make things worse. He says that it was
for that same reason that Rosen in his first edition did not add an apparatus.
This argument is invalid, because at the time of Rosen, many editors did not
publish a collation of variants. It was a contemporary of Rosen, Christian
Lassen, who first criticized this practice in Indology. But Müller here twists the
argument, for he knew that Rosen’s work appeared only after his premature
death and that most of his notes were never printed.154
So in effect all is based on the notion that the Brahmins were the true
transmitters of the Veda, and that their knowledge of the correct text went
153 Max Müller, for instance, thought that the edition of the Hitopadeśa by Schlegel and Lassen
was ecclectic and that they should have used one manuscript as the basis. See the Review by
N. Delius in Zeitschrift für die Kunde des Morgenlandes 7 (1850), p. 231. 154 Lassen, when
reporting on Rosen’s work in the Zeitschrift für die Kunde des Morgenlandes 3 (1840), p. 471f., says
that he had seen Rosen’s papers, which contained a collation of variants from manuscripts not
used for the edition. Lassen in general may be the first to have stated that the Veda has no true
variants. He says: “Es kommen nämlich kaum wirkliche Varianten vor.” (ibid., p. 472). Later he
states after saying that the variants are mostly writing mistakes: “somit erklärt es sich hinreichend,
dass auch die Hymnen, trotz ihrer Heiligkeit, keineswegs von Schreibfehlern frei geblieben sind.
In der That sind die meisten Varianten, die mir vorliegen, dieser Art. Bei der grossen Anzahl von
Handschriften, der Ausführlichkeit der Scholien und den vielen Ausführungen der Veda-Stellen
in grammatischen und anderen Büchern sind jedoch diese Fehler unschwer zu beseitigen. Man
sieht zugleich, dass solche Abweichungen der Authenticität der alten Texte keinen Eintrag thun
können und dass ein gewissenhafter und so reich ausgerüsteter Herausgeber, wie Rosen, sich
die Berechtigung geben durfte, solche Varianten nach sicherer Herstellung des Richtigen mit
Stillschweigen zu übergehen, wenn dieses, wie ich vermuthe, seine Absicht war.” (ibid., p. 473).
beyond manuscript evidence. I am of course not arguing against this theory
of oral transmission. But one tends to forget that Müller’s edition is not based
on such a theory of Vedic orality, but on a very peculiar understanding of
textual criticism. This is made clear when he comes to his edition of Sāyan.a’s
commentary:
For the first Asht.aka I had twelve MSS. However, we have learnt
from Greek and Latin philology that a great number of MSS. is
not at all desirable for critical purposes. In most cases those nu-
merous MSS. which have been collated for classical authors have
only served to spoil the text; to make the reading of doubtful pas-
sages still more doubtful; and to give rise to a mass of conjectural
readings, based either upon authority of the transcriber of a MS.,
or upon that of an ingenious editor. In this manner an immense
deal of labour has been wasted in classical philology; so that now,
after the simple rules for using MSS. have been laid down by a new
school of critical philologers, such as Bekker, Dindorf, Lachmann,
and others, almost all the old editions of classical authors have
become useless for criticial purposes, with the exception of some
of the editiones principes, which, as they simply reproduced one
MS., though generally a very bad one, can claim for themselves at
least a certain degree of authenticity. (Vol. 1, p. xix–xv.)
Müller had found this conviction already in 1844 after two years of, as he says,
Sanskrit study in his leasure time,155 when he translated the Hitopadeśa into
German. Although the work clearly shows that Müller’s knowledge of Sanskrit
was not up to the task,156 he criticizes the editors Schlegel and Lassen for
producing an eclectic text instead of sticking to one manuscript.157 While his
knowledge of Sanskrit seems much improved in his edition of the Hitopadeśa
some decades later in England, his text-critical convictions remain unchanged.
He dismissed the text of Schlegel in favour of an uncritical Indian print of
1830. In the introduction we read:158
155 Thus in a letter to A. W. Schlegel of May 1844. 156 Nicolaus Delius in his review in Zeitschrift
für die Kunde des Morgenlandes 7 (1850), p. 230–244, lists his translation mistakes. 157 Hitopa-
desa. Eine altindische Fabelsammlung aus dem Sanskrit zum ersten Male in das Deutsche übersetzt
von Max Müller. Leipzig 1844. 158 The First Book of the Hitopadeśa. London 1864, p. vii.
That text was chosen as an authority, partly because it was desir-
able to have, as far as possible, the same text in the examinations
in India and England, partly because an eclectic text, even one
so carefully elaborated as that of Schlegel and Lassen, seemed
to be incompatible with those principles of diplomatic criticism
which are now adopted by all sound scholars, not only in Greek
and Latin, but likewise in Sanskrit and Oriental literature. No
attempt has yet been made to arrange the numerous MSS. of the
Hitopadeśa genealogically, and there is hardly another work with
which each copyist has ventured to take such liberties as with
this, the most popular story-book of India. Until MSS. have been
genealogically arranged, a selection of certain plausible readings
from this or that MS. is worse than useless. In my translation
of the Hitopadeśa, published in the year 1844, I pointed out that
an eclectic restoration of the text, even if carried out by men of
taste and profound scholarship, could never satisfy the demands
of modern criticism. As the labour of collating and classifying
the MSS. of the Hitopadeśa would have been very great, and as,
owing to the nature of this popular work, the result would always
have been problematical, I determined to make no attempt at a
critical restoration of the text, but to adhere throughout to one
native authority.
The critique is particularly inept in this work, where Müller has changed the
text of his single source – as he says – for reasons of grammar and »decency«.
Müller’s text-critical ideas coincide with the observation that Lachman-
nians often merely tried to reduce their sources in order to keep it simple.
This has by the way not much to do with what we today understand as the
method of Lachmann. The danger involved in this method of such a rapid
and convenient exclusion of witnesses is that one is likely to misjudge their
value.
In indological editing one further complication arises through the variety
of scripts employed in the transmission of Sanskrit texts. Müller has not
heeded one of the rules of Sanskritic textual criticism, to use not just De-
vanāgarı̄ manuscripts.159 Regularly manuscripts in local scripts transmit local
recensions and a proper recensio is incomplete as long as one limits oneself to
Nāgarı̄ sources. Of course, Müller is not to blame for the fact that there were
so few manuscripts of the R.gveda accessible to him in Europe.
But when his colleague Georg Bühler, who was travelling India in search
of manuscripts, found what is considered one of the oldest Vedic manuscripts,
he thought it useful to alert Max Müller to his findings:160
159 Actually in his edition the script is not stated, but it is understood that all his Veda manuscripts
are in Devanāgarı̄. He seems to mention the script only when a manuscript is not written in
Devanāgarı̄. One manuscript of the commentary (C6) is in Bengali characters. 160 Georg
Bühler: Detailed report of a tour in search of Sanskrit Mss. made in Kásmîr, Rajputana, and
Central India, p. 35. 161 Preface to Vol. 3, p. 26. Quoted in I. Scheftelowitz: Die Apokryphen
des R.gveda (Khilāni). Breslau 1906, p. 35.
contained no independent readings, and that it had by no means
been carefully copied.
162 Max Müller: Vedic Hymns. (Sacred Books of the East 32). Oxford 1891, p. 31f. 163 “The
critical principles by which I have been guided in editing for the first time the text of the Rig-veda,
require a few words of explanation, as they have lately been challenged on grounds which, I think,
rest on a complete misapprehension of my previous statements on this subject.” Max Müller:
Vedic Hymns, p. 43.
fixing on three representative MSS., as described in the preface
to the first volume of my edition of the Rig-veda. Even these
MSS. are not free from blunders,—for what MS. is?—but these
blunders have no claim to the title of various readings. They are
lapsus calami, and no more; and, what is important, they have
not become traditional.
Here finally we have arrived at real textual criticism, for the examples Müller
gives are more than telling. He says in the foonote: “Thus X, 101, 2, one of
the Pada MSS. (P2) reads distinctly yagñám prá krinuta sakhâyah, but all the
other MSS. have nayata, and there can be little doubt that it was the frequent
repetition of the verb kri in this verse which led the writer to substitute krinuta
for nayata. No other MS., as far as I am aware, repeats this blunder.” This and
the following examples hidden in this footnote show that there were indeed
variants even in the few manuscripts used by Müller, where he established
the text by text critical reasoning, while at the same time claiming that there
were no variants. And when these variants were becoming too threatening,
he introduced the absurd category of a “traditional” variant and argues with
what “all the other manuscripts” read and that there are readings “no other
manuscript” transmits. But let us recall that the total number of manuscripts
used is very small and the “tradition” invoked here is just the European based
moks.amūlaraśākhā he takes as the final word in editing.
Coming back to the Kashmirian birch-bark manuscript, perhaps Bühler
was not too happy with the decision, in any case he forwarded the manuscript
to a scholar who took more care to analyse it: Isidor Scheftelowitz164 closely ex-
amined the manuscript in his work on the Khilas of the R.gveda, and criticized
Müller for excluding that source:165
164 The following passage has been brought to my attention by my colleague Walter Slaje.
165 I. Scheftelowitz: Die Apokryphen des R.gveda (Khilāni). Breslau 1906, p. 35.
not relied on Wenzel, he would surely have arrived at exactly the
opposite verdict.166
166 “Doch Müllers abfälliges Urteil über die Kaśmir-Handschrift, welches sich auf einen ungenü-
genden Auszug des Dr. Wenzel stützt, ist unberechtigt. Hätte Müller nur diese 11 sogenannten
Vālakhilyās, deren Lesarten er als ein Beispiel von dem Charakter dieser Handschrift anführt,
selbst nachgeprüft und sich nicht auf Wenzel verlassen, so würde er sicherlich gerade zum ent-
gegengesetzten Resultat gelangt sein.” 167 Op. cit., p. 176. 168 Op. cit., p. 36. 169 Op. cit.,
p. 46. 170 Op. cit., p. 46. 171 “Der Schreiber dieses Kaśmir-Ms. hat den R.V. oder die Khilāni
nicht aus dem Gedächtnis niedergeschrieben, sondern ihm lag ebenfalls ein im Śāradā-Alphabet
geschriebener Text vor, was mit Deutlichkeit aus verschiedenen Schreibfehlern hervorgeht.” Op.
when he discussed in his Prolegomena the Khilas. Luckily a transcript of the
Khilas by the late Dr. Wenzel came into the hands of Professor Macdonell, and
was used with important results in constituting the text of the Br.haddevatā
[. . .]”172 Macdonell, in the introduction to his edition,173 describes how he by
chance came into possession of Wenzel’s notebook containing his transcript of
the manuscript, when he met Leumann in 1902 in Hamburg at an Orientalist
conference. His estimate is quite different from the one by Müller: “It has
proved of utmost service, since it has saved me from several mistakes, besides
giving certainty in various passages where all would otherwise have been
conjecture.”174
It is to the credit of Scheftelowitz, who was supported by Hillebrandt,175
not to have given in. The following year he published an article, in which he
tried to spell out the results of his text-critical studies176 and listed alternative
and better readings from “Bühler’s” manuscript. For a comparatively young
scholar without a university position to demonstrate that Müller and the larger
part of the Vedic scholars had not only ignored the oldest Veda manuscript,
but were wrong in methodically excluding its readings, was quite daring. We
would want to think that this could not happen nowadays, but actually the
peer review system governing many journals could have easily prevented such
a publication. And of course it did not do Scheftelowitz much good. After
publishing the two interesting analyses of the Pune R.gveda manuscript in
1906 and 1907, he did not remain in the University, but worked as a Rabbi in
Cologne from 1908 onwards. When the university of Cologne was founded,
he became the first teacher, but only as an unsalaried “Honorarprofessor”, in
“Indo-Germanic Philology”. In the year 1933, when the Nazis rose to power, he
was barred from lecturing, with the ban on Jewish scholars his venia legendi
was eventually rescinded, and he narrowly escaped with his family to England
in 1934, received a position in Oxford, but died in the same year.
172 In JRAS 1907, p. 224–229. 173 Arthur Anthony Macdonell: The Br.had-Devatā attributed
to Śaunaka. Cambridge, Massachusetts 1904, p. xxxi. 174 Op. cit., p. xxxi. Although, as
Scheftelowitz has shown in his review (ZDMG 59, p. 420ff.), his ignorance of the Khilas has led
him to unnecessary emendations. The most recent critical edition on a broader manuscripts
base is Muneo Tokunaga: The Br.haddevatā. Kyoto 1997. 175 Scheftelowitz: Apokryphen,
p. 176. 176 “Zur Textkritik und Lautlehre des R.gveda.” In: Wiener Zeitschrift für die Kunde des
Morgenlandes 21 (1907), p. 85–142.
For Vedic Studies the issue of variant readings had long been put to rest.
Müller himself had written that he did not anticipate that “the text, as restored
by me from a collation of the best MSS. accessible in Europe, will ever be
materially shaken” and that “our text of the Veda is better authenticated, and
supported by a more perfect apparatus criticus, than the text of any Greek or
Latin author, and I do not think that diplomatic criticism can ever go beyond
what has been achieved in the constitution of the text of the Vedic hymns.”177
Then Oldenberg had already in his Prolegomena178 formulated the rule that
even Vedic parallels, as variants in the Sāmaveda, should only be considered
if all else fails, otherwise there should be “a categorial presumption for the
readings transmitted in the R.gveda or those conjectured from them”.179 Here
Oldenberg already walks a fine line. The stricture that the “text of the R.gveda”
is in fact the one edited by Max Müller from few Devanāgarı̄ manuscripts,
and that additional evidence is a priori inferior is absurd and can only be
explained by the fact that so much manpower had gone into this text, its
explanation, and conjecture. It seems the last thing the Vedicist establish-
ment wanted to hear was about variants of a newly discovered manuscript
not only in a regional script, but also much older than the sources surveyed
before. So when Scheftelowitz presented his findings two decades after Olden-
berg’s pronouncement, Oldenberg promptly dismissed all Vedic variants as
useless:180
177 Max Müller: Vedic Hymns, p. 45. 178 Hermann Oldenberg: Metrische und
There are many other interpretations which differ especially in the rendering
of ari, which has been interpreted as “Sippenangehöriger” (Witzel), “Fremder”
(Thieme), or “Standesherr” (Geldner).182 The obvious problem is that we
have a singular subject as well as a plural one, not connected by “and”, but
that these have to be construed with a plural verb. Of course, it is difficult
to say, whether this is a problem, or a feature of the Vedic language. The
reading is confirmed by the commentators, but also the unease with the
incongruence: Sāyan.a glosses arir uta as śatravo ’pi and also Skandasvāmin
and Veṅkat.amādhava comment on the problem,183 for the former the singular
181 Stephanie W. Jamison and Joel P. Brereton: The Rigveda. The Earliest Religious Poetry
of India. New York 2014, p. 93. 182 For an overview, see Michael Witzel and Toshifumi
Gotō: Rig-Veda. Das Heilige Wissen. Erster und zweiter Liederkreis. Frankfurt 2007, p. 491.
183 Rgveda. Ed. Vishva Bandhu. Hoshiarpur 1965, p. 27f.
.
is used for, or substituted for plural (vyatyayena caikavacanam), the latter
devotes half of his brief commentary on the problem to explain that there is
congruence. What else can one do if that is the text.
Here Scheftelowitz comes up with a different solution from “his” manu-
script, from the background of which he states that construing the text of
Müller is “quite impossible”:
1,4,6 abhí voceyuh., M[ax]M[üller] arír vocéyuh., doch ist arír ganz
unmöglich, da ja das Substantiv, worauf es sich beziehen könnte,
N. pl. f. kr.s.t.ayah. ist.184
This reading would have been worth considering, but since all experts had
already proposed solutions,185 the susceptibility for a new reading was low.
The reading of the older manuscript analysed by Scheftelowitz apparently
came too late. His fresh look at the readings had to compete with elaborate
justifications of what now became a canonised lectio difficilior. In 1909 Olden-
berg, when dealing with the passage, said that he wanted to keep to the text
as transmitted,186 without even mentioning the alternative reading. Since he
refers exclusively to works produced before the publication of Scheftelowitz as
justification,187 his adherence to the status quo without serious consideration
is unsound.
So what about later Indian editions of the R.gveda, which were based on
more manuscripts as well as on scholars who had memorized the text, the
often invoked oral tradition of the text: For instance, there was an exten-
sive effort to produce a new edition of the text188 involving—as it is stated
in the introduction—a team of ten (Mahārās.t.rian) Vaidikāh., of Śāstrins,
184 “Zur Textkritik und Lautlehre des R.gveda.” In: Wiener Zeitschrift für die Kunde des Morgen-
landes 21 (1907), p. 85. 185 Max Müller refers to the problem without mentioning the variant in
his Vedic Hymns (Sacred Books of the East 32), p. xli. Already Rosen wanted to solve the problem
by declaring arih. a plural: “possitne arih pluralis esse contracta terminatione, pro arayah?”.
186 “[. . .] halte ich am Ueberlieferten fest”. Rgveda. Textkritische und exegetische Noten, p. 4.
.
187 Pischel in ZDMG 40, Pischel and Geldner in Vedische Studien 3 and Oldenberg in ZDMG 54.
188 Rgveda-Samhitā. Sāntabalekarakulajena Dāmodarabhattasūnunā Śrı̄pādaśarmana sampāditā.
. . .. . .
Aundh sam . 1994. The details are given in the “Prastāvah.”, p. 3ff. I am grateful to Shrikant Bahulkar
for pointing me to this edition. The name of the author is—without Sanskritisation—spelled
Sātaval.ekara.
Sam . śodhakas and Adhyāpakas. We can therefore assume that the oral tradi-
tion available at the time was duly considered.
Under the heading “Ādarśa-granthāh.” the editor lists his sources. He
starts with the edition of Max Müller, which he says was printed very correctly
and free of faults.189 He makes this text the basis of his edition.190 Next
follow “some manuscripts” from a variety of places, but there is no listing of
individual manuscripts and the number itself remains unclear.191 The editor
then says that he first took the edition of moks.amūlarabhat.t.a, to which so
many scholars had contributed, as the final word on the constitution of the
text (antimapramān.atvena svı̄kr.tam), but in the course of editing several
errors and misprints became obvious.192
The editor then criticizes two misreadings of Müller: Firstly, he writes
consistently—there are six passages—syandra for the correct spandra. Fur-
thermore, he writes mathnā for mathrā in 1.181.5. The editor says that Bloom-
field in his Vedic Concordance as well as Wilson in his Sanskrit-English Dic-
tionary have this form of the word and expresses his confusion, but then
accepts what all his other sources—oral and handwritten—transmitted.193
Sāntabalekar devotes considerable space in his brief introduction to this error
of Max Müller and cannot enough wonder how this could happen.194 But
Formulated in this manner it does not address the question. There may not
be variant readings that “vaidikas” would take seriously, but there may still
be variants. These we may assume were again not worth reporting.
For the unsuspecting reader it must be said that in recent years the in-
tellectual climate for an academic study of the Veda in India has changed.
