This chapter takes up the sixth class of roots, the tudādi, "the roots beginning with tud". Their class-sign (vikaraṇa) is śa, heard as a. It stands where the bhū class has śap. The difference is small but decisive: śa has no p as marker, so it counts as ṅit, and the root vowel takes no guṇa. So you get tudati, not todati. After this chapter you will be able to derive the present of any tudādi root. You will also see where the root goes back to its own rules, outside the present. You will meet bhrasj "to roast", with its optional augment ram (babharja or babhrajja). You will meet vij, whose iṭ-initial affixes are ṅit (udvijitā), and the optional ṣ of niḥṣphurati. The root mṛ "to die" takes the ātmanepada only in some lakāras (mriyate but mamāra). Then come the augment s of upaskirati, the optional l of gilati, and the n before śa in muñcati. Bhaṭṭoji walks the tudādi list of the Dhātupāṭha root by root, and teaches each rule where it first bites. The chapter has 9 sūtras, SK 2534 to SK 2542.
tud + śa + ti → tudati (no guṇa); bhrasj + śa + ti → bhṛsj-a-ti → bhṛjjatiThe first rule gives the class-sign. After tud and the roots listed after it, śa comes where śap would. Because śa has no p, SK 2234 makes it ṅit. Two things follow at once. The root vowel takes no guṇa: tud gives tudati. And roots that take saṃprasāraṇa before a ṅit affix take it here: bhrasj gives bhṛjjati. The first six roots of the list are svaritet: they take both sets of endings.
Rule. After tud and the roots that follow it, the affix śa comes, in its own sense, when a sārvadhātuka affix denoting the agent follows. It is an exception (apavāda) to śap. Bhaṭṭoji's commentators supply this setting.
The ś of śa is an it, so śa is śit and sārvadhātuka; only a is heard. It has no p, so by SK 2234 it is ṅit.
Example. The root tud "to pain".
tud + laṭ → tud + tip (endings as in chapter 44)
→ tud + śa + ti (this rule; śa instead of śap)
→ śa is apit, so ṅit (SK 2234) → no guṇa of the light penultimate u
→ tudati "he pains". With the ātmanepada: tudate.
Outside the present system śa is gone, and the root behaves normally:
tutoda, with guṇa. In thal, tutoditha: iṭ is obligatory here by the krādi restriction, since the root neither ends in a vowel nor has a. Ātmanepada tutude.tottā: the root takes no iṭ (it is aniṭ).atautsīt, with the vṛddhi taught for a consonant-final root; ātmanepada atutta.Applies when / blocked by. Why does guṇa not come first, before śa is added? Bhaṭṭoji's commentators answer that śa is nitya: it applies whether or not guṇa has applied. So it is made first, and then, being ṅit, it blocks guṇa. Scholars discuss a second ground for the same result.
Watch out. The present is tudati, never todati. That missing guṇa is the mark of the whole class.
The roots that follow. Bhaṭṭoji says six roots from tud on are svaritet. They are tud, ṇud, diś, bhrasj, kṣip and kṛṣ.
nudati, nudate, nunoda, nunude, nottā.deṣṭā, with ṣ for ś by SK 294 and then ṭ for t. Āśīrliṅ ātmanepada dikṣīṣṭa: the affix is kit by SK 2300, so there is no guṇa. luṅ adikṣat, adikṣata, with ksa by SK 2336.bhrasj + śa + ti
→ śa is ṅit, so SK 2412 gives saṃprasāraṇa: ṛ for the r, and the following a merges into it → bhṛsj + a + ti
→ the s becomes ś before the palatal j (ścutva) → bhṛśj + a + ti
→ the ś becomes j (jaśtva) → bhṛjjati "he roasts". Ātmanepada bhṛjjate.
Bhaṭṭoji's commentators close by setting up bhrasj + ṇal, the liṭ. That leads to the next rule.
bhrasj + ṇal → babharja / babhrajja; bhrasj + tās → bharṣṭā / bhraṣṭā; bhrasj + yāt → bhṛjjyātOutside the present, bhrasj has two shapes. Before an ārdhadhātuka affix, an augment ram may replace its r and penultimate s, giving bharj. Without it, the root stays bhrasj. So most forms come in pairs: babharja or babhrajja. But before a kit or ṅit affix, saṃprasāraṇa wins over ram: bhṛjjyāt. The group then runs on through kṣip and kṛṣ, with the many luṅ forms of kṛṣ.
