Udbhaṭa (fl. c. 800) is one of the most underappreciated heroes of South Asian intellectual history. This was not always so. In the twelfth-century River of Kings (Rājataraṅgiṇī), Kalhaṇa celebrates him in no uncertain terms as the lead intellectual in what he sees as Jayāpīḍa’s self-conscious reshaping of Kashmir as a capital of letters.1 Udbhaṭa is said to have been the president of Jayāpīḍa’s star-decked royal academy that caused an acute brain-drain in all neighboring countries; he is also said to have been the recipient of an astronomical renumeration of 100,000 dinars per diem.2

  • 1Rājataraṅgiṇī 4.402–502, cf. Bronner 2013.

  • 2Rājataraṅgiṇī 4.495, cf. Bronner 2013: 173–174 n. 38. The salary of no other academy member is mentioned.

What may explain Udbhaṭa’s acclaimed status in Kalhaṇa’s eyes? We know that he is the author of at least four works in the field of Sanskrit letters. There is his Exposition (Vivaraṇa) on Bhāmaha’s Ornament of Poetry (Kāvyālaṅkāra)—the first known commentary in the field of Sanskrit poetics, and highly erudite at that. Then there is a commentary on Bharata’s vast Treatise on Theater (Nāṭyaśāstra), likely the first such work in a long line of interpretations. Both commentaries are now lost save for fragments of the former and some mentions of and citations from both. Udbhaṭa is also the author of The Essential Compendium of the Ornament(s) of Literature (Kāvyālaṅkārasārasaṅgraha), a short, versified manual on alaṅkāras that seems to be fully extant, as well as of his accompanying Kumārasambhava , some 94 verses of which—all in seemingly continuous narrative—were preserved as the illustrations of his Compendium. This is not the place to discuss the combined impact of this corpus in Sanskrit poetics. Let me just note that Udbhaṭa forever changed the field’s course: in particular, by leading its deep foray into semantics and its partial merger with dramaturgy.3 From this point onward, when writers wanted to present their ideas as innovative, they would contrast them with those of Udbhaṭa or of “the followers of Udbhaṭa” (audbhaṭāḥ).4 To stake a claim in post-ninth century Kashmiri poetics meant the contender first had to emerge from Udbhaṭa’s long shadow.

  • 3Bronner 2016; Bronner forthcoming.

  • 4I cannot provide here a complete survey of the mentions and citations of Udbhaṭa’s works (the latter are often unnamed) in subsequent works on poetics. Let me just briefly mention that in the opening sentence of his Alaṅkārasarvasva, which starts with a historical prelude, Ruyyaka says “Here, to begin with, the ancient makers of ornaments, Bhamaha, Udbhaṭa, and the others” (iha hi tāvad bhāmahōdbhaṭaprabhr̥tayaḥ cirantanālaṅkārakārāḥ…; AS p. 3). The term audbhaṭāḥ in the plural is found, for example, in Abhinavagupta’s Abhinavabhāratī. See Treatise on Theater, p. 258 of vol. 1.

Moreover, there is a growing agreement that the Udbhaṭa who authored works of poetics, poetry, and dramaturgy is none other than the one who composed treatises on the heterodox materialist Cārvāka philosophy (and its relations to logic) and on grammar, thus making him an authority on an astonishing array of disciplines.5 I briefly return to this in my conclusions.

If Udbhaṭa, as I propose, is one of the unsung heroes of South Asia’s history of thought, this is epitomized by his Kumārasambhava. To the best of my knowledge, not a single study of this work exists. In fact, I doubt very much that, in the century and a half since modern scholars have discovered Udbhaṭa’s Compendium, anyone has even read his illustration verses as a continuous poem and examined them as such.6 I believe that this is not a coincidence, and that the neglect of modern scholars reflects that of their premodern predecessors. Udbhaṭa’s Kumārasambhava was intentionally ignored.

  • 6One exception is Narayana Daso Banhatti, who in his introduction to his edition of the Compendium dedicates two pages (Banhatti 1982: xiv–xv) to the poem, and notes that “sometimes the natural order of verses seems to be changed.” He also provides in an appendix “[a] list of all examples given by Udbhaṭa in the order in which they occur in the text” (appendix iii, p. xii). But other than a brief summary of the text and a note that it is “a great pleasure to read the poem even more than once” (xiv–xv), he offers no analysis.

The purpose of the current essay is therefore simple: to begin the process of thinking about Udbhaṭa’s forgotten poem. This I do in four sections. In the first, I discuss its status as an independent work and briefly introduce the narrative structure of the verses preserved in the Compendium. The second section consists of a translation of these verses, excerpted from my forthcoming reader on Sanskrit figurative theory.7 But whereas there, as in all previous treatments, the poetic illustrations are subordinated to the figurative phenomena they serve to illustrate, here I let the reader appreciate the poetry without interruption (references to and explanations of the relevant alaṅkāras are relegated to the footnotes). The third section is dedicated to an initial analysis of some aspects of the poem, and finally, the fourth, to speculations about the reasons that might have led to its cold reception.

Udbhaṭa’s Kumārasambhava as an Independent Poem

Are Udbhaṭa’s verses part of a larger independent poem? They are certainly not presented as such in the Compendium. There is no reference to a larger poem, and there is no marked beginning or end. That said, the Compendium itself is also strangely not introduced by its author: there is no benediction, and no opening verse presenting the work’s title or its goals. TheCompendium simply begins unannounced with the list of its first group of ornaments, and it ends unceremoniously sixty-something verses later with the last ornament in the sixth and final group.

The commentators, for their part, see their task as explicating just one work, the Compendium. For example, the third and last verse in Pratīhārēndurāja’s brief introduction of his commentary (c. 900) simply states that having studied with his learned teacher Mukula, the author now has elucidated the Compendium. He does not acknowledge here that the examples in Udbhaṭa’s work form a cohesive narrative that belong in a separate work, and that, therefore, commenting on the Compendium also means commenting on Udbhaṭa’s poem. Thus, when he comes to the example of the first ornament, he merely introduces it by saying: “here is its example” (tasyōdāharaṇam), a style he repeats throughout. The approach of Tilaka (c. 1100), who knew his predecessor’s commentary, is identical.

The commentators are, however, aware of Udbhaṭa’s poem as a separate entity. This we learn a bit later when they address a minor textual discrepancy in the Compendium. In the opening of the work, where the first group of ornaments is listed (in what forms, in essence, the chapter’s table of contents), “simile” (upamā) precedes “illumination” (dīpaka) but later, when the ornaments are defined and illustrated, “illumination” precedes “simile.” Why this change of order? Pratīhārēndurāja raises this objection himself and answers as follows:

anēna granthakrtā svōparacitakumārasambhavaikadēśō ’trōdāharaṇatvēnōpanyastaḥ. tatra pūrvaṁ dīpakasyōdāharaṇānī. tadanusandhānāvicchēdāyātrōddēśakramaḥ parityaktaḥ. uddēśas tu tathā na kr̥tō vr̥ttabhaṅgabhayāt. ēvam uttaratrāpi lakṣaṇēṣūddēśakramānanusārēṇa samādhir vācyaḥ.

Kāvyālaṅkārasārasaṅgraha (Banhatti 1982: 16, ll. 22–26)

A portion of our author’s own poem, Kumārasambhava, is used here for illustration. There, the illustrations of “illumination” come first. So as not to depart from that poem’s flow, he abandoned the order found in the [Compendium’s] initial list. In fact, the only reason that the order in that list is not changed accordingly is to avoid breaking the meter. Later, too, we will likewise state adjustments by ignoring the original list when it comes to the actual definitions.

Pratīhārēndurāja asserts that there is a separate work by “our author,” that it has its own title, Kumārasambhava, that it includes more than is found in the Compendium (only a “portion” from it “is used here for illustration”), that it has its own “flow,” or integral sequence, and that this sequence overrides whatever order is found, for purely metrical reasons, in the tables of contents of the different chapters of the Compendium. Indeed, it seems important for him to convey this information. After all, referring to the metrical constraints of the initial list of ornaments is a perfectly sufficient explanation for the noted change in the order of their illustration. It thus may be that Pratīhārēndurāja was looking for an opportunity to acknowledge the existence of this poem qua poem and then to quickly move on. Tilaka, the other commentator, says something to the same effect: “Although ‘simile’ was listed first for metrical considerations, it is defined after ‘illumination’ in agreement with the sequence of the poem he composed, entitled Kumārasambhava . In later changes in order, too, the same distinction is to be realized.”9 Note that Tilaka does not mention that the poem was larger, though it is unclear if this omission has any significance.10

  • 9Kāvyālaṅkārasārasaṅgraha (Ramaswami Sastri 1931: 12, ll. 15–16): vr̥ttānurōdhād upamā prāgupadiṣṭāpi svayaṁkr̥takumārasambhavākhyakāvyasaṁgatyanurōdhād dīpakasya paścāl lakṣitā. parivartanē cōttaratrāyam ēva viśēṣō jñēyaḥ.

  • 10Note that, in Pratīhārēndurāja’s reading, the order in which Udbhaṭa illustrates the five subcategories of Gujarati alliteration is also different from that in which he defines them (verses 6-10 below; Tilaka, however, provides the examples in the order of the categories). Here a different order would have changed little in the meaning of the poem, and hence Pratīhārēndurāja’s insisting on a different order may have been inspired by having access to a manuscript of the poem itself, but this is of course speculative and inconclusive.

Some later writers also knew about Udbhaṭa’s poem. An example is Māṇikyacandra, who cites an illustration verse from the Compendium in his commentary on Mammaṭa’s Light on Literature (Kāvyaprakāśa). In introducing it he says: “For example, in the praise of Gaurī in Udbhaṭa’s Kumārasambhava” (yathōdbhaṭakumārasambhavē gaurīstutau).11 So Māṇikyacandra refers to Udbhaṭa’s poem by its name, and for him, it is the source of the citation rather than the Compendium. Note also the need to insert the name Udbhaṭa into the compound naming the source. Just to say Kumārasambhava is to refer strictly to Kālidāsa’s far more famous and original work of the same name, where no mention of the author is needed, whereas a reference to the less celebrated competitor work requires differentiation, if not a trigger warning.12 That said, the verse cited here is from the part of the poem illustrated in the Compendium. Did Māṇikyacandra have access to a more complete, separate copy? 13 All I can say is that this is a rare named citation of Udbhaṭa’s poem, and that I know of no such explicit reference to any verse that is not already included in the Compendium.14 In other words, whatever portions that preceded and followed the section preserved in the Compendium were, they disappeared without trace.15 This disappearance calls for an explanation, and I come back to this in my conclusions.

  • 11Kāvyaprakāśa, p. 252, l. 18.

  • 12For example, Dhvanyālōka p. 539: yathā bhagavatī pārvatī kumārasambhavē.

  • 13He cites two illustration verses from Udbhaṭa’s stanzas already in the Compendium, both already precited: the first (nētraiḥ, p. 228) by multiple sources, and the second at least by Sōmēśvara (Saṅkēta, p. 294, although without the explicit reference to Udbhaṭa that Māṇikyacandra gives). On the relative chronology of Māṇikyacandra and Sōmēśvara, see Bronner and Ollett 2024.

  • 14A. K. Warder noted in passing that of the few verses attributed to Udbhaṭa in the anthologies, “one describing Umā,” “is probably from his Kumārasambhava” (Warder 1983: 472). Note, however, that the verse is typically cited amidst others that depict Brahma’s over-the-top investment in the beauty of certain young women, and that it, too, does the same, without naming its heroine (see, for example, Subhāṣitaratnakōṣa of Vidyākara v. 455). Thus, while the verse depicts the young woman’s creation as offering a new life to Kāma whom Śiva has burnt, there is nothing to force us to see it as belonging in the narrative of Śiva and Umā’s falling in love, let alone to be from Udbhaṭa’s lost poem on this topic.

  • 15One unique reference to Udbhaṭa’s poem is in the Telugu Kumārasambhavamu of Nanne Cōḍa: “Udbhaṭa composed a Kumāra-sambhava / on the theme of Śiva’s play / and pleased the god with this poem, / which is the whole of figuration / with kāvya deep inside” (translation from Narayana Rao and Shulman 2002: 116). A study of the relationship between this Telugu poem and Udbhaṭa’s work is a desideratum, but note that here, too, the reference is only to the illustration of ornaments, that is, presumably the part preserved in the Compendium.

What, then, do we find in Udbhaṭa’s verses in the Compendium? To begin with, they are all in the same textbook carrying meter (anuṣṭubh). To A. K. Warder, this suggested that they originally belonged in the same chapter but without that chapter’s closing verses, usually set aside by the introduction of a new meter.16 Moreover, the text begins in media res, its first word being tadāprabhr̥ti: "from then on," not exactly the most common opening line of a Sanskrit work. The reader is thrown right into the middle of the Śiva-Umā love story, at a point that, we later realize, comes soon after Śiva has burnt the love god Kāma to ashes, thereby causing the lovelorn Umā to begin her penance. Indeed, in Kālidāsa’s intertext, the same phrase, tadāprabhr̥ti, begins a verse that depicts Umā’s love madness following that very traumatic event (here in Hank Heifetz’s translation):

From then on, filled with love, the curls of her hair
dusty gray from the sandal paste smeared on her forehead
to cool her, she could never find relief even
lying on the high mountain ice of her father’s home.

KKS 5.55 (here in a flashback, when Umā’s friend narrates this to the sage who is really Śiva in disguise), Heifetz 1985: 76.

In Udbhaṭa’s namesake poem, however, the same phrase marks the depiction not of Umā’s suffering but of Śiva's, as we shall see shortly. I believe that this is not a coincidence, and I would like to draw some initial conclusions from this choice of words. First, that with it, Udbhaṭa wishes to juxtapose his poem with that of his great predecessor. Second, if I am right in my interpretation of the adverbtadāprabhr̥ti as keyed to Kālidāsa’s, it may be that the Compendium’s commentator, Pratīhārēndurāja, was wrong in taking it as meaning “ever since the separation from Satī,” an event that took place much earlier in Śiva’s story.18 If this is correct, and if the section just prior to what we have preserved narrated the burning of Kāma, this may indicate that already Pratīhārēndurāja did not necessarily have access to Udbhaṭa’s complete poem, if one indeed existed, leading him to misidentify the immediate context. This is only a speculation, but it is one worth keeping in mind.

  • 18Pratīhārēndurāja on verse 1 (UKS p. 3): tadāprabhr̥ti satīviyōgād ārabhya. Tilaka, by contrast, rightly identifies this moment as following the burning of Kāma (kāmadahanād ārabhya, p. 4, l. 1; he also has a slightly different reading: tataḥ prabhr̥ti).

At any rate, with this unheralded opening mid-plot, the first 30 verses, corresponding to the first group of ornaments in Udbhaṭa’s Compendium, are dedicated to two primary topics: Śiva’s lonely meditation and the onset of autumn. It is autumn rather than apring that is described here as enticing and, hence, posing a threat to God’s yogic resolve.19 At some point during this description, the poet switches without warning to worries and self-doubts in Śiva’s voice, before turning to the narrator, only to shift back, in the end of the section, to God’s inner thoughts. These carry over to the next section of 16 verses, illustrating the Compendium’s second group of ornaments—note that there is no clear narrative or other boundary between the verses attached to the Compendium’s different sections. Here Śiva eventually focuses on Umā in meditation, realizes that she has become an ascetic, and decides to seek her out in disguise. He finds her in Gaurī Peak (gaurīśikhara), another direct echo of Kālidāsa’s namesake poem.20 Umā is the focus of his gaze, inner as well as actual, and the verses continue to alternate between the narrator’s and God’s voice.

  • 19Note that in Kālidāsa’s version, the burning of Kāma is preceded by the advent of the spring season, yet out of order (ākālikīm … madhupravr̥ttim; KKS 3.34). The cycle of seasons depicted in Umā’s penance in Kālidāsa’s poem seems to bypass the autumn.

  • 20KKS 5.7; UKS 37.

The description of ascetic Umā carries over to the 4 verses that are used to illustrate the three ornaments of the Compendium’s short third section. The last of these verses is a depiction of a fawn Umā befriends. This fawn is also the topic of the first verse in the fourth section, consisting of 14 verses, again indicating that the verses have their own flow that is not necessarily dictated by the logic of the manual. Here the focus immediately shifts from Umā to Śiva, whose dramatic emotional reaction to her sight the narrator turns to depict. Śiva has a hard time restraining himself, but once he manages to regain composure, he greets her, heaps praise on her father, Himālaya (with echoes of Kālidāsa’s opening chapter), and identifies her as this mountain’s daughter.

The last two sections, 18 verses each, all continue the speech of God while still in disguise. Śiva tries to talk Umā out of her austerities, which, he says, do not befit her delicate beauty. Somewhere midpoint the fifth section he launches into a long laudation of Viṣṇu, technically a set of relative clauses that keeps the identity of the target of praise at bay, until the main clause is provided, verses later, in the sixth and final section. Here Śiva notes that even Viṣṇu, whom he had praised at length, would have chosen Umā as his consort. With this strange compliment he nears the end of his speech, and the poem, as we have it, ends with a wry recommendation: “Go get yourself a husband.”

A note on the translation method used here and the text that serves as its basis. Since the point of this essay is to examine the poem itself, the translation provided below is continuous, without the definitions that frame the verses in the Compendium. Readers who are interested in reading the verses as illustrations of their respective ornaments, are advised to consult my forthcomingReader, where they are translated side by side with their definitions. I also tried my best to translate the verses as poetry, taking some liberties and avoiding over-literalization. That said, I did my best to capture the figurative effect each verse is meant to illustrate, an attempt that is often prone to failure, especially when it comes to sound effects and wordplays that are language specific. I nonetheless tried to recreate these effects using lexical choices and sounds in the target language. To offset the consequences of these choices, I provide, for each verse, a somewhat lengthy footnote, consisting of the Sanskrit original, an explanation of the figure it serves to illustrate, and, in cases of considerable departures, an explanation of the liberties I took while translating it. This annotation notwithstanding, I strongly recommend reading the translation uninterruptedly, verse after verse, as it is primarily based on such reading that the arguments in the following sections hinge.

As for the text itself, some issues need to be flagged. First, for the sake of convenience, I number the verses consecutively, from 1 to 94, although this exact enumeration is not found in any of the printed edition (in the footnotes, I also provide the verse numbers in the existing editions). Second, there are textual discrepancies in wording and in the sequence of the verses, and these are also noted. Finally, it is not always clear to me that a second version of one and the same verse that is repeated with only very minute changes, to fit another subvariety of the same figure, is in the voice of the author, Udbhaṭa, or of his commentators—the two main ones relied on here are Pratīhārēndurāja (hereafter PIR) and Tilaka (T)—who try to make sure that the text of the poem complies perfectly with the classificatory apparatus of the Compendium. In the absence of anything like a critical edition of the text (Banhatti’s edition of the text with PIR’s commentary is based primarily on a single manuscript), I made my own judgment calls, and these, too, are noted as such in the notes.

