revisiting śakuntalā 1

The Inarticulate Nymph and the Eloquent King

When Kālidāsa is celebrated as a Sanskrit poet it is easy to forget that a substantial portion of the
Abhijñānaśakuntalā1 is written in Middle Indo-Aryan or Prakrit languages.2 Any attempt to transpose such
a work, originating in a simultaneously multilingual and polyglossic3 cultural paradigm (the Gupta
empire in Kālidāsa’s time), into theoretical abstractions deriving from a much more (though not entirely)
monolingual and monoglossic culture (academic modernity), is beset with conceptual and practical
problems at nearly every step. The particular issue investigated in the present paper is the tension
between expression that is saṃskṛta, refined or “articulate,” and expression that is vibhraṣṭa, gone astray or
“inarticulate.” By focussing our attention and theorization primarily on the seemingly more glamorous
rhetorical power of Sanskrit we are in danger of underestimating the capacity of the inarticulate to
engender innovation and change. Unfortunately, the role of the inarticulate in either Sanskrit or Prakrit
literature has so far not, to my knowledge, been the primary subject of any scholarly work. This is
regrettable, for in her study of inarticulate language usage in the European Renaissance, a period where
eloquence was valued in ways comparable to the Gupta period Kālidāsa wrote in, Mazzio (2008) has
shown that there, rhetorical brilliance had an undeniable corollary that must be admitted into a wider,
more inclusive definition of expressive language.

Since these problems are complex, it may be helpful briefly to rehearse some of the terminology
involved. One of the most pervasive mischaracterizations of the relationship between Sanskrit and
Prakrit seeks to present a simple dichotomy between a moribund high language and a living low
language. This is simultaneously too simplistic and too misleading to be a meaningful analytical
abstraction. Firstly, just as there are many Prakrits,—the three most important Prakrits in drama are
Māhārāṣṭrī, Śaurasenī, and Māgadhī—, there are also many Sanskrits: Vaidika, Pāṇinian or “Classical”
Sanskrit with many styles (scientific, various kāvya rītis etc.), Pāṇini’s “Bhāṣā,” Epic Sanskrit, Buddhist
Hybrid Sanskrit, Ārṣa Sanskrit, Aiśa Sanskrit, Buddhist Tantric Sanskrit (yogabhāṣā, yoginīmlecchabhāṣā, “the
barbaric language of the Yoginīs”), colloquial Sanskrit,4 etc.5 Secondly, we need to consider that both
Sanskrit and Prakrit exist somewhere between the two poles of the “actual” and the “normative”. Here
the degree to which Sanskrit was, can be, or is even now in the present “actual” is heavily disputed,
though further progress can be made as more empirical data becomes available.6 When the Prakrits (for
example inscriptional Prakrits) are treated as “actual” languages, they are sometimes named aer the
regions they reputedly originated in or where they were commonly used in, such as Gāndhārī Prakrit.
This actuality, furthermore, can admit of the two modifying categories of dialect and idiolect, whereas
the normativity is presented to us in grammars and lexicons as stereotypical, in the sense that normative

1 The exact form and interpretation of the title is disputed, see most recently Levitt (2005).
2More specifically we are concerned with the middle phase of the long (4th cent BC – 12th cent AD) Middle Indo-Aryan
period (4th cent BC – 12th cent AD), sometimes called “Middle Middle Indo-Aryan.”
3Pollock (2006:50) goes even further and characterizes the situation as one of “hyperglossia.” On the long history of diglossia
and bilinguality in India see Hock & Pandharipande (1976).
4 See Wezler (1996).
5Of these, the last three have so far received the least scholarly attention, see Szanto (2008) for a successful edition of such
language.
6Definitions of “mother tongue,” “parent tongue,” or “most commonly spoken language,” are notoriously difficult in Indian
contexts. Most recently, see the insightful discussion in Hanneder (2009:211–213).
 revisiting śakuntalā 2

usage can be labelled as correct or incorrect.7 For example, the Māhārāṣṭrian songs in the
Abhijñānaśakuntalā would have been composed with at least some intent to reflect a normative language of
the stage, and Kālidāsa presumably observed something resembling the stylistic strictures of Nāŀaśāstra
chapter 17—for recited (pāṭhya) Prakrit that is nāŀadharmī (“artistically drama-normative”)—and chapter
32—for sung (geya) Prakrit—, and Prakrit grammars such as Vararuci’s Prākṛtaprakāśa if he knew it, the
date of its author Vararuci being still uncertain. By contrast, the core verses of a Prakrit work such as the
Sattasaī (or Gāhākosa) of Hāla Sātavāhana, an anthology of early Māhārāṣṭrī Prakrit songs, would have been
composed in a language much closer to actual Māhārāṣṭrī Prakrit dialects and idiolects of perhaps the
second to fourth centuries AD. We end up talking, in this example, of two languages that are conceived
of in very different terms but that bear the same name: Māhārāṣṭrī Prakrit. The practical problem
associated with this situation is philological, for even a work composed in an actual language would
eventually undergo normative redaction at the hands of commentators, scribes and redactors who
transmitted it over the centuries in different geographical regions and cultural contexts, giving rise to
numerous divergent recensions. Even the very idea of what normative Māhārāṣṭrī Prakrit is, is then
contingent on place, time and a complex set of subjective authorial motives.

The source of much of the Sanskrit śāstric terminology used to describe normative Prakrit stage
languages is the Nāŀaśāstra (before 400 AD?). While providing a general definition of spoken or recited
Prakrit, the Nāŀaśāstra8 stigmatises it as “chaotic” (viparyasta), as “lacking the virtue of
refinement” (saṃskāraguṇavarjita),9 and divides it into three types: 1. words that are morphologically
identical to Sanskrit words (samānaśabda or the later category of tatsama), 2. words derived from Sanskrit
(vibhraṣṭa, “fallen,” or the later category of tadbhava),10 and 3. words of local origin (deśī). The relationship
between these categories is again subject to dispute. Disagreement lies firstly in the implications of the
two possible taddhita derivations of the term prākṛta from prakṛti either by Pāṇini 4.3.74: tata āgataḥ (“come
from X”) or 4.3.53: tatra bhavaḥ (“located in X”), and secondly in the exact nature of the prakṛti, or “source”
either as Sanskrit or something else. That the rival derivations have real hermeneutic consequences is

