PREMABASESiddhānta-kaumudī
Learner edition · Prakaraṇa 55

55. Tiṅanta san-prakaraṇa — Verbs: the desiderative, san in the sense of wishing

This chapter teaches the desiderative: "he wishes to read", "he wishes to eat". Pāṇini forms it with the affix san, which you met in chapter 44 in the root's own sense (jugupsate). Here san adds the sense of wishing. Bhaṭṭoji calls the section the sannanta-prakriyā, "the working of what ends in san". After this chapter you will be able to build pipaṭhiṣati "he wishes to read" from paṭh, step by step. You will know when san takes the augment iṭ and when it does not (bubhūṣati, pipṛcchiṣati, dudyūṣati / dideviṣati). You will know when san is treated as marked with k, so that the root vowel takes no guṇa (rurudiṣati, bibhitsati, didyutiṣate / didyotiṣate). You will lengthen the stem before san (cicīṣati, jighāṃsati), replace iṇ "to go" and iṅ "to study" by gam (jigamiṣati, adhijigāṃsate), and drop the reduplicated syllable in a short list of roots (īpsati, pitsati, ditsati, mokṣate). Last, you will learn which roots keep s after the reduplicated syllable and which take ṣ (tuṣṭūṣati, but sisikṣati, sisvedayiṣati). The chapter has twenty-one sūtras, SK 2608 to SK 2628.

Before you start

A. The desiderative: san in the sense of wishing — running example: paṭhitum icchati: paṭh + san → paṭh-iṣa (iṭ) → pa-paṭh-iṣa (doubling) → pi-paṭh-iṣa (sanyataḥ) → pipaṭhiṣati

One sūtra founds the desiderative. SK 2608 adds san after a root when the action of that root is what someone wishes, and the one who wishes is the one who will act. The running example is paṭh "to read". Paṭhitum icchati, "he wishes to read", becomes pipaṭhiṣati.

SK 2608 · धातोः कर्मणः समानकर्तृकादिच्छायां वा · dhātoḥ karmaṇaḥ samānakartṛkādicchāyāṃ vā (A 3.1.7)

धातोः कर्मणः समानकर्तृकादिच्छायां वा
dhātoḥ karmaṇaḥ samānakartṛkādicchāyāṃ vā · A 3.1.7
इषिकर्मणः इषिणैककर्तृकाद्धातोः सन्प्रत्ययो वा स्यादिच्छायाम् । धातोर्विहितत्वादिह सन आर्धधातुकत्वम् । इट् । द्वित्वम् । सन्यतः SK 2317 । पठितुमिच्छति पिपठिषति । कर्मणः किम् । गमनेनेच्छतीति करणान्मा भूत् । समानकर्तृकात्किम् । शिष्याः पठन्त्वितीच्छति गुरुः । वाग्रहणात्पक्षे वाक्यमपि । लुङ्सनोर्घसॢ SK 2427 एकाच उपदेशे - SK 2246 इति नेट् । सस्य तत्वम् । अत्तुमिच्छति जिघत्सति ॥ ईर्ष्यतेस्तृतीयस्येति यिसनोर्द्वित्वम् । ईर्ष्यियिषति । ईर्ष्यिषिषति ॥

Rule. The affix san comes optionally, in the sense of wishing (icchā), after a dhātu. Two conditions hold. The dhātu's action must be the object (karman) of the wishing. And it must have the same agent (samānakartṛka) as the wishing. Bhaṭṭoji's gloss: "let the affix san optionally come, in the sense of wishing, after a dhātu that is the object of iṣ 'to wish' and has one and the same agent as iṣ".

The word san is carried forward (anuvṛtti) from guptijkidbhyaḥ san (SK 2393), one of Bhaṭṭoji's commentators notes. Both commentators say that "object" and "same agent" are measured against the wishing alone, since the sūtra names the wishing.

Bhaṭṭoji adds the key point. This san is prescribed "after a dhātu" (dhātoḥ). So it is ārdhadhātuka, and so it takes iṭ. One of his commentators explains why the word dhātoḥ matters. Only an affix taught with that word is ārdhadhātuka. The san of jugupsate (SK 2393) is not taught so. If it were ārdhadhātuka, its light penultimate would take guṇa.

Example.

paṭhitum icchati, "he wishes to read":

paṭh + san (this rule); n is a marker → paṭh + sa

→ san is ārdhadhātuka and begins with s, a val consonant; it takes iṭ (SK 2184) → paṭh + i + sa

→ s after i becomes ṣ (SK 212) → paṭh-iṣa

→ the form ends in san, so it is doubled (SK 2395) → paṭh paṭhiṣa

→ only the first consonant stays in the abhyāsa (SK 2179) → pa paṭhiṣa

→ a of the abhyāsa becomes i before san (SK 2317, sanyataḥ) → pi paṭhiṣa

→ pipaṭhiṣa ends in san, so it is a dhātu (SK 2304)

→ laṭ, the ending tip, the class-sign śap → pipaṭhiṣa + a + ti

→ a + a gives one a (pararūpa, SK 191) → pipaṭhiṣati.

One of Bhaṭṭoji's commentators gives the sense: "he wishes reading done by himself". The san shows that the reading is what is wished, and that the reader is the one who wishes.

Applies when / blocked by. Bhaṭṭoji tests each word of the sūtra.

Example 2: jighatsati. attum icchati, "he wishes to eat". Bhaṭṭoji recalls two rules from earlier chapters.

ad + sa

→ ghas replaces ad before san (SK 2427, luṅsanor ghasḷ) → ghas + sa

→ no iṭ: ghas is a one-vowelled root that refuses it (SK 2246) → ghas + sa

→ the s of ghas becomes t before an ārdhadhātuka in s (SK 2342) → ghat + sa

→ doubling (SK 2395) → ghat ghatsa → the first consonant stays (SK 2179) → gha ghatsa

→ in the abhyāsa gh becomes a palatal (SK 2245) and then unaspirated (SK 2182) → ja ghatsa

→ i for a (SK 2317) → jighatsa → jighatsati.

One of Bhaṭṭoji's commentators gives these steps, with the two changes in the abhyāsa.

Example 3: īrṣyiyiṣati, īrṣyiṣiṣati. From īrṣy "to envy". In chapter 54 you met the vārttika īrṣyates tṛtīyasya, "for īrṣy, [doubling] of the third". Bhaṭṭoji now applies it to san: "there is doubling of yi or of san". Both his commentators match each form to one reading of the vārttika.

How the first form should be spelled, scholars discuss this.

Watch out. Two *san*s now stand side by side. The san of SK 2393 (jugupsate) is in the root's own sense and is not ārdhadhātuka. The san of this rule is the desiderative, and it is ārdhadhātuka. A third use, san in the sense of apprehension, comes as a vārttika under SK 2622. Whether a desiderative san can follow another san is settled at the end of the chapter, under SK 2628.

B. When san is kit: no guṇa of the stem — running example: rud + san → rud-iṣa (kit, 2609: no guṇa) → ru-rud-iṣa → rurudiṣati; bhū + sa → bhū-ṣa (kit, 2612) → bubhūṣati; bhid + sa → bibhitsati (2613); optional didyutiṣate / didyotiṣate (2617)

Normally san, being ārdhadhātuka, brings guṇa to the root: a light penultimate i, u, ṛ or a final ik vowel is raised. Four sūtras stop this by treating san as kit, "marked with k". An affix that is kit blocks guṇa (SK 2217). SK 2609 makes san kit after six named roots, even with iṭ. SK 2612 and SK 2613 make it kit when it begins with a consonant of the jhal class, after certain roots. SK 2617 makes a san with iṭ optionally kit after another kind of root. SK 2617 comes later in Bhaṭṭoji's order; it is placed here with its kind. The running examples are rurudiṣati "he wishes to weep", bubhūṣati "he wishes to be", and bibhitsati "he wishes to split".

SK 2609 · रुदविदमुषग्रहिस्वपिप्रच्छः सँश्च · rudavidamuṣagrahisvapipracchaḥ sam̐śca (A 1.2.8)

रुदविदमुषग्रहिस्वपिप्रच्छः सँश्च
rudavidamuṣagrahisvapipracchaḥ sam̐śca · A 1.2.8
एभ्यः सन् क्त्वा च कितौ स्तः । रुरुदिषति । विविदिषति । मुमुषिषति ॥

Rule. After rud "to weep", vid "to know", muṣ "to steal", grah "to seize", svap "to sleep" and pracch "to ask", the affix san is kit. So is the affix ktvā, the -tvā of the gerund, taught later. Bhaṭṭoji's gloss: "after these, san and ktvā are kit".

One of Bhaṭṭoji's commentators gives the carried words. The word ca, "and", gathers in ktvā from the preceding sūtra, mṛḍamṛda… (SK 3323, taught later). The word kit is carried forward from asaṃyogāl liṭ kit (SK 2242).

Example.

rud + san → iṭ (SK 2184) → rud + i + sa → ṣ (SK 212) → rud-iṣa

→ guṇa of the light penultimate u (SK 2189) would give rod-; but san is kit (this rule), so guṇa is blocked (SK 2217) → rud-iṣa

→ doubling (SK 2395) → rud rudiṣa → first consonant only (SK 2179) → ru rudiṣa

→ sanyataḥ changes only an a, so the u stays → rurudiṣa → rurudiṣati.

Likewise vividiṣati "he wishes to know" and mumuṣiṣati "he wishes to steal".

Applies when / blocked by. Why name these six? Bhaṭṭoji's commentators agree on the purpose, root by root.

The san forms of grah, svap and pracch need further rules. They come in group C: jighṛkṣati, suṣupsati (SK 2610) and pipṛcchiṣati (SK 2611).