Here, then, is the translation of the verses in sequence.

Udbhaṭa’s Kumārasambhava: A Translation

From then on withdrawn, concealed in the hide
of a mammoth elephant, the god whose throat
was blackened, and who was charred
by the grief of Satī’s demise, killed time.21 (1)
In a strict crypt on the brow of the world,
with only his fiend friends—a wild lot
if ever there was one—he
whiled away his days.22 (2)
Then, when even shallow plashes brandished
shoots, shoal to shore, and, in every direction,
sheets of brownish awns shimmered—
autumn showed up in a flash.23 (3)

  • 21tadāprabhr̥ti niḥsaṅgō nāgakuñjarakrttibhr̥tśitikaṇṭhaḥ kālagalatsatīśōkānalavyathaḥ[PIR 1.*1; T 1.4; T reads tataḥ for tadā]

    The ornament is “apparent repetition” (punaruktavadābhāsa), likely Udbhaṭa’s own invention. The idea is that, at first, certain phrases appear redundant, and then, upon further reflection, the reader realizes that the second of each pair has a unique, different meaning or reading (Bronner 2016: 113–114). In this verse, both nāga and kuñjara seem, at first blush, to signify “elephant,” until one realizes that the latter here means “the best of its kind”; and both śitikaṇṭhaḥ and kālagala- seem to signify “black throat,” until one realizes that the latter is really kālagalat-, referring to the gradual passing of the pain of separation by time. I have tried to recreate the effect in English (at the cost of deviating from the literal meaning of the second half), with “mammoth elephant,” “concealed … hide,” and “blackened … charred.”

  • 22sa dēvō divasān ninyē tasmiñ śailēndrakandarēgariṣṭhagōṣṭhīprathamaiḥ pramathaiḥ paryupāsitaḥ[PIR 1.*2; T 1.5]

    The ornament is “enticing alliteration” (chēkānuprāsa), where the effect is based on several pairs of similar sound patterns (as in sa dēvō divasān etc.). Again, I tried to recreate the effect in English with partial success.

  • 23tatra tōyāśayāśēṣavyākōśitakuśēśayācakāśē śālikiṁśārukapiśāśāmukhā śarat[PIR 1.*3; T 1.7]

    The ornament here is “alliteration” (anuprāsa), which Udbhaṭa divides according to the various euphonic modes. The verse here falls under the “harsh euphonic mode” (paruṣavr̥tti), which abounds in ś and ṣ sounds, consonant clusters that include r, retroflex consonants, and combinations such as hl, hv, and hy. The example verse is primarily dominated by sh sibilants, as is my translation.

Bands of bees suddenly landed
on the blooming bundles, indulging
in the dense nectar that inundated
in uncommon abundance.24 (4)

  • 24sāndrāravindavrndōtthamakarandāmbubindubhiḥsyandibhiḥ sundarasyandaṁ nanditēndindirā kvacit[PIR 1.*4; T 1.9]

    The second type of alliteration is based on the “urbane euphonic mode” (upanāgarikavr̥tti), which is replete with duplicated letters or consonant clusters that contain nasals followed by homorganic stops. The original is dominated by the nd cluster, something I tried to replicate in the translation.

What a lovely hullabaloo: long lines
of bumblebees playing wiles in the lotus beds.
It clanged like the jingling anklets
on Lady Blossom’s legs.25 (5)

  • 25kēlilōlālimālānāṁ kalaiḥ kōlāhalaiḥ kvacitkurvatī kānanārūḍhaśrīnūpuraravabhramam[PIR 1.*5; T 1.11]

    The third and last type of alliteration is based on the “rustic euphonic mode” (grāmyavr̥tti), which is said to abound in the remaining sounds as appropriate. In the Sanskrit the predominant sound is l, which I tried to replicate in the English.

White grass like white grass
shone, lakes like lakes,
and rivulets stole young hearts
just by being rivulets.26 (6)

  • 26kāśāḥ kāśā ivōdbhāṁsi sarāṁsīva sarāṁsi cacētāṁsy ācikṣipur yūnāṁ niṁnagā iva niṁnagāḥ[PIR 1.*6; T 1.19; T reads ābhānti for udbhāṁsi]

    This is the first example of the so-called “Gujarati alliteration” (lāṭānuprāsa), the exact repetition of stems or words with no change in form or meaning, so long as each instance serves a different purpose. This ornament has five subtypes. Here is illustrated the fourth subtype, where standalone words such as “white grass” (kāśāḥ), “lakes” (sarāṁsi) and “rivulets” (niṁnagāh) get repeated. Note that T gives the verses illustrating “Gujarati alliteration” in a different order that agrees with that of their definition.

Wives whose husbands strayed big time
made no angry scene.
And husbands, too, if their wives strayed,
made no angry scene.27 (7)

  • 27striyō mahati bhartrbhya āgasy api na cukrudhuḥbhartārō ’pi sati strībhya āgasy api na cukrudhuḥ[PIR 1.*7; T 1.20]

    “Gujarati alliteration” type 5, when the entire metrical unit is repeated.

There were full-blown-lotuses
and lotus-baffled-bees. The air was filled
with bee-hum and the humdrum
of frantic cranes.28 (8)

  • 28kvacid utphullakamalā kamalabhrāntaṣaṭpadāṣaṭpadakvāṇamukharā mukharasphārasārasā[PIR 1.*8; T 1.16]

    “Gujarati alliteration,” type 1, namely, the repetition of two identical words (kamala, ṣaṭpada, mukhara) while each forms part of a different compound. I partially replicated this in the translation with “lotus” and “hum.”

With filaments unlike the filaments of any other flower,
dark blue water lilies unfolded,
becoming earrings on the moon
that women wear for a face.29 (9)

  • 29jitānyapuṣpakiñjalkakiñjalkaśrēṇiśōbhitamlēbhē ’vataṁsatāṁ nārīmukhēnduṣv asitōtpalam[PIR 1.*9; T 1.18]

    “Gujarati alliteration,” type 3, where the repeated words form part of a single compound. In the original, the first line (“With filaments unlike the filaments of any other flower”) constitutes a single compound word, something that I did not replicate in the English.

And it was gander-time: white ganders,
fervently nesting in lotus-ponds,
made the ponds jut out, as it were,
with numerous milk teeth.30 (10)

  • 30padminīṁ padminīgāḍhaspr̥hayāgatya mānasātantardanturayām āsur haṁsā haṁsakulālayāt[PIR 1*10; T1.17, reads padminīḥ]

    “Gujarati alliteration” type 2, when only one of the repeated pair of words is part of a compound and the other stands alone (eg., padminīm vs. padminī-). I replicated this with “gander” and “pond.”

Pouring moonlight-water
from moon-jars, the night-maidens
gradually sprinkled the sky-garden,
whose blossoms are stars.31 (11)

  • 31jyōtsnāmbunēndukumbhēna tārākusumaśāritamkramaśō rātrikanyābhir vyōmōdyānam asicyata[PIR 1.*11; T 1.23]

    The figure is “identification” (rūpaka) of the “full-set type” (samastavastuviṣaya), wherein every subject (moonlight etc.) is explicitly identified with its standard (water etc.). For a discussion of Udbhaṭa’s groundbreaking understanding of rūpaka, see Bronner 2016: 92–99, 106–110, and Bronner forthcoming.

With regal geese soaring up and down,
bearing a white mop of feathers,
autumn was fanning
the kings of lakes.32 (12)

  • 32utpatadbhiḥ patadbhiś ca picchālīvālaśālibhiḥrājahaṁsair avījyanta śaradaiva sarōnr̥pāḥ[PIR 1.*12; T 1.24; reads piñcha for piccha]

    This is the second of four types of “identification,” one that is “confined in presence” (ēkadēśavivarti). Here some of the identifications are explicit (e.g., the king is explicitly identified with the lake), but some are only implied (e.g., autumn as the king’s whisk-lady).

Braids of forest deities, sword blades
of Love’s special troops, or Death’s iron shackles
for forsaken wives—rows of black bees
were everywhere.33 (13)

  • 33vanāntadēvatāvēṇyaḥ pānthastrīkālaśr̥ṅkhalāḥmārapravīrāsilatā bhr̥ṅgamālāś cakāśirē[PIR 1.*14; T 1.26]

    This is “chain identification” (mālārūpaka), the third subtype of rūpaka. Here the same subject (“rows of black bees”) is repeatedly identified with a variety of standards (braids of forest deities etc.).

Then the white clouds
illuminating the horizon
poured a rain of arrows to redeem
the kingdom of heaven.34 (14)

  • 34āsāradhārāviśikhair nabhōbhāgaprabhāsibhiḥprasādhyatē sma dhavalair āśārājyaṁ balāhakaiḥ[PIR 1.*14; T 1.27]

    This is the last type of “identification,” “confined in capacity” (ēkadeśavr̥tti). This is a somewhat obscure category, and it partly depends on the dual meaning of the verb prasādhyate, which could be taken to have both aesthetic and military connotations. I tried to replicate this ambiguity with “redeemed.” See Bronner 2016: 108–109 and Bronner forthcoming for a discussion.

Autumn liquidated all assets
of the kadamba blossoms,
as well as the entire store of happiness
of women far away from their love.35 (15)

  • 35saṁjahāra śaratkālaḥ kadambakusumaśriyaḥprēyōviyōginīnāṁ ca niḥśēṣasukhasampadaḥ[PIR 1.*15; T 1.29]

    This is the first of three illustrations of dīpaka, or “illumination,” an ornament based on ellipsis. Here the verb (“liquidated”) appears only once yet construes with two separate objects (“the kadamba blossoms” and “the store of happiness” of separated women). It is found in the beginning, which makes it a case of “beginning illumination” (ādidīpaka).

Life in foreign parts, and the sorry sight
of women whose men are away,
crushed the hearts of travelers,
as did this autumn.36 (16)

  • 36vidēśavasatir yātapatikājanadarśanamduḥkhāya kēvalam abhūc charac cāsau pravāsinām[PIR 1.*15; T 1.30]

    In this “middle illumination,” the verbal element “crushed the hearts of travelers” (duḥkhāya kēvalam abhūt) is situated in the middle, from where it construes with different entities (“Life in foreign parts,” etc.) in different parts of the verse.

Then, overflowing with radiance
and beaming with endless allure,
the moon and women’s moon-faces
filled every heart with bliss.37 (17)

  • 37tadānīṁ sphītalāvaṇyacandrikābharanirbharaḥkāntānanēndur induś ca kasya nānandakō ’bhavat[PIR 1.*17; T 1.31]

    In this “end illumination,” filling every heart with bliss, which construes with both moon-faces and the actual moon, is found in the end.

Pure moonlight sparked instant fever
and added fuel to the flames
of love for the lonely, as if it were
cool sandalwood lotion.38 (18)

  • 38kṣaṇaṁ kāmajvarōtthityai bhūyaḥ santāpavr̥ddhayēviyōginām abhūc cāndrī candrikā candanaṁ yathā[PIR 1.*18; T 1.39]

    This is the first of many verses illustrating “simile” (upamā), here with explicit syntax and the word yathā (“as if it were”).

Dark nympheas like eyes, lotuses
like faces—shimmering lakes shone like girls
coming of age, and the sheldrakes
were like their budding breasts.39 (19)

  • 39nētrair ivōtpalaiḥ padmair mukhair iva saraḥśriyaḥtaruṇya iva bhānti sma cakravākaiḥ stanair iva[PIR 1.*19; T 1.40]

    “Simile,” explicit syntax, this time using the word “like” (iva). Note that this is what Daṇḍin called “reverse simile” (viparyāsopamā), where the established standards become subjects and vice versa (KĀ 2.17–18). In Udbhaṭa’s purely formal analysis of the simile, the identity of the standard is immaterial, but he does gesture, by way of examples, that such reversals are possible.

Fully open, white at night,
filaments busy with bees—
the lily pond became
the full moon’s equal.40 (20)

  • 40prabōdhād dhavalaṁ rātrau kiñjalkālīnaṣaṭpadampūrṇēndubimbapratimam āsīt kumudakānanam[PIR 1.*20; T 1.41, reads: śaśāṅkabimbēna samaṁ babhau]

    “Simile,” nonexplicit syntax, using the word “equal” (here pratimā). PIR offers the following reading as an alternative to the second half: akhaṇḍēnēndunā tulyam āsīt kumudakānanam (which he enumerates 1.*29), likely to squeeze out of this example verse an additional linguistic method of expressing the same simile. T’s alternative reading (śaśāṅkabimbapratimam) serves a similar purpose. I ignore both alternatives in my translation.

“Will she somehow present herself—
her perfect face to my thirsty eyes—
emerging miraculously, rain-like
from cloudless skies?41 (21)

  • 41api sā sumukhī tiṣṭhēd dr̥ṣṭēḥ pathi kathañcanaaprārthitōpasampannā patitānabhravr̥ṣṭivat[PIR 1.*22; T 1.43; the following verse precedes this one in T’s edition]

    This verse continues the manifold analysis of simile, here using the nominal suffix vat (“-like”), and having the actual semblance implied.

Is she, too, aching boundlessly
me-like, another victim
of the sudden savagery
of Love?”42 (22)

  • 42kiṁ syur utkalikā madvat tasyā api nirargalāḥakāṇḍōḍḍāmarānaṅgahatakēna samarpitāḥ[PIR 1.*23; T 1.42]

    This is an illustration of a simile where the actual semblance is explicit, based again on the nominal suffix vat.

Thus, when the air was filled
with the gripping cries of geese,
the tiger god of all gods,
hair tangled in knots, remorsefully
set his thoughts on Gaurī,
with her moon-hue-face,
lily-petal-eyes, and the gold
of a lotus cup.43 (23–24)

  • 43iti kālē kalōllāpikādambakulasaṅkulētridaśādhīśaśārdūlaḥ paścāttāpēna dhūrjaṭiḥtāṁ śaśicchāyavadanāṁ nīlōtpaladalēkṣaṇāmsarōjakarṇikāgaurīṁ gaurīṁ prati manō dadhau[PIR 1.24-25; T 44-45]

    This illustrates cases of simile which are concise, in the sense that various elements are elided thanks to the use of compounds. There is a play on the word gaurī (both the color in question and the goddess’s personal name).

He had it made, rich with power
unbound, yet he was vagabonding.
The garden that offers all one could wish for
was hell all around.44 (25)

  • 44sa duḥsthīyan kr̥tārthō ’pi niḥśēṣaiśvaryasampadānikāmakamanīyē ’pi narakīyati kānanē[PIR 1.*26, T 46]

    The similes in this verse likewise illustrate the elision of the word such as “like."” Here this is done thanks to the denominative kyac suffix, an effect which I tried to replicate in my translations (“vagabonding,” “was hell”).

Without her the world was a-pyring,
and right before his gazing eyes
the mighty light of perfect knowledge
began to firefly.45 (26)

  • 45kr̥śānavaj jagat tasya paśyatas tāṁ priyāṁ vinākhadyōtāyitum ārabdhaṁ tattvajñānamahāmahaḥ[PIR 1.*27; T 1.47]

    The similes in this verse likewise illustrate various elisions based on the kvip and kyaṅ denominative suffixes. I tried to recreate the effect with “a-pyring” and “began to firefly.”

Scorching the heart of everyone,
the flames of Love scorched his,
which was drawn to Umā
by the pull of her penance.46 (27)

  • 46tasyētaramanōdāham adahat prajvalanmanaḥumāṁ prati tapaḥśaktyākr̥ṣṭabuddhēḥ smarānalaḥ[PIR 1.*28; T 1.48]

    In this simile, a ṇamul ending (dāham) implies the word “like.” As Pratīhārēndurāja notes [p. 26, l. 16]: “Here the heart of everyone else, in the sense of ordinary people, is the standard of comparison, and Śiva’s heart, the subject. Being scorched is the shared attribute.”

Love, whose limbs he set ablaze,
mustered courage, touched him nonetheless.
He kept thinking the thought
of every ordinary man: 47 (28)

  • 47sa dagdhavigrahēṇāpi vīryamātrasthitātmanāspr̥ṣṭaḥ kāmēna sāmānyaprāṇicintam acintayat[PIR 1.*29; T 1.49]

    In this simile, a ṇamul ending referringto an agent (cintam) suggests the word “like.”

“When I, all but a pariah,
burned Love to ashes and disappeared,
did she, consumed by peerless anguish,
die of grief, then and there? 48 (29)

  • 48caṇḍālakalpē kandarpaṁ pluṣṭvā mayi tirōhitēsaṁjātātulanairāśyā kiṁ sā śōkān mr̥tā bhavēt[PIR 1.*30; T 1.50]

    Simile based on the suffix kalpa, here translated as “all but.”

Such riches of character and beauty
are rare in this world.
How many nights in a year
have a perfect full moon?49 (30)

  • 49viralās tādr̥śō lōkē śīlasaundaryasampadaḥniśāḥ kiyatyō varṣē ’pi yāsv induḥ pūrṇamaṇḍalaḥ[PIR 1.*31; T 1.53]

    This last example of simile is based on parallelism (prativastūpamā). This is probably still Śiva speaking in his own voice about the beauty and character of Umā.

O the power of Love! Even Rudra
is in such a state… But enough of that!
Can one take the measure of the ocean
by a bucket?”50 (31)

  • 50ahō smarasya māhātmyaṁ yad rudrē ’pi daśēdr̥śīiyad āstāṁ samudrāmbhaḥ kumbhair mānē tu kē vayam[PIR 2.*1; T 2.4]

    This is the first of two illustrations of “dismissal” (ākṣēpa). In this variety, the dismissal is future oriented.

Such worry led Him to further worries
without end. A wonder! But wait:
Isn't musing on love endless
like time itself? 51 (32)

  • 51iti cintayatas tasya citraṁ cintāvadhir na yatkva vā kāmavikalpānām antaḥ kālasya cēkṣitaḥ[PIR 2.*2; T 2.5]

    This “dismissal” is oriented towards what was already stated.

After all, there’s nothing a guy won't do,
when hell-bent on a mission.
Even the God of the Mighty Bow,
became a schoolboy.52 (33)

  • 52tan nāsti yan na kurutē lōkō hy atyantakāryikaḥēṣa śarvō ’pi bhagavān baṭūbhūya sma vartatē[PIR 2.*3; T 2.8]

    This is the first example of the ornament “citing another case” (arthāntaranyāsa). In this variety, the corroborating sentence (the general statement about “"a guy” doing whatever it takes) appears before the corroborated (Śiva’s taking on the guise of a schoolboy), and there is use of the particle hi (“after all”). T reads lōkē for lōkaḥ.

