This chapter teaches the causative: "he makes someone become", "he makes someone go". Pāṇini forms it with the same affix ṇic you met in chapter 53, but in a new sense. There ṇic came in the root's own sense; here it comes when the activity of a prompter is to be expressed. Bhaṭṭoji calls the section the ṇyanta-prakriyā, "the working of what ends in ṇi". After this chapter you will be able to form bhāvayati "he makes [someone] be" from bhū, with its ātmanepada bhāvayate. You will know who, in such a sentence, is the hetu and who the kartṛ. You will build the reduplicated aorist of causatives, and learn how this chapter adjusts the frame of chapter 44 for it. You will build these aorists: abībhavat, asisravat / asusravat, aśūśavat / aśiśvayat, asūṣupat, ajūhavat, ararambhat, avaveṣṭat / aviveṣṭat, apīpyat, atiṣṭhipat, ajighripat / ajighrapat, adhyajīgapat, prāṇiṇat. You will know three kinds of root. Some take an augment other than puk before ṇi: śāyayati, vājayati, vilīnayati, pālayati, bhīṣayate. Some change a final sound: sphāvayati, śātayati, ropayati. Some take ā for their vowel: krāpayati, adhyāpayati, sādhayati, vismāpayate. You will meet the roots whose causative takes the ātmanepada even when the fruit is not the agent's (lāpayate, bhāpayate). Last, you will see gam replace iṇ (gamayati), and Bhaṭṭoji's remark on a causative "from which the prompting has ceased". The chapter has thirty-three sūtras, SK 2575 to SK 2607.
bhavantaṃ prerayati: bhū + ṇic → bhau-i → bhāv-i → bhāvi + śap + ti → bhāvayati; ātmanepada bhāvayateTwo rules found the causative. SK 2575 names the participant who prompts the agent: he is the hetu, and he too is a kartṛ. SK 2576 then adds ṇic after a root when the prompter's activity is to be expressed. The running example is bhū "to be, to become". "He impels someone who is becoming" is said bhāvayati; with the ātmanepada, bhāvayate.
Rule. The one who prompts (prayojaka) the agent is given the name hetu. He is also given the name kartṛ, "agent". Bhaṭṭoji's gloss: "what prompts the agent receives the designation hetu and also the designation kartṛ".
The word tat, "that", points back to the agent of the preceding sūtra, svatantraḥ kartā (SK 559). Bhaṭṭoji's commentators agree on this. The one prompted is called the prayojya, "the one to be prompted".
Example.
"Yajñadatta makes Devadatta cook": pac + ṇic.
→ Devadatta does the cooking; he is the agent by SK 559
→ Yajñadatta sets that agent going; by this rule he is hetu
→ by this rule he is also kartṛ.
In bhāvayati likewise: the one who impels is hetu and kartṛ; the one who becomes is the prompted agent.
Applies when / blocked by. Why two names? One of Bhaṭṭoji's commentators explains. Only Devadatta is the locus of the cooking, so without this rule the prompter would not be kartṛ. And without the name hetu, his being a cause would be credited to the prompted agent.
The other commentator gives each name its use. The name hetu is used in SK 2594 and SK 2595 (hetubhaye, "when the fear comes from the hetu"). The name kartṛ lets the lakāra express the prompter, by SK 2152 (laḥ karmaṇi ca…).
How does the word ca add the second name? One commentator says ca gathers in the name kartṛ from SK 559. The other says ca sets aside the heading by which a thing takes only one name. Both reach the same double name.
Watch out. There are now two agents in the sentence. The prompted one is the agent of the root's own action. The prompter, being kartṛ too, is the one the lakāra expresses: in bhāvayati the ending -ti speaks of him.
Rule. The affix ṇic comes after a root when the hetumat is to be expressed. Hetumat means "that which has the hetu", namely the prompter's activity: directing (preṣaṇa) and the like. Bhaṭṭoji's gloss: "when the activity of the prompter, directing and the like, is to be expressed, ṇic comes after a dhātu".
"Ṇic" is carried forward from SK 2563, and "after a dhātu" (dhātoḥ) from SK 2629, taught later. One of Bhaṭṭoji's commentators gives both, and explains hetumat as the activity that has the prompter as its locus.
Example.
bhavantaṃ prerayati, "he impels [someone] who is becoming":
bhū + ṇic (this rule); ṇ and c are markers → bhū-i
→ ṇic is marked with ṇ, so the final ū takes vṛddhi (SK 254) → bhau-i
→ au before a vowel becomes āv (SK 61) → bhāvi
→ bhāvi ends in ṇic, so it is a dhātu (SK 2304)
→ laṭ, the ending tip, the class-sign śap (SK 2167) → bhāvi + a + ti
→ guṇa of the i before śap (SK 2168) → bhāve + a + ti
→ ay for e before a vowel (SK 61) → bhāvayati.
With the ātmanepada:
bhāvi + laṭ, fruit for the agent → ātmanepada by ṇicaś ca (SK 2564) → bhāvayate.
Liṭ: ām after the ṇic-stem (SK 2306), then the auxiliary bhū in liṭ (chapter 44) → bhāvayāṃ babhūva.
Applies when / blocked by. "And the like" (ādi) widens the sense. Both of Bhaṭṭoji's commentators list what it covers:
One of them, the Bālamanoramā, illustrates upadeśa: telling a fever-patient to drink a decoction. It adds that even holding back someone fleeing from being killed is a prompter's activity. The other, the Tattvabodhinī, adds anugraha, a favour. Which kind is meant is known from context.
The Tattvabodhinī also says that the prompter may be sentient or not. It may also be a result: bhikṣā vāsayati, "alms make him stay". He also contrasts ṇic with loṭ. Ṇic expresses the prompter's activity together with the prompted one's activity. Loṭ expresses the prompter's activity alone.
Why bhāvayāṃ babhūva, without the ātmanepada? One of Bhaṭṭoji's commentators answers: the auxiliary bhū is not an ātmanepadin root. Whether the ātmanepada should appear here, scholars discuss this.
Watch out. This ṇic is not the curādi ṇic of SK 2563. That one adds no meaning. This one, called hetumaṇṇic, adds the prompter's activity. Both make a ṇic-final dhātu, and both use the same machinery after that.
bhāv-i + caṅ → (bhū doubled first) bu-bhāv-i-a-t → (shortening) bu-bhav-a-t → (u → i, 2577, via sanvadbhāva) bi-bhav → (lengthening) abībhavatThe causative takes the reduplicated aorist with caṅ (SK 2312). In chapter 44, sanvadbhāva gave the abhyāsa i for a (SK 2317). But what if the abhyāsa has u, as in bubhū-? SK 2577 gives i for that u before a labial, a semivowel or j followed by a. SK 2578 makes this optional for six roots. Under SK 2578 Bhaṭṭoji also recalls two earlier points: the roots that refuse the shortening, and a ṇic added to a stem that already ends in ṇic. The running example is the aorist of bhāvayati: abībhavat.
Rule. The u of an abhyāsa is replaced by i before san. The condition: the next sound is a labial (pu, the p-class), a yaṇ (y, v, r, l) or j, and an a-vowel follows that sound. Bhaṭṭoji's gloss: "for the u of an abhyāsa that is part of an aṅga followed by san, i is substituted when a labial, a yaṇ or j, followed by an a-vowel, follows".
Carried forward: "before san" (sani) from SK 2317; "of the abhyāsa" from SK 2620, taught later; "i" (it) from SK 2496. "Of the aṅga" is the heading; the aṅga is the base to which an affix is added. Bhaṭṭoji's commentators give all of these.
Example.
bhū + ṇic + luṅ → caṅ replaces cli (SK 2312) → bhū i a t
→ doubling is due (SK 2315). While it is due, no vowel-substitute is made before ṇi. So bhū is doubled before its vṛddhi → bhū bhū i a t
→ the abhyāsa's vowel is shortened (SK 2180), and bh becomes b (SK 2182) → bu bhū i a t
→ vṛddhi (SK 254) and āv (SK 61) in the second member → bu bhāv i a t
→ the ṇi is deleted (SK 2313) → bu bhāv a t
→ the penultimate ā is shortened (SK 2314) → bu bhav a t
→ a light a follows the abhyāsa: sanvadbhāva (SK 2316). But SK 2317 needs an a in the abhyāsa, and bu has u
→ this rule: the u of bu stands before bh, a labial, followed by a → bi bhav a t
→ the light i is lengthened (SK 2318) → bī bhav a t
→ the augment aṭ (SK 2206) → abībhavat.
Bhaṭṭoji's other forms: apīpavat (pūṅ, labial p), amīmavat (mūṅ, labial m), ayīyavat (yu), arīravat (ru), alīlavat (lūñ), ajījavat (ju). Bhaṭṭoji's commentators identify the roots. One of them notes that ju is a root found only in the sūtras (sautra).
Applies when / blocked by. The step "doubled before vṛddhi" rests on a principle the commentators use throughout this chapter. It is dvitve kārye ṇāv ajādeśo na: "while doubling is due, no vowel-substitute is made before ṇi". One of Bhaṭṭoji's commentators gives it here, for abībhavat.
Bhaṭṭoji's commentators raise an objection. The shortening of ā to a is a substitute. If it were treated like the original (sthānivadbhāva), no light syllable would follow the abhyāsa, and sanvadbhāva would fail. They answer: "by the very undertaking of the rule" (ārambhasāmarthya), sthānivadbhāva does not apply here. Otherwise the rule would have nothing to do.
One of Bhaṭṭoji's commentators gives two counter-examples. Why "followed by a"? bubhūṣati: the bh is followed by ū. Why "a labial, a yaṇ or j"? ūrṇunaviṣati: the u stands before n.
Watch out. The rule is taught for san. In the aorist it works only through sanvadbhāva. So all the conditions of SK 2316 must hold too.
Rule. For the six roots sru, śru, dru, pru, plu and cyu, the i for the abhyāsa's u is optional. The condition: before san, a syllable of the root followed by an a-vowel comes next (dhātvakṣara). "Followed by a" (apare) is carried forward from SK 2577, but not "a labial, a yaṇ or j". One of Bhaṭṭoji's commentators says why: after these roots' u a labial or j is impossible, and a yaṇ is always there.
Example.
sru + ṇic + luṅ → caṅ (SK 2312) → sru i a t
→ doubling comes before the vowel-substitute, as with bhū → sru sru i a t
→ only the first consonant stays in the abhyāsa (SK 2179) → su sru i a t
→ vṛddhi (SK 254) and āv (SK 61) in the second member → su srāv i a t
→ the ṇi is deleted (SK 2313), the penultimate shortened (SK 2314) → su srav a t
→ this rule, optionally: i for that u → si srav a t, or su srav a t
→ the abhyāsa stands before the cluster sr, so it is heavy (SK 32) and is not lengthened
→ the augment aṭ → asisravat or asusravat.