Which vid is meant, scholars discuss this. One of Bhaṭṭoji's commentators takes only vid "to know", since it is named beside rud.

Watch out. Here san is kit even though it begins with iṭ. The next two rules need a san that begins with a consonant. This one does not.

SK 2612 · इको झल् · iko jhal (A 1.2.9)

इको झल्
iko jhal · A 1.2.9
इगन्ताज्झलादिः सन् कित्स्यात् । बुभूषति । दीङ् । दातुमिच्छति दिदीषते । एज्विषयत्वाभावात् मीनातिमिनोति - SK 2508 इत्यात्वं न । अत एव सनि मीमा - SK 2623 इति सूत्रे माधातोः पृथङ्मीग्रहणं कृतम् ॥

Rule. After a dhātu ending in an ik vowel (i, u, ṛ, ḷ), a san that begins with a jhal consonant is kit. A jhal is any non-nasal stop, a sibilant, or h; the s of san is one. Bhaṭṭoji's gloss: "let a san beginning with a jhal be kit after [a dhātu] ending in ik".

The words san and kit are carried forward from SK 2609 and SK 2242. Bhaṭṭoji's commentators explain "ending in ik". The dhātu is implied by san, and ik qualifies it. So the rule covers a dhātu that ends in ik (tadantavidhi).

Example.

bhū + san → bhū ends in u, an uk vowel, so it takes no iṭ (SK 2610) → bhū + sa

→ san begins with s, a jhal, after an ik-final root → kit (this rule)

→ guṇa of the final ū (SK 2168) is blocked (SK 2217) → bhū-sa → ṣ (SK 212) → bhūṣa

→ doubling (SK 2395) → bhū bhūṣa → the abhyāsa vowel is shortened (SK 2180) → bhu bhūṣa

→ bh becomes b in the abhyāsa (SK 2182) → bubhūṣa → bubhūṣati "he wishes to be".

Example 2: didīṣate. From dīṅ "to waste away". Dātum icchati, "he wishes to waste away".

dī + sa (the form has no iṭ) → kit (this rule): no guṇa → dī-sa → ṣ → dīṣa

→ doubling and shortening → di dīṣa → didīṣa

→ dīṅ is marked with ṅ, so the desiderative takes the ātmanepada, by pūrvavat sanaḥ (SK 2734, taught later) → didīṣate.

Bhaṭṭoji draws a lesson. Dīṅ takes ā for its vowel by SK 2508, but only where an e would stand (the ej-viṣaya). Here san is kit, so no guṇa e can arise. Bhaṭṭoji's commentators gloss it so: "guṇa does not obtain, because of kittva". So the ā does not come.

"For this very reason", Bhaṭṭoji adds, SK 2623 names mī apart from mā. His commentators explain. If ā came even outside the domain of e, mī would become mā, and the mention of mā would cover it. The separate mention of mī would then be pointless.

Applies when / blocked by. One of Bhaṭṭoji's commentators gives counter-examples.

Watch out. Iṭ decides kittva. Whenever san takes iṭ, it begins with a vowel, and this rule cannot apply.

SK 2613 · हलन्ताच्च · halantācca (A 1.2.10)

हलन्ताच्च
halantācca · A 1.2.10
इक्समीपाद्धलः परो झलादि सन् कित्स्यात् । गुहू । जुघुक्षति । बिभित्सति । इकः किम् । यियक्षते । झल्किम् । विवर्धिषते । हल्ग्रहणं जातिपरम् । तृन्हू । तितृक्षति । तितृंहिषति ॥

Rule. A san beginning with a jhal is also kit after a consonant (hal) that stands near an ik vowel. Bhaṭṭoji's gloss: "let a san beginning with a jhal, that follows a hal standing near an ik, be kit".

The words ik and jhal are carried forward from SK 2612, san from SK 2609, and kit from SK 2242. Bhaṭṭoji's commentators explain the word anta. Usually it means "end". But a consonant cannot be the end of a vowel. So here anta means "near".

Example.

bhid + san (the form has no iṭ) → bhid + sa

→ d is a hal next to the ik vowel i; san begins with the jhal s → kit (this rule)

→ guṇa of the light penultimate (SK 2189) is blocked → bhid-sa

→ d becomes t before s (SK 121) → bhitsa

→ doubling, first consonant only → bhi bhitsa → bh → b (SK 2182) → bibhitsa → bibhitsati "he wishes to split".

Example 2: jughukṣati. From guh (guhū) "to hide".

guh + sa: guh takes no iṭ before san (SK 2610), though it is an ūdit root → kit (this rule): no guṇa

→ h → ḍh (SK 324) → g takes aspiration, gh (SK 326) → ḍh → k before s (SK 295) → s → ṣ after k (SK 212) → ghukṣa

→ doubling (SK 2395) → ghu ghukṣa → gh → palatal jh (SK 2245) → unaspirated j (SK 2182) → jughukṣati.

One of Bhaṭṭoji's commentators gives these steps.

Applies when / blocked by. Bhaṭṭoji tests the words.

Bhaṭṭoji then says: "the word hal refers to the class (jāti)". It means one or more consonants of the class, not just one. He shows why with tṛṃh (written tṛnhū) "to hurt".

In tṛṃh, the anusvāra and the h both stand between ṛ and san. Taken one by one, neither is "a hal near an ik" with san right after it. Taken as a class, the two together are. Tṛṃh is ūdit, so its iṭ is optional (SK 2279).

Without iṭ: tṛṃh + sa

→ kit (this rule, on the class reading)

→ the penultimate n is lost before a kit affix (SK 415) → tṛh-sa

→ ḍh, k, ṣ (SK 324, 295, 212) → tṛkṣa

→ doubling → tṛ tṛkṣa → a for ṛ in the abhyāsa (SK 2244) → i (SK 2317) → titṛkṣati.

With iṭ: tṛṃh-iṣa → not jhal-initial, so not kit, and the n stays → titṛṃhiṣati.

One of Bhaṭṭoji's commentators gives this reasoning; the other agrees on the loss of n.

Watch out. The class reading returns under SK 2621, with dambh: dhipsati.

SK 2617 · रलो व्युपधाद्धलादेः संश्च · ralo vyupadhāddhalādeḥ saṃśca (A 1.2.26)

रलो व्युपधाद्धलादेः संश्च
ralo vyupadhāddhalādeḥ saṃśca · A 1.2.26
उश्च इश्च वी ते उपधे यस्य तस्माद्धलादे रलन्तात्परौ क्त्वासनौ सेटौ वा कितौ स्तः । द्युतिस्वाप्योः संप्रसारणम् SK 2344 । दिद्युतिषते । दिद्योतिषते । रुरुचिषते । रुरोचिषते । लिलिखिषति । रलः किम् । दिदेविषति । व्युपधात्किम् । विवर्तिषते । हलादेः किम् । एषिषिषति । इह नित्यमपि द्वित्वं गुणेन बाध्यते । उपधाकार्यं हि द्वित्वात्प्रबलम् । ओणेर्ऋदित्करणस्य सामान्यापेक्षाज्ञापकत्वत् ॥

Rule. Take a dhātu that begins with a consonant (halādi), ends in a ral consonant, and has u or i as its penultimate (vyupadha). After it, san and ktvā that take iṭ (seṭ) are optionally kit. Ral is the pratyāhāra of every consonant except y and v. Bhaṭṭoji's gloss: "u and i are 'vī'. After [a dhātu] that begins with a hal and ends in ral, whose penultimate is one of these, ktvā and san with iṭ are optionally kit".

Bhaṭṭoji's commentators give four carried words. Ktvā is gathered by ca from pūṅaḥ ktvā ca (SK 3051). "With iṭ" comes from na ktvā seṭ (SK 3322). "Optionally" comes from nopadhāt thaphāntād vā (SK 3324). Kit comes from SK 2242. The first three are taught later.

Example.

dyut "to shine": dyut + iṣa

→ kit (this rule): no guṇa → dyut-iṣa

→ doubling → dyut dyutiṣa

→ the abhyāsa's y takes saṃprasāraṇa (SK 2344): diut, and i + u gives i → dit dyutiṣa

→ first consonant only (SK 2179) → di dyutiṣa

→ ātmanepada, as for the bare root (SK 2734, taught later) → didyutiṣate.

Not kit: guṇa of the light penultimate (SK 2189) → dyot-iṣa → didyotiṣate.

Likewise ruc "to shine": ruruciṣate / rurociṣate; likh "to write": lilikhiṣati / lilekhiṣati. In each pair the first form has a kit san.

Applies when / blocked by. Bhaṭṭoji tests the three conditions.

The other commentator asks why iṭ is a condition. His counter-examples are bhittvā, chittvā and bubhukṣate.

One of Bhaṭṭoji's commentators states the doubt that Bhaṭṭoji answers next, about eṣiṣiṣati. Doubling is nitya: it applies whether or not a rival applies first. If doubling came first, the i of iṣ would no longer be penultimate. Guṇa could never arise, and the word "beginning with a consonant" would do nothing. Bhaṭṭoji's answer: "Here doubling, though nitya, is set aside by guṇa. For an operation on the penultimate (upadhā) is stronger than doubling." His proof (jñāpaka) is that Pāṇini marks the root oṇ with ṛ. That sign, he says, speaks of the whole class of penultimate operations.

One of his commentators sets out the proof. With ṇic and the caṅ-aorist, the ṛ-marking forbids shortening of oṇ's penultimate (SK 2572). That prohibition would be pointless if doubling always came first. So shortening, a penultimate operation, must come before doubling. The same commentator notes that the Bhāṣya's proof speaks only of shortening. Whether it covers guṇa too, scholars discuss this.