7 There are commentators, most notably Ghanaśyāma of Tanjore, who will censure even the most celebrated of poets for using
such “incorrect” Prakrit forms that deviate from received grammars. See Chadhuri (1943:3–5, note).
8 Nāŀaśāstra 17.1–3.
9 But see below for Abhinavagupta’s different interpretation of this compound.
10Abhinavagupta, commenting on the Nāŀaśāstra, gives a general derivation of the word prākṛta as follows: saṃskṛtam eva
saṃskāraguṇena yatnena parirakṣārūpeṇa varjitaṃ prākṛtam, prakṛter asaṃskārarūpāyā āgatam, “Sanskrit itself, devoid of the quality of
saṃskāra, which is an effort that has the nature of protection, is Prakrit. It is derived from the source that lacks saṃskāra.”
Protection (parirakṣā), is here a reference to a frequently invoked passage of the Mahābhā﬇a on the purpose of grammar:
rakṣohāgamalaghvasandehāḥ prayojanam.
 revisiting śakuntalā 3

demonstrated by Kahrs (1992).11 Subsequently, 12 the Nāŀaśāstra divides this dramatic, spoken language13
into four types according to the speaker: 1. atibhāṣā, “transcendent language” used by Gods, 2. āryabhāṣā,
“noble language” used by kings, 3. jātibhāṣā, colloquial “type-language,”14 and 4. yonyantarībhāṣā, the
“non-human language” of wild and domestic animals and birds.15 For the commentator Abhinavagupta
atibhāṣā and āryabhāṣā seem to differ in the degree of their saṃskāra: atibhāṣā retains Vedic accents, āryabhāṣā
retains merely many Vedic words. Both (?) are universal languages, they exist in all seven continents.16
The jātibhāṣās, as type or class languages, on the other hand, which can be either Sanskrit or Prakrit
depending on context and status, may be justifiably identified as stage sociolects, and they are in the
Nāŀaśāstra indeed also associated with the four classes of society, the caturvarṇa.17 They appear to form a
stratum of dramatic colloquial languages that may contain “barbarian” (mleccha) usages and are confined
to India. Other classes of society not included in the caturvarṇa scheme are denied their own voice in
Sanskrit drama.18 This initial setup of four languages is complicated by a secondary set of prescriptions
relating to languages of regional origin (bhāṣā) and ethnic dialects (vibhāṣā) (see Pollock (2006:93-4)). The
seven bhāṣās or deśabhāṣās (so Abhinavagupta)19 are: 1. Māgadhī, 2. Avantī, 3. Prācyā, 4. Śaurasenī, 5.
Ardhamāgadhī, 6. Bāhlīkā, 7. Dakṣīṇātyā.20 The seven vibhāṣās are 1. Śakāra, 2. Ābhīra, 3. Caṇḍāla, 4.
Śabara, 5. Dramila, 6. Āndhra, 7. Vanacārabhāṣā.21 This appears to contradict what was stated earlier, for

11In later grammatical and commentarial literature the category deśī is usually reserved for nomina alone, and the term dhātvādeśa
(“root-suppletion”) is used for verbal forms. Even though an origin in provincialisms may be assumed, the śāstric category of
deśī has a greater universality and therefore must not be conflated with vernacular language, see Pollock (2006). For example,
the 11th cent, commentator Tribhuvanapāla uses the expression lokaprasiddhi to distinguish such provincial usage from deśī (see
Vasudeva & Chiarucci, NGMCP newsletter 7, forthcoming). As Jacobi (1886:xiii) puts it: “Die deçî-Worte sind nicht, wie man
aus dem Namen schließen könnte, Provinzialismen, wenn sie es auch ursprünglich gewesen sein mögen; wenigstens werden
diese von Hemacandra (D. K. 1, 4) ausdrücklich als nicht zu den deçî gehörend bezeichnet. Denn die deçî müssen allgemeine
Gültigkeit haben, wie denn in der That viele derselben auch in anderen Dialekten nachgewiesen werden können.”
12 Nāŀaśāstra 17.26cd-28ab. See Ghosh (1951:326-327), Nitti-Dolci (1972:75).
13 That is, language that is both pāṭhya, “spoken” and not sung, and nāŀadharmī “artistically drama-normative” as opposed to
lokadharmī, “conventionally drama-normative.” On the terms nāŀadharmī and lokadharmī, both of which articulate different
types of of dramatic normativity, see Raghavan (1933, 1934). These strictures serve to exclude from the current discussion a
special complex of musicological issues when Sanskrit and Prakrit are sung. It is noteworthy that Nāŀaśāstra 17.26ab further
explicitly introduces this section with the comment ata ūrdhvaṃ pravak﬇āmi deśabhāṣāvikalpanam. One could simply take this to
mean that Sanskrit (even that assigned to stage representations of Gods) and Prakrit were in some sense considered languages
of place by the redactor of the Nāŀaśāstra. This is the interpretation of Nitti-Dolci (1972:67,75) and Ghosh (1951:326), but
against this see Pollock (2006:189). Alternatively, the compound deśabhāṣāvikalpanaṃ could be analyzed as deśaś ca bhāṣā ca tayor
vikalpanam, or more likely, the reference ata ūrdhvaṃ could point to a later section such as 17.47.
14Ghosh (1951:327) notes that Bhoja in his Śṛṅgāraprakāśa reinterprets atibhāṣā as śrauta (Vedic), āryabhāṣā as ārṣa, and jātibhāṣā as
laukika or “colloquial” language.
15Abhinavagupta explains: paśupakṣiprabhṛtīnāṃ yad rutaṃ tan naŀaprayoge kutrāpy avasare saṃbhāvyam ity āha naŀadharmīti, “He
says ‘artistically drama-normative’ because the sound of animals and birds etc. is in certain dramatic contexts possible.”
Compare here Hanumān’s “choice of language” conundrum when he has to address Sītā as a stranger in human language and
does not wish to frighten her, as analyzed by Goldman (2000:87-89).
16 Reading with Ghosh and MS Bha saptadvīpapratiṣṭhitā.
17 Nāŀaśāstra 17.31cd–32ab: jātibhāṣāśrayaṃ pāṭhyaṃ dvividhaṃ samudāhṛtam / prākṛtaṃ saṃskṛtaṃ caiva cāturvarľasamāśrayam.
18 Nāŀaśāstra 17.46-47.
19 Abhinavagupta explicitly glosses these bhāṣās as “languages of place”: deśabhāṣāṃ saṃkṣipyāha māgadhīm ityādi.
20 Nāŀaśāstra 17.49.
21 Nāŀaśāstra 17.50.
 revisiting śakuntalā 4

Dravidian (Dramila) is forbidden as a class language but permitted as an ethnic dialect. In this scheme
too, the vibhāṣā dialects are indicative of lower social status than the bhāṣā languages.22 There are many
other, later, theorizations that associate similar and more extensive catalogues of languages and dialects
with social groups ranked by status, the Prakrits being usually assigned to the lowest (nīca) social classes
of dramatic representation.