One of Bhaṭṭoji's commentators gives the other pairs: aśiśravat / aśuśravat, adidravat / adudravat, apipravat / apupravat, acicyavat / acucyavat. He gives a counter-example for "followed by a": śuśrūṣati.
One of Bhaṭṭoji's commentators adds that two consonants intervene here, so there is no sanvadbhāva. How that fits the i of asisravat, scholars discuss this. What is certain is the point above: the abhyāsa is heavy, so it is not lengthened.
Recalled: SK 2572, nāglopiśāsvṛditām. Bhaṭṭoji recalls the rule that forbids the shortening of SK 2314 for an aglopin, for śās, and for a root marked with ṛ.
śās + ṇic + luṅ → śās i a t → the ṇi is deleted
→ SK 2572: the ā is not shortened → śās a t
→ doubling; the abhyāsa's vowel is shortened → śa śās a t
→ a heavy ā follows the abhyāsa, so no sanvadbhāva → aśaśāsat.
Likewise aḍuḍhaukat, from ḍhaukṛ, which is marked with ṛ.
From cakāsṛ, also marked with ṛ, Bhaṭṭoji gives acīcakāsat and, "on another view", acacakāsat. These are the two construals of SK 2316 met in chapter 44. Bhaṭṭoji's commentators say which view gives which form. In both, SK 2572 keeps the ā long: ca cakās a t.
acīcakāsat.acacakāsat.The aglopin, Bhaṭṭoji says, "will be exemplified in the section on nominal roots" (subdhātu). One commentator, the Bālamanoramā, notes that one example, mā bhavān ūnanat, was given earlier. Both say more, such as atitirāyat, will follow there.
Recalled: ṇic after a ṇic-final. Bhaṭṭoji now takes a causative of a curādi root.
cur + ṇic (SK 2563) → guṇa (SK 2189) → cori
→ the causative ṇic (SK 2576) → cori + i
→ two rules meet. Vṛddhi of the first i (SK 254) would give corāyayati. Deletion of that ṇi (SK 2313) would give corayati
→ the deletion wins, "by pūrvavipratiṣedha, or because it is an exception (apavāda)" → cor + i
→ corayati.
Pūrvavipratiṣedha is the reverse of the normal order: in a conflict the earlier rule wins. One of Bhaṭṭoji's commentators identifies the source: the vārttika ṇyallopāv iyaṅyaṇguṇavṛddhidīrghebhyaḥ pūrvavipratiṣedhena, which you met under SK 2313. The other commentator says "or because it is an exception" serves those who reject the ṇi-deletion part of that vārttika.
In luṅ:
cor i a t → the second ṇi is deleted
→ shortening by SK 2314 → cur a t
→ doubling → cu cur a t
→ lengthening by SK 2318 → cū cur a t
→ acūcurat.
But the stem has lost a ṇi, whose i is an ak. So it is an aglopin, and SK 2572 and SK 2316 should block both operations. Bhaṭṭoji answers: "ṇi" in SK 2314 is mentioned as a class (ākṛti). Bhaṭṭoji's commentators explain. The first ṇi is not followed by caṅ; the second has an aglopin base. But "ṇi-hood", shared by both, is followed by caṅ. Otherwise, one of them says, the form would be acucorat. Whether the class-reading holds, scholars discuss this.
Watch out. For these six roots the option is only between i and u in the abhyāsa. Neither is lengthened, because a cluster follows the abhyāsa in every one of them.
śvi-i-a-t → (saṃprasāraṇa + pūrvarūpa first) śu-i-a-t → śu-śu → śu-śāv-i → śu-śav → aśūśavat; the other option aśiśvayatThree roots have a semivowel that turns into its vowel in the causative aorist. SK 2579 makes this optional for śvi "to go, to grow". SK 2584 makes it obligatory for svāpi, the causative of svap "to sleep". SK 2586 does the same for hvā (hveñ "to call"). In all three the saṃprasāraṇa comes first, before vṛddhi, and so the root is doubled in its changed form. The running example is the aorist of śvi + ṇic: aśūśavat, or, without saṃprasāraṇa, aśiśvayat. Bhaṭṭoji teaches SK 2584 and SK 2586 later in his order; they are brought here because they work in the same way.
Rule. Before a ṇi that is followed by san or by caṅ, saṃprasāraṇa optionally replaces the v of śvi (śvayati). Bhaṭṭoji's gloss: "when ṇi followed by san, or ṇi followed by caṅ, follows, saṃprasāraṇa optionally replaces [the v of] śvi".
The preceding sūtra, vibhāṣā śveḥ (SK 2420), is carried forward whole; it gives "optionally" and "of śvi". Bhaṭṭoji's commentators agree on this. Where the word "saṃprasāraṇa" is carried from, scholars discuss this.
The same words, ṇau ca saṃścaṅoḥ, also form another sūtra of the Aṣṭādhyāyī, A 2.4.51. Bhaṭṭoji teaches that one at SK 2601, for a different root. Keep the two apart by their numbers.
Example.
śvi + ṇic + luṅ → caṅ (SK 2312) → śvi i a t
→ this rule: saṃprasāraṇa, v → u → śui i a t
→ pūrvarūpa (SK 330): u + i → u → śu i a t
→ śu is doubled → śu śu i a t
→ vṛddhi (SK 254) and āv (SK 61) in the second member → śu śāv i a t
→ the ṇi is deleted (SK 2313), the penultimate shortened (SK 2314) → śu śav a t
→ a light a follows the abhyāsa: sanvadbhāva, and the abhyāsa's u is lengthened (SK 2318) → śū śav a t
→ the augment aṭ → aśūśavat.
The other option, without saṃprasāraṇa:
śvi + ṇic + luṅ → śvi i a t
→ doubling comes first, before the vowel-substitute → śvi śvi i a t
→ the abhyāsa keeps only ś (SK 2179) → śi śvi i a t
→ vṛddhi (SK 254) and āy (SK 61) in the second member → śi śvāy i a t
→ the ṇi is deleted (SK 2313), the penultimate shortened (SK 2314) → śi śvay a t
→ the abhyāsa stands before the cluster śv, so it is not light. Bhaṭṭoji: "because it is not light, there is no lengthening"
→ the augment aṭ → aśiśvayat.
Applies when / blocked by. Why does saṃprasāraṇa come before vṛddhi and āy? One of Bhaṭṭoji's commentators states the problem. Vṛddhi and āy are inner (antaraṅga) and would normally go first. Then the form would not be aśūśavat. Bhaṭṭoji settles it with a maxim (paribhāṣā): saṃprasāraṇaṃ tadāśrayaṃ ca kāryaṃ balavat, "saṃprasāraṇa, and the operation that depends on it, is stronger". So saṃprasāraṇa and then pūrvarūpa come first.
Bhaṭṭoji's commentators tie the maxim to the sūtra liṭy abhyāsasyobhayeṣām. One says it is stated in the Bhāṣya on that sūtra. The other says it is drawn from the word ubhayeṣām in it, which would otherwise be superfluous. He adds the case it protects: in vavraśca the v of the abhyāsa would otherwise take saṃprasāraṇa.
Watch out. The saṃprasāraṇa is optional. Always give both forms: aśūśavat and aśiśvayat.
Rule. Saṃprasāraṇa replaces the v of svāpi, the ṇic-final of svap, when caṅ follows. "Saṃprasāraṇa" is carried forward from SK 1003, one of Bhaṭṭoji's commentators says.
Example.
svap + ṇic → vṛddhi of the penultimate a (SK 2282) → svāpi; in luṅ:
svāp i a t → this rule: saṃprasāraṇa, v → u, first, by the maxim of SK 2579 → suāp i a t
→ doubling, with pūrvarūpa (SK 330): u + ā → u → su sup i a t
→ the ṇi is deleted (SK 2313) → su sup a t
→ a light u follows the abhyāsa: sanvadbhāva, and the abhyāsa's u is lengthened (SK 2318) → sū sup a t
→ ṣ for the s after the abhyāsa's ū (SK 212; the commentators' ṣatva), as in liṭ suṣvāpa → sū ṣup a t
→ the augment aṭ → asūṣupat.
This is the order one of Bhaṭṭoji's commentators gives: saṃprasāraṇa, then doubling, pūrvarūpa, sanvadbhāva with lengthening, and ṣ.
Applies when / blocked by. Bhaṭṭoji's commentators stress that the doubling comes after the saṃprasāraṇa. One of them gives the counter-example for "when caṅ follows": svāpayati, with no change.
Watch out. Unlike SK 2579, this rule is not optional, and it names only caṅ, not san.
Rule. Saṃprasāraṇa replaces the v of hvā before a ṇi followed by san or by caṅ. Hvā is the root hveñ "to call" with ā already put for its e (SK 2370). Bhaṭṭoji's gloss begins "before a ṇi followed by san and before a ṇi followed by caṅ". Bhaṭṭoji's commentators explain why: ṇau ca saṃścaṅoḥ (SK 2579) continues into this rule.
Example.
hvā + ṇic + luṅ → caṅ (SK 2312) → hvā i a t
→ this rule: saṃprasāraṇa, v → u, and pūrvarūpa (SK 330), done first by the maxim of SK 2579 → hu i a t
→ no hvā is left, so the augment yuk (SK 2585) has nothing to attach to
→ hu is doubled → hu hu i a t
→ the h of the abhyāsa becomes j (SK 2245, 2182) → ju hu i a t
→ vṛddhi (SK 254) and āv (SK 61) in the second member → ju hāv i a t
→ the ṇi is deleted (SK 2313), the penultimate shortened (SK 2314) → ju hav a t
→ sanvadbhāva, and the abhyāsa's u is lengthened (SK 2318) → jū hav a t
→ the augment aṭ → ajūhavat.
Bhaṭṭoji also gives ajuhāvat. One of his commentators explains it. The shortening of SK 2314 is optional here by the vārttika kāṇyādīnāṃ vā, under SK 2583. Its list includes hvāyi. Without the shortening, a heavy ā follows the abhyāsa. So there is no sanvadbhāva and no lengthening: ajuhāvat.
Applies when / blocked by. In laṭ there is no san or caṅ, so no saṃprasāraṇa. The root keeps hvā and takes yuk (SK 2585): hvāyayati.