Watch out. This option applies only to a san with iṭ. Without iṭ, SK 2612 and SK 2613 decide.

C. iṭ on san: none, obligatory, or optional — running example: grah + sa (no iṭ, 2610) → gṛh-sa (kit, saṃprasāraṇa) → ḍh, bhaṣbhāva, k, ṣ → jighṛkṣati; div + sa (no iṭ, 2618) → dudyūṣati / with iṭ dideviṣati

Being ārdhadhātuka, san takes iṭ by SK 2184, unless the root refuses it (SK 2246). Five sūtras adjust this for san. SK 2610 denies iṭ after grah, guh and roots ending in uk. SK 2611 and SK 2626 give iṭ to roots that would not otherwise have it, or would have it only optionally. SK 2618 and SK 2625 make iṭ optional for lists of roots. Bhaṭṭoji teaches SK 2618, SK 2625 and SK 2626 later; they are placed here with their kind. As group B showed, iṭ matters twice over: without it, san begins with a jhal and may be kit. The running examples are jighṛkṣati "he wishes to seize" and the pair dudyūṣati / dideviṣati "he wishes to play".

SK 2610 · सनि ग्रहगुहोश्च · sani grahaguhośca (A 7.2.12)

सनि ग्रहगुहोश्च
sani grahaguhośca · A 7.2.12
ग्रहेर्गुहेरुगन्ताच्च सन इण्न स्यात् । ग्रहिज्या - SK 2412 इति संप्रसारणम् । सनः षत्वस्यासिद्धत्वाद्भष्भावः । जिघृक्षति । सुषुप्सति ॥

Rule. Before san, iṭ is not added after grah and guh, nor after a dhātu ending in an uk vowel (u, ṛ, ḷ). Bhaṭṭoji's gloss: "let iṭ not come to san after grah, after guh, and after [a dhātu] ending in uk".

The words "not" and "iṭ" are carried forward from neḍ vaśi kṛti (SK 2981, taught later). The word ca draws in "ending in uk" from śryukaḥ kiti (SK 2381). Bhaṭṭoji's commentators agree that śri is not drawn in with it. SK 2618 gives śri an option of its own. They also say why the rule is needed: otherwise grah would take iṭ always, and guh, being ūdit, optionally.

Example.

grahītum icchati, "he wishes to seize":

grah + san → no iṭ (this rule) → grah + sa

→ san is kit after grah (SK 2609)

→ r takes saṃprasāraṇa before a kit affix (SK 2412), and ṛ + a gives ṛ → gṛh + sa

→ the s now follows h, an iṇ sound, so ṣ for it (SK 212) is due. But that ṣ is asiddha: treated as not yet made for earlier rules.

→ so the form still ends in h before s: h → ḍh (SK 324) → gṛḍh + sa

→ g takes aspiration before s, since the root now ends in an aspirate (bhaṣbhāva, SK 326) → ghṛḍh + sa

→ ḍh → k before s (SK 295) → ghṛk + sa → s → ṣ after k (SK 212) → ghṛkṣa

→ doubling (SK 2395) → ghṛk ghṛkṣa → first consonant only (SK 2179) → ghṛ ghṛkṣa

→ a for ṛ in the abhyāsa (SK 2244) → i for a (SK 2317) → ghi ghṛkṣa

→ gh → jh (SK 2245) → j (SK 2182) → jighṛkṣa → jighṛkṣati.

Bhaṭṭoji's own reason is "because the ṣ of san is asiddha, there is bhaṣbhāva". Both his commentators say "because the change to k is asiddha" would have been the better wording. How to read his words, scholars discuss this.

Example 2: suṣupsati. From svap "to sleep", without iṭ.

svap + sa → kit (SK 2609) → saṃprasāraṇa of v (SK 2409): sup-sa; no guṇa

→ doubling → sup supsa → first consonant only → su supsa

→ the root's s follows the u of the abhyāsa and becomes ṣ (SK 212) → suṣupsati.

One of Bhaṭṭoji's commentators gives these steps.

Example 3. One of Bhaṭṭoji's commentators gives bubhūṣati as the example for "ending in uk". The other adds jughukṣati for guh (worked under SK 2613).

Watch out. The "no iṭ" of this rule feeds SK 2612. Bhū has no iṭ, so its san begins with s and is kit.

SK 2611 · किरश्च पञ्चभ्यः · kiraśca pañcabhyaḥ (A 7.2.75)

किरश्च पञ्चभ्यः
kiraśca pañcabhyaḥ · A 7.2.75
कॄ गॄ दृङ् धृङ् प्रच्छ् एभ्यः सन इट् स्यात् ॥ पिपृच्छिषति । चिकरिषति । जिगरिषति । जिगलिषति । अत्रेटो दीर्घो नेष्टः ॥ दिदरिषते । दिधरिषते । कथमुद्दिधीर्षुरिति । भौवादिकयोर्धृङ्धृञोरिति गृहाण ॥

Rule. After five tudādi roots, san takes iṭ. The five are kṝ "to scatter", gṝ "to swallow", dṛṅ "to respect", dhṛṅ "to stay" and pracch "to ask". Bhaṭṭoji's gloss: "let iṭ come to san after these: kṝ, gṝ, dṛṅ, dhṛṅ, pracch".

One of Bhaṭṭoji's commentators gives the carried words. "Iṭ" comes from iḍ atyartivyayatīnām (SK 2384). "Before san" comes from SK 2626, taught later in this chapter. He also gives the purpose. For kṝ and gṝ, SK 2625 would make iṭ optional. For the others, SK 2246 would deny it. This rule overrides both.

Example.

praṣṭum icchati, "he wishes to ask":

pracch + san → iṭ (this rule) → pracch + i + sa → ṣ → pracch-iṣa

→ san is kit after pracch (SK 2609), so r takes saṃprasāraṇa (SK 2412) → pṛcch-iṣa

→ doubling → first consonant only → pṛ pṛcchiṣa → a for ṛ (SK 2244) → i (SK 2317) → pipṛcchiṣati.

kṝ: kṝ + i + ṣa → san with iṭ is not kit, so guṇa of the final ṝ (SK 2168), with r after it → kar-iṣa

→ doubling → ka kariṣa → i (SK 2317) → k → c (SK 2245) → cikariṣati.

gṝ: jigariṣati; and with l for r, optional before a vowel by aci vibhāṣā (SK 2541), jigaliṣati.

Vārttika: atreṭo dīrgho neṣṭaḥ. "Here the lengthening of iṭ is not wanted." After a root in ṝ, SK 2391 would make iṭ optionally long, which would give an unwanted long ī in cikariṣati. This statement forbids it here. Bhaṭṭoji sets it apart as a vārttika. Where it comes from, scholars discuss this.

dṛṅ and dhṛṅ: didariṣate, didhariṣate. One of Bhaṭṭoji's commentators says they take the ātmanepada by pūrvavat sanaḥ (SK 2734), since the roots are marked with ṅ.

Applies when / blocked by. Bhaṭṭoji raises a doubt. How then is uddidhīrṣuḥ "wishing to hold up" correct, with no iṭ? His answer: "Take it as from dhṛṅ and dhṛñ of the bhvādi class." The tudādi dhṛṅ is in this rule; the bhvādi roots are not. Bhaṭṭoji's commentators agree that the lengthening then comes from SK 2614. One of them gives the steps:

dhṛ + sa (no iṭ) → long vowel before a jhal-initial san (SK 2614) → dhṝ-sa

→ i for ṝ, with r after it (SK 2390) → dhir-sa → i lengthened before a consonant (SK 354) → dhīr-sa → ṣ → dhīrṣa

→ doubling, short vowel, dh → d → didhīrṣa → with ud and the affix u (taught later) → uddidhīrṣuḥ.

Watch out. The rule speaks of the tudādi roots only. The same sounds in another class need not take iṭ.

SK 2618 · सनीवन्तर्धभ्रस्जदम्भुश्रिस्वृयूर्णुभरज्ञपिसनाम् · sanīvantardhabhrasjadambhuśrisvṛyūrṇubharajñapisanām (A 7.2.49)

सनीवन्तर्धभ्रस्जदम्भुश्रिस्वृयूर्णुभरज्ञपिसनाम्
sanīvantardhabhrasjadambhuśrisvṛyūrṇubharajñapisanām · A 7.2.49
इवान्तेभ्य ऋधादिभ्यश्च सन इड्वा स्यात् । इडभावे हलन्ताच्च SK 2613 इति कित्वम् । छ्वोः - SK 2561 इति वस्य ऊठ् । यण् द्वित्वम् । दुद्यूषति । दिदेविषति । स्तौतिण्योरेव - SK 2627 इति वक्ष्यमाणनियमान्न षः । सुस्यूषति । सिसेविषति ॥

Rule. Before san, iṭ is optional after these: roots ending in iv; ṛdh "to prosper"; bhrasj; dambh; śri; svṛ; yu; ūrṇu; bhṛ (bhara); jñapi; and the root san. Bhaṭṭoji's gloss: "let iṭ optionally come to san after [roots] ending in iv, and after ṛdh and the rest".

One of Bhaṭṭoji's commentators gives the carried words. "Optionally" comes from svaratisūti… (SK 2279). "Iṭ" comes from iṇ niṣṭhāyām (SK 3045, taught later).

Example.

devitum icchati, "he wishes to play", without iṭ:

div + sa → v is a hal next to the ik vowel i; san begins with a jhal → kit (SK 2613)

→ v is replaced by ūṭh, that is ū (SK 2561) → di ū sa

→ y for i before a vowel (yaṇ, SK 47) → dyū sa → ṣ → dyūṣa

→ doubling (SK 2395) → dyū dyūṣa → first consonant only (SK 2179) → short vowel (SK 2180) → du dyūṣa → dudyūṣati.

