वासवदत्ता (महाकवि सुबन्धु - श्लेष-प्रधान शास्त्रीय गद्यकाव्य सान्वय सटीक)

20 VĀSAVADATTĀ
for (Buddhist) ascetics at their mealtime, and was like a mass of
Yamunā's foam by night, and resembled a fragment of stone for
the polishing of Mēnakā's nails' (ed. Hall, p. 44).¹
The yamaka, 'repetition' or 'chiming,' is the repeating of
words or parts of words of similar sound but divergent meaning,
which the Kāvyādarśa (1. 61; cf. 3. 1–37; Sāhityadarpaṇa,
No. 640; Kāvyaprakāśa, tr. Jhā, pp. 185–188, Benares, 1898;
Kuvalayānandakārikā, 4. 6) describes as
āvṛttiṁ varṇasaṅghātagōcarāṁ yamakaṁ viduḥ,
'a repetition consisting of a combination of sounds they know as
yamaka.' This is illustrated by the following passage from the
Vāsavadattā: āndōlitakusumakēsarē kēsarēṇumuṣi raṇitamadhu-
ramaṇīnāṁ ramaṇīnāṁ vikacakamudākarē mudākarē, '(when
there blew a wind that) rocked the filaments of the flowers and
removed their pollen from the hair of damsels wearing delight-
somely tinkling jewels, whilst it had an abundance of expanded
white lotuses, and caused pleasure' (ed. Hall, pp. 52–53).
The prauḍhōkti, or 'pomposity,' is thus defined by the Kuva-
layānandakārikā (1. 124):
prauḍhōktir uktā 'rthāhētōs taddhētutvaprakalpanam,
'in the absence of a cause for a thing, the invention of a cause
for it is called prauḍhōkti.' It is exemplified in the Vāsavadattā
by the passage describing the heroine's lip as mukhacandra-
sannihitasandhyārāgēṇa dantamaṇirakṣāsindūramudrānukāriṇā,
'which had the glow of eventide in close proximity to her moon-
like face; which had what seemed to be a minium seal as a guard
for the jewels of her teeth' (ed. Hall, p. 58).
The rūpakātiśayōkti, or 'hyperbolical metaphor,' is merely an
exaggerated form of the preceding alaṁkāra. It is thus defined
in the Kuvalayānandakārikā (1. 34):
rūpakātiśayōktiḥ syān nigīryādhyavasānataḥ,
'identification so that (the object identified) should be swallowed

¹ On the similes in the Vāsavadattā see also the examples collected by Krishna-
machariar, Introd., pp. 20–22.

INTRODUCTION 21
up (and thus completely disappear) would be rūpakātiśayōkti;
and as an example may be cited, from the description of Vāsa-
vadattā just quoted, the passage vilōcanēndīvarabhramarapaṅk-
tibhyāṃ mukhamadanamandiratōraṇābhyāṃ rāgasāgaravēlābh-
yāṃ yāuvananartakalāsikābhyāṃ bhrūlatābhyāṃ virāiamānām,
‘adorned with delicate brows which were clusters of bees about
her blue-lotus eyes; portals of her face that formed the abode of
Love; the shores of Passion’s sea; wantoning in youthful dancing’
(ed. Hall, p. 61).
The akramātiśayōkti, or ‘fused hyperbole,’ is closely akin to
the preceding rhetorical figure, of which it is merely an intensi-
fication. It receives the following definition in the Kuvalayā-
nandakārikā (1. 39):
akramātiśayōktiḥ syāt sahatvē hētukāryayōḥ,
‘akramātiśayōkti would be in the unity of cause and effect.’
Śivarāma cites but one instance of the figure in the Vāsavadattā,
this being samaṃ dviṣāṃ dhanuṣāṃ ca jīvākṛṣṭiṃ yōdhāś cakruḥ,
‘the warriors drew at once the ‹lives (jīva)› of their foes and the
‹strings (jīvā)› of their bows’ (ed. Hall, p. 295).
Two other forms of hyperbole are mentioned by Śivarāma as
occurring in Subandhu’s romance. The first of these is bhēdakā-
tiśayōkti, or ‘hyperbole of differentiation.’ It is defined as follows
in the Kuvalayānandakārikā (1. 36):
bhēdakātiśayōktis tu tasyāi ”vā ’nyatvavarṇanam,
‘bhēdakātiśayōkti is the description of that (which is the subject
under discussion) by means of differentiation,’ and it is exemplified
in the Vāsavadattā by pṛthur api gōtrasamutsāraṇavistāri-
tabhūmaṇḍalaḥ, ‘Pṛthu ‹levelled the earth by banishing the
mountains› (but Cintāmaṇi ‹covered the earth by sending forth
his offspring›)’ (ed. Hall, p. 22).
The remaining form of hyperbole in the Vāsavadattā is sam-
bandhātiśayōkti, or ‘hyperbole of connexion,’ which is thus
defined in the Kuvalayānandakārikā (1. 37):
sambandhātiśayōktiḥ syād ayōgē yōgakalpanam,
‘sambandhātiśayōkti would be the invention of connexion when