Watch out. The order matters. Saṃprasāraṇa first removes the very form hvā that SK 2585 names. That is why the aorist has no y.
rabh-i + caṅ → rambh-i (num, 2581) → ra-rambh-a-t → ararambhat (heavy, so no sanvadbhāva); veṣṭ-i + caṅ → avaveṣṭat / aviveṣṭatThese four sūtras, and the remarks Bhaṭṭoji hangs on them, settle the aorist of particular stems. SK 2580 stops an upasarga from bringing ṣ into stambh, siv and sah. Under it Bhaṭṭoji works through the roots that begin with a vowel. SK 2581 and SK 2582 give the augment num to rabh and labh. SK 2583 lets veṣṭ and ceṣṭ keep a in the abhyāsa, and a vārttika under it makes the shortening optional for a list of roots. The running examples are ararambhat and avaveṣṭat / aviveṣṭat.
Rule. When caṅ follows, ṣ does not replace the s of stambh, siv and sah, where the conditioning sound stands in an upasarga. Bhaṭṭoji's gloss: "when the conditioning [sound] stands in an upasarga, ṣ does not replace the s of these when caṅ follows".
The limit "in an upasarga" is not in the sūtra's words. One of Bhaṭṭoji's commentators says it comes from a vārttika, upasarganimittasya pratiṣedhaḥ, "the prohibition is of [ṣ] conditioned by an upasarga". He also gives the carried words: "not" (na) from SK 3168, taught later; "of s" (saḥ) from SK 335. "A retroflex" is the heading.
Which rules would give the upasarga-ṣ? For siv and sah, SK 2275 (parinivibhyaḥ…), Bhaṭṭoji's commentators say. For stambh, one commentator, the Tattvabodhinī, names SK 2272 (stambheḥ). The other, the Bālamanoramā, names SK 2273 (avāc cālambanāvidūryayoḥ).
Example.
ava + stambh + ṇic + luṅ → ava stambh i a t
→ the ṇi is deleted (SK 2313); the penultimate is a consonant, m, so nothing is shortened
→ doubling; of the cluster st the abhyāsa keeps t (SK 2259) → ava ta stambh a t
→ the a of stambh stands before a cluster, so it is heavy: no sanvadbhāva
→ the augment aṭ, and a + a → ā (SK 85) → avātastambhat
→ SK 2273 would give ṣ after the upasarga ava; this rule blocks it.
In laṭ, by contrast, the ṣ comes: avaṣṭambhayati, one of Bhaṭṭoji's commentators notes.
pari + siv + ṇic + luṅ → siv i a t
→ doubling comes first (SK 2315); the abhyāsa keeps s (SK 2179) → si siv i a t
→ guṇa in the second member (SK 2189) → si sev i a t
→ the ṇi is deleted (SK 2313), the e shortened (SK 2314) → si siv a t
→ a light i follows the abhyāsa: sanvadbhāva (SK 2316); the abhyāsa's i is lengthened (SK 2318) → sī siv a t
→ ṣ for the s after the abhyāsa's ī (SK 212; the commentators' ṣatva) → sī ṣiv a t
→ the augment aṭ, and pari + a → parya (SK 47) → paryasīṣivat.
The ṣ that the abhyāsa conditions still comes. Only the ṣ that pari would bring to the first s is blocked. One of Bhaṭṭoji's commentators makes this point.
ni + sah + ṇic + luṅ → sah i a t
→ doubling comes first (SK 2315); the abhyāsa keeps s (SK 2179) → sa sah i a t
→ vṛddhi of the penultimate a in the second member (SK 2282) → sa sāh i a t
→ the ṇi is deleted (SK 2313), the ā shortened (SK 2314) → sa sah a t
→ a light a follows the abhyāsa: sanvadbhāva (SK 2316), i for the abhyāsa's a (SK 2317) → si sah a t
→ the light i is lengthened (SK 2318) → sī sah a t
→ ṣ for the s after the abhyāsa's ī (SK 212; the commentators' ṣatva) → sī ṣah a t
→ the augment aṭ (SK 2206), and ni + a → nya (SK 47) → nyasīṣahat.
The vowel-initial roots. Bhaṭṭoji next gives āṭiṭat (aṭ) and āśiśat (aś). He states a point of order: "the shortening of the penultimate, though bahiraṅga, takes place before doubling itself".
Bahiraṅga means "outer": a rule whose condition lies further out. The shortening needs ṇi followed by caṅ; doubling needs caṅ alone. So, one of Bhaṭṭoji's commentators explains, the shortening is the outer one and would normally wait.
Bhaṭṭoji's proof is a sign (liṅga): the root oṇ is marked with ṛ (oṇṛ). That marker exists to block the shortening by SK 2572. If ṇi were doubled first, the o would no longer stand before ṇi, and the marker would have nothing to block. Bhaṭṭoji's commentators both draw this.
aṭ + ṇic → vṛddhi (SK 2282) → āṭ i; in luṅ:
āṭ i a t → shortening first (SK 2314) → aṭ i a t
→ the second syllable, ṭi, is doubled together with the i of ṇic (SK 2176) → a ṭi ṭi a t
→ the ṇi of the second member is deleted (SK 2313) → aṭiṭ a t
→ the augment āṭ (SK 2254), and ā + a → ā (SK 269) → āṭiṭat.
One of Bhaṭṭoji's commentators notes that ṭi is doubled with the ṇic, so there is no sanvadbhāva and no lengthening.
One of Bhaṭṭoji's commentators adds why Bhaṭṭoji needs a further example. āṭiṭat and āśiśat come out the same either way, by SK 269. The order shows only with edh:
edh + ṇic + luṅ, with mā → edh i a t
→ shortening first: e becomes i → idh i a t
→ dhi is doubled (SK 2176); the abhyāsa's dh becomes d (SK 2182) → i di dhi a t
→ the ṇi is deleted → ididh a t
→ with mā there is no augment (SK 2228) → mā bhavān ididhat.
Had doubling come first, the e would not have been shortened.
With pra: mā bhavān predidhat. There is no vṛddhi, Bhaṭṭoji says. SK 73 (etyedhatyūṭhsu) is taught for edh beginning with an ec, and idh does not. So pra + i takes plain guṇa (SK 69). One of Bhaṭṭoji's commentators adds that doubling is "uttering twice" (dviḥprayoga), as the Bhāṣya settles.
No doubling of n, d, r. By na ndrāḥ saṃyogādayaḥ (SK 2446), an n, d or r at the head of a cluster is not doubled:
und → und i a t → n is not doubled, so di is doubled → un di di a t → undid a t → āṭ and vṛddhi → aundidat;aḍḍ → only ḍi is doubled → āḍḍiḍat;arc → only ci is doubled → ārcicat.For aḍḍ, one of Bhaṭṭoji's commentators says the root's penultimate is really d, made ḍ by a retroflex rule that has not yet taken effect.
ubj. The root ubja "in straightness" has d as its penultimate in upadeśa, Bhaṭṭoji says: udj. The b comes by nipātana: the root appears in the form ubj in the sūtra bhujanyubjau pāṇyupatāpayoḥ (SK 2877). This substitution, though inner, waits until ji has been doubled. Bhaṭṭoji's reason: the root is pronounced with d, and that pronouncing must serve a purpose.
udj + ṇic + luṅ → udj i a t
→ SK 2446 stops the d from being doubled, so only ji is doubled → ud ji ji a t
→ the ṇi is deleted → udjij a t
→ now b for d, by the nipātana → ubjij a t
→ āṭ and vṛddhi → aubjijat.
Without this order, one of Bhaṭṭoji's commentators says, bji would be doubled, giving the wrong aubibjat. The other commentator reports that the vārttika-maker took the root's penultimate differently, and that, with the Bhāṣya, it is taken as d.
Only after an initial vowel. SK 2446 applies only to a root that begins with a vowel (ajādeḥ ity eva). Bhaṭṭoji's commentators explain: ajādeḥ is carried into it from SK 2176. "Not here", says Bhaṭṭoji: adidrapat. The d stands at the head of the root, so the abhyāsa simply keeps it. One of Bhaṭṭoji's commentators takes adidravat, from dru, as the intended form, and adidrapat as from drā with puk; scholars discuss this.
Watch out. This rule blocks only the upasarga's ṣ. Do not drop the ṣ that the abhyāsa brings: paryasīṣivat, not paryasīsivat.
Rule. The augment num is added to rabh before a vowel-initial affix, but not before śap or liṭ. Num is carried forward from SK 2262 (idito num dhātoḥ), and "before a vowel" (aci) from SK 2302. Bhaṭṭoji's commentators give both.
Example.
rabh + ṇic: ṇic begins with a vowel, and it is neither śap nor liṭ
→ this rule: num, its n placed after the root's last vowel (SK 37) → ranbh-i
→ n before bh becomes anusvāra (SK 123), and the anusvāra becomes m, the nasal of bh's class (SK 124) → rambh-i
→ after num the penultimate is m, so SK 2282 does not lengthen the a.
Bhaṭṭoji gives the aorist under the next sūtra: ararambhat.
Applies when / blocked by. One of Bhaṭṭoji's commentators gives the counter-examples. Not before a vowel: ārabdham. Before śap: rabhate. Before liṭ: rebhe.
Watch out. Num is an augment, not a substitute: nothing of rabh is replaced. It adds n after the root's vowel; the sandhi steps above turn that n into m.
Rule. Labh likewise takes num before a vowel-initial affix other than śap and liṭ. The rest is carried from SK 2581. One of Bhaṭṭoji's commentators says Bhaṭṭoji gives no gloss because the meaning is clear.
Example.
rambh-i + luṅ → rambh i a t
→ the ṇi is deleted (SK 2313); there is no long penultimate to shorten → rambh a t
→ doubling → ra rambh a t
→ after num the a of rambh is followed by a cluster, so it is not light: no sanvadbhāva
→ the abhyāsa keeps a, and is not lengthened
→ the augment aṭ → ararambhat.
Likewise alalambhat from labh.
One of Bhaṭṭoji's commentators says "no sanvadbhāva" of ararambhat; it holds equally for alalambhat.
Applies when / blocked by. Why a separate sūtra for labh? One of Bhaṭṭoji's commentators says the rule is split off (yogavibhāga) so that labh alone, not rabh, continues into the next rule, āṅo yi.
