This chapter takes the fourth class of roots, the divādi, "div and the rest". It runs from root 1107 divu "to play" to root 1246 gṛdhu "to covet". These roots take the class-sign śyan, heard as ya, in place of śap: dīvyati "he plays", nṛtyati "he dances", jāyate "he is born". After this chapter you will be able to build the sārvadhātuka forms of a divādi root with śyan, and say why the root vowel takes no guṇa before it. You will know four rules that act on the root in its śyan forms: lengthening by hali ca (dīvyati), loss of o (śyati), jā (jāyate; before any śit affix), and the long vowel of the śamādi (śāmyati). You will also know how three sub-groups of the class behave: the puṣādi take aṅ in the aorist (apuṣat); the radhādi take iṭ optionally (rarandhitha / raraddha); the śamādi lengthen their vowel (śāmyati). As in chapter 44, Bhaṭṭoji walks the root list and teaches each rule where it first bites. The rules therefore come from all over the Aṣṭādhyāyī. The chapter has 18 sūtras, SK 2505 to SK 2522.
div + śap + ti → div + śyan + ti → div + ya + ti → dīv + ya + ti → dīvyatiThis group has one rule. After the roots of the div class, the class-sign is śyan, not śap. Because śyan counts as ṅit, the root vowel takes no guṇa. The running example is div "to play" in the present: dīvyati "he plays".
Rule. After div and the other roots of its class, the affix śyan comes before a sārvadhātuka affix in the sense of the agent. The ś and the n of śyan are it-markers, so it is heard as ya.
Two words are carried forward (anuvṛtti). "In the sense of the agent" comes from kartari śap (SK 2167). "Before a sārvadhātuka" comes from sārvadhātuke yak (SK 2756). Bhaṭṭoji calls śyan an apavāda, an exception, to śap.
Example.
div "to play", laṭ, third person singular: div + śap + ti
→ śyan replaces śap (this rule) → div + śyan + ti
→ ś and n are it-markers → div + ya + ti
→ śyan is not p-marked, so it counts as ṅit (SK 2234): no guṇa of the i
→ the i is lengthened before the consonant by hali ca (SK 354) → dīvyati.
The other sārvadhātuka forms are built the same way: loṭ dīvyatu, laṅ adīvyat, vidhiliṅ dīvyet. The āśīrliṅ is dīvyāt.
Applies when / blocked by. Śyan comes only before a sārvadhātuka affix. In liṭ, luṭ, lṛṭ, luṅ and lṛṅ the root stands without śyan and shows guṇa and iṭ: liṭ dideva, luṭ devitā, lṛṭ deviṣyati, luṅ adevīt, lṛṅ adeviṣyat. One of Bhaṭṭoji's commentators notes that devitā shows the root is seṭ.
Bhaṭṭoji opens the class with a note on the endings. The roots "ending with jhṝṣ" are parasmaipadin. One of Bhaṭṭoji's commentators fixes the range: roots 1107 to 1131.
Watch out. Do not write devyati. The e would be guṇa, and śyan, being ṅit, blocks it. The long ī of dīvyati is not guṇa either: it comes from hali ca.
The root list after the rule gives several points.
devitvā / dyūtvā, dyūtam. Bhaṭṭoji's commentators both say so.pariṣīvyati, liṭ pariṣiṣeva. The ṣ after pari comes by SK 2275. The restriction of SK 2277 does not reach this case, Bhaṭṭoji's commentators explain. SK 2277 carries the words "before sita". Sita is a word in SK 2275, so SK 2277 reaches only what is taught before that word, and SK 2275 itself is outside its reach. So the ṣ of the abhyāsa (the first copy of the doubled root) also comes by SK 2275. Luṅ: nyaṣevīt or nyasevīt, optionally, by SK 2359.snusyati, suṣṇosa. Ṣṇasu "to spit": snasyati, sasnāsa. Knasu "to be crooked, to shine": caknāsa. Vyuṣa "to burn": vuvyoṣa. Pluṣa "to burn".nṛtyati, liṭ nanarta. Its ī-marker denies iṭ in the past participle (SK 3039). The next rule starts from this root.nṛt + sya + ti → nart + i + ṣya + ti → nartiṣyati / nart + sya + ti → nartsyati; but luṅ anartīt (iṭ invariable)This group has one rule. Five roots, nṛt among them, take iṭ only optionally before an ārdhadhātuka affix that begins with s. The aorist affix sic is left out of the rule. The running example is nṛt "to dance" in the future: nartiṣyati or nartsyati "he will dance", but always anartīt "he danced".
Rule. After kṛt, cṛt, chṛd, tṛd and nṛt, an ārdhadhātuka affix that begins with s optionally takes iṭ. The affix sic is excepted.
"Iṭ" and "of an ārdhadhātuka" are carried forward from SK 2184. "Optionally" (vā) comes from udito vā (SK 3328). One of Bhaṭṭoji's commentators explains that the rule makes optional an iṭ that would otherwise be obligatory.
Example.
nṛt, lṛṭ: nṛt + sya + ti, with the root vowel in guṇa: nart + sya + ti
→ with iṭ (this rule allows it) → nart + i + sya + ti → s becomes ṣ after i → nartiṣyati
→ without iṭ → nartsyati.
Vidhiliṅ nṛtyet, āśīrliṅ nṛtyāt.