Rule. In place of the r and the penultimate (upadhā) of bhrasj, the augment ram optionally comes, when an ārdhadhātuka affix follows. "Ārdhadhātuka" is in force as a heading from SK 2307.
The a of ram is only for pronunciation, and the m is an it. So ram adds an r. Being marked with m, it comes after the last vowel (SK 37). The sūtra names the r and the penultimate "in place of" (a sthānaṣaṣṭhī, a sixth case of place). So those two cease to be pronounced.
Example. liṭ, with ram:
bhrasj → ram after the last vowel → bhra + r + sj
→ the old r and the penultimate s cease → bharj
→ doubling for liṭ, then ṇal → babharja "he roasted".
Likewise babharjatuḥ, babharje (ātmanepada).
In thal, iṭ is optional by Bhāradvāja's restriction: babharjitha. Without iṭ, j becomes ṣ (SK 294) and th becomes ṭh: babharṣṭha.
liṭ, without ram:
bhrasj → doubling → babhrasj + a
→ s becomes ś, then ś becomes j → babhrajja.
Likewise babhrajjatuḥ, babhrajjitha, babhrajje.
Without iṭ: the s is lost by SK 380, and j becomes ṣ by SK 294 → babhraṣṭha.
luṭ: bhraṣṭā (without ram) or bharṣṭā (with ram). lṛṭ: bhrakṣyati or bharkṣyati, with k by SK 295.
Applies when / blocked by. Why no saṃprasāraṇa in babharjatuḥ, though atus is usually kit? Bhaṭṭoji's commentators explain: after a root ending in a consonant cluster, a liṭ affix is not kit. So SK 2412 does not apply.
How can an augment take the "place" of two sounds? One of Bhaṭṭoji's commentators reads "place" as "occasion": where r and s would be pronounced, the r after the a is used instead. Scholars discuss this further.
A vārttika then settles a conflict.
Vārttika kṅiti ramāgamaṃ bādhitvā saṃprasāraṇaṃ pūrvavipratiṣedhena. Before a kit or ṅit affix, saṃprasāraṇa sets aside the augment ram. This is pūrvavipratiṣedha: in this conflict the earlier rule wins. Without the vārttika, ram, the later rule, would win and give the wrong bharjyāt. Bhaṭṭoji's commentators say so.
bhrasj + yāt (āśīrliṅ; the augment yāsuṭ is kit)
→ saṃprasāraṇa by SK 2412, not ram → bhṛsj + yāt
→ ś, then j → bhṛjjyāt "may he roast". Likewise bhṛjjyāstām.
In the ātmanepada āśīrliṅ the affix is not kit. So there is no saṃprasāraṇa, and ram is free: bharkṣīṣṭa or bhrakṣīṣṭa.
luṅ: abhārkṣīt or abhrākṣīt. Ātmanepada abharṣṭa or abhraṣṭa; the s of sic is lost by SK 2281.
Watch out. ram comes only before an ārdhadhātuka affix. It never touches bhṛjjati, and before a kit or ṅit affix it yields to saṃprasāraṇa.
The roots that follow.
kṣipati, kṣipate, kṣeptā, luṅ akṣaipsīt, akṣipta. Bhaṭṭoji's commentators add that by the krādi restriction iṭ is still obligatory in thal: cikṣepitha.kṛṣati, kṛṣate.kraṣṭā or karṣṭā. The augment am is optional by SK 2402. With am the root is kraṣ; without it, guṇa ar gives karṣ.kṛṣyāt and kṛkṣīṣṭa take neither am nor guṇa, since the affix is kit (the second by SK 2300).akrākṣīt (with am) or akārkṣīt (without, with vṛddhi). With ksa: akṛkṣat.akṛṣṭa, akṛkṣātām, akṛkṣata.akṛkṣata (singular), akṛkṣātām, akṛkṣanta. So akṛkṣata is both the sic plural and the ksa singular.ṛṣati, liṭ ānarṣa.juṣate.udvijate. In udvijitā the guṇa of the light penultimate would be due. The next rule stops it.ud-vij + iṭ + tā → udvijitā (not *udvejitā)One short sūtra makes the iṭ-initial affixes of vij ṅit, so its light penultimate i takes no guṇa. Bhaṭṭoji then walks a long stretch of the root list, about a hundred roots. Most entries give only a meaning and a present form. Where he gives more, the forms teach a handful of recurring lessons. Saṃprasāraṇa comes before śa and in the liṭ abhyāsa (vṛścati, vivyāca). Some roots keep their n before śa (tṛmphati). Some take iṭ optionally (lobhitā or lobdhā). And the kuṭādi roots begin, whose affixes are ṅit (cukuṭitha).