Studies or editions now mainly serve religious purposes, whereas an academic
context is only adduced, when it serves the religious one. In other words, the
perspective of many sections of Vedic Studies have, fostered by the current
political climate, changed into a kind of fundamentalist theology. As an exam-
ple I shall adduce one publication198 by the organisation for Vedic Studies that
is directly financed and controlled by the central Indian government, founded
as Veda Vidya Pratishthan “to preserve and protect the Vedic tradition”.199
The perspective of these publications has tangibly changed from that of many
earlier Vedic Studies produced in India, which were more academic in out-
look. In them a religious view of these texts as a valid revelation is mixed with
195 In 7.44.3 mam 196 iti pātho bahusammatah. Op. cit., p. 7. It is not clear
. ś versus mām
. ś. . .
whether this means most editions, manuscripts, or Vedic scholars. 197 Op. cit., p. 7. 198 The
R.gveda Saṁhitā of Śāṁkhyāyana-Śākhā. Ed. Amal Dhari Singh Gautam. Ujjain 2012. 199 Op.
cit., Preface, p. 1.
pseudo-scientific statements that have recently become extremely popular.
Let me give one example from this publication:200
Vedas are the earliest literary treasure of mankind. These are not
composed but have been realized by seers in penance. So these
are authentic, free from all blemishes and thus are the treasure-
house of all true knowledge as also have been eulogized by foreign
scholars.
They are the oldest of books in the library of mankind.
The oldest literary monument of the Indo-European world.
The same can be read in the Sanskrit introduction, but with a nationalistic un-
dertone: vedānām . kāran.ād eva asmākām
. bhāratı̄yā sam
. skr.tir asti viśvavārā
201
viśvavandyā viśvavaren.yā. In other words, the perspective is theological
and nationalistic. In order to underline the universalistic claim, distorted
pieces from foreign scholars, usually only Max Müller, are adduced, but
merely as proof that even the West has accepted the doctrines of Mı̄mām . sā,
the apaurus.eyatva and svatah.prāmān.ya of the Veda. Sound academic sources,
whether Indian or non-Indian, are not, it seems, on the reading list. This new
religious nationalism has of course been analysed and I shall not deal with
it any further, but for our text-critical problem the question is whether it is
likely that in this intellectual climate research scholars will be particularly
open to the discovery of actual variants in manuscripts?
In the publication mentioned above unsurprisingly the same picture pre-
vails. While it is based supposedly on a large number of manuscripts the
editor says: mantres.u pāt.habhedas tu na prāpyate.202 It is difficult for an
outsider to Vedic studies like the present writer, without having seen many
manuscripts of the text, to make sense of this situation. At least the three
variants mentioned by Sāntabalekar—two misreadings of Müller and one true
variant—should have been worth mentioning.
But the picture given by the Ujjain edition is the following: (1) syandra,
supposed to be a misreading of Max Müller, occurs in this edition in 1.80.9.
However, it reads spandra in 5.52.03, 5.52,8 and 5.87.3. In two instances, 6.12.5
200 Op. cit., Preface, p. 1. 201 Op. cit., p. v. 202 Op. cit., p. xi.
and 10.42.5, it reads syandra in the text, and spandra in the Padapāt.ha. (2) The
case of mathnā is different, since that reading of Müller was already changed
to mathrā in the edition of Aufrecht.203 In 8.46.23 the Śāṅkhāyana edition
reads mathrā, as does Aufrecht. In 1.181.5, where Aufrecht and Sāntabalekar
read mathrā, it reads manthā in the text, but mathnā in the Padapāt.ha.
We cannot leave the discussion here without briefly mentioning that Euro-
pean Vedicists had already dealt with these and other misprints of the editio
princeps as well as with insecurities of readings. Sāntabalekar mentions just
two misprints of Müller, whereas Aufrecht, who had access to merely a few
manuscripts from the Chambers collection, identified a longer list of the blun-
ders of “Herr Müller”.204 The polemical tone is not untypical in these days of
the controversy on the interpretation of the Veda, and I shall give just a few
passages in translation:
Among Vedicists the merit of individual readings had been already discussed,
as for instance: “According to my assessment the inner probability as well as
VIII, 46, 23 is strongly in favour of mathrā´ ”.205
Other errors or misreadings of Max Müller have escaped the attention of
later editors. For instance, in 2.11.10 Müller reads nijūrvı̄t. Aufrecht says that
all manuscripts and the commentator read nijūrvāt. The Ujjain edition reads
nijūrvı̄t. So it seems that we do not even have to look at other manuscripts
to get the impression that the notion of a variant-free Veda is not to be taken
literally. Vedic criticism is far from being able to provide a text consistent
203 Theodor Aufrecht: Die Hymnen des R.igveda. 2 vols. Bonn 1877. 204 Op. cit., vol. 1,
p. 4f. 205 “M. E. spricht innere Wahrscheinlichkeit und VIII, 46, 23 entschieden für mathrā´ .”
Oldenberg: R.gveda. Textkritische und exegetische Noten, p. 180.
enough to warrant the claim that there are no variants. The “Vaidikas” may
agree that there is no pāt.habheda, but we know since Scheftelowitz that there
are manuscript readings. So the common statement may mean no more than
there is no accepted variance. But because of the dogma that there are no
variants, it seems none could be reported and the claim, as in the Ujjain
edition of the R.gveda that is supposedly based on 25 manuscripts, cannot be
taken at face value. It seems that textual criticism of the Veda has been put to
rest, before it could even start.
The a priori conviction of the uselessness of all variants is a strange type
of eliminatio of all sorts of known and unknown manuscripts, and the dec-
laration that conjectures based on the Devanāgarı̄ manuscripts are superior
to other readings is even worse than the text-critical blunder Bentley is of-
ten charged with, namely to blur the distinction between conjectures and
transmitted readings. From the perspective of normal textual criticism it is
as clear a case of editorial madness as those described by the famous Cam-
bridge classical scholar Housman in his article “The application of thought
to textual criticism”.206 There he makes the point that professed textual critics
sometimes stop using their common sense and their ability to think, but
follow blindly maxims and principles that do not stand to reason. One of
the examples given by Housman is the case of an Italian scholar, whom he
calls simply “Mr. Vallauri” or even merely “one Vallauri”, I suppose it was
the nineteenth-century classicist Tommaso Vallauri from Torino. Vallauri
noticed that one of his theories was contradicted by a manuscript reading, but
instead of remodelling his theory to fit the new evidence, he chose to ignore
it altogether by stating that the manuscript was not trustworthy, since it was
generally in a bad condition. This by the way referred only to its physical
condition, the readings where they were legible were not in doubt, but Val-
lauri did not want this evidence to disturb his theory and thus decided that a
manuscript so much “tattered and battered” did not deserve attention. The ar-
gumentation is not worth of serious academic consideration; old manuscripts
tend not to look as neat as new ones and one would not stop quoting from a
book, just because one has spilt one’s morning coffee on it in a hurry. And
206 “The Application of Thought to Textual Criticism”. In: The Classical Papers of A. E. Housman.
Collected and Edited by J. Diggle and F. R. D. Goodyear. Cambridge: University Press 1972,
p. 1061.
this is Housman’s point: we do not need that much theory, just common sense
to see that Vallauri was wrong.
My reason for referring to Housman is not because of one singular exam-
ple, but because of Housman’s résumé. For he claims that: “If you engage in
textual criticism, you may come upon a second Mr Vallauri at any turn.”207
In my humble opinion Oldenberg can claim the same rank. He speaks of
the persistent inferiority of the Kashmirian manuscript, which leads him to
discard its evidence completely—clearly the error of the “best” versus the
“bad” manuscript.
Unfortunately we are not so well-informed about Indian reactions to the
edition of the R.gveda by Max Müller. According to one author there were
resentments in Mahārās.t.ra. Müller himself reports that “attempts were made
in various quarters to taboo it, as having been printed by a Mleccha [impure
foreigner or barbarian] and with cow’s blood.”208 In Pune it is reported that
Brahmins still wanted to know the edition and had it read out by a non-
Brahmin, so that no one had to touch it. It is unclear what to make of these
anecdotes. But if another report should be true, that later on Pandits corrected
their private manuscripts of the R.gveda with the help of Müller’s edition, then
this is hardly short of a text-critical fiasco.209
And surely this is not the only example. If we look too closely at some
widely used editions, we might be surprised. There is, for instance, another
important early genre in Indian literature, where we find a mismatch between
public interest, translations and reprints on the one side and critical studies
of the sources on the other: The text of the Upanis.ads we read is still that of
vulgate editions produced from very few manuscripts. Here too the rule is
confirmed that reading text-critical introductions is not to be advised, since
it raises too many questions about the quality of the text presented. Take for
instance the remark in the well-known edition of Limaye and Vadekar:210
207 Op. cit., p. 1062. 208 Max Müller: Auld Lang Syne. Second Series. My Indian Friends.
New York 1899, p. 25. 209 Walter Slaje just alerted me to Johannes Bronkhorst: How the
Brahmins Won (Leiden 2016), where the author does indeed use Scheftelowitz’ edition (p. 450),
assigns it to a different śākhā. 210 Eighteen Principal Upanis.ads. Vol. 1. Poona 1958, p. vi.
corrections—which in its turn follows the text of the Śāṅkara-
bhās.ya, being the oldest available; we have also noted variants
in the other Bhās.yas whenever important; but this treatment
cannot be said to be complete. We consulted old MSS of the
Br.hadāran.yaka and the Śvetāśvatara but our collation was not
found to be useful: in many cases we found the MSS more faulty
than the printed editions [. . .] So we gave up the attempt.
212 The articles are now easily accessible in: Wilhelm Rau. Kleine Schriften. Hrsg. Konrad Klaus
und Joachim Friedrich Sprockhoff. 2 vols. Wiesbaden 2012, p. 1313–1328. The passage I sum-
marize is: “Śaṅkara las nicht anders als die gesamte indische Tradition, von der wir Kunde haben,
die Upanis.ads in fehlerhaften Handschriften und konnte sich bei ihrer sprachlichen Erklärung
von keinem verlässlichen Lehrer Rat holen.” (p. 1327) 213 Patrick Olivelle: Unfaithful
Transmitters. Philological Criticism and Critical Editions of the Upanis.ads. In: Journal of Indian
Philosophy 26 (1998): 173–187. 214 “[. . .] der Überlieferung durch Konjekturen aufzuhelfen. In
letzter Beziehung hat er nicht selten über das Ziel geschossen, indem er nicht die Überlieferung,
sondern den Autor verbesserte.” Delbrück: “Otto Böhtlingk”. In: Anzeiger für indogermanische
Sprach- und Altertumskunde. Beiblatt zu den Indogermanischen Forschungen. 17 (1904), p. 136.
215 Richard Hauschild: Die Śvetāśvatara-Upanisad. Leipzig 1927. 216 Thomas Oberlies:
.
»Die Śvetāśvatara-Upanisad«. In: WZKS 39 (1995), p. 61–65.
anyone had ever collated for the first time a number of the actual manuscripts
of this Upanis.ad?
One could go on. Even from my own limited experience it appears that
reading manuscripts often means opening a can of worms. The result is
sometimes catastrophic—in the sense that the text does not actually contain
what we thought, but something else. For practising textual critics it would
be a great pleasure to read a printed text without wondering at every unclear
point, whether this is really the correct reading. Who wants to ponder on
passages and put all ingenuity of interpretation into a phrase, when a first
glance at a manuscript of the text shows that the editor made a stupid mistake.
This for me is the great disadvantage of textual criticism: it bereaves one of
some of the joy of just reading printed texts.
217 Oskar von Hinüber: “Der vernachlässigte Wortlaut”. In: Kurt Gärtner (ed.): Zur
Überlieferung, Kritik und Edition alter und neuerer Texte. Mainz: Steiner 2000, p. 17f. The article
lists conveniently the indological literature on textual criticism.
editions. Lachmann had like August Wilhelm Schlegel, the first indological
editor, studied in Göttingen with Heyne, but much later. Before Lachmann
had developed his “method”, indologists had published quite a few critical
editions of Sanskrit works. All of these are criticial in the sense that they are
based on a number of manuscripts that were fully collated—not just on one
single copy—and that they use text-critical reason. And most are in some way
or another connected to the school of Schlegel in Bonn, in some cases also to
Bopp in Berlin. Most travelled to London and Paris in search of manuscripts,
thereby renewing the contacts between German, French, and British Indology
that August Wilhelm Schlegel had inaugurated. From the introductions one
gets the impression that this was a group of pioneering indological editors,
connected by various students’ friendships. In the introductions the fellow
students are often mentioned, and it seems that Stenzler, Lenz, Bohlen, Lassen,
Brockhaus, Tullberg, and Böhtlingk formed a close group. The following is a
list of these early editions, not counting the more well-known editions of the
Bhagavadgı̄tā, the Hitopadeśa, and the Rāmāyan.a by Schlegel himself:
218 It is in such cases sometimes possible to reconstruct the text commented upon even when the
text itself is not transmitted separately, just from the passages quoted in the commentary.
mixed, and sometimes the other readings were marked as “ks.epakah.
‘distichon rejiciendum’ ” or “pāt.hāntaram
. ‘alia lectio’ ” (p. iv). This
is certainly what we expect: Commentators were obviously aware of
variant readings, had presumably collected manuscripts and produced
their own edition, where they ruled which verse was original and which
one had to be rejected as spurious.
Stenzler was however aware that he should not produce from the mass
of variants a new recension, but stick to one that presented the best text
closely. After comparing several aspects of these, he settled more or less
on the text of Mallinātha.219
219 “Ad textum ex his subsidiis constituendum, aut alterutram, sive Mallināthae sive reliquorum
duorum commentatorum recensionenm religiose sequi debebam, aut ex ampla ista lectionum
varietate, quam codices suppeditabant, nova aliqua recensio erat concinnanda. Hoc ne facerem,
deterrebat me rei ambiguitas; quare cum utriusque recensionis, scholiis stabilitatae, comparatione
accurata instituta, nullum mihi dubium relinqueretur, quin ea, in quam Mallināthas commen-
tarium suum concinnavit, altera esset antiquior, carminis nostri textum ex ejus commentario
emendare potius duxi.” Op. cit., p. iv.
3. Christianus Lassen: Gymnosophista sive Indicae Philosophiae Docu-
menta. Bonn 1832. This edition of the Sām
. khyakārikā is based on three
manuscripts, the variants are discussed in the running commentary.
220 “Fateor, me summopere miratum, quum primum Gı̄tagovindam perlegissem, tam pauca
reconditoris sensus vestigia in carmine repperiri, [. . .]”, p. xiii. 221 “Tenet ea testum criti-
cis praesidiis omnino nudum [. . .]”, p. xxxv. 222 “Congerendam semper esse totam variae
scripturae nubem ex libris reor [. . .]” Op. cit., p. xxxv. 223 Adolf Friedrich Stenzler:
These are not even all critical editions published before 1850 by authors con-
nected to the “Bonn school”,226 and we shall return to this date in the next
chapter in connection with Lachmann. What we can see from this survey is
the following: for typographical reasons readings could not easily be printed
at the bottom of the page. There was a smaller Devanāgarı̄ typeface, but
possibly not in sufficient quantities. Usually variants were printed either in
an appendix or discussed in the commentary. Often readings were selectively
discussed in the critical notes, only few authors list a sizeable number of vari-
ant readings. The authors seem to have tried to utilize as many manuscripts
as they could get in Europe. Only when all manuscripts failed, that is, yielded
an unsatisfactory text, one resorted to “Konjekturalkritik”, even though, as in
the case of Brockhaus, this approach was sometimes rejected. The practice to
conjecture cannot have been what Timpanaro has described as emendatio ope
codicum, that is, a practice to use one manuscript and insert better readings
from other manuscripts whenever necessary. At least in the case of Schlegel,
we know from his instructions to Lassen in his letters that the basis of the
text was a complete collation of all manuscripts. Since this was the method
practised by the two first indologists in Bonn, it is fair to assume that their stu-
dents worked in the same way. We may therefore not exactly know about the
225 Jacob S. Speyer: Studies about the Kathāsaritsāgara. Amsterdam 1908. 226 We could add
other editions produced in the same period: Böhtlingk’s edition of the As.t.ādhyāyı̄ (Bonn 1839),
his Śakuntalā (Bonn 1842), or Stenzler’s edition of the Mr.cchakat.ikā (Bonn 1847). Then there are
lesser known ones, as Bernary’s edition of the Devı̄māhātmya (1830).
theory of editing in early Indology—the introductions are mostly too short
and too practical to permit a reconstruction of the theoretical background—
but we know that there was a practical school of editing. And if this much
guess-work may be allowed here: it was—in a time when reading classes had
to use manuscripts, because there were not enough books—probably taught
in the same way, as translating from Sanskrit is taught in Indology now: as
an established practice with its own history, but without much theory to fall
back on. The introductions provide sufficient examples of a keen awareness
of text-critical problems specific to the transmisson of Sanskrit texts.
Lachm ann
227 “[. . .] die jetzt im Druck vollendete, wenn auch noch nicht ins Publikum gekommene Aus-
gabe des Properz”. Martin Herz: Karl Lachmann. Eine Biographie. Berlin 1851, p. 34. 228 Sex.
Aurelii Propertii Carmina emendavit ad codicum meliorum fidem et annotavit Carolus Lach-
mannus. Lipsiae 1816. 229 “Lachmann hat keine feste Theorie seiner Methode hinterlassen,
obwohl sie bis in unsere Tage als solche immer wieder beschworen wird. Man muß sie aus
den einzelnen Hinweisen und Ansätzen zusammentragen, zumal auch seine Editionen äußerst
auskunftsfaul gewesen sind. Die beste Quelle ist eine ausführliche Rezension aus dem Jahre 1817
in der Jena’schen Literaturzeitung [. . .]” Hans-Gert Roloff: “Karl Lachmann, seine Methode
und die Folgen”. In: Hans Gert Roloff (ed.): Geschichte der Editionsverfahren vom Altertum
bis zur Gegenwart im Überblick. Berlin 2003, p. 66.
to Roloff, is the review of van der Hagen’s edition of the Nibelungenlied,
to which Witzel also refers: “Lachmann had established the rules of textual
criticism already in his Habilitationsschrift (1816), and in his early review of
Hagen’s Nibelungen and Benecke’s Bonerius, contributed in 1817.”230
Lachmann then describes his own approach. First he wants to establish his
text with the help of a sufficient number of good manuscripts and arrive at the
text that is the basis of all of them, a text, which is either already the original
text or at least one very near to it. Roloff adds that “unfortunately Lach-
mann has never sufficiently explained, what constitutes a good manuscript
[. . .] From his practical decisions we can gather that it is mostly the oldest
manuscript, but that according to his decisions also younger manuscripts can
hold the better readings.”232
Up to this point Lachmann’s approach is fairly clear: One should not print
one manuscript with occassional improvements, but systematically search for
the oldest state of the text, we can arrive at with the help of all sources. In his
230 Witzel: Textual Criticism, p. 19. 231 The review is reprinted in Kleinere Schriften zur
deutschen Philologie von Karl Lachmann. Ed. Karl Müllenhoff. Berlin 1876, p. 81ff. The trans-
lation is mine. 232 “Leider hat Lachmann nie konkret verdeutlicht, was eine gute Handschrift
ist [. . .] Aus der Praxis seiner Entscheidungen läßt sich entnehmen, daß das zwar meistens die
ältesten Handschriften sind, aber daß durchaus nach seiner Entscheidung auch jüngere Hand-
schriften die besseren Lesarten haben können.” Roloff: “Karl Lachmann, seine Methode und
die Folgen”, p. 67.
review Lachmann now suddenly introduces “simple rules” for establishing
the text of the Nibelungenlied:233
The paradoxical situation here is that we may see such rules as the outcome
of stemmatic considerations, only there is not even an attempt to explain the
rationale for these rules. One needs considerable hindsight bias to see the
start of the stemmatic method here.