When confronted with this complexity, Ghosh (1967:lvii) asks some obvious questions: “What do all
these things signify? How can we believe that the author of the Nāŀaśāstra could have treated the subject
in such a glaring disregard of logic and common sense?”23 There are no easy answers to these questions.
A sober philologer must retain a skeptical view of any broader theorization here: the redactional history
of the Nāŀaśāstra is still far too unclear to warrant more than provisional hypotheses. Nevertheless, it is
certainly not implausible that the fourfold classification of stage language in 17.26cd-28ab might be a
later interpolation. More careful philological work like that of Srinivasan (1981) is needed to determine
hypothetical stratifications and eventually provide access to the Nāŀaśāstra as the “text of a work” rather
than just a “text of a document,” to use the terminology of Tanselle (1992).

Another vexed question is the degree to which this theoretical framework impacted Kālidāsa’s practice.
Goodall & Isaacson (2006:introduction) have demonstrated that Kālidāsa’s śiṣṭaprayoga usage did not
slavishly follow the rules of grammar and rhetorics that the later commentarial tradition would try to
impose. Would it therefore not seem equally plausible that his attitude to dramatic theory was
comparable?

Can then the hierarchical gradations of the Nāŀaśāstra’s normative, dramatic jātibhāṣā languages be read
into the Abhijñānaśakuntalā? In the Abhijñānaśakuntalā it is certainly the case that Māgadhī Prakrit, spoken
by the fisherman and the policemen etc., appears as a socially stigmatising lowest pole. But it is also true
that the situation of the regional, actual, source languages is more difficult to assess, and the attested
usage of Buddhist and Jain monks may be cited as obvious instances of high status speakers preferring
such Eastern Prakritic or Magadhan languages.24 Mutual intelligibility in a drama moreover demands
that all characters are shown to possess at least a passive knowledge of all of these stage languages. A
failure to comprehend, especially where Prakrit is concerned, can even be construed as comical. For
example, when the buffoon and the king listen to Queen Haṃsapadikā’s Prakrit song in the
Abhijñānaśakuntalā, it is the Prakrit speaking buffoon who does not understand the suggestive significance
of the song, while the Sanskrit speaking king does. Of course, even the king fails to perceive the far-
reaching, subtle suggestions directed at the spectator. The use of inarticulate Prakrit for comical
characters is moreover not inevitably a sign of low social standing. In the Pādatāḍitaka of Śyāmilaka a
certain Bhadrāyudha is portrayed as ridiculous not merely because of his curious appearance—he wears
miniature wooden jugs as earrings—and his bizarre manner, but especially because of his strange accent.
“Accosting people with ‘ja’ sounds, he seems to be mocking the Lāṭas,” he “assaults everyone he meets

22Locanā commentary to Sāhityadarpaṇa 6.288, citing the Bhā﬇ārṇava: hīnair bhā﬇ā vibhāṣā syāt sā ca saptavidhā smṛtā, “Lowly
characters should use vibhāṣā dialect, it is of seven kinds.”
23See Ghosh (1967:lviii–lxiii) for an attempt to explain these discrepancies as evidence for the composite character of the
Nāŀaśāstra, the scheme of seven bhāṣās are in this view an artifact of the redactors’ accommodating a linguistic landscape
altered by the diachroneity of language.
24For the special set of issues associated with the transmission of such actual languages in Buddhist literature, e.g. from
Eastern Prakrit to Western Prakrit to Gāndhārī to Buddhist Hybrid Sanskrit etc., see Norman (2006).
 revisiting śakuntalā 5

with ‘śa’ sounds,” and then, “placing his hands on his chest in the pigeon-gesture, yelling ‘jaja’ without
‘ya’ sounds, ... he promenades as if he were touching mud with his fingers.” Evidently this parodies
speakers of the Western Prakrits that use ja in place of ya. Despite this, it is highly probable that this
Bhadrāyudha is actually the Pratīhāra of King Vikramāditya of Ujjain.25

In the Abhijñānaśakuntalā, moreover, it is oen the most memorable episodes that are set in contexts
where Prakrit languages predominate. Kālidāsa’s Prakrits are rich in speech artifacts and manifest a
superabundance of particles (and idiomatic compounds thereof) such as avvo, vi, avi, khu, cea, ccia, piva,
miva, ṇāma, yyeva. No doubt this is intended to evoke a homely rusticity (it can be misleading to translate
grāmya as “vulgar” in this context). Some of these modern editors have erased as redundant, even though
they are present in the manuscript transmissions.26 This erasure is even more common in translation.
Sometimes it is admittedly fully justifiable, for example when particles can be diagnosed as no more than
versefillers. Sanskrit grammatical and literary theory judges this to be an issue stemming from the
comparatively great lexical polyvalence of particles and hence responds with the production of
descriptive lexical works, such as the Gaṇaratnamahodadhi of Vardhamāṇa (Eggeling (1879)), that list
multiple meanings for each particle and exemplify standard usage in literary works. One issue of concern
here is that Prakrit passages where such particles function as glossolalia are almost impossible to translate
without introducing some sort of a second linguistic or emotive register that differs from that used to
translate Sanskrit passages lacking such forms.