Recalled: SK 2531, her acaṅi. That rule gives a velar for the h of hi after an abhyāsa, "but not before caṅ". So before caṅ there is no velar (kutva):
hi + ṇic + luṅ → hi i a t → hi is doubled first → hi hi i a t
→ the abhyāsa's h becomes j (SK 2245, 2182) → ji hi i a t
→ vṛddhi and āy in the second member → ji hāy i a t
→ the ṇi is deleted, the penultimate shortened → ji hay a t
→ sanvadbhāva and lengthening → jī hay a t
→ no gh, by acaṅi → ajīhayat.
Recalled: SK 2566, at smṛdṝtvaraprathamradastṝspaśām. This rule puts a in the abhyāsa of seven roots before caṅ.
smṛ + ṇic + luṅ → smṛ i a t
→ doubling comes first (SK 2315) → smṛ smṛ i a t
→ a for the abhyāsa's ṛ (SK 2244), and only the first consonant kept (SK 2179) → sa smṛ i a t
→ vṛddhi in the second member (SK 254), with r following → sa smār i a t
→ the ṇi is deleted (SK 2313), the penultimate shortened (SK 2314) → sa smar a t
→ sanvadbhāva would give i (SK 2317); SK 2566 gives a instead
→ the augment aṭ → asasmarat. Likewise adadarat from dṝ.
Bhaṭṭoji adds: "here, by force of the t-marking of at, there is no lengthening of the light vowel". At with t names only short a. One of Bhaṭṭoji's commentators takes "here" with adadarat. In asasmarat the abhyāsa stands before the cluster sm and is not light anyway. He adds the other roots' forms: atatvarat, apaprathat, amamradat, atastarat, apaspaśat.
Watch out. Num changes the weight of the syllable. Rabh has a light a; rambh has a heavy one. That one change removes sanvadbhāva.
Rule. For veṣṭ and ceṣṭ, a optionally replaces the vowel of the abhyāsa before a ṇi followed by caṅ. Bhaṭṭoji's gloss: "a is optionally substituted for the abhyāsa when ṇi followed by caṅ follows".
One of Bhaṭṭoji's commentators gives the carried words: "of the abhyāsa" from SK 2620, taught later; "a" (at) from SK 2566; "followed by caṅ" (caṅpare) from SK 2316. "Ṇi" is understood from the sense.
Example.
veṣṭ + ṇic + luṅ → veṣṭ i a t
→ the ṇi is deleted (SK 2313); the penultimate is a consonant, so nothing is shortened → veṣṭ a t
→ doubling; the abhyāsa keeps v and a short vowel, i (SK 2180) → vi veṣṭ a t
→ the e of veṣṭ is a long vowel, so the syllable is heavy: no sanvadbhāva
→ this rule, optionally: a for the abhyāsa's vowel → va veṣṭ a t
→ the augment aṭ → avaveṣṭat or aviveṣṭat.
Likewise acaceṣṭat or aciceṣṭat.
Recalled: SK 2565, bhrājabhāsa…. By that rule the shortening is optional for bhrāj:
bhrāj i a t → bhraj a t → ba bhraj a t (SK 2179, 2182) → a light a follows, so sanvadbhāva: bi bhraj a t → the abhyāsa stands before the cluster bhr, so no lengthening → abibhrajat;ba bhrāj a t → a heavy ā follows, so no sanvadbhāva → ababhrājat.One of Bhaṭṭoji's commentators explains the i and the absence of lengthening in abibhrajat.
Vārttika: kāṇyādīnāṃ veti vaktavyam. "It should be stated that for kāṇi and the rest it is optional." One of Bhaṭṭoji's commentators supplies what is optional: the shortening of SK 2314.
Bhaṭṭoji lists the roots:
Vāṇi already stands among the Nyāsa's four, so which two roots the last line names, scholars discuss this.
kaṇ + ṇic → vṛddhi of the penultimate a (SK 2282) → kāṇi; luṅ kāṇ i a t → the ṇi is deleted (SK 2313) → kāṇ a t
kaṇ a t → ka kaṇ a t → sanvadbhāva: i (SK 2317), lengthened (SK 2318) → kī kaṇ a t → the k of the abhyāsa becomes c (SK 2245) → acīkaṇat;ka kāṇ a t → a heavy ā follows, so no sanvadbhāva → ca kāṇ a t → acakāṇat.Watch out. Two different options sit in this section. SK 2583 makes the abhyāsa vowel optional. The vārttika makes the shortening of the stem optional. Do not mix them.
śā + ṇi → śā-y-i (yuk, not puk) → śāyayati; vā → vāj-i → vājayati; lī → līn-i → vilīnayatiIn chapter 53 you met puk (SK 2570): before ṇi, a root ending in ā adds p, as in sthāpayati. This group gives other augments to particular roots instead. An augment (āgama) is a sound added to a form, not put in place of anything. SK 2585 adds y (yuk), SK 2590 adds j (juk), and SK 2591 adds n (nuk) or l (luk). A vārttika under SK 2589, taught in group F, adds l to pā "to protect". The running examples are śāyayati, vājayati and vilīnayati. Bhaṭṭoji teaches SK 2585 before group F; it is placed here with the other augments.
Rule. The augment yuk is added to śā, chā, sā, hvā, vyā, ve and pā when ṇi follows. Bhaṭṭoji supplies "when ṇi follows", which the sūtra leaves unstated. He adds that yuk is an exception (apavāda) to puk.
One of Bhaṭṭoji's commentators identifies the roots. Śo, cho, ṣo, hveñ and vyeñ are cited with their ā already made (SK 2370). Veñ "to weave" and pā "to drink" are the other two. The k of yuk is a marker and the u is for pronunciation. Being marked with k, the y is added at the end of the root.
Example.
śo → ā for the o in upadeśa (SK 2370) → śā
→ śā + ṇic → puk (SK 2570) would add p; this rule adds y instead → śāy-i
→ laṭ, tip, śap, guṇa of i, and ay (SK 2167, 2168, 61) → śāyayati.
Likewise hvāyayati.
One commentator, the Tattvabodhinī, gives aorists for these roots: aśīśayat, acicchayat, asīṣayat, avivyayat, avīvayat, and, for pā, apīpayat. Bhaṭṭoji's own aorist of pā, under SK 2587, is apīpyat.
Applies when / blocked by. Pā here is pā "to drink", not pā "to protect". Bhaṭṭoji's commentators explain by a maxim. Where a root with luk of its class-sign and one without could both be meant, the one without is taken. Pā "to protect" is such a root: its class-sign is lost. Its form, pālayati, comes under SK 2589.
One of Bhaṭṭoji's commentators makes a further point. The roots are cited with ā already made. That shows that puk would otherwise apply to an ā made by a rule. So in this section the maxim lakṣaṇapratipadokta does not operate. That maxim prefers a form named directly to one made by a rule. Hence dhyāpayati and kāpayati, with puk.
Watch out. Hvā takes yuk in laṭ (hvāyayati) but not in the aorist, where SK 2586 removes hvā first (ajūhavat).
Rule. The augment juk is added to vā before ṇi, in the sense of shaking (vidhūnana). Bhaṭṭoji's gloss: "let juk be added to vāti before ṇi, in the sense of shaking". "Before ṇi" (ṇau) is carried forward from SK 2570, one of Bhaṭṭoji's commentators says.
Example.
vā + ṇic, "shaking" → this rule: juk → vāj-i
→ vājayati.
Applies when / blocked by. Why "in the sense of shaking"? Otherwise the root takes puk: keśān vāpayati. One of Bhaṭṭoji's commentators glosses it "he makes the hair fragrant".
Which vā is meant? Bhaṭṭoji's commentators take vai "to dry" with its ā made. The adādi vā (vā gatigandhanayoḥ) is excluded, being a root with luk of its class-sign. One of them notes that Vāmana too agrees. Bhaṭṭoji's own word is vāti; whether it names the right root, scholars discuss this.
The same commentator adds: vājayati would come anyway from vaja "to go". So the only result of juk is to keep puk out.
Recalled: SK 2509, vibhāṣā līyateḥ. Bhaṭṭoji recalls the rule that makes ā optional for lī. Bhaṭṭoji's commentators note that it cites the root with yak (līyati), so it covers both lī and līṅ. It prepares the next sūtra.
Watch out. The augment depends on the sense. One root, two causatives: vājayati "makes shake", vāpayati otherwise.
Rule. Before ṇi, lī optionally takes the augment nuk, and lā optionally takes the augment luk, in the sense of liquefying a fat (sneha). Bhaṭṭoji's gloss: "let the augments nuk and luk optionally be added, respectively, to līyati and to lāti, before ṇi, in the melting of a fat". "Before ṇi" is carried from SK 2570, one of Bhaṭṭoji's commentators says.
Example.
Bhaṭṭoji's four forms, all with vi and the object ghṛtam, "he melts the ghee":
lī + ṇic → this rule: nuk → līn-i → vilīnayati;lī + ṇic without nuk → vṛddhi (SK 254) → lai-i → āy (SK 61) → lāy-i → vilāyayati;lā + ṇic → this rule: luk → lāl-i → vilālayati;lā + ṇic without luk → puk (SK 2570) → lāp-i → vilāpayati.Bhaṭṭoji's commentators explain the roots. Lī covers lī and līṅ. Lā covers the adādi lā and lī or līṅ with ā made by SK 2509.
Applies when / blocked by. Bhaṭṭoji adds a point about the wording. An ī is fused into the sūtra (praśleṣa): read "lī ī". So nuk goes only to lī ending in ī. In the option with ā, there is no nuk. One of Bhaṭṭoji's commentators gives the reason this is needed. By the maxim "what is altered in one part is not another thing" (ekadeśavikṛta), lā would still count as lī. The fused ī blocks that.
Why "liquefying a fat"? lohaṃ vilāpayati, vilāyayati, "he melts iron". Bhaṭṭoji's commentators explain: no nuk or luk. With ā there is puk; without ā, vṛddhi and āy.
Vārttika: pralambhanābhibhavapūjāsu liyo nityam ātvam aśiti vācyam. "In the senses of deceiving, overpowering and honouring, it should be stated that lī takes ā invariably before an affix not marked with ś." SK 2509 made the ā optional; this vārttika makes it obligatory in three senses. One of Bhaṭṭoji's commentators gives the counter-example for "not ś-marked": līyate, lināti. Its forms come under SK 2592.
Watch out. Keep the four forms apart by two questions: is the vowel ī or ā, and is the augment taken or not?
pāy-i + caṅ → a-pa-pāy-a-t → (penultimate ā deleted) a-pa-py-a-t → (ī for the abhyāsa final) apīpyat; sthāp-i → a-ti-ṣṭhap-a-t → atiṣṭhipatThree rules change the stem of the causative aorist for three roots with ā. SK 2587 deletes the ā of pā and gives the abhyāsa ī. SK 2588 puts i for the penultimate of sthā. SK 2589 makes the same i optional for ghrā. Under them Bhaṭṭoji recalls puk and ur ṛt, and a vārttika gives pā "to protect" its own augment. The running examples are apīpyat and atiṣṭhipat.