To see into a woman’s heart,
one must go undercover.
Thus the God of Tangled Dreadlocks
set out in a schoolboy’s body.53 (34)

  • 53pracchannā śasyatē vr̥ttiḥ strīṇāṁ bhāvaparīkṣaṇēpratasthē dhūrjaṭir atas tanuṁ svīkr̥tya bāṭavīm[PIR 2.*4; T 2.9]

    The second example of “citing another case,” corroboration again first, but no use of the particle hi.

He then began to meditate,
fixing himself on Himself.
The eye, after all, may err,
but not the inner eye.54 (35)

  • 54harō ’tha dhyānam ātasthau saṁsthāpyātmānam ātmanāvisaṁvadēd dhi pratyakṣaṁ nirdhyātaṁ dhyānatō na tu[PIR 2.*5; T 2.10]

    “Citing another case,” corroboration statement later, use of the particle hi.

He saw Umā inflicting on Herself
the harshest torture.
To win an inconceivable match,
a girl has no other option.55 (36)

  • 55apaśyac cātikaṣṭāni tapyamānāṁ tapāṁsy umāmasambhāvyapatīcchānāṁ kanyānāṁ kāparā gatiḥ[PIR 1.*6; T 2.11]

    “Citing another case,” corroboration later, no hi.

He came to Gaurī Peak and saw Umā:
an ascetic, all skin and bones, shaming
the sliver left over when Rāhu
gobbled down the moon.56 (37)

  • 56sa gaurīśikharaṁ gatvā dadarśōmāṁ tapaḥkr̥śāmrāhupītaprabhasyēndōr jayantīṁ dūratas tanum[PIR 2.*7; T 2.13]

    The figure is “distinction” (vyatireka), and in this case, the cause for the subject’s outdoing its standard is implied.

The lackluster day lotus at night,
and the moon, lusterless by day—
her face disgraced both with a sheen
that shone nonstop.57 (38)

  • 57padmaṁ ca niśi niḥśrīkaṁ divā candraṁ ca niṣprabhamsphuracchāyēna satataṁ mukhēnādhaḥ prakurvatīm[PIR 2.*8; T 2.14]

    The second illustration of “distinction”; this time the cause (“with a sheen that shone nonstop”) is explicit.

Dried leaves, water, and wind
were her sole diet in this harsh ordeal,
and yet, she bore not a trace of pride
as your typical sadhu would.58 (39)

  • 58śīrṇaparṇāmbuvātāśakaṣṭē ’pi tapasi sthitāmsamudvahantīṁ nāpūrvaṁ garvam anyatapasvivat[PIR 2.*9; T 16]

    “Distinction” based on dissimilarity. Compare KKS 5.28.

The winter month, “Penance,”
is known for its austere allure.
She put its fame to shame
with an unabated penance.59 (40)

  • 59yā śaiśirī śrīs tapasā māsēnaikēna viśrutātapasā tāṁ sudīrghēṇa dūrād vidadhatīm adhaḥ[PIR 2.*10; T 2.16]

    “Distinction” based on “embrace” (ślēṣa), in this case the double meaning of the word tapas as a name of a month and the word for penance. Compare to KKS 5.28.

Her lean limbs all gold—
no Kashmir saffron rubbed.
Her sealed lips red—
no rouge applied.60 (41)

  • 60aṅgalēkhām akāśmīrasamālambhanapiñjarāmanalaktakatāmrābhām ōṣṭhamudrāṁ ca bibhratīm[PIR 2.*11; T 2.20]

    The ornament is “evocation” (vibhāvanā), where the result is manifest despite the absence of its normal cause.

Bright bud teeth,
tender hand twigs,
slender, standing in the wild,
her matted locks a swarm of bees.61 (42)

  • 61dantaprabhāsumanasaṁ pāṇipallavaśōbhinīmtanvīṁ vanagatāṁ līnajaṭāṣaṭcaraṇāvalim[PIR 2.*12; T 2.22]

    The ornament is “condensed speech” (samāsokti). The verse is meant to imply her resemblance to a forest creeper. According to T, the bees cling to the matted locks, rather than being identified with it.

Though grown gaunt, she no doubt
looked no less shapely,
her innate beauty flashing forth
from the radiance of her ordeal.62 (43)

  • 62tapastējaḥsphuritayā nijalāvaṇyasampadākr̥śām apy akr̥śām eva dr̥śyamānām asaṁśayam[PIR 2.*13; T 2.26]

    This is the first illustration of “intensification” (atiśayōkti), here in the type of identity given difference. With this verse ends a string of verses (38-43) that depict Umā as the object of Śiva’s gaze.

God then thought: “Amazing!
The splendor born of self-restraint
has turned her into another person,
no longer a girl.63 (44)

  • 63acintayac ca bhagavān ahō nu ramaṇīyatātapasāsyāḥ kr̥tānyatvaṁ kaumārād yēna lakṣyate[PIR 2.*14; T 2.27]

    The second type of “intensification,” defined as difference given identity.

If a streak of moonlight fell
right into an open day lotus,
this rosary of pearls in her palm
would find an equal.64 (45)

  • 64patēd yadi śaśidyōtacchaṭā padmē vikāśinimuktāphalākṣamālāyāḥ karē 'syāḥ syāt tadōpamā[PIR 2.*15; T 2.28]

    The third type of “intensification” in Udbhaṭa’s scheme involves a hypothetical, imagined entity.

I could swear her eyes shot sideways
only later. Yes, the God of Love
must have hit me first
with a barrage of arrows.65 (46)

  • 65manyē ca nipatanty asyāḥ kaṭākṣā dikṣu pr̥ṣṭhataḥprāyēṇāgre tu gacchanti smarabāṇaparamparāḥ[PIR 2.*16; T 2.29]

    Udbhaṭa understands the last type of “intensification” as a case where the effect (falling in love) precedes its cause (eye contact with the flirtatious love interest).

With her arms, gait, face,
she patently defeats
the stalk, gander, lotus—
outshining all lotus ponds.66 (47)

  • 66mr̥ṇālahaṁsapadmāni bāhucaṅkramaṇānanaiḥnirjayantyānayā vyaktaṁ nalinyaḥ sakalā jitāḥ[PIR 3.*1; T 3.3]

    The ornament is “respective enumeration” (yathāsaṅkhya). T reads tarjayantyā for nirjayantyā.

She whispers her prayers, eyes steadily drinking in
the rays of the sun.
A dark tan mark like I’m not sure this entirely works in English.You are right, but "like" has to be there, as Udbhaṭa insists... landed on her face
mistaking it for the moon.67 (48)

  • 67asyāḥ sadārkabimbasthadr̥ṣṭipītātapair japaiḥśyāmikāṅkēna patitaṁ mukhē candrabhramād iva[PIR 3.*2; T 3.6]

    This is “seeing as” (utprēkṣā) type 1. Udbhaṭa says, in his definition, that words such as “like” can express “seeing as.” Yet in both of his examples use only the word “like” (iva) itself. To indicate this, I chose to translate iva as “like” in this example. Note the closeness of this verse to KKS 5.21, also involving a “seeing as” of the śyāmikā (“dark tan mark”) on Umā’s face.

Oh dear, what has become
of her cheeks? Both have wasted away
as if from losing sight
of each other.68 (49)

  • 68kapōlaphalakāv asyāḥ kaṣṭaṁ bhūtvā tathāvidhauapaśyantāv ivānyōnyam īdr̥kṣāṁ kṣāmatāṁ gatau[PIR 3.*3; T 3.7]

    This is Udbhaṭa’s second illustration of “seeing as.”

One moment pulling away, another turning back,
halfway, prodding her with the tip of its horn—
a fawn with great affection
fills her with yearning.69 (50)

  • 69kṣaṇaṁ naṁṣṭvārdhavalitaḥ śr̥ṅgēṇāgrē kṣaṇaṁ nudanlōlīkarōti praṇayād imām ēṣa mr̥gārbhakaḥ[PIR 3.*4; T 3.9]

    This is Udbhaṭa’s example of “factual statement” (svabhāvōkti). I take the root naś/naṁś in its sense of running away. T’s explanation (ruditvā) makes little sense in this context. The illustration of the third group of ornaments ends here, but the depiction of the fawn carries over to the next verse, with which the fourth group begins.

With yearning not different
from love for a son,
Umā drew him to her chest,
immersing herself in soothing speech.”70 (51)

  • 70iyaṁ ca sutavāllabhyān nirviśeṣā spr̥hāvatīullāpayitum ārabdhā kr̥tvēmaṁ krōḍa ātmanaḥ[PIR 4.*1; T 4.3]

    The ornament is “endearing” (prēyasvat). Compare to KKS 5.15 (similar theme-wise, but not in the actual wording).

When Śiva thus conceived of Pārvatī,
virtue by virtue, His passion,
fed by no few visions,
grew strong.71 (52)

  • 71iti bhāvayatas tasya samastān pārvatīguṇānsambhr̥tānalpasaṅkalpaḥ kandarpaḥ prabalō ’bhavat[PIR 4.*2; T 4.6]

    This is the first of three verses illustrating the ornament “flavored” (rasavat). The rasa is, of course, the erotic. Here it is evoked by its proper term and that for its underlying stable emotion (sthāyibhāva), namely, “passion,” and by the mention of Pārvatī “thus conceived … virtue by virtue,” as the “foundational factor” (ālambanavibhāva).

Although His body was covered with sweat,
His hair was standing on end
resembling the mass of filaments
on the pistil of the kadamba bud.72 (53)

  • 72svidyatāpi sa gātrēṇa babhāra pulakōtkaramkadambakalikākōśakēsaraprakarōpamam[PIR 4.*3; T 4.7]

    In this second illustration of “flavored,” the emphasis is on Śiva’s sweat and hair standing on end as accompanying bodily states of the erotic rasa.

One moment pregnant with longing,
another, frozen with worry,
a third, languid with delight—His eyes
were the ornament of His face.73 (54)

  • 73kṣaṇam autsukyagarbhiṇyā cintāniścalayā kṣaṇamkṣaṇaṁ pramōdālasayā dr̥śā ’syāsyam abhūṣyata[PIR 4.*4; T 4.8]

    In this third and last illustration of “flavored,” the emphasis is on Śiva’s secondary emotions of longing, worry, and delight that help evoke the erotic rasa and on the acting registers of Śiva ’s eye movements.

The more His passion grew, the closer He drew
to grabbing the Daughter of the Mountain
by force, forgetting all about
the proper path.74 (55)

  • 74tathā kāmō ’sya vavr̥dhē yathā himagirēḥ sutāmsaṅgrahītuṁ pravavr̥tē haṭhēnāpāsya satpatham[PIR 4.*5; T 4.10]

    This is Udbhaṭa's ’s illustration of the ornament “prideful” (ūrjasvin), more on which in the following section.

He had the wives of Demon Gaja
wear their hair disheveled, cry,
bruise their breasts with their fists,
lose their bangles.75 (56)

  • 75yēna lambālakaḥ sāśraḥ karaghātāruṇastanaḥakāri bhagnavalayō gajāsuravadhūjanaḥ[PIR 4.*6; T 4.12]

    This is Udbhaṭa’s illustration for “roundabout speech” (paryāyōkta), where the effect (the lament of the demon’s wives) suggests its cause (his death at Śiva’s hands). Note that this verse alone illustrates the ornament in question, and the next one, which is syntactically part of the same sentence, has no illustrative purpose (I discuss this fact below).

Yet this god too is now tormented
by someone He Himself has burnt to ashes.
Offer homage to him, Bearer of the Fish Banner,
whose power cannot be restrained.76 (57)

  • 76sō ’pi yēna kr̥taḥ pluṣṭadēhēnāpy ēvam ākulaḥnamō ’stv avāryavīryāya tasmai makarakētavē[PIR 4.*7; T 4.13, where the reading is apāravīryāya and kusumadhavanē]

    The “Bearer of the Fish Banner” is Kāma, the god of love.

The passionate gaze, flirtatious eyes,
whirling brows, enchanted expression,
hair standing on end, beading sweat,
limbs burning with the fever of love—
The God of Mountains
laid it all to rest, approached
the Daughter of the Mountain, said
“Greetings.”77 (58–59)

  • 77atha kāntāṁ dr̥śaṁ dr̥ṣṭyā vibhramāṁś ca bhramaṁ bhruvōḥprasannaṁ mukharāgaṁ ca rōmāñcaṁ svēdasaṅkulamsmarajvarapradīptāni sarvāṅgāni samādadhatupāsarpad girisutāṁ giriśaḥ svastipūrvakam[PIR 4.*8-9; T 4.15-16]

    Like Pollock (2016: 347n150), I read dr̥ṣṭyā vibhramāṁś ca and rōmāñcaṁ with Tilaka, instead of dr̥ṣṭvā vibhramāc ca and rōmāñca- with PIR. This is Udbhaṭa’s illustration of “coterminous” (samāhita), which he understands, very uniquely, as the termination of rasa (“lay it all to rest”).

He said: “He has pearls,
born of wild boars, bamboos, elephants—
ornaments tribal beauties
would die for.78 (60)

  • 78uvāca ca yataḥ krōḍē vēṇukuñjarajanmabhiḥmuktāphalair alaṅkāraḥ śabarīṇām apīcchayā[PIR 4.*10; T 4.18]

    This is the first of five verses that illustrate the ornament “magnificent” (udātta): first based on riches (likely the four first verses) and then on deeds. Pollock, following the commentators, reads krōḍē[a-] in the sense of wild boars, one of the eight traditional sources of pearls along with bamboos and elephants (Pollock 2016: 347n153). The word might also mean “breast” or “lap,” either with reference to the women or the mountain.

He is home to Intoxicating Scents,
a ridge whose heads—
plush sapphires, rubies, lapis—
scrape the sky.79 (61)

  • 79puṣṭyēndranīlavaiḍūryapadmarāgamayair viyatśirōbhir ullikhad yatra śikharaṁ gandhamādanam[PIR 4.*11; T 4.19]

    “Magnificent riches” (udātta), second example. T reads yasya for yatra.

The ground near his northern slopes
is solid gold,
and a mighty mound of emerald
takes shelter at his foot.80 (62)

  • 80uttarōpatyakā yasya pradhānasvarṇabhūmayaḥmahān marakatōrvīdhraḥ pādōpāntaṁ ca saṁśritaḥ[PIR 4.*12; T 4.20]

    “Magnificent riches” (udātta), third example.

When Earth dived down on doomsday,
he didn’t fall along. Hell no:
his true dimensions
were fully exposed.81 (63)

  • 81babhūva yasya pātālapātinyāṁ saṅkṣayē kṣitaupatanaṁ na tayā sārdham āyāmas tu prakaṭy abhūt[PIR 4.*13; T 4.21]

    The commentators classify this verse as the fourth example of “magnificence” based on riches, although like the following one, it could also be based on deeds.

And when Viṣṇu, the Primordial Boar, battered him
broad-shouldered, blow after blow,
he stood still. He is Himālaya.
You must be his daughter.82 (64)

  • 82tasyādikrōḍapīnāṁsanigharṣē ’pi punaḥ punaḥniṣkampasya sthitavatō himādrēr bhavatī sutā[PIR 4.*14; T 4.22]

    “Magnificent deeds” (udātta).

Moreover, you yourself glow
with the reddish hue of the morning sun:
with light fingers that shine like shoots
you grant the fruit of awakening.83 (65)

  • 83svayaṁ ca pallavātāmrabhāsvatkaravirājinīprabhātasandhyēvāsvāpaphalalubdhēhitapradā[PIR 4.*15; T 4.25]

    This is the first of three verses illustrating bitextual “embrace” (ślēṣa). In Sanskrit, the “embrace” in the second half works as follows: Umā grants the wish of those who desire the fruit that is impossible to obtain, i.e., liberation (a-su-āpa-phala-lubdha-īhita-pradā); but segmented differently, the sun is that which gives good advice, or a boon to a person who does not desire the fruit of sleep (a-svāpa-phala-lubdhē hitapradā). I tried to very partially allude to this with “the fruit of awakening.” In the first half, kara may mean either hand or ray, and bhāsvat, either shining or the sun. I tried to replicate this pun with “light fingers.”

Your face is friend to the moon, your hair
sapphire, your glamour perfect nay-care, [read: nacre]
and your toes, ruby. You’re the gem
of the triple world.84 (66)

  • 84indukāntamukhī snigdhamahānīlaśirōruhāmuktāśrīs trijagadratnaṁ padmarāgāṅghripallavā[PIR 4.*16; T 4.26]

    “Embrace” (ślēṣa). Indukānta in Sanksrit means “friend to the moon” (in the sense of resembling it) and, hence, also the moonstone. Muktā-śrī may refer to the shine of a pearl, but read as muktā-aśrī, to something devoid of anything not shining or beautiful (I tried to replicate this with nacre/nay-care). The rest of the puns depend on whether the precious stones are taken to modify Umā’s body parts or to be identified with them.

Though no one comes to pick a quarrel, [read: coral]
you're the tree of heaven on earth.
Your looks are rupture [read: rapture] to one’s eyes,
yet your beauty swells and constantly flows.85 (67)

  • 85apārijātavārtāpi nandanaśrīrbhuvisthitāabindusundarī nityaṁ galallāvaṇyabindukā[PIR 4.*17; T 4.27]

    Third illustration of “embrace” (ślēṣa). Here is how the original works: apārijātavārtā may mean either that the heaven is bereft of the Pārijāta (a coral wish-granting tree that, according to legend, grows in heaven), or someone who has no enemies left. I tried to replicate this with coral/quarrel. A-bindu-sundarī means containing not even a drop of beauty, or, if we read ab-indu-sundarī: beautify like the moon reflected in water. Nandana-śrī may mean either the riches of Indra’s heaven or beauty that delights. At first blush, the verse creates a contradiction: the riches of heaven are without its wish-granting tree, are found on earth, and with not even a drop of beauty. To resolve this antithesis, a second reading is supplied, referring to Umā, who is indeed on earth, delights with her beauty, and has no enemies. I tried to replicate this antithesis in the translation, though, of course, with limited success.

This surely is not asceticism.
In truth, it is the deadliest poison,
especially for ladies such as yourself,
tender like a sliver of the moon.86 (68)

  • 86ētad dhi na tapaḥ satyam idaṁ hālāhalaṁ viṣamviśēṣataḥ śaśikalākōmalānāṁ bhavādr̥śām[PIR 5.*1; T 5.4]

    The ornament is “denial” (apahnuti).