Applies when / blocked by. The rule says "not sic" (asici). So before sic, iṭ stays obligatory. Bhaṭṭoji's commentators both point to this: luṅ anartīt, always with iṭ, because the rule excepts sic.
One of Bhaṭṭoji's commentators adds more forms from the five roots: ninartiṣati / ninṛtsati, and kartiṣyati / kartsyati from kṛt.
The root list after the rule gives several points.
trasyati or trasati. Liṭ: tresatuḥ or tatrasatuḥ. The e with loss of the abhyāsa is optional by SK 2356.kṣipyati, kṣeptā. Puṣpa "to blossom": puṣpyati, pupuṣpa. Tima, ṣṭima, ṣṭīma "to be wet": timyati, stimyati, stīmyati. Vrīḍa "to impel, to be ashamed": vrīḍyati. Iṣa "to go": iṣyati. Ṣaha and ṣuha, "in the sense of cak": sahyati, suhyati. "The sense of cak is satisfaction."jīryati: the ṝ becomes i (SK 2390), and hali ca lengthens it. Liṭ jajaratuḥ or jeratuḥ: guṇa by SK 2383, then the option of SK 2356. Luṭ jaritā or jarītā, by SK 2391. Vidhiliṅ jīryet, āśīrliṅ jīryāt. Luṅ: aṅ is optional by SK 2291, with guṇa by SK 2406: ajarat; otherwise ajāriṣṭām. From jhṝṣ: jhīryati, jajharatuḥ, ajhārīt.sūyate, liṭ suṣuve. In liṭ the second person is suṣuviṣe, with iṭ. SK 2279 would make iṭ optional here, and it is the later rule. But the prohibition of SK 2381 sets that option aside, because the section of prohibitions was begun earlier. Bhaṭṭoji's commentators both explain it so. Then the krādi restriction (SK 2293) makes iṭ obligatory: suṣuviṣe (second person singular), suṣuvivahe (first person dual). Luṭ sotā or savitā.dūyate. Dīṅ "to waste away": dīyate. The next rule starts from dīṅ.dī-dī + e → didī + yuṭ + e → didīye; dī + tā → (e due) dā + tā → dātā; lī + tā → letā / lātāThis group has three rules about roots in ī and i. After dīṅ "to waste away", a vowel-initial kit or ṅit affix takes the augment yuṭ: didīye. Then mī, mi and dīṅ take ā where an affix would give them e: dātā. For lī this ā is optional: letā or lātā.
Rule. After dīṅ, an ārdhadhātuka affix that begins with a vowel (ac) and is kit or ṅit receives the augment yuṭ. The ṭ of yuṭ is an it-marker and the u is for pronunciation, so it is heard as y. Being ṭ-marked, it goes at the front of the affix (SK 36).
"Ārdhadhātuka" is the heading of SK 2307. One of Bhaṭṭoji's commentators explains that "vowel" qualifies it: "an ārdhadhātuka that begins with a vowel".
Example.
dīṅ, liṭ, third person singular: the ending ta becomes e (SK 2241); it is kit (SK 2242)
→ the root is doubled → didī + e
→ yaṇ (ī → y) by SK 272 is due, and it is the later rule
→ but yuṭ applies first, because it is nitya: it applies whether or not the other has applied
→ didī + y + e → didīye.
Applies when / blocked by. A vārttika, a supplementary statement on a sūtra, follows the rule: vugyuṭāv uvaṅyaṇoḥ siddhau vaktavyau, "vuk and yuṭ should be stated to be valid with respect to uvaṅ and yaṇ". Bhaṭṭoji's commentators both explain why it is needed. Yuṭ falls in the ābhīya zone, so it would count as not yet done (asiddha). Then yaṇ would still apply, and give *\didyye, with the ī lost and two *y*s heard. The vārttika makes yuṭ count as done, so didīye stands.
One of Bhaṭṭoji's commentators gives counter-examples. Not vowel-initial: dedīyate. Not kit or ṅit: upadānam.
Rule. Ā replaces the vowel of mī (the root of mīnāti), mi (the root of minoti) and dīṅ before the affix lyap, the absolutive affix after a prefixed root. By the word ca "and", it also replaces it in the domain of an affix that is not ś-marked (aśit) and that would bring about an ec (e, o, ai, au).
Ā is carried forward from SK 2370. One of Bhaṭṭoji's commentators explains that ca gathers in two more words from SK 2370, "ec" and "not śit". Bhaṭṭoji's commentators construe the case-ending of "ec-causing" differently, "before" or "in the domain of", and scholars discuss this.
Example.
dīṅ, luṭ: dī + tā
→ tā would bring guṇa, e, which is an ec; so this is the domain of an ec-causing affix
→ ā replaces the ī (this rule) → dā + tā → dātā.
Lṛṭ dāsyate.
Applies when / blocked by. Luṅ adāsta, adāsthāḥ. Once dīṅ has ā, one of Bhaṭṭoji's commentators counts it as ghu (the name of dā and dhā). Then sthāghvor ic ca (SK 2389) would give *\adita, but a vārttika in the Bhāṣya bars that rule for dīṅ. Another commentator reports a view that dīṅ is never ghu at all; scholars discuss this.
One of Bhaṭṭoji's commentators gives forms with lyap: pramāya (from mī) and upadāya (from mi).