In light of these problems, how are we to understand these differing registers of poetic expression, and
how would a reader in ancient or medieval India have perceived them? Let me provide a couple of
examples to illustrate the issues, the first taken from the late twelh century AD, from the Sanskrit
Āryāsaptaśatī of the Gauḍa poet Govardhana, a high status brahmin court poet who worked at the court
of the Lakṣmaṇasena. This is a Sanskrit counterpart or reworking of a Māhārāṣṭrī Prakrit work that
predated it by a thousand years, the Sattasaī or Gāhākosa of Hāla Sātavāhana mentioned above. When
Govardhana justifies his work he tells us incidentally a lot about his attitude to Prakrit.27

Speech with the poetic sentiments associated with Prakrit

: that has sentiments relished by vulgar people
has been by great effort
turned into Sanskrit
: brought to the sophisticated man as by Balarāma
the Yamunā river28 whose water was flowing downwards was raised to the sky.
(transl. F. Hardy)

As Hardy points out, Govardhana reveals that he thinks Prakrit is appreciated by vulgar people (prākṛta,
the commoners), that refining it required great effort (balena), that Prakrit poetry runs
“downward” (nimna) like a river, and that in so doing, he, Govardhana himself, is acting rashly like the
drunk Balarāma (bala) and doing violence to the Sanskrit language. A further suggestion is that Prakrit is

25 See Dezsō & Vasudeva (2008:69–71).

26 See especially Steiner (1997:167-208) & (2001).
27 Āryāsaptaśatī 52: vāṇī prākṛtasamucitarasā Balenaiva saṃskṛtaṃ nītā / nimnānurūpanīrā Kalindakanyeva gaganatalam //.
28 An allusion to the Yamunā story of the Bhāgavatapurāṇa 10.65.
 revisiting śakuntalā 6

like the murky, dark Yamunā and Sanskrit is like the pure, celestial Gaṅgā. In view of the Nāŀaśāstra
passages adduced above this attitude might be unsurprising. However, a counter-valuation can be found
in a number of Gaṇikā plays that predate Kālidāsa. In Śūdraka’s Mṛcchakaṭikā (ca. 200 AD) the buffoon
sees the roles of Sanskrit and Prakrit as follows:29

Two things make me laugh: a woman reciting Sanskrit and a man singing a so and sweet Kākalī
song. Firstly, a woman reciting in Sanskrit makes too many ‘su-su’ sounds like a young cow who has
just been given a nose-rope. A man singing a Kākalī30 song is like an aged priest wearing a garland
of dried up flowers who mumbles Mantras, and definitely not to my liking.

This appears, at first sight, to present a view of gendered, maybe even biologically inherent capacity or
incapacity for Sanskrit recitation and sweet singing respectively. But the intended predication seems to
me to be the other way round: Sanskrit is too harsh for so voiced women, Kākalī songs are too sweet
for harsh voiced men. It actually extolls Prakrit as sweeter than Sanskrit. But we must bear in mind that
this is the voice of a Brahmin speaking Prakrit, as such it would be deviant and not normative, to use our
earlier categories. But similar judgments can be found placed into the mouth of Sanskrit speakers too. In
Īśvaradatta’s Dhūrtaviṭasaṃvāda, “Rogue and Pimp Confer,” a Bhāṇa dating to the Gupta era (?), it is the
great sanctity conferred upon Sanskrit that has been construed as a negative trait. As part of a long
impassioned speech deriding the pleasures of heaven against simple worldly joys we hear that women
speaking in Sanskrit were perceived as a turn-off:

Ah! I’d rather bed down with senile bigots than with heavenly nymphs. Scriptures teach that they
are ancient, speak Sanskrit, and are majestic. They gave birth to the great sages headed by Vasiṣṭha
and Agastya: who could let himself go in their company?31

Beside the perhaps expected labeling of Sanskrit as majestic, implying that Prakrit is perhaps more
ingenuous, do such evaluations legitimize an understanding of Sanskrit dramatic expression as
“articulate” and sharp, in contrast with Prakritic modes of expression that are somehow sweet and so
but “inarticulate”? And if so can we further clarify the status of the “inarticulate” in Kālidāsa’s works?
Why did he use so much of it?

Before exploring this further let us first look at the other side of the coin and see how Kālidāsa portrays
the king as the ideal, eloquent speaker of Sanskrit. When the king first sees Śakuntalā he desires her but
fears that she is the daughter of a brahmin sage and therefore beyond his aspirations. He resolves his
doubts about an illegitimate, inter-caste marriage by uttering the following verse:32

29 Mṛcchakaṭikā act 3, p.79 vidūṣakaḥ: mama dāva duvehiṃ jjevva hassaṃ jāädi. itthiāe sakkaäṃ paṭhaṃtīe, maṇusseṇa a kāalīṃ gāaṃteṇa.
itthiā dāva sakkaäṃ paṭhaṃtī, diṇṇaṇavaṇassā via giṭṭī, ahiaṃ susuāädi. maṇusso vi kāälīṃ gāäṃto, sukkhasumaṇodāmaveṭṭido buḍḍhapurohido
via maṃtaṃ javaṃto, diḍhaṃ me ṇa roadi.
30
Nāradamāhāpurāṇa 50.57ab: kākalir dṛśyate yatra prādhānyaṃ paṃcamasya tu. Agnipurāṇa 3.359.70 ab: kākalī tu kale sūkṣme dhvanau tu
madhurāsphuṭe. Amarakoṣa 1.7.409: kākalī tu kale sūkṣme dhvanī tu madhurāsphuṭe.
31See Dezsō & Vasudeva (2008:408–9): bho māṃ prati varaṃ śrotriyair vṛddhaiḥ sahāsituṃ nāpsarobhiḥ! tās tu dīrghāyuṣmatyaḥ
saṃskṛtabhāṣiľo mahāprabhāvāś ca śrūyante. yāsu Vasiṣṭhāgastyaprabhṛtayo maharṣayaḥ samutpannās tāsu ko visrambhaḥ?
32Abhijñānaśakuntalā 1.19: asamśayaṃ kṣatraparigrahakṣamā yad evam asyām abhilāṣi me manaḥ / satāṃ hi saṃdehapadeṣu vastuṣu
pramāṇam antaḥkaraṇapravṛttayaḥ //.
 revisiting śakuntalā 7

Doubtlessly she is fit to be wed by a warrior,

since my heart desires her so.
For in matters of doubt the inclinations
of their inner faculties are authority for the good.

The rational king is here verbalizing as a parārthanumāna (“inference for another”), for the readers’
benefit, what is in fact his own internalised svārthānumāna, “inference for oneself”. Since he has made the
śāstracakṣuḥ, the “eye of science” which reveals what is beyond the range of normal knowledge, his own, he
presumably actually thinks with this paradigmatic rationality usually employed by Pramāṇa philosophers.
We can easily formalize his inference:

1. Proposition (pratijñā): Śakuntalā (pakṣa) is fit to be married by a warrior (sādhya).

2. Reason (hetu): Because my heart desires her.
3. Example (dṛṣṭānta): When there is doubt, the inner faculties of the good are trustworthy.