Rule. An affix beginning with iṭ that follows vij is treated as ṅit (ṅidvat). "ṅit" is carried forward (anuvṛtti) from SK 2461.
How does the bare word iṭ come to mean "an affix beginning with iṭ"? Bhaṭṭoji's commentators agree on the sense, but scholars discuss how the sūtra yields it. Some later grammarians, as one commentator reports, point out that the augment iṭ by itself being ṅit would achieve nothing.
Example.
ud + vij + tā (the luṭ ending, as in chapter 44) → iṭ is added → ud + vij + i + tā
→ the affix begins with iṭ, so it is ṅit (this rule) → no guṇa of the penultimate i
→ udvijitā "he will fear". Likewise lṛṭ udvijiṣyate. Bhaṭṭoji's commentators add udvijiṣīṣṭa and udavijiṣṭa.
Applies when / blocked by. One of Bhaṭṭoji's commentators says the vij meant is ovijī of this class (and of the rudhādi). It is not the vij of the juhotyādi class. So that root keeps guṇa: vivejitha.
Watch out. Not udvejitā. The guṇa is blocked only because the affix begins with iṭ.
The roots that follow: laj, lasj. laj and lasj (taught olajī, olasjī) mean "to be ashamed". They close the four ātmanepadin roots. laj gives lajate, liṭ leje; lasj gives lajjate, liṭ lalajje.
Saṃprasāraṇa before śa and in the abhyāsa: vraśc. From vraśc on, the roots are parasmaipadin. vraśc (taught ovraścū) means "to cut".
vraśc + śa + ti → śa is ṅit, so SK 2412 gives ṛ for the r → vṛścati "he cuts".
liṭ singular:
vraśc + ṇal → doubling → abhyāsa vra-
→ saṃprasāraṇa in the abhyāsa by SK 2408: ṛ for the r
→ that ṛ becomes a by urat (SK 2244)
→ that a is treated as the ṛ it replaced, by SK 50
→ so the v takes no saṃprasāraṇa: SK 363 forbids a saṃprasāraṇa before a saṃprasāraṇa
→ vavraśca "he cut".
Bhaṭṭoji's commentators ask why the r takes saṃprasāraṇa first, and not the v. One answers that the prohibition of SK 363 itself shows it (it is a jñāpaka). Scholars discuss this route.
vavraścatuḥ: the affix follows a cluster, so it is not kit, and there is no saṃprasāraṇa.vavraścitha or vavraṣṭha. Without iṭ, the s is lost by SK 380 and the c becomes ṣ by SK 294.vraścitā or vraṣṭā; lṛṭ vraściṣyati or vrakṣyati.vṛścyāt: the affix is kit, so saṃprasāraṇa returns.avraścīt (with iṭ) or avrākṣīt (without). In the second, the s is lost by SK 380, ṣ comes by SK 294, and k by SK 295.vyac. vyac "to feign".
vicati: SK 2412 before śa.vivyāca: the y of the abhyāsa takes saṃprasāraṇa, not the v, by SK 363.vivicatuḥ: the affix is kit, so SK 2412 applies before doubling.vyacitā, vyaciṣyati; āśīrliṅ vicyāt.avyācīt or avyacīt: vṛddhi is optional by SK 2284.Vārttika vyaceḥ kuṭāditvam anasi: "vyac counts as a kuṭādi root, except before as." Bhaṭṭoji says this does not operate here. anasi "not before as" is an exclusion of the kind called paryudāsa. Because of that, the vārttika covers kṛt affixes only. That is why vyacitā has no saṃprasāraṇa. One of Bhaṭṭoji's commentators gives udvicitum as a kṛt form where it does come.