22 VĀSAVADATTĀ

connexion is absent,' as when Subandhu describes trees as being
anūrukarakaśābhighātaparavaśaravirathaturagagrāsaviṣamitapa-
llavāis, 'with shoots made uneven by the feeding of the horses
of the chariot of the sun which are obedient when lashed by the
whip in the hands of Anūru' (ed. Hall, p. 120).
The ratnāvalī, or 'jewel necklace,' is defined as follows in the
Kuvalayānandakārikā (1. 139) :
kramikāprakṛtārthānāṃ nyāsaṃ ratnāvalīṃ viduḥ,
'an arrangement of objects serially irrelevant they know as ratnā-
valī,' and is exemplified in the Vāsavadattā where the heroine is
described as vikacēna nētrakamalēna śanāiścarēṇa pādēna tamasā
kēśapāśēna grahamayīm iva, 'she seemed to be made of planets:
of <Venus>, for she had <wide-open> lotus eyes; of <Saturn>, for
she had <slow-moving> steps; of <Rāhu>, for she had <dark> heavy
hair' (ed. Hall, p. 64).
The kāvyaliṅga, or 'poetic reason,' is thus defined by the
Sāhityadarpaṇa (No. 710; cf. Kāvyaprakāśa, tr. Jhā, pp. 238–
239, Benares, 1898; Kuvalayānandakārikā, 1. 120) :
hētōr vākyapadārthatvē kāvyaliṅgaṃ nigadyatē,
'kāvyaliṅga is applied to the implication of a cause in a sentence or
word,' and finds exemplification in the Vāsavadattā in the passage
khalaḥ punas tad aniṣṭam anucitam ēvā 'vadhārayanty aniṣṭōd-
bhāvanarasōttaraṃ hi khalahṛdayam, 'the wicked, on the other
hand, make it (thy conduct) out to be undesirable and indecorous;
for the heart of the wicked man finds its highest delight centred
in bringing to light what is undesirable' (ed. Hall, p. 70).
The mīlita, or 'vanished,' denotes a complete loss of distinction
between two objects because of their superficial resemblance, as
is expressed by the definition of the Kuvalayānandakārikā (1. 145;
cf. Kāvyaprakāśa, tr. Jhā, pp. 253–254, Benares, 1898; Ruyyaka's
Alaṃkārasarvasva, ed. Durgāprasāda and Parab, p. 167, Bombay,
1893):
mīlitaṃ yadi sādṛśyād bhēda ēva na lakṣyatē,
'if, because of similarity, a distinction is not observed, it is
mīlita.' As an example from Subandhu I may cite mādhurya-

INTRODUCTION 23

śāityasucitvasantāpaśāntibhiḥ payaḥ paya ivē 'ti, 'fancying that
“water is as milk because of its sweetness, coolness, purity, and
healing of distress”' (ed. Hall, p. 80).
The anuprāsa, or 'alliteration,' a rhetorical figure found with
considerable frequency in Subandhu's romance, is thus defined
by the Sāhityadarpaṇa (No. 633; cf. Kāvyādarśa, 1. 55; Kāvya-
prakāśa, tr. Jhā, pp. 182–184, Benares, 1898; Kuvalayānandakā-
rikā, 4. 2–5) :
anuprāsaḥ śabdasāmyaṁ vāiṣamyē 'pi svarasya yat,
‘anuprāsa (is) a similarity of sound, despite a dissimilarity of the
vowel.’ As an example from the Vāsavadattā may be cited these
two adjectives descriptive of the River Rēvā: madakalakalahaṁ-
sasārasarasitōdbhrāntabhāḥkūṭavikaṭapucchacchaṭāvyādhūtavikaca-
kamalakhaṇḍavigalitamakarandabindusandōhasurabhitasalilayā..
. . . upakūlasañjātanalanikuñjapuñjitakulāyakukkuṭaghaṭāghūtkā-
rabhāiravatīrayā, 'whose waters were perfumed by the abundance
of the drops of liquid which had fallen from the fragments of full-
blown lotuses shaken by many monstrous tails of bhāḥkūṭa-fish
that had been terrified by the notes, indistinct for passion, of geese
and herons; ..... whose banks were strident with the screams
of multitudes of wild cocks whose nests thronged the bowers of
reeds that had sprung up near its shores' (ed. Hall, pp. 95, 98).
The sama, or 'equal,' is thus defined in the Kāvyādarśa (1. 47;
cf. Sāhityadarpaṇa, No. 618; Kāvyaprakāśa, tr. Jhā, pp. 175–
176, Benares, 1898):
samaṁ bandhēṣv aviṣamaṁ tē mṛdusphuṭamadhyamāḥ
bandhā mṛdusphuṭōnmiśravarṇavinyāsayōnayaḥ,
‘sama is not uneven in collocations (of words); these collocations,
smooth, rough, (and) middling, depend on the arrangement of
smooth, rough, and mixed (sounds).’ It is illustrated by the
passage kāmadāruṇa madāruṇanētrāsmaramayaṁ ramayantaṁ
tvām adayaṁ madayantī param akam itāraṁ param akamitāraṁ
vāñchati, 'what gentle-eyed woman who fervently delighteth
thee, that art not inflamed with passion, (but art) the essence of
love, delightsome, (and) a most excellent lover, desireth another

24 VĀSAVADATTĀ
that is no lover [cruel with passion ! red-eyed with lust ! alas, an
unlovely dame desireth thee, the essence of lovelessness, hot,
pitiless, absolutely no lover, (and) bound for utmost woe !]’ (ed.
Hall, pp. 213–214).
The vidhi, or ‘rule,’ is defined as follows by the Kuvalayānan-
dakārikā (I. 167) :
siddhasyāi "va vidhānaṃ yat tad āhur vidhyalaṃkṛtim,
‘what (is) a precept of a thing well established, that they call the
vidhi adornment,’ and is exemplified by kuraṅgikē kalpaya
kuraṅgaśāvakēbhyaḥ śaspāṅkuraṃ kiśōrikē kāraya kiśōrakēbhyaḥ
pratyavēkṣāṃ, ‘Kuraṅgikā, prepare a blade of young grass for
the antelope fauns ! Kiśōrikā, have the young colts looked after’ !
(ed. Hall, pp. 230–231).
The sambhāvana, or ‘supposition,’ is thus defined by the
Kuvalayānandakārikā (I. 125) :
sambhāvanaṃ yadī 'tthaṃ syād ity ūhō 'nyasya siddhayē,
‘sambhāvana is a conjecture for the attainment of something
else with the thought, “if it were so.”’ It is illustrated in the
Vāsavadattā by the passage tvatkṛtē yā "nayā vēdanā "nubhūtā
sā yadi nabhaḥ patrāyatē sāgarō mēlānandāyatē brahmāyatē
lipikarō bhujagarājāyatē kathakas tadā kim api katham apy
anēkāir yugasahasrāir abhilikhyatē kathyatē vā, ‘the pain that
hath been felt by this maiden for thy sake might be written or
told in some wise or in some way in many thousands of ages
if the sky became paper, the sea an ink-well, the scribe Brāhma,
(and) the narrator the Lord of Serpents’ (ed. Hall, pp. 238–239).
The kāraṇamālā, or ‘chain of causes,’ is given the following
definition in the Kuvalayānandakārikā (I. 103 ; cf. Kāvyaprakāśa,
tr. Jhā, p. 246, Benares, 1898) :
gumphaḥ kāraṇamālā syād yathāprakrāntakāraṇāiḥ,
‘a series (made) by causes proceeding one after the other is
a kāraṇamālā,’ and an example is seen by Śivarāma in the
description of Vāsavadattā’s palace as ajñātataṭasphāṭikapaṭ-
ṭasukhaniṣaṇṇanidrāyamāṇaprāsādapārāvatābhiḥ, ‘with palace