Rule. Before a ṇi followed by caṅ, the penultimate of pā (pibati) is deleted, and ī replaces the final of the abhyāsa. "Before a ṇi followed by caṅ" and "the penultimate" are carried from SK 2314. One of Bhaṭṭoji's commentators gives this.
Example.
pā + ṇic → yuk (SK 2585) → pāyi; in luṅ:
pāy i a t → the ṇi is deleted (SK 2313) → pāy a t
→ doubling; the abhyāsa keeps p and a short vowel (SK 2179, 2180) → pa pāy a t
→ this rule: the penultimate ā of pāy is deleted → pa py a t
→ this rule: ī replaces the final of the abhyāsa → pī py a t
→ the augment aṭ → apīpyat.
Applies when / blocked by. Why does ī replace only the final of the abhyāsa? A maxim says a substitute for a meaningless item does not take its final (nānarthake 'lo 'ntyavidhiḥ). Bhaṭṭoji's commentators note that the maxim excepts a change made to an abhyāsa. So ī replaces the a of pa.
Why is ī taught at all? Bhaṭṭoji's commentators explain. Once the ā is deleted, the stem has lost an ak: it is an aglopin. And the stem has no light penultimate. So neither the i nor the lengthening of sanvadbhāva is available. This rule supplies the ī directly.
How is pāy, not py, doubled? One of Bhaṭṭoji's commentators gives two ways. Either the deletion is held back while doubling is due, or, once made, it is treated like the original (sthānivat) for doubling.
Recalled: SK 2570, puk. Bhaṭṭoji now illustrates puk in the order of that sūtra. One of Bhaṭṭoji's commentators supplies the steps:
ṛ + ṇic → puk comes before vṛddhi, being nitya: it would apply whether or not vṛddhi had applied first → ṛp-i → guṇa (SK 2189) → arp-i → arpayati;hrī → hrīp-i → guṇa, the stem ending in puk (SK 2189) → hrepayati; likewise vlepayati (vlī), repayati (rī);knūyī (ī is a marker; the root is knūy) → puk → knūyp-i → y deleted before a consonant (SK 873) → knūp-i → guṇa (SK 2189), though the penultimate is not light, because the stem ends in puk → knopayati;kṣmāyī (ī again a marker; the root is kṣmāy) → puk, y deleted → kṣmāp-i → kṣmāpayati;sthā → sthāp-i → sthāpayati.Watch out. Pā "to drink" meets two rules of this chapter. Before ṇi it takes yuk (SK 2585), so its laṭ is pāyayati. In the aorist this rule then deletes the ā of pāy: apīpyat.
Rule. Before a ṇi followed by caṅ, i replaces the penultimate of sthā (tiṣṭhati). "The penultimate" and the condition are carried from SK 2314, one of Bhaṭṭoji's commentators says.
Example.
sthā + ṇic → puk (SK 2570) → sthāpi; in luṅ:
sthāp i a t → the ṇi is deleted, the penultimate shortened (SK 2313, 2314) → sthap a t
→ doubling; of the cluster sth the abhyāsa keeps th (SK 2259), which becomes t (SK 2182) → ta sthap a t
→ sanvadbhāva: i for the abhyāsa's a (SK 2317) → ti sthap a t
→ ṣ for the s after the abhyāsa's i (SK 212; the commentators' ṣatva), as in tiṣṭhati, and ṭh for th after ṣ (SK 113) → ti ṣṭhap a t
→ the abhyāsa stands before the cluster ṣṭh, so it is not light and is not lengthened
→ this rule: i for the penultimate a → ti ṣṭhip a t
→ the augment aṭ → atiṣṭhipat.
One of Bhaṭṭoji's commentators sets up the stage atiṣṭhap a t, to which this rule applies.
Applies when / blocked by. The root is cited as tiṣṭhati, with its present stem. One of Bhaṭṭoji's commentators says this excludes the stem of the intensive with luk (yaṅluk): atāsthāpat.
Watch out. The i here replaces the stem's penultimate, not the abhyāsa's vowel. The abhyāsa already has its i from sanvadbhāva.
Rule. For ghrā (jighrati), the i for the penultimate before a ṇi followed by caṅ is optional. The rest is carried as in SK 2588, as one of Bhaṭṭoji's commentators supplies it.
Example.
ghrā + ṇic → puk → ghrāpi; in luṅ:
ghrāp i a t → the ṇi is deleted, the penultimate shortened → ghrap a t
→ doubling; the abhyāsa keeps gh, which becomes j (SK 2245, 2182) → ja ghrap a t
→ sanvadbhāva: i (SK 2317); before the cluster ghr there is no lengthening → ji ghrap a t
→ this rule, optionally: i for the penultimate → ji ghrip a t
→ ajighripat or ajighrapat.
Recalled: SK 2567, ur ṛt. Before a ṇi followed by caṅ, ṛ optionally replaces a penultimate ṛ-vowel. Bhaṭṭoji gives three pairs; one of his commentators derives each.
kṝt i a t → ṛ for ṝ (SK 2567), the ṇi deleted → kṛt a t → doubling; a for the abhyāsa's ṛ (SK 2244), one consonant kept → ka kṛt a t → c for k (SK 2245) → sanvadbhāva: i, lengthened → cī kṛt a t → acīkṛtat.kirt → doubling → ki kirt a t → lengthening in the second member by SK 2265 (upadhāyāṃ ca) → ki kīrt a t → a heavy kīr follows the abhyāsa: no sanvadbhāva → c for k → acikīrtat.vṛt a t → va vṛt a t → i, lengthened → avīvṛtat.va vart a t → no light syllable follows the abhyāsa, so neither sanvadbhāva nor lengthening → avavartat.mṛj a t → ma mṛj a t → i, lengthened → mī mṛj a t → amīmṛjat. Without ṛ: vṛddhi in the second member by SK 2473 (mṛjer vṛddhiḥ) → ma mārj a t → no sanvadbhāva → amamārjat.Vārttika: pāter ṇau lug vaktavyaḥ. "Luk should be stated for pā [to protect] before ṇi." Here luk is an augment, l, not a deletion; one of Bhaṭṭoji's commentators says so. Bhaṭṭoji: "it is the exception to puk".
pā "to protect" + ṇic → puk would add p; the vārttika adds l → pāl-i → pālayati.
Bhaṭṭoji's commentators add that pālayati would come anyway from the root pāla "to protect". So the vārttika's only result is to keep puk out.
Watch out. "Luk" usually means deletion of an affix. In this vārttika, and in SK 2591, it is an augment: the sound l.
bhī + ṇic (fear from the prompter) → bhai-i → bhā-i (2593) → bhāp-i (puk) → muṇḍo bhāpayate (ātmanepada, 2594); or bhī-ṣ-i (ṣuk, 2595) → bhīṣayateBy ṇicaś ca (SK 2564) a ṇic-stem takes the ātmanepada when the fruit of the action goes to the agent. This group gives the ātmanepada to three causatives even when it does not. SK 2592 does so for lī in three senses. SK 2594 does so for bhī "to fear" and smi "to wonder", when the fear or wonder comes from the hetu, the prompter. Around it, SK 2593 and SK 2596 give these two roots ā, and SK 2595 gives bhī the augment ṣ. The running example is muṇḍo bhāpayate, "the shaven-headed man frightens [people]", or, with ṣuk, bhīṣayate.
Rule. After lī and līṅ ending in ṇi, the ātmanepada comes in the senses of honouring and of overpowering, and, by ca, of deceiving. It comes even when the fruit does not go to the agent. Bhaṭṭoji's gloss: "let ātmanepada come after lī and līṅ ending in ṇi, even when the fruit does not go to the agent, in the senses of honouring and overpowering, and in the sense of deceiving".
One of Bhaṭṭoji's commentators gives the carried words: "ātmanepada" from SK 2157; "ending in ṇi" (ṇeḥ) from SK 2738, taught later. Ca gathers in "deceiving" (pralambhana) from SK 2739, also taught later. The other commentator also reads ca as "deceiving".
Example.
lī + ṇic, sense "honouring" → vṛddhi (SK 254) → lai-i
→ ā invariably for that ec, by the vārttika under SK 2591 → lā-i
→ puk (SK 2570) → lāp-i
→ this rule: ātmanepada, the ending ta; śap (SK 2167) → lāpi + a + ta
→ laṭ is marked with ṭ, so e replaces the final a of the ending (SK 2233) → lāpi + a + te
→ guṇa of the i before śap (SK 2168), ay (SK 61) → lāpayate.
The steps after lāp-i are those of bhāvayati under SK 2576.
Bhaṭṭoji's three sentences:
jaṭābhir lāpayate: "he gains honour by his matted locks";śyeno vartikām ullāpayate: "the hawk overpowers the quail";bālam ullāpayate: "he deceives the child".Applies when / blocked by. Why "even when the fruit does not go to the agent"? One of Bhaṭṭoji's commentators explains: when it does go to the agent, ṇicaś ca (SK 2564) already gives the ātmanepada. So this rule is for the other case.
Watch out. Outside these three senses, lī with ṇic follows SK 2564 like any ṇic-stem. The rule is tied to its senses.
Rule. Before ṇi, ā optionally replaces the ec of bhī (bibheti), when the fear comes from the hetu. Bhaṭṭoji's gloss: "let ā optionally be substituted for the ec of bibheti, if the fear is from the prompter".
Bhaṭṭoji's commentators give the carried words. "Ec" and "ā" come from SK 2370; "optionally" from SK 2509; "before ṇi" from SK 2569. They gloss hetubhaya as fear whose source is the prompter.
Example.
bhī + ṇic, fear from the prompter → vṛddhi (SK 254) → bhai-i
→ this rule, optionally: ā for the ec ai → bhā-i
→ puk (SK 2570), the root now ending in ā → bhāp-i.
The finished forms come under the next two sūtras: bhāpayate with ā, bhīṣayate without it.
Applies when / blocked by. The fear must come from the prompter. If it comes from something else, such as an instrument, there is no ā. Bhaṭṭoji shows this under SK 2596.
Watch out. Bhaṭṭoji's gloss says ec. The vowel ai arises here by vṛddhi before ṇic; the derivation above takes that ai as the ec replaced.