Roots read twice: uch, ucch. uch "to glean" gives uñchati; ucch "to end" gives ucchati. Both are also read in the bhū class. Bhaṭṭoji's commentators say they are read here to get śa. Then the feminine of the present participle gets an optional n by SK 445: uñchatī or uñchantī. The bhvādi reading serves the accent. (The same holds for kaḍa "to be intoxicated", later in the run.)
Guṇa and nuṭ: ṛcch. ṛcch means "to go", "to fail in the senses", "to become solid".
ānarccha, ānarcchatuḥ.ṛcchitā has no guṇa, because SK 2383 is limited to liṭ.Bhaṭṭoji also gives the liṭ ānarca from ṛc "to praise", and later ānarpha from ṛph "to harm".
Other liṭ forms. micch "to afflict": mimiccha, luṅ amicchīt. tvac "to cover": tatvāca.
Optional iṭ.
lubhati, luṭ lobhitā or lobdhā (iṭ optional by SK 2340), lobhiṣyati.cartitā; lṛṭ cartiṣyati or cartsyati (optional by SK 2506); acartīt.tṛhati, tatarha; stṛhati, tastarha; luṭ starhitā or starḍhā. luṅ atṛṃhīt or atārṅkṣīt, dual atārṇḍhām. In atārṅkṣīt the h becomes ḍh and then k, and the anusvāra becomes ṅ (Bhaṭṭoji's commentators).icchati. luṭ eṣitā or eṣṭā; eṣiṣyati; iṣyāt; aiṣīt.miṣati, meṣitā.sic obligatory: tṛp. tṛp "to be satisfied": tṛpati, tatarpa, tarpitā. The vārttika spṛśa-mṛśa… makes sic optional, but only for the tṛp of the puṣādi group (a group of divādi roots that take aṅ, SK 2343). The vārttika is an exception to aṅ, so it works only where aṅ would come. Here sic is obligatory: atarpīt.
num restored before śa: tṛmph.
tṛmph + śa + ti → śa is ṅit, so SK 415 deletes the n → tṛph + a + ti …
Vārttika śe tṛmphādīnāṃ num vācyaḥ: before śa, tṛmph and the rest get num. So the n returns: tṛmphati.
"And the rest" here means "of that kind": the roots in this stretch that have an n. So tumpati, tumphati, dṛmphati, ṛmphati, gumphati, umbhati, śumbhati. Outside śa the n stays deleted: tatṛmpha, tṛphyāt.
The roots of this kind that begin with a vowel and are heavy take ām in liṭ (SK 2237): ṛmphāṃcakāra, umbhāṃcakāra.
Variant readings. Bhaṭṭoji records several.
riphati, rirepha. Some read rih, as in the line śiśuṃ na viprā matibhī rihanti.thuḍati, tuthoḍa, tusthoḍa): some read khuḍ, chuḍ.Lengthening in āśīrliṅ, and its bar. sur "to rule, to shine": surati, suṣora, āśīrliṅ sūryāt. kur "to sound": kurati, kūryāt. The vowel is lengthened by SK 354, hali ca.
SK 1629, na bhakurchurām, forbids that lengthening for kur and chur. Yet it does not touch this kur. Bhaṭṭoji reports the view that the kur there means kṛ "to do" only. One of his commentators admits that only the commentators' explanation settles this. The bar does reach chur "to cut": churyāt, not lengthened.
likh. likh "to write": lilekha. One of Bhaṭṭoji's commentators says it is not a kuṭādi root. So lekhitā stands, and likhitum is an error.
The kuṭādi begin: kuṭ. From kuṭ "to be crooked" the kuṭādi roots begin. By SK 2461 their affixes, unless ñit or ṇit, are ṅit. So there is no guṇa.
cukuṭitha (liṭ, thal); luṭ kuṭitā.cukoṭa or cukuṭa. ṇal is ṇit, so it escapes SK 2461 and guṇa comes; in the first person it is optionally not ṇit, and then guṇa is blocked.cakarḍa, kṛḍitā; tuṭ "to quarrel": tuṭati, tutoṭa, tuṭitā.truṭyati or truṭati; liṭ tutroṭa.sphuṭati, pusphoṭa.The run closes with sphur and sphul, which lead into the next rule.
nis + sphurati → niḥṣphurati / niḥsphuratiAfter the prefixes nir, ni and vi, the s of sphur and sphul may become ṣ. Then Bhaṭṭoji finishes the kuṭādi roots, with gu (jugutha, gutā) and kuṅ, and marks their end. He closes with pṛ "to exert oneself", whose ṅit present gives vyāpriyate.