INTRODUCTION 25
doves sleeping comfortably (because) perched on slabs of crystal
from shores unknown' (ed. Hall, pp. 217–218).
The udātta, or 'exalted,' is thus defined in the Sāhityadarpaṇa
(No. 752; cf. Kāvyaprakāśa, tr. Jhā, p. 240, Benares, 1898;
Kuvalayānandakārikā, 1. 162–163):
lōkātiśayasampattivarṇanō "dāttam ucyatē
yad vā "pi prastutasyā 'ṅgaṁ mahatāṁ caritaṁ bhavēt,
'the portrayal of extraordinary prosperity is called udātta, or
it would even be the deeds of the great, (if they form) part of
the subject in hand.' An illustration of this figure is found in
another portion of the description just quoted: karpūrapūra-
viracitapulinataṭaniviṣṭaninadānumīyamānarājahaṁsībhir, 'with
flamingoes whose noise would imply that they had settled near
the sand bank formed by the stream of camphor' (ed. Hall,
p. 218), only one of extreme wealth being able to possess such
a river.¹
The kāitavāpahnuti, or 'false concealment,' is defined by the
Kuvalayānandakārikā (1. 28) as follows:
kāitavāpahnutir vyaktāu vyājādyāir nihnutēḥ padāiḥ,
'kāitavāpahnuti (consists) in the manifestation of concealment by
words of pretext and the like,' and finds exemplification in the
Vāsavadattā in the passage ativēganipītajaladhijalaśaṅkhamālām
iva balākācchalād udvamann adṛśyata jaladaḥ, 'the cloud seemed
to vomit forth, like a crane, what appeared to be a series of
ocean shells that had been drunk down too hastily' (ed. Hall,
pp. 283–284).
The lōkōkti, 'popular expression,' is thus defined by the
Kuvalayānandakārikā (1. 156):
lōkapravādānukṛtir lōkōktir iti kathyatē,
'the imitation of a popular colloquialism is called lōkōkti,' and
finds an illustration in Subandhu's romance in the exhortation
tad adhunā yadi tvaṁ sahapāṁśukrīḍanasamaduḥkhasukhō 'si
¹ Śivarāma rightly notes that this passage also contains the rhetorical figure anumāna,
or 'inference' (cf. Kāvyaprakāśa, tr. Jhā, pp. 243-244, Benares, 1898; Kuvalayānan-
dakārikā, 2. 10).

26 VĀSAVADATTĀ
tadā mām anugaccha, 'now, therefore, if thou didst share the
sorrows and joys of our playing together in the dust, then follow
me' (ed. Hall, p. 82).
The svabhāvōkti, or 'natural description,' receives the following
definition in the Kuvalayānandakārikā (1. 160; cf. Kāvyapra-
kāśa, tr. Jhā, p. 235, Benares, 1898) :
svabhāvōktiḥ svabhāvasya jātyādisthasya varṇanam,
'svabhāvōkti is the description of inherent nature consisting of
characteristics and the like.' As an example from the Vāsava-
dattā may be cited :
paśyō 'dañcadavāñcadañcitavapuḥ paścārddhapūrvārddhabhāk
stabdhōttānitapṛṣṭhaniṣṭhitamanāgbhugnāgralāṅgūlabhṛt
daṃṣṭrākōṭiviśaṅkaṭāsyakuharaḥ kurvan saṭām utkaṭām
utkarṇaḥ kurutē kramaṃ karipatāu krūrākṛtiḥ kēsarī,
'lo, with his bending body bending up and bending down, now
with his hind quarters and now with his fore quarters, with the
tip of his tail slightly bent along his hard, arched back, with his
cavernous mouth monstrous with the tips of his fangs, making
his mane huge, (and) with his ears erect, the horrible lion doth
make attack upon the lord of elephants' (ed. Hall, p. 103).
The kāvyārthāpatti, or 'poetic inference,' is defined as follows
in the Kuvalayānandakārikā (1. 119):
kāimutyēnā 'rthasaṃsiddhiḥ kāvyārthāpattir iṣyatē,
'an a fortiori attainment of a matter is regarded as kāvyārthā-
patti.' It is exemplified by Subandhu in his heroine's letter to
Kandarpakētu :
pratyakṣadṛṣṭabhāvā 'py asthirahṛdayā hi kāminī bhavati
svapnānubhūtabhāvā draḍhayati na pratyayaṃ yuvatiḥ,
'a loving maid is of unsteady heart even when she hath seen the
feelings (of her lover) with her eyes ; a girl who hath learned his
feelings only from a dream hath no assurance' (ed. Hall, p. 164).
Literary and Ethical Merit of the Vāsavadattā. The Vāsa-
vadattā apparently being written to display its author's skill in
rhetoric, rather than his inventive powers in fiction, we are