Birth in a prosperous home,
dazzling beauty and enchanting youth,
yet no happiness.
Who would fail to be amazed?87 (69)

  • 87maharddhini gr̥hē janma rūpaṁ smarasuhr̥dvayaḥtathāpi na sukhaprāptiḥ kasya citrīyatē na dhīḥ[PIR 5.*2; T 5.7]

    This is the first instance of “exceptionality” (viśēṣōkti), where the effect does not materialize despite the presence of its causes. Here the cause for this failure is not mentioned.

Seeing you behave
so recklessly, my tongue,
though typically quick,
fell silent.88 (70)

  • 88itthaṁ visaṁṣṭhulaṁ dr̥ṣṭvā tāvakīnaṁ vicēṣṭitamnōdēti kimapi praṣṭuṁ satvarasyāpi mē vacaḥ[PIR 5.*3; T. 5.8]

    “Exceptionality” (viśēṣōkti). Tilaka reads “my mind” (manas) instead of “speech” (vacas), translated as “tongue.” Here the explanation is displayed (“seeing you behave so recklessly”).

Still, what can I do? My amazement
has me speak: the shape of your body
and the severity of your penance
are worlds apart.89 (71)

  • 89yad vā māṁ kiṁ karōmy ēṣa vācālayati vismayaḥbhavatyāḥ kvāyam ākāraḥ kvēdaṁ tapasi pāṭavam[PIR 5.*4; T 5.10]

    The ornament, based on the incongruity between her delicate body and her severe penance, is “antithesis” (virodha); later theorists would label such instances viṣama.

Can anyone observe your fragile limbs
without realizing at once
how sturdy are jasmine sprigs,
moon slivers, and banana shoots?90 (72)

  • 90tvadaṅgamārdavaṁ draṣṭuḥ kasya cittē na bhāsatēmālatīśaśabhr̥llēkhākadalīnāṁ kaṭhōratā[PIR 5.*5; T 5.12. T reads dr̥ṣṭvā for draṣṭuḥ]

    This is “the yoke of equivalence” (tulyayogitā), and here, unlike in the following example, the standards (jasmine sprigs etc.) are not pertinent to the context.

Yoga gear, matted locks,
tree-skin, deer-hide—
can you please explain
how these suit your limbs?91 (73)

  • 91yōgapaṭṭō jaṭājālaṁ tāravī tvaṅmr̥gājinamucitāni tavāṅgasya yady amūni tad ucyatām[PIR 5.*6; T 5.13]

    In this second instance of “the yoke of equivalence,” the standards are pertinent to the situation.

They decay then and there
with no one to enjoy them,
lush fruits, flowers, and all—
the riches of forests beyond man’s reach.92 (74)

  • 92yānti svadēhēṣu jarām asamprāptōpabhōktr̥kāḥphalapuṣparddhibhājō ’pi durgadēśavanaśriyaḥ[PIR 5.*7; T 5.15]

    This is “praise of the irrelevant” (aprastutapraśaṁsā): the speaker depicts the fate of fruit unpicked in the forest, while the real intention is Pārvatī herself.

Damn this beauty of yours
that has no equal!
For nowhere in the triple world
will you find a match.93 (75)

  • 93dhig ananyōpamām ētāṁ tāvakīṁ rūpasampadamtrailōkyē ’py anurūpō yad varas tava na labhyatē[PIR 5.*8; T 5.17]

    This is “praise in disguise” (vyājastuti), where the literal damning of Umā’s beauty ends up being a compliment (albeit a somewhat left-handed one).

A lady without a suitable husband,
beautiful though she may be,
forever bears the gloom
of a moonless night.94 (76)

  • 94vinōcitēna patyā ca rūpavaty api kāminīvidhuvandhyavibhāvaryāḥ prabibharti viśōbhatām[PIR 5.*9; T 5.19]

    The ornament is “illustration” (nidarśanā).

You want the moon for a husband.
But suppose you can’t get that.
Somewhere out there, I’m sure, there’s a husband
you can learn to respect.95 (77)

  • 95yady apy atyantam ucitō varēndus tē na labhyatētathāpi vacmi kutrāpi kriyatām ādarō varē[PIR 5.*10; T 5.21]

    T reads ādarō ’parē. This is Udbhaṭa’s first type of “fusion” (saṅkara), based on a “doubt” (sandēha) as to which ornament out of several captures the reader’s attention.

If you stand like this to pursue a spouse,
this useless pursuit makes no sense.
Your beauty is such that every young man
is a humble slave at your feet.96 (78)

  • 96itthaṁ sthitir varārthā cēn mā kr̥thā vyartham arthitāmrūpēṇa tē yuvā sarvaḥ pādabaddhō hi kiṅkaraḥ[PIR 5.*11; T 5.23]

    “Fusion” of both sound and sense ornaments.

Enough. Stay put. You delight
like a lotus in a painting,
with bright colors and beautiful ears,
just by being seen.97 (79)

  • 97maivam ēvāstha sacchāyavarṇikācārukarṇikāambhōjinīva citrasthā dr̥ṣṭimātrasukhapradā[PIR 5.*12; T 5.24]

    “Fusion,” of two ornaments, “simile” and “embrace,” in one portion of the sentence. The “ears” of the lotus are its central seed pods, and Umā has a pair of beautiful ears. The first part could also mean “Don't just sit there!” But I translated it the way I did based on the notion that she delights “like a lotus in a painting,” that is, by simply being idle.

This hunter, Love,
though you, like Śiva, had him disembodied,
never lets go of your body
out of audacity, I presume.98 (80)

  • 98harēṇēva smaravyādhas tvayānaṅgīkr̥tō ’pi santvadvapuḥ kṣaṇam apy ēṣa dhārṣṭyād iva na muñcati[PIR 5.*13; T 5.26]

    “Fusion,” mutual dependency between ornaments. It is hard to recreate the pun on anaṅgīkr̥taḥ (not embraced, rendered bodiless). Recall that the speaker, Śiva, is disguised. He refers to himself as if he were a different person.

He whose discus is equal to his hand,
and whose hand to the discus
in swiftly plucking, flower-like,
the heads of his enemies in battle, 99 (81)

  • 99śirāṁsi paṅkajānīva vēgōtpātayatō dviṣāmājau karōpamaṁ cakraṁ yasya cakrōpamaḥ karaḥ[PIR 5.*14; T 5.28]

    T reads vēgāt pātayatō. This is “comparison with the standard of comparison” (upameyopamā), where the subject and the standard are compared with each other in succession. The speaker, Śiva in disguise, has now begun a long depiction of Viṣṇu in a series or relative clauses. Viṣṇu’s identity as the subject of this praise will only be revealed in verse 90 below.

and whose servant, Mr. Discus,
fulfilled the wishes of the gods,
together with Death, in the battle
that bode ill for Demon Tāraka,100 (82)

  • 100dyujanō mr̥tyunā sārdhaṁ yasyājau tārakāmayēcakrē cakrābhidhānēna praiṣyēṇāptamanōrathaḥ[PIR 5.*15; T 5.30]

    T reads prēṣya-. The ornament is “concurrence” (sahōkti), in this case, of two actions in one.

and who gave his bare chest
to the enemies of the gods, and gained,
by slaying Hiraṇyākṣa and the others,
fame and glory in battle,101 (83)

  • 101urō datvāmarārīṇāṁ yēna yuddhēṣv agr̥hyatahiraṇyākṣavadhādyēṣu yaśaḥ sākaṁ jayaśriyā[PIR 5.*16; T 5.32]

    The ornament is “reciprocity” (parivr̥tti), where some action is figuratively portrayed as an act of give and take (“gave his bare chest … gained fame and glory”). For PIR, this is a barter of equal elements (chest and fame, he says, are equals).

and who filled the milky ocean with precious stones,
that fell from the crest of Mount Mandara,
which Snake Vāsuki spun forcefully,
to get the Kaustubha gem in return,102 (84)

  • 102nētrōragavalabhrāmyanmandarādriśiraścyutaiḥratnair āpūrya dugdhābdhiṁ yaḥ samādatta kaustubham[PIR 5.*17; T 5.33]

    The second illustration of “reciprocity” where the thing sacrificed (precious stones) is inferior to that won (the Kaustubha gem).

and who, when Bali reigned to the ends of Earth
and was set to gain the sky through ritual,
gave safety to the residents of heaven
by taking on smallness,103 (85)

  • 103yō balau vyāptabhūsīmni makhēna dyāṁ jigīṣatiabhayaṁ svargasadmabhyō datvā jagrāha kharvatām[PIR 5.*18; T 5.34]

    In this third illustration of “reciprocity,” the thing given (safety) is far superior to that gained (smallness, by becoming Vāmana, the dwarf).

‘There, in His hand: Could that be a mass of white fame
born from tearing every demon’s heart?
But how has it turned
into a solid lump?104 (86)

  • 104hastē kim asya niḥśēṣadaityahr̥ddalanōdbhavaḥyaśaḥsañcaya ēṣa syāt piṇḍībhāvō ’sya kiṁ kr̥taḥ[PIR 6.* 1; T 6.3]

    This is the first illustration of the ornament “in doubt” (sasandēha).

Is it a goose, then, drawn to the lotus
growing from His navel? But it moves not!’
That’s how the innocent take in His conch
with confusion.105 (87)

  • 105nābhipadmaspr̥hāyātaḥ kiṁ haṁso naiṣa cañcalaḥiti yasyābhitaḥ śaṅkham aśaṅkiṣṭārjavō janaḥ[PIR 6.*2; T 6.4]

    “In doubt” continued.

‘A black cloud atop Mount Meru?
Smoke from the fire of doomsday?’
This Dark One on the radiant King of Birds
has one confused.106 (88)

  • 106nīlābdaḥ kim ayaṁ mērau dhūmō ’tha pralayānalēiti yaḥ śaṅkyatē śyāmaḥ pakṣīndrē ’rkatviṣi sthitaḥ[PIR 6.*3; T 6.6]

    A second case of “in doubt” that leads to the suggestion of another ornament, in this case, a pair of similes.

His speech is like His speech,
His spotless deeds like His very own deeds,
and His beauty entices the eyes of the world
just like His beauty.107 (89)

  • 107yasya vāṇī svavāṇīva svakriyēva kriyāmalārūpaṁ svam iva rūpaṁ ca lōkalōcanalōbhanam[PIR 6.*4; T 6]

    “Inimitability” (ananvaya).

Yes, I’ve been talking about Viṣṇu.
Even He, as the moon to moonlight at dawn,
would any day ditch his very own Lakṣmī,
that endless rain of immortality, for you.108 (90)

  • 108tvatkr̥tē sō ’pi vaikuṇṭhaḥ śaśīvōṣasi candrikāmapy adhārāṁ sudhāvr̥ṣṭiṁ manyē tyajati tāṁ śriyam[PIR 6.*5; T 6.10]

    “Mixture” (saṁsr̥ṣṭi) of independent ornaments, in this case of “simile,” “identification,” and host of other figures.

So enough, lotus-eyed girl,
find some lucky boy,
then go home and enjoy
your youth with him.109 (91)

  • 109tad uttiṣṭhātidhanyēna kēnāpi kamalēkṣaṇēvarēṇa saha tāruṇyaṁ nirviśantī gr̥hē vasa[PIR 6.*6; T 6.11]

    Another illustration of “mixture.”

You give pain, pleasure:
eyes untouched by kohl,
yet the sheen of many ornaments
vividly visible on your every limb.110 (92)

  • 110karōṣi pīḍāṁ prītiṁ ca nirañjanavilōcanāmūrtyānayā samudvīkṣya nānābharaṇaśōbhayā[PIR 6.*7; T 6.13]

    This is Udbhaṭa’s illustration of “integrity” (bhāvika), where objects of the past or future (in this case Umā’s jewels and ornaments) are visible as if they were there, in front of one’s eyes.

The tan, all around a bit darker,
betrays the spots
where ornaments were borne
and breaks my heart.111 (93)

  • 111chāyēyaṁ tava śēṣāṅgakāntēḥ kiñcid anujjvalāvibhūṣāghaṭanādēśān darśayantī dunōti mām[PIR 6.*8; T 6.15, reads vibhūṣāghaṭanōddēśān]

    This is an illustration of the “inferential sign of poetry” (kāvyahētu), which is here the scope for poetic inference by the speaker. Compare to KKS 5.48.

Why keep on heaping up words?
Go get yourself a husband.
Do great rivers stay put
before finding the ocean?”112 (94)

  • 112kiṁ cātra bahunōktēna vraja bhartāram āpnuhiudanvantam anāsādya mahānadyaḥ kim āsatē[PIR 6.*9; T 6.17, reads vātra]

    The ornament here is the “analogy of poetry” (kāvyadr̥ṣṭānta), which again applies a tool from logic to poetry, in this case the analogy. Compare to KKS 5.49.

Analysis

On the face of it, Udbhaṭa’s Kumārasambhava, at least as we have it, is just a slightly adapted version of Kālidāsa’s famous namesake. The characters are the same, so is the basic narrative, and the choice of language and imagery ring unmistakably familiar. In several cases, Udbhaṭa even employs the same ornaments as in the parallel passage from the source poem.113 But the impression that Udbhaṭa’s is a faithful adaptation that merely retools Kālidāsa’s classic for the pedagogical purposes of introducing the ornaments is false, and intentionally so.

  • 113One example are the verses based on "distinction" (vyatireka). Compare UKS 37–40 to KKS 5.27–29.

Consider, first, the dramatic shift in focus. As Gary Tubb has shown, in Kālidāsa’s work “Pārvatī is the principal protagonist”: her union with Śiva is “a process of spiritual maturing brought about … through her own efforts.” Thus, “the whole narrative focus of the poem is on the events of her life.”114 Indeed, in Kālidāsa’s rendering, Śiva is “out of the frame” from the moment he burns Kāma to the moment he wanders, disguised as an ascetic, into Pārvatī’s penance grove; the narrative focus is entirely on Pārvatī (and her parents). In Udbhaṭa’s poem, by contrast, Śiva is the unquestionable hero. True, if what we have is just a portion of a larger work, it is dangerous to generalize from it to whatever was lost. But it is significant that it is precisely in the section where, in Kālidāsa’s work, Umā is in the spotlight (through her performance of tapas and by withstanding God’s test), Udbhaṭa’s Śiva is the sole subject. Of the 94 verses, 28 depict Śiva, 13 are his words to himself, 18 are dedicated to the autumn as he experiences it, and the remaining 34 consist of his speech to Umā while he is disguised. Umā does not utter a single word in these verses, and she is described only as seen by him.115

  • 114Tubb 1984: 229, 231, 225.

  • 115Thus, the key scene of Umā’s decision to perform tapas after taking permission from her parents and the loving description of her tapas by the narrator (KKS 5.1-29) have no parallel in Udbhaṭa’s poem as we have it.

This is not just a matter of who is at the center of the narrative. For Kālidāsa, Umā is the emotional focus of the poem and the locus of the primary rasas of heroism (in hertapas) and love. She, perhaps, represents us readers: as a subject enduring immense difficulty before uniting with her beloved, and as a human obtaining union with the divine. Kālidāsa’s Śiva, by contrast, is more aloof, and for the most part, calm, restrained, meditative, and withdrawn: the locus of the rasa of peace (śāntarasa).116 Even when he momentarily loses his complete calm, we have very little access to his feelings and thoughts.117 Likewise, in the last and eighth chapter, depicting the newlyweds' honeymoon, it is Umā’s emotions that primarily interest the poet.

  • 116As Tubb notes (1984: 229): “The commentators on the Kumārasambhava come close to recognizing this status of Pārvatī’s when they connect the description of her austerities with the heroic mood (vīra-rasa) while associating Śiva’s own austerities with the mood of peace (śānta-rasa).”

  • 117KKS 3.69 tells of the unrest of his senses (indriyakṣōbha) while meditating and of his success in restraining of them; 3.70–71, of his anger (manyu, krōdha) in response to Kāma’s attack; and 3.67 that his steadiness was slightly disturbed (kiñcitpariluptadhairya) when facing Umā.

The opposite is true of Udbhaṭa’s version of the same events. To begin with, here Śiva is the unmistakable nāyaka of Sanskrit love poetry. He falls in love with Umā even before he comes to know about her tapas, feels remorse for burning Kāma (there is nothing of the sort in Kālidāsa’s version), and like all lovers, is totally uncertain as to whether Umā feels as he does (“Is she, too, aching boundlessly / me-like, another victim / of the sudden savagery / of Love?” UKS 22). More important, if for Kālidāsa Śiva is the paragon of self-restraint, Udbhaṭa’s Śiva is the poster child of emotional excess. Udbhaṭa makes sure that readers familiar with Kālidāsa’s classic—a group that is presumably synonymous with all educated readers—will notice this twist by his redeployment of imagery and language. For instance, whereas Kālidāsa’s Śiva is said to have seen “the highest light, also known as the highest self” (paramātmasañjñaṁ dr̥ṣṭvā param jyōtir, KKS 3.58),118 for Udbhaṭa’s Śiva, "the mighty light of perfect knowledge / began to firefly" (khadyōtāyitum ārabdhaṁ tattvajñānamahāmahaḥ, UKS 26).

The most visible example of Śiva as an emotional wreck is in the section dedicated to the rasa ornaments. Udbhaṭa dramatically retheorized these devices: of the original categories only the name remained, and they now represented the evolution of emotional flavors as understood in dramaturgy (or at least in Udbhaṭa’s version thereof), from basic emotions (prēyasvat, his name for prēyas), to fully evolved rasas (rasavat), then to rasas whose production was hampered by a socially inappropriate excess of emotions (ūrjasvin), and finally, the cessation of emotion (samāhita).119 That Śiva is the subject chosen to exemplify most of this entire arc is highly consequential (note that for the first ornament, “endearing” or prēyasvat, Udbhaṭa depicts Umā’s fondness for a fawn, but for the more fervent feelings, he turns to Śiva). His passion towards Umā is intense to begin with, which makes it a rasa, and most significantly, it is Śiva who illustrates the unacceptable, violent emotional excess that violates rasa, unthinkable of Kālidāsa’s Śiva: “The more His passion grew, the closer He drew / to grabbing the Daughter of the Mountain / by force, forgetting all about / the proper path” (UKS 55). Thus, it is He who is forced to abort the process of rasa altogether in the end (samāhita, UKS 56).