Rule. After nir, ni and vi, a retroflex (mūrdhanya) optionally replaces the s of sphur and sphul. The retroflex for s is ṣ. "Retroflex" is in force as a heading from SK 210. "Optionally" is carried forward from SK 2359, sivādīnāṃ vā. Bhaṭṭoji's commentators say so.
Example.
nis + sphurati (the prefix nir, split as nis in the sūtra, appears as niḥ before the s) → s becomes ṣ optionally (this rule)
→ niḥṣphurati or niḥsphurati.
Likewise with sphul, and after ni and vi.
Watch out. The ṣ is optional, so both forms are correct. And the rule covers only these two roots, after only these three prefixes.
Some read the root as sphar, with a as its penultimate. Its liṭ is pasphāra.
The roots that follow. sphuḍ, cuḍ, vruḍ mean "to cover"; kruḍ and bhṛḍ mean "to sink", according to some.
gur "to exert" (taught gurī) has an anudātta it, so it is ātmanepadin: gurate, jugure, guritā.
ṇū "to praise" ends in a long vowel. Bhaṭṭoji cites pariṇūtaguṇodayaḥ "one whose rise of virtues has been praised". There is no iṭ in pariṇūta, by SK 2381. Its ṇ is from the upadeśa.
From here four roots are parasmaipadin.
nuvati (with uvaṅ), luṅ anuvīt.dhuvati.juguvitha or jugutha; luṭ gutā; lṛṭ guṣyati.aguṣīt; dual agutām, where the s of sic is lost by SK 2369; plural aguṣuḥ.dhruvati and the rest, like guvati. dhruva takes iṭ: dudhruvitha, dhruvitā, dhruviṣyati, dhrūvyāt, adhruvīt, adhruviṣṭām.kuṅ "to sound". Kaiyaṭa and others say it ends in a long vowel, so it takes iṭ: kuvitā, akuviṣṭa. The author of the Nyāsa says it ends in a short vowel: kutā, akuta.
Then vṛt: the kuṭādi roots end here.
pṛ "to exert oneself" (taught pṛṅ) is ṅit, so ātmanepadin. It is generally used with vi and āṅ.
vi + ā + pṛ + śa + te → riṅ: ri for ṛ (SK 2367) → iyaṅ: iy for i (SK 271)
→ vyāpriyate "he exerts himself".
liṭ vyāpapre, vyāpaprāte. lṛṭ vyāpariṣyate, with iṭ by SK 2366. luṅ vyāpṛta, vyāpṛṣātām.
Then comes mṛ "to die" (taught mṛṅ), which leads into the next rule.
mṛ + śa + te → mri-y-a-te → mriyate; but mamāra, mariṣyati; and mṛṣīṣṭa, amṛtamṛ is marked with ṅ, so it should take the ātmanepada everywhere. The sūtra restricts that. mṛ takes the ātmanepada only in luṅ, in liṅ, and before a śit affix, such as śa. Everywhere else it takes the parasmaipada. Then come seven parasmaipadin roots, ending with kṝ "to scatter".
Rule. After mṛ, when it is the base of luṅ, of liṅ, or of a śit affix, the ātmanepada comes; not elsewhere. "Ātmanepada" is carried forward from SK 2157. The word ca "and" draws in "śit" from SK 2362, śadeḥ śitaḥ.
Bhaṭṭoji writes taṅ, meaning the ātmanepada as a whole. So the participle affix āna is included: mriyamāṇa.
The ṅ-marking alone would give the ātmanepada everywhere. So this sūtra is a restriction (niyama). Why then is mṛ marked with ṅ at all? Bhaṭṭoji answers: for the accent. One of his commentators cites SK 3730 and the form mā hi mṛta.
Example.
mṛ + laṭ → śa is śit, so the ātmanepada (this rule) → mṛ + śa + te
→ riṅ: ri for ṛ (SK 2367) → mri + a + te
→ iyaṅ: iy for i (SK 271) → mriyate "he dies".