With iṭ: div + iṣa → the san does not begin with a jhal, so it is not kit; no ūṭh

→ guṇa of the light penultimate (SK 2189) → dev-iṣa → doubling, short vowel → dideviṣati.

One of Bhaṭṭoji's commentators explains the second form so.

siv "to sew": susyūṣati / siseviṣati. Here the s after the abhyāsa stays s. Bhaṭṭoji cites the restriction stautiṇyor eva… (SK 2627), to be stated: only stu and a ṇi-final stem take that ṣ.

The other roots of the list. Bhaṭṭoji works most of them under SK 2620 and SK 2621. Each comes in two forms: without iṭ and with it.

The rest are worked here.

śri: without iṭ, śri + sa

→ long vowel (SK 2614)

→ śrī-sa

→ kit (SK 2612): no guṇa

→ śrīṣa

→ doubling, first consonant, short vowel

→ śiśrīṣati.

With iṭ: no lengthening and no kittva. The doubling sees the original śri (SK 2243)

→ śi śri-iṣa

→ guṇa and ay in the second part (SK 61)

→ śiśrayiṣati.

One of Bhaṭṭoji's commentators notes the lengthening and the kittva in the first form, and their absence in the second.

svṛ: without iṭ, svṛ + sa

→ long ṝ (SK 2614)

→ kit (SK 2612)

→ u for ṝ after the labial v, with r (SK 2494)

→ svur-sa

→ u lengthened (SK 2265)

→ svūrṣa

→ doubling

→ su svūrṣa

→ susvūrṣati.

With iṭ: svṛ doubled

→ a for ṛ in the abhyāsa (SK 2244)

→ i (SK 2317)

→ si

→ guṇa in the second part

→ sisvariṣati.

yu: without iṭ, yu + sa → long vowel (SK 2614) → kit → yūṣa → yuyūṣati.

With iṭ: guṇa is held back while doubling is due (SK 2243)

→ yu yu-iṣa

→ guṇa and av in the second part

→ yu yav-iṣa

→ the abhyāsa u becomes i before y + a (SK 2577)

→ yiyaviṣati.

ūrṇu: without iṭ, ūrṇu + sa

→ long vowel (SK 2614)

→ kit (SK 2612): no guṇa

→ the root begins with a vowel, so its second part, nū with the s, is doubled; r is not doubled (SK 2446)

→ short vowel in the abhyāsa

→ ūrṇunūṣati.

With iṭ, san is optionally ṅit by vibhāṣorṇoḥ (SK 2447). One of Bhaṭṭoji's commentators explains both results.

Why the abhyāsa of ūrṇunaviṣati shows u. Bhaṭṭoji meets an objection. Guṇa and av, being later rules (para), would apply first. Then nav would be doubled, its a would become i, and no u would be heard. He answers: when doubling is to be done, dvirvacane 'ci (SK 2243) works in one of two ways. Either the substitute takes on the original's form (sthānirūpātideśa), or the substitute is not made at all (ādeśaniṣedha). One of his commentators explains: on either reading, nu is what gets doubled. So u is heard in the abhyāsa.

A second objection follows. SK 2243 works only when a vowel causing the doubling follows. Here that vowel is the i of san. But the whole sannanta undergoes doubling. Can a thing that undergoes an operation (kāryin) also be its cause (nimitta)? Bhaṭṭoji answers with a maxim (paribhāṣā): "it is only while undergoing an operation that the undergoer is not taken as its cause". "And here," he adds, "san does not undergo the doubling." Bhaṭṭoji's commentators gloss: it is nu that is doubled.

bhṛ: without iṭ, bhṛ + sa

→ long ṝ (SK 2614)

→ kit

→ u with r after the labial bh (SK 2494)

→ lengthened (SK 354)

→ bhūrṣa

→ doubling, short vowel, and bh becomes b (SK 2182)

→ bubhūrṣati.

With iṭ: doubling, a for ṛ in the abhyāsa, i → guṇa in the second part → bibhariṣati.

One of Bhaṭṭoji's commentators gives these steps.

The root san: without iṭ, san + sa

→ ā for the n before a jhal-initial san (janasana…, SK 2504)

→ sā-sa

→ doubling, short vowel, i

→ si sāsa

→ the s after the abhyāsa's i becomes ṣ (SK 212)

→ siṣāsati.

With iṭ: no ā, since san does not begin with a jhal → sisaniṣati.

One of Bhaṭṭoji's commentators explains the s in each. In siṣāsati the san has not become ṣ (it follows ā), so SK 2627 does not apply, and the ṣ stands. In sisaniṣati the san has become ṣ, so SK 2627 keeps the s.

Vārttika under SK 2621: tanipatidaridrātibhyaḥ sano vā iḍ vācyaḥ. "It should be stated that san optionally takes iṭ after tan, pat and daridrā." One of Bhaṭṭoji's commentators calls it an option where iṭ would otherwise apply. Its forms come later: titaniṣati (SK 2622), pipatiṣati (SK 2622), and didaridriṣati / didaridrāsati (SK 2623).

Watch out. Each root in the list has two desideratives, and they differ in more than iṭ. Without iṭ, san is usually kit, and lengthening (SK 2614) may apply. With iṭ, san is not kit, so guṇa usually applies, and there is no lengthening.

SK 2625 · इट् सनि वा · iṭ sani vā (A 7.2.41)

इट् सनि वा
iṭ sani vā · A 7.2.41
वृङ्वृञ्भ्यामॄदन्ताच्च सन इड्वा स्यात् । तितरिषति । तितरीषति । तितीर्षति । विवरिषति । विवरीषति । वुवूर्षति । वृङ् । वुवूर्षते । विवरिषते । दुध्वूर्षति ॥

Rule. After vṛṅ and vṛñ, and after a root ending in ṝ, san optionally takes iṭ. Bhaṭṭoji's gloss: "let san optionally take iṭ after vṛṅ and vṛñ and after [a root] ending in ṝ".

One of Bhaṭṭoji's commentators gives the carried word: "after vṛ and ṝ" from vṝto vā (SK 2391). He calls the rule an exception (apavāda) to SK 2610. Both commentators add a limit. The word "in upadeśa" is carried forward from SK 2246. So the root must end in ṝ as first taught.

Example.

tṝ:

with iṭ: tṝ + iṣa

→ not kit

→ guṇa with r

→ tar-iṣa

→ doubling

→ ta tariṣa

→ i (SK 2317)

→ titariṣati;

with iṭ made long by SK 2391 → titarīṣati;

without iṭ: tṝ + sa

→ kit (SK 2612): no guṇa

→ i for ṝ, with r (SK 2390)

→ tir-sa

→ lengthened (SK 354)

→ tīrṣa

→ doubling, short vowel

→ titīrṣati.

One of Bhaṭṭoji's commentators gives these steps.

vṛñ: vivariṣati, vivarīṣati, and without iṭ vuvūrṣati (u by SK 2494 after the labial v).

vṛṅ: vuvūrṣate, vivariṣate; the ātmanepada comes because the root is marked with ṅ.

Applies when / blocked by. In cikīrṣati "he wishes to do", from kṛ, the ṝ is made by SK 2614. It was not there in upadeśa, so this rule does not apply. For the same reason, one of Bhaṭṭoji's commentators says, Bhaṭṭoji gives dudhvūrṣati without iṭ. Dhvṛ "to be crooked" ends in short ṛ in upadeśa.

dhvṛ + sa

→ long ṝ (SK 2614)

→ u with r (SK 2494)

→ lengthened (SK 354)

→ doubling

→ dudhvūrṣati.

Watch out. Vṛ and the ṝ-final roots end in uk. Without this rule, SK 2610 would deny them iṭ outright.

SK 2626 · स्मिपूङ्रञ्ज्वशां सनि · smipūṅrañjvaśāṃ sani (A 7.2.74)

स्मिपूङ्रञ्ज्वशां सनि
smipūṅrañjvaśāṃ sani · A 7.2.74
स्मि पूङ् ऋ अञ्जू अश् एभ्यः सन इट् स्यात् । सिस्मयिषते । पिपविषते । अरिरिषति । इह रिस्शब्दस्य द्वित्वम् । इस् इति सनोऽवयवः कार्यभागिति कार्यिणो निमित्तत्वायोगात् द्विर्वचनेऽचि - SK 2243 इति न प्रवर्तते । अञ्जिजिषति । अशिशिषते । उभौ साभ्यासस्य SK 2606 । प्राणिणिषति । उच्छेस्तुक् । चुत्वम् । पूर्वत्रासिद्धीयमद्विर्वचने इति चछाभ्यां सहितस्येटो द्वित्वम् । हलादिः शेषः SK 2179 । उचिच्छिषति । निमित्तापाये नैमित्तिकस्याप्यपायः इति त्वनित्यम् । च्छ्वोः - SK 2561 इति सतुग्ग्रहणाज्ज्ञापकात् । प्रकृतिप्रत्यापत्तिवचनाद्वा । णौ च संश्चङोः - SK 2579 इति सूत्राभ्यामिङो गाङ् श्वयतेः संप्रसारणं च वा । अधिजिगापयिषति । अध्यापिपयिषति । शिश्वाययिषति । शुशावयिषति । ह्वः संप्रसारणम् SK 2586 जुहावयिषति । णौ द्वित्वात्प्रागच आदेशो नेत्युक्तत्वादुकारस्य द्वित्वम् । पुस्फारयिषति । चुक्षाययिषति । ओः पुयण्ज्यपरे SK 2577 । पिपावयिषति । बिभावयिषति । जिजायिषति । पुयण्जि किम् । नुनावयिषति । अपरे किम् । बुभूषति । स्रवति - SK 2578 इतीत्वं वा । सिस्रावयिषति । सुस्रावयिषतीत्यादि । अपर इत्येव । शुश्रूषते ॥

Rule. San takes iṭ after smi "to smile", pūṅ, ṛ, añjū "to anoint" and aś "to pervade". Bhaṭṭoji's gloss: "let san take iṭ after smi, pūṅ, ṛ, añjū, aś". "Iṭ" is carried forward from SK 2384.