The root list after the rule gives several points.
ḍīyate, liṭ ḍiḍye. Dhīṅ "to hold": dhīyate, didhye, dhetā.mīyate. "Injury here is separation from life."rīyate. Līṅ "to cling": the next rule starts from it.Rule. For lī and līṅ, ā optionally replaces the vowel, in the domain of an ec-causing affix and before lyap. Ā is carried forward from SK 2370, and the other two conditions from SK 2508.
The root is cited as līyati, with the affix yak (SK 2756), not with śyan. Bhaṭṭoji says so in his first sentence. Yak is the passive class-sign, and both roots take it. So the rule covers the divādi līṅ and also lī "to cling" of the ninth class, which takes śnā. One of Bhaṭṭoji's commentators adds: had śyan been meant, the sūtra would have said just līṅaḥ.
Example.
lī, luṭ: lī + tā → tā would bring guṇa, e
→ with ā (this rule) → lā + tā → lātā
→ without it, guṇa → letā.
Lṛṭ leṣyate or lāsyate. With lyap, one of Bhaṭṭoji's commentators gives vilāya or vilīya.
Applies when / blocked by. Why "in the domain of an ec"? Where no affix brings e, there is no ā: līyate, liṭ lilye.
The root list after the rule gives several points.
vrīyate, vivriye.prasūnam, ḍīnaḥ.pīyate. Māṅ "to measure": māyate, liṭ mame. One of Bhaṭṭoji's commentators notes that SK 2462 does not give ī in māyate, since that rule is limited to ārdhadhātuka affixes.īyate, liṭ ayāṃcakre. Prīṅ "to please": prīyate, pipriye. Bhaṭṭoji calls it transitive (sakarmaka).śo + śyan + ti → ś + ya + ti → śyati; outside śyan śo → śā: śaśau, śātāThis group has one rule. Four roots end in o. Before śyan that o is deleted: śyati "he sharpens". Elsewhere SK 2370 turns the o into ā: śaśau, śātā.
Rule. The final o of a root is deleted (lopa) before śyan. "Deletion" is carried forward from ghor lopo leṭi vā (SK 3584).
Example.
śo, laṭ: śo + śyan + ti → śo + ya + ti
→ the o is deleted before śyan (this rule) → ś + ya + ti → śyati.
Dual śyataḥ, plural śyanti.
Applies when / blocked by. The rule needs śyan. Outside the śit domain, ā replaces the o by SK 2370. In liṭ the ending ṇal becomes au after ā (SK 2371):
liṭ śaśau, śaśatuḥ; luṭ śātā; lṛṭ śāsyati.
In luṅ there are two series. By SK 2376, sic is optionally deleted by luk: aśāt, aśātām, aśuḥ. In aśuḥ the ending is us (SK 2227). Where sic is kept, SK 2377 gives iṭ and the augment sak: aśāsīt, aśāsiṣṭām.
The root list after the rule gives several points.
chyati, by this rule.syati, liṭ sasau. With abhi: abhiṣyati (ṣ by SK 2270), abhyaṣyat (ṣ by SK 2276). But abhisasau, with no ṣ, by the restriction of SK 2277.dyati, liṭ dadau. praṇidātā has ṇ by SK 2285. Āśīrliṅ deyāt, with e by SK 2374. Luṅ adāt, with sic deleted by SK 2223.jan + śyan + te → jā + ya + te → jāyate; a-jan + ciṇ + ta → a-jan-i-(ta) → ajani (not *ajāni); pra-ni-a-pad + ciṇ + ta → praṇyapādiThis group has three rules, met as Bhaṭṭoji walks the ātmanepada roots. Jan "to be born" becomes jā before śyan: jāyate. In the aorist with ciṇ, jan keeps a short vowel: ajani, not *\ajāni. And after pad "to go", ciṇ is obligatory before the ending ta: praṇyapādi. The block ends with the roots that are parasmaipadin to the end of the class.
Rule. Jā replaces jñā "to know" and jan "to be born" before a śit affix. "Before a śit affix" is carried forward from SK 2320.
Example.
jan, laṭ: jan + śyan + te → jan + ya + te
→ śyan is śit, so jā replaces jan (this rule) → jā + ya + te → jāyate "he is born".
For jñā, which takes śnā: jānāti "he knows".
Applies when / blocked by. Outside the śit domain there is no jā. One of Bhaṭṭoji's commentators gives jñātā for jñā. For jan: liṭ jajñe, jajñāte, jajñire; luṭ janitā; lṛṭ janiṣyate. In jajñe, one of Bhaṭṭoji's commentators explains, the penultimate a is deleted by SK 2363, and the n becomes ñ after j.
In luṅ, ciṇ replaces cli optionally, by SK 2328. With a-jan-i-ta standing, vṛddhi of the penultimate would then be due. The next rule stops it.
Watch out. Why is the substitute long jā, not short ja? A short ja would be lengthened before ya by SK 2170 anyway. Bhaṭṭoji's commentators both treat the long jā as a jñāpaka, a sign in the wording. It shows a principle: once an operation on the stem (aṅga) has been done, another stem operation is not done on top of it.
Rule. For jan and vadh, vṛddhi of the penultimate does not take place before ciṇ, or before a kṛt affix (one that makes a noun from a root) marked with ñ or ṇ.