Elsewhere, the parasmaipada:
mamāra. thal mamartha, with no iṭ even on Bhāradvāja's view. mamriva, with iṭ by the krādi restriction.martāsi. lṛṭ mariṣyati, with iṭ by SK 2366.In liṅ and luṅ, the ātmanepada again: āśīrliṅ mṛṣīṣṭa, luṅ amṛta.
Watch out. mriyate and mamāra are both correct. The same root changes its set of endings with the lakāra.
The roots that follow. Seven parasmaipadin roots.
riyati, piyati.retā, petā.suvati, luṭ savitā.kirati, kirataḥ: i for ṝ by SK 2390, with r after it.cakāra, and cakaratuḥ: guṇa although the affix is kit, by SK 2383.karitā or karītā: the iṭ is optionally long, by SK 2391.kīryāt; luṅ akārīt.upa + kirati → upa-s-kirati → upaskirati; upa + a-kirat → upāskiratTwo sūtras add an s before the k of kṝ. After upa, it comes in the sense of cutting. After prati, and again after upa, it comes in the sense of injuring. A vārttika lets the s stand before the k even when the augment a of the past or the reduplicated syllable comes between: upāskirat, upacaskāra.
Rule. In the sense of cutting (lavana), the augment suṭ is added to kirati (the root kṝ) after upa. It is placed before the k. "After upa" is carried forward from SK 2552, and "suṭ before k" from SK 2553. Both are taught later in the Kaumudī. Bhaṭṭoji's commentators say so.
suṭ is an augment marked with ṭ; only s is heard.
Example.
upa + kirati (sense: cutting)
→ suṭ before the k (this rule) → upa + s + kirati
→ upaskirati "he cuts".
Applies when / blocked by. In any other sense there is no s: upakirati "he scatters". One of Bhaṭṭoji's commentators gives this counter-example.
Vārttika aḍabhyāsavyavāye 'pi suṭ kāt pūrva iti vaktavyam. It should be stated that suṭ comes before the k even when aṭ or the abhyāsa intervenes (vyavāya). aṭ is the augment a of laṅ and luṅ.
laṅ: upa + a + kirat → suṭ before the k, across the a → upāskirat.
liṭ: upa + ca + kāra → suṭ before the k of the root, across the abhyāsa → upacaskāra.
Rule. In the sense of injuring (hiṃsā), suṭ is added to kirati after prati, and, by ca "and", after upa. Bhaṭṭoji's own gloss supplies kirati and suṭ. Bhaṭṭoji's commentators read ca as bringing in upa.
Example.
prati + kirati (sense: injuring) → suṭ before the k → pratiskirati "he injures".
upa + kirati (sense: injuring) → upaskirati.
One of Bhaṭṭoji's commentators cites the poet Māgha: uro vidāraṃ praticaskare nakhaiḥ. There praticaskare is liṭ in the object sense, with guṇa by SK 2383.
The root that follows. gṝ "to swallow". One of Bhaṭṭoji's commentators glosses it as "to eat". It leads into the next rule.
gṝ + śa + ti → gir-a-ti → girati / gilatiThe r of gṝ may become l before a vowel-initial affix: girati or gilati. Then come dṛ "to respect" and dhṛ "to remain", two more ṅit roots with riṅ and iyaṅ. Then pracch, the last of the five kirādi roots. Sixteen parasmaipadin roots follow, among them sṛj, masj and spṛś. The group ends at muc.
Rule. Before a vowel-initial affix, l optionally replaces the r of gṝ. "Of gṝ" is carried forward from SK 2639, gro yaṅi, taught later. "l for r" is carried forward from SK 2350, kṛpo ro laḥ.
The word aci "before a vowel" describes the affix. So the rule looks at an affix that begins with a vowel (a tadādividhi). One of Bhaṭṭoji's commentators says it therefore does not reach a noun form like girau. There the vowel au is a case ending, not an affix added after the root.
Example.
gṝ + śa + ti → ir for ṝ (SK 2390) → gir + a + ti
→ the r stands before the vowel-initial affix a: l optionally (this rule)
→ girati or gilati "he swallows".
liṭ jagāra or jagāla; thal jagaritha or jagalitha.
luṭ garitā, garītā, galitā, galītā. The l is optional, and the iṭ is optionally long by SK 2391.