Example.

smi: smi + iṣa

→ guṇa held back while doubling is due (SK 2243)

→ smi smi-iṣa

→ first consonant only

→ si smi-iṣa

→ guṇa and ay in the second part → sismayiṣa → sismayiṣate.

The s after the abhyāsa stays s, by SK 2627. One of Bhaṭṭoji's commentators notes that smiṅ would otherwise take no iṭ.

pūṅ: pū doubled (SK 2243)

→ short vowel

→ guṇa and av in the second part

→ pu pav-iṣa

→ the abhyāsa u becomes i before p + a (SK 2577)

→ pipaviṣate.

Bhaṭṭoji's commentators note that pūñ, the root with ñ, is not named. It gives pupūṣati, with no iṭ by SK 2610.

ṛ: ṛ + iṣa

→ guṇa, with r

→ ar-iṣa

→ the root begins with a vowel, so the second part, ris, is doubled (SK 2176)

→ a ris risa

→ first consonant only

→ a ri riṣa

→ aririṣati.

Why guṇa comes first in aririṣati. Bhaṭṭoji says: "here ris is doubled". Does SK 2243 not hold guṇa back? No. "Since is, a part of the san, shares in the operation (kāryabhāk), and since what undergoes an operation cannot be its cause, dvirvacane 'ci does not apply." One of his commentators explains: through its part is, the whole isa undergoes the doubling. So it cannot be the vowel that causes the doubling.

añjū: añj + iṣa → the n is not doubled (SK 2446), so jis is doubled → añjijiṣati.

aś: aś + iṣa → śis doubled → aśiśiṣate.

Bhaṭṭoji's commentators note that both roots are ūdit, so iṭ would otherwise be optional (SK 2279). Here it is obligatory. Aś "to eat" of the kryādi class takes iṭ in any case.

Further forms Bhaṭṭoji gives here.

prāṇiṇiṣati "he wishes to breathe", from pra + an. By ubhau sābhyāsasya (SK 2606), the n becomes ṇ in both the abhyāsa and the second part.

ucicchiṣati, from ucch (ucchī):

→ the augment tuk (t) before ch (SK 146, che ca) → the t becomes c (SK 111) → ucch + iṣa

→ the second part, cchis, is doubled with its c and ch

→ only the first consonant stays in the abhyāsa (SK 2179) → u ci cchiṣa → ucicchiṣati.

Why is c, not t, doubled? Bhaṭṭoji cites the statement pūrvatrāsiddhīyam advirvacane: for doubling, the later rule's result is not treated as unmade. So the c made by SK 111 counts at the doubling.

A second doubt follows. In the abhyāsa, the ch is lost. The tuk was there only because of ch. Should its c not vanish too? Bhaṭṭoji answers that the maxim nimittāpāye naimittikasyāpy apāyaḥ, "when the cause goes, what it caused goes too", is not binding (anitya). He gives two signs (jñāpaka). One is SK 2561, which names ch together with its tuk. The other is the statement that prescribes restoring a base (prakṛtipratyāpatti). One of his commentators explains both. Each statement would be useless if the maxim always held.

The remaining forms of this kaumudī are san after ṇi-final stems. They are set out in group G, beside the rules on ṣ that they feed.

Watch out. Smi, ṛ and the rest would not take iṭ, or would take it only optionally. This rule makes it obligatory. So none of these *san*s is kit, and guṇa applies wherever the root has an i, u or ṛ (sismayiṣate, pipaviṣate, aririṣati).

D. The long vowel before a jhal-initial san — running example: ci + sa → cī-sa (2614) → ci-cī-ṣa → cicīṣati / cikīṣati; han + sa → hān-sa → jighāṃsati; tan + sa → titāṃsati / titaṃsati

Before a san without iṭ, some stems lengthen a vowel. SK 2614 lengthens a final vowel, and the penultimate a of han "to kill" and of a certain gam. SK 2622 makes the same lengthening optional for tan "to stretch". Bhaṭṭoji teaches SK 2622 later; it is placed here with SK 2614. The running examples are cicīṣati "he wishes to gather", jighāṃsati "he wishes to kill", and titāṃsati / titaṃsati "he wishes to stretch".

SK 2614 · अज्झनगमां सनि · ajjhanagamāṃ sani (A 6.4.16)

अज्झनगमां सनि
ajjhanagamāṃ sani · A 6.4.16
अजन्तानां हन्तेरजादेशगमेश्च दीर्घः स्याज्झलादौ सनि । सन्लिटोर्जेः SK 2331 जिगीषति । विभाषा चेः SK 2525 चिकीषति । चिचीषति । जिघांसति ॥

Rule. Before a san beginning with a jhal, a long vowel replaces the final of a stem (aṅga) ending in a vowel (ac). It also replaces the penultimate of han, and of gam when gam is a substitute for a vowel-root (ajādeśa). Bhaṭṭoji's gloss: "let there be a long vowel for [stems] ending in ac, for han, and for gam that is a substitute for an ac, before a san beginning with a jhal".

One of Bhaṭṭoji's commentators gives the carried words.

The words "beginning with a jhal" are not carried from any sūtra. The same commentator traces them to the Bhāṣya on cchvoḥ… (SK 2561), which says "we shall qualify san by jhal".

Example.

cetum icchati, "he wishes to gather":

ci + sa (no iṭ) → the stem ends in a vowel; san begins with a jhal → long vowel (this rule) → cī-sa

→ kit (SK 2612): no guṇa → ṣ (SK 212) → cīṣa

→ doubling, short vowel in the abhyāsa → ci cīṣa

→ optionally a velar for the c after the abhyāsa, by vibhāṣā ceḥ (SK 2525) → cicīṣati / cikīṣati.

ji "to conquer": jī-sa

→ doubling

→ ji jīṣa

→ a velar for j after the abhyāsa, by sanliṭor jeḥ (SK 2331)

→ jigīṣati.

Bhaṭṭoji recalls both rules here.

hantum icchati, "he wishes to kill":

han + sa (no iṭ) → the penultimate a is lengthened (this rule) → hān-sa

→ doubling → hān hānsa → first consonant only, short vowel (SK 2179, 2180) → ha hānsa → i (SK 2317) → hi hānsa

→ h → jh (SK 2245) → j (SK 2182) → ji hānsa

→ gh for the h after the abhyāsa (abhyāsāc ca, SK 2430) → ji ghānsa

→ n → anusvāra before s (SK 123) → jighāṃsati.

One of Bhaṭṭoji's commentators gives these steps. He adds that the a is lengthened, not the n. The word "long" calls up "a vowel" by itself.

Applies when / blocked by. Gam here means only the gam that replaces a vowel-root. SK 2615 and SK 2616 put it in place of iṇ and iṅ, and a vārttika in place of ik. Bhaṭṭoji's commentators agree on this. One says the Bhāṣya makes it clear. The other draws it from the wording: the word ac in the sūtra is not needed for the stems, so it is split off to qualify gam.

So gamḷ "to go", the ordinary root, is not lengthened: jigaṃsyate, saṃjigaṃsate (see SK 2615). One of Bhaṭṭoji's commentators also gives a counter-example to "jhal-initial": jigamiṣati, where iṭ comes by SK 2401.

Watch out. The lengthening needs a san without iṭ. With iṭ, the san begins with i, and there is no long vowel: compare śiśrīṣati and śiśrayiṣati (SK 2618).

SK 2622 · तनोतेर्विभाषा · tanotervibhāṣā (A 6.4.17)

तनोतेर्विभाषा
tanotervibhāṣā · A 6.4.17
अस्योपधाया दीर्घो वा स्याज्झलादौ सनि । तितांसति । तितंसति । तितनिषति ।
आशङ्कायां सन्वक्तव्यः
(वार्तिकम्) ॥ श्वा मुमूर्षति । कूलं पिपतिषति ॥

Rule. The penultimate of tan is optionally lengthened before a san beginning with a jhal. Bhaṭṭoji's gloss: "let the penultimate of this [root] optionally be lengthened before a san beginning with a jhal". One of his commentators gives the carried words: "penultimate" from SK 370, "long" from SK 174, and "before a jhal-initial san" from SK 2614.

Example.

tan + sa (no iṭ, by the option of the vārttika under SK 2621)

→ long penultimate (this rule) → tān-sa → anusvāra (SK 123) → tāṃsa

→ doubling → ta tāṃsa → i (SK 2317) → titāṃsati.

Without the lengthening → titaṃsati.

With iṭ, by that vārttika → titaniṣati.

Vārttika: āśaṅkāyāṃ san vaktavyaḥ. "San should be stated in the sense of apprehension (āśaṅkā)." One of Bhaṭṭoji's commentators explains. This san comes in its own sense (svārthe), after a root whose action is feared to be about to happen. Bhaṭṭoji's examples:

The same commentator glosses the first as "it is one whose death is apprehended". How to analyse the sense, scholars discuss this.

Watch out. The dog does not wish anything. This san is not the desiderative of SK 2608, though its forms look the same.