The focus on Śiva’s succumbing to his emotions is visible not just in Udbhaṭa’s illustration of the rasa ornaments. Consider the ornament “roundabout speech” (paryāyōkta), which following Bhāmaha’s order, is stuck in the middle of a tight group of affectual figures. Udbhaṭa dedicates two verses to illustrating this ornament rather than one. This in itself is not unusual in Udbhaṭa’s poem; there are several ornaments that get a more elaborate illustration, even in the absence of explicit mention of subtypes. What is highly unusual, however, indeed unique in the entire poem, is that only the first of the two verses illustrates the ornament in question, and the second is pedagogically redundant:120

  • 120Verse 23 also does not illustrate an ornament, but it supplies the seeing subject (Śiva), and hence sets the stage for the set of similes he sees in Umā in verse 24.

He had the wives of Demon Gaja
wear their hair disheveled, cry,
bruise their breasts with their fists,
lose their bangles. (56)
Yet this god, too, is now tormented
by someone He Himself has burnt to ashes.
Offer homage to him, Bearer of the Fish Banner,
whose power cannot be restrained. (57)

The first of the two verses (56) depicts Śiva’s triumph over Gaja in a roundabout way: instead of stating that he killed this demon, the poet dwells on the effect this act had on Gaja’s newly widowed wives. This is par for the course for what becomes, after Udbhaṭa, the classical understanding of this ornament. But the verse is not a complete sentence: it is a relative clause, and the expectation is that the next verse will reveal and extol its yet-unmentioned subject. Instead, verse 57 names this triumphant god only to subject him (in yet another relative construction) to a superior power, the Bearer of the Fish Banner (Love). As noted, there is no ornament in this verse, and it thus has no purpose as an illustration. Rather, Udbhaṭa literally goes out of his programmatic way to establish the fact that Kāma is superior to Śiva. The pair of verses, after all, culminate in the demand to offer homage not to Śiva, the hero of the relative clause that illustrates “roundabout speech” (paryāyōkta), but to Kāma, who defeats him in the main and technically unnecessary sentence.

This is not the only passage in the poem that is meant to establish Śiva’s complete surrender to Love. Consider the following variety of “dismissal” (ākṣēpa), here likely in Śiva’s own voice (albeit in the third person):121

  • 121The following verse identifies this as Śiva’s thought.

O the power of Love! Even Rudra
is in such a state… But enough of that!
Can one take the measure of the ocean
by a bucket? (31)

The first half of the verse posits as striking the fact that Śiva is under Love’s power. But the second half, as standard in this ornament, dismisses the first. There is nothing at all striking or unusual in this half-finished thought (its being incomplete makes this a dismissal of the unsaid). The second half of the verse corroborates this dismissal by resorting to an implied comparison of Love’s power to that of the infinite ocean. In Kālidāsa’s poem, Śiva’s mind is compared to an ocean whose steadiness is only slightly diminished with the moonrise of Umā’s face (haras tu kiñcitpariluptadhairyaś candrōdayārambha ivāmburāśiḥ; KKS 3.67). The choice of imagery must be meaningful: not only is it Love who is cast as the ocean, rather than Śiva, but the latter, by implication, is a mere bucket-full in comparison.122

  • 122I am grateful to Sheldon Pollock for a conversation on this verse.

Once we train our eyes to see it, the belittling of Śiva is all over Udbhaṭa’s poem. The god who in Kālidāsa’s work completely blocked his breath and controlled his senses (antaścarāṇāṁ marutāṁ nirōdhāt, jitēndriyē śūlini: KKS 3.48, 3.57), now has his body “covered with sweat, / His hair was standing on end” (UKS 43). He is “vagabonding” (UKS 25), his mind and limbs are scorched by Love, and he behaves just like any ordinary guy (UKS 27–28). He is, moreover, totally unaware of Umā’s fate and whereabouts (KKS 29) and is forced to turn to meditation to find these out, because even His eye “may err” (but not the inner eye; KKS 35). Finally, note the pair of verses that repeat the image of Śiva turning into a schoolboy (KKS 33–34), even if for the purpose of disguising himself.123 Our poet is clearly invested in portraying Śiva as small and powerless, certainly against a far-superior Kāma.

  • 123In Kālidāsa’s poem, the disguised Brahmin, while a youth, is depicted far more impressively (KKS 5.30).

Finally, consider the concluding long section in which, as in Kālidāsa’s poem, Śiva approaches Umā in disguise. ’s There the stranger who enters the ascetic grove, a glowing young Brahmin of eloquent speech (pragalbhavāg jvalann iva brahmamayēna tējasā; KKS 5.30), begins with what we may describe as “ascetic shoptalk” (KKS 5.33–35), flattering words for Umā (KKS 5.36–39), and questions about the purpose of her penance, which he depicts as incommensurate with her delicate beautiful body (KKS 5.40–50). Then he suggests that her goal is winning a husband and asks who he may be. In response, he learns from her friend that Umā seeks Śiva, something she briefly corroborates (KKS 5.52–64). At this point, the young Brahmin lambasts her plans: Śiva has snakes swirling on the hand she wishes to hold, he wears a grizzly elephant skin, his matted hair is scattered with ashes, as is his chest, his mount is a decrepit old bull, the moon on his crest is a mere sliver, a third eye distorts his face, and his ancestry is unknown; thus, she is advised to forget all about him, who is like an impaling stake in the cremation ground, and seek a more traditional husband, likened to a Vedic sacrificial post (KKS 5.65–73). When Umā withstands this verbal test and counters his criticism in no uncertain terms, the speaker finally reveals himself to her as Śiva and promises to be her slave (KKS 5.75–86).

The parallel passage in Udbhaṭa’s poem is markedly different. Most significantly, the disguised god, an ascetic boy, avoids any mention of Śiva, thus allotting no airtime to the sort of self-criticism that, in Kālidāsa’s work, is nothing but praise with a smile. Instead, after expressing admiration for Himālaya, Umā’s father, he launches into a eulogy of Viṣṇu (UKS 81–90).124 This laudatory speech ends with a surprising twist:

  • 124Note that in Kālidāsa’s intertext Himālaya is compared with Viṣṇu (KKS 6.67).

Yes, I’ve been talking about Viṣṇu.
Even He, as the moon to moonlight at dawn,
would any day ditch his very own Lakṣmī,
that endless rain of immortality, for
you.

On the one hand, this is a compliment to Umā: she is more attractive than even the radiant Lakṣmī. On the other, it is a dig at the subject of the apparent extolment, Viṣṇu, who turns out to be a fool who would give up his most precious treasure (“that endless rain of immortality”) and a serial betrayer (“as the moon to moonlight at dawn”). Moreover, the compliment to Umā, already inherently left-handed (“Damn this beauty of yours / that has no equal! / For nowhere in the triple world / will you find a match”125), culminates in a cruel recommendation that she move on: “Go get yourself a husband” (94).

  • 125KKS 75. Narayana Daso Banhatti, comments on this verse with disapproval: “This is not a very good example of vyājastuti” he says, for Umā’s beauty “is really censurable if it hinders her union with a fit husband … and this kind of meaning indicating the reality of nindā [blame] lingers in our mind when we read the verse” (KKS p. 129 of the annotation). For more left-handed compliments of Udbhaṭa’s Śiva consider, for example, verses 69 (“Birth in a prosperous home, / dazzling beauty and enchanting youth, yet no happiness”) and 74 (“They decay then and there…”). For cases of outright criticism, see for instance verse 70 (“Seeing you behave/ so recklessly…”).

Thus, the poem consistently belittles its main hero, Śiva, includes a barb against the other main god, Viṣṇu, and concludes on the hero’s sardonic and poignant address to his beloved. Furthermore, it ends before Śiva drops His guise and announces Himself her slave. Of course, we know this must be coming. But Udbhaṭa consciously ended his illustration of ornaments—the poem at least as it is included in his Compendium—on this sour note. This cannot have been an accident. But what does it all mean?

Concluding Thoughts

One explanation is that this has to do with the possible identity of Udbhaṭa as a Cārvāka materialist. According to this line of thought, he may have composed his Kumārasambhava to counter the theology of his Śaiva colleagues and promote his own doctrine. After all, a materialist will score a doctrinal point by making the gods appear far less transcendent and subject to the same this-worldly powers that govern ordinary human beings. This might explain why in Udbhaṭa’s version, there is a constant belittling of Śiva and Viṣṇu. It might also account for the fact that in his poem, it is not Umā’s penance, of which Śiva is initially unaware, that attracts God to her, but a far more mundane and total succumbing to love. Recall that one of the recovered sūtras in Br̥haspati’s foundationalLōkāyata text is “Desire (kāma, i.e., the fulfilment of desire) is the only aim of life.”126

  • 126Franco 2018: 635. As Franco notes, there is a variant quotation: “desire and wealth [artha] are the only aims of life.”

This is a theory certainly worth considering. But I think there exists another explanation. I see in Udbhaṭa’s verses less a calculated undermining of Śaiva theology than a poetic subversion of Kālidāsa’s classic, though perhaps with the added benefit of provoking its religious tenets. By the ninth century, Kālidāsa’s Kumārasambhava already enjoyed a canonical status likely unparalleled by that of any other Sanskrit literary work. Consider, in this context, Ānandavardhana Light on Dhvani (Dhvanyālōka), written in Kashmir only a generation or two after Udbhaṭa: in the final chapter of the work, where Ānandavardhana discusses innovation (the ability of poetry to make something infinitely new), he demonstrates this with three verses dedicated to the same topic (Umā’s beauty), all from one work, Kālidāsa’s Kumārasambhava; no other poem is given such an honor in his treatise.127 To compose an eponymous poem retelling the same narrative as Kālidāsa’s required considerable chutzpah. To do so by lifting language and images from the parent poem only to present a far less flattering version of the falling in love of the couple who are, “as the opening verse of the Raghuvaṁśa reminds us, the parents of us all,” is poetic patricide on steroids.128

  • 127Dhvanyāloka p. 539. By the same token, Kālidāsa’s Kumārasambhava comes under attack for its depiction of Śiva and Umā’s honeymoon in chapter eight, and although Ānandavardhana notes that this flaw of impropriety is concealed by Kālidāsa’s skill, it is clear that this was perceived as a problem perhaps already before Udbhaṭa’s time. It thus might be that Udbhaṭa found the idea of outdoing Kālidāsa in this domain appealing. I am grateful to Lawrence McCrea for suggesting this point to me.

  • 128The quote is from Tubb 2014: 73.

Note, moreover, that the portrayal of Śiva as succumbing to Kāma is not unheard of in Purāṇic sources. “In several texts, Śiva is said to faint with lust, to be full of desire, or to be tortured by Kāma […] Śiva himself muses upon the phenomenon of his excitement: ‘How can I lust to make love to Pārvatī when she has not performed a vow of tapas? And how is it that I wish to rape her? How can I have been excited by desire when I do not wish it now? For some reason I seem to be attracted to this young girl and to wish to unite with her.’”129 The Kālikāpurāṇa likely postdates both Kālidāsa and Udbhaṭa, but I think it is safe to say that versions of the story in which Śiva is overcome by Kāma’s influence and arrows were known to both authors, and these were not written by Cārvākas.130 There are likewise versions in which Śiva is self-controlled or simply pretends to be excited.131 Regardless of one’s understanding of Śiva—as a god who oscillates between complete asceticism and wild erotic urges, as presented by Wendy Doniger (O’Flaherty 1973), or as a Yogi whose tapas or meditative “internalization is always erotic,” as seen by Don Handelman and David Shulman (1997)—it is clear that contemporary texts offered a spectrum of characterizations of Śiva as either in control or being controlled.132 In his Kumārasambhava, Kālidāsa chose to portray Śiva as calm, self-possessed, and entirely superior to Kāma. It is not his burning of the love god that proves this (indeed, it seems to prove the opposite, as others have noted).133 Rather, as Gary Tubb has shown, the overall structure of the poem in two parallel halves, one featuring Kāma’s attempt to get the couple together as a failure, and the other presenting Umā’s successful winning of Śiva’s hand through her tapas, drives this point home.134 But in his telling, Udbhaṭa deliberately belittles Śiva, and goes out of his way to portray him as inferior to Kāma, all while presenting his poem as anotherKumārasambhava.

Rather than reduce this move to Udbhaṭa’s materialist identity, then, I wish to suggest that it fits well with his overall intellectual profile as a grammarian, philosopher, and literary theorist. In his former hat, he has been portrayed as “someone who felt almost completely free from the traditional interpreters of Pāṇini’s grammar […] He split rules where this suited him, and gave forced interpretations where this helped him to obtain the results that he wanted. In a way he behaved in the same way as Patañjali had behaved many centuries earlier, but he did so at a time when many other grammarians had opted to recognize Patañjali as an authority.” Indeed, if for Jayanta Bhaṭṭa, “Pāṇini’s grammar contains the words of God himself […] if only interpreted in accordance with Patañjali’s and Bhartṛhari’s comments,” Udbhaṭa “did not abide by these rules.”135

Likewise, also “within the Cārvāka movement, [Udbhaṭa] was a bit of a rogue (dhūrta)” as portrayed by Jayanta Bhaṭṭa’s commentator Cakradhara.136 For example, he turned on its head the common interpretation ofbhūtēbhyaś caitanyam: “Earlier Cārvākas had interpreted this to mean ‘Consciousness out of the elements,’ taking the word bhūtebhyaḥ to be an ablative.” But “Udbhaṭa preferred to read the sūtra from the foundational Cārvāka text as containing a dative, ‘Consciousness for the elements,’ which profoundly changed a fundamental tenet of the system.”137

Something similar happens in Udbhaṭa’s output as a literary theorist. Consider the Compendium, the treatise which his Kumārasambhava serves to illustrate. Here Udbhaṭa presents pretty much the same set of ornaments as Bhāhama, defines them by redeploying the language of his predecessor, and even reuses Bhāmaha’s title as part of his (note that the name Kāvyālaṅkārasārasaṅgraha could read as the Essential Compendium of [Bhāmaha's] Ornament of Literature). Yet Udbhaṭa departs from Bhāmaha’s views time and again, especially in key points such as the work’s beginning and end, and in fact uses his predecessor’s legacy to present a radically new theory of figuration. Thus, Udbhaṭa’s use of Bhāmaha is opportunistic, and he frequently proves his master wrong: new ornaments are coined while many of Bhāmaha’s are dramatically changed or unceremoniously cancelled; categories that Bhāmaha explicitly and vehemently rejected get promptly reinstated; and there are also occasional nods to Bhāmaha’s nemesis, Daṇḍin. Moreover, as noted by his commentators, Udbhaṭa often calls attention to many of his own altercations with Bhāmaha.138

  • 138For an example of cancelled categories, see “twinning” (yamaka), the second on Bhāmaha’s list, and which Udbhaṭa simply removes (likely to be subsumed by punaruktavadābhāsa, now the first ornament on his list despite not being there in Bhāmaha; see Bronner 2016: 113–114). For an example of a newly coined ornament see his discussion of "fusion" (saṅkara) added. For a case where Udbhaṭa accepts an ornament that Bhāmaha rejects, see the case of kāvyahētu (in Udbhaṭa, whereas Bhāmaha famously rejected hētu). For more on the contrarian mode of Udbhaṭa’s presentation of ornaments when compared to Bhāmaha, See Bronner forthcoming.

In short, while enshrining Bhāmaha as the founder of the tradition and pretending merely to provide the gist of his work, Udbhaṭa uses his predecessor’s text as a springboard for a radical beginning that is entirely his. Moreover, he seems to enjoy, here as elsewhere, the persona of an academic mischief-maker with a wink to his good scout guise. I believe that what he does to Pāṇini and Patañjali in grammar, Br̥haspati and his traditional commentators in Cārvāka land, and Bhāmaha in literary theory, he does to Kālidāsa, the grand patriarch of Sanskrit poetry, and presumably with the same deliberate method and sense of satisfaction.

I realize that this, too, is perhaps not a full explanation regarding the conception of Udbhaṭa’s poem, about which so much is still unknown: Were there other sections to this work? Where does it begin and how does it end? Did the additional sections, if they existed, illustrate other aspects of linguistic and literary theory in the manner of Bhaṭṭi’s Poem (Bhaṭṭikāvya)? Many of these key questions remain unanswerable for now. But I believe that the above analysis helps to account for the poem’s reception. If there was, indeed, a longer Kumārasambhava by Udbhaṭa, from which the extended passage preserved in the Compendium is an excerpt, it quickly disappeared with hardly a trace. Not a single verse from it is ever quoted as such, and not a single manuscript is known to have survived, this while all the other known works by Udbhaṭa were still current in Kashmir and beyond and were being cited by prominent thinkers in the centuries after his death.

Indeed, when considering the commentarial practices of Pratīhārēndurāja and Tilaka, the two readers who engaged with Udbhaṭa’s verses as part of their commentary on his Compendium, a striking picture emerges. The commentators decidedly ignore the contents of all illustration verses. Sure, they occasionally gloss certain words that seem in need of explication, but they do the bare minimum in that domain and instead explain the verses merely as illustrations of the figures in question. Yet they never explain the poetry as poetry or comment upon Udbhaṭa’s subversive portrayal of his protagonist. So, to give just one random example, when Udbhaṭa says that “[Śiva's] mighty light of perfect knowledge / began to firefly,” Pratīhārēndurāja glosses the word “firefly,” explains the denominative grammatical formation in question, runs through the different elements of the simile as provided here (the “mighty light” is the subject of comparison, the light produced by a firefly, the standard, and so on), and explains how this illustration fits Udbhaṭa’s complex scheme of simile subtypes. But he completely avoids the implications of this belittling comparison, certainly against the background of Kālidāsa’s intertext.139 This is his policy throughout: he offers no observation regarding any of the individual verses and has nothing whatsoever to say about the overall picture they portray or, indeed, about the larger poem, of which he is presumably aware.

  • 139Pratīhārēndurāja p. 25. Tilaka does exactly the same (p. 18).

Pratīhārēndurāja is not alone. Udbhaṭa’s Compendium remained an important reference point for later poetic theory, and thus the illustration verses preserved in it could not be entirely ignored. But as far as I can tell, the verses were dealt with strictly as illustrations of their figurative categories and nothing more, and as noted at the outset, this practice is continuous even today in studies and editions of Udbhaṭa’s legacy. Udbhaṭa has succeeded, by virtue of including these verses in his Compendium, to ensure the survival of 94 verses on the theme of Śiva’s falling in love, with their subversion of Kālidāsa and an upside-down version of God. But the flip side of his success was the weird status that his poetry has gained, somewhat akin to the deadly poison that Śiva prominently stores in his throat from the beginning of time: neither spit, nor swallow.

Notes

  1. ↑Rājataraṅgiṇī 4.402–502, cf. Bronner 2013.