INTRODUCTION 27
naturally led to consider what literary value we may assign to it.
Here the ‘personal equation’ must inevitably play a part, and
here the fundamental difference between Oriental and Occidental
concepts must be duly recognised. In the West the subject-
matter comes first in nearly every form of literary composition ;
and the more tense and nervous the people, the more simple and
direct is the style. In the East, on the contrary, the form is
often more important than the matter, especially in periods of
hyper-civilisation, such as was that during which Subandhu
wrote. We must, therefore, consider the Vāsavadattā from the
luxuriant atmosphere of the land of its author, not from the
‘practical’ point of view of the West. To me, at least, there is
true melody in the long, rolling compounds, a sesquipedalian
majesty which can never be equalled save in Sanskrit ; and the
alliterations have a lulling music all their own to ears weary of
the blatant discords of vaunted modern ‘progress.’ There is, on
the other hand, a compact brevity in the paronomasias, which
are, in most cases, veritable gems of terseness and twofold
appropriateness, even though some are manifestly forced and are
actually detrimental to the sense of the passages in which they
occur. Yet in judging Subandhu for his faults, it must be re-
membered that he created, at least so far as we now know, a new
literary genre in India ; and if this fact be borne in mind, his
blemishes appear to be marvellously few. In estimating his
literary merits special stress should be laid on his descriptions.
These are, it must be confessed, cloying from their abundance.
They form the preponderating part of the entire romance, and
the slender framework of the story is wellnigh lost beneath
them. Yet despite this tropical luxuriance, the descriptions are
not without beauty and appropriateness, whether they set forth
the charms of mountain, forest, and stream, or portray the
rāja’s valour and the loveliness of the heroine herself. The
entire romance may, in a sense, be likened to India’s own
architecture, where the whole structure is so overlaid with minute
detail that the eye forgets the outlines of the building in amaze-
ment at the delicate traceries which cover it.

28 VĀSAVADATTĀ
Nor does it seem to me that the ethical standard of the
Vāsavadattā can be objectionable to one of healthy mind.
True, the East is not as the West; and there are personal
descriptions more detailed than would be desirable in Occidental
literature, together with evident approval of relations and ideals
which the less sensual Western mind rightly condemns. There
are passages, too, which I would gladly have omitted, had I felt
that a faithful translator could do so. And yet, despite all this,
I find in the romance no evidence of delight in uncleanness, such
as nauseates, for example, in Petronius or in Martial. It is not
pornographic; it is, at worst, unmoral, though its rigid adherence
to all conventions, both in letter and, I think, in spirit, renders
even unmorality almost too harsh an accusation. From an
Indian point of view, unlightened by the radiance of Christianity
and the morality which it inculcates, I should not hesitate to
term the Vāsavadattā a moral work, especially in view of the
conditions of life in mediæval India. Its atmosphere, luxuriant
though it be, has never seemed to me to be debasing.
It is by no means impossible that some will dissent from the
views here expressed regarding the literary and moral quality of
Subandhu's romance. If so, they may turn from the first
Western translator of the Vāsavadattā to the first Western
editor of the romance, Fitzedward Hall, who, in his Introduction,
has unsparingly condemned the entire production both in its
literary and in its ethical aspects—a precedent followed by
Krishnamachariar in his sarcastic critique of the whole plot of
Subandhu's work (Introd., pp. 50–66).
Outline of the Plot of the Romance. The outline of the story
of the Vāsavadattā is as follows : A king named Cintāmaṇi had
a son Kandarpakētu, who was, like his father, the embodiment
of all virtues. Once upon a time toward dawn, when true
dreams come, the young prince saw in his sleep a vision of
a maiden of some eighteen years, whose loveliness could not be
surpassed. Jealous sleep forsook Kandarpakētu, who, with his
friend Makaranda, left the city in his love-longing for the
unknown princess. In their wanderings the pair came to the

INTRODUCTION 29
Vindhya mountains, and there, in the watches of the night,
the sleepless prince overheard the conversation of two birds
perched on a branch of the tree beneath which he lay. To the
story of the husband-bird, trying to explain his late hours to his
suspicious wife, Kandarpakētu listened, and was richly rewarded
by what he heard. In the city of Pāṭaliputra on the Ganges, so
the maina recounted, reigned the mighty monarch Śṛṅgāraśē-
khara, who had an only daughter named Vāsavadattā. In the
spring she, too, had met her fate in a dream—a youth of
matchless beauty, whose name was Kandarpakētu. The con-
fidante of the princess at this juncture was her maid, Tamālikā,
who had volunteered to seek Kandarpakētu and bear to him
a missive from the princess telling of her love. The lovers were
now soon united at Pāṭaliputra, where Kandarpakētu was informed
that Śṛṅgāraśēkhara, dismayed at his daughter’s unwedded state,
had determined to marry her the very next day to the Vidyādhara
prince Puṣpakētu. Kandarpakētu and Vāsavadattā accordingly
returned almost immediately, by means of a magic steed, to the
Vindhyas; but when the prince awoke in the morning, his
beloved was no longer in the bower. Mad with sorrow, he was
restrained from suicide only by a voice from heaven which
promised him reunion with the princess. After many months of
weary searching and waiting, he found Vāsavadattā turned to
stone. His touch gave the statue life again, and she told him
how, while two armies destroyed each other to gain her for their
leaders, she had unwittingly intruded in the garden of a hermit,
who laid upon her the curse of petrification until her lover should
come. Thus, at last, the woes of the lovers were over, and
returning to Kandarpakētu’s capital, delight was theirs ever
afterward (for other summaries see Hall, Introd., pp. 29–43;
Stréhly, Revue politique et littéraire, 44. 305–308; Krishnama-
chariar, Introd., pp. 9–14; and the references given in the
bibliography, below, pp. 197–199).
From this brief outline of the plot of the Vāsavadattā it will
be seen that Subandhu alludes to several incidents widely found
in literature and folk-tales, such as talking birds, magic steeds,