Watch out. The l is optional everywhere the condition holds. So one luṭ has four correct forms.
The roots that follow: dṛ, dhṛ. dṛ "to respect" (taught dṛṅ) is ṅit and takes no iṭ. Bhaṭṭoji gives its forms with ā (āṅ):
ādriyate, ādriyete: riṅ and iyaṅ, as in mriyate.ādadre, ādadriṣe; luṭ ādartā; lṛṭ ādariṣyate, with iṭ by SK 2366.ādṛṣīṣṭa: the affix is kit by SK 2368, so there is no guṇa.ādṛta, ādṛṣātām. In ādṛta the s of sic is lost by SK 2369.dhṛ "to remain" (taught dhṛṅ): dhriyate.
pracch. Sixteen parasmaipadin roots follow. The first is pracch "to wish to know".
pṛcchati: śa is ṅit, so SK 2412 gives ṛ for the r.papraccha, papracchatuḥ. The affix follows a cluster, so it is not kit, and there is no saṃprasāraṇa.papracchitha or papraṣṭha. iṭ is optional by Bhāradvāja's restriction. Without it, ch becomes ṣ by SK 294 and th becomes ṭh.praṣṭā; lṛṭ prakṣyati; luṅ aprākṣīt.Then vṛt: the kirādi roots end here. They are five: kṝ, gṝ, dṛ, dhṛ and pracch. Another rule names them as five (kiraś ca pañcabhyaḥ); Bhaṭṭoji's commentators say the number serves the rule after it.
sṛj. sṛj "to emit" takes no iṭ.
sasarjitha or sasraṣṭha: iṭ is optional by SK 2404.sraṣṭā, lṛṭ srakṣyati, luṅ asrākṣīt. The a after the r is the augment am, by SK 2405.sṛjet, āśīrliṅ sṛjyāt.masj. masj "to cleanse" (taught ṭumasjo).
majjati: the s becomes ś, then j. liṭ mamajja.Vārttika masjer antyāt pūrvo num vācyaḥ. The num of masj stands before its final sound. Bhaṭṭoji then gives the loss of the s at the head of the cluster (saṃyogādilopa, SK 380).
mamasj + tha (liṭ, thal, no iṭ)
→ the s is lost (SK 380) → mamaj + tha
→ j becomes g (kutva) → num before the final → mamang + tha
→ the n becomes anusvāra, then ṅ to match the g (parasavarṇa) → g becomes k (cartva)
→ mamaṅktha. With iṭ: mamajjitha.
Likewise luṭ maṅktā, lṛṭ maṅkṣyati, luṅ amāṅkṣīt, amāṅktām, amāṅkṣuḥ.
Applies when / blocked by. mamaṅktha would come out the same even if num stood after the vowel. Bhaṭṭoji's commentators give the vārttika's real use: in magna the n must be penultimate, so that it can be deleted.
ruj to spṛś.
roktā, rokṣyati, araukṣīt, arauktām. bhuj "to bend" goes like ruj.choptā, acchaupsīt.roṣṭā, rokṣyati; reṣṭā, rekṣyati.alikṣat, with ksa.spraṣṭā or sparṣṭā, the augment am being optional by SK 2402; lṛṭ sprakṣyati or sparkṣyati. luṅ asprākṣīt, aspārkṣīt (with sic, by the vārttika spṛśa-mṛśa…) or aspṛkṣat (with ksa).vicch. vicch "to go" takes the affix āya by SK 2303, and in the ārdhadhātuka it does so optionally.
vicchāyati.vicchāyāṃcakāra (with āya) or viviccha (without).viś, mṛś. viś "to enter": viśati, veṣṭā. mṛś "to touch": luṅ amrākṣīt or amārkṣīt (with sic, am optional), or amṛkṣat (with ksa).