Five words are carried forward, one of Bhaṭṭoji's commentators explains. "Of the penultimate" comes from SK 2282, "vṛddhi" from SK 2473, and "not" from SK 2763. "Before ciṇ or a kṛt" comes from SK 2761, and "marked with ñ or ṇ" from SK 254.
Example.
jan, luṅ, third person singular: a + jan + cli + ta
→ ciṇ replaces cli (SK 2328) → a-jan-i-ta
→ vṛddhi of the penultimate a would be due (SK 2282); this rule blocks it
→ ta is deleted after ciṇ (SK 2329) → ajani "he was born".
With sic instead of ciṇ: ajaniṣṭa.
One of Bhaṭṭoji's commentators gives the kṛt forms: janaḥ with ghañ, and janakaḥ, with a ṇ-marked affix.
The root list after the rule gives several points.
dīpyate, didīpe, luṅ adīpi or adīpiṣṭa, by the option of SK 2328. Pūrī "to swell": pūryate, apūri or apūriṣṭa. Tūrī, dhūrī, gūrī, ghūrī, jūrī, śūrī and cūrī follow: tūryate, dhūryate, dudhūre, gūryate, jugūre. One of Bhaṭṭoji's commentators notes that these are ī-marked, so the past participle has no iṭ: dīptaḥ, pūrṇaḥ.tapyate "he is master", taptā, tapsyate. Otherwise it takes śap and is parasmaipadin: tapati. Some take the vā as the start of the next root, reading it vāvṛtu. A variant reads the root pata, with t and p swapped; the Vedic word patyamānaḥ "being master" supports it.vṛtyate; on the other reading, vāvṛtyate. Bhaṭṭoji quotes a line of the poet Bhaṭṭi with vāvṛtyamānā "seeking".kliśyate, kliśitā. Kāśṛ "to shine": kāśyate. Vāśṛ "to sound": vāśyate, vavāśe.mṛṣyati, mṛṣyate, mamarṣa, mamṛṣe. Īśucir "to become moist": śucyati, śucyate, śuśoca, śuśuce. Its luṅ is aśucat, with aṅ because it is irit, or aśocīt; ātmanepada aśociṣṭa.nahyati, nahyate, liṭ nanāha. Second person liṭ: nanaddha or nehitha, since iṭ in thal, the second person singular ending of liṭ, is optional by Bhāradvāja's view (SK 2296). Without iṭ, h becomes dh by SK 440. With iṭ, e and loss of the abhyāsa come by SK 2261. Also nehe, naddhā, natsyati, anātsīt.rajyati, rajyate, with the n deleted by SK 415. Śapa "to curse": śapyati, śapyate.padyate, pede, pattā, padyeta, patsīṣṭa. The next rule is about its luṅ.Rule. After pad, ciṇ replaces cli when the ending ta follows. Cli is carried forward from SK 2222. Bhaṭṭoji's commentators both say that ta here means the ātmanepada third person singular.
Example.
pad with pra and ni, luṅ, third person singular: pra-ni + a + pad + cli + ta
→ ciṇ replaces cli (this rule) → pra-ni-a-pad-i-ta
→ vṛddhi of the penultimate a before the ṇ-marked ciṇ (SK 2282) → pād
→ ta is deleted after ciṇ (SK 2329)
→ n becomes ṇ after pra by SK 2285, and i + a gives ya → praṇyapādi.
Applies when / blocked by. Only before ta. In the dual and plural, sic comes: apatsātām, apatsata. One of Bhaṭṭoji's commentators gives the second person singular, apatthāḥ, as another counter-example.
Watch out. The rule is not optional. Contrast SK 2328, where ciṇ is only optional for jan, budh and others.
The root list after the rule gives several points.
khidyate, cikhide, khettā, akhitta. Vida "to exist": vidyate, vettā, without iṭ.budhyate, bubudhe, boddhā, bhotsyate, bhutsīṣṭa. Luṅ abodhi or abuddha, by the option of SK 2328; dual abhutsātām.yudhyate, yuyudhe, yoddhā, ayuddha. How then is there a parasmaipada yudhyati? Bhaṭṭoji gives two answers. It may be from the noun yudh "battle", with the affix kyac, "desires battle". Or, by a paribhāṣā (a rule of interpretation), ātmanepada due to an anudātta it-marker is not invariable. One of his commentators rejects the second answer, and scholars discuss this.anurudhyate. One of Bhaṭṭoji's commentators notes that, read in this class, it takes śyan instead of śnam.aṇyate, āṇe, aṇitā. Some say the root ends in a dental: ana. Mana "to know": manyate, mene, mantā.yujyate, yoktā.sasṛjiṣe, with iṭ; sraṣṭā, srakṣyate. In āśīrliṅ and luṅ the affixes are kit by SK 2300. So there is no guṇa, and no augment am either, since SK 2405 gives am only before an affix that is not kit: sṛkṣīṣṭa, asṛṣṭa, asṛkṣātām.liśyate, leṣṭā, lekṣyate, likṣīṣṭa. Luṅ alikṣata, alikṣātām, with ksa by SK 2336.aparādhyati "offends", virādhyantam "one who offends", rādhyaty odanaḥ "the rice gets done", and kṛṣṇāya rādhyati "examines Kṛṣṇa's fortune". In the last, the fortune is included in the root's meaning, so the root is intransitive, like jīvati.rarādha, rarādhatuḥ, rarādhitha. SK 2532 gives e and loss of the abhyāsa to rādh "to injure". It does not act here, because rādh "to injure" is transitive and so cannot belong to this class. Also rāddhā, rātsyati. The root is also in the fifth and tenth classes.vidhyati, with saṃprasāraṇa, y → i, by SK 2412 before the ṅit