30 VĀSAVADATTĀ
and transformation. To all these I have sought to give parallels,
especially from modern Indian folk-tales, in footnotes to the
passages in which they occur. There seems to be no parallel,
however, to the story of the Vāsavadattā as a whole, and, as
already stated (p. 2), Subandhu evidently invented the slender
plot of his own romance (cf. also Hall, Introd., pp. 1–6).
Sanskrit ‘Estilo Culto’ Previous to Subandhu, and in Later
Epigraphy. This allusion to comparative literature brings us to
possible quasi-parallels to the style of the Vāsavadattā. The
commingling of prose and verse which is characteristic of the
ākhyāyikā, kathā, and other forms of gadya, or poetical prose (cf.
Regnaud, Rhétorique sanskrite, pp. 74–77, Paris, 1884), was by
no means an innovation of Subandhu. It is found, for instance,
in the Pañcatantra and the Jātakas, as well as in the gāthās of
the Brāhmaṇas and the Northern Buddhists, even though these
latter be more archaic than the prose in which they are set.
Obscure as is the date of the beginning of the kāvya style in
India, a flood of light has been cast on its early history by
Bühler in his ‘Die indischen Inschriften und das Alter der
indischen Kunstpoesie,’ in SWAW. 122, Abhandlung 11. There
he has shown that the eulogy of Vatsabhaṭṭi, preserved in an
inscription in the Temple of the Sun at Maṇḍasōr, and dated
473–474 A.D., contains descriptions of cities, natural phenomena,
and the like, together with compound words of inordinate length,
and the rhetorical devices of anuprāsa, upamā, utprēkṣā, rūpaka,
and (possibly) virōdha, all of which even fulfil the requirements
laid down, for instance, by Daṇḍin in his Kāvyādarśa. The term
kāvya itself occurs at least as early as 375–390 A. D. in Hariṣēṇa’s
panegyric on Samudragupta, inscribed on a pillar at Allāhābād,
which also contains long compounds and the rhetorical figures
of varṇānuprāsa, rūpaka, upamā, and ślēṣa. The kāvya style
is carried back to the early second half of the second century A.D.
by the Girnār inscription of Rudradāman, which has long com-
pounds and numerous anuprāsas, together with two upamās and
one utprēkṣā. It is thus clear that a fairly developed kāvya was
known in India as early as the second century of our era, not

INTRODUCTION [p. 31]

forgetting that the epic of the Rāmāyaṇa contains many approxi-
mations to the kāvya style (Jacobi, Das Râmâyaṇa, pp. 119–126,
Bonn, 1893).
In the course of time kāvya inscriptions became more elaborate,
particularly in their use of the śleṣa. Mention can here be made
only of the śleṣas and virōdhas in the Valabhī grant of Dhruva-
sēna III, dated in 653–654 A.D. (Hultzsch, EI. 1. 85–92) ; the
Balōda plates of Tīvaradēva, of the middle of the eighth century
(Hultzsch, ib. 7. 102–107); a grant of Indrarāja III, dated in
915 A.D. (Bhandarkar, ib. 9. 24–41); the Cambay plates of
Gōvinda IV, dated in 929–930 A.D. (Bhandarkar, ib. 7. 26–47) ;
and the Dēvulapalli plates of Immaḍi Nṛsiṃha, dated in 1504
(Ramayya, ib. 7. 74–85); though allusion should also be made
to the general style of such an inscription as the Kaḍaba plates
of Gōvinda III, dated in 812–813 A.D. (Lüders, EI. 4. 332–349).
In the inscriptions of the Vāillabhaṭṭasvāmin temple at Gwaliar,
dated in 874–875 A.D. (Hultzsch, EI. 1. 154–162), which abound
in virōdhas, Kielhorn (apud Hultzsch, EI. 1. 157, note 23) has
already called attention to a possible reminiscence of the dhana-
dēnā ’pi pracetasā, ‘which is ⟨Kubēra⟩, yet ≪Varuṇa≫, for it is
⟨generous⟩ and ≪wise≫,’ of the Vāsavadattā (ed. Hall, p. 111) in
the dhanadō ’pi na pramattō, ‘he was ⟨Kubēra⟩, but not ≪Varuṇa≫,
for he was ⟨generous⟩ and not ≪inattentive≫’ (line 6). Kielhorn,
moreover, in his edition and translation of the Rādhanpur plates
of Gōvinda III, dated in 807–808 A.D. (EI. 6. 239–251), expressly
declares (p. 240) that ‘an examination of the language and
general style of most of these verses can leave no doubt that
their author or authors are greatly indebted for their expressions
and poetical devices to such works as Subandhu’s Vāsavadattā
and Bāṇa’s Kādambarī and Harṣacarita.’ There is no need here
to repeat the parallels which Kielhorn has drawn between the
inscription in question and the romances of Subandhu and Bāṇa
(EI. 6. 247–250), nor is it necessary to make more than a passing
allusion to the fact that a close examination of Indian epigraphy
would probably reveal many more parallels to the Vāsavadattā
and other productions of the same genre. It would by no means

32 VĀSAVADATTĀ
follow, however, that such resemblances necessarily imply
borrowing from the romances of Subandhu and Bāṇa, for the
same research would, in all probability, show an equal, or even
greater, affinity with kāvya literature in general. The kāvya of
the inscriptions must, therefore, be regarded as an integral part
of the vast mass of artificial Sanskrit literature, its development
being attained by a process of natural growth.
The Commingling of Prose and Verse and Paronomasia Out-
side the Vāsavadattā. Outside of India the commingling of
prose and poetry in the same composition is found in the Chinese
romance P'ing Shan Lêng Yen (tr. Julien, P'ing-Chan-Ling-Yen,
Les Deux Jeunes Filles lettrées, 2 vols., Paris, 1860); in Sa'dī's
Gulistān; in The Thousand Nights and One Night; in the Old
Picard Aucassin et Nicolette; in Norse Sagas and in Middle Irish
tales and histories (cf. Windisch, Irische Texte, 3. 447–449, Leipzig,
1891–1897); and in Boccaccio's L'Ameto; as well as in the Saturæ
Menippeæ of Varro; Petronius; the author of the Historia
Apollonii regis Tyri; and among Basutos and Eskimos (cf.
MacCulloch, Childhood of Fiction, London, 1905, pp. 480–481;
Teuffel-Schwabe, Geschichte der römischen Literatur, 5 ed.,
pp. 43–44, Leipzig, 1890). In like manner, the elaborate parono-
masias which are so essentially a part of the style of the Vāsava-
dattā, and which later led to such productions as Kavirāja's Rāgha-
vapāṇḍavīya, which in identical words celebrates the deeds of the
Raghu princes by one reading and those of the Pāṇḍava heroes
by the other reading, or Rāmacandra's Rasikarañjana (ed. and
tr. Schmidt, Stuttgart, 1896), which may be read as a laudation
either of asceticism or of eroticism, were in vogue, as we have
seen, long before the time of Subandhu.¹ They also occur in the
Chinese romance Ch'in P'ing Mei (cf. Giles, History of Chinese
Literature, p. 309, London, 1901) and in the writings of Lyly.
¹ To this list Krishnamachariar (Introd., p. 18) adds Dhanamjaya's Dvisandhāna-
kāvya (ed. Bombay, 1895 = Kāvyamālā, No. 49) and the Rāghavayādavapāṇḍavīya,
or Kathātrayīvyākhyāna, of Cidambara and his son Anantanārāyaṇa, which gives the
substance of the Rāmāyaṇa, Mahābhārata, and Bhāgavatapurāṇa. Another literary
curiosity of this general type is the elegy of Leon of Modena on his teacher, Moses
Bassola, which may be read either in Hebrew or in Italian (cf. Jewish Encyclopedia,
2. 576, New York, 1902).