Furthermore, the type of stemmatics we can infer from Lachmann’s treat-
ment is hardly satisfactory. We have in our transmission a corrector, as
Lachmann says, whose activity has spread unevenly to the following genera-
tions of manuscripts. In other words, the transmission is contaminated, but
the establishment of the text is still put down to simple rules. The first of these
is merely numerical, which cannot work, if conscious modification is part
of the transmission. The second rule cannot be the outcome of stemmatic
233 “Die weitere Untersuchung, die wir jedoch hier nicht ausführen können, ergibt nämlich, dass
die übrigen Handschriften, die erwähnte Umarbeitung E und die jüngere münchner (M), eben wie
G, aus einem Exemplare, das B sehr ähnlich war, geflossen sind, alle drey aber nicht unmittelbar,
und dass diese Urschrift der drey genannten nicht eine ganz neue gewesen, sondern eine alte,
welcher der Verbesserer seine Änderungen beygeschrieben hatte. Diese Änderungen, welche bald
dieser, bald jener Schreiber übersehen, und jeder mit neuen vermehrt hat, herauszufinden, das
ist die Aufgabe des Herausgebers. Die Gesetze sind, so viel wir gefunden haben, folgende: 1) Drey
Handschriften unter unseren vieren überstimmen alle Mal eine. 2) Wo je zwei überein stimmen,
ist BG < EM [. . .]” Kleinere Schriften zur deutschen Philologie von Karl Lachmann, p. 87.
considerations, since Lachmann puts B near to the source of E, M, and G. If
these are the two branches of our imagined stemma then the rule that E and
M, where they agree, are stronger than B and G, is not at all what we expect
from stemmatics, rather than any agreement, even of a single manuscript with
B would have to be preferred.
In view of these facts it is problematic to speak of a method of Lachmann at
all, and even more so to date it to his early works. Timpanaro and Schmidt234
come to the conclusion that his method was only formulated in his latest
works.
It is also interesting to note that in his earlier works Lachmann never had
to collate himself, for he relied on other’s apparatuses, and for some Middle
German editions he in fact used the best manuscript method. Schmidt main-
tains that despite all claims to the contrary, he had no patience for collation or
complicated transmissional scenarios, and quotes him as saying: “Through
theoretical doubts the work that is already difficult, is only complicated fur-
ther.”235 It seems he was particularly gifted and lucky in making most of
very few manuscripts. But neither did his conjectures make it into moder-
nity, nor did his analyses of the transmission of many texts stand the test of
time. What is interesting is how students of Lachmann understood and used
his method. Schmidt talks of an “imitation of Lachmann’s lack of method
in choosing as quickly as possible a ‘best’ manuscript and even renouncing
earlier, broader attempts at a stemmatic reconstruction.”236 In this sense, but
not in any later sense, Max Müller in his edition of the R.gveda was, as he
himself says, following Lachmann! According to Schmidt, the connection of
Lachmann with the stemmatic method is a back-projection. And finally the
idea of a “méthode de Lachmann” occurs first in the writings of his French
adversary Bédier, who has rightly critized the arbitrariness in constructing
stemmas. But actually Lachmann’s method of excluding manuscripts as apo-
graph or contaminated and practically using only very few or even one source
is very near to Bédier’s method, and this discrepancy between what was later
understood as the method of Lachmann and what Lachmann had practised
can only be described as a historical misunderstanding.
234 P. L. Schmidt: “Lachmann’s Method: On the History of a Misunderstanding”. In: The Uses of
Greek and Latin. London 1988, p. 227–236. 235 “Durch theoretische Zweifel verwirrt man die
an sich schwierige Arbeit nur noch mehr”. Op. cit., p. 231. 236 Op. cit., p. 233.
Indological Solutions
If one reviews recent editions of Sanskrit texts, one notices a wide spectrum of
approaches to textual criticism. There is firstly the meticulous application of
stemmatics. Editors know that a critical edition needs a stemma, so it seems
they produce one at all costs, that is, even where it has no effect on the editing
process. Bronkhorst, for instance, has shown that the stemma given in the
edition of the Yuktidı̄pikā237 has serious logical defects.238 The summary of
his deliberations is revealing. After correcting the stemma, he says: “What
difference would this modified stemma make to the edited text. Very little, of
course.”239 In other words, the stemma is erected in the introduction like the
sign post of a critical edition, but the effect on the edition itself seems less
pronounced.
An edition of a Sanskrit text according to the stemmatic method is that
of the Vākyapadı̄ya by Wilhelm Rau.240 It has taken into account the more
than twenty manuscripts known at the time. In a long process Rau produced
a stemma of all manuscripts which is characterized by a division of the whole
transmission into a Southern and a Northern branch. The quality of the
Southern branch Rau considers as higher and bases his edition on it. The
Northern transmission he uses only in cases of doubt.
There is no reason to question this decision, but we should note that here
the stemma is used in a function quite different from that in Maas’ handbook
and thereby shows one of the practical complications involved. Rau’s stemma
has two fundamentally unequal branches, the Southern transmission, which
is characterized by a higher quality of the text, and a Northern generally
more faulty one. Apparently the discrepancy is so clear that he formulates a
general rule that he will follow the Southern transmission wherever possible.
The stemma is not used as a tool to find out the reading of the archetype
by following the mechanical rules laid down by Maas, but has only a minor
function in establishing the text. Rau states that he refrained from construing
a stemma from variant readings, for this—as he says—does not go beyond
p. 242ff. 239 Op. cit., p. 244. 240 Bhartr.haris Vākyapadı̄ya. Wiesbaden 1977.
subjective impressions;241 his stemma is based solely on lacunae. In this
respect it is an exemplarily sober edition.
But strangely Rau’s edition gives only a thin selection of readings,242 so
that the reader cannot gain an impression of the manuscripts’ quality. More
importantly any further work on the text is thereby severly impeded, since as
soon as new manuscripts would appear, a new complete collation would be
necessary. Perhaps this is the reason why Witzel did not grant Rau’s edition a
place in his A-list.
I adduce this example, because this is one of vexing practical questions
any editor of a Sanskrit text has to address. Obviously it is the task of the
editor to present the complete evidence contained in the manuscripts, but
should this include really everything? First, the magisterial answer from our
favourite handbook:
Editors who are not very familiar with the behaviour of an-
cient and medieval scribes often report many trivial mistakes
of spelling. These are not valuable unless one is making a sur-
vey of scribal habits, which is an important but highly special-
ized branch of study, not part of the brief of the average editor.
Sometimes, however, an editor may feel justified in adopting a
compromise position: he will perhaps come to the conclusion
that if there is one manuscript of much greater importance than
any other single witness to the text, even the minor errors of this
manuscript should be recorded.243
The formulation is flawless, but for editorial practice some more hints are
needed:
241 Wilhelm Rau: “Über sechs Handschriften des Vākyapadı̄ya”. In: Oriens, 15 (1962),
p. 376. 242 He says in the introduction that some readers would have probably wanted less.
243 Reynolds and Wilson: Scribes and Scholars, p. 239.
that of texts in nineteenth-century German Kurrent script. An edition
depends in a high degree on the amount of training an editor has in the
relevant script and on his ability to make sense of the text while deci-
phering it. We shall see below from examples that this is an important
cause for erroneous editing.
What this means for the practice of reporting variants is this: if we
are sure that an ambiguity disappears as soon as we understand the
text, then there is no point in reporting it. For instance, scribes write
the Śāradā letters ma and sa as almost indistinguishable. If in a given
passage we know that the word in the context will have to be samānı̄ya
we need not report that the first sa looks perhaps more like a ma.
• Some manuscripts are so faulty or irregular in their placement of, for in-
stance, visargas or anusvāras, that one might opt for not reporting these
errors. This should be mentioned in the introduction. The reasoning is
that neither will these errors contribute in any way to the establishment
of the text, nor to the construction of a stemma.244
• The great danger with the non-reporting of seemingly trivial errors
is that one has to keep the eyes open for possibly original features, as
for instance regional variations of Sanskrit. An editor, who knows that
Kashmirian scribes seem particularly prone to interchanging vovel com-
binations as makura - mukura, might want not to report such errors and
just quote the correct form. But this would mean that regional spellings
could be lost. Walter Slaje has provided me with the following evidence
of one “Kashmirian” aorist form aśiśrayat, which he has accepted as
the critical text in his edition of Jonarāja’s Rājataraṅgin.ı̄ 322.245 The
manuscript evidence there oscillates between this and the correct(ed)
form aśiśriyat. But in Kalhan.a’s Rājataraṅgin.ı̄ there are cases where we
have aśiśriyat without variants,246 cases where the manuscripts have
one or the other,247 but also passages where we find aśiśrayat as the only
reading.248 Here the rule of the lectio difficilior should alert us to the
244 Naturally such individual errors cannot be regarded as connecting errors. 245 See his
Kingship in Kashmir. From the pen of Jonarāja [. . .] Halle 2014. 246 3.216. 4.49. 8.458; 544; 860;
864; 923; 938; 1431. 247 4.148. 7.762;770. 8.106; 927; 1438; 1485. 248 5.132. 7.560.
most likely genesis of error, which is the correction of the non-standard
form: aśiśrayat → aśiśriyat. In other words, the un-Pān.inian form
seems to be the original, Kashmirian variant, gradually weeded out
from the tradition by correctors. The fault of an unsuspecting editor
would not be to relegate this reading to the footnotes—without exten-
sive collection of additional materials he could not dare to postulate an
“ungrammatical” form as the original—but not to report it at all as a
trivial mistake.249
• A further example: In editions of Kashmirian Sanskrit texts I was puz-
zled to find the reading vis.ada for viśada. According to the Böhtlingk
dictionary it is simply “erroneous”, a mistake of spelling facilitated
by pronunciation.250 Of course, but a fairly persistent one; in many
manuscripts of the Moks.opāya it is well represented, but why accept
vis.ada, when other manuscripts read the correct viśada? Only recently
I found a compelling argument for elevating vis.ada to the status of a
true variant spelling:
Coming back to the edition of Rau, we have to add one more important point
of criticism. Rau wants to reconstruct the archetype of the kārikā-manuscripts
and comes to the conclusion that it was already deficient,252 whereas some
249 For the phenomenon of regional Sanskrit, see Madhav Deshpande: “On Vernacular San-
skrit”. In: Sanskrit & Prakrit. Sociolinguistic Issues. Delhi 1993, p. 33ff. 250 Monier-Williams
does not even record it. 251 Here in a feminine bahuvrı̄hi. 252 Bhartrharis Vākyapadı̄ya,
.
p. 24.
versions with commentary—which are excluded as sources for the edition—
according to Rau preserve original readings. Thus he considers his edition
merely as one step on the way to a “final edition of the Vākyapadı̄ya”,253 which
would have to gain the original work with the help of the Kārikā-Text, the
commentator’s version and the testimonia. Two things are puzzling here: Why
does Rau, if he considers his edition preliminary and expects others to carry
on the work, give only few readings, thereby making it extremely difficult
for others to build upon his edition? And secondly, what are the arguments
for keeping the transmission of the kārikās and the commentated version
separate?
In other compartments of Indology we find a declaration that stemmatics
is in view of the contamination of many transmissions often impossible.254
This diagnosis has been arrived at frequently and in view of the activity of pre-
modern Indian scholars transmitting and editing Sanskrit works as described
above this scenario is so likely that the burden of proof—usually one expects
the editor to prove that the transmission is contaminated in order to absolve
him or her from producing a stemma—should be reversed.
As a reaction to this some Sanskrit editors have adopted an entirely differ-
ent type of classical textual criticism to deal with contaminated recensions,255
the “Kontaminationskritik” as developed by Pasquali and others. This method
may appear to be a roll-back before the time of Lachmann: from objective,
stemmatic editing back to subjective decisions. But actually this impression is
created only by the undue propaganda of Lachmann. As we have seen, even
stemmatic editing has to depend very often on decisions according to other,
non-stemmatic criteria. This fact is easily overlooked by theorists and only
hesitantly disclosed by practitioners, because here the scientific approach of
stemmatics comes to an abrupt halt. Unless the editor silently follows a “best”
manuscript, he has to decide according to the so-called inner criteria, which
Lachmann wanted to avoid. Fränkel has indicated that in practice this admis-
sion is avoided through the dirty tricks of the trade: Editors tend to produce
a stemma, but silently follow a Leithandschrift in order to avoid having to
253 “[. . .] endgültigen Edition des Vākyapadı̄ya”. Op. cit. p. 37. 254 See the quotation of
Coulson above, p. 74. 255 See Jürgen Hanneder: “Introduction”. In: Jürgen Hanneder
and Philipp A. Maas (eds.): Text Genealogy, Textual Criticism and Editorial Technique. [=WZKS
LII–LIII (2009–2010)], p. 12.
decide between readings. Fränkel, Pasquali and others therefore argue for
a return of the inner criteria, which were forgotten, or rather pushed out of
sight, through the claim of the stemmatic method to be more scientific.
The rationale of “Kontaminationskritik” can be explained by starting
from Maas, who says that, if contamination is diagnosed, stemmatic exclusion
techniques cannot work, because every variant becomes a presumptive variant.
This means every contaminated manuscript can alone, that is, against all
other witnesses, preserve the original reading. As a result no manuscript
can be excluded as an apograph, singular readings cannot be excluded, the
whole instrumentarium of simplifying the editing process cannot operate
any more. This observation is quite old. The first indological editor, August
Wilhelm Schlegel, will have read it in a preface by his colleague in Jena: “Lectio
paucorum codicum, imo vnius codicis, genuina esse potest.”256 Thus every
reading has to be judged for its merits, and criteria have to be found in order
to avoid that this process is simply one of individual taste. It may for some be
an unpleasant realisation that under these conditions we are suddenly much
nearer to the practitioners of a pre-modern ars critica, be it European or
South-Asian. Phrased in a more theoretical way, we are producing an eclectic
text, a confession that has, as one would expect, led to extensive discussions,257
but again not at all new:258
256 Libri Historici Novi Testamenti Graece. Pars prior [...] Emendavit et lectionis varietatem
adiecit Io. Iac. Griesbach. Halae apud Io. Iac. Curt. 1774, Praefatio, p. xv. 257 Eldon Jay
Epp: “The Eclectic Method in New Testament Textual Criticism: Solution or Symptom?”. In:
Harvard Theological Review 69 (1976), p. 211–257. 258 Paul de Lagarde: Anmerkungen zur
Griechischen Übersetzung der Proverbien. Leipzig: Brockhaus 1863, p. 3. 259 “die manuscripte
der griechischen übersetzung des alten testaments sind alle entweder unmittelbar oder mittelbar
das resultat eines eklektischen verfahrens: darum muß, wer den echten text wiederfinden will,
ebenfalls eklektiker sein.”
is as good as to not often enough present bad variants, and none
so bad as to not present occasionally a good “grain”.260
This sums up succinctly the dilemma of the editor, and suggested solutions
are many. There are now advocates for rigorous (computer-)stemmatics
on one hand of the spectrum and those who advocate a “thoroughgoing
eclecticism” where “the cult of the best manuscript gives way to the cult of the
best reading”.261
For the latter the most detailed treatment of such a method for Sanskrit
literature is found in Srinivasan’s introduction to his edition of the Sām . khya-
262
tattvakaumudı̄, a work without which a treatment of indological textual
criticism is I think not complete.263 What Srinivasan has done was to reformu-
late the once degraded internal criteria, as the lectio difficilior, usus scribendi,
etc.264 with regard to Sanskrit texts. The following is a brief summary.265
Srinivasan starts by looking at individual constellation of readings. If there
is no variance, the text is in principle accepted as “primary” (§ 1.4.1), only
occasionally the primary reading was gained through emendation (§ 1.4.2).
Herein the different schools of editing will agree, but in case of variance there
is one general principle, on which most others depend:
260 “so gut ist, daß sie nicht oft genug schlechte lesarten, keine so schlecht, daß sie nicht mitunter
ein gutes körnchen böte.” 261 J. K. Elliot: “Rational criticism and the text of the New
Testament.” (Theology 75 (1972), p. 340f.) as quoted in Eldon Jay Epp: “The Eclectic Method”,
p. 253. 262 Vācaspatimiśras Tattvakaumudı̄. Ein Beitrag zur Textkritik bei kontaminierter
Überlieferung. Hamburg 1967. 263 Witzel would probably not agree, since he does not mention
the work. 264 See Jürgen Hanneder: “Introduction”, p. 12. 265 The English rendering of
some rules was prepared long ago by Dominic Goodall and myself. 266 “Gefälle zur Korruptel
hin”.
a very lucid expression of the same idea, the old term may have been lectio
media, the middle reading that explains how the others developed. If we
search the history of textual criticism for early formulations of the principle,
we at least arrive at Erasmus:267
In many cases, Erasmus does not apply the principle of the harder
reading as such, but the commonsense principle that governs this
text-critical rule. He asks, so to speak, which one of two readings
was more likely to give rise to the other one. He then uses the
possible offence taken by a scribe as an indication. In this way, it
is not simply a criterion by means of which he decides between
two rival readings, but an effort to imagine the possible process
by which a text was altered [. . .]
It is not simply a matter of choosing the harder reading, but an
attempt to ascertain the most likely historical process. In this
case, an argument from style and context also comes into play.
The example of the harder reading shows that this genetic rule is really a
meta-rule of Kontaminationskritik, from which a host of others derive.
Many authors have used or formulated this principle, but the question
is rather, when to apply it. Sukthankar, to whom Srinivasan refers, tried “to
find a reading which best explains how the other readings may have arisen.”
But only when all other stemmatic methods failed. Srinivasan in his text
diagnoses total contamination and therefore starts with this method right
from the beginning and applies it to all cases. Srinivasan situates his method
within the broader field of textual criticism, referring to Pasquali, Dawe,
Kosambi, but singles out a statement of Vandelli, who in his work on the
criticism of Dante’s Divina Comedia made this principle “il più importante”,
the most important one.
One characteristic of a contaminated recension is often that there are in a
given passage variants that make almost equal sense, as if a long transmission
had already corrected all the inevitable scribal errors and distilled a selection
of plausible readings. We cannot reject one of them because of their position
in the wrong manuscript or on the wrong side of the stemma, and we cannot
267 Jan Krans: Beyond what is written. Leiden: Brill 2006, p. 41.
use the sense as the guiding criterion, because the transmitters may have
thought of that already. According to Srinivasan these readings are, however,
not equal as regards their textcritical value, and this is the chance for the critic
to arrive at the primary reading (§ 1.4.4).