  2. ↑Rājataraṅgiṇī 4.495, cf. Bronner 2013: 173–174 n. 38. The salary of no other academy member is mentioned.

  3. ↑Bronner 2016; Bronner forthcoming.

  4. ↑I cannot provide here a complete survey of the mentions and citations of Udbhaṭa’s works (the latter are often unnamed) in subsequent works on poetics. Let me just briefly mention that in the opening sentence of his Alaṅkārasarvasva, which starts with a historical prelude, Ruyyaka says “Here, to begin with, the ancient makers of ornaments, Bhamaha, Udbhaṭa, and the others” (iha hi tāvad bhāmahōdbhaṭaprabhr̥tayaḥ cirantanālaṅkārakārāḥ…; AS p. 3). The term audbhaṭāḥ in the plural is found, for example, in Abhinavagupta’s Abhinavabhāratī. See Treatise on Theater, p. 258 of vol. 1.

  5. ↑Shah 1972: 7, Solomon 1977–1978: 992 (where the question of the identity between the different Udbhaṭas is left open), Bronkhorst 2008: 297–298, Pollock 2016: 65–66, Bronner 2016: 139–141.

  6. ↑One exception is Narayana Daso Banhatti, who in his introduction to his edition of the Compendium dedicates two pages (Banhatti 1982: xiv–xv) to the poem, and notes that “sometimes the natural order of verses seems to be changed.” He also provides in an appendix “[a] list of all examples given by Udbhaṭa in the order in which they occur in the text” (appendix iii, p. xii). But other than a brief summary of the text and a note that it is “a great pleasure to read the poem even more than once” (xiv–xv), he offers no analysis.

  7. ↑Bronner forthcoming.

  8. ↑Kāvyālaṅkārasārasaṅgraha (Ramaswami Sastri 1931: 12, ll. 15–16): vr̥ttānurōdhād upamā prāgupadiṣṭāpi svayaṁkr̥takumārasambhavākhyakāvyasaṁgatyanurōdhād dīpakasya paścāl lakṣitā. parivartanē cōttaratrāyam ēva viśēṣō jñēyaḥ.

  9. ↑Note that, in Pratīhārēndurāja’s reading, the order in which Udbhaṭa illustrates the five subcategories of Gujarati alliteration is also different from that in which he defines them (verses 6-10 below; Tilaka, however, provides the examples in the order of the categories). Here a different order would have changed little in the meaning of the poem, and hence Pratīhārēndurāja’s insisting on a different order may have been inspired by having access to a manuscript of the poem itself, but this is of course speculative and inconclusive.

  10. ↑Kāvyaprakāśa, p. 252, l. 18.

  11. ↑For example, Dhvanyālōka p. 539: yathā bhagavatī pārvatī kumārasambhavē.

  12. ↑He cites two illustration verses from Udbhaṭa’s stanzas already in the Compendium, both already precited: the first (nētraiḥ, p. 228) by multiple sources, and the second at least by Sōmēśvara (Saṅkēta, p. 294, although without the explicit reference to Udbhaṭa that Māṇikyacandra gives). On the relative chronology of Māṇikyacandra and Sōmēśvara, see Bronner and Ollett 2024.

  13. ↑A. K. Warder noted in passing that of the few verses attributed to Udbhaṭa in the anthologies, “one describing Umā,” “is probably from his Kumārasambhava” (Warder 1983: 472). Note, however, that the verse is typically cited amidst others that depict Brahma’s over-the-top investment in the beauty of certain young women, and that it, too, does the same, without naming its heroine (see, for example, Subhāṣitaratnakōṣa of Vidyākara v. 455). Thus, while the verse depicts the young woman’s creation as offering a new life to Kāma whom Śiva has burnt, there is nothing to force us to see it as belonging in the narrative of Śiva and Umā’s falling in love, let alone to be from Udbhaṭa’s lost poem on this topic.

  14. ↑One unique reference to Udbhaṭa’s poem is in the Telugu Kumārasambhavamu of Nanne Cōḍa: “Udbhaṭa composed a Kumāra-sambhava / on the theme of Śiva’s play / and pleased the god with this poem, / which is the whole of figuration / with kāvya deep inside” (translation from Narayana Rao and Shulman 2002: 116). A study of the relationship between this Telugu poem and Udbhaṭa’s work is a desideratum, but note that here, too, the reference is only to the illustration of ornaments, that is, presumably the part preserved in the Compendium.

  15. ↑Warder 1983: 469.

  16. ↑Pratīhārēndurāja on verse 1 (UKS p. 3): tadāprabhr̥ti satīviyōgād ārabhya. Tilaka, by contrast, rightly identifies this moment as following the burning of Kāma (kāmadahanād ārabhya, p. 4, l. 1; he also has a slightly different reading: tataḥ prabhr̥ti).

  17. ↑Note that in Kālidāsa’s version, the burning of Kāma is preceded by the advent of the spring season, yet out of order (ākālikīm … madhupravr̥ttim; KKS 3.34). The cycle of seasons depicted in Umā’s penance in Kālidāsa’s poem seems to bypass the autumn.

  18. ↑KKS 5.7; UKS 37.

  19. ↑
    tadāprabhr̥ti niḥsaṅgō nāgakuñjarakrttibhr̥tśitikaṇṭhaḥ kālagalatsatīśōkānalavyathaḥ
    [PIR 1.*1; T 1.4; T reads tataḥ for tadā]

    The ornament is “apparent repetition” (punaruktavadābhāsa), likely Udbhaṭa’s own invention. The idea is that, at first, certain phrases appear redundant, and then, upon further reflection, the reader realizes that the second of each pair has a unique, different meaning or reading (Bronner 2016: 113–114). In this verse, both nāga and kuñjara seem, at first blush, to signify “elephant,” until one realizes that the latter here means “the best of its kind”; and both śitikaṇṭhaḥ and kālagala- seem to signify “black throat,” until one realizes that the latter is really kālagalat-, referring to the gradual passing of the pain of separation by time. I have tried to recreate the effect in English (at the cost of deviating from the literal meaning of the second half), with “mammoth elephant,” “concealed … hide,” and “blackened … charred.”

  20. ↑
    sa dēvō divasān ninyē tasmiñ śailēndrakandarēgariṣṭhagōṣṭhīprathamaiḥ pramathaiḥ paryupāsitaḥ
    [PIR 1.*2; T 1.5]

    The ornament is “enticing alliteration” (chēkānuprāsa), where the effect is based on several pairs of similar sound patterns (as in sa dēvō divasān etc.). Again, I tried to recreate the effect in English with partial success.

  21. ↑
    tatra tōyāśayāśēṣavyākōśitakuśēśayācakāśē śālikiṁśārukapiśāśāmukhā śarat
    [PIR 1.*3; T 1.7]

    The ornament here is “alliteration” (anuprāsa), which Udbhaṭa divides according to the various euphonic modes. The verse here falls under the “harsh euphonic mode” (paruṣavr̥tti), which abounds in ś and ṣ sounds, consonant clusters that include r, retroflex consonants, and combinations such as hl, hv, and hy. The example verse is primarily dominated by sh sibilants, as is my translation.

  22. ↑
    sāndrāravindavrndōtthamakarandāmbubindubhiḥsyandibhiḥ sundarasyandaṁ nanditēndindirā kvacit
    [PIR 1.*4; T 1.9]

    The second type of alliteration is based on the “urbane euphonic mode” (upanāgarikavr̥tti), which is replete with duplicated letters or consonant clusters that contain nasals followed by homorganic stops. The original is dominated by the nd cluster, something I tried to replicate in the translation.

  23. ↑
    kēlilōlālimālānāṁ kalaiḥ kōlāhalaiḥ kvacitkurvatī kānanārūḍhaśrīnūpuraravabhramam
    [PIR 1.*5; T 1.11]

    The third and last type of alliteration is based on the “rustic euphonic mode” (grāmyavr̥tti), which is said to abound in the remaining sounds as appropriate. In the Sanskrit the predominant sound is l, which I tried to replicate in the English.

  24. ↑
    kāśāḥ kāśā ivōdbhāṁsi sarāṁsīva sarāṁsi cacētāṁsy ācikṣipur yūnāṁ niṁnagā iva niṁnagāḥ
    [PIR 1.*6; T 1.19; T reads ābhānti for udbhāṁsi]

    This is the first example of the so-called “Gujarati alliteration” (lāṭānuprāsa), the exact repetition of stems or words with no change in form or meaning, so long as each instance serves a different purpose. This ornament has five subtypes. Here is illustrated the fourth subtype, where standalone words such as “white grass” (kāśāḥ), “lakes” (sarāṁsi) and “rivulets” (niṁnagāh) get repeated. Note that T gives the verses illustrating “Gujarati alliteration” in a different order that agrees with that of their definition.

  25. ↑
    striyō mahati bhartrbhya āgasy api na cukrudhuḥbhartārō ’pi sati strībhya āgasy api na cukrudhuḥ
    [PIR 1.*7; T 1.20]

    “Gujarati alliteration” type 5, when the entire metrical unit is repeated.

  26. ↑
    kvacid utphullakamalā kamalabhrāntaṣaṭpadāṣaṭpadakvāṇamukharā mukharasphārasārasā
    [PIR 1.*8; T 1.16]

    “Gujarati alliteration,” type 1, namely, the repetition of two identical words (kamala, ṣaṭpada, mukhara) while each forms part of a different compound. I partially replicated this in the translation with “lotus” and “hum.”

  27. ↑
    jitānyapuṣpakiñjalkakiñjalkaśrēṇiśōbhitamlēbhē ’vataṁsatāṁ nārīmukhēnduṣv asitōtpalam
    [PIR 1.*9; T 1.18]

    “Gujarati alliteration,” type 3, where the repeated words form part of a single compound. In the original, the first line (“With filaments unlike the filaments of any other flower”) constitutes a single compound word, something that I did not replicate in the English.

  28. ↑
    padminīṁ padminīgāḍhaspr̥hayāgatya mānasātantardanturayām āsur haṁsā haṁsakulālayāt
    [PIR 1*10; T1.17, reads padminīḥ]

    “Gujarati alliteration” type 2, when only one of the repeated pair of words is part of a compound and the other stands alone (eg., padminīm vs. padminī-). I replicated this with “gander” and “pond.”

  29. ↑
    jyōtsnāmbunēndukumbhēna tārākusumaśāritamkramaśō rātrikanyābhir vyōmōdyānam asicyata
    [PIR 1.*11; T 1.23]

    The figure is “identification” (rūpaka) of the “full-set type” (samastavastuviṣaya), wherein every subject (moonlight etc.) is explicitly identified with its standard (water etc.). For a discussion of Udbhaṭa’s groundbreaking understanding of rūpaka, see Bronner 2016: 92–99, 106–110, and Bronner forthcoming.

  30. ↑
    utpatadbhiḥ patadbhiś ca picchālīvālaśālibhiḥrājahaṁsair avījyanta śaradaiva sarōnr̥pāḥ
    [PIR 1.*12; T 1.24; reads piñcha for piccha]

    This is the second of four types of “identification,” one that is “confined in presence” (ēkadēśavivarti). Here some of the identifications are explicit (e.g., the king is explicitly identified with the lake), but some are only implied (e.g., autumn as the king’s whisk-lady).

  31. ↑
    vanāntadēvatāvēṇyaḥ pānthastrīkālaśr̥ṅkhalāḥmārapravīrāsilatā bhr̥ṅgamālāś cakāśirē
    [PIR 1.*14; T 1.26]

    This is “chain identification” (mālārūpaka), the third subtype of rūpaka. Here the same subject (“rows of black bees”) is repeatedly identified with a variety of standards (braids of forest deities etc.).

  32. ↑
    āsāradhārāviśikhair nabhōbhāgaprabhāsibhiḥprasādhyatē sma dhavalair āśārājyaṁ balāhakaiḥ
    [PIR 1.*14; T 1.27]

    This is the last type of “identification,” “confined in capacity” (ēkadeśavr̥tti). This is a somewhat obscure category, and it partly depends on the dual meaning of the verb prasādhyate, which could be taken to have both aesthetic and military connotations. I tried to replicate this ambiguity with “redeemed.” See Bronner 2016: 108–109 and Bronner forthcoming for a discussion.

  33. ↑
    saṁjahāra śaratkālaḥ kadambakusumaśriyaḥprēyōviyōginīnāṁ ca niḥśēṣasukhasampadaḥ
    [PIR 1.*15; T 1.29]

    This is the first of three illustrations of dīpaka, or “illumination,” an ornament based on ellipsis. Here the verb (“liquidated”) appears only once yet construes with two separate objects (“the kadamba blossoms” and “the store of happiness” of separated women). It is found in the beginning, which makes it a case of “beginning illumination” (ādidīpaka).

  34. ↑
    vidēśavasatir yātapatikājanadarśanamduḥkhāya kēvalam abhūc charac cāsau pravāsinām
    [PIR 1.*15; T 1.30]

    In this “middle illumination,” the verbal element “crushed the hearts of travelers” (duḥkhāya kēvalam abhūt) is situated in the middle, from where it construes with different entities (“Life in foreign parts,” etc.) in different parts of the verse.

  35. ↑
    tadānīṁ sphītalāvaṇyacandrikābharanirbharaḥkāntānanēndur induś ca kasya nānandakō ’bhavat
    [PIR 1.*17; T 1.31]

    In this “end illumination,” filling every heart with bliss, which construes with both moon-faces and the actual moon, is found in the end.

  36. ↑
    kṣaṇaṁ kāmajvarōtthityai bhūyaḥ santāpavr̥ddhayēviyōginām abhūc cāndrī candrikā candanaṁ yathā
    [PIR 1.*18; T 1.39]

    This is the first of many verses illustrating “simile” (upamā), here with explicit syntax and the word yathā (“as if it were”).

  37. ↑
    nētrair ivōtpalaiḥ padmair mukhair iva saraḥśriyaḥtaruṇya iva bhānti sma cakravākaiḥ stanair iva
    [PIR 1.*19; T 1.40]

    “Simile,” explicit syntax, this time using the word “like” (iva). Note that this is what Daṇḍin called “reverse simile” (viparyāsopamā), where the established standards become subjects and vice versa (KĀ 2.17–18). In Udbhaṭa’s purely formal analysis of the simile, the identity of the standard is immaterial, but he does gesture, by way of examples, that such reversals are possible.

  38. ↑
    prabōdhād dhavalaṁ rātrau kiñjalkālīnaṣaṭpadampūrṇēndubimbapratimam āsīt kumudakānanam
    [PIR 1.*20; T 1.41, reads: śaśāṅkabimbēna samaṁ babhau]

    “Simile,” nonexplicit syntax, using the word “equal” (here pratimā). PIR offers the following reading as an alternative to the second half: akhaṇḍēnēndunā tulyam āsīt kumudakānanam (which he enumerates 1.*29), likely to squeeze out of this example verse an additional linguistic method of expressing the same simile. T’s alternative reading (śaśāṅkabimbapratimam) serves a similar purpose. I ignore both alternatives in my translation.

  39. ↑
    api sā sumukhī tiṣṭhēd dr̥ṣṭēḥ pathi kathañcanaaprārthitōpasampannā patitānabhravr̥ṣṭivat
    [PIR 1.*22; T 1.43; the following verse precedes this one in T’s edition]

    This verse continues the manifold analysis of simile, here using the nominal suffix vat (“-like”), and having the actual semblance implied.

  40. ↑
    kiṁ syur utkalikā madvat tasyā api nirargalāḥakāṇḍōḍḍāmarānaṅgahatakēna samarpitāḥ
    [PIR 1.*23; T 1.42]

    This is an illustration of a simile where the actual semblance is explicit, based again on the nominal suffix vat.

  41. ↑
    iti kālē kalōllāpikādambakulasaṅkulētridaśādhīśaśārdūlaḥ paścāttāpēna dhūrjaṭiḥ
    tāṁ śaśicchāyavadanāṁ nīlōtpaladalēkṣaṇāmsarōjakarṇikāgaurīṁ gaurīṁ prati manō dadhau
    [PIR 1.24-25; T 44-45]

    This illustrates cases of simile which are concise, in the sense that various elements are elided thanks to the use of compounds. There is a play on the word gaurī (both the color in question and the goddess’s personal name).

  42. ↑
    sa duḥsthīyan kr̥tārthō ’pi niḥśēṣaiśvaryasampadānikāmakamanīyē ’pi narakīyati kānanē
    [PIR 1.*26, T 46]

    The similes in this verse likewise illustrate the elision of the word such as “like."” Here this is done thanks to the denominative kyac suffix, an effect which I tried to replicate in my translations (“vagabonding,” “was hell”).

  43. ↑
    kr̥śānavaj jagat tasya paśyatas tāṁ priyāṁ vinākhadyōtāyitum ārabdhaṁ tattvajñānamahāmahaḥ
    [PIR 1.*27; T 1.47]

    The similes in this verse likewise illustrate various elisions based on the kvip and kyaṅ denominative suffixes. I tried to recreate the effect with “a-pyring” and “began to firefly.”

  44. ↑
    tasyētaramanōdāham adahat prajvalanmanaḥumāṁ prati tapaḥśaktyākr̥ṣṭabuddhēḥ smarānalaḥ
    [PIR 1.*28; T 1.48]

    In this simile, a ṇamul ending (dāham) implies the word “like.” As Pratīhārēndurāja notes [p. 26, l. 16]: “Here the heart of everyone else, in the sense of ordinary people, is the standard of comparison, and Śiva’s heart, the subject. Being scorched is the shared attribute.”

  45. ↑
    sa dagdhavigrahēṇāpi vīryamātrasthitātmanāspr̥ṣṭaḥ kāmēna sāmānyaprāṇicintam acintayat
    [PIR 1.*29; T 1.49]

    In this simile, a ṇamul ending referringto an agent (cintam) suggests the word “like.”

  46. ↑
    caṇḍālakalpē kandarpaṁ pluṣṭvā mayi tirōhitēsaṁjātātulanairāśyā kiṁ sā śōkān mr̥tā bhavēt
    [PIR 1.*30; T 1.50]

    Simile based on the suffix kalpa, here translated as “all but.”

  47. ↑
    viralās tādr̥śō lōkē śīlasaundaryasampadaḥniśāḥ kiyatyō varṣē ’pi yāsv induḥ pūrṇamaṇḍalaḥ
    [PIR 1.*31; T 1.53]

    This last example of simile is based on parallelism (prativastūpamā). This is probably still Śiva speaking in his own voice about the beauty and character of Umā.

  48. ↑
    ahō smarasya māhātmyaṁ yad rudrē ’pi daśēdr̥śīiyad āstāṁ samudrāmbhaḥ kumbhair mānē tu kē vayam
    [PIR 2.*1; T 2.4]

    This is the first of two illustrations of “dismissal” (ākṣēpa). In this variety, the dismissal is future oriented.