INTRODUCTION 33
Subandhu and Lyly. It is Lyly, indeed, with whom Subandhu
may perhaps best be compared for a general parallel with the
style and spirit of the Vāsavadattā. Neither the Euphues nor
Subandhu's romance is concerned so much with the matter as
with the form, although the English author manifests a didactic
purpose which finds no counterpart in the Sanskrit writer. Like
the Vāsavadattā, the Euphues and its continuation contain epi-
sodes, or stories within stories, as the tale of Callimachus (Lyly,
ed. Arbor, pp. 227–245, Birmingham, 1868), which itself contains
the story of Cassander the hermit (pp. 235–239), thus parallelling
the arrangement not only of the Vāsavadattā and the Kādambarī,
but also of the Pañcatantra, Śukasaptati, Vetālapañcaviṁśati,
The Thousand Nights and One Night, Decameron, and other similar
works.¹ Lyly's romance is curiously like Subandhu's in yet other
respects. Here I may note particularly from his Euphues the use
of paronomasia, alliteration, antithesis (corresponding respectively
to Subandhu's employment of śleṣa, anuprāsa, and virōdha), and
learned allusions. From the mass of possible citations under
each category I quote but two of each. (1) Paronomasia : ‘and
though they be commonly in a great cholar that receiue the mate,
yet would I willingly take euery minute tenne mates to enioy Liuia
for my louing mate’ (Lyly, ed. Arbor, p. 66, Birmingham, 1868) ;
‘did not Iupiters egge bring forth as well Helen a light huswife
in earth, as Castor a light Starre in heauen ?’ (p. 208).² (2) Alli-
teration : ‘these subtill shiftes, these painted practises (if I wer to
be wonne) would soone weane me from the teate of Vesta to the
toyes of Venus’ (p. 76) ; I am Philautus no Italian lady, who
commonly are woed with leasings, and won with lust, entangled
with deceipt, and enioyed with delight, caught with sinne, and
cast off with shame’ (p. 360). (3) Antithesis : ‘how curious were
we to please our Lady, how carelesse to displease our Lorde ?
Howe devout in seruing our Goddesse, how desperate in forget-
¹ For a charming modern imitation see Bain, Digit of the Moon (London, 1899).
² It is well known that paronomasia occurs not infrequently in Shakespeare (e. g.
Julius Caesar, 1. 2. 155 ; Merchant of Venice, 4. 1. 123 ; Taming of the Shrew, 2. 1.
190 ; Richard II, 2. 1. 74) ; and it seems to have been a favourite device of Ben Jonson
(e. g. Mermaid ed., 1. 15–16, 20, 68, 106, 344 ; 2. 91 ; 3. 178, 402, London, n. d.).
D

ting our God'? (p. 106); 'thou weepest for the death of thy
daughter, and I laugh at the folly of the father, for greater vanitie
is there in the minde of the mourner, then bitternesse in the death
of the deceased. But shee was amiable, but yet sinful, but she
was young and might haue liued, but she was mortall and must
haue dyed. I but hir youth made thee often merry, I but thine
age shold once make thee wise. I but hir greene yeares wer unfit
for death, I but thy hoary haires should dispyse life' (pp. 182–
183). (4) Learned allusions: 'the filthy Sow when she is sicke,
eateth the Sea-Crab, and is immediatly recured: the Torteyse
hauing tasted the Viper, sucketh Origanum and is quickly reuiued:
the Beare ready to pine licketh vp the Ants, and is recouered:
the Dog hauing surfetted to procure his vomitte, eateth grasse
and findeth remedy: the Hart beeing perced with the dart, run-
neth out of hand to the hearb Dictanum, and is healed. . . . Ah
well I perceiue that Loue is not unlyke the Figge tree, whose
fruite is sweete, whose roote is more bitter then the clawe of
a Bitter: or lyke the Apple in Persia, whose blossome sauoreth
lyke Honny, whose budde is more sower then Gall' (p. 61);
'Loue knoweth no laws: Did not Iupiter transforme himselfe
into the shape of Amphitrio to embrace Alcmæna? Into the
forme of a Swan to enioy Læda: Into a Bull to beguile Iò: Into
a showre of golde to winne Danae: Did not Neptune chaunge
himselfe into a Heyfer, a Ramme, a Floud, a Dolphin, onely for
the loue of those he lusted after? Did not Apollo conuert him-
selfe into a Shephearde, into a Birde, into a Lyon: for the desire
he had to heale his disease'? (p. 93).
Parallels such as these might be multiplied indefinitely, not
only between the Vāsavadattā and the Euphues, but between
Subandhu's romance and many other productions both of the
East and of the West.¹ Here, however, it must suffice to note
that Dunlop, in his History of Prose Fiction (ed. Wilson, London,
1896), records the following incidents and motifs which may now

¹ For parallels in the Vāsavadattā and Bāṇa's romances with The Thousand Nights
and One Night see Gray, 'Literary Studies on the Sanskrit Novel,' in WZKM.
18. 39–48.