śyan. Liṭ vivyādha, vividhatuḥ, vivyaddha or vivyadhitha; vyaddhā, vyatsyati, vidhyet, vidhyāt, luṅ avyātsīt.puṣyati, pupoṣa, pupoṣitha, poṣṭā, pokṣyati. Luṅ apuṣat, with aṅ by SK 2343. Puṣ heads the puṣādi, the roots from puṣ to the end of this class, which take aṅ in the parasmaipada luṅ.aśuṣat. Tuṣa "to be pleased". Duṣa "to be spoiled". Śliṣa "to embrace": śliṣyati, śiśleṣa, śleṣṭā, ślekṣyati. Its luṅ is the subject of the next rule.a-śliṣ + ksa + t → a-śliṣ-sa-t → aślikṣat (ṣ → k before s, s → ṣ) "embraced"; not embracing: sam-a-śliṣ + aṅ + t → samaśliṣatThis group has one rule, read as two. Śliṣ is a puṣādi root, so its parasmaipada luṅ would take aṅ. This rule gives it ksa instead, but only in the sense of embracing: aślikṣat kanyāṃ devadattaḥ "Devadatta embraced the girl". Where the root means "stick together", aṅ stays: samaśliṣaj jatu kāṣṭham "the lac and the wood stuck together".
Rule. Bhaṭṭoji splits the sūtra into two rules (yogavibhāga).
1. Śliṣaḥ: after śliṣ, when the root is aniṭ, ksa replaces cli.
2. Āliṅgane: after śliṣ, ksa replaces cli only in the sense of embracing, not elsewhere.
"Cli" is carried forward from SK 2222. "When aniṭ" and "ksa" come from SK 2336. One of Bhaṭṭoji's commentators explains why "aniṭ" is needed. It keeps out the seṭ root śliṣu "to burn" of the first class, whose luṅ is aśleṣīt.
Example.
śliṣ "to embrace", luṅ, third person singular: a + śliṣ + cli + t
→ the sense is embracing, so ksa replaces cli (this rule) → a + śliṣ + sa + t
→ ṣ becomes k before s, and the s becomes ṣ → aślikṣat.
aślikṣat kanyāṃ devadattaḥ "Devadatta embraced the girl".
Applies when / blocked by. SK 2336 would already give ksa to śliṣ. Why teach it again? One of Bhaṭṭoji's commentators answers: to set aside the puṣādi aṅ (SK 2343). Bhaṭṭoji says the same: this ksa is an exception to the puṣādi aṅ, not to ciṇ. He grounds this on the purastād apavāda maxim, a maxim on how far an exception reaches.
The second rule is a restriction (niyama). Bhaṭṭoji says that, by the force of the split, it binds the ksa of SK 2336 too. His commentators both explain why: if only the first rule's ksa were limited, the split would serve no purpose.
So, where the sense is not embracing, there is no ksa:
sam + a + śliṣ + aṅ + t → samaśliṣaj jatu kāṣṭham "the lac and the wood stuck together".
Here śliṣ means contact, and the puṣādi aṅ comes. One of Bhaṭṭoji's commentators reads samāśliṣat, with the prefix ā; scholars discuss this.
In the passive, in a sense other than embracing, only sic comes, not ksa. One of Bhaṭṭoji's commentators explains: the puṣādi aṅ is for the parasmaipada only, and the passive takes ātmanepada. The singular takes ciṇ: aśleṣi. The other forms take sic:
aślikṣātām — here sic and ksa would give the same form.aślikṣata — with sic, jha becomes at.aśliṣṭhāḥ, aśliḍhvam — with ksa these would differ. Bhaṭṭoji's commentators write the last one aśliḍḍhvam.Watch out. The root must be aniṭ, and the sense must be embracing. The aniṭ condition is carried into the first part from SK 2336. The embracing condition is the second part.
The root list after the rule gives several points.
śakyati, śakyate hariṃ draṣṭuṃ bhaktaḥ "the devotee is able to see Hari". Liṭ śaśāka, śekitha or śaśaktha, śeke. Also śaktā, śakṣyati, śakṣyate, luṅ aśakat, aśakta. "Some say this root is seṭ": śakitā, śakiṣyati. The verses on aniṭ roots read śaki with the marker ḷ, and that follows their view.svidyati, siṣveda, siṣveditha, svettā, asvidat. The author of the Nyāsa and others make it ñī-marked; Haradatta and others do not.kroddhā, krotsyati. Kṣudha "to be hungry": kṣoddhā. How then is there kṣudhita "hungry", with iṭ? Mādhava derives it from the noun kṣudh with the affix itac. "In truth", says Bhaṭṭoji, iṭ comes by SK 3046, taught later.śudhyati, śuśodha, śoddhā. Ṣidhu "to be accomplished": the reading with ū as marker is an error. sidhyati, seddhā, setsyati, asidhat.radhyati, liṭ rarandhatuḥ, with the augment num (an n after the root vowel) by SK 2302. The next rule starts from radh.ra-radh + tha → rarandhitha / raraddha; radh + i + tā → radhitā (no num); na-naś + tha → na-naṃś + tha → nanaṃṣṭha; pra + naṃṣ + tā → pranaṃṣṭā (no ṇ)This group has four rules. Eight roots, the radhādi, take iṭ only optionally: rarandhitha or raraddha. Then two rules govern the augment num, an n added after the root vowel (written ṃ before ś and ṣ). Radh refuses it before an iṭ outside liṭ: radhitā. Naś "to disappear, to perish" takes it before a consonant of the jhal class: nanaṃṣṭha. Last, the n of naś does not become ṇ when naś ends in ṣ: pranaṃṣṭā.