Rule. After bhī and smi ending in ṇi, the ātmanepada comes when the fear (or wonder) comes from the hetu. Bhaṭṭoji's gloss: "let ātmanepada come after these two when they end in ṇi, if fear or wonder is from the hetu". "Ātmanepada" is carried from SK 2157, and "ending in ṇi" from SK 2738, taught later.
But the sūtra says only "fear". Bhaṭṭoji explains: the word bhaya stands for the sense of each root (upalakṣaṇa). So for smi it means wonder (smaya). One of Bhaṭṭoji's commentators raises the objection this answers.
Example.
bhī + ṇic → bhāp-i (SK 2593 and puk, as above)
→ this rule: ātmanepada, ta → śap, te (SK 2167, 2233) → bhāpi + a + te → guṇa, ay (SK 2168, 61) → bhāpayate
→ muṇḍo bhāpayate, "the shaven-headed man frightens [people]".
One of Bhaṭṭoji's commentators lists the three operations in this form: ā, puk, and the ātmanepada.
Applies when / blocked by. Bhaṭṭoji's commentators say this rule, like SK 2592, serves the case where the fruit is not the agent's. One of them notes that the sūtra's sixth-case form bhīsmyoḥ is used for the fifth, "after these two" (vyatyaya, interchange).
Watch out. This rule gives the ātmanepada; it does not give ā. The ā of bhāpayate comes from SK 2593, and the p from puk.
Rule. Before ṇi, the augment ṣuk is added to bhī ending in ī, when the fear comes from the hetu. Bhaṭṭoji explains the words "ending in ī": "an ī is fused in [the sūtra, read as] bhī ī". "Before ṇi" is carried from SK 2570.
Example.
bhī + ṇic, fear from the prompter, without the ā of SK 2593
→ this rule: ṣuk, ṣ at the end of the root → bhīṣ-i
→ the penultimate ī is long, so it takes no guṇa
→ ātmanepada by SK 2594 → bhīṣayate.
Applies when / blocked by. What does the fused ī achieve? One of Bhaṭṭoji's commentators answers. It limits the rule to bhī ending in ī. In the option with ā (SK 2593), the root no longer ends in ī, so there is no ṣuk: bhāpayate, with puk. He adds that the Bhāṣya makes this reading clear, in its discussion of the sthānivat sūtra.
Watch out. With fear from the prompter, bhī has two causatives: bhāpayate (ā + puk) and bhīṣayate (ṣuk). Never both ā and ṣ.
Rule. Before ṇi, ā invariably replaces the ec of smi (smayati), when the wonder comes from the hetu. Bhaṭṭoji's gloss: "let ā invariably be substituted for the ec of smayati before ṇi, when the wonder is from the hetu".
One of Bhaṭṭoji's commentators gives the carried words: "ec" and "ā" from SK 2370, "before ṇi" from SK 2569, "from the hetu" from SK 2593. There "fear" stood for wonder too; here only wonder is meant. The word nitya, "invariably", stops "optionally" from carrying on from SK 2509.
Example.
vi + smi + ṇic, wonder from the prompter → vṛddhi → smai-i
→ this rule: ā → smā-i
→ puk → smāp-i
→ ātmanepada by SK 2594 → vismāpayate
→ jaṭilo vismāpayate, "the ascetic with matted locks astonishes [people]".
Applies when / blocked by. "If fear or wonder is from the hetu", Bhaṭṭoji recalls from SK 2594. So when an instrument causes it, these rules do not apply:
kuñcikayainaṃ bhāyayati, vismāyayati, "he frightens him, astonishes him, with a kuñcikā".
Here bhī → bhai-i → āy (SK 61) → bhāy-i → bhāyayati: no ā, no ṣuk, and the parasmaipada. One of Bhaṭṭoji's commentators glosses kuñcikā as a kind of hair-knot. Being an instrument, it is not a prompting agent.
Bhaṭṭoji then raises an objection from poetry. How, then, does one account for vismāpayan vismitam ātmavṛttau, "[the lion,] astonishing [the king], already astonished at what had befallen him"? There the wonder comes from an instrument, as the words manuṣyavācā, "by human speech", show. One of Bhaṭṭoji's commentators identifies the verse as from the Raghuvaṃśa. Bhaṭṭoji adds: "otherwise śānac too would come". One of his commentators explains. If the speech were taken as the source of the wonder, SK 2594 would give the ātmanepada, and its participle affix is śānac, not śatṛ.
Bhaṭṭoji's reply has two parts.
vismāpayan has the participle affix śatṛ after a second ṇi, added to the ṇic-final.One of Bhaṭṭoji's commentators explains why the second option takes the parasmaipada. SK 2594 gives the ātmanepada only after a ṇic-stem whose base is bhī or smi. Smāpi + a second ṇi has a ṇic-stem as its base. How the word prayojya-kartrī is to be analysed, scholars discuss this.
Watch out. The four rules SK 2593–2596 all hang on "from the hetu". Change the source of the fear or wonder to an instrument, and all four fall away together.
adhi + iṅ + ṇic → adhi-ai-i → (ā, 2600) adhi-ā-i → (puk) adhyāpayati; luṅ adhyajīgapat (gāṅ, 2601) / adhyāpipatThis group changes particular roots before ṇi. SK 2597 to SK 2599 replace a final consonant: v for sphāy, t for śad, p for ruh. SK 2600, SK 2602 and SK 2603 put ā for the ec of krī, iṅ, ji, sidh and vī, each under its own condition. SK 2601 replaces iṅ "to study" by gāṅ in the aorist. The running example is adhi + iṅ, "to study": causative adhyāpayati "he teaches", aorist adhyajīgapat or adhyāpipat.
Rule. Before ṇi, v replaces the final of sphāy. Bhaṭṭoji's whole gloss is "ṇau", "when ṇi follows". One of his commentators says ṇau completes the sūtra, carried from SK 2570. It replaces the final sound (by the general rule SK 42; Bhaṭṭoji and his commentators do not say so here).
Example.
sphāy + ṇic → this rule: v for the final y → sphāv-i
→ sphāvayati.
Watch out. The root keeps its ā. Only the y changes.
Rule. Before ṇi, t replaces the final of śad, except in the sense of going. Bhaṭṭoji's gloss: "let t be the substitute for the final (antādeśa) of śad before ṇi, but not in the sense of going". One of his commentators adds that the a of taḥ is only for pronunciation: the substitute is t.
Example.
śad + ṇic → this rule: t for d → śat-i
→ vṛddhi of the penultimate a (SK 2282) → śāt-i
→ śātayati.
Applies when / blocked by. In the sense of going there is no t: gāḥ śādayati govindaḥ, "Govinda drives the cows". The meaning, Bhaṭṭoji says, is "makes them go" (gamayati).
Watch out. The same root gives śātayati and śādayati. The sense decides.
Rule. Before ṇi, p optionally replaces the final of ruh. That it replaces the final sound follows from the general rule SK 42; Bhaṭṭoji and his commentators do not say so here. Bhaṭṭoji's gloss again is "ṇau", completing the sūtra.
Example.
ruh + ṇic → this rule: p for h → rup-i → guṇa of the light penultimate (SK 2189) → rop-i → ropayati
→ without it: ruh-i → guṇa → rohayati.
Watch out. Anyatarasyām means "optionally". Both ropayati and rohayati are correct.
Rule. Before ṇi, ā replaces the ec of krī, iṅ and ji. Bhaṭṭoji's gloss: "let ā be substituted for the ec of these before ṇi". "Ec" and "ā" are carried from SK 2370, one of Bhaṭṭoji's commentators says. Bhaṭṭoji's commentators identify the roots: ḍukrīñ "to buy", iṅ "to study", ji "to conquer".
Example.
adhi + iṅ + ṇic → vṛddhi (SK 254) → adhi ai-i
→ this rule: ā for the ec → adhi ā-i
→ puk (SK 2570), the root now ending in ā → adhi āp-i
→ y for the i of adhi before a vowel (SK 47) → adhyāpayati, "he teaches".
Likewise krāpayati (krī) and jāpayati (ji). One of Bhaṭṭoji's commentators lists the steps: ā, then puk. Bhaṭṭoji's commentators give the aorists acikrapat and ajījapat.
Applies when / blocked by. Bhaṭṭoji does not say which ec is meant. The roots end in ī and i, which are not *ec*s. The derivation above takes the ec as the ai made by vṛddhi before ṇi.
Watch out. After the ā, puk follows automatically, as for any root ending in ā. Do not forget the p.
Rule. When a ṇi followed by san or by caṅ is intended, gāṅ optionally replaces iṅ. Bhaṭṭoji's gloss: "when a ṇi followed by san, or a ṇi followed by caṅ, [is intended], let gāṅ optionally be substituted for iṅ".
These are the same words as SK 2579 (A 6.1.31), but a different sūtra, A 2.4.51. One of Bhaṭṭoji's commentators gives the carried words: "for iṅ" from SK 2616, taught later; "gāṅ" from SK 2459; "optionally" from SK 2460. He reads the locative ṇau as one of domain (viṣayasaptamī): "when ṇi is intended", not "when ṇi is present".
Example.
With gāṅ:
adhi + iṅ, ṇi + caṅ intended → this rule: gā for i → adhi gā i a t
→ puk (SK 2570) → adhi gāp i a t
→ the ṇi is deleted (SK 2313), the penultimate shortened (SK 2314) → adhi gap a t
→ doubling; the abhyāsa keeps g (SK 2179) → adhi ga gap a t
→ sanvadbhāva: i (SK 2317), lengthened (SK 2318) → adhi gī gap a t
→ j for the abhyāsa's g (SK 2245) → adhi jī gap a t
→ the augment aṭ, and y for the i of adhi (SK 47) → adhyajīgapat.
Without gāṅ:
adhi + iṅ + ṇic → ā (SK 2600) and puk → adhi āp i a t
→ the shortening (SK 2314) comes before doubling, as Bhaṭṭoji says under SK 2580 → adhi ap i a t
→ the root begins with a vowel, so the second syllable pi, with the i of ṇic, is doubled (SK 2176) → adhi a pi pi a t
→ the ṇi of the second member is deleted (SK 2313) → adhi apip a t
→ the augment āṭ (SK 2254), and ā + a → ā (SK 269) → adhi āpip a t
→ y for the i of adhi (SK 47) → adhyāpipat.
Bhaṭṭoji's commentators both say that here pi is doubled.