be compared with those in the Vāsavadattā : story within story
(1. 30, 37 ; 2. 389, 405, 409, 416, 436, 438) ; love from a dream
(1. 253, 258–259, 380) ; magic steeds (1. 339, 342, 375) ; love
letters (2. 382–383 ; cf. Schlegel, ‘Feuilles d’arbre comme lettres
d’amour,’ in Mélanges Charles de Harlez, pp. 271–274, Leiden,
1896) ; and birds playing a part in affaires de cœur (2. 482). It
should also be observed that Dunlop repeatedly alludes to Indian
parallels with Occidental literature (1. 12, 64–65, 74, 76, 109–110,
210, 306, 318, 413, 459–461 ; 2. 8, 30, 49, 53–54, 81, 87–88, 90,
118–120, 127–128, 272, 509, 558–559), although the most of these
refer to the novel, or ‘short story,’ and not to the romance.
The Sanskrit Romance Compared with the Greek. The sole
national romance literature of the West which has been regarded
as connected with that of India is the Greek. The view has been
advanced by Peterson (Kādambarī, 2 ed., Introd., pp. 101–104,
Bombay, 1889, where several parallels are adduced between the
Kādambarī and Achilles Tatios), Weber (Sitzungsberichte der
Berliner Akademie, phil.-hist. Klasse, 37. 917), and Goblet
d'Alviella (Ce que l’Inde doit à la Grèce, p. 136, Paris, 1897) that
the Indian romance was a direct importation from Greece. This
hypothesis, however, is rejected by Lévi (Quid de Graecis veterum
Indorum monumenta tradiderint, p. 60, Paris, 1890) with what
seems to me to be perfect justice. It is, of course, true that many
parallels of incident and even of literary device exist between the
Sanskrit and the Greek romance. Among parallels of incident
I may note the following, restricting the list to similarities be-
tween the Greek romances and the Vāsavadattā : falling in love
from a dream (story of Zariadres and Odatis as recorded by
Athenaios, Deipnosophistai, 13. 35 ; this story is clearly of Eastern
origin ; cf. Rohde, Griechischer Roman, 2 ed., pp. 47–54, Leipzig,
1900, where the parallel with the Vāsavadattā is expressly noted ;
for dreams in general in the Greek romances see Rohde, op. cit.
pp. 508, 524, 531–533, 547, 561, 566) ; svayaṃvara, or ‘self-
choice’ of a husband (also in the story of Zariadres and Odatis ;
cf. Rohde, op. cit. p. 52) ; letters between lovers (Achilles Tatios,
Τὰ κατὰ Λευκίππην καὶ Κλειτοφῶντα, 5. 18, 20 ; Niketas Eugenîanos,
D 2

36 VĀSAVADATTĀ
Τὰ κατὰ Δρόσιλλαν καὶ Χαρικλέα, 2. 169–179, 202–223, 240–277,
284–314; 5. 199–247; Eustathios, Τὸ καθʼ Ὑσμίνην καὶ Ὑσμινίαν
δρᾶμα, 9. 8; 10. 2; cf. Rohde, op. cit. pp. 513, 566); fainting
(Chariton, Τὰ περὶ Χαιρέαν καὶ Καλλιρρόην, 1. 4; 2. 7; 3. 6; 4. 5;
8. 1; cf. Rohde, op. cit. p. 172, where the parallel with the Vāsa-
vadattā is noted); long-winded lamentations (Niketas Eugenianos,

1. 226–257, 289–352; 2. 8–23; 4. 110–150; 5. 131–168, 183–
   193; 6. 34–94, 204–234, 306–327; 8. 84–130, 196–238; 9. 37–
   107; cf. Rohde, op. cit. p. 566); and threatened suicide (Chari-
   ton, 1. 5; 5. 10; 6. 2; 7. 1; 8. 8; cf. Rohde, op. cit. p. 527). The
   following parallels of literary device, similarly restricted, may also
   be noted : stories within stories and episodes (the adventures of
   Derkullis and the sub-episode of Astraios in the lost Τὰ ὑπὲρ
   Θύλην ἄπιστα; the story of Aigialeus in Xenophon Ephesaios, Τὰ
   κατὰ Ἄνθειαν καὶ Ἀβροκόμην Ἐφεσιακά, 5. 1; and the stories of
   Knemon in Heliodoros's Αἰθιοπικά, 1. 9–14; 2. 8–10; and of Kala-
   siris, ib. 2. 24–5. 1; 5. 17–33; with the sub-episode of Charikles,
   ib. 2. 29–33); descriptions of nature (Niketas Eugenianos, 1. 77–
   115; Achilles Tatios, 1. 15; Longos, Ποιμενικὰ τὰ κατὰ Δάφνιν
   καὶ Χλόην, 4. 2–3); detailed personal descriptions (Achilles Tatios,
2. 1; Niketas Eugenianos, 1. 120–158; cf. Rohde, op. cit. pp. 160–
   166); learned allusions and citations of precedents (Longos, 1.
   16; 4. 17; Achilles Tatios, 1. 8; 2. 36–37); and alliteration,
   parisoi, homoioteleuta, and similar figures of rhetoric which cor-
   respond in a sense to the Sanskrit anuprāsa (Achilles Tatios, 1.
   1–2; 2. 7; 3. 2, 7; 4. 4; 5. 1, 21; Longos, preface; 1. 18, 30;
3. 35; 3. 3, 24, 33, 34; 4. 18; Eustathios, 1. 4; 2. 13; 4. 1, 3, 12,
   21; 5. 11; 7. 13; cf. Rohde, op. cit. pp. 552–553, 561), although
   they often form rhymes (Achilles Tatios, 1. 1, 11; Longos, pre-
   face; 2. 26; 3. 13, 23; 4. 17, 28; Eustathios, 1. 9) and strained
   compounds (Niketas Eugenianos, 1. 133; 2. 143; 3. 121; 5. 341;
4. 48; cf. Rohde, op. cit. p. 567).
   Yet all these parallels, and many more which might be cited,
   seem to me to prove nothing. In the first place, a large number
   of them can be considered parallels only by straining the sense
   of the term; and in the second place, they are obviously the out-