This satisfies the classical condition of possessing a trairūpya liṅga:

1) pakṣadharmatā: the sādhya (marriagability) is a property of the pakṣa (Śakuntalā)

2) sapakṣe sattvam: sapakṣa, congruent case = Haṃsapadikā, possesses sādhya and hetu
3) vipakṣe ‘sattvam: vipakṣa, incongruent case= the brahmin nun Gautamī, possesses neither
sādhya nor hetu

Not only is the king’s verse formulated as a correct anumāna inference, but it also is supported by
scriptural testimony, for it restates the sanction of self-satisfaction (ātmanas tuṣṭiḥ) as one of four valid
sources of dharma as taught in two verses of the Manusmṛti, an authoritative law-book.33

2.6. The whole Veda is the (first) source of the sacred law, next the tradition and the virtuous
conduct of those who know the (Veda further), also the customs of holy men, and (finally) self-
satisfaction. (translation by G. Bühler)

This association of Abhijñānaśakuntalā 1.19 with the Manusmṛti’s fourth source of dharma was already made
by the seventh century Mīmāṃsaka philosopher Kumārila, who cites the verse to demonstrate that the
pleasure principle is not invariably a source of dharma. Rather, only in the case of extraordinary
individuals whose self has become permeated by dharma through long Vedic study can self-satisfaction
be truly counted as a source of dharma.34 The king is thus presented as sharp, articulate, rational, and
competent to expound the law with a remarkable degree of philosophical precision.

33 Manusmṛti 2.6, 2.12.

34 See Davis (2007:288-9).
revisiting śakuntalā 8
 revisiting śakuntalā 9

5UbƷ[d- c]d b9ĭQř75ƫ b5ƫ_\_Ż bc6\ŵbQ< eBSPdZż5ŵb

_PSmebPm PmZdƑ]6r

 `BưĭQ]d- 9VDŋ[d_]m¯[G
c]d\ZPf< Bdř7Zŵb VdSXƑZźPŵb
_eS5\m bƫ_Ʋămr 75ƫ P_BưbƲZImŨ_Pdr 55ƫeX XĚW]Sd< 9_Ym5¯CZm
bc6\mr VŬ[ĭQfeQűeQ

 eIJ[ƫ_Sd- c]d 5PbƷ<IdPdeb ƎBƑPƑZăƫ b9ĭQ]d _PSmebPƫ 5eSŅăƫ

ı¯CeS Ƒăr

5UbƷ[d- P¯CƲ e_YdŤƑZ r

 eIJ[ƫ_Sd- ITd _PSmebPd 5PƲbeSųP VdSĺP bÌSd PmZdƑ]6 5e_

 $PEE  ķ #PE .4 4BOTL E  
 ķ #PE .4 4UFJO PS E  JJ
 ķ #VSLIBSET
CPSJ NT ķ 4ISJOBHBS BDD OP  #FE #VSLIBSET FEJUJPO

 c]d > ķķķ cř ķ  75ƫ > ķķ 75 ķ 7[ƫ ķ  b5ƫ_\1 > ķ
ŵ_5ƫ_\1 ķķ bcƫ_\1 ķ  1_Ż > ķķķ 1_źbp ķ  Q< > ķķķ QƲŏ
ķ  1PdZż5ŵb > ķ 1PdZTƞ5ŵb ķķķ  1SmebPm > ķķķķ 1SmebPf
#FE  PmZdƑ]6 > ķ P_ZdƑ]Bd ķķ P_ZdZdƑ]6 ķ  c]d > ķķ
cř ķ c] ķ  \ZPf< > ķķķ \ZPfŏ ķ  7Zŵb > ķķ QD <Sŵb ķ
7cŵb ķ 7ŵbZ ķ BD  VdSX1 > ķķķ VdSV1 ķ  1ƑZźPŵb > ķķķķ
1ƑZTƲPŵb #FE  _eS5\m > ķķķ _eS_\m ķ  bƫ_Ʋăm > ķķķ b_Ʋăm ķ
 75ƫ > ķķ 75 ķķ  1ImŨ_Pd > ķķ 1IƲŨ_PdP_ZdƑ]6 ķ IƲŨ_ƲPd
U_ZdƑ]6 ķ  55ƫeX > ķ 55ƫeV ķ 55ƫeXP_ZdƑ]6 ķ 55ƑŵZ ķ 
XĚW]Sd< > ķķķ#FE XĚS< ķ  9_Ym51 > ķ 9VYm51 ķ 95Ym51 ķ
9_cm51 ķ  1¯CZm > ķ 1®Zm ķ 1¯CŋZm ķ  bc6\m > ķķ bcm6\m
ķ bcd6\ ķ  c]d > PNķ  IdPdeb > ķķ IdPdƑZ ķ IdUdeb ķ 
ƎBƑPƑZăƫ > ķ ƎBeUƑZăƫ ķ ƎBƑõPƑZăƫ ķ eBƑĮƑZăZp ķ  1SmebPƫ > ķķ
1SmebƒP #FE 1QmebPƫ ķ 1ĊbPƫ ķ  5eSŅăƫ > ķķ 5eSZăƫ ķ 5eTŅăƫ ķ
 ı¯CeS > #FE ıƑ¯CċSm ķ IJƞƑ¯CeS ķķ  Ƒă > ķķ eQ ķ  ¯CƲ > ķ
CƲ ķķķ  e_YdŤƑZ > ķķķ#FE e_X46_ƑZ ķ  ITd > ķķķ ISd ķ
 1SmebPd > ķķķ 1SmebPf #FE 1ĊeaPd ķ  5PƲbeSųP > ķķ bbeSųP
ķķ  VdSĺP > ķķ VdSıU ķ VdSŤP ķ  bÌSd > ķķķ#FE bƫDQd
ķ  PmZdƑ]6 > #FE P_ZdƑ]6 ķķķķ  5e_ > ķķ 5eX ķķ
 revisiting śakuntalā 10

Let us now turn to the Prakrit passage that immediately precedes the king’s display of Sanskritic
rationality to see how Śakuntalā and her friends are portrayed as speakers of Prakrit. Image 1 shows a
page from my forthcoming edition of the Kashmirian recension, based on five Śāradā manuscripts.35 The
apparatus covers more than half of the page, a considerably larger portion than it would in a comparable
passage of Sanskrit, demonstrating graphically the uncertainty of the scribal transmission of Kālidāsa’s
Prakrit in Kashmir. Image 2 shows the Shrinagar manuscript (acc. no. 1435) with its partial and sometimes
incorrect Sanskrit chāyā for the Prakrit passages. So far this recension has not been available in a reliable
form (nor do the other recensions fare much better), since Burkhardt published it from what he thought
was a codex unicus. The passage contains a well known episode where Śakuntalā, Anasūyā and
Priyaṃvadā are watering trees when they spot a jasmine vine and a mango tree that they look upon as a
married couple.36 Śakuntalā is teased about her keen interest in this tree-vine couple, and Priyaṃvadā
insinuates that Śakuntalā secretly also yearns to be married.