Rule. After radh and the rest, an ārdhadhātuka affix beginning with val (any consonant but y) takes iṭ optionally. The eight roots are radh, naś, tṛp, dṛp, druh, muh, ṣṇuh and ṣṇih. A root with optional iṭ is called veṭ.
"Iṭ for a val-initial ārdhadhātuka" is carried forward from SK 2184. "Optionally" comes from SK 2279. One of Bhaṭṭoji's commentators notes that iṭ would otherwise be obligatory.
Example.
radh, liṭ, second person singular: ra-radh + tha
→ with iṭ → ra-radh + i + tha → before the vowel, num by SK 2302 → rarandhitha
→ without iṭ → raraddha.
First person dual: rarandhiva or redhva.
Applies when / blocked by. One of Bhaṭṭoji's commentators adds a point about the intensive. Because the eight roots are named as a group, the intensive with luk of yaṅ has only rāradhitā.
In luṭ, with iṭ, num by SK 2302 would be due before the vowel. The next rule stops it.
Rule. Radh does not take the augment num before an iṭ that is not part of liṭ. "Num" is carried forward from SK 2262.
Example.
radh, luṭ, with iṭ: radh + i + tā
→ num would come before the vowel (SK 2302); this rule blocks it
→ radhitā.
Without iṭ: raddhā. Lṛṭ: radhiṣyati or ratsyati.
Applies when / blocked by. The iṭ of liṭ is excepted, so there num comes: rarandhiva. One of Bhaṭṭoji's commentators explains why the n then stays. The affix follows a conjunct, so it is not kit (SK 2242), and the n is not deleted. He also gives randhakaḥ: there is no iṭ there, so num comes.
Before aṅ, num comes by SK 2302, and then its n is deleted by SK 415: luṅ aradhat.
The root list goes on with ṇaśa "to disappear, to perish": naśyati, liṭ nanāśa, neśatuḥ. Second person neśitha, with iṭ (optional by SK 2515), and e with loss of the abhyāsa by SK 2261. Without iṭ, nanaś + tha stands; the next rule takes it up.
Rule. Masj "to sink" and naś "to disappear, to perish" take the augment num before an affix beginning with a jhal. "Num" is carried forward from SK 2262.
Example.
naś, liṭ, second person singular, without iṭ: na-naś + tha
→ tha begins with a jhal, so num comes (this rule) → na-naṃś + tha (the n of num is written ṃ before ś)
→ ś becomes ṣ, and th becomes ṭh after it → nanaṃṣṭha.
Our text and the printed kaumudī read nanaṣṭa, which has neither the num nor the ṭh. One of Bhaṭṭoji's commentators and other texts read nanaṃṣṭha, the form that shows the rule; scholars discuss this.
Applies when / blocked by. Where iṭ comes, the affix no longer begins with a jhal, and there is no num. Nor is there num before v or m, which are not jhal. First person dual neśiva or neśva; plural neśima or neśma. Luṭ naśitā or naṃṣṭā; lṛṭ naśiṣyati or naṅkṣyati. Vidhiliṅ naśyet, āśīrliṅ naśyāt, luṅ anaśat.
One of Bhaṭṭoji's commentators adds a point on naṃṣṭā. The velar substitute of the rule naśer vā is taught only at the end of a word. So it does not act here.
With pra, the n becomes ṇ: praṇaśyati. The next rule says when it does not.
Rule. The n of naś does not become ṇ when naś ends in ṣ. Bhaṭṭoji's gloss, "ṇatva does not take place", spells out two carried-forward words. "Ṇ for n" comes from SK 235, and "not" from SK 2840.
Example.
naś with pra, luṭ, without iṭ: pra + naś + tā
→ num (SK 2517) → pra + naṃś + tā → ś becomes ṣ → pra + naṃṣ + tā
→ the root now ends in ṣ, so the n does not become ṇ (this rule)
→ pranaṃṣṭā.
Applies when / blocked by. Why "ending in ṣ"? Where the root ends in ś, ṇ comes: praṇaśyati.
Why the word anta "ending in"? It brings in a root that was ṣ-final (bhūtapūrva). In pranaṅkṣyati the ṣ has become k, yet the n stays. One of Bhaṭṭoji's commentators adds that anta would otherwise be superfluous, so it is a jñāpaka. With iṭ the root keeps its ś: naśiṣyati.
Watch out. Compare pranaṃṣṭā with praṇaśyati. The same root after the same prefix takes ṇ in one and not the other.