  49. ↑
    iti cintayatas tasya citraṁ cintāvadhir na yatkva vā kāmavikalpānām antaḥ kālasya cēkṣitaḥ
    [PIR 2.*2; T 2.5]

    This “dismissal” is oriented towards what was already stated.

  50. ↑
    tan nāsti yan na kurutē lōkō hy atyantakāryikaḥēṣa śarvō ’pi bhagavān baṭūbhūya sma vartatē
    [PIR 2.*3; T 2.8]

    This is the first example of the ornament “citing another case” (arthāntaranyāsa). In this variety, the corroborating sentence (the general statement about “"a guy” doing whatever it takes) appears before the corroborated (Śiva’s taking on the guise of a schoolboy), and there is use of the particle hi (“after all”). T reads lōkē for lōkaḥ.

  51. ↑
    pracchannā śasyatē vr̥ttiḥ strīṇāṁ bhāvaparīkṣaṇēpratasthē dhūrjaṭir atas tanuṁ svīkr̥tya bāṭavīm
    [PIR 2.*4; T 2.9]

    The second example of “citing another case,” corroboration again first, but no use of the particle hi.

  52. ↑
    harō ’tha dhyānam ātasthau saṁsthāpyātmānam ātmanāvisaṁvadēd dhi pratyakṣaṁ nirdhyātaṁ dhyānatō na tu
    [PIR 2.*5; T 2.10]

    “Citing another case,” corroboration statement later, use of the particle hi.

  53. ↑
    apaśyac cātikaṣṭāni tapyamānāṁ tapāṁsy umāmasambhāvyapatīcchānāṁ kanyānāṁ kāparā gatiḥ
    [PIR 1.*6; T 2.11]

    “Citing another case,” corroboration later, no hi.

  54. ↑
    sa gaurīśikharaṁ gatvā dadarśōmāṁ tapaḥkr̥śāmrāhupītaprabhasyēndōr jayantīṁ dūratas tanum
    [PIR 2.*7; T 2.13]

    The figure is “distinction” (vyatireka), and in this case, the cause for the subject’s outdoing its standard is implied.

  55. ↑
    padmaṁ ca niśi niḥśrīkaṁ divā candraṁ ca niṣprabhamsphuracchāyēna satataṁ mukhēnādhaḥ prakurvatīm
    [PIR 2.*8; T 2.14]

    The second illustration of “distinction”; this time the cause (“with a sheen that shone nonstop”) is explicit.

  56. ↑
    śīrṇaparṇāmbuvātāśakaṣṭē ’pi tapasi sthitāmsamudvahantīṁ nāpūrvaṁ garvam anyatapasvivat
    [PIR 2.*9; T 16]

    “Distinction” based on dissimilarity. Compare KKS 5.28.

  57. ↑
    yā śaiśirī śrīs tapasā māsēnaikēna viśrutātapasā tāṁ sudīrghēṇa dūrād vidadhatīm adhaḥ
    [PIR 2.*10; T 2.16]

    “Distinction” based on “embrace” (ślēṣa), in this case the double meaning of the word tapas as a name of a month and the word for penance. Compare to KKS 5.28.

  58. ↑
    aṅgalēkhām akāśmīrasamālambhanapiñjarāmanalaktakatāmrābhām ōṣṭhamudrāṁ ca bibhratīm
    [PIR 2.*11; T 2.20]

    The ornament is “evocation” (vibhāvanā), where the result is manifest despite the absence of its normal cause.

  59. ↑
    dantaprabhāsumanasaṁ pāṇipallavaśōbhinīmtanvīṁ vanagatāṁ līnajaṭāṣaṭcaraṇāvalim
    [PIR 2.*12; T 2.22]

    The ornament is “condensed speech” (samāsokti). The verse is meant to imply her resemblance to a forest creeper. According to T, the bees cling to the matted locks, rather than being identified with it.

  60. ↑
    tapastējaḥsphuritayā nijalāvaṇyasampadākr̥śām apy akr̥śām eva dr̥śyamānām asaṁśayam
    [PIR 2.*13; T 2.26]

    This is the first illustration of “intensification” (atiśayōkti), here in the type of identity given difference. With this verse ends a string of verses (38-43) that depict Umā as the object of Śiva’s gaze.

  61. ↑
    acintayac ca bhagavān ahō nu ramaṇīyatātapasāsyāḥ kr̥tānyatvaṁ kaumārād yēna lakṣyate
    [PIR 2.*14; T 2.27]

    The second type of “intensification,” defined as difference given identity.

  62. ↑
    patēd yadi śaśidyōtacchaṭā padmē vikāśinimuktāphalākṣamālāyāḥ karē 'syāḥ syāt tadōpamā
    [PIR 2.*15; T 2.28]

    The third type of “intensification” in Udbhaṭa’s scheme involves a hypothetical, imagined entity.

  63. ↑
    manyē ca nipatanty asyāḥ kaṭākṣā dikṣu pr̥ṣṭhataḥprāyēṇāgre tu gacchanti smarabāṇaparamparāḥ
    [PIR 2.*16; T 2.29]

    Udbhaṭa understands the last type of “intensification” as a case where the effect (falling in love) precedes its cause (eye contact with the flirtatious love interest).

  64. ↑
    mr̥ṇālahaṁsapadmāni bāhucaṅkramaṇānanaiḥnirjayantyānayā vyaktaṁ nalinyaḥ sakalā jitāḥ
    [PIR 3.*1; T 3.3]

    The ornament is “respective enumeration” (yathāsaṅkhya). T reads tarjayantyā for nirjayantyā.

  65. ↑
    asyāḥ sadārkabimbasthadr̥ṣṭipītātapair japaiḥśyāmikāṅkēna patitaṁ mukhē candrabhramād iva
    [PIR 3.*2; T 3.6]

    This is “seeing as” (utprēkṣā) type 1. Udbhaṭa says, in his definition, that words such as “like” can express “seeing as.” Yet in both of his examples use only the word “like” (iva) itself. To indicate this, I chose to translate iva as “like” in this example. Note the closeness of this verse to KKS 5.21, also involving a “seeing as” of the śyāmikā (“dark tan mark”) on Umā’s face.

  66. ↑
    kapōlaphalakāv asyāḥ kaṣṭaṁ bhūtvā tathāvidhauapaśyantāv ivānyōnyam īdr̥kṣāṁ kṣāmatāṁ gatau
    [PIR 3.*3; T 3.7]

    This is Udbhaṭa’s second illustration of “seeing as.”

  67. ↑
    kṣaṇaṁ naṁṣṭvārdhavalitaḥ śr̥ṅgēṇāgrē kṣaṇaṁ nudanlōlīkarōti praṇayād imām ēṣa mr̥gārbhakaḥ
    [PIR 3.*4; T 3.9]

    This is Udbhaṭa’s example of “factual statement” (svabhāvōkti). I take the root naś/naṁś in its sense of running away. T’s explanation (ruditvā) makes little sense in this context. The illustration of the third group of ornaments ends here, but the depiction of the fawn carries over to the next verse, with which the fourth group begins.

  68. ↑
    iyaṁ ca sutavāllabhyān nirviśeṣā spr̥hāvatīullāpayitum ārabdhā kr̥tvēmaṁ krōḍa ātmanaḥ
    [PIR 4.*1; T 4.3]

    The ornament is “endearing” (prēyasvat). Compare to KKS 5.15 (similar theme-wise, but not in the actual wording).

  69. ↑
    iti bhāvayatas tasya samastān pārvatīguṇānsambhr̥tānalpasaṅkalpaḥ kandarpaḥ prabalō ’bhavat
    [PIR 4.*2; T 4.6]

    This is the first of three verses illustrating the ornament “flavored” (rasavat). The rasa is, of course, the erotic. Here it is evoked by its proper term and that for its underlying stable emotion (sthāyibhāva), namely, “passion,” and by the mention of Pārvatī “thus conceived … virtue by virtue,” as the “foundational factor” (ālambanavibhāva).

  70. ↑
    svidyatāpi sa gātrēṇa babhāra pulakōtkaramkadambakalikākōśakēsaraprakarōpamam
    [PIR 4.*3; T 4.7]

    In this second illustration of “flavored,” the emphasis is on Śiva’s sweat and hair standing on end as accompanying bodily states of the erotic rasa.

  71. ↑
    kṣaṇam autsukyagarbhiṇyā cintāniścalayā kṣaṇamkṣaṇaṁ pramōdālasayā dr̥śā ’syāsyam abhūṣyata
    [PIR 4.*4; T 4.8]

    In this third and last illustration of “flavored,” the emphasis is on Śiva’s secondary emotions of longing, worry, and delight that help evoke the erotic rasa and on the acting registers of Śiva ’s eye movements.

  72. ↑
    tathā kāmō ’sya vavr̥dhē yathā himagirēḥ sutāmsaṅgrahītuṁ pravavr̥tē haṭhēnāpāsya satpatham
    [PIR 4.*5; T 4.10]

    This is Udbhaṭa's ’s illustration of the ornament “prideful” (ūrjasvin), more on which in the following section.

  73. ↑
    yēna lambālakaḥ sāśraḥ karaghātāruṇastanaḥakāri bhagnavalayō gajāsuravadhūjanaḥ
    [PIR 4.*6; T 4.12]

    This is Udbhaṭa’s illustration for “roundabout speech” (paryāyōkta), where the effect (the lament of the demon’s wives) suggests its cause (his death at Śiva’s hands). Note that this verse alone illustrates the ornament in question, and the next one, which is syntactically part of the same sentence, has no illustrative purpose (I discuss this fact below).

  74. ↑
    sō ’pi yēna kr̥taḥ pluṣṭadēhēnāpy ēvam ākulaḥnamō ’stv avāryavīryāya tasmai makarakētavē
    [PIR 4.*7; T 4.13, where the reading is apāravīryāya and kusumadhavanē]

    The “Bearer of the Fish Banner” is Kāma, the god of love.

  75. ↑
    atha kāntāṁ dr̥śaṁ dr̥ṣṭyā vibhramāṁś ca bhramaṁ bhruvōḥprasannaṁ mukharāgaṁ ca rōmāñcaṁ svēdasaṅkulam
    smarajvarapradīptāni sarvāṅgāni samādadhatupāsarpad girisutāṁ giriśaḥ svastipūrvakam
    [PIR 4.*8-9; T 4.15-16]

    Like Pollock (2016: 347n150), I read dr̥ṣṭyā vibhramāṁś ca and rōmāñcaṁ with Tilaka, instead of dr̥ṣṭvā vibhramāc ca and rōmāñca- with PIR. This is Udbhaṭa’s illustration of “coterminous” (samāhita), which he understands, very uniquely, as the termination of rasa (“lay it all to rest”).

  76. ↑
    uvāca ca yataḥ krōḍē vēṇukuñjarajanmabhiḥmuktāphalair alaṅkāraḥ śabarīṇām apīcchayā
    [PIR 4.*10; T 4.18]

    This is the first of five verses that illustrate the ornament “magnificent” (udātta): first based on riches (likely the four first verses) and then on deeds. Pollock, following the commentators, reads krōḍē[a-] in the sense of wild boars, one of the eight traditional sources of pearls along with bamboos and elephants (Pollock 2016: 347n153). The word might also mean “breast” or “lap,” either with reference to the women or the mountain.

  77. ↑
    puṣṭyēndranīlavaiḍūryapadmarāgamayair viyatśirōbhir ullikhad yatra śikharaṁ gandhamādanam
    [PIR 4.*11; T 4.19]

    “Magnificent riches” (udātta), second example. T reads yasya for yatra.

  78. ↑
    uttarōpatyakā yasya pradhānasvarṇabhūmayaḥmahān marakatōrvīdhraḥ pādōpāntaṁ ca saṁśritaḥ
    [PIR 4.*12; T 4.20]

    “Magnificent riches” (udātta), third example.

  79. ↑
    babhūva yasya pātālapātinyāṁ saṅkṣayē kṣitaupatanaṁ na tayā sārdham āyāmas tu prakaṭy abhūt
    [PIR 4.*13; T 4.21]

    The commentators classify this verse as the fourth example of “magnificence” based on riches, although like the following one, it could also be based on deeds.

  80. ↑
    tasyādikrōḍapīnāṁsanigharṣē ’pi punaḥ punaḥniṣkampasya sthitavatō himādrēr bhavatī sutā
    [PIR 4.*14; T 4.22]

    “Magnificent deeds” (udātta).

  81. ↑
    svayaṁ ca pallavātāmrabhāsvatkaravirājinīprabhātasandhyēvāsvāpaphalalubdhēhitapradā
    [PIR 4.*15; T 4.25]

    This is the first of three verses illustrating bitextual “embrace” (ślēṣa). In Sanskrit, the “embrace” in the second half works as follows: Umā grants the wish of those who desire the fruit that is impossible to obtain, i.e., liberation (a-su-āpa-phala-lubdha-īhita-pradā); but segmented differently, the sun is that which gives good advice, or a boon to a person who does not desire the fruit of sleep (a-svāpa-phala-lubdhē hitapradā). I tried to very partially allude to this with “the fruit of awakening.” In the first half, kara may mean either hand or ray, and bhāsvat, either shining or the sun. I tried to replicate this pun with “light fingers.”

  82. ↑
    indukāntamukhī snigdhamahānīlaśirōruhāmuktāśrīs trijagadratnaṁ padmarāgāṅghripallavā
    [PIR 4.*16; T 4.26]

    “Embrace” (ślēṣa). Indukānta in Sanksrit means “friend to the moon” (in the sense of resembling it) and, hence, also the moonstone. Muktā-śrī may refer to the shine of a pearl, but read as muktā-aśrī, to something devoid of anything not shining or beautiful (I tried to replicate this with nacre/nay-care). The rest of the puns depend on whether the precious stones are taken to modify Umā’s body parts or to be identified with them.

  83. ↑
    apārijātavārtāpi nandanaśrīrbhuvisthitāabindusundarī nityaṁ galallāvaṇyabindukā
    [PIR 4.*17; T 4.27]

    Third illustration of “embrace” (ślēṣa). Here is how the original works: apārijātavārtā may mean either that the heaven is bereft of the Pārijāta (a coral wish-granting tree that, according to legend, grows in heaven), or someone who has no enemies left. I tried to replicate this with coral/quarrel. A-bindu-sundarī means containing not even a drop of beauty, or, if we read ab-indu-sundarī: beautify like the moon reflected in water. Nandana-śrī may mean either the riches of Indra’s heaven or beauty that delights. At first blush, the verse creates a contradiction: the riches of heaven are without its wish-granting tree, are found on earth, and with not even a drop of beauty. To resolve this antithesis, a second reading is supplied, referring to Umā, who is indeed on earth, delights with her beauty, and has no enemies. I tried to replicate this antithesis in the translation, though, of course, with limited success.

  84. ↑
    ētad dhi na tapaḥ satyam idaṁ hālāhalaṁ viṣamviśēṣataḥ śaśikalākōmalānāṁ bhavādr̥śām
    [PIR 5.*1; T 5.4]

    The ornament is “denial” (apahnuti).

  85. ↑
    maharddhini gr̥hē janma rūpaṁ smarasuhr̥dvayaḥtathāpi na sukhaprāptiḥ kasya citrīyatē na dhīḥ
    [PIR 5.*2; T 5.7]

    This is the first instance of “exceptionality” (viśēṣōkti), where the effect does not materialize despite the presence of its causes. Here the cause for this failure is not mentioned.

  86. ↑
    itthaṁ visaṁṣṭhulaṁ dr̥ṣṭvā tāvakīnaṁ vicēṣṭitamnōdēti kimapi praṣṭuṁ satvarasyāpi mē vacaḥ
    [PIR 5.*3; T. 5.8]

    “Exceptionality” (viśēṣōkti). Tilaka reads “my mind” (manas) instead of “speech” (vacas), translated as “tongue.” Here the explanation is displayed (“seeing you behave so recklessly”).

  87. ↑
    yad vā māṁ kiṁ karōmy ēṣa vācālayati vismayaḥbhavatyāḥ kvāyam ākāraḥ kvēdaṁ tapasi pāṭavam
    [PIR 5.*4; T 5.10]

    The ornament, based on the incongruity between her delicate body and her severe penance, is “antithesis” (virodha); later theorists would label such instances viṣama.

  88. ↑
    tvadaṅgamārdavaṁ draṣṭuḥ kasya cittē na bhāsatēmālatīśaśabhr̥llēkhākadalīnāṁ kaṭhōratā
    [PIR 5.*5; T 5.12. T reads dr̥ṣṭvā for draṣṭuḥ]

    This is “the yoke of equivalence” (tulyayogitā), and here, unlike in the following example, the standards (jasmine sprigs etc.) are not pertinent to the context.

  89. ↑
    yōgapaṭṭō jaṭājālaṁ tāravī tvaṅmr̥gājinamucitāni tavāṅgasya yady amūni tad ucyatām
    [PIR 5.*6; T 5.13]

    In this second instance of “the yoke of equivalence,” the standards are pertinent to the situation.

  90. ↑
    yānti svadēhēṣu jarām asamprāptōpabhōktr̥kāḥphalapuṣparddhibhājō ’pi durgadēśavanaśriyaḥ
    [PIR 5.*7; T 5.15]

    This is “praise of the irrelevant” (aprastutapraśaṁsā): the speaker depicts the fate of fruit unpicked in the forest, while the real intention is Pārvatī herself.

  91. ↑
    dhig ananyōpamām ētāṁ tāvakīṁ rūpasampadamtrailōkyē ’py anurūpō yad varas tava na labhyatē
    [PIR 5.*8; T 5.17]

    This is “praise in disguise” (vyājastuti), where the literal damning of Umā’s beauty ends up being a compliment (albeit a somewhat left-handed one).

  92. ↑
    vinōcitēna patyā ca rūpavaty api kāminīvidhuvandhyavibhāvaryāḥ prabibharti viśōbhatām
    [PIR 5.*9; T 5.19]

    The ornament is “illustration” (nidarśanā).

  93. ↑
    yady apy atyantam ucitō varēndus tē na labhyatētathāpi vacmi kutrāpi kriyatām ādarō varē
    [PIR 5.*10; T 5.21]

    T reads ādarō ’parē. This is Udbhaṭa’s first type of “fusion” (saṅkara), based on a “doubt” (sandēha) as to which ornament out of several captures the reader’s attention.

  94. ↑
    itthaṁ sthitir varārthā cēn mā kr̥thā vyartham arthitāmrūpēṇa tē yuvā sarvaḥ pādabaddhō hi kiṅkaraḥ
    [PIR 5.*11; T 5.23]

    “Fusion” of both sound and sense ornaments.