In the following the author gives examples for two groups of non-original
readings: some secondary readings result from deliberate alteration, some are
accidental (§ 1.4.5). This classification is an important hint to the stemmatic,
an observation that questions the foundations of the model of Maas, according
to which every scribe just copied, but without intervention.
At this point we find what is perhaps the most problematic of Srinivasan’s
rules:
268 Again, not a new discovery. See Griesbach’s description: “Praeferatur lectio breuior, obscurior,
durior, sensum paradoxum aut apparenter falsum [. . .]”. Libri Historici Novi Testamenti Graece.
Pars prior, p. xv. 269 Here one might argue that this method can, as every other, arrive at
no more than the reading of the archetype, but not necessarily the one intended by the author.
270 For the following, see Alexis Sanderson: “The Śaiva Age”. In: Shingo Einoo: Genesis and
Now, the objection that a version which is less clear in this sense
must have preceded one that is freer of these defects, proceeds
from a serious misunderstanding of how the rule of the lectio
difficilior is to be applied. Firstly, like all other ‘rules’ of tex-
tual criticism, it should never be put to work mechanically and
in advance, without the application of thought to the weighing
of probabilities in each case; and secondly, it should never be
invoked to give precedence to readings that are grammatically
defective, incoherent, or contextually awkward.
271 This was stated in his article “Vajrayāna: Origin and Function” (in: Buddhism into the Year
2000. Bangkok 1994, p. 87–102) in a most unambiguous manner: “The present author’s view is
that almost everything concrete in this system is non-Buddhist in origin even though the whole is
entirely Buddhist in function.” (p. 92.) 272 Praud.ha Manoramā with Commentary Śābdaratna.
Critically edited by Venkatesh Laxman Joshi. Vol. 1. (Deccan College Monograph Series 31)
Poona 1966, p. 30.
It is perhaps even an unwanted side effect of the eclipse of the inner criteria
as formulated in the pre-Lachmann era, for instance by Wettstein, that some
editors became so focussed on the lectio difficilior that warnings against its
overuse are now frequent. As usual Reynolds and Wilson have found the
common-sensical formulation: “Many references to the maxim difficilior lectio
will be found in commentaries, and there is no doubt of its value. But it has
probably been overworked, for there is a temptation to use it as a defense of
anomalous syntax or usage; in such cases the more difficult reading may be
more difficult because it is wrong.”273
Then Srinivasan continues with a set of rules that describe the activity of
scribes. These rules quite obviously cannot be applied mechanically but only
with circumspection, and tailored to specific cases. They are techniques used
according to situation and thus contain inevitably a subjective moment. Here
the claim of objective editing has been or has to be given up and be replaced
by a training of taste. This is, however, also not a new insight, but described
already by Paul Maas in his treatment of the stemmatic method.274 The reason
for this is that the stemmatic method is far from being fundamentally more
objective, it merely separates a supposedly mechanical retrieval of the readings
of the archetype from the phase of conjecture. The following are examples for
further rules of Srinivasan:275
273 Scribes and Scholars, p. 221f. 274 “[. . .] in matters of style he alone is responsible, and
it must be his keenest endeavour throughout his life to perfect his feeling for style, even if he
realizes that one man’s lifetime is not long enough to allow a real mastery in this field to reach
maturity.” Textual Criticism, p. 10. 275 The translation is not very literal, since we tried to
Much could be said about these maxims. For instance, emendation may
be necessary in many cases, but where the manuscripts already supply a
wealth of readings from which to select, it is difficult to assess whether a
non-transmitted reading is really called for. The threshold here varies with
individual editors. Then there are those texts which display an irregular
Sanskrit, ranging from Epic to Buddhist “Hybrid”, and as a more recent
addition “Aiśa”, i.e. Śaiva Sanskrit. For the editor it is very difficult to say
whether these texts have suffered more through scribes adding more mistakes,
or through correctors who tried to move towards a standardized Sanskrit
thereby effacing the grammatical peculiarities. Here the editor is always in
danger of emending an irregular, but “original” form, or leaving an error by
the scribe in the text as a supposed feature of the language. As often in textual
criticism, there is no easy solution and perhaps none that would satisfy a
theoretical approach. One example would be a recent edition of three early
Pañcarātra texts from codices unici.276 There is a line in one of these texts that
seems to have a parallel in another:
These lines as found in the manuscripts are ungrammatical, and the question
is how much of it is original. Is ekamūrti just a scribal error, or is it a feature
of this language to use stem forms for various “cases”.277
In the first instance the editor emends to ekamūrtim . vijānı̄yā brahma-
vis.n.umaheśvaram . , in the second to ekā mūrtis trayo devā brahmavis.n.u-
maheśvarāh., which is understandable, but amounts to a rewriting of these
lines to yield the required sense. Since the texts are not translated, a third
theoretical option is not explored which would be to explain the grammatical
construction through a translation, but refrain from adding case endings.
Reading ekamūrti in both cases, and interpreting it variously as an accusative
or a nominative, is an option that would at least allow us to remain a little
nearer to the transmitted texts. For it is not the extremely high number of
emendations and conjectures in each and every verse of these texts that makes
the reader wonder,—these may be unavoidable when editing from a single
manuscript that is very faulty—but he may wonder whether the reconstruc-
tion of the sense has more or less effaced the original wording in “Tantric”
Sanskrit.278
The last rule of Srinivasan also shows the limitation of the method. In
some transmissions of Sanskrit texts we regularly find in a given passage
synonyms that work equally well in the context and cannot convincingly be
selected by internal criteria. Here a mechanical criterium, as for instance a
stemma produced by whatever method, in the worst case a best manuscript
276 Diwakar Acharya: Early Tantric Vais.n.avism. Three Newly Discovered Works of the Pañ-
carātra. Pondichéry, Hamburg 2015. 277 This would not be so unusual for a speaker of a
Sino-Tibetan language. 278 The editor treats the Sanskrit of these Pañcarātra scriptures as
parallel to that in Śaiva scriptural texts where it is called “Aiśa” (pertaining to Īśa, i. e. Śiva) and
uses the term for it, but if the language was not limited to that milieu, a wider term might be
more appropriate.
would be a relief, unless one wants to roll a dice with manuscripts sigla for
numbers.
Thus, as every practitioner knows, the theory may be fascinating, but
the practice can be messy, or, to put it in a more sophisticated way: “While
general principles are undoubtedly of great use, specific problems have an
unfortunate habit of being sui generis, and similarly it is rare to find two
manuscript traditions which respond to exactly the same treatment.”279
And so I must close this chapter without a clear methodological suggestion.
As usual British pragmatism has found the right words:280
The more open a tradition is, the less fruitful the stemmatic ap-
proach is likely to be, and other methods must be tried. These
range from empirical, common-sense approaches which accept
the necessities of an imperfect world, to elaborate statistical tech-
niques which aim at more objective results. In some cases it
is possible to adopt a flexible modification of the genealogical
method. The manuscripts are classified as far as is possible into
broad groups and the editor chooses his readings eclectically,
persuaded more by their intrinsic merit than by considerations of
affiliation and authority and taking care to balance these factors
to suit the nature of the tradition. But if contamination has gone
so far that, in the words of Housman, ‘the true line of division is
between the variants themselves, not between the manuscripts
which offer them,’ various approaches may be adopted which all
tend to concentrate on the variants themselves rather than on
the manuscripts which carry them [. . .] Ultimately, the basic
essential equipment is taste, judgement, common sense and the
capacity to distinguish what is right from what is wrong in a given
context; and these remain the prerequisite of human wit.
In practice the text-critical work remains difficult, and with its individual
solutions often proves somewhat theory-resistant. Here one might side with
279 Scribes and Scholars, p. 239. 280 Scribes and Scholars, p. 239f.
Lachmann—“Through theoretical doubts the work that is already difficult, is
only complicated further”.281
Phrased in a more positive manner, we could also say that editors of San-
skrit texts have at their disposal a rather differentiated repertoire of techniques
and methods of criticism, and need to test and weigh the application of one or
the other carefully. To make stemmatics a hard and fast rule is in my opinion
unnecessary, if not counterproductive. The unpleasant side effect of this is
that the type of objectivity and “Wissenschaftlichkeit” that the “Lachmann
method” had promised has to be given up.
But I think we need not officially reintroduce the subjective factor in edit-
ing, or even—like Donaldson—denounce, “with amiable severity, the entire
project of a “scientific” textual criticism”,282 for the selection of a reading
may not be automatic, but it must follow sound arguments. That such argu-
ments must be philosophical or literary rather than stemmatic can neither be
avoided nor should it be regretted. Or, in the words of Bédier: « Il est permis
d’estimer que la poésie a des raisons que l’arithmétique ne connaît pas ».283
281 “Durch theoretische Zweifel verwirrt man die an sich schwierige Arbeit nur noch mehr”.
282 Jerome McGann: A Critique of Modern Textual Criticism. Chicago 1983, p. 107. 283 Roma-
nia 64 (1938), p. 215.
The Pitfalls of Editing
1 I think the present chapter will clarify that I am far from defending this practice, as Witzel
thought: “Hanneder (2009/10) still defends this practice as the results of such editions were
normally reliable.” Witzel: Textual criticism in Indology, , p. 52. 2 See Lao Zi (Laotse). Der
Urtext. Übersetzt und kommentiert von Wolfgang Kubin. Freiburg 2011, p. 11ff.
of Theravāda texts, who thought they knew that early Buddhism was found in
Ceylonese sources and nowhere else, and had to accept gradually the existence
of other materials.3 The conviction about the validity of some sources was
apparently so strong that editions produced in the Pali Text Society were
sometimes based on merely a single manuscript, not because the text was
transmitted to posteriority in a codex unicus, but because more material was
deemed unnecessary.4
But the Pāli canon and some Buddhist Sūtras in Sanskrit seem to present
again a problem sui generis, which some editors have tried to circumvent with
the help of synoptic editions, where all versions are printed in parallel texts.
Let us assume that there are ten witnesses, then the first ten lines will give
ten versions of the text, very much like an old-style collation sheet. Peter
Skilling has written on this practice and asked the question whether this could
be seen as the solution to the dilemma of editing Pāli literature.5 There is
also one example for such a synoptic edition,6 which Peter Skilling in his
introduction to this work has called a test edition. Such editions may be
valuable tools for scholars, but are subject to the same criticism as multi-text
editions in European philologies produced in the wake of the so-called “new
philology”. Such editions have one clear advantage for the editor in that
they avoid problematic decisions, but with the effect that these are left to the
reader.7
In the following I shall merely give one example for a first edition from
one manuscript that has produced a problematic text. The other concerns a
text, where there is a strange discrepancy between the edition and the text in
the manuscript.
Bhāskara’s Cittānubodhaśāstra
3 Nalini Balbir: “Thoughts about ‘European editions’ of Pāli texts.” In: Thai International
Journal for Buddhist Studies 1 (2009), p. 1–19. 4 Op. cit., p. 3f. 5 “An impossible task? The
classical “edition” and Thai Pāli Literature.” In: Thai International Journal for Buddhist Studies
1 (2009), p. 33–43. 6 Santi Pakdeekham: Jambūpati-sūtra. A synoptic romanized edition.
Kachru: Kashmiri Literature. Wiesbaden: Harrassowitz 1981 (A History of Indian Literature 8.4).
p. 15ff. 12 Pandey, Sushama: Cittānubodhaśāstram śāradātah. devanāgarāks.ares.v anuvartya
prathamam . mudran.am. Rājānaka-Bhāskarakan.t.hapran.ı̄tam
. . Sampādikā sam
. skartı̄ ca Sus.amā
Pān.d.eya. Vārān.ası̄: Tārā Buka Ejensı̄ 1990.
apart from one notable exception,13 has received no attention. Many readers
must have noticed that the sense of this work seems often opaque, and as a
textual critic one must have suspected the usual problems with an edition of a
complicated philosophical text from a codex unicus.
Since I had read Bhāskarakan.t.ha’s commentary on the Moks.opāya repeat-
edly, and was fascinated by this versatile author, I kept my eyes open for
manuscripts of the Cittānubodhaśāstra for many years and finally the critical
mass was reached by combined efforts: we had a couple of Śāradā manuscripts
for the text, and a newly discovered auto-commentary, which seems lost in
Kashmir. An analysis of the sources yielded an unexpected result. While
the Śrı̄nagar manuscript, the one geographically closest to the author, some-
times presents an idiosyncratic text, it is further removed from the wording
of the mūla text as fixed through the auto-commentary, which is preserved
in a single Nāgarı̄ manuscript kept in Alwar. Unfortunately the manuscript
of the commentary is an editor’s nightmare. The scribe did obviously not
understand Sanskrit; it abounds in simple spelling mistakes, non-words, lost
or inserted nonsensical syllables and the like.
But taken together with the other materials, it is still possible to restore the
wording of the mūla text beyond doubt. Since the work demands a few years
of sustained attention, I decided to set up and apply for a research project,
but first I had to demonstrate to the funding institution, why a new edition
mattered. I expected a frantic search for those passages that would show my
point most clearly and in such a concentrated way that would not test the
patience of the unsalaried referees.
But it was absurdly simple: when I had read the first three verses, I had
more than enough material to diagnose a total failure of the printed edition.
What follows is a brief comparison of the first three verses of Bhāskarakan.t.ha’s
Cittānubodhaśāstra as given in the printed edition with the edition prepared
first for the application for the research project (and then with more sources in
the project itself). For proving my point I had used merely a few manuscripts
then at hand, now we have in addition an auto-commentary that allows us to
ascertain the original readings beyond any reasonable doubt.
vis.ayanicayam
. deham . prān.am. svarūpatayā matam .
dhiyam atha nabho nyakkr.tyaitat prameyatayā tatah. /
padam upagataś cinmātrākhyam . prakāśamayam . punas
tadanu sakalam. svāntar dhr.tvā sthito jayatād vibhuh. //
The wording and sense are not in doubt,15 and the text is largely secured
through the pratı̄kas quoted in the auto-commentary. Now let us see the
printed edition:
vis.ayanicayam
. deham . prān.asvarūpatayā matam .
dhiyam atha nabho nyakkr.tyaitat prameyatayā tatah. /
padam upagatiś cinmātrākhyam . prakāśamayam . punas
tadanu sakalam. svāntam. vr.tvā sthito jayatād vibhuh. //
From this only the bold conjecturer would have been able to gain the intended
sense. The changes are small, but the noun upagatih. cannot be construed
and renders padam cinmātrākhyam . doubtful, the reading vr.tvā spoils the
sense, and without the reading prān.am . , we easily fail to grasp that a doctrinal
sequence—usually the last item is termed śūnya—is meant here. The tragedy
here lies in the fact that the erroneous readings are merely misreadings of
the Śāradā manuscript. No textual criticism is needed, merely experience in
reading Śāradā and some knowledge of the Śaiva background of the author.
————————————————
————————————————
yad ādyam . pı̄yūs.am
. nijakaragatam. maunam abhito
vihāyetaś cintārasarasikatām
. yāmi vivaśah. //
Cittānubodhaśāstra 1.3 (Crit. edition)
The Maṅkhakośa
The following example does not concern an insufficient edition in the obvious
sense. It is, however, a text that needs to be re-edited, and a case where only
text-critical curiosity unearthed an astonishingly different text behind the
edition.
The author Maṅkha (middle of the twelfth century) was discovered for
European Indology at a rather late time. It was only in 1877 that Bühler
drew attention to “a hitherto unknown poet”,19 whose Śrı̄kan.t.hacarita20 was
considered unique, especially for its historical and biographical passages, but
nevertheless not much noticed outside his homeland Kashmir. Even later
16 The Indian tradition not only recognized Vedic (ārs.a) Sanskrit, but also Śaiva (aiśa), then
there are Epic, Buddhist, and in later times “vernacular” varieties. 17 This is the whole point
of discussing a perhaps bewildering or boring diversity of transmissional scenarios. 18 For
the author, his revised date, and many other details, see Stanislav Jager: Bhāskarakan.t.has
Cittānubodhaśāstram: Kritische Edition des ersten Kapitels nebst Erstedition des Autokom-
mentares. Ph.D. dissertation, Marburg 2016. 19 Georg Bühler: Detailed Report [. . .], p. 50.
20 Śrı̄maṅkhakakaviviracitam śrı̄kanthacaritam. Jonarājakrtayā tı̄kayā sametam. Ed. by Pandit
. .. . . ..
Durgāprasād and Kāśīnāth Pan. d. urang Parab. Bombay 1887. Reprint. Delhi 1983.
indological research did not treat his work in much detail,21 a fact noted in the
most recent work on the author’s poem.22 Maṅkha is at the same time author
of a homonymic Sanskrit lexicon called Anekārthakos.a or Maṅkhakos.a, which
was edited by Zachariae.23 The dictionary is important, in so far as—as Vogel
states—“Quite a few words and senses are registered here for the first time.”24
The programmatic passage in the introductory section of the Maṅkhakośa
is relevant in many respects. We hear that Maṅkha has used the lexicogra-
phers Bhāguri, Kātya, Halāyudha, Hugga, Amarasim . ha and Śāśvata, but also
25
Dhanvantari’s Nighan.t.u. He says that he used rules for grammatical gender,
but also the “examples of good poets”.26 It is therefore more than likely that
the commentary, in which these sources are documented, is from the pen of
the author, especially since the commentary only starts with the dictionary
part from stanza 9 onward, but ignores the introduction.
Fortunately most manuscript sources utilized by Zachariae for the edition
of his text are still available and accessible. Those he received from M. A. Stein
are in the Stein collection in the Bodleian Library Oxford, those procured
by Georg Bühler and deposited by him in the Deccan College, Pune, are now
available in the Bhandarkar Oriental Institute. Furthermore, the library of the
Deutsche Morgenländische Gesellschaft, Halle, holds the literary bequest of
Zachariae, where we find further materials on the Maṅkhakośa.
Zachariae was undoubtedly an excellent Sanskritist, and his edition pro-
duces a very reliable text. This can best be explained with an example.
21 With the following exceptions: Elisabeth Kreyenborg: Der XXV. Gesang des Śrı̄kan.t.ha-
caritam des Maṅkha: ein Beitrag zur altindischen Literaturgeschichte. Münster 1930. Bankim
Chandra Mandal, “Authorship and date of Maṅkhakośa.” Annals of the Bhandarkar Ori-
ental Research Institute 57 (1976): 160–166. Bankim Chandra Mandal: Śrı̄kan.t.hacarita.
A Mahākāvya of Maṅkhaka. Literary Study with an Analysis of Social, Political and His-
torical Data of Kashmir of the 12th Century A.D. Calcutta 1991. Bhagavatprasad Natvar-
lal Bhatt: Śrı̄kan.t.hacaritam. A Study. Baroda 1973. 22 See Walter Slaje: Bacchanal
im Himmel und andere Proben aus Maṅkha. Wiesbaden 2015. 23 Theodor Zachariae:
Der Maṅkhakośa. Herausgegeben mit Auszügen aus dem Commentare und drei Indices.