The root list after the rule gives several points.
nāgnis tṛpyati kāṣṭhānām "fire is not satisfied with logs", and Bhaṭṭi's pitṝn atārpsīt "he satisfied the ancestors".tatarpitha with iṭ, as a radhādi root; tatraptha with the augment am by SK 2402; tatarptha without either.tarpitā, tarptā, traptā.atārpsīt (with vṛddhi, SK 2267), atrāpsīt (with am), atarpīt (with iṭ). Otherwise the puṣādi aṅ: atṛpat.dṛpyati. Both tṛp and dṛp are veṭ as radhādi roots. Their listing among the aniṭ roots, Bhaṭṭoji says, is for the sake of am.dudrogdha, dudroḍha, dudrohitha. Luṭ drohitā, drogdhā, droḍhā. Lṛṭ drohiṣyati, dhrokṣyati; the form is the same with gh or with ḍh. Luṅ adruhat.muhyati, mumohitha, mumogdha, mumoḍha, mogdhā, moḍhā, mohitā, mohiṣyati, mokṣyati, amuhat.snuhyati, suṣṇoha, suṣṇohitha, suṣṇogdha, suṣṇoḍha, suṣṇuhiva, suṣṇuhva, snohitā, snogdhā, snoḍhā, snohiṣyati, snokṣyati, asnuhat.snihyati, siṣṇeha. Then "vṛt": the radhādi end here. "The puṣādi, however, run to the end of the class: this is the settled view (siddhānta)."pra-ni + śam + śyan + ti → pra-ni + śām + ya + ti → praṇiśāmyatiThis group has one rule. Eight roots, the śamādi, lengthen their vowel before śyan: śāmyati "he becomes calm", mādyati "he rejoices". Outside śyan the vowel stays short: śamitā.
Rule. The vowel of the eight roots śam and the rest is lengthened before śyan. The plural śamām means "of śam and the rest". Bhaṭṭoji's commentators list the eight: śam, tam, dam, śram, bhram, kṣam, klam and mad.
Example.
śam with pra and ni, laṭ: pra-ni + śam + śyan + ti
→ the a is lengthened before śyan (this rule) → pra-ni + śām + ya + ti
→ śāmyati is named in SK 2285, so the n of ni becomes ṇ → praṇiśāmyati.
Applies when / blocked by. Only before śyan. Elsewhere the vowel is short: liṭ śematuḥ, śemitha; luṭ śamitā; luṅ aśamat, with the puṣādi aṅ.
The root list after the rule gives several points.
tāmyati, tamitā, atamat.dāmyati means "he calms (someone)"; luṅ adamat.śrāmyati, aśramat.bhrāmyati. Luṅ with aṅ: abhramat. "The rest is as in the first class."kṣāmyati. As an ūdit root it is veṭ: cakṣamitha or cakṣantha; cakṣamiva or cakṣaṇva; cakṣamima or cakṣaṇma; kṣamitā or kṣantā. In cakṣaṇva the m becomes n by SK 2309, and then ṇ.kṣāntiḥ comes from the divādi root, kṣamā from kṣamūṣ.klāmyati or klāmati. SK 2320 already lengthens its vowel before śyan, just as before śap. So it is read among the śamādi for another purpose: the affix ghinuṇ (SK 3121). Luṅ with aṅ: aklamat.mādyati, amadat. "The eight, śam and the rest, are finished."asyati, āsa, asitā. The next rule is about its luṅ.Watch out. Bhram and klam also stand in SK 2321, so śyan is optional for them. For bhram, the long vowel of bhrāmyati comes from this rule, and only once śyan has come. For klam, SK 2320 gives the long vowel before śyan and śap alike: klāmyati, klāmati.
as + aṅ + t → as-thuk + a + t → (āṭ) āsthat; yas + śyan/śap + ti → yasyati / yasati; sam + yas → saṃyasyati / saṃyasati; but pra + yas → prayasyati onlyThis group has three rules and closes the class. As "to throw" takes the augment thuk before aṅ: āsthat "he threw". Then yas "to strive" takes śyan only optionally, without a prefix or after sam: yasyati or yasati. After any other prefix, śyan is obligatory: prayasyati.
Rule. As "to throw" receives the augment thuk when aṅ follows. "When aṅ follows" is carried forward from SK 2406. The k of thuk is an it-marker and the u is for pronunciation. Being k-marked, it goes at the end of the root (SK 36).
Example.
as, luṅ, third person singular: as + cli + t → aṅ replaces cli, as as is a puṣādi root → as + a + t
→ thuk is added to the root (this rule) → as-th + a + t
→ with the augment ā of a vowel-initial root → āsthat.
Applies when / blocked by. As already gets aṅ as a puṣādi root. Why then is it named in SK 2438, which also gives aṅ? Bhaṭṭoji answers: for aṅ with ātmanepada endings, where the puṣādi aṅ cannot reach. Those endings come optionally after a prefix, by the vārttika upasargād asyatyūhyor vā, given under SK 2701. So with pari: paryāsthata.
The next root is yasu "to strive".
Rule. After yas, when no prefix (upasarga) precedes it, śyan comes optionally. "Śyan" is carried forward from SK 2505, and "optionally" from SK 2321. Bhaṭṭoji writes no gloss of his own for this rule; he glosses it together with the next.
Example.
yas + ti, no prefix → with śyan → yasyati
→ with the ordinary śap → yasati.