Roots read again. Several roots of this stretch were read before. Each second reading has a purpose.
sīdati and the rest, as in the bhū class. The reading here gives śa, and with it an optional n in the feminine present participle: sīdatī or sīdantī. Its reading among the jvalādi roots is for the affix ṇa: sādaḥ. The two readings also differ in accent: śap is anudātta, but śa is udātta.milati, milate; mimela, mimile.The next root, muc "to release" (taught mucḷ), leads into the last sūtra.
muc + śa + ti → mu-n-c-a-ti → muñcatiThe last sūtra gives an n to muc and the roots after it, before śa: muñcati, lumpati, vindati, limpati, siñcati, kṛntati, khindati, piṃśati. These are the mucādi. With them the tudādi class ends.
Rule. When śa follows, the augment num is added to muc and the roots after it. "num" is carried forward from SK 2262, idito num dhātoḥ.
num is marked with m, so it comes after the last vowel (SK 37). Only n is heard.
Example.
muc + śa + ti → num after the vowel (this rule) → mu + n + c + a + ti
→ the n becomes anusvāra, then ñ to match the palatal c (parasavarṇa, as in mamaṅktha)
→ muñcati "he releases". Ātmanepada muñcate.
Outside śa there is no n:
moktā; āśīrliṅ mucyāt.mukṣīṣṭa: the affix is kit by SK 2300, so there is no guṇa.amucat: aṅ, because the root is marked with ḷ (SK 2343).amukta, amukṣātām. In amukta the s of sic is lost by SK 2281.Watch out. The n belongs to the present system only. muñcati, but moktā and amucat.
The roots that follow.
lumpati, lumpate; luṅ alupat, alupta.vindati, vindate; liṭ viveda, vivide.veditā. On the view of the Bhāṣya and others it does not: vettā.parivettā: pari means passing over. A parivettā is one who, passing over his eldest brother, has taken wives and the sacred fires. The affix is tṛn or tṛc.limpati, limpate, leptā. luṅ takes aṅ by SK 2418: alipat. In the ātmanepada aṅ is optional (SK 2419): alipata or alipta.siñcati, siñcate.asicat; ātmanepada asicata or asikta.abhiṣiñcati: ṣ after the prefix, by SK 2270.abhyaṣiñcat: ṣ even with aṭ between, by SK 2276.abhiṣiṣeca: ṣ by SK 2277.Then three parasmaipadin roots.
kṛntati, cakarta, kartitā; lṛṭ kartiṣyati or kartsyati (iṭ optional by SK 2506); luṅ akartīt.khindati, cikheda, khettā. In the sense of dejection it is read in the divādi and rudhādi classes.piṃśati, peśitā. It is also used for "to light up": tvaṣṭā rūpāṇi piṃśatu "may Tvaṣṭṛ light up the forms".Then vṛt: the mucādi roots end, and with them the tudādi roots. Here end the tudādi roots.
| SK | what it does | example |
|---|---|---|
| 2534 | śa for śap after tud and the rest; śa is apit, so ṅit, so no guṇa | tud + śa + ti → tudati; bhrasj → bhṛjjati |
| 2535 | bhrasj: augment ram optionally for its r and s, before an ārdhadhātuka | babharja / babhrajja; bharṣṭā / bhraṣṭā |
| vārttika on 2535 | before a kit or ṅit affix, saṃprasāraṇa beats ram | bhṛjjyāt |
| 2536 | an iṭ-initial affix after vij is ṅit | udvijitā |
| vārttika on 2536 | vyac is kuṭādi before kṛt affixes only | vyacitā (no saṃprasāraṇa) |
| vārttika on 2536 | num before śa for tṛmph and the like | tṛmphati, but tṛphyāt |
| 2537 | optional ṣ for the s of sphur, sphul after nir, ni, vi | niḥṣphurati / niḥsphurati |
| 2538 | mṛ takes the ātmanepada only in luṅ, liṅ and before a śit affix | mriyate, mṛṣīṣṭa, amṛta; but mamāra |
| 2539 | suṭ before the k of kṝ after upa, in the sense of cutting | upaskirati |
| vārttika on 2539 | suṭ even across aṭ or the abhyāsa | upāskirat, upacaskāra |
| 2540 | suṭ after prati and upa, in the sense of injuring | pratiskirati |
| 2541 | optional l for the r of gṝ before a vowel-initial affix | girati / gilati |
| vārttika on 2541 | the num of masj stands before its final | mamaṅktha |
| 2542 | num before śa for muc and the rest | muñcati, lumpati, siñcati |