E. gam for iṇ, iṅ and ik before san — running example: iṇ + san → gam + iṣa (gamer iṭ) → jigamiṣati; passive gam + sa (no iṭ before taṅ) → gām-sa (2614) → jigāṃsyate

In chapter 54, SK 2607 put gam in place of iṇ "to go" before ṇi. Two sūtras now do the same before san. SK 2615 does it for iṇ, except in the sense of making known. SK 2616 does it for iṅ "to study". A vārttika adds ik "to remember". The running example is jigamiṣati "he wishes to go". In the passive, the same root gives jigāṃsyate "it is wished to be gone to".

SK 2615 · सनि च · sani ca (A 2.4.47)

सनि च
sani ca · A 2.4.47
इणो गमिः स्यात्सनि ने तु बोधने । जिगमिषति । बोधने प्रतीषिषति ॥
इण्वदिकः
(वार्तिकम्) ॥ अधिजिगमिषति । कर्मणि तङ् । परस्मैपदेषु इत्युक्तेर्नेट् । झलादौ सनि इति दीर्घः । जिगांस्यते । अधिजिगांस्यते अजादेशस्येत्युक्तेर्गच्छतेर्न दीर्घः । जिगंस्यते । संजिगंसते ॥

Rule. Before san, gam replaces iṇ, but not in the sense of making known (bodhana). Bhaṭṭoji's gloss: "let gam (gami) replace iṇ before san, but not in the sense of making known". One of Bhaṭṭoji's commentators gives the carried words. "Of iṇ" comes from iṇo gā luṅi (SK 2458). "Gam, except in making known" comes from ṇau gamir abodhane (SK 2607). The other confirms that "except in making known" comes from SK 2607.

Example.

iṇ + san → gam replaces i (this rule) → gam + sa

→ iṭ by gamer iṭ parasmaipadeṣu (SK 2401) → gam-iṣa

→ the san begins with i, not a jhal, so SK 2614 does not lengthen

→ doubling → ga gamiṣa → i (SK 2317) → g → j (SK 2245) → jigamiṣati.

Applies when / blocked by. In the sense of making known, there is no gam: pratīṣiṣati, "he wishes to make known". One of Bhaṭṭoji's commentators gives the steps.

prati + i + sa → the root begins with a vowel, so the second part, sa, is doubled (SK 2176) → i sa sa

→ the abhyāsa takes i (SK 2317) → i si sa

→ i of prati + i → long ī (SK 85) → both s become ṣ → pratīṣiṣati.

Whether the abhyāsa should take i here, scholars discuss this; one commentator holds that it does.

Vārttika: iṇvad ikaḥ. "Ik is like iṇ." One of Bhaṭṭoji's commentators says it is stated for ik "to remember". So ik too takes gam before san: adhijigamiṣati, "he wishes to remember".

The passive. Bhaṭṭoji adds the passive (karmaṇi). There the endings are the ātmanepada ones (taṅ), by bhāvakarmaṇoḥ (SK 2679, taught later). Gamer iṭ speaks only of parasmaipada endings. So now there is no iṭ.

iṇ + san, passive → gam + sa (no iṭ) → jhal-initial, so long penultimate (SK 2614) → gām-sa → anusvāra → gāṃsa

→ doubling → ga gāṃsa → i, j → jigāṃsa

→ the passive sign yak (SK 2756, taught later) with te → jigāṃsyate, "it is wished to be gone to".

From ik: adhijigāṃsyate, "it is wished to be remembered". One of Bhaṭṭoji's commentators gives both glosses.

The ordinary root gamḷ is not a substitute for a vowel-root. So, Bhaṭṭoji says, it is not lengthened: jigaṃsyate; and with sam, saṃjigaṃsate. One of his commentators says the ātmanepada of saṃjigaṃsate comes by samo gamyṛcchibhyām through SK 2734.

Watch out. Iṭ depends on the endings. Active jigamiṣati has iṭ and a short a. Passive jigāṃsyate has no iṭ and a long ā.

SK 2616 · इङश्च · iṅaśca (A 2.4.48)

इङश्च
iṅaśca · A 2.4.48
इङो गमिः स्यात्सनि । अधिजिगांसते ॥

Rule. Gam also replaces iṅ "to study" before san. Bhaṭṭoji's gloss: "let gam replace iṅ before san". One of his commentators says "gam" is carried forward from SK 2607 and "before san" from SK 2615.

Example.

adhi + iṅ + san → gam (this rule) → adhi gam + sa

→ iṅ is marked with ṅ, so the desiderative takes the ātmanepada (SK 2734)

→ gamer iṭ holds only before parasmaipada endings, so no iṭ

→ jhal-initial san: long penultimate (SK 2614) → gāṃsa → doubling → jigāṃsa → adhijigāṃsate, "he wishes to study".

One of Bhaṭṭoji's commentators gives these steps; Bhaṭṭoji gives only the form.

Watch out. With a ṇi between root and san, iṅ may instead take gāṅ. That form, adhijigāpayiṣati, is set out in group G.

F. A vowel substitute, and the abhyāsa dropped — running example: āptum icchati: āp + sa → īp-sa (ī, 2619) → ī-psa-psa (psa doubled) → ī-psa (abhyāsa dropped, 2620) → īpsati; pat + sa → pist-sa (is, 2623) → abhyāsa dropped → s lost by skoḥ → pitsati

Five sūtras give a few roots a new vowel before a san that begins with s. SK 2619 puts ī in āp, jñap and ṛdh. SK 2621 puts i or ī in dambh. SK 2623 puts is in mī, mā, the ghu roots and five more. SK 2624 gives guṇa to intransitive muc. And SK 2620 deletes the abhyāsa wherever these four have applied. In Pāṇini's order the block runs SK 2623, 2619, 2621, 2624, 2620. Bhaṭṭoji scatters it, so several of these rules carry words from SK 2623, which he teaches later. The running examples are īpsati "he wishes to obtain" and pitsati "he wishes to fall".

SK 2619 · आप्ज्ञप्यृधामीत् · āpjñapyṛdhāmīt (A 7.4.55)

आप्ज्ञप्यृधामीत्
āpjñapyṛdhāmīt · A 7.4.55
एषामच ईत्स्यात्सादौ सनि ॥

Rule. Ī replaces the vowel of āp "to obtain", jñap and ṛdh "to prosper", before a san beginning with s. Bhaṭṭoji's gloss: "let ī be substituted for the ac of these before a san beginning with s".

Bhaṭṭoji's commentators give the carried words. "Of the vowel" (acaḥ) comes from sani mīmā… (SK 2623); both commentators say so. One of them adds that "before san" comes from it too. That sūtra stands just before this one in Pāṇini's order. "S" comes from saḥ sy ārdhadhātuke (SK 2342). It qualifies san, so the rule means a san that begins with s.

Example. The forms come under the next two sūtras: īpsati, īrtsati (SK 2620) and jñīpsati (SK 2621).

With iṭ, the san begins with i, not s, so there is no ī: ardidhiṣati (SK 2620).

Watch out. "Beginning with s" in effect means "without iṭ". The option of SK 2618 decides which form you get for ṛdh and jñap.

SK 2620 · अत्र लोपोऽभ्यासस्य · atra lopo'bhyāsasya (A 7.4.58)

अत्र लोपोऽभ्यासस्य
atra lopo'bhyāsasya · A 7.4.58
सनिमीमा - SK 2623 इत्यारभ्य यदुक्तं तत्राभ्यासस्य लोपः स्यात् । आप्तुमिच्छति ईप्सति । अर्धितुमिच्छति । रपरत्वम् । चर्त्वम् । ईर्त्सति । अर्दिधिषति । बिभ्रज्जिषति । बिभर्जिषति । बिभ्रक्षति । बिभर्क्षति ॥

Rule. "Here" the abhyāsa is deleted (lopa). "Here" means wherever SK 2623, 2619, 2621 or 2624 has applied. Bhaṭṭoji's gloss: "wherever [an operation] is stated from sani mīmā… onwards, let there be deletion of the abhyāsa". One of his commentators explains: once the operation of those four rules has been carried out.

Example.

āptum icchati, "he wishes to obtain":

āp + sa (no iṭ) → ī for ā (SK 2619) → īp-sa

→ the root begins with a vowel, so the second part, psa, is doubled (SK 2176) → ī psa psa

→ the abhyāsa is deleted (this rule) → ī psa → īpsati.

Bhaṭṭoji's commentators give the doubling of psa.

ardhitum icchati, "he wishes to prosper", without iṭ:

ṛdh + sa → ī for ṛ (SK 2619), with r after it (raparatva, SK 70) → īrdh-sa

→ dh → t before s (cartva, SK 121) → īrt-sa

→ r is not doubled (SK 2446), so tsa is doubled → īr tsa tsa

→ the abhyāsa is deleted (this rule) → īrtsati.

One of Bhaṭṭoji's commentators adds why t, not dh, is doubled. The statement pūrvatrāsiddhīyam advirvacane lets the doubling see the result of cartva.

With iṭ (SK 2618): no ī, since the san does not begin with s

→ guṇa with r → ardh-iṣa → dhis is doubled → ar dhi dhiṣa → dh → d in the abhyāsa → ardidhiṣati.

bhrasj gives four forms. The iṭ option of SK 2618 combines with the optional augment ram of SK 2535. One of Bhaṭṭoji's commentators sets them out.

These four have no deleted abhyāsa. Bhrasj is not among the four rules.

Watch out. The abhyāsa goes only where one of the four rules has acted. In ardidhiṣati SK 2619 did not act, so the abhyāsa stays.