Applies when / blocked by. An objection: "while doubling is due, no vowel-substitute is made before ṇi". Should gāṅ not wait until after the doubling? Bhaṭṭoji's commentators answer. That bar holds only where the part after the abhyāsa has an a. If gāṅ waited, only the ṇic would be doubled, by SK 2176. Then no a would stand after the abhyāsa, either during the derivation or in the result. So gāṅ comes first.
Watch out. Gāṅ changes the whole root, so in adhyajīgapat the abhyāsa is made from gāp, not from i. In adhyāpipat the root stays vowel-initial, and SK 2176 doubles the second syllable.
Rule. Before ṇi, ā replaces the ec of sidh (sidhyati), when the sense does not concern the other world. Bhaṭṭoji's gloss: "let ā be substituted for the ec of sidhyati, when it stands in a sense concerning this world, before ṇi".
One of Bhaṭṭoji's commentators gives the carried words: "ec" and "ā" from SK 2370, "before ṇi" from SK 2600. The other notes that the root is cited with its class-sign śyan (sidhyati), so the bhvādi sidh is not meant.
Example.
sidh + ṇic → guṇa of the light penultimate (SK 2189) → sedh-i
→ this rule: ā for the ec e → sādh-i
→ annaṃ sādhayati, "he produces the food".
Applies when / blocked by. Why "not concerning the other world"? Bhaṭṭoji: tāpasaḥ siddhyati, "the ascetic ascertains the truth". Austerity prompts him to it: sedhayati tāpasaṃ tapaḥ, "austerity makes the ascetic ascertain [the truth]". Here there is guṇa and no ā. Bhaṭṭoji's commentators explain that ascertaining the truth of the self is of use in the other world. One of them adds that producing food serves bodily satisfaction, in this world.
Watch out. The rule depends on the sense of the sentence, not on the form. The same stem gives sādhayati and sedhayati.
Rule. Before ṇi, ā optionally replaces the ec of vī, in the sense of conceiving (prajana). Bhaṭṭoji's gloss: "let ā optionally be substituted for its ec before ṇi, in the sense of conceiving". One of Bhaṭṭoji's commentators gives the carried words: "ec" and "ā" from SK 2370, "before ṇi" from SK 2569, "optionally" from SK 2509.
Example.
vī + ṇic → vṛddhi (SK 254) → vai-i
→ this rule, optionally: ā → vā-i → puk → vāp-i → vāpayati
→ without ā: āy (SK 61) → vāy-i → vāyayati.
vāpayati vāyayati vā gāḥ purovātaḥ, "the east wind makes the cows conceive". Bhaṭṭoji glosses: "makes them take an embryo". One of his commentators adds that cows are known to conceive in the season of the east wind.
Applies when / blocked by. Which vī is meant? Bhaṭṭoji's commentators agree it is the adādi vī. Whether vyeñ is also covered, scholars discuss this.
Recalled: SK 2364, ūd upadhāyā gohaḥ. Ū replaces the penultimate of guh:
guh + ṇic → this rule gives ū instead of guṇa → gūh-i → gūhayati.
One of Bhaṭṭoji's commentators calls the ū an exception to guṇa of a light penultimate. The recall leads into SK 2604.
Watch out. The rules simply say "the ec". In the derivations of this chapter, that ec is the guṇa vowel e for sidh (SK 2602), and the vṛddhi vowel ai for krī, iṅ, ji, bhī, smi and vī, and for lī under SK 2592.
duṣ + ṇic → dūṣ-i (ū, not guṇa) → dūṣayati; pra + an-i + caṅ → ('ni' doubled) → prāṇiṇat; iṇ + ṇic → gam-i → gamayatiThe last four sūtras deal with three roots. SK 2604 and SK 2605 give duṣ "to be spoiled" a long ū before ṇi, obligatory or optional by sense. Under SK 2605 Bhaṭṭoji recalls the short-vowelled (mit) roots and adds a vārttika on rañj. SK 2606 brings ṇ into both halves of the doubled an "to breathe". SK 2607 replaces iṇ "to go" by gam. Two vārttikas follow it, and Bhaṭṭoji closes with a remark on causatives "from which the prompting has ceased". The running examples are dūṣayati, prāṇiṇat and gamayati.
Rule. Before ṇi, ū replaces the penultimate of duṣ (duṣyati). Bhaṭṭoji's gloss: "let ū be substituted for the penultimate of duṣyati before ṇi". "Ū" and "of the penultimate" run on from SK 2364, one of Bhaṭṭoji's commentators says.
Bhaṭṭoji opens with a remark: "duṣaḥ would have been easy to say". The sūtra cites the root as doṣ, its guṇa form. One of his commentators explains the point of that. The rule applies only where guṇa of the light penultimate is due; it is an exception to that guṇa.
Example.
duṣ + ṇic → guṇa of the light penultimate (SK 2189) would apply
→ this rule, its exception, gives ū for the penultimate u instead → dūṣ-i
→ dūṣayati.
Applies when / blocked by. One of Bhaṭṭoji's commentators gives the counter-example for "before ṇi": doṣaḥ, with plain guṇa.
Watch out. The rule is an exception to guṇa, like SK 2364 for guh (gūhayati). Both give a long ū where o would be expected.
Rule. The ū of SK 2604 is optional when the mind (citta) is estranged (virāga). Bhaṭṭoji glosses virāga as "absence of affection" (aprītatā). "Ū", "of the penultimate" and "doṣo ṇau" are carried forward; one of Bhaṭṭoji's commentators gives this.
Example.
duṣ + ṇic, sense "estranging the mind"
→ with this rule: dūṣ-i → dūṣayati
→ without it: guṇa (SK 2189) → doṣ-i → doṣayati.
cittaṃ dūṣayati doṣayati vā kāmaḥ, "desire estranges the mind". Bhaṭṭoji's commentators explain: the mind turns away from the daily rites, such as bathing and the twilight worship, and desire prompts it.
Recalled: SK 2568, mitāṃ hrasvaḥ. The mit roots keep a short penultimate before ṇi. Bhaṭṭoji: "the mit roots have been stated in the bhvādi and the curādi".
ghaṭ + ṇic → vṛddhi (SK 2282) → ghāṭ-i → mit, so shortened (SK 2568) → ghaṭayati.jan + ṇic → vṛddhi (SK 2282) → jān-i → shortened (SK 2568) → jan-i → janayati;
jṝ + ṇic → vṛddhi (SK 254) → jār-i → shortened (SK 2568) → jar-i → jarayati.
jārayati, with the vṛddhi kept. One of Bhaṭṭoji's commentators explains: the jṝ of the ninth class is not marked with ṣ, so it is not mit.Vārttika: rañjer ṇau mṛgaramaṇe nalopo vaktavyaḥ. "For rañj before ṇi, loss of n should be stated in the sense of hunting (mṛgaramaṇa)." Bhaṭṭoji glosses mṛgaramaṇa as the chase (ākheṭaka).
rañj + ṇic, hunting → this vārttika: the n is lost → raj-i
→ vṛddhi (SK 2282) → rāj-i
→ shortened (SK 2568) → raj-i → rajayati mṛgān, "he hunts the deer".
Why shortened? Rañj is among the mit roots of the list-rule above. That reason is drawn from the list itself; Bhaṭṭoji and his commentators do not state it here.
Why the vārttika? One of Bhaṭṭoji's commentators answers. SK 415 deletes a penultimate n only before an affix marked with k or ṅ. Ṇi is neither, so that rule does not reach here.
Bhaṭṭoji's two counter-examples:
rañjayati pakṣiṇaḥ, of birds, with the n kept;rañjayati mṛgāṃs tṛṇadānena, "he draws the deer by giving them grass". One of his commentators adds that this is done in order to keep them.Recalled: ci and sphur. Among the curādi roots, ciñ belongs to the jñapādi, which are mit. With cisphuror ṇau (SK 2569): capayati, cayayati, "as has been stated" in chapter 53. "But of cinoti", the ci of the fifth class, which is not mit:
ci + ṇic → ā (SK 2569) → cā-i → puk → cāp-i → no shortening → cāpayati. SK 2569 puts ā for the root vowel itself, as chapter 53 taught, so no vṛddhi step is shown;cāy-i → cāyayati.Bhaṭṭoji's commentators explain the difference: one root is mit, the other is not.
Sphur: sphārayati (with ā by SK 2569), sphorayati (with guṇa). In the aorist:
sphur + ṇic + luṅ → sphur i a t
→ doubling comes before the vowel-substitute, so sphu is doubled with its u → sphu sphur i a t
→ the abhyāsa keeps ph (SK 2259), which becomes p (SK 2182) → pu sphur i a t
→ in the second member, ā optionally (SK 2569)
→ with ā: pu sphār i a t → the ṇi is deleted, the ā shortened (SK 2314) → pu sphar a t → apuspharat
→ without ā: guṇa, the ṇi deleted, the o shortened → pu sphur a t → apusphurat
In both, s follows the abhyāsa's u. S is not a labial, so SK 2577 gives no i. And the abhyāsa stands before sph, so it is not lengthened.
Bhaṭṭoji's commentators derive both forms this way. One of them adds that the vowel-substitute is either barred while doubling is due, or treated like the original.
Watch out. Two kinds of roots look alike here: mit ones end up with a short vowel (ghaṭayati, capayati), others keep their vṛddhi (jārayati, cāyayati). Check the class before forming the causative.
Rule. Both n's of an (aniti), together with its abhyāsa, become ṇ when the condition is present. Bhaṭṭoji's gloss: "both n's of an together with its abhyāsa undergo ṇatva when the conditioning [sound] is present". One of Bhaṭṭoji's commentators explains the condition: an r standing in an upasarga. He gives the carried words: "of an" (aniteḥ) from SK 2478, "after an upasarga" from SK 2287, under the heading SK 235.
Example.
pra + an + ṇic + luṅ → vṛddhi (SK 2282) → ān i a t
→ the penultimate is shortened (SK 2314) → an i a t
→ the ṇ that SK 2478 would bring counts as not yet in effect (asiddha) for doubling, so the second syllable, ni, is doubled (SK 2176) → a ni ni a t
→ the ṇi of the second member is deleted (SK 2313) → anin a t
→ the augment āṭ and vṛddhi (SK 2254, 269) → ānin a t
→ pra + ā → prā (SK 85) → prāninat
→ now the second n is separated from the r of pra by the abhyāsa's n. SK 2478 alone would not reach it
→ this rule: ṇ for both n's → prāṇiṇat.
This is the derivation one of Bhaṭṭoji's commentators sets out.