This passage shows Priyaṃvadā produce a quasi-inference in some ways comparable to that of the king.
It is not a formal Sanskritic naiyāyayika or pramāṇa type of inference for another (parārthānumāna) though,
rather it is a less articulate, vague poetic intuition (pratibhā), a guess of someone else’s hidden yearning.
Priyaṃvadā has intuited Śakuntalā’s secret—a secret even Śakuntalā herself may not be fully cognizant of
—but it would be nonsensical if transposed into a formal anumāna syllogism as we did before. That
Prakrit is the language par excellence for such suggestivity is borne out by its reception as such in much of
later Sanskrit rhetorical literature, one might even see Ānandavardhana’s radical theory of suggestion
(dhvani) as not simply inspired by Prakrit gāthās, but rather as an attempt to explain what can be an
initially unsettling, temporary aesthetic aphasia induced by reading such Prakrit lines, as one seeks
answers to questions such as: How exactly does Priyaṃvadā make her guess? Why do we follow her so
easily? Why do we enjoy it so much? The less sharply focussed and less articulate Prakrit has in this case
proven itself to be more productive for aesthetic theory than the precise rhetorical eloquence of the
king’s speech that rather invites comparison with syllogistic reasoning, and analysis with the system of
figures of speech elaborated by the ālaṅkārikas.

Now here we run into a philological issue that can serve to further illustrate how the problems of
practicality alluded to above can impact interpretation. In the Devanāgarī recension Priyaṃvadā asks
Anasūyā if she knows why Śakuntalā is staring at Vanajyotsnā (fem.), “moonlight of the forest,” as she has
named the navamālikā jasmine vine. This is so well known that one hardly thinks about it, but it is
curious. The scene does not take place at night, nor is the forest described as being very dark, and
metaphors or similes combining jasmine and moonlight are actually quite rare. When they do occur they
rather involve jasmine petals strewn on the ground or night-flowering jasmine. In the Bengal recension
the vine is simply called Vanatoṣiṇī (fem.), “delighter of the forest,” with no implications of nighttime. In
the Kashmirian recension we have a very different situation. Śakuntalā is not focussing her attention on
the jasmine vine at all, but rather at the mango tree. It is this “male” tree that she has named, Vaṇadosī
(masc., = Skt. Vanatoṣin), “delighter of the forest.” Evidently this must have changed the perceived

35Ś1 = Bodleian MS sansk. d. 88 (5); Ś2 = Bodleian MS Stein. or. d. 42(ii); Ś3 = Bodleian MS sansk. d. 87; Ś4 = the Puṇe MS
published by Burkhard; Ś5 = Shrinagar acc. no. 1435. The testimony of Ś3 is not available for the passage in question here. Of
these manuscripts Ś1-3 were partially used by Kanjilal (1980) for his attempted reconstruction, which as Wright (1999:531 note
10) notes is largely congruent with Pischel’s second edition.
36See Pieruccini (2005) for this motif. She has traced the image of the vine embracing the tree in a range of kāvyas and nāṭakas
as far back as Aśvaghoṣa’s Saundarananda, the Mahābhārata, the Ṛgveda, and the Atharvaveda.
 revisiting śakuntalā 11

character of Śakuntalā, she is less a self-absorbed young woman looking at the vine who represents a
mirror image of herself, but at the tree, who represents her future husband. Recall here that she also had
a male deer that she adopted as her son, mirroring her future son Bharata. I personally believe that both
the Devanāgarī and the Kashmirian readings can in this case be explained as corruptions of the Bengal
readings (jo for do). That these are true “recensions” and not just “versions” is evident from the fact that
the redactor had to make this change consciously throughout the text wherever the name Vaṇadosī recurred
(in image 2 we see three such occurrences).

Such variation should not come as a great surprise. The manuscript transmission of secular Prakrit, for
want of a better word, is usually quite poor, particularly for texts that are transmitted without a Sanskrit
commentary or at least a chāyā paraphrase. Bhayani (1977), introducing his work on the Bhāṣālakṣaṇa
chapter of the Gītālaṅkāra, that lists and exemplifies 42 varieties Prakrit, voices a complaint that could be
repeated, at least in part, for many other Prakrit works.

The text as it stands appears mostly chaotic. On the face of it, it does not seem possible to make
out any connected sense even from single lines or their portions. Under the circumstances, any
attempt to make word-division is also in danger of being considered nothing but futile
guesswork.

In the face of this various methods of editing Prakrit texts have been attempted, see most recently Steiner
(2006). For example, the Bengal recension mentioned above was edited by Richard Pischel who
disregarded the manuscript evidence as unreliable, and standardized according to the rules of much
later, Eastern Prakrit grammarians. For the Kashmirian recension I have weighted the manuscript
readings more heavily, while I have simultaneously attempted to derive morphological patterns or rules
from readings found in the earliest witnesses. See Dezső (2007:12–15) for a similar method applied to
the Prakrit passages of the Kundamālā of Dhīranāga. For the Kashmirian text these are comparatively late,
namely the 17th cent. Nevertheless, the conservative nature of the Kashmirian scribal tradition has meant
that the Kashmirian recension might reach back nearly a thousand years. We also have explicit statements
from Abhinavagupta concerning which Prakrit grammar was in vogue in Kashmir a thousand years ago.