SK 2621 · दम्भ इच्च · dambha icca (A 7.4.56)

दम्भ इच्च
dambha icca · A 7.4.56
दम्भेरच इत्स्यादीच्च सादौ सनि । अभ्यासलोपः । हलन्ताच्च SK 2613 इत्यत्र हल्ग्रहणं जातिपरमित्युक्तम् । तेन सनः कित्त्वान्नलोपः । धिप्सति । धीप्सति । दिदम्भिषति । शिश्रीषति । शिश्रयिषति । उदोष्ठ्यपूर्वस्य SK 2494 । सुस्वूर्षति । सिस्वरिषति । युयूषति । यियविषति । ऊर्णुनूषति । ऊर्णुनुविषति । ऊर्णुनविषति । न च परत्वाद्गुणावादेशयोः सतोरभ्यासे उकारो न श्रूयेतेति वाच्यम् । द्विर्वचनेचि SK 2343 इति सूत्रेण द्वित्वे कर्तव्ये स्थानिरूपातिदेशादादेशमनिषेधाद्वा । न च सन्नन्तस्य द्वित्वं प्रति कार्यित्वान्निमित्तता कथमिति वाच्यम् ॥ [(परिभाषा - ) कार्यमनुभवन्हि कार्यी निमित्ततया नाश्रीयते न त्वननुभवन्नपि] ॥ न चेह सन् द्वित्वमनुभवति । बुभूर्षति । बिभरिषति । ज्ञपिः पुगन्तो मित्संज्ञकः पकारान्तश्चौरादिकश्च । इडभावे इको झल् SK 2612 इति कित्त्वान्न गुणः । अज्झन - SK 2614 इति दीर्घः परत्वाण्णिलोपेन बाध्यते । आप्ज्ञप् - SK 2619 इति ईत् । ज्ञीप्सति । जिज्ञपयिषति । अमितस्तु । जिज्ञापयिषति । जनसन - SK 2504 इत्यात्वम् । सिषासति । सिसनिषति ।
तनिपतिदरिद्रातिभ्यः सनो वा इड्वाच्यः
(वार्तिकम्) ॥

Rule. I, and by the word ca also ī, replaces the vowel of dambh before a san beginning with s. Then the abhyāsa is deleted (SK 2620). Bhaṭṭoji's gloss: "let i, and also ī, be substituted for the ac of dambh before a san beginning with s". One of his commentators gives the carried words: "before san" and "of the vowel" from SK 2623, ī from SK 2619, and s from SK 2342.

Example.

dambh + sa (no iṭ, SK 2618) → i for a (this rule) → dimbh-sa

→ is the san kit? The anusvāra and the bh stand between the i and the san. Bhaṭṭoji recalls that hal in SK 2613 means the class. So the san is kit (SK 2613)

→ the penultimate n is lost before a kit affix (SK 415) → dibh-sa

→ doubling → the abhyāsa is deleted (SK 2620) → dibh-sa

→ d takes aspiration before s, since the root ends in bh (SK 326) → dhibh-sa

→ bh → p before s (SK 121) → dhipsa → dhipsati.

With ī: dhīpsati. With iṭ: the san is not kit, so the n stays and there is no i → didambhiṣati.

jñapi. Bhaṭṭoji explains the jñapi of SK 2618. It is the short-vowelled (mit) ṇic-stem of jñā, with the augment puk. It is also the curādi root that ends in p of itself. One of his commentators adds that the first is jñā of the ghaṭādi list, whose penultimate is shortened before ṇi.

jñapi + sa (no iṭ) → the stem ends in i, an ik → kit (SK 2612), so the i of ṇi takes no guṇa

→ SK 2614 would lengthen it; but deletion of ṇi (SK 2313), being the later rule, applies instead → jñap-sa

→ ī for a (SK 2619) → jñīp-sa → doubling → the abhyāsa is deleted (SK 2620) → jñīpsati.

With iṭ: jñapi + iṣa → guṇa and ay of the ṇi → doubling → jijñapayiṣati.

From the stem that is not mit, jñāpi: jijñāpayiṣati. One of Bhaṭṭoji's commentators says this stem is not covered by SK 2618, so its iṭ is obligatory.

The other roots of SK 2618, which Bhaṭṭoji works in this kaumudī, are set out under SK 2618. So is the vārttika he states here on tan, pat and daridrā.

Watch out. In dhipsati three earlier rules do the work after this one: SK 2613 (kit by the class reading), SK 415 (n lost) and SK 2620 (abhyāsa gone).

SK 2623 · सनि मीमाघुरभलभशकपतपदामच इस् · sani mīmāghurabhalabhaśakapatapadāmaca is (A 7.4.54)

सनि मीमाघुरभलभशकपतपदामच इस्
sani mīmāghurabhalabhaśakapatapadāmaca is · A 7.4.54
एषामच इस् स्यात्सादौ सनि । अभ्यासलोपः । स्कोः - SK 380 इति सलोपः । पित्सति । दिदरिद्रिषति । दिदरिद्रासति । डुमिञ् मीञ् आभ्यां सन् । कृतदीर्घस्य मिनोतेरपि मीरुपत्वाविशेषादिस् । सः सि - SK 2342 इति तः । मित्सति । दोदाणोः । दित्सति । देङ् । दित्सते । दाञ् । दित्सति । दित्सते । धेट् । धित्सति । धाञ् । धित्सति । धित्सते । रिप्सते । लिप्सते । शकि । शिक्षति । शक मर्षण इति दिवादिः । स्वरितेत् । शिक्षति । शिक्षते । पित्सते ।
राधो हिंसायां सनीस् वाच्यः
(वार्तिकम्) ॥ रित्सति । हिंसायां किम् । आरिरात्सति ॥

Rule. Before a san beginning with s, is replaces the vowel of mī, mā, the ghu roots, rabh, labh, śak, pat and pad. The ghu roots are the dā and dhā roots. Bhaṭṭoji's gloss: "let is be substituted for the ac of these before a san beginning with s". "S" is carried forward from SK 2342. Then the abhyāsa is deleted (SK 2620).

Example.

patitum icchati, "he wishes to fall", without iṭ:

pat + sa → is for a (this rule) → p-is-t-sa → pist-sa

→ doubling, then the abhyāsa is deleted (SK 2620) → pist-sa

→ the s at the head of the cluster st is lost (skoḥ…, SK 380) → pitsa → pitsati.

Bhaṭṭoji gives the steps in this order.

daridrā, with and without iṭ (the vārttika under SK 2621): didaridriṣati, didaridrāsati.

mī: Bhaṭṭoji takes two roots, ḍumiñ and mīñ. Ḍumiñ has a short i, but SK 2614 lengthens it before san. "Even [the mi of] minoti, once lengthened, takes is, because its form mī is no different."

mī + sa

→ is

→ mis-sa

→ the s of is becomes t (SK 2342)

→ mitsa

→ doubling, abhyāsa deleted

→ mitsati, mitsate.

One of his commentators says the root is marked with ñ, so both padas are used (SK 2734). The other adds that the san is kit by SK 2613, so there is no guṇa.

That commentator also meets an objection. A maxim says that a form named directly is preferred to one made by a rule (lakṣaṇa-pratipadokta). Minoti's mī is made by a rule. He answers, citing Kaiyaṭa, that the maxim holds where the product of a rule is relied on, not where only the bare form is. So it is not binding (anitya).

mā "to measure": mitsati. Māṅ and meṅ: mitsate. One of Bhaṭṭoji's commentators notes that meṅ takes the form mā before san, so it is covered too.

The ghu roots:

dā + sa

→ is

→ dis-sa

→ t for the s of is (SK 2342)

→ ditsa

→ abhyāsa deleted

→ ditsati.

rabh: ripsate. labh: lipsate. śakḷ "to be able": śikṣati.

śak "to endure" (marṣaṇe) is a divādi root, marked by a svarita vowel, so it takes both padas: śikṣati, śikṣate.

One of Bhaṭṭoji's commentators asks whether both śak roots are meant. By a maxim about items named without their markers (niranubandhaka), only śak "to endure" would be taken. He answers that the maxim concerns affixes. So both roots are taken.

pad: pitsate.

Vārttika: rādho hiṃsāyāṃ sanīs vācyaḥ. "Is should be stated [for the ā] of rādh before san, in the sense of injuring (hiṃsā)." ritsati, "he wishes to injure". Why "in the sense of injuring"? In another sense: ārirātsati.

Watch out. Is here is a substitute for a vowel, not the augment iṭ. Its s may change: to t in mitsati (SK 2342), or drop in pitsati (SK 380).

SK 2624 · मुचोऽकर्मकस्य गुणो वा · muco'karmakasya guṇo vā (A 7.4.57)

मुचोऽकर्मकस्य गुणो वा
muco'karmakasya guṇo vā · A 7.4.57
सादौ सनि । अभ्यासलोपः । मोक्षते मुमुक्षते वा वत्सः स्वयमेव । अकर्मकस्य किम् । मुमुक्षति वत्सं कृष्णः ।न वृद्भ्यश्चतुर्भ्यः SK 2348 । विवृत्सति । तङि तु । विवर्तिषते । सेऽसिचि - SK 2506 इति वेट् । निनर्तिषति । निनृत्सति ॥

Rule. Muc "to free", when intransitive (akarmaka), optionally takes guṇa before a san beginning with s. Bhaṭṭoji's gloss adds only "before a san beginning with s". One of his commentators gives the carried words, "before san" from SK 2623 and "s" from SK 2342.

Both commentators give the purpose. San after muc is kit by SK 2613, and kittva would bar guṇa. This rule restores it as an option.

Example.

vatsaḥ svayam eva, "the calf [wishes to get free] of its own accord":

muc + sa → guṇa (this rule) → moc-sa → c → k (SK 378) → ṣ (SK 212) → mokṣa

→ doubling, then the abhyāsa is deleted (SK 2620) → mokṣate.

Without guṇa: mukṣa → doubling; the abhyāsa stays → mumukṣate.