Applies when / blocked by. "Not after a prefix" (anupasargāt) is a paryudāsa: it names the case "other than after a prefix". So after a prefix the option does not hold, and the general rule SK 2505 gives śyan obligatorily: prayasyati. One of Bhaṭṭoji's commentators reports a view that the two sūtras could be worded more briefly; scholars discuss this.
Rule. After yas preceded by sam, śyan also comes optionally. Bhaṭṭoji's gloss for both sūtras is "śyan comes optionally".
Example.
sam + yas + ti → with śyan → saṃyasyati
→ with śap → saṃyasati.
Applies when / blocked by. Without this rule, sam would count as a prefix under SK 2521, and śyan would be obligatory. One of Bhaṭṭoji's commentators explains that this is why the rule is made.
Watch out. Three cases: no prefix, optional; sam, optional; any other prefix, obligatory.
The root list closes the class.
jasyati. Tasu, dasu "to waste away": tasyati, atasat, dasyati, adasat.vasyati, liṭ vavāsa, vavasatuḥ. Na śasadada… (SK 2263) prohibits e with loss of the abhyāsa (SK 2260). But on the view that the root begins with b: besatuḥ, besuḥ.avyuṣat. The earlier root has sic only: avyoṣīt. Others read this root as pyus, still others as yus.apluṣat. "They say" its earlier reading is for sic. Bhaṭṭoji replies that the first-class reading already serves that purpose.bisyati, abisat. Kusa "to join": akusat. Busa "to discharge". Musa "to cut". Masī "to change"; some read samī. Luṭha "to stir".ucyati, uvoca, ūcatuḥ, and with mā, mā bhavān ucat.babharśa, abhṛśat. For bhraṃś, SK 415 deletes the n: bhraśyati, abhraśat.vṛśyati, avṛśat. Kṛśa "to become thin": kṛśyati. Ñitṛṣā "to thirst". Hṛṣa "to be content": it differs from the first-class hṛṣ by śyan and aṅ.roṣitā / roṣṭā, reṣitā / reṣṭā.yupyati, rupyati, lupyati. Lup here is seṭ: lopitā. The verses on aniṭ roots mean only the sixth-class lup, because they pair it with lip.lobhitā / lobdhā, lobhiṣyati, lubhyet, lubhyāt, alubhat. "They say" lobhati is also correct, since the first class is not closed off by vṛt.klidyati, cikleditha / ciklettha, ciklidiva / ciklidva, ciklidima / ciklidma, kleditā / klettā.medyati, with guṇa by SK 2346; luṅ amidat. Its reading in the dyutādi would already give amidat and amediṣṭa. It is read here so that *\amedīt does not arise. Bhaṭṭoji grants that a reading among the ātmanepada roots outside the dyutādi would be the proper course.ānardha, ārdhat. Gṛdhu "to covet": agṛdhat. Then "vṛt".The class closes: "the puṣādi and the divādi are finished". Some say the vṛt marks only the end of the puṣādi. On that view the divādi is, like the bhvādi, an ākṛtigaṇa, a class known by its type rather than a closed list. That accounts for forms like kṣīyate and mṛgyati. "Here end the divādi roots."
devitvā.dyūtam, ḍīnaḥ.pramāya.jānāti.janaḥ.yudhyati "desires battle".vidhyati (SK 2412).kṣudhita.| SK | What it does | Example |
|---|---|---|
| 2505 | After the div class, śyan replaces śap; being ṅit, it blocks guṇa | dīvyati, nṛtyati |
| 2506 | After kṛt, cṛt, chṛd, tṛd, nṛt: optional iṭ before an s-initial ārdhadhātuka other than sic | nartiṣyati / nartsyati; but anartīt |
| 2507 | After dīṅ: augment yuṭ before a vowel-initial kit or ṅit affix | didīye |
| 2508 | Mī, mi, dīṅ: ā before lyap and in the domain of an ec-causing aśit affix | dātā, adāsta, pramāya |
| 2509 | Lī, līṅ: the same ā, optionally | letā / lātā |
| 2510 | Root-final o deleted before śyan | śyati, chyati, syati, dyati |
| 2511 | Jñā, jan → jā before a śit affix | jāyate, jānāti |
| 2512 | Jan, vadh: no vṛddhi of the penultimate before ciṇ or a ñ/ṇ-marked kṛt | ajani, janaḥ |
| 2513 | After pad: ciṇ for cli before ta | praṇyapādi; but apatsata |
| 2514 | After aniṭ śliṣ: ksa, and only in the sense of embracing | aślikṣat; but samaśliṣat, passive aśleṣi |
| 2515 | After the eight radhādi: optional iṭ | rarandhitha / raraddha |
| 2516 | Radh: no num before a non-liṭ iṭ | radhitā; but rarandhiva |
| 2517 | Masj, naś: num before a jhal | nanaṃṣṭha, naṃṣṭā |
| 2518 | Naś ending in ṣ (even formerly): no ṇ | pranaṃṣṭā, pranaṅkṣyati; but praṇaśyati |
| 2519 | The eight śamādi: long vowel before śyan | praṇiśāmyati, mādyati |
| 2520 | As: augment thuk before aṅ | āsthat, paryāsthata |
| 2521 | Yas without a prefix: śyan optional | yasyati / yasati; but prayasyati |
| 2522 | Yas after sam: śyan optional too | saṃyasyati / saṃyasati |