But let us get back to the problem of the registers of language and reconsider the issue of mutual
intelligibility. As should be abundantly clear now, the artificial stage languages do not simply mirror the
linguistic and political realities of the usage and status of the actual languages common in the Gupta
empire. For a majority readership the dramatic situation would have presented an inversion of relative
intelligibility, Prakritic stage diction actually being more intelligible because it was closer to a
linguistically imperfect lingua franca. Full, authentic membership in the Sanskrit cosmopolis might have
been for large sections of the population at best aspirational. Let me exemplify this with a final example,
a short passage from the Padmaprābhṛtaka of Śūdraka, composed probably just before Gupta era. Here we
encounter Dattakalaśi, a character who ought to epitomize the mastery of language familiar from
brahminical self-representations, a Pāṇinian scholar of Sanskrit, the foremost of the learned.37 In the

37 Dhvanyāloka: prathame hi vidvāṃso vaiyākaraṇāḥ, vyākaraṇamūlatvāt sarvavidyānām.

 revisiting śakuntalā 12

Padmaprābhṛtaka this grammarian is satirized as a rather pitiful character, for nobody can fully understand
his hyper-correct Sanskrit that is full of obscure forms.38

What are you saying? “Art thou hale?”39 Good grief! …

What are you saying? “Whither art thou desirous of precipitating thyself?40 Rest awhile.
Why art thou put to flight?”41 Ah! Have mercy! Please do not strike me with such word-
thunderbolts, harsh with painful blows. Speak to me in plain language. I am not conversant with
the speech-vice of you grammarians, a dose of poison to the ears, uncouth like the bellowing of a
camel.

The Sanskrit spoken by Dattakalaśi abounds in recherché desiderative formations with harsh alliterations
that the hero of the play, a viṭa or pimp, likens to thunderbolts. This in itself is a misappropriation and
mockery of the Manusmṛti’s eulogy of speech as the weapon of brahmins, with which they smite their
enemies through curses. 42 In a complete reversal of what happens in the Abhijñānaśakuntalā, this Bhāṇa
paradoxically depicts the polished Sanskrit speech of an elite language expert as macaronic and therefore
ineffectual. “Oh how helpless you are! (aho anāthaḥ khalv asi)” exclaims the viṭa aer Dattakalaśi himself
admits that he cannot “make my diction as sweet and so as a woman’s body,” and even confesses that
his style of speech is “a hundred-killer war-machine with its countless verbal bases, which I mastered as a
result of clashing with the bulls among garrulous disputants.”43 Admittedly, a deeper contextualization is
again called for in this case. The whole scene is narrated by the viṭa, who though a brahmin Sanskrit
speaker, is impoverished but tolerated and in many depictions even respected in society. From the
perspective of the intended reader, it is of course also a satire, meant to amuse, and so it must not be
looked upon uncritically as a “native informant” for the Gupta era’s sociolinguistic milieu any more than
the Abhijñānaśakuntalā.

Does Kālidāsa’s usage reflect these language tensions in the Gaṇikā plays adduced above that preceded
him? How does he configure the relationship between Sanskrit and Prakrit as stage languages?

I propose that a layered phenomenology of language is involved: the Sanskrit speakers are transcendent
subjects who have complete access to the Prakrit languages arrayed below them in a scheme of ever
greater incoherence. To use the terminology of the Nāŀaśāstra, they are more vibhraṣṭa, a term I extend
beyond the purely morphological to include also sense that is inarticulate. The fisherman’s Prakrit would
rank at the very end of the spectrum. It is in this qualified sense that I think it is legitimate to speak of
the dramatic jātibhāṣās, or class-languages, as sociolects. Nevertheless, it would be too reductionist to

38 Padmaprābhṛtaka of Śūdraka in Dezsō & Vasudeva (2008:240–243):: kim āha bhavān—“api sukham aśayiṣṭhāḥ?” kā gatiḥ! … kiṃ
bravīṣi—“kva saṃcicariṣuḥ? tiṣṭha tāvat. kim asi dudrūṣuḥ?” hā dhik prasīdatu bhavān. nārhaty asmān evaṃvidhaiḥ kaṣṭaprahāraniṣṭhurair
vāgaśanibhir abhihantum! sādhu vyāvahārikayā vācācaratv asmān. abhājanaṃ hi vayam īdṛśānāṃ śarabhodgāradurbhagānāṃ
śrotraviṣaniṣekabhūtānāṃ vaiyākaraṇavāgvyasanānām.
39 Lit.: “Didst thou sleep well?” -iṣ-aorist of śī, “to lie down, to sleep,” 2nd pers. sg. Ātmanepada.
40 Desiderative adj. in -u of the root car.
41 Desiderative adjective of dru I P (dravati), “to flee.”
42 Manusmṛti 11.33cd: vākśastraṃ vai brāhmaṇasya tena hanyād arīn dvijaḥ.
43See Dezsō & Vasudeva (2008:242–243): katham aham idānīm anekavāvadūkavādivṛṣabhavighaṭṭanopārjitām anekadhātuśataghnīṃ
vācam utsṛjya strīśarīram iva mādhuryakomalāṃ kari﬇āmi?
 revisiting śakuntalā 13

claim that Kālidāsa’s Prakrits exhaust themselves by reifying, as artificial stage sociolects, an idealized
vision of the social stratification promoted by the Gupta polity, for Kālidāsa was not just a propagandist
for the Gupta state (though he was that too). Rather, the phenomenological layering accounts for the
language-inherent lessening of injunctive force with which speeches in lower ranked languages are
empowered. Their domain of full efficacy is their own level and whatever lies below. The inarticulate
speech of the lowest ranked languages is the least “effectual” in the dramatic portrayal of the social
domain, but it can also be much more evocative and suggestive than Sanskrit in the literary domain. This
naturally involves a degree of mimicry, the king talks differently to his ministers and to lower characters,
he is capable of adjusting his diction and matching the listener’s level. This means that the Sanskrit
speaker is uniquely privileged with the capacity and freedom to slide up and down the phenomenological
hierarchy at will. This allows Kālidāsa to draw from a much wider range of tacit knowledge to integrate
the Gupta empire’s moral codes and value judgments into what Pollock has called the social aesthetic
that one needs to understand to be able to appreciate kāvya literature.

While the existence of early Indian phenomenologies is still a matter of dispute,44 it might seem
anachronistic to introduce the concept here. I do so nevertheless since at least with the literary critic
Abhinavagupta we find a fully developed layered phenomenology of apperceptive perceivers
(pramātṛbheda) of the type proposed here for languages that owes nothing to either Heidegger or
Merleau-Ponty.45