Wien/Bombay 1897. (Reprint). Varanasi 1972. Theodor Zachariae, Epilegomena zu der Ausgabe
des Maṅkhakośa. Wien 1899. [= Kl. Schr. 387–440]. 24 Claus Vogel: Indian Lexicography.
Revised and Enlarged Edition. Ed. by Jürgen Hanneder and Martin Straube. München 2015,
p. 334. 25 bhāgurikātyahalāyudhahuggāmarasim . haśāśvatādikr.tān / kośān nirı̄ks.ya nipun.am
.
dhanvantarinirmitam 26
. nighan.t.um. ca (3). liṅgānuśāsanāni ca vicārya laks.yam . mahākavı̄nām.
ca / kurute ’nekārthānām
. śabdānām. maṅkhakah . kośam (4).
Ms. BORI 337/1875–6
In the edition we read that Zachariae has merely given extracts from the
commentary,28 and a glance at the actual presentation of the edition makes
one wonder about the principles of extraction or indeed the original form of
the commentary.
I shall give 40cd as an example. In the edition it runs as follows:
mecakam
. tris.u kr.s.n.e syāt pumām
. s tu śikhicandrake
“mecaka, in all three [genders] means “black”, only the masculine
means the eye on the peacock’s tail.”
Here the reader may rightly wonder how the full commentary may have
looked like, for the extract is merely an abbreviation or condensation of the
information contained therein. In the example mecaka in brackets has been
added by the editor, but what about “śikhi” and the elision indicated by the
27 Lit. “lower mouth”. The context here seems different from the tantric one of picuvaktra as
summarized by Shaman Hatley in: Dominic Goodall and Marion Rastelli (eds.): Tāntrikā-
bhidhānakośa III. Vienna 2013, sub voce. 28 Thus the English title page: The Maṅkhakośa
29 I omit the details on the manuscripts and refer to the forthcoming edition. 30 Bombay 1893,
p. xv.
result. But the unusual cikuramecaka occurs in Moks.opāya 3.34.41b in the
phrase kālı̄cikuramecakı̄. Since the Moks.opāya originated in Kashmir two
decades before Maṅkha, perhaps the compound cikuramecaka was a known
phrase at least in Kashmirian Sanskrit.
Sometimes Maṅkha justifies meanings by what seems to be a common
saying, not a literary quotation: the word stoka is well-attested in the sense of
“little” etc., but Maṅkha gives a meaning putra,31 which lexicographers might
have regarded as doubtful, but the commentary suggests that the kośakāra
considered this meaning to be in popular use—at least the quotation does not
sound like a literary one:
These are but tiny examples to demonstrate how treacherous the thought
that we already are in possession of Maṅkha’s dictionary can be. Only an
unabridged edition of Maṅkha’s auto-commentary would enable us to utilize
this lexicon as a source, thus no more effort need be made to explain why a
re-edition is direly necessary.
What makes things worse or more interesting is that many of the suppos-
edly 3400 quotations need to be traced. Since Maṅkha’s time is fixed, we may
even have a firm terminus ante quem for the texts quoted, and we may gain
an impression of the kind and range of literature known to a highly educated
author in twelfth-century Kashmir.
Examinatio
If the first case applies, the editor need not act, in the last case the passage has
to be marked as corrupt or an emendation or conjecture has to be applied.
The other cases are more complex. Here it is at the discretion of the editor
to suggest a solution, or to assume that the style of the author is not as one
suspected, or that he is using unusual words, has a lapse in style etc. Here
conjecture can be a dangerous instrument that may weed out historically
correct readings and produce ahistorical ones. And here the maxim of the
lectio difficilior can be usefully applied to provide the argument for keeping
the unusual reading in the text. But we can see, following Maas’ distinctions,
that we can apply this argument only in the middle range, never to case (4),
for which Chadwick’s rule as quoted above should apply.33
The same kind of reasoning is applicable to the rule sthitasya gatir cin-
tanı̄yā. It can be understood as a cautionary remark not to emend too quickly,
not to change, for instance, a phrase in Kālidāsa or Bhartr.hari’s Śatakas, be-
cause it contains an un-Pān.inian form, for it might still be original, or to
change its wording to suit poetological taste.34 Unfortunately, the maxim
was misused to allow for and even justify implausible texts. The project of
editing the Moks.opāya has yielded many instances, where the text of the
Yogavāsis.t.ha is actually meaningless, but nevertheless commented upon by
Ānandabodhendra, as if it would make sense. Such a commentary may be no
more than a desperate attempt to squeeze the remaining sense out of passage
that was disfigured in the course of transmission and would have to be put
into the fourth category in Maas’ list. The detrimental effect of this is that the
32 Op. cit., p. 10. 33 “The principle lectio difficilior potior does not extend to nonsense”.
34 See the work of Goodall mentioned above, p. 99.
reader is tricked into believing that this is still a meaningful text, since the
commentator apparently managed to make sense of it.
A similar phenomenon is found in modern editing too. Sometimes editors
print a text that simply does not make sense and cannot possibly have been
written by a sane author. Such a text is bound to be confusing. Here the rules
demand that a text which is doubtful and cannot be “healed by” conjecture
should be marked as corrupt.
In the following we shall look at some examples, which show that by
ignoring this examinatio one can go wrong as an editor in presenting a text
that cannot possibly have been written by the author, and cannot be used for
further studies of the work. In such cases a re-edition of a text can become as
urgent a task as the first edition.
Ratnakan.t.ha’s Sūryaśataka
35 The text of the Sūryaśataka was edited by Stanislav Jager and myself, together with two
more Stotras by the same author: that of the Sūryarahasya was the result of Jager’s master
thesis, and in the end Alexis Sanderson added one further text by the same author. See Jürgen
Hanneder, Stanislav Jager and Alexis Sanderson: Ratnakan.t.has Stotras. Sūryastutirahasya,
Sūryaśataka und Śambhukr.pāmanoharastava. München 2012. 36 See above, p. 76. 37 West:
Textual Criticism and Editorial Technique, p. 61.
in problematic passages, especially since the text is often difficult and has
lacunae. I will give her edition of the third verse as an example:
The relevant folio of the manuscript is printed in her book and we can see that
the cause of the lacunae in the third verse is a broken margin.
Now a simple check of the metre shows that her first conjecture cannot be
correct. She reads prakhelatimira and while timira is the obvious guess, the
preceding syllable has to be long. Furthermore, the conjecture produces
an awkward compound: are we to understand that the elephant (kari) of
darkness (timira) is the play (prakhela—by the way not an attested form) in
the wilderness of space? Do we really think the author wrote such a clumsy
line? Was Ratnakan.t.ha, writing in the sixteenth century, still able to produce
high quality classical Sanskrit? Some adherents of a death of Sanskrit would
perhaps start to doubt.
Of course, Ratnakan.t.ha never wrote such a line. One has to read prakhe-
lattimira and will thereby restore a more convincing meaning: the elephant
of darkness playing in the wilderness called space. Metre, compound and
content are without blemish. But this is not all. As even the manuscript
page reproduced in Ghai’s work shows, we have to read krudhyad for kupyad,
spardhino for sr.tvarā, and for the metrically wrong conjecture dı̄rghatara we
propose dı̄vyattara. The result is a perfectly understandable verse:
vyomāran.yaprakhelattimirakarighat.āghātaraktacchat.ārdra-
krudhyaddvı̄pı̄ndradı̄vyattarakharanakharaspardhino ye nitāntam
kausumbhenāmbaren.āvr.tam iva kr.payā tanvate digvadhūnām
aṅgam
. dr.s.t.vaiva nagnam
. divasakarakarāh. santu te vah. śivāya (3)
With the previous edition being full of misreadings and wrong conjectures,
we hardly needed justification for producing a new edition. Without access to
the same manuscript, we would have perhaps even assumed that the editor
had other readings. From this perspective the use of previous editions as a
mine for readings can be quite problematic.38
When looking at the manuscripts, we had one promising experience. The
real accolade of the sophisticated editor comes when one of his conjectures is
confirmed. Witzel gives an example in his article on textual criticism,39 where
38 Dimitrov has given one impressive example to caution against assuming a simple relation-
ship between manuscript and edition. In the edition of the Kāvyādarśa by Thakur and Jha,
he has counted more than sixty instances of differences between the edition and the codex
unicus, on which it is based—on one page! Especially in editions without critical apparatus,
it is virtually impossible to establish the text of the manuscript. See Dragomir Dimitrov:
Śabdālam. kārados.avibhāga. Die Unterscheidung der Lautfiguren und der Fehler. Vol. 1. Wies-
baden 2011, p. 51. 39 Michael Witzel: How to enter the Vedic mind. online publication:
urn:nbn:de:bsz:16-savifadok-1090
a conjecture of his turned out to be exactly true: After conjecturing the begin-
ning of a text he had edited from microfilm, he could finally see the original
manuscript and noticed that a page had folded in when being microfilmed.
When he opened it, it confirmed his conjecture. In such instances one feels
philology is not just readings and guess work, but seems to open up, as far as
possible, a direct path to the past.
The following example affords an interesting case for discussing the validity
of one of the important rules of stemmatic editing, which is usually termed
eliminatio codicum descriptorum. One aspect of this rule is easily understood
and seems logical, but some of its implications are so problematic that we may
again question its general and universal applicability.
The rule is formulated by Maas as follows:40
This formulation has sparked critical comments, like that of Pasquali that he
misses the word “accident”, which nowhere occurs in Maas booklet. He says
that a single error is unlikely to prove anything.
The problem with this rule is that it is unclear what is gained by it. If one
has to collate the whole manuscript in order to assess that it has one more
error than its source, then the whole process will merely relieve the apparatus
from an irrelevant siglum (all readings except one would be identical with its
source) and not much is gained in terms of efficiency.
With an additional remark Maas shows where the main application of the
rule could lie:43
Methodological Problem”. In: Editing Greek and Latin Texts. New York 1989, p. 1.
We felt exhilarated when some time after preparing the edition we sud-
denly noticed that there was a second manuscript in the Welcome Library in
London and that this manuscript confirmed some of our conjectures, even
though they were minor and hardly sophisticated ones. Manuscript “W”, by
the way, was not only a fairly complete manuscript, almost without lacunae,
the text and the orthography seemed to be of a superior quality. In the course
of editing the readings of the new manuscript were almost always adopted.
After reading on, however, our joy was somewhat dampened by the fol-
lowing observations: In 43a ms. B has the lacuna ja . . . ti. W reads jayati,
which cannot be correct, since we have another verb stāt—no doubt easily
overlooked—in the verse. There are also some parts of the aks.ara left, which
suggest more likely jagati, which was our reading. Since the margins of all
folios of B are broken, it has regularly recurring lacunae. W, on the con-
trary, presents a complete text and has fewer errors. An application of Maas’
rules would lead us to a stemma, in which B could be an apograph of W, but
since it is clearly much younger, we would rather construe something like the
following:
The clue to the real state of affairs came later in the text. In 67c, we found a
matching lacuna in both manuscripts: dı̄pti . . . sāpti. This would no doubt
be a binding error in a normal transmission, the common sources must have
had exactly this gap, which makes the verse unmetrical, only that in one
manuscript it is caused by a physical defect, namely, the broken margin. The
last piece of evidence needed we encountered near the end of the text, where
in 91c both sources read mā . . . pala. In B the reason is the usual broken
margin, in W we find the following:
In the middle of the third line the scribe has apparently left and thus marked
a gap in his exemplar, a gap, which is identical with the one caused by the
broken margin in B.
Now we could put the pieces together: Our almost complete manuscript
W was an apograph of B, otherwise it could not have exactly the same gap as
the damaged manuscript. A stemma would look quite unexpected, actually
the opposite of what the normal rules about the distribution of error would
dictate:
Here the outcome is that we may stop editing or searching for a meaningful
text, we should rest content with reading manuscripts as individual instances
of a “Redeereignis”. But the main critique one could level at this type of
presentation of the problems faced by an editor of medieval literature is that
1 I will not attempt a translation, because it is nearly impossible to imitate the terminology. The
passage contains an abstract formulation utilizing all the catchwords of this school of thought
for what might just be expressed as: Do not edit critically multi-archetype texts. The text can be
found at www.edkomp.uni-muenchen.de/CD1/A1/Altgerm-A1-MB.html.
it uses rhetoric in order to veil any possible outcome that could otherwise be
of practical value.
It probably has become clear in the preceding pages that it is very instruc-
tive, if not to some extent indispensible, to become acquainted with the litera-
ture on the theory of editing and textual criticism from different disciplines.
Even if some of these works now strike us as too transient, especially with
the rhetoric of close anticipation, of pronouncements, or of breakthroughs—
which according to what we know now may not have happened—, these works
are mostly useful, because they force us to rethink indological practices in the
light of discussions in other fields. We may tend to think that other philolo-
gies will have not much to offer for Sanskrit philology, that their situation is
entirely different, and this may be often the case. But there are interesting
exceptions.
A good example for this is McGann’s A Critique of Modern Textual Criti-
cism. The author, who wrote in 1983, thinks that this was a time, when scholars
were “too busy exploring the fault lines of what they already know and experi-
menting with new models and ideas”2 as a consequence of which no reliable
guides were being written.
Some decades later, from our present perspective, this anticipation of
imminent developments can be put to rest. It seems to be a dynamic pattern
similar to, if not unconsciously modelled after Christian eschatological ex-
pectations. There too the predictions are not fulfilled, but after some time
renewed. Still only the cynic would have anticipated the speed or frequency of
new announcements and the short expiry dates of those already announced.
Some years after McGann the so-called “New Philology”, which will be dealt
with later, arose and changed again everything, or so it seemed.
However, having said that, McGann’s keen observations are still highly
instructive.3 The main thrust of his “critique” concerns the concept of autho-
rial intention in editing modern works. For understanding the concept, its
problems and limitations, we need to take as our main examples not works
from antiquity, where we have manuscripts written many centuries after the
author, and can only speculate on how the author wished to present his text,
but works, where we do have the author’s manuscripts as submitted to the
2 Jerome McGann: A Critique of Modern Textual Criticism. Chicago 1983, p. 2. 3 This is not
so say that these observations are spectacularly novel.
publisher, a printed version produced by a publisher, a second revised edition,
and any amount of additional versions, for all of which “normal”, meaning
here classical, textual criticism cannot possibly apply, because the author has
written all of them. For a classical editor all manuscripts, prints, etc. by the
author would have to be regarded as archetypes or even autographs, and the
problem is suddenly not one of misprints—scribal errors being of course
mostly irrelevant—, but of versions with equal authenticity. Apparently one
of the striking examples for this are some of Byron’s works, which exist in
“multiple manuscripts, multiple corrected and uncorrected proofs, a trial edi-
tion, a whole series of early editions at least three of which are known to have
been proofed and revised by Byron”.4 In such cases one “finds it difficult to
accept the idea that one of these texts, and presumably an early manuscript,
represents the author’s final intentions”.5
We have already spoken about the problem that specific genres and trans-
missional scenarios need different methods and rationales for editing. Follow-
ing the “best manuscript” or indeed a single manuscript might be extremely
naïve in editing a classical Sanskrit text, but can be the last resort when editing
some Buddhist Sūtras. In indological editing one has rightly distinguished
different areas with their own rules.6 In Vedic and Epic literature we have
to work and struggle with the concept of oral literature versus manuscript
culture, in a large portion of “classical” Sanskrit literature the criticism of
Latin classics is the closest parallel and therefore may apply. And we know
that there is no real counterpart to modern, in the European sense of print
culture, philology for Sanskrit.7
Since we know so little about the background of Indian authors, we have
become used to this type of compartmentalization and stopped looking fur-
ther. Thus, in the field of classical Sanskrit editing, where we have no auto-
graphs, we work on the assumption that the author’s will can still be inferred
and that this assumption may guide us through the transmission. But we
are not equally ready to work on the assumption of individual authorship in
4 McGann: A Critique of Modern Textual Criticism, p. 31f. 5 Op. cit., p. 32. 6 Oskar
von Hinüber: “Der vernachlässigte Wortlaut. Die Problematik der Herausgabe buddhistischer
Sanskrit-Texte.” In: Kurt Gärtner (ed.): Zur Überlieferung, Kritik und Edition alter und neuerer
Texte. Mainz 2000, p. 17–36. 7 An exception would be the question taken up recently by Manu
Francis on the influence of early prints on Indian manuscript culture.
Purān.ic or Epic literature which appear more likely to be multi-archetypal
“bardic” texts, slightly changed and adapted to the circumstances at every new
recitation.8
McGann proves his point by taking as an example an author, in whose
case we—unlike in Sanskrit editing—actually know the circumstances of his
activity: the case of Byron’s “Windsor Poetics”,9 which has a quite unexpected
transmission.
8 About the question whether Purān.as can be critically edited, see the prolegomena to the critical
edition of The Skandapurān.a 1. Adhyāyas 1–25. Ed. R. Adriaensen, H. T. Bakker and H.
Isaacson. Groningen 1998. 9 Jerome McGann: A Critique of Modern Textual Criticism.
“Final authority” for literary works rests neither with the author
nor with his affiliated institution; it resides in the actual structure
of the agreements which these two cooperating authorities reach
in specific cases.
In editorial practice this does not help much. And even worse, in McGann’s
final analysis the concept of authorial intention itself vanishes:14
Perhaps this may come as a shock for a Sanskrit editor. If we cannot really
assume that the author has written one version, why should we strive for find-
ing the best reading? But even if we knew in a specific case that an old-Indian
author had not written a text for publication, but given it as a lecture three
times in slightly different ways and that students had noted down the three
versions. We would still try to give the reader a single reading edition, rather
than a genetic edition showing the development of the topic from one lecture
to the next. If you think that with respect to Sanskrit literature this is pure
theory, I would like to point to the analysis of some Buddhist texts from the
so-called Pramān.a tradition, for which Helmut Krasser has developed the idea
that their peculiarities are best explained by the assumption that “that these
texts, like the digressions in the P[rajñā]P[radı̄pa], were written down by stu-
dents on the basis of a teacher’s oral instructions. Thus, these texts represent
first-hand information about daily life in Indian Buddhist monasteries of the
sixth century, giving us a fascinating glimpse of the education system of that
time.”16
Again, let me close with the observation that this does not fundamentally
change our approach: we still have to edit on the assumption that there was
only one text. But in some cases the above observations can heal an overt
security that “this only can be the text of the author”. It seems, we have to
reckon with the possibility that he may have changed his mind—a cautionary
tale against entertaining too simplistic ideas about authors and their texts,
and just one example how one can benefit from reading widely in textual
criticism.
“New Philology”
16See Helmut Krasser: “How to Teach a Buddhist Monk to Refute the Outsiders. Text-critical
Remarks on some Works by Bhāviveka”. In: Dhı̄h. 51 (2011), p. 49. I am very grateful to Isabelle
Ratié for alerting me to this.
stable, because newer philologies have appeared since. But the “New Philology”
has produced quite a stir and is an interesting didactic episode about the rise
and fall of (text-critical) theories.
