What only her husband knows, and two crowns of ruby

The first name is put together oddly, and the oddity is the point. Jñāta is a past participle, known; saubhāgya and mārdava, grace and softness, are what is known; and Kāmeśa, in compound, is the one who knows them. The tradition reads the name two ways at once and does not choose. On the plain reading it is modesty: this is the part of the body a husband sees, and the hymn says so instead of describing it. On the other reading jñāta is doing technical work — what is known here is known by direct acquaintance rather than by report, and Kāmeśa is not a censor but the only qualified witness. Both senses are in the word and neither cancels the other.

Śaṅkara makes the same claim about exclusivity at a different point on the body and states it as a comparison between goddesses. How many and how many poets do not enjoy Brahmā's wife — Sarasvatī is taken by anyone who writes. Who fails to become the husband of Śrī, given a little money? But contact with this goddess is not to be had by anybody except Mahādeva, not even by the kuravaka tree; and the kuravaka is the tree that, by long convention, flowers when a beautiful woman embraces it, so it is the one thing in the poetic world that gets to touch every woman there is. Denying it access is the strongest form the claim can take.

Then the knees, and here the parallel is exact rather than thematic. Śaṅkara devotes one verse to thighs and knees together and gives both their comparisons. Her two thighs have defeated the trunks of lordly elephants and the golden stems of the plantain — the two stock similes for a smooth tapering limb, disposed of in half a line. And her two knees, well-rounded and hardened by bowing to her husband, are the pair of frontal globes on the elephant of the gods.

Praṇatikaṭhina, hardened by prostration, is the detail to stop on. The knees of a goddess whom Brahmā, Viṣṇu and Śiva are all said to worship are calloused because she kneels. The first verse of the Saundaryalaharī says that without her Śiva cannot so much as stir; the eighty-second says her knees are hard from bowing to him. The tradition holds the two together without any sense of strain, and so does this hymn, which will seat her on Śiva's lap at name fifty-two and then, two names later, call her the woman whose beloved is entirely under her own control.

A textual note, since the manuscripts disagree here. The third line of that Saundaryalaharī verse is transmitted in two forms: the Śṛṅgeri text reads vidhijñe, a vocative meaning you who know the rite, while other printed editions read vijigye, you have conquered, which makes the knees the subject of a second verb. The picture is the same either way. It is worth knowing that this poem, like the sahasranāma, has been copied for a very long time and shows it.

The ruby is the fourth red stone in the hymn and the third distinct word for one. Māṇikya is the ordinary word and the first meditation verse used it for her crown; kuruvinda was the stone strung in her diadem at name fourteen; padmarāga was her cheek at name twenty-one. What is new here is the shape. The knee is not compared to a ruby but to a crown made of ruby, so an ornament that sat on top of her head at the beginning of the description turns up again at the knee in miniature. Monier-Williams defines mukuṭa as a tiara or diadem and adds that it is said to be crescent-shaped, distinguishing it from the pointed kirīṭa and the three-pointed mauli. A crescent is exactly the outline of a kneecap seen straight on.

What the two names do together is worth naming, because it is a structural problem solved rather than an image invented. The head-to-foot form obliges the poet to keep going. Between the waist and the ankle there is a stretch of the body that a hymn addressed to a mother cannot describe in the manner of erotic poetry and cannot skip either. This verse is the way out, and it is a good one: one name that says the information exists and is not being handed over, and one name that returns to plain visual description as soon as it decently can.

Two sixteen-syllable names, both with the caesura at a clean joint: kāmeśajñātasaubhāgya, then mārdavorudvayānvitā; māṇikyamukuṭākāra, then jānudvayavirājitā.

The conjunct ज्ञ in kāmeśajñāta is the same one the pilot drilled at mahārājñī in the first verse, and this is the shape in which you will meet it most often. It is a palatal j followed by the palatal nasal ñ, both made with the tongue-body flat against the hard palate. Reciters trained in the north approximate it as something like gy and reciters trained in the south as something like jñ; the grammarians describe the latter. You will need it constantly from here on — jñāna, prajñā, vijñāna, sarvajñā all take it.

Saubhāgya is a vṛddhi formation and so is mārdava, and having two of them side by side makes the process visible. Subhaga, fortunate, strengthens its u to au and becomes saubhāgya, good fortune. Mṛdu, soft, strengthens its syllabic ṛ to ār and becomes mārdava, softness. The last verse gave you kusumbha becoming kausumbha by the same rule; the fourth verse gave you sugandha becoming saugandhika. Collect them, because once vṛddhi is audible, half the abstract nouns in Sanskrit stop looking arbitrary.

Mārdava plus ūru gives mārdavoru: a short a and a long ū have merged into a single o, written मार्दवोरु with one ो and no trace of the join. This is the same merger that produced tapanoḍupa in the eighth verse and it has the same consequence — ūru will not be found in the dictionary in the shape it has here, and you have to undo the sandhi before you can look it up. Note also that ūru begins with a long ū, which English speakers shorten.

Dvaya occurs twice in the verse, at ūrudvaya and jānudvaya, and the dv is a dental d followed by a labiodental v with both audible. English tends to drop one or the other.

The second line is where the tongue goes back. Māṇikya has the retroflex ṇ, mukuṭa has the retroflex ṭ, and then the line comes forward again for jānu, dvaya and virājitā, all of which are dental or palatal. Curl back twice and return, and do not let the ṇ of māṇikya soften into the n of jānu three syllables later — they are the same drill the pilot set at sindūrāruṇa and trinayanāṃ.

Mukuṭa plus ākāra gives mukuṭākāra, one more of the invisible a-plus-ā mergers that this hymn is built out of.

The verse ends on virājitā, the same past participle that ended the first name of the twelfth verse — cibukaśrīvirājitā there, jānudvayavirājitā here. Five verses apart, the identical closing word, and it is a good marker to have in the ear: whenever a line ends -virājitā, something has been made resplendent by something else, and the something else is named just before it.

For the archanā, kāmeśajñātasaubhāgyamārdavorudvayānvitāyai namaḥ and māṇikyamukuṭākārajānudvayavirājitāyai namaḥ.

On what to hold: the verse gives a real instruction about the limits of looking, and it is the more effective for being unmarked. The first name tells you that there is something here you are not going to be told. It does not apologize for it, hint at it, or invite you to imagine it. Most descriptive poetry works by promising more than it delivers; this one works by declining, and then goes straight on to the knees, which is exactly what the reciter should do.