Bibliography

Bhayani, H. C. (1977): “The Bhāṣā-lakṣaṇa chapter of Gītalaṅkāra,” in: Sambodhi 5, 1977, S. 1-29.
Chaudhuri, J.B. (1943): “The Camatkāra-Taraṅgiṅī of Sundarī and Kamalā and the Prānapratiṣṭhā of their
husband Ghanaśyāma,” in: The contribution of women to Sanskrit literature, Volume 1 (Drama), Calcutta
Oriental Series No. 30.
Davis, D. R. Jr. (2007): “On Atmatuṣṭi as a Source of Dharma,” Journal of the American Oriental Society; Jul-
Sep 2007; 127, 3; pp. 279-296.
Dezső, Cs. (2007): “A Parallel Edition of the Nepalese and South Indian Recensions of the First Act of
the Kundamālā (Part I),” Newsletter of the NGMCP, ed. H. Isaacson, Number 3, January-February
2007.
Dezső, Cs. & Vasudeva S. (2009): The quartet of causeries, by Syāmilaka, Vararuci, Śūdraka & Īśvaradatta. New
York, NY : New York University Press, Clay Sanskrit Library, JJC Foundation, 2009.
Ghosh, M. (1967): The Nāŀaśāstra, ascribed to Bharata-Muni. Edited with an introd. and various readings,
Calcutta, Manisha Granthalaya, 1967.
Goldman, R. (2000): “Language, Gender and Power: The Sexual Politics of Language and Language
Acquisition in Traditional India,” in: Invented Identities : The Interplay of Gender, Religion and Politics in
India,” edited by Julia Leslie, Mary McGee. New Delhi: Oxford University Press.
Goodall D., & Isaacson H. ( 2003): The Raghupañcikā of Vallabhadeva: being the earliest commentary on the
Raghuvaṁśa of Kālidāsa. Groningen, E. Forsten, 2003.

44More recently Yogācāra Buddhism has been labelled as Buddhist phenomenology. But see also the criticism in Mayer
(2009).
45 See Vasudeva (2004:191ff).
 revisiting śakuntalā 14

Hahn, M. (2005): “Preliminary Remarks on Two Manuscripts Written in the So-called “Arrow-headed”
Script”, in: Buddhism and Jainism. Essays in Honour of Dr. Hojun Nagasaki on His Seventieth Birthday. Kyoto
2005, pp. 712–701, [119–130].
Hanisch, A. (2006): “Progress in Deciphering the So-called “Arrow-head” Script Allowing Access to
Sarvarakṣita’s Maṇicūḍajātaka, a Text of the Buddhist Sāṃmitīya School”, in: Journal of Buddhist
Studies, Vol. IV. Colombo 2006, pp. 109–161.
Hanneder, J. (2009): “Modernes Sanskrit, eine vergessene Literatur,” in: Pāsādikadānam: Festschri für
Bhikkhu Pāsādika. Ed. Martin Straube, Roland Steiner, Jayandra Soni, Michael Hahn and Mitsuyo
Demoto. Marburg : Indica et Tibetica Verlag, 2009. (Indica et Tibetica ; 52).
Hardy, F. (2009): Āryāsaptaśatī (Seven Hundred Elegant Verses) by Govardhana. Clay Sanskrit Library, 2009.
Hock, H.H. & Pandharipande R. (1976): “The sociolinguistic position of Sanskrit in pre-muslim South
Asia,” Studies in Language Learning, Volume I, Number II, Spring 1976.
Jacobi, H. (1886): Ausgewählte Erzählungen in Mâhârâshṭrî, Leipzig.
Kahrs, E. (1992): “What is a tadbhava word?,” Indo-Iranian Journal, Volume 35, Numbers 2-3 / July, 1992 pp.
225-249.
Levitt, S. H. (2005): “Why Are Sanskrit Play Titles Strange?”, Indologica Taurinensia, Volume XXXI: 195–
232.
Mayer, A. L. (2009): Review of: “Buddhist Phenomenology: A Philosophical Investigation of Yogācāra
Buddhism and the Ch'eng Wei-shih Lun,” Journal of Buddhist Ethics, Volume 16, 2009.
Mazzio, C. (2008): The Inarticulate Renaissance, Language Trouble in an Age of Eloquence, University of
Pennsylvania Press, 2008.
Norman, K. R., (2006): “Translation problems in early Buddhist literature,” in: Jaina-Itihāsa-Ratna
Festschri für Gustav Roth zum 90. Geburtstag, ed. Ute Hüsken, Petra Kieffer-Pülz Und Anne
Peters, Indica Et Tibetica, e. Michael Hahn unter Mitwirkung von Jens-Uwe Hartmann, Konrad
Klaus und Roland Steiner, Band 47, Indica Et Tibetica Verlag, Marburg 2006
Pieruccini, C. (2005): “The Loving Creeper and the Tree: Kālidāsa, Botany and Ancient Spells”, in
Cracow Indological Studies, Vol. 7, Love and Nature in Kāvya Literature, Proceedings, ed. by Lidia
Sudyka, Kraków 2005, pp. 29-40.
Pollock, S. (2008): “Toward a Political Philolo: D. D. Kosambi and Sanskrit,” Economic and Political Weekly.
Srinivasan, S.A. (1981): On the Composition of the Nāŀaśāstra, (Studien zur Indologie und Iranistik
Monographie 1) Reinbeck: Dr. Inge Wezler Verlag für Orientalische Fachpublikationen, 1980.
Steiner, R. (1997): Untersuchungen zu Harṣadevas Nāgānanda und zum indischen Schauspiel
(“Investigation into Harṣadeva's Nāgānanda and Indian Drama”). Swisttal-Odendorf 1997. Indica
et Tibetica. 31.
Steiner, Roland, (2001): “Play Editing and Prakrit Grammarians,” Les sources et le temps. Sources and time. A
colloquium. Pondicherry 11-13 January 1997. Edited by François Grimal. Pondichéry 2001 (Publications
du département d'indologie. 91.), pp. 63-76.
Szanto, P-D. (2008): “A Critical Edition of Catuṣpīṭhatantra I.3. with Three Sanskrit
Commentaries,” Tantric Studies 1.2008. Hamburg: Centre for Tantric Studies, 2008, p. 2-21.
Tanselle, G. T., (1992): A Rationale of Textual Criticism, University of Pennsylvania Press 1992.
Vasudeva, S. (2004): The Yoga of the Mālinīvijayottaratantra. Chapters 1-4, 7, 11-17. Critical Edition, Translation
& Notes. (Collection Indologie 97). Pondichéry 2004.
Wezler, A. (1996). “Do you speak Sanskrit? on a class of Sanskrit texts composed in the late middle ages.”
Ideolo and Status of Sanskrit, Contributions to the History of the Sanskrit Language, ed. J. Houben (1996)
Wright, J. C., (1999): “The Supplement to Ludwig Alsdorf’s ‘Kleine Schrien’: A Review Article,” Bulletin
of the School of Oriental and African Studies, University of London, Vol. 62, No. 3 (1999), pp. 529-542