The origin of “New Philology” is usually traced to one monographical
work by Cerquiligni,17 but mainly in hindsight. The founding document
of the new school was rather an issue of the journal Speculum,18 and the
details are not without interest. In an Editor’s Note Luke Wenger explains
that the editor Stephen G. Nichols wished to provoke a programmatic
positioning of invited authors by asking questions like: “Have medieval studies
become irrelevant? Do medievalists speak a (conservative) language of their
own, addressing antiquarian concerns of interest to no one but themselves?”
Eventually these questions were put to a group of participants that were surely
not aware of forming a new school,19 and the whole exercise converged in
an over-arching topic “Is there a new philology?” This genesis may sound
familiar to anyone who has planned interdisciplinary symposia or similar
events that had to be fitted into a catchy, innovative sounding title. In the
end the editor was convinced that there was a “New Philology” and he could
adduce as proof that so many scholars had dealt with the topic. That they were
responding to questions about a new philology, and importantly that their
responses were not always positive, did not seem to matter.
Whether we regard this process as an orchestrated political move to pro-
duce a new school, or an accidental combination of events that were then
cleverly exploited, in any case an analysis of the process is instructive for
every academic, for one may easily oneself become the founding member of a
new movement without being aware of it. The process started with a question,
merely a Leitmotiv for a conference. But then the hardly surprising fact that
participants dealt with the theme was taken as proof for its virulent existence,
although participants disagreed on many points. The perception of coherence
and thereby the creation of a school was part of the process of reception, but
cleverly projected back by the editor, who spoke of a “New Philology”.
The politics of this process is not without interest. Firstly, common sense
tells us that when you assemble a group of scholars it is not likely that a
We get the impression that the new movement is a desperate attempt to defend
philology in an untoward academic environment: “Medievalists are frequently
viewed by modernist colleagues as hostile or indifferent to contemporary
theory.”22 New Philology appears above all to be a phenomenon within an
academic culture, in which philology has become a term of abuse, where a
complete abandoning of the term has been recommended in 1948. This at least
highly ambivalent, if not overtly hostile attitude has to this date led scholars
to pretend that they are not philologists, whereas according to European
standards they would naturally be.
This North-American ban on the term is producing a schizophrenic sit-
uation in some German University departments, where we find the term
philology everywhere except in the English title of the section dealing with
English and American literature. Apparently American scholars would be
as confused to be called “philologists”, as German visitors to the USA, who
21Malicious gossip has it that the real cause for this selection of sources is the availability of
some key texts in English translations in University textbooks. 22 Nichols, Introduction, p. 1.
when asked about their “race” at the imigration office are usually at a loss to
identify it.23
As an indologist teaching in Marburg I was especially amused about the
notion in the Speculum volume that one had to leave behind the narrow
confines of old European study rooms as in Marburg, as an epitome of the
old philology. But at the same time romance scholars continuously refer to
Spitzer, Auerbach, and Curtius, who have their fixed places in many of the
“new(er) philologies”. All of them taught in Marburg.24
In the end the arguments come down to a very simple point, which is
visible already in the subtitle to the introduction,25 and which Nichols puts into
the mouth of Auerbach: “philology represented a technological scholarship
made possible by a print culture. It joined forces with the mechanical press in
a movement away from the multiplicity and variance of a manuscript culture
[. . .] The high calling of philology sought a fixed text as transparent as
possible, one that would provide the vehicle for scholarly endeavor but, once
the work of editing accomplished, not the focus of inquiry. It required, in
short, a printed text.”26 After all the new philology wanted above all to issue
the call ad fontes, a new reading of the sources:27
23 The concept of “race” itself has in post World War II education become an expression of Nazi
ideology and consequently it has been dropped. By the way, the word and eighteenth-century
concept of a “Caucasian race” is unknown to Germans. 24 “Gautier lamented that more
Germans in a single town (Marburg) were working on the chanson de geste than were French
scholars in all of France.” Howard Bloch: “New Philology and Old French” . In: Speculum 65.1
(1990), p. 40. 25 “Philology in a Manuscript Culture”. 26 Nichols: “Introduction”, p. 2.
27 Op. cit., p. 7.
the work of different artists or artisans—poet, scribe, illuminator,
rubricator, commentator—who projected collective social atti-
tudes as well as interartistic rivalries onto the parchment. The
manuscript folio contains different systems of representation:
poetic or narrative text, the highly individual and distinctive
scribal hand(s) that inscribe that text, illuminated images, col-
ored rubrications, and not infrequently glosses or commentaries
in the margins or interpolated in the text. Each system is a unit
independent of the others and yet calls attention to them; each
tries to convey something about the other while to some extent
substituting for it.
What the author criticizes is the tendency to use manuscript sources merely for
the texts they contain and ignore them for the further information they may
provide, whether through their provenance, artistic quality, etc. An important
observation, but not new, rather forgotten or ignored. For Nichols this
emphasis on the material aspects of the texts is the gist of the New Philology;
in another, later article he even substitutes the term in favour of a material
philology.28 So in a sense the creator of New Philology soon coined a newer,
clearer term.
The rationale behind this emphasis on the material aspect of manuscripts
is transparent and should be consensual. We might argue, for instance, that
without having the text as a physical object in the original form in front of us,
we cannot really recreate and appreciate the experience of reading it. Put in
this abstract way, it sounds and probably is quite exaggerated. But there are
rather concrete applications.
Let me first adduce an example from German philology. You may know
that German was written and printed in the so-called German script, as
opposed to the Latin script. Nowadays we associate this old German script
with the NS regime, but it was actually abolished by the Nazis in the early
1940s with the absurd argument of having been created by Jews. From that
time on German books had to be printed in Latin script, and old German
handwriting was not taught any more.
28 Stephen G. Nichols: »Why Material Philology«. In: Zeitschrift für Deutsche Philologie 116,
p. 10–30.
No one would suspect that anything could be lost through this transcrip-
tion, since except a few details (there are two shapes for the “s” depending on
whether it is within a word or at the end) both scripts have the same repertoire
of characters. But as experts tell us, there is a pun in a text of Goethe, where
a bird of prey, in German “Greif ”, says he does not want to be called an old
man, German “Greiß”. Only with the old spelling and in the old script it
becomes clear that this is a visual joke: In German script, in which Goethe’s
work was printed, and which he obviously had in mind when writing this
passage, “Greiff ” and “Greiss” looked almost the same and could easily be
confused.29
The old philologists and also Sanskritists were aware of this problem,
and one should remember that Schlegel preferred the printing of Sanskrit in
Devanāgarı̄, since he thought that the script is the skin of a language which
we cannot remove without disfiguring it.30
Coming back to the “New Philology”, we can see that behind the rhetoric
of novelty lies a true concern that philology itself could be neglected.31
Phrased in this way, new philology does not concern so much the invention
of a new method, but rather the application of the known method. Similarly
Howard Bloch:32
In this paper I will argue not only that there is nothing new in the
term “New Philology” (viz. Michele Barbi’s Nuova filologia, Flo-
rence, 1938), but that the old philology was in fact a new philology
29 Of course, if the ligatures were executed correctly by the printer, no confusion was possible,
since at the end of the word “ss” coalesce into “ß”. 30 Schlegel may not have thought about
the complication that Sanskrit texts were written in a variety of Indian scripts. 31 Siegfried
Wenzel: “Reflections on (New) Philology”. In: Speculum 65.1 (1990), p. 18. 32 “New Philology
and Old French”. In Speculum 65.1 (1990), p. 38.
(viz. the Neo-Grammarians) with respect to that which had pre-
ceded. Use of the labels “new” and “old”, applied to the dialectical
development of a discipline, is a gesture sufficiently charged ideo-
logically as to have little meaning in the absolute terms—before
and after, bad and good—that it affixes. On the contrary, to the
extent that calling oneself “new” is a value-laden gesture which
implies that something else is “old” and therefore less worthy, it
constitutes a rhetorical strategy of autolegitimation—with little
recognition, of course, that the process itself of declaring oneself
“new” is indeed very old, or at least as old, where the present
case is concerned, as Vico’s Nuova Scienza, which some see as
the beginning of philological science. The qualifier “new” is by
definition a relational term. Vico conceived of his science as new
with respect to the philosophy of Descartes; Meyer-Lübke and the
Neo-Grammarians, with respect to the Romantics; the Italian New
Philologists, with respect primarily to the textual methodology
of Joseph Bedier.
33 Suzanne Fleischmann: “Philology, Linguistics, and the Discourse of the Medieval Text”. In:
In: Speculum 65.1 (1990), p. 37.
by the pressures of discourse. It is through these and similar
gestures that we might ultimately reformulate philology’s role in
the field of medieval studies, adapting its praxis to the challenges
of postmodernism.
If we step back, we can see that the so-called New Philology is a defensive
movement within the coordinates of an interdisciplinary US-American en-
vironment and that it is supported mainly by scholars of Romance studies.
Some details can be understood as a specific reaction to editorial problems in
Old French Poetry.34 Despite its implicit claim to describe a more universal
phenomenon, it is based on observations in a rather limited field, but even
within its confines these theories are by no means accepted. Especially in
Germany this movement provoked a clear dismissal. The movement was some
time after its inception analysed, and the result was that it was by no means
“new”, but presented merely a distorted view of the old, at least the authors
appreciated that the discussion surrounding it was worthwhile.35
Apart from the problems concerning what is new and what is not, it seems
above all the term philology that is causing most misunderstandings. The
openness of the term has given rise to many unnecessary fights. For the
opponents of philology it is a narrow term, usually limited to studying merely
the language of a text, but not even its content, and never its context. For
its adherents it includes all these fields and any other field that might shed
light on the text. At least in the case of Sanskrit philology the practice of
philology is of no worth unless it is informed by the specialized disciplines
that suggest itself, whether it may be medicine, botany, history or archery.
Philology has no fixed spectrum and it is certainly not limited to language or
linguistics. The critique of philology is in my humble opinion an unfortunate
misunderstanding.
But for some authors movements like the “New Philology” are counterpro-
ductive:
The ‘New Philology’ has moved much (and many), it was (and
remains) at the center of an ongoing debate. But in the German
34 Martin-Dietrich Gleßgen and Franz Lebsanft: “Von alter und neuer Philologie”. In:
Martin-Dietrich Gleßgen et. al: Alte und neue Philologie. Tübingen 1997, p. 3. 35 Gleßgen
und Lebsanft: “Von alter und neuer Philologie”, p. 2.
research scene, which is orientated towards the history of trans-
mission its methods have long been developed and applied. But
this type of research has not been noticed by French and Amer-
ican “new philologists”. It may have to do with the language
barrier, but also that research in German tried its method on
text that did not belong to the canon (prose, mainly religious
prose). Perhaps the main reason is that research orientated to-
wards the history of transmission in Old German Studies did not
operate on the basis of theory, but orientated towards philological
practice, and has therefore not produced manifests like Bernhard
Cerquilignis ‘Eloge de la Variante’.36
36 “Die ‘New Philology’ hat viel (und viele) bewegt, sie stand (und steht) im Zentrum einer regen
Debatte. Aber: in der deutschen überlieferungsgeschichtlich orientierten Forschung sind ihre
Methoden längst ausgeprägt und angewandt worden. Allerdings ist diese Forschung durch die
französischen und amerikanischen ‘neuen Philologen’ kaum wahrgenommen worden. Das mag
an der Sprachbarriere liegen, das kann auch daran liegen, daß die deutsch-sprachige Forschung
ihre neuen Methoden zunächst an Texten erprobte, die nicht zum Kanon gehören (Prosa, vor allem
geistliche Prosa). Vielleicht liegt der eigentliche Grund darin, daß die überlieferungsgeschichtlich
orientierte Forschung der deutschen Altgermanistik kaum theoriebezogen operierte und – an
der philologischen Praxis orientiert – keine Texte von Manifestcharakter hervorgebracht hat wie
sie Bernhard Cerquilignis ‘Eloge de la Variante’.” Werner Schröder: “Die ‘Neue Philologie’
und das ‘Moderne Mittelalter”’. In: Germanistik in Jena. Reden anläßlich des 70. Geburtstags
von Heinz Mettke. Jena 1996, p. 33–50. 37 “[. . .] arrogante und selbstironische Postulat”.
Martin-Dietrisch Gleßgen und Franz Lebsanft: “Von alter und neuer Philologie”, p. 1.
38 Op. cit., p. 2.
For the academic subject such tendencies always result in a larger
problem, since philology—to paraphrase Luther—takes revenge
on her slanderers. The experience gained through many years
of handling sources cannot be substituted by anything. There
is a considerable editorial minimum, which presupposes paleo-
graphical as well as codicological practice, a readiness to deal
with philological details and a basic work on the epoch [. . .] That
this minimum is not everywhere given, is our main problem.39
39 “Für das Fach entsteht aus solchen Richtungen immer ein größeres Problem, denn die Philolo-
gie rächt sich – gewissermaßen mit Luther zu reden – an ihren Verächtern. Die nur in jahrelangem
Umgang mit den Quellen erworbene Erfahrung läßt sich durch nichts ersetzen: Es gibt ein hoch
anzusetzendes editorisches Minimum, das paläographische und kodikologische Übung, philolo-
gische Detailfreude und epochenbezogene Grundlagenarbeit voraussetzt [. . .] Daß ein solches
‘Minimum’ nicht überall gegeben ist, das ist unser eigentliches Problem.” Martin-Dietrich
Gleßgen und Franz Lebsanft: “Von alter und neuer Philologie”, p. 8.
reader to the actual transmission of the work. But in practice the approach
also creates a problem of its own:
Summarily we can say that most theories about editing are likely to widen
our perspective and are useful for practical editing. And it is true, as Pollock
states,41 that indologists do not tend to notice developments in other fields.
If one, for instance, writes that it is “a remarkable and highly important phe-
nomenon in the history of religio-philosophical literature that has still to be
fully addressed by modern scholarship” that the Saddharmapun.d.arı̄kasūtra
has no “Urtext”, but consists of a multitude of closely related recensions,42 it
must be said that the phenomenon is already well-known from Indian narra-
tive literature, and that textual critics of European medieval texts have long
known and analysed the problem, and discussed editorial solutions to it.
There are also, rarely quoted, instances of an overview on the different
genres of Indian textual criticism with their respective methods,43 which
contain a great deal of indological experience, without being framed by noisy
theoretical claims. I am therefore more doubtful that, as Pollock has suggested,
a “higher-order hypothesis” for Indian textual criticism is such an urgent
need:
40 “Zweifellos kann eine elektronische Mehrtextedition nicht von den wissenschaftlichen Auf-
gaben einer kritischen Textausgabe entbinden, geschweige denn diese ersetzen. Der Blick auf die
handschriftliche Varianz eines mittelalterlichen Textes mag faszinierende Lektüreerfahrungen
vermitteln, doch wird die große Mehrzahl derjenigen, die mit Texteditionen zu tun haben,
nach wie vor einen verbindlichen Lesetext und eine Buchausgabe [. . .] wünschen.” Thus
Michael Stolz in an internet review of several electronic editions of the works of Chaucer.
(http://computerphilologie.uni-muenchen.de/jg01/stolz.html). 41 See below. 42 See Gre-
gory Schopen: “On the Absence of Urtexts and Otiose Ācāryas”. In: Gérard Colas and Gerdi
Gerschheimer: Écrire et transmettre en Inde classique. Paris 2009, p. 192, quoting Ruegg.
43 Oskar von Hinüber: “Der vernachlässigte Wortlaut. Die Problematik der Herausgabe bud-
dhistischer Sanskrit-Texte.” In: Kurt Gärtner (ed.): Zur Überlieferung, Kritik und Edition alter
und neuerer Texte. Mainz: Steiner 2000, p. 17–36.
Related to this monadization is a widely-shared reluctance to
compare phenomena across eras, genres, languages, or traditions,
to draw broader conclusions from case studies, or to think self-
critically about the kinds of theory—concerning language, texts,
society, or history—that each of the essays, however superficially
pretheoretical, actually embodies (none of the essays engages
with contemporary work on textuality and philology; none, ex-
cept for the introduction, betrays any familiarity with it). While I
agree that it is too early to offer a general account of premodern
Indian philology, it is not too early to begin to develop general
principles about old and new philological practices. Many of
the authors very helpfully gesture in that direction, but they too
often refrain from explicitly formulating higher-order hypothe-
ses – even the introduction seems indifferent to exploring larger
findings, and to asking whether any connections or tendencies
unite the essays.44
44 Sheldon Pollock, Review of Écrire et transmettre en Inde classique. In: Journal Asiatique
299.1 (2011), p. 440. 45 “Occasionally the reader stares glassy-eyed at the presentation of a
47 “Zukunftsphilologie oder die nächste M[eth]ode”. In: ZDMG 163.1 (2013), p. 159–172.
48 Quoting from an internet abstract of a lecture, liberation philology implies: “balancing
the claims of the inside and the outside of the text; supplementing postcolonialism with post-
capitalism, or a concern over past wrongs with a concern for future rights; and finding way to
meet, from our small philological locations as specific intellectuals, the obligation to construct ‘a
planet-wide inclusivist community.’ ” In the lecture, which is at present available on youtube,
Pollock says that the topic could have also been called “post-capitalist philology” and it becomes
clear that here philology is gearing up for improving the world, a “scholarship [. . .] aiming
towards some different future”. Since our aim here is much more modest, I will not elaborate, but
merely mention that there is a further variation on the theme in Pollock’s “Liberating Philology”.
In: Verge: Studies in Global Asias 1.1 (2015), p. 16–21.
E d i t i n g S ā h i b K au l’s Wo r k s
�
The Author and H is Works
Seeing that many of these works are apparently Stotras, we can perhaps
explain why an edition has not been a priority for any editor. I am emphasizing
this, because there is a tendency with some scholars to relegate such works to
a second rank, behind proper philosophy or behind proper Kāvya. Stotras, it
seems, are commonly viewed as a popular, less sophisticated genre.
49 The forthcoming edition referred to is my critical edition of the works of Sāhib Kaul.
In the eyes of many research scholars Sāhib Kaul may have a further
shortcoming: it is clear that his own philosophy was not pure “Kashmir
Śaivism”, his works rather represent a mixture of Pratyabhijñā and Advaita
Vedānta. This is admittedly not the favourite of many research scholars
specializing on Kashmir Śaivism, or indeed some other branches of Indian
philosophy, who had to free themselves and their field of study from the
modern domination of Vedānta.50
On the other hand, Sāhib Kaul’s works are quite interesting for the reli-
gious history of Kashmir. The clan of the Kauls migrated from Mithila to
Kashmir, and fused their own ritual system and possibly also their theology
or philosophy with the local Kashmirian counterparts.51 This can be seen in
the fact that Sāhib Kaul wrote a ritual manual on Śyāmā, i. e. Daks.in.akālı̄, a
deity from his old home, but also a Stotra containing a Mantroddhāra of the
Kashmirian lineage deity Śārikā. On a philosophical level we will see that he
is using Advaita Vedānta and Pratyabhijñā side by side.