  95. ↑
    maivam ēvāstha sacchāyavarṇikācārukarṇikāambhōjinīva citrasthā dr̥ṣṭimātrasukhapradā
    [PIR 5.*12; T 5.24]

    “Fusion,” of two ornaments, “simile” and “embrace,” in one portion of the sentence. The “ears” of the lotus are its central seed pods, and Umā has a pair of beautiful ears. The first part could also mean “Don't just sit there!” But I translated it the way I did based on the notion that she delights “like a lotus in a painting,” that is, by simply being idle.

  96. ↑
    harēṇēva smaravyādhas tvayānaṅgīkr̥tō ’pi santvadvapuḥ kṣaṇam apy ēṣa dhārṣṭyād iva na muñcati
    [PIR 5.*13; T 5.26]

    “Fusion,” mutual dependency between ornaments. It is hard to recreate the pun on anaṅgīkr̥taḥ (not embraced, rendered bodiless). Recall that the speaker, Śiva, is disguised. He refers to himself as if he were a different person.

  97. ↑
    śirāṁsi paṅkajānīva vēgōtpātayatō dviṣāmājau karōpamaṁ cakraṁ yasya cakrōpamaḥ karaḥ
    [PIR 5.*14; T 5.28]

    T reads vēgāt pātayatō. This is “comparison with the standard of comparison” (upameyopamā), where the subject and the standard are compared with each other in succession. The speaker, Śiva in disguise, has now begun a long depiction of Viṣṇu in a series or relative clauses. Viṣṇu’s identity as the subject of this praise will only be revealed in verse 90 below.

  98. ↑
    dyujanō mr̥tyunā sārdhaṁ yasyājau tārakāmayēcakrē cakrābhidhānēna praiṣyēṇāptamanōrathaḥ
    [PIR 5.*15; T 5.30]

    T reads prēṣya-. The ornament is “concurrence” (sahōkti), in this case, of two actions in one.

  99. ↑
    urō datvāmarārīṇāṁ yēna yuddhēṣv agr̥hyatahiraṇyākṣavadhādyēṣu yaśaḥ sākaṁ jayaśriyā
    [PIR 5.*16; T 5.32]

    The ornament is “reciprocity” (parivr̥tti), where some action is figuratively portrayed as an act of give and take (“gave his bare chest … gained fame and glory”). For PIR, this is a barter of equal elements (chest and fame, he says, are equals).

  100. ↑
    nētrōragavalabhrāmyanmandarādriśiraścyutaiḥratnair āpūrya dugdhābdhiṁ yaḥ samādatta kaustubham
    [PIR 5.*17; T 5.33]

    The second illustration of “reciprocity” where the thing sacrificed (precious stones) is inferior to that won (the Kaustubha gem).

  101. ↑
    yō balau vyāptabhūsīmni makhēna dyāṁ jigīṣatiabhayaṁ svargasadmabhyō datvā jagrāha kharvatām
    [PIR 5.*18; T 5.34]

    In this third illustration of “reciprocity,” the thing given (safety) is far superior to that gained (smallness, by becoming Vāmana, the dwarf).

  102. ↑
    hastē kim asya niḥśēṣadaityahr̥ddalanōdbhavaḥyaśaḥsañcaya ēṣa syāt piṇḍībhāvō ’sya kiṁ kr̥taḥ
    [PIR 6.* 1; T 6.3]

    This is the first illustration of the ornament “in doubt” (sasandēha).

  103. ↑
    nābhipadmaspr̥hāyātaḥ kiṁ haṁso naiṣa cañcalaḥiti yasyābhitaḥ śaṅkham aśaṅkiṣṭārjavō janaḥ
    [PIR 6.*2; T 6.4]

    “In doubt” continued.

  104. ↑
    nīlābdaḥ kim ayaṁ mērau dhūmō ’tha pralayānalēiti yaḥ śaṅkyatē śyāmaḥ pakṣīndrē ’rkatviṣi sthitaḥ
    [PIR 6.*3; T 6.6]

    A second case of “in doubt” that leads to the suggestion of another ornament, in this case, a pair of similes.

  105. ↑
    yasya vāṇī svavāṇīva svakriyēva kriyāmalārūpaṁ svam iva rūpaṁ ca lōkalōcanalōbhanam
    [PIR 6.*4; T 6]

    “Inimitability” (ananvaya).

  106. ↑
    tvatkr̥tē sō ’pi vaikuṇṭhaḥ śaśīvōṣasi candrikāmapy adhārāṁ sudhāvr̥ṣṭiṁ manyē tyajati tāṁ śriyam
    [PIR 6.*5; T 6.10]

    “Mixture” (saṁsr̥ṣṭi) of independent ornaments, in this case of “simile,” “identification,” and host of other figures.

  107. ↑
    tad uttiṣṭhātidhanyēna kēnāpi kamalēkṣaṇēvarēṇa saha tāruṇyaṁ nirviśantī gr̥hē vasa
    [PIR 6.*6; T 6.11]

    Another illustration of “mixture.”

  108. ↑
    karōṣi pīḍāṁ prītiṁ ca nirañjanavilōcanāmūrtyānayā samudvīkṣya nānābharaṇaśōbhayā
    [PIR 6.*7; T 6.13]

    This is Udbhaṭa’s illustration of “integrity” (bhāvika), where objects of the past or future (in this case Umā’s jewels and ornaments) are visible as if they were there, in front of one’s eyes.

  109. ↑
    chāyēyaṁ tava śēṣāṅgakāntēḥ kiñcid anujjvalāvibhūṣāghaṭanādēśān darśayantī dunōti mām
    [PIR 6.*8; T 6.15, reads vibhūṣāghaṭanōddēśān]

    This is an illustration of the “inferential sign of poetry” (kāvyahētu), which is here the scope for poetic inference by the speaker. Compare to KKS 5.48.

  110. ↑
    kiṁ cātra bahunōktēna vraja bhartāram āpnuhiudanvantam anāsādya mahānadyaḥ kim āsatē
    [PIR 6.*9; T 6.17, reads vātra]

    The ornament here is the “analogy of poetry” (kāvyadr̥ṣṭānta), which again applies a tool from logic to poetry, in this case the analogy. Compare to KKS 5.49.

  111. ↑One example are the verses based on "distinction" (vyatireka). Compare UKS 37–40 to KKS 5.27–29.

  112. ↑Tubb 1984: 229, 231, 225.

  113. ↑Thus, the key scene of Umā’s decision to perform tapas after taking permission from her parents and the loving description of her tapas by the narrator (KKS 5.1-29) have no parallel in Udbhaṭa’s poem as we have it.

  114. ↑As Tubb notes (1984: 229): “The commentators on the Kumārasambhava come close to recognizing this status of Pārvatī’s when they connect the description of her austerities with the heroic mood (vīra-rasa) while associating Śiva’s own austerities with the mood of peace (śānta-rasa).”

  115. ↑KKS 3.69 tells of the unrest of his senses (indriyakṣōbha) while meditating and of his success in restraining of them; 3.70–71, of his anger (manyu, krōdha) in response to Kāma’s attack; and 3.67 that his steadiness was slightly disturbed (kiñcitpariluptadhairya) when facing Umā.

  116. ↑For a discussion of this passage in Kālidāsa, see Handelman and Shulman 1997: 167.

  117. ↑Bronner 2016: 129–136.

  118. ↑Verse 23 also does not illustrate an ornament, but it supplies the seeing subject (Śiva), and hence sets the stage for the set of similes he sees in Umā in verse 24.

  119. ↑The following verse identifies this as Śiva’s thought.

  120. ↑I am grateful to Sheldon Pollock for a conversation on this verse.

  121. ↑In Kālidāsa’s poem, the disguised Brahmin, while a youth, is depicted far more impressively (KKS 5.30).

  122. ↑Note that in Kālidāsa’s intertext Himālaya is compared with Viṣṇu (KKS 6.67).

  123. ↑KKS 75. Narayana Daso Banhatti, comments on this verse with disapproval: “This is not a very good example of vyājastuti” he says, for Umā’s beauty “is really censurable if it hinders her union with a fit husband … and this kind of meaning indicating the reality of nindā [blame] lingers in our mind when we read the verse” (KKS p. 129 of the annotation). For more left-handed compliments of Udbhaṭa’s Śiva consider, for example, verses 69 (“Birth in a prosperous home, / dazzling beauty and enchanting youth, yet no happiness”) and 74 (“They decay then and there…”). For cases of outright criticism, see for instance verse 70 (“Seeing you behave/ so recklessly…”).

  124. ↑Franco 2018: 635. As Franco notes, there is a variant quotation: “desire and wealth [artha] are the only aims of life.”

  125. ↑Dhvanyāloka p. 539. By the same token, Kālidāsa’s Kumārasambhava comes under attack for its depiction of Śiva and Umā’s honeymoon in chapter eight, and although Ānandavardhana notes that this flaw of impropriety is concealed by Kālidāsa’s skill, it is clear that this was perceived as a problem perhaps already before Udbhaṭa’s time. It thus might be that Udbhaṭa found the idea of outdoing Kālidāsa in this domain appealing. I am grateful to Lawrence McCrea for suggesting this point to me.

  126. ↑The quote is from Tubb 2014: 73.

  127. ↑O’Flaherty 1973: 145. The text translated here is from the Kālikāpurāṇa 44.110–112:

    yōnijāṅ girijāṅ kālīn tapōvratavivarjitāmkathaṁ saṅgamakāmō ’han dhartum icchāmi vai haṭhāt
    tapōvratapavitrāṅgīn tapaścaraṇasatkr̥tāmsvayam ēva grahīṣyāmi satīn dakṣāyaṇīm iva
    katham iva kr̥takāmō ’ham anicchann iva sāmpratamkēnāpi cākr̥ṣṭa iva cikīrṣuḥ saṅgamōdbhavam

    For an overview of the story in the Purāṇas with an emphasis on Kāma, see Benton 2006: 39–63.

  128. ↑See, for instance, the versions of Matsyapurāṇa and Skandapurāṇa discussed in O’Flaherty 1973: 149.

  129. ↑O’Flaherty 1973: 147; Handelman and Shulman 1997: 57.

  130. ↑The quote is from Handelman and Shulman 1997: 164. For this duality, see also Benton 2006: 47.

  131. ↑O’Flaherty 1973: 148–151; Handelman and Shulman 1997: 169.

  132. ↑Tubb 1984: 222, 2014: 73.

  133. ↑Bronkhorst 2008: 293–294.

  134. ↑Ibid. 296. Franco (2018: 638) takes this to mean “cunning/fraudulent,” also referring to Udbhaṭa. For more on his highly innovative approach, see Solomon 1977–1978.

  135. ↑Bronkhorst 2008.

  136. ↑For an example of cancelled categories, see “twinning” (yamaka), the second on Bhāmaha’s list, and which Udbhaṭa simply removes (likely to be subsumed by punaruktavadābhāsa, now the first ornament on his list despite not being there in Bhāmaha; see Bronner 2016: 113–114). For an example of a newly coined ornament see his discussion of "fusion" (saṅkara) added. For a case where Udbhaṭa accepts an ornament that Bhāmaha rejects, see the case of kāvyahētu (in Udbhaṭa, whereas Bhāmaha famously rejected hētu). For more on the contrarian mode of Udbhaṭa’s presentation of ornaments when compared to Bhāmaha, See Bronner forthcoming.

  137. ↑Pratīhārēndurāja p. 25. Tilaka does exactly the same (p. 18).

AS Ruyyaka’s Alaṅkārasarvasva
KKS Kālidāsa’s Kumārasambhava
PIR Pratīhārēndurāja (see Banhatti 1982)
T Tilaka (see Ramaswami Sastri 1931)
UKS Udbhaṭa’s Kumārasambhava (quoted in the commentaries to his Kāvyālaṅkārasārasaṅgraha)
Table 1
Abbreviations

Primary Sources

  • Alaṅkārasarvasva of Ruyyaka, with the Sañjīvanī commentary of Vidyācakravartin. Edited by S. S. Janaki. Delhi: Meharchand Lachhmandas, 1965.
  • Kāvyālaṅkārasārasaṅgraha of Udbhaṭa: See Banhatti 1992 (with the Laghuvr̥tti of Pratīhārendurāja) and Ramaswami Sastri 1931 (with the Vivr̥ti of Tilaka).
  • Dhvanyāloka of Ānandavardhana, with commentaries by Abhinavagupta and Śrīrāmaśāraka. Ed. Pt. Pattābhirāma Śāstrī. Kashi Sanskrit Series, v. 135. Varanasi: Chowkhamba Sanskrit Series Office, 1940.
  • Kālikā Purāṇa. Bombay: 1891 (publisher not mentioned).
  • Kāvyaprakāśa of Mammaṭa with the Saṅketa commentary of Māṇikyacandra. Ed. Vāsudeva Śāstrī Abhayaṅkara. [Pune]: Ānandāśramamudrālaya, 1921.
  • Kumārasambhava of KālidāsA with the Sañjīvinī and Śiśuhitaiṣiṇī commentaries of Mallinātha and Sītārām Kavi. Edited Kanaklāl Ṭhakkur. Varanasi: Chaukhambha, 1987 (second edition).
  • Rājataraṅgiṇī of Kalhaṇa, 2 v. Ed. V. Bandhu. Hoshiarpur: Woolner Indological Series 8, 1963–1965.
  • Saṅketa (a.k.a. Mirror on [The Light on] Literature, Kāvyādarśa) of Someśvara = Rasiklal C. Parikh (ed.). Kāvyaprakāśa of Mammaṭa with the Saṁketa named Kāvyādarśa of Someśvara Bhaṭṭa (Son of Devaka of the Bhāradvāja Family). First Part – The Text. Jodhpur: Rajasthan Oriental Research Institute, 1959.
  • Subhāṣitaratnakoṣa of Vidyākara. Eds. D. D. Kosambi and V. V. Gokhale. Cambridge MASS: Harvard University Press, 1957.
  • Treatise on Theater ( Nāṭyaśāstra) of Bharata: The Nāṭyaśāstra of Bharatamuni. With the Commentary Abhinavabhāratī by Abhinavaguptācārya. 4 vols. Vadodara [Baroda]: Oriental Institute, 1992–2006. Vol. 1 (chaps. 1–7), 4th ed. by K. Krishnamoorthy, 1992. Vol. 2 (chaps. 8–18), 2nd ed,.by V. M. Kulkarni and T. Nandi, 2001. Vol. 3 (chaps. 19–27), 2nd ed. by V. M. Kulkarni and T. Nandi, 2003. Vol. 4 (chaps. 28–36), 2d ed. by V. M. Kulkarni and T. Nandi, 2006.

Secondary Sources

  • Banhatti, Narayana Daso (ed.). 1982. Kāvyālaṁkāra-sāra-saṁgraha of Udbhaṭa with the Commentary, the Laghuvṛitti of Indurāja. Poona: Bhandrakar Oriental Research Institute. 2nd ed. (1st ed. 1925).
  • Benton, Catherine. 2006. God of Desire: Tales of Kāmadeva in Sanskrit Story Literature. Albany, NY: SUNY.
  • Bronkhorst, Johannes. 2008. “Udbhaṭa, a Grammarian and Cārvāka.” In Linguistic Traditions of Kashmir: Essays in Memory of Paṇḍit Dinanath Yaksha, ed. M. Kaul and A. Aklujkar. New Delhi: D.K. Printworld, 281–299.
  • Bronner, Yigal. 2013. “From Conqueror to Connoisseur: Kalhaṇa’s Account of Jayāpīḍa and the Fashioning of Kashmir as the Kingdom of Learning.” Indian Economic and Social History Review 50 (2): 161–77. DOI 10.1177/0019464613487098.
  • Bronner, Yigal. 2016. “Understanding Udbhaṭa: The Invention of Kashmiri Poetics in the Jayāpīḍa Moment.” In Eli Franco and Isabelle Ratié (eds.) Around Abhinavagupta: Aspects of the Intellectual History of Kashmir from the Ninth to the Eleventh Century. Leipzig: Lit Verlag, 81–147.
  • Bronner, Yigal. Forthcoming. An Alaṅkāra Reader: Classical Indian Poetics..
  • Bronner Yigal and Andrew Ollett. 2024. “The First Hundred Light Years: The Wave of Twelfth-century Responses to the Kāvyaprakāśa.” Journal of the American Oriental Society 144 (4): 807–831. DOI 10.7817/jaos.144.4.2024.ar029.
  • Franco, Eli. 2018. “Lokāyata.” Brill’s Encyclopedia of Hinduism vol. 3, 629–642.
  • Handelman, Don and David Shulman. 1997. God Inside Out: Śiva’s Game of Dice. New York: Oxford University Press.
  • Heifetz, Hank. 1985. The Origin of the Young God: Kālidāsa’s Kumārasambhava. Berkeley: University of California Press.
  • Narayana Rao, Velcheru and David Shulman. 2002. Classical Telugu Poetry: An Anthology. New York: Oxford University Press.
  • O’Flaherty, Wendy Doniger. 1973. Asceticism and Eroticism in the Mythology of Śiva. London: Oxford University Press.
  • Pollock, Sheldon. 2016. A Rasa Reader: Classical Indian Aesthetics. New York: Columbia University Press.
  • Ramaswami Sastri Siromani, K. S. (ed.). 1931. Kāvyālaṅkārasārasaṅgraha of Udbhaṭa with the ‘Vivṛti.’ Baroda: Oriental Institute.
  • Shah, Nagin J. (ed.). 1972. Cakradhara’s Nyāyamañjarī-Granthibhaṅga. L. D. Series 35. Ahemdabad: L. D. Institute of Technology.
  • Solomon, E.A. 1977–1978. “Bhaṭṭa Udbhaṭa” Annals of the Bhandarkar Oriental Research Institute 58–59: 985–992.
  • Tubb, Gary. 1984. “Heroine as Hero: Pārvatī in the Kumārasaṃbhava and the Pārvatīpariṇaya.” Journal of the American Oriental Society 104: 219–36.
  • Tubb, Gary. 2014. “Baking Umā.” In Innovations and Turning Points: Toward a History of Kāvya Literature, ed. Y. Bronner, D. Shulman, and G. Tubb. New Delhi: Oxford University Press, 71–85.
  • Warder, A. K. 1983. Indian Kāvya Literature: Volume Four The Ways of Originality (Bāṇa to Dāmodaragupta). Delhi: Motilal Banarsidass. (Reprint, 1994)

Revision history

  • 2026-05-04T16:58:07Z (Yigal Bronner):