working of independent, though partially similar, processes in the
development of Greek and Sanskrit literature respectively, and
should be interpreted accordingly. But even were an essential
resemblance granted, it would still be difficult, I think, to prove
the dependence of the Sanskrit romance on the Greek, the latter
being, of course, the earlier. The romances of the two peoples
are totally different both in plan and in spirit, as even a cursory
reading will show. The least part of the Sanskrit romance is the
thread of the story or the adventures of its characters ; all the
stress is laid on rhetorical embellishment, minute description of
nature, detailed characterisation of exploits and of mental, moral,
and physical qualities. In the Greek romance, on the other
hand, as in Latin (if we may judge from the Satiræ of Petronius),
the story is everything. The reader is hurried from adventure
to adventure, the wilder and more improbable the better ; fine
writing is practically disregarded ; description and appreciation
of nature are, to all intents and purposes, non-existent. The only
Greek romance, it seems to me, that can, by the utmost stretch
of imagination, be compared even superficially with the works of
Subandhu and Bāṇa is the Ποιμενικά of Longos ; but even there
the sole real similarity is a longing for nature rather than for
feverish adventure, a longing which may be traced back to
Theokritos, Bion, and Moschos on the one hand, and to
Bhartṛhari and his congeners on the other. Even the Daśakumā-
racarita, which, as a picaresque romance, one might be tempted
to compare with the works of Achilles Tatios, Heliodoros, and
Chariton, has a totally different plan from any Greek romance,
tracing its ‘box arrangement’ of stories to the peculiarly Indian
scheme which may be seen, for instance, in the Pañcatantra, the
Kathāsaritsāgara, or the Jātakas, and which was later carried to
Persia, where it was incorporated in The Thousand Nights and
One Night, ultimately appearing in the Occident in the Decameron
of Boccaccio.¹ The adventures narrated in Daṇḍin’s romance of
¹ For the ‘box arrangement’ in Egyptian see King Khufui and the Magicians
(about 2000 B.C. ; tr. Maspero, Contes populaires de l’Égypte ancienne, 3 ed., pp. 23–42,
Paris, 1906) ; and in Japanese cf. Bakin’s Shichiya no Kura (1810 A.D. ; Aston,
History of Japanese Literature, pp. 359–360, London, 1901).

38 VĀSAVADATTĀ
roguery, moreover, bear no resemblance, either in plot or in
episode, to the amorphisms of Eustathios and his fellows. To
sum up, the spirit of the Sanskrit and of the Greek romancers is
as divergent as the audiences of scholars on the one hand, and of
weaklings on the other, for whom they wrote ; nor can any real
affinity, much less any direct connexion, be traced between the
romances of India and of Greece.
The Double Recension of the Vāsavadattā. Like the
Mahābhārata and the Rāmāyaṇa, the Vāsavadattā appears to be
preserved in two distinct recensions, which I would tentatively
call ‘northern’ and ‘southern.’ The first of these is represented
by the text of Hall (Calcutta, 1859) and the second by the
Madras edition of 1862, reprinted as an appendix to the present
volume, as well as by Krishnamachariar’s text (Srirangam, 1906–
1908). The ‘southern’ text, to which Hall’s manuscript D seems
also to belong, differs from the ‘northern’ chiefly in showing a
large number of interpolations, particularly śleṣas, virōdhas, and
other forms of similes appended to the series given by the
‘northern’ text. The spirit, however, of the ‘southern’ recension
is identical with that of the ‘northern.’ There is, indeed, one
omission of a complimentary allusion to the Buddhists (ed. Hall,
p. 235) and of one derogatory reference to the Jains (ed. Hall,
p. 297 ; cf. above, p. 7), but the retention of all other deprecia-
tory mentions of both sects shows that the ‘southern’ redactor
was not led to modify his text by any circumstances surrounding
him ; nor is there any added allusion in the ‘southern’ text which
seems to be specifically South Indian. It may be regarded as
practically certain that the ‘northern’ text represents Subandhu’s
original composition more closely than does the ‘southern’
recension, not only in view of the fact that the author of the
romance was, in all probability, a native of northern India, where
he almost indubitably wrote (cf. above, pp. 12–13), but also in
consideration of the familiar principles of textual criticism that
the shorter text is, other things being equal, to be preferred to
the longer, and that the more difficult reading is to be preferred
to the easier, since the ‘southern’ recension evidently seeks, by

INTRODUCTION 39
substitution or omission, to avoid many difficulties of word,
phrase, or allusion which the 'northern' text boldly retains.
Manuscripts and Editions of the Romance. For the manu-
scripts of the Vāsavadattā and its commentators reference may
be made to the catalogues of Sanskrit manuscripts enumerated
by Aufrecht (Catalogus Catalogorum, 1. 566, 726; 2. 133–134,
224; 3. 120, Leipzig, 1891–1903). The principal commentators
on the Vāsavadattā have been discussed by Hall (Introd.,
pp. 44–48), and I myself have written on 'Śivarāma's Com-
mentary on the Vāsavadattā' (in JAOS. 24. 57–63), so that the
problems there treated need not be repeated in this place.
Although the editions of Subandhu's romance will be duly
listed in the bibliography (see below, pp. 197–198), a more detailed
discussion seems appropriate here. These editions are seven in
number. The first and best-known, which represents the
'northern' recension, is that of Fitzedward Hall (Calcutta, 1859),
based on eight manuscripts and accompanied by the gloss of
Śivarāma Tripāṭhin, who flourished early in the eighteenth
century. The introduction to this edition has become a classic
among Sanskritists, and the learning which it evinces on every
page renders it amply deserving of this honour. Hall's text, how-
ever, represents a form really older than any of his manuscripts,
being that of Śivarāma, in whose favour the editor repeatedly
discarded the united authority of all his manuscripts, a method of
textual criticism which may perhaps be considered open to objec-
tion. The text of Hall has been reprinted verbatim by Jīvānanda
Vidyāsāgara (Calcutta, 1874; 3d ed., 1907). The remaining five
editions belong to the 'southern' recension. Of these the earliest
is one in Telugu script (Madras, 1861), which was reprinted, with
a correction of misprints, in Telugu (Madras, 1862) and Grantha
script (Madras, 1870). The second of these, duly collated with
the India Office copies of the other two, which Mr. F. W. Thomas,
librarian of the India Office, London, most kindly placed at my
disposal, is transcribed below (pp. 145–195). The text of this
'southern' recension is evidently later than the 'northern', since
it abounds in interpolations, especially in the lengthened ślēṣas and