Applies when / blocked by. A maxim, pūrvatrāsiddhīyam advirvacane, says that the "not in effect" status does not hold for doubling. If it held here, ṇ would come first, be doubled, and give prāṇiṇat with no need for this rule. Bhaṭṭoji's commentators conclude: this very rule shows that the maxim is not invariable (anitya). Hence in ūrṇunāva nu is doubled before the ṇ, as you saw in chapter 45.
Why the word ubhau, "both"? One of Bhaṭṭoji's commentators explains, citing Kaiyaṭa. Without it, the rule might give ṇ only to the separated n, not to both together. He also gives the counter-example for "together with its abhyāsa": prāṇ namati.
Watch out. After doubling, the abhyāsa's n stands between the r and the second n. The second ṇ of prāṇiṇat is what this rule secures.
Rule. Before ṇi, gam replaces iṇ "to go", except in the sense of making known (bodhana). Bhaṭṭoji's gloss: "let gam be substituted for iṇ before ṇi". "For iṇ" is carried from SK 2458, one of Bhaṭṭoji's commentators says; the i of gami is only for pronunciation.
Example.
iṇ + ṇic → this rule: gam for i → gam-i
→ vṛddhi (SK 2282) → gām-i
→ shortened (SK 2568, chapter 53), since roots ending in am are mit by the list-rule of chapter 44 → gam-i
→ gamayati, "he makes [him] reach".
Bhaṭṭoji and his commentators do not explain the short a here; the reason is drawn from that list-rule.
Applies when / blocked by. In the sense of making known, there is no gam: pratyāyayati, "he makes [it] known".
prati + i + ṇic → vṛddhi (SK 254) → ai-i → āy (SK 61) → āy-i → y for the i of prati (SK 47) → pratyāyayati.
Bhaṭṭoji's commentators give its aorist, pratyāyiyat:
prati + i + ṇic + luṅ → i i a t
→ vṛddhi and āy, being the later rule (para), set aside iṇo yaṇ (SK 2455) → āy i a t
→ the shortening (SK 2314) comes before doubling, as Bhaṭṭoji says under SK 2580 → ay i a t
→ the second syllable, yi, is doubled (SK 2176) → a yi yi a t
→ the ṇi of the second member is deleted (SK 2313) → ayiy a t
→ the augment āṭ, and ā + a → ā (SK 2254, 269) → āyiy a t
→ y for the i of prati (SK 47) → pratyāyiyat.
Should vṛddhi not wait until after the doubling? No, they answer: if it waited, the ṇic would be doubled, and no a would follow the abhyāsa. The bar does not apply.
Vārttika: iṇvad ikaḥ. "Ik is like iṇ." Ik "to remember" also takes gam before ṇi: adhigamayati, "he reminds". One of Bhaṭṭoji's commentators glosses it so.
Recalled: SK 2574 and SK 358. han + ṇic → t for the final n (SK 2574) → hat-i → vṛddhi (SK 2282) → hāt-i → gh for h (SK 358) → ghātayati. One of Bhaṭṭoji's commentators adds the aorist ajīghatat. Bhaṭṭoji then gives īrṣyayati, the causative of īrṣy.
Vārttika: īrṣyates tṛtīyasyeti vaktavyam. "For īrṣy, '[doubling] of the third' should be stated." Bhaṭṭoji gives two meanings: "of the third consonant", or "of the third single-vowelled unit (ekāc)". One of his commentators says the Bhāṣya on SK 2446 makes both clear; the other says both rest on the source works.
On the first meaning, the rule stops the ṣ from being doubled; the r is already barred by SK 2446. Bhaṭṭoji and the Tattvabodhinī give the form airṣyiyat, but neither derives it. The Bālamanoramā doubles y alone and reads airṣyyat. Scholars discuss which is right.
Bhaṭṭoji's form results if y is doubled together with the i of ṇic. This route is the chapter's own, not a commentator's:
īrṣy + ṇic + luṅ → īrṣy i a t
→ r is not doubled (SK 2446); ṣ is not doubled (this vārttika); the third consonant, y, is doubled with the i of ṇic → īrṣ yi yi a t
→ the ṇi of the second member is deleted → īrṣyiy a t
→ the augment āṭ and vṛddhi (SK 2254, 269) → airṣyiyat.
On the second meaning, the vārttika works only in the desiderative (sannanta), as an exception to SK 2176. With iṭ, īrṣy i sa has three *ekāc*s (īrṣ, ṣyis, sa), and the third, sa, is doubled: īrṣyiṣiṣati (one of Bhaṭṭoji's commentators). Bhaṭṭoji gives both desiderative forms under SK 2608 (chapter 55). With caṅ there is no third ekāc: īrṣ is the first and ṣyi the second. So SK 2176 doubles the second, ṣyi:
īrṣy i a t → īr ṣyi ṣyi a t
→ only the first consonant stays in the abhyāsa (SK 2179), Bhaṭṭoji says → īr ṣi ṣyi a t
→ the ṇi of the second member is deleted → īrṣiṣy a t
→ āṭ and vṛddhi → airṣiṣyat.
Nivṛttapreṣaṇa. Bhaṭṭoji ends with a remark. "From a dhātu from which the prompting has ceased (nivṛttapreṣaṇa), with the hetumaṇṇic, the meaning is the same as with the plain dhātu." This accounts for the line prārthayanti śayanotthitaṃ priyāḥ, "the beloveds entreat [him], risen from his bed". Bhaṭṭoji's commentators place it in Māgha's poem. "The same is to be worked out for all transitive (sakarmaka) dhātus."
Both of Bhaṭṭoji's commentators set out the problem. Arth is an āgarvīya root of chapter 53; with its own curādi ṇic it would take the ātmanepada. The Bālamanoramā adds that the causative ṇic cannot be meant either, since no one is prompted. It explains Bhaṭṭoji's answer: the causative is formed by taking the prompting as something that once was, and the sense is the plain one. Both commentators quote a half-verse to this effect. The Bālamanoramā adds that the point is clear in the Bhāṣya on SK 2738 (ṇer aṇau). The Tattvabodhinī rejects a different explanation, which takes prārtha as a noun.
Here ends the ṇyanta-prakriyā, the section on the ṇic-final.
Watch out. Gam replaces iṇ only before ṇi, and only outside "making known". gamayati and pratyāyayati come from the same root.
vājayati (SK 2590).| SK | What it does | Example |
|---|---|---|
| 2575 | the prompter of the agent is named hetu, and also kartṛ | Yajñadatta in "makes Devadatta cook" |
| 2576 | ṇic after a root when the prompter's activity is expressed | bhāvayati, bhāvayate |
| 2577 | i for the abhyāsa's u before a labial, yaṇ or j + a (before san, so via sanvadbhāva) | abībhavat, apīpavat |
| 2578 | the same, optional, for sru, śru, dru, pru, plu, cyu; recalls SK 2572 and ṇic on a ṇic-stem | asisravat / asusravat; aśaśāsat; corayati, acūcurat |
| 2579 | optional saṃprasāraṇa for śvi before ṇi + san/caṅ | aśūśavat / aśiśvayat |
| 2584 | saṃprasāraṇa for svāpi before caṅ | asūṣupat |
| 2586 | saṃprasāraṇa for hvā before ṇi + san/caṅ | ajūhavat / ajuhāvat |
| 2580 | no upasarga-ṣ for stambh, siv, sah before caṅ; the vowel-initial roots | avātastambhat, paryasīṣivat; āṭiṭat, mā bhavān ididhat, aubjijat |
| 2581 | num for rabh before a vowel, not before śap or liṭ | rambh-i |
| 2582 | num for labh too; recalls SK 2531, 2566 | ararambhat, alalambhat; ajīhayat, asasmarat |
| 2583 | optional a in the abhyāsa of veṣṭ, ceṣṭ; vārttika kāṇyādīnāṃ vā | avaveṣṭat / aviveṣṭat; acīkaṇat / acakāṇat |
| 2585 | yuk, not puk, for śā, chā, sā, hvā, vyā, ve, pā | śāyayati, hvāyayati |
| 2590 | juk for vā in the sense of shaking | vājayati; but keśān vāpayati |
| 2591 | optional nuk for lī, luk for lā, melting a fat; vārttika: invariable ā for lī in three senses | vilīnayati, vilāyayati, vilālayati, vilāpayati |
| 2587 | pā: penultimate deleted, ī for the abhyāsa's final; recalls puk | apīpyat; arpayati … sthāpayati |
| 2588 | i for the penultimate of sthā before ṇi + caṅ | atiṣṭhipat |
| 2589 | the same, optional, for ghrā; recalls ur ṛt; vārttika: luk augment for pā "protect" | ajighripat / ajighrapat; acīkṛtat / acikīrtat; pālayati |
| 2592 | ātmanepada for ṇic-final lī in honouring, overpowering, deceiving, even without the agent's fruit | jaṭābhir lāpayate |
| 2593 | optional ā for the ec of bhī, fear from the hetu | bhāpayate (with SK 2594) |
| 2594 | ātmanepada for ṇic-final bhī, smi, fear or wonder from the hetu | muṇḍo bhāpayate |
| 2595 | ṣuk for bhī ending in ī, fear from the hetu | bhīṣayate |
| 2596 | invariable ā for smi, wonder from the hetu; none of this with an instrument | jaṭilo vismāpayate; kuñcikayainaṃ bhāyayati |
| 2597 | v for the final of sphāy before ṇi | sphāvayati |
| 2598 | t for the final of śad, except in going | śātayati; gāḥ śādayati |
| 2599 | optional p for the final of ruh | ropayati / rohayati |
| 2600 | ā for the ec of krī, iṅ, ji before ṇi | krāpayati, adhyāpayati, jāpayati |
| 2601 | optional gāṅ for iṅ when ṇi + san/caṅ is intended | adhyajīgapat / adhyāpipat |
| 2602 | ā for the ec of sidh, not in an other-worldly sense | annaṃ sādhayati; sedhayati tāpasaṃ tapaḥ |
| 2603 | optional ā for the ec of vī, in conceiving; recalls SK 2364 | vāpayati / vāyayati; gūhayati |
| 2604 | ū for the penultimate of duṣ before ṇi | dūṣayati |
| 2605 | the same, optional, when the mind is estranged; recalls mit, ci, sphur; vārttika on rañj | dūṣayati / doṣayati; rajayati mṛgān; apusphurat / apuspharat |
| 2606 | ṇ for both n's of an with its abhyāsa | prāṇiṇat |
| 2607 | gam for iṇ before ṇi, except in making known; vārttikas on ik and īrṣy; nivṛttapreṣaṇa | gamayati, pratyāyayati; adhigamayati; airṣyiyat / airṣiṣyat |