When discussing Stotra literature we need to speak about another, partly
subjective issue in editing, one that should not matter for an editor, but
sometimes nevertheless does. We tend to be more enthusiastic about editing
a specific text, when we like the author and enjoy reading his work, or at least
hold that his œuvre is in some ways important. Editing might also be based
on a more detached academic motivation: we may want to find a piece of a
historical puzzle (the text itself may be no more than a means to that end).
There are also other motivations that involve still more distance to one’s object
of study. Oldenberg explained his interest in the Veda like the interest of a
doctor in the delirious talk of lunatics. So there is quite a range of views an
editor might adopt regarding a text.
Partly overlapping with such perceptions are more official notions about
the value of the text one is dealing with. Since I have neither French nor
English examples at hand, I will give you one from German literature. It is
probably unnecessary to explain that the academic study of national poets
like Goethe is commonly seen as worthwhile. But our views are partly a back
projection, and naturally they represent a minority view fixated on upper
50 Some scholars were first inclined to believe that non-dualism in Kashmir derived from
Śaṅkarācārya’s supposed visit there, as reported in hagiographies. 51 For details, see my
forthcoming edition.
class, high culture, a view open to all sorts of criticism. If we go beyond this
and include data on what was actually popular at the time, suddenly names
like Kotzebue surface. About this author a consensus has built that he is not a
worthy object of study, he just happened to be “popular” at the time, but had
no lasting value and was denied admission into the halls of world literature or
even of German literature. Only recently the ban was lifted, because it would
have been quite irrational to deny that a study of popular culture may tell us
something about the mentalities of the age. I adduce this perhaps far-fetched
example, because similar presuppositions may shape the way we understand
and read Stotras. They represent popular religion, and thus to some they
deserve no place except far below high philosophy or theology.
Recently Hamsa Stainton has, in his doctoral thesis on Poetry and Prayer:
Stotras in the Religious and Literary History of Kashmir,52 made a fresh attempt
at understanding Stotras. Faithful to the North-American tradition of trying
to explain the larger view on the topic, he has studied an impressive spectrum
of Kashmirian Stotras53 and produced a positive view of the genre.
Stainton explains that the typical framework for understanding that the
same authors wrote philosophical treatises as well as Stotras, is to talk of
those scholars we like to see as philosophers as passing from a “noetic” to an
“emotional” register (Torella), where they could express themselves as freed
from the constraints of philosophical discourse. Stotras, as Fürlinger says,
referring to the Śivastotrāvalı̄, were often outpourings, and therefore follow
another logic; they do not, for instance, display the expected logical order
etc.54 Stainton carefully disagrees and assumes that such ideas are more due to
a “Romantic legacy” in interpretation—poetry conceived as the spontaneous
expression of a poetic genius—but that such explanations are not sufficient:
52 Columbia University 2013. 53 This includes unpublished materials and some Stotras of
Sāhib Kaul, which I had sent to him in a preliminary edition. 54 Although in the case of the
Śivastotrāvalı̄ this may be due to an editorial failure, for which see above, p. 80.
expository treatises. While his philosophical writings aim to
persuade his audience with logic and argumentation, his poetry
exemplifies his teachings and demonstrates for his human au-
dience how his non-dualistic theology can be implemented in
speech. This can be seen in how both Utpaladeva and Ks.emarāja
deal with key issues relevant to the study of stotras, including
poetry and poetics, the acts of praise and prayer, and bhakti.55
I would agree that there is more to explaining Stotras than a simple reference
to “emotion” as opposed to the “rational”. Let me quote just one example
from Utpaladeva’s Śivastotrāvalı̄:56
jayanti bhaktipı̄yūs.arasāsavavaronmadāh.
advitı̄yā api sadā tvaddvitı̄yā api prabho (1.5)
There are different modes to read verses like these. If one looks at Ks.emarāja’s
commentary, one is treated with the expected doctrinal background and
interpretation:
55 Op. cit., p. 192. 56 All references are to the edition by Rājānaka Laksmana: The Śiva-
. .
stotrāvalı̄ of Utpaladevācārya with the Sanskrit Commentary of Ks.emarāja. Varanasi 1964.
contradiction”. Are we to think that this is just an alam
. kāra Utpaladeva was
particularly fond of?
Let me first quote how Stainton would interpret these texts:57
Again, I would agree, but add that this genre even allows the author and
reader to free himself from the usual doctrinal modes of understanding the
experience that might lie behind the doctrine. Utpaladeva speaks of a drunk-
enness, which Ks.emarāja tones down in his reinterpretation by giving it a
technical meaning: unmada for him means hars.a, whereby he loses the image
of drunkenness, but—as I would argue—also the point of the verse that like a
drunken person cannot properly distinguish objects, one drunk with bhakti
cannot really say whether he is advitı̄ya or tvaddvitı̄ya. The image of drunken-
ness is crucial for understanding that in this uncommon state the boundaries
of subject and object have changed.58 But not only does the subject-object
division seem blurred, apparently different modes of experience that to the
sober philosopher are mutualy contradicting suddenly seem to coexist in this
state: a non-dual state (advitı̄ya) and one where the subject is in some relation
to Śiva that is not specified.
57 Poetry and Prayer, p. 196. 58 For the image, compare Moks.opāya 5.34.94, and in the transla-
tion Der Weg zur Befreiung. Das Fünfte Buch. Übersetzung von Roland Steiner. Wiesbaden
2015, p. 197, fn. 7.
The philosopher would immediately demand to know whether this is one
of śivasāmya, śivaikya or -sādr.śya, but the advantage of the genre of Stotra
is that there is no place for this. What Ks.emarāja does, is to rationalize away
this ambivalence. He seems to tell us that although Utpaladeva acts in the
guise of the drunken devotee, he is still our well-known philosopher, who
exactly knows what the dvitı̄ya is, and who presents us with a verbal riddle,
which seems, but actually is not contradictory, for this is the point of the
virodhacchāyā. What if Utpaladeva actually meant it as he wrote it, that is,
expressed his wonder about this strange coincidentia oppositorum?
Ks.emarāja does not only try to pull back the free expressions of Utpaladeva
into solid doctrinal categories and treats poetry no different than Śāstra, but
he also severly criticizes the editor of the Stotras59 on doctrinal grounds. After
commenting briefly on Śivastotrāvalı̄ 17.47, and showing that it expresses an
idea similar to Bhagavadgı̄tā 12.15, he has to vent his anger on the editor(s),
for the first time in his commentary rather explicitly:
Is this verse the only case of irritation60 in this Stotra full of doc-
trines of the glorious great essence? True, one has to remember
things like “the knot of all vāsanās” (17.15). [Here follow seven
other examples]—but here and in other cases the verses are not
in accord.61 This is the “grace” of Śrı̄viśvāvarta alone, who has
spread out the collections incorrectly (or: full of faults). Thus
there is, in other Stotras also, much that is similarly inappropriate,
but we have not revealed it.
nanu ca śrı̄manmahāsāroktimaye ’mutra stotre ’yam . śloko dadru-
sthānı̄yah. ? satyam “aśes.avāsanāgranthi” ityādikasyāpi smar-
tavyam / [. . .] ityādayas tv anugun.ā apy atra ślokā na santi /
tad ayam asamañjasaśayyāprastārin.ah. śrı̄viśvāvartasyaiva pra-
sādah. / evam anyes.v api stotres.v evam
. prāyam
. bahv anucitam asti,
tat tu asmābhir nodghāt.itam.
59For this, see above, p. 79ff. 60 Literally: of a “rash, skin eruption”. 61 Supposedly with
what he identifies as the main doctrine that should be in the text.
be another possibility to understand Stotras, namely as a poetical refugium
where one can express ideas that in a purely śāstric setting would be much
more difficult to express and defend, a place for religious or mystical poetry.
M ystical Experience
Here the fixation on “gender” and “power” has pushed alternative interpre-
tations out of sight. It is likely that Lallā suffered from male domination
and that we can interpret many episodes and statements in accounts of her
62 Michelle Voss Roberts: “Power, Gender, and the Classification of a Kashmir Śaiva ‘Mystic’.”
In: The Journal of Hindu Studies 2010, p. 285. The verse corresponds to no. 94 in th edition of
George Grierson and Lionell Barnett: Lallā-vākyāni. London 1920.
life from that angle, but it is not a holy law that each and every statement
must be interpreted in that manner and that all alternative interpretations are
henceforth unimaginable. In the passage under review I think that a viable
alternative would be an interpretation of the passage as describing not a social,
but a psychic process. Her “guru” might not be a person after all, but an inner
voice that directed her inwards.
I shall quote another very instructive example from this text. As is
well-known a corpus of her sayings has been translated into Sanskrit by
Bhāskarakan.t.ha, and here we encounter yet another facet of the versatile
author introduced above. The following verse is Grierson’s translation into
which I have inserted the (tatsama) terms from the original Kashmiri in order
to facilitate a comparison. Then follows the Sanskrit version of Bhāskara:63
Holy books (tantra) will disappear, and then only the mystic
formula (mantra) will remain.
When the mystic formula (mantra) departed, naught but mind
was left.
When the mind disappeared, naught was left anywhere,
And a void became merged within the Void.
64 Oliver Freiberger and Christoph Kleine: Buddhismus. Handbuch und kritische Ein-
führung. Göttingen 2011, p. 233ff. 65 Robert H. Sharf: “The Rhetoric of Experience and the
Study of Religion”. In: Journal of Consciousness Studies 7.11–12 (2000), p. 267–87.
quite different type of academic attention.66 Here we find research conducted
on the effects of meditation or, to be more exact, altered states of conscious-
ness, from the perspective of neuro-sciences and thus from a diametrically
opposed angle. Contents or doctrine are only marginally interesting, what
is in focus are physiological data to describe the short term and long term
effects of meditation on the brain. Details are not only beyond the expertise
of the Sanskrit philologist, they belong to a habitat of specialized neurological
journals, which they leave only when health effects can be claimed, as when it
is found that “meditation delays Alzheimer’s disease” and so on. Here contents
of experiences do not really come into play, since there is no framework to
deal with them. For a description of meditative experiences the language
of the practitioners who form the object of study is used, which is usually
(Tibetan) Buddhist.
Since neuroscientists are often not aware of the complications surrounding
a supposedly “neutral” description of “experience”, but use terminology
uncritically, it is quite difficult to establish a dialogue between the fields.
The communication becomes even more problematic, when the results of
research are then fed back into modern Western Buddhism on a popular level.
For instance, in a special issue of Buddhismus aktuell (2012.3) devoted to
“Buddhismus und Wissenschaft im Dialog. Meditation und Gehirnforschung”,
we find that European Buddhists unknowingly employ the peculiar rhetoric
of the supremacy of “Eastern Religion” or Buddhism once established by
Buddhists as well as Hindus to counteract European colonial emphasis on
“science”. When we read in one article that a brain scientist explains how
research benefits today from the methods of Buddhist meditation, the context
evoked is that of—thus another headline—“ancient wisdom and modern
science”. If we add the substantial body of literature recently produced on
the much-promoted topic of “Buddhism and Science”67 the whole matter
becomes hopelessly unmanageable, and the philologist or scholar of religion
will often wonder where and how to start untangling this net. But since the
topic has become part of an indological discussion about the relationship
66 A popular overview of the research is given in: Ulrich Ott: Meditation für Skeptiker.
München 2010. 67 See only Donald Lopez: Buddhism and Science: A Guide for the Perplexed.
Chicago 2008.
between doctrine and mystical experience, we can not altogether ignore it
here.
Buddhologists are aware of some articles written by one of the great
authorities in the field, Lambert Schmithausen, who explained in an early
article in 1976 “On the relation of spiritual practice and philosophical theory
in Buddhism” in which way some items of Buddhist Philosophy are most
likely based on spiritual experiences. To this Eli Franco has recently reacted
in much detail in an article published in an interdisciplinary volume devoted
to meditation.68 Therein Franco dismisses the thesis that there is a close
correspondence in Buddhism between meditation and metaphysics, an idea
expressed in the most sweeping way by Regamey and Conze and in a more
careful manner by Schmithausen. The two former scholars play no role in
the article by Franco, who concentrates on Schmithausen. Franco starts with
the case of the four meditations,69 which he thinks confirms Schmithausen’s
thesis. The reason is not given in clear terms, but it seems that the fact that
these insights are described in the texts as originating from meditation is
crucial.70
Franco conducts his discussion without any explanation of the complexity
of the terms of “mystic, spiritual, meditative” or other similar experiences.
We are, it seems, inferring the experience from the texts. Surely mystical expe-
rience is not a realm where our normal experience will give us a background
from which to judge with any amount of certainty whether a text is describing,
or rather rationalizing, an experience that might be called mystical or not.
While in many other fields of philological study the specific practical back-
ground can be ignored according to the dictum of Terenz homo sum, humani
nihil a me alienum puto, it is in this case more difficult. Even a psychiatrist
may not be entirely confident that he knows exactly how an acute psychosis
feels.
To give an easier example: When reading Sanskrit texts on metallurgy,
how can we know whether the author was describing (1) his own experience of
71 See Jürgen Hanneder: Der »Schwertgleiche Raum«. Zur Kulturgeschichte des indischen Stahls.
Stuttgart 2005.
The reason for this seems to be72 that there is no emic claim to the effect, and
that arguments in its favour are purely rational. A few more examples of a
similar kind are given, then the argumentation shifts to the points actually
raised by Schmithausen. The first would be the anātman doctrine. According
to Franco “the evidence he adduces for the hypothesis that it has its origin
in meditative experience is rather meagre. Schmithausen is one of the most
learned scholars of Buddhism of our time, and yet for the negation of the soul
(ātman) in meditation he could find no earlier testimony than Candrakı̄rti’s
Madhyamakāvatāra [. . .]”73 He adduces Steinkellner’s and Vetter’s explana-
tion of this doctrine as a reaction to the early pudgalavāda and dismisses
Schmithausen’s theory as inconclusive. A similar treatment is given to the
philosophical extension of the idea of anātmatva into the doctrine that “all
things lack substance”. Here Franco asks and concludes: “Could one maintain
that the development of this more sweeping doctrine is due to meditation?
Again: evidence is lacking and one could make up various scenarios all equally
speculative.”74
This is probably true, but we may also ask the question, what type of
evidence Franco would admit for proving Schmithausen’s thesis. It seems only
the clear statement in the texts that a certain doctrine is the direct outcome of
meditational experience will do. But even this criterium cannot always apply,
as the example of the four noble truths shows. The four noble truths surely
satisfy the criterium that they are presented as originating in meditation, but
here the composite nature of this doctrine (as analysed by Bareau) is enough
to disprove the idea. Spiritual experience, it seems, at least has to be logical. As
a result “it seems that in Conservative Buddhism most philosophical doctrines
did not originate directly from meditative practice.”75
Before following Franco into some details and identify problems of his
approach, we must note that he uses a variety of arguments to dismantle
Schmithausen’s thesis. For most doctrines under scrutiny for possibly orig-
inating in meditation experiences, he adduces an alternative philosophical
explanation, mostly working with the typical logic of Frauwallner, that doc-
trines respond to flaws in the argumentation.76 When this is the case, a
72In many cases Franco’s reasons have to be inferred. 73 Franco: “Meditation and Meta-
physics”, p. 103. 74 Op. cit., p. 104. 75 Op. cit., p. 105. 76 Op. cit., p. 104.
meditation experience can be ruled out. In order to satisfy Franco’s demands,
the Buddha or his later voices would have had to leave some prominent doc-
trine unexplained, for it apparently needed to be “counterintuitive”. In other
words, we are looking for failures to explain the supposed meditational experi-
ence in a coherent doctrine—not only a high, but also a fairly arbitrary hurdle.
Furthermore, one might suspect that the search for such unexplainable breaks
in a religious doctrine that has given rise to one of the most sophisticated and
impressive philosophical reflections is, if not doomed to failure, at least an
unrealistic endeavour.
Predictably also a whole host of other theories cannot be proven to be
the result from meditative experience. For the question of a Buddha-nature
residing in all beings, Franco calls Zimmermann as a witness to say: “Of
course, we cannot know whether the idea of the Buddha-nature in living
beings resulted from a novel meditative experience [. . .]”77 We may add:
Of course, one cannot possibly know about experiences made by persons
millennia ago and how they committed them to writings. Did they use literary
forms, paradoxes, or philosophy? If the type of reasoning applied by Franco
were enough to exhaust the relationship between literature and the real world,
we could dispense with many disciplines of interpretation.
The article of Franco ends with the well-known observation that medita-
tive experience does not imply a state of tabula rasa and its interpretation is
couched in images of one’s own culture: The Jewish mystic, he says, does not
experience the anātman, and a Buddhist Yogi will not meet God the Creator.78
In the end this approach will not tell us much about the relation between
“experience” and doctrine, for “it would even be hard to prove that theories
about meditation arise from meditative practice.”79
But perhaps the argumentation can solved in a much easier way. One cru-
cial qualification introduced by Franco lies in the word “directly”. He does not
deny that an indirect connection between doctrine and meditation is possible,
77 Op. cit., p. 119. 78 This is a funny claim, since if one looks at actual descriptions of religious
experiences one does find the occassional mismatch, for instance, Christians faced with an egoless
ultimate reality. See, for instance, Bernadette Roberts: The Experience of No-Self. Shambala
1982. 79 Franco: “Meditation and Metaphysics”, p. 117.
but considers such a relation, since “meditation is a central phenomenon in
Buddhism”, as “trivial”.80
The question is: would it make a difference if we were to understand
“originating directly” not in the manner of Franco in disproving Schmithausen,
but in a more multidimensional manner? For instance: does the Kāmasūtra
“originate directly” from the sexual experience of the author? Why should
it? No reader would doubt that his general experience would be sufficient
qualification for writing the book. The question would have to be called at
least pedantic, or—if one would use it to prove philologically that he never
had sex—sophistic.
Sometimes Franco simply states that there is no connection: “Now, what
is this perfection of wisdom that is repeatedly praised in this Sūtra? It is
the insight that all final elements of existence (dharmas) are unreal, and
this insight is realized during a meditation that causes the suppression of all
consciousness and feelings. In other words, when the perfection of wisdom
is attained, the world disappears [. . .]”81 What confuses Franco is that “the
content of this meditation corresponds to absolute reality. When the yogi
emerges from the meditative state, he generalizes his experience: [. . .] the
whole world is but an illusion [. . .]”82 The passage is now dissected according
to the method described above: “Can we conclude that this counterintuitive
doctrine has arisen from meditative practice? I fail to see that there is evidence
for such a conclusion. There are at least three possible hypotheses that may
account for the development of the perfection of Wisdom [. . .]”83 Here, too,
the fact that a doctrine has been explained philosophically in various ways, in
this case by Western scholars, precludes that it originates in an experience. I
see no reason why it should not, but agree that it cannot be proven within the
argumentative framework set up by Franco. To my mind, a scientifically valid
proof that Buddhist doctrine is derived from “experience” would necessarily
operate within what we should understand as a theological argumentation.
I therefore agree with Franco that the theory that Buddhist doctrines stem
directly from experiences cannot be proven by philological and philosophical
given in Ulrich Ott: Meditation für Skeptiker. München 2010, p. 112–117. 87 Ibid., p. 117–125.
teristics of such an experience. Whether the authors of the respective texts
are really those, who had the experience, or whether they are in various ways
removed from the “Yogic perception”,88 we may never find out. But there are
interesting phenomenological parallels. For instance, the above-mentioned
Carl Albrecht, who uses a neutral, but sometimes also Christian language to
describe mystical experiences, has given a characterization of the mystical
state, which could provide a bridge to the Buddhist notion of non-self:
88 Thus the title of the volume edited by Franco. 89 “Dem Ankommenden, welches hier
erscheint, fehlt jede Ich-Qualität. Es ist weder ein Bereich von Erlebnissen, die früher dem Ich
zugehörig waren, es ist kein schon einmal Gehabtes oder Gewußtes, noch ist es ein Bereich,
den man in Zukunft einmal zur Verfügung haben könnte. Es ist kein unbekannter, etwa im
Symbol personifizierter Teilbereich des Selbstes, sondern es wird erlebt, als ob es ein schlechthin
Fremdes und Anderes wäre.” Quoted in Simon Peng-Keller: “Präsenzschau in Versunkenheit
und Ekstase. Carl Albrechts Phänomenologie der Mystik.” In: ZMR 90 (2006), p. 90–102. (p. 93)
The translation is mine.
�
The reason for this excursus is that some of Sāhib Kaul’s Stotras are par-
ticularly interesting in the context of religious experience. It is also worth
mentioning that Sāhib Kaul not only “signs” some of his Stotras in the con-
cluding verses, he also destroys the illusion of a poetic subject different from
the author by speaking of himself as “Sāhib Kaul”. The author moreover
appears from his works as having an individual style, with a peculiar fondness
for word-plays, for which I give one example:90
A good example for the way he describes what we might understand as reli-
gious experiences is the following passage from the Svātmabodha:
When I was born, all things were born, and I was the source of
myself and of all things, and had I so wished, I would not have
been born and neither all things. And if I were not, God would
not be either. That God is God, for this I am the cause. And if I
were not, God would not be God.91
91 “In mîner geburt dâ wurden alliu dinc geboren, und ich was sache mîn selbes und aller dinge;
und hæte ich gewolt, ich enwære niht, noch alliu dinc enwæren niht; und enwære ich niht, sô
enwære ouch got niht. Daz got got ist, des bin ich ein sache; enwære ich niht, sô enwære got
nicht got.” 92 See Meister Eckhart. Deutsche Predigten. Eine Auswahl. Mittelhochdeutsch.
Sāhib Kaul’s works are not merely devotional Stotras: the so-called Citsphāra-
sārādvaya is a soteriological dialogue, a discourse to enlighten a student. The
work has been edited before by Janardan Pandey in his collection Śaivādvaya-
94
vim . śatikā. Unfortunately the editor’s sources are not disclosed, and there
is reason to believe that in the process of transcription many mistakes were
added. It is again, and I shall give examples later, a text edition in which often
even the basic sense is in doubt.
First I shall give a summary. Without introduction or invocation, the
Citsphārasārādvaya reports a religious dialogue: The disciple, seeing the
suffering of existence (vs. 1c), speaks to his teacher (vs. 2), describes his
mental state and requests to be shown how his own self is an abode of being,
consciousness and bliss. The reader cannot really get the sense from Pandey’s
edition, which for sacciddhars.a (3d) “being, consciousness and bliss”, reads
saccidvars.a “being, consciousness and rain”, thereby reminding one of Rowan
Atkinson’s jokes, who when playing a nervous vicar in Four Weddings and a
Funeral prays in church to “the father, the son, and the holy goat".
The teacher, who is said to have found his own luminous self, listens to
the words of the disciple, and now starts to teach from his own experience
(4). Since the mind of the pupil is fortunately already pure, has escaped like a
lotus the mud through his own power and has been washed with the waters of
detachment, he is merely in need of the touch of the Sun, that is the teaching,
to blossom (5). The setting reminds one of the Moks.opāya, where Rāma, in
a religious depression, has all the prerequisites to be liberated through this
kind of religious didactic intervention.
93Kurt Flasch: Die Geburt der “Deutschen Mystik” aus dem Geist der arabischen Philosophie.
München 2006. 94 New Delhi 1997.
At this point the author suddenly reveals his identity. It is Sāhib Kaul
himself, or rather Śiva acting through him, who teaches the non-duality of the
expanse of consciousness—hence the title of the work—to his pupil (6).
Having thus consoled him with true words, the knower of the free
self, the great devotee of Śiva, residing in the last state of Sāhib
Kaul, taught a true disciple, whose mind was pacified, about the
supreme nature of his own self. Thus he taught this (prakr.ta) free,
incomparable non-duality of the expanse of consciousness.
In his disquisition the author uses the Vedāntic triple methodology of hearing,
thinking and meditating, in the course of which insight is said to appear on its
own accord (svatah. 7d). According to the essence of the Vedic revelation, Śiva
is always liberated, awakened, pure, etc. (8a), whereas the world is identical
with him, consists of him, and appears thus, or is, real (8c). We find here in
other words a curious blend of ideas from monistic Śaivism and Vedānta, and
similarly in the next verse, where the eternal reality brahman, from which
everything is born like waves from the ocean, is said to have an egoity (ahantā),
and its realization is the recognition of something forgotten (10). Here we do
not find the term, but idea of pratyabhijñā.
It was already known from the Devı̄nāmavilāsa that Sāhib Kaul used
especially the popular Pratyabhijñāhr.daya of Ks.emarāja alongside other ideas,
but the extent to which he mixes different strands of thought is well worth
noting. Also in the Citsphāra we find pieces from the Kashmirian monistic
Śaivism: Śiva is said to appear as someone else (asah. 12c), like an actor who
plays different roles and cannot be recognized. The student is repeatedly urged
to give up his limitations, the “contraction” (sam . koca) that causes suffering.
All fears will be groundless, once one’s own nature as Mahābhairava is realized
(14).
In the text then follow appeals to the disciple in his state of religious
despair to understand that he himself is Śiva, is all-pervading, pure, etc. Much
of the work is just a rephrasing of this central message with the help of all kinds
of religious doctrines and images. The emphasis is here not philosophical or
systematical, it is like an individual instruction aiming at bringing the student
to the intended realization of his true nature.
It is very unfortunate that Pandey has in his edition regularly obscured
the text. Let us look at one verse:
I am not sure, how one could possibly understand the verse and instead of
trying I shall merely give my edition in order to demonstrate the difference.
anvis.t.am
. sad athāsti yan nahi kadā kutrāpi vā pus.pavat
khe chāyāsv iva yaks.avac ca marubhūtr.s.n.āvad artham . mudhā
hānau tasya kiyatprayāsagan.anam . kı̄dr.k ca vā syād aho
yaddhānena nutasya cidghanavapur bhı̄to yathāsi sthitah. (45)
Unfortunately these are not singular instances. The reader may also pity Sāhib
Kaul for being suspected to have written phrases like “the spilling of poison
in a picture, etc.” (citrādau vis.aphūtkr.tir). The idea is that poison painted in
a picture cannot harm anyone, but the example is rather ill-suited because no
one would recognize poison in a picture in the first place. And what would
be the scope of -ādi? The intended image, as far as we can say, is not too
complicated and to suspect an accomplished poet like Sāhib Kaul to spoil it
in that manner should arouse doubts in the reader and textual critic. For all
we know Sāhib Kaul never wrote such a line. All manuscripts collated for the
present edition read citrāhau vis.aphūtkr.tir, so we are talking about the world
being as ineffective to harm the subject as a poisonous snake in a painting.
Here is one further example, which shows that Sāhib Kaul’s thought world
is gradually lost in these recurring misreadings:
The verse mentions king Lavan.a,95 who in one episode in the Moks.opāya is
magically transformed into a Can.d.āla world, lives there, but returns to his
kingdom and takes up his reign again. What he loses is his fear of being really
a Can.d.āla, an interesting interpretation by the way. The reading svāmim .s
tvam removes the object on which three adjectives depend, the correlative
tadvat is lost to a misreading as tattvam . , and the pupil is surely not asked
to give up māyām ātmamatim . , but māyāmātr.mitim. “the experience of the
illusionary subject of perception (māyı̄yapramātr.)”.
In the 64 verses of the Citsphārasārādvaya there are around hundred
such instances, in which the re-edition based on all available sources differs
substantially from that of Pandey. I will conclude with one last example. In
verse 30 the author tells us that we rightly recognize a tree, when we see leaves,
branches, flowers, etc. and likewise we recognize Śiva or the self, when we see
the world. The first two Pādas of the verse are given in the edition of Pandey
as:
95 Peter Stephan: Die Lavan.a-Episode im Moks.opāya: Über den illusionären Charakter per-
sonaler Identität. Textkritische Edition, Erstübersetzung, Studie. (urn:nbn:de:gbv:3:4-2891) Halle
2008.
The obvious problem here is that vr.ks.e sāvasukāhva makes no sense. Unless,
of course, we approach the text in the way adopted by some pre-modern
Sanskrit commentators, who work according to the famous maxim sthitasya
gatir cintanı̄yā, “we must think about [how to] understand what is there”, in
other words, make sense of what is in the manuscript. This is then sometimes
done without thinking about probabilites. In the present case we might divide
vr.ks.e asau asu-ka-āhva; “something called life”, which fits somehow for a tree,
and in such cases commentators might produce long-winded explanations,
simply because they cannot make good sense of the wording. Since I do not
want to be accused of colonial, or worse, German Orientalist arrogance, I will
adduce a quote from an Indian, even a royal source, namely, Bhoja, who says
in his commentary on the Yogasūtras:
2 Reynolds and Wilson: Scribes and Scholars, p. 234. 3 For details of sources, see my
The work starts with a standard description of the morning schedule, and
as soon as the mūlamantra of Tripurasundarı̄ is mentioned we know that
this is a handbook for the Śrı̄vidyā ritual. Then follows the Nāthastotra:
namas te nātha bhagavan śivāya gururūpin.e [. . .] and so forth. All these
elements occur also literally in the Śyāmāpaddhati and presumably also a
larger number of similar works. The Śyāmāpaddhati has the advantage that
it is more detailed and allows us to understand the different parts of the
ritual much clearer. The Śyāmāpaddhati also mentions optional practices,
more intensive and time-consuming as well as simpler ones. With all this
intertextuality and fluidity, it was difficult to maintain the ascription of the
In this way the manuscripts often contain a second version of the text, which
goes far beyond normal variation, it is a systematic re-writing for practical
reasons. Another feature of the work is that some rituals are given even in one
manuscript in different variations, or there are options for the performance
according to the intensity of one’s religious life. In one instance a multitude of
versions of one ritual are enlisted:
mahāmāheśvarācāryavaryacaryātiviśrutaih.
śrı̄matsāhibakaulākhye sthitimadbhih. parāśrame
śrı̄mitrānandanāthāya svātmajāya vinirmitā
Is this claim at all realistic? Again we may ask “which text was written for his
son”.
There are some passages in the text that suggest that this ascription may
be genuine. These are the instances where the practitioner has to address his
own guru in ritual and adapt the mantras accordingly. For the present purpose
not much context for the following quotation is needed: In the manual, after
the completion of one ritual action, the adept has to recite the mūla-mantra
of the Śrı̄vidyā, then follows the passage under consideration, where the adept
has to worship the sandals of his Guru.
om. aim
. hrı̄m
. śrı̄m
. hasakhaphrem
. /
hasaraks.amalavaraya ūm . /
sahakhaphrem . sahaks.amalavaraya ūm
. /
hsaum . h. shaum. h. śrı̄macchrı̄vidyādharakaulānandanāthaśrı̄pādu-
kām
. śrı̄bhavānyām . bāśrı̄pādukām
. pūjayāmi namah. /
iti daśadhā vimr.śya manasā dan.d.apran.āmam
. kuryāt /
An editor publishing the text from one manuscript would not have to change
anything. But let us look at the second paragraph in another manuscript:
Now one realizes that Bhavānı̄ was not a name of Pārvatı̄, it was rather, like
Vidyādhara a personal name. If we know that initiation names for Śrı̄vidyā
initiates end in -ānandanātha for men and deduce from the text that those of
the spouses or tantric consorts end in -ambā, the text gives the impression
that it was a personal copy of someone whose tantric gurus bore those names.
It was in other words an individualized prayer book. Naturally every such
personal copy had to differ.
Before asking how one should edit such a text, we might first ask how
such a text was copied for someone else. In a living tradition reproducing
individual names of Gurus would not make any sense unless one’s guru’s name
was also Prakāśānanda Kaul. One would have to indicate that this is to be filled
in with one’s own data. In one manuscript there is exactly such a correction
and the corrected text reads as follows:
Coming to the end of these lectures I entertain the hope to have answered
the question, whether producing first editions makes a difference. Even the
brief overview on Kashmirian texts having been produced or currently being
edited in Marburg has displayed a variety of subjects and a considerable time
frame. The following is a chronological overview.
5. The late nineteenth-century authors Sāhib Rām and his son Dāmodara
both worked on continuing the genre of Kashmirian historiography.
Sāhib Rām’s historiographic materials have been edited in a project
just completed by Anett Krause, Dāmodara’s “fifth Rājataraṅgin.ı̄” was
edited by Bidur Bhattarai.
7 See Jürgen Hanneder, Stanislav Jager and Alexis Sanderson: Ratnakan.t.has Stotras.
Sūryastutirahasya, Sūryaśataka und Śambhukr.pāmanoharastava. München: Kirchheim Verlag
2012 (Indologica Marpurgensia 5).
6. A study of Sāhib Rām’s Vı̄raratnaśekharaśikhā, an adaptation of a Per-
sian “nı̄ti” text, provides insights not only into the reception of Persian
works in a Sanskrit environment in the late nineteenth century, but
also into Indo-Persian translation studies and the cultural history of
multilingual Kashmir.
So, does it make a difference to the big picture whether these works are known
or not? I hope the lectures have given the impression that it does, no more
was possible in such a summary.
The demonstration that some first editions were so deficient that they
should better not be used highlights one other aspect of the examples on
which these lectures were based. It makes one painfully aware that the time
of the great Kashmirian editors is long gone, and after the destruction of
the Kashmirian Pandit culture in the 1990s, when their houses and private
libraries were burned down, it has become only too obvious that for the
case of Kashmir we lack local Sanskrit scholars who might be able to explain
to us many details, from geography to customs. Having local editors may
not be particularly relevant for editing texts from, for instance, early Nyāya
or Buddhism, but for early modern literature in a specific area, one would
have often appreciated a helping hand. In the case of Kashmir things are
therefore more difficult, but—as I have also tried to demonstrate—with good
philological practice not all is lost.
Further Reading on Editing
European Philology
Hilarius Emonds: Zweite Auflage im Altertum. Leipzig 1941.
Hermann Fränkel: Einleitung zur kritischen Ausgabe der Argonautika des Appolo-
nius. Göttingen 1964.
A. E. Housman: “The Application of Thought to Textual Criticism”. In: The Classical
Papers of A. E. Housman. Collected and Edited by J. Diggle and F. R. D. Goodyear.
Cambridge: University Press 1972, p. 1058–69.
Hermann Kantorowitz: Einführung in die Textkritik. Leipzig 1921.
Paul Maas: Textkritik. Leipzig 4 1960 (1 1927). English translation by Barbara
Flower. Oxford 1958.
Jerome McGann: A Critique of Modern Textual Criticism. Chicago 1983.
Michael D. Reeve: “Eliminatio Codicum Descriptorum: A Methodological Prob-
lem”. In: Editing Greek and Latin Texts. New York 1989, p. 1–35.
L. D. Reynolds and N. G. Wilson: Scribes and Scholars. Oxford 1991.
L. D. Reynolds (ed.): Texts and Transmission. Oxford 1983.
Hans Gert Roloff (ed.): Geschichte der Editionsverfahren vom Altertum bis zur
Gegenwart im Überblick. Berlin 2003.
P. L. Schmidt: “Lachmann’s Method: On the History of a Misunderstanding”. In:
The Uses of Greek and Latin. London 1988.
Otto Stählin: Editionstechnik. Ratschläge für die Anlage textkritischer Ausgaben.
Leipzig 1914.
Sebastiano Timpanaro: La Genesi del metodo del Lachmann. Torino 2003. German
translation (“Die Entstehung der Lachmannschen Methode”) Hamburg 1971.
Klara Vanek: Ars corrigendi in der frühen Neuzeit. Berlin 2007.
Martin L. West: Textual Criticism and Editorial Technique. Stuttgart 1973.
Indian Philology
R. Adriaensen, H. T. Bakker and H. Isaacson (eds.): The Skandapurān.a 1.
Adhyāyas 1–25. Groningen 1998.
Ram Shankar Bhattacharya: “Use of manuscripts in textual criticism by our
commentators.” In: V. V. Dvivedi, J. Pandeya (ed.): Sampādana ke Siddhānta
aur Upādāna. Samyak-Vāk Series 5 (1990), p. 200–221.
Gérard Colas: “The Criticism and Transmission of Texts in Classical India”. In:
Diogenes 47 (1999), p. 30–43. French version: “Critique et transmission des
textes dans la littérature sanskrite”. In: L. Chiard and Chr. Jacob: Des Alexan-
dries I. Du livre au texte. Paris 2001.
Gérard Colas: “Relecture et techniques de correction dans les manuscrits indiens”.
In: Christian Jacob: Lieux de Savoir 2. Les mains de l’intellect. Paris 2011.
Michael Coulson: A Critical Edition of the Mālatı̄mādhava. Delhi 1989.
Franklin Edgerton: The Panchatantra reconstructed. New York 1924.
Venkatesh Laxman Joshi: Praud.ha Manoramā with Commentary Śabdaratna.
Vol. 1. (Deccan College Monograph Series 31). Poona 1966.
Jürgen Hanneder and Philipp A. Maas (eds.): Text Genealogy, Textual Criticism
and Editorial Technique. [=WZKS LII–LIII (2009–2010)].
Oskar von Hinüber: “Der vernachlässigte Wortlaut. Die Problematik der Her-
ausgabe buddhistischer Sanskrit-Texte.” In: Kurt Gärtner (ed.): Zur Über-
lieferung, Kritik und Edition alter und neuerer Texte. Mainz 2000.
S. M. Katre: Introduction to Indian Textual Criticism. Bombay 1941.
D. D. Kosambi: Epigrams attributed to Bhartr.hari. Bombay 1948.
Patrick Olivelle: “Unfaithful Transmitters. Philological Criticism and Critical
Editions of the Upanis.ads”. In: Journal of Indian Philosophy 26 (1998), p. 173–187.
Wilhelm Rau: Bhartr.haris Vākyapadı̄ya. Wiesbaden 1977.
Srinivasa Ayya Srinivasan: Vācaspatimiśras Tattvakaumudı̄. Ein Beitrag zur Text-
kritik bei kontaminierter Überlieferung. Hamburg 1967.