Why "intransitive"? mumukṣati vatsaṃ kṛṣṇaḥ, "Kṛṣṇa wishes to free the calf".

Further forms Bhaṭṭoji gives here.

Watch out. The abhyāsa is lost only in mokṣate, where this rule has acted. mumukṣate keeps it.

G. san on ṇi-final stems, and the ṣ after the abhyāsa: only stu and the ṇi-final — running example: stu + san → stū-ṣa (2614) → tu-ṣṭū-ṣa → tuṣṭūṣati; but sic + sa → si-sik-ṣa → sisikṣati; and svedi + iṣa → sisvedayiṣati (2628)

After an iṇ vowel such as i or u, SK 212 normally turns an s into ṣ. So one would expect ṣ after every abhyāsa in i or u, as in a supposed suṣyūṣati from siv. SK 2627 restricts this. Once the san has itself become ṣ, that s becomes ṣ only in stu "to praise" and in a stem ending in ṇi. SK 2628 then takes three ṇi-final stems back out. Stems in ṇi take san too, giving desideratives of causatives. Bhaṭṭoji recalls rules from chapter 54 for them under SK 2626, so they are set out first. The running examples are tuṣṭūṣati "he wishes to praise", sisikṣati "he wishes to sprinkle", and sisvedayiṣati "he wishes to make sweat".

Recalled under SK 2626: san after a ṇi-final stem. These are not sūtras of this chapter. Bhaṭṭoji repeats rules of the ṇic section (chapter 54) and shows them working before san. Here each ṇi-final stem takes san with iṭ. The ṇi then takes guṇa and ay: -ayiṣati.

The two sūtras ṇau ca saṃścaṅoḥ (SK 2601 and SK 2579). Before ṇi followed by san or caṅ, gāṅ optionally replaces iṅ, and śvi optionally takes saṃprasāraṇa.

One of Bhaṭṭoji's commentators gives these steps. He says saṃprasāraṇa comes before the other operations in the last form.

hvaḥ saṃprasāraṇam (SK 2586). Hve takes saṃprasāraṇa: hu. "Since it has been stated that, before ṇi, no substitute replaces a vowel before doubling", Bhaṭṭoji says, "the u is doubled."

hve + ṇic, san and iṭ → saṃprasāraṇa → hu-i-iṣa

→ hu is doubled → hu hu-i-iṣa

→ h becomes j in the abhyāsa (SK 2245, SK 2182) → ju hu-i-iṣa

→ the second part takes vṛddhi and āv before ṇi → ju hāv-i-iṣa

→ guṇa and ay of the ṇi → juhāvayiṣati.

Likewise pusphārayiṣati, from sphur with ā by cisphuror ṇau (SK 2569). One of Bhaṭṭoji's commentators adds why its abhyāsa keeps u: an s stands between that u and the ph, so no labial follows the u. And cukṣāvayiṣati, from kṣu.

oḥ puyaṇjy apare (SK 2577). The u of an abhyāsa becomes i before a labial (pu), a semivowel (yaṇ) or j that is followed by a.

One of Bhaṭṭoji's commentators gives the steps. The u-final root itself is doubled, since no vowel substitute is made before ṇi while doubling is due. Then the abhyāsa u becomes i. Both commentators note that ju is a root found only in the sūtras (sautra).

sravati… (SK 2578). For sru and five other roots the i is optional: sisrāvayiṣati / susrāvayiṣati, and so on. "Followed by a" still holds: śuśrūṣate "he wishes to hear". Bhaṭṭoji's commentators give its ātmanepada by jñāśrusmṛdṛśāṃ sanaḥ (SK 2731, taught later).

SK 2627 · स्तौतिण्योरेव षण्यभ्यासात् · stautiṇyoreva ṣaṇyabhyāsāt (A 8.3.61)

स्तौतिण्योरेव षण्यभ्यासात्
stautiṇyoreva ṣaṇyabhyāsāt · A 8.3.61
अभ्यासेणः परस्य स्तौतिण्यन्तयोरेव सस्य षः स्यात्षभूते सनि नान्यस्य । तुष्टूषति । द्युतिस्वाप्योः - SK 2344 इत्युत्वम् । सुष्वापयिषति । सिषाधयिषति । स्तौतिण्योः किम् । सिसिक्षति । उपसर्गात्तु स्थादिष्वभ्यासेन च -SK 2277 इति षत्वम् । परिषिषिक्षचि । षणि किम् । तिष्ठासति । सुषुप्सति । अभ्यासादित्युक्तेर्नेह निषेधः । इण् । प्रतीषिषति । इक् । अधीषिषति ॥

Rule. When the san has itself become ṣ (ṣaṇ), an s that follows an iṇ vowel of the abhyāsa becomes ṣ only in stu and in a stem ending in ṇi. Bhaṭṭoji's gloss: "let ṣ be substituted for an s that follows an iṇ of the abhyāsa only in stu and in a ṇi-final [stem], when the san has become ṣ; not for [the s] of any other".

One of Bhaṭṭoji's commentators places the rule.

He also explains the word eva, "only". Without it, the rule might be read as "only before ṣaṇ". Then stu would lose its ṣ in the perfect tuṣṭāva.

Example.

stotum icchati, "he wishes to praise":

stu + sa → no iṭ, since stu ends in uk (SK 2610) → long vowel (SK 2614) → stū-sa → kit (SK 2612)

→ the s of san becomes ṣ after ū (SK 212) → stūṣa: the san has become ṣ

→ doubling → stū stūṣa → in the abhyāsa the t after s stays (śarpūrvāḥ khayaḥ, SK 2259), and the vowel is short → tu stūṣa

→ the root's s follows u, an iṇ of the abhyāsa; the root is stu, so ṣ (this rule) → tu ṣtūṣa

→ t → ṭ after ṣ (SK 113) → tuṣṭūṣati.

One of Bhaṭṭoji's commentators names the lengthening by SK 2614.

ṇi-final stems:

Applies when / blocked by. Bhaṭṭoji tests each word.

Watch out. Three conditions must all hold for the restriction to bite: the s follows an abhyāsa, the vowel is an iṇ, and the san has become ṣ. Then only stu and the ṇi-final take ṣ. This is why you met susyūṣati, sisaniṣati and sismayiṣate with plain s.

SK 2628 · सः स्विदिस्वदिसहीनां च · saḥ svidisvadisahīnāṃ ca (A 8.3.62)

सः स्विदिस्वदिसहीनां च
saḥ svidisvadisahīnāṃ ca · A 8.3.62
अभ्यासेणः परस्य ण्यन्तानामेषां सस्य स एव न षः षणि परे । सिस्वेदयिषति । सिस्वादयिषति । सिसाहयिषति । स्थादिष्वभ्यासेनेति नियमान्नेह । अभिसुसूषति । शैषिकान्मतुबर्थीयाच्छैषिको मतुबर्थिकः । सरूपः प्रत्ययो नेष्टः सन्नन्तान्न नसिष्यते ॥ (भाष्यं) । शैषिकाच्छैषिकः सरूपो न, तेन शालीये भव इति वाक्यमेव । न तु छान्ताच्छः । सरूपः किम् । अहिच्छत्रे भव आहिच्छत्रः । आहिच्छत्रे भव आहिच्छत्रीयः । अण्णन्ताच्छः । तथा मत्वर्थात्सपरूपः स न । धनवानस्यास्ति । इह मतुबन्तान्मतुप् न । विरूपस्तु स्यादेव । दण्डिमती शाला । सरूप इत्यनुषज्यते । अर्थद्वारा सादृश्यं तस्यार्थः । तेन इच्छासन्नन्तात्सन्न । स्वार्थसन्नन्तात्तु स्यादेव । जुगुप्सिषते । मीमांसिषते ॥

Rule. For svid "to sweat", svad and sah, when they end in ṇi, the s after an iṇ of the abhyāsa stays s, not ṣ, when ṣaṇ follows. Bhaṭṭoji's gloss: "let the s of these, when they end in ṇi, following an iṇ of the abhyāsa, be s only, not ṣ, when ṣaṇ follows".

One of Bhaṭṭoji's commentators gives the carried words. SK 2627 continues without stu, and "of s" comes from SK 335. He adds that prescribing s for s is itself a restriction: "s only".

Example.

svedayitum icchati, "he wishes to make [someone] sweat":

svid + ṇic → guṇa of the light penultimate (SK 2189) → sved-i

→ san with iṭ; guṇa and ay of the ṇi → svedayiṣa: the san has become ṣ

→ doubling → first consonant only, short vowel → si svedayiṣa

→ SK 2627 would allow ṣ for a ṇi-final stem; this rule keeps s → sisvedayiṣati.

Likewise sisvādayiṣati (svad) and sisāhayiṣati (sah).

Applies when / blocked by. Bhaṭṭoji adds abhisusūṣati, from su with abhi. The s of the second part stays s by SK 2627, one of Bhaṭṭoji's commentators notes. Might the abhyāsa take ṣ after the upasarga, by SK 2270? No: "because of the restriction sthādiṣv abhyāsena". SK 2277 lets the abhyāsa take ṣ only in sthā and the roots listed with it. The same commentator explains the suspicion so. The other adds that SK 2524 is still needed, for the noun abhisusūḥ.

The Bhāṣya verse. Bhaṭṭoji ends the chapter with a verse he labels as the Bhāṣya's:

"After a śaiṣika [affix], a śaiṣika, and after one in the sense of matup, one in the sense of matup, of the same form (sarūpa), is not wanted; after a sannanta, san is not wanted."

Whether it is the Bhāṣya's own or a vārttika in verse, scholars discuss this.

A śaiṣika is an affix taught under the heading śeṣa, "the remainder", such as those meaning "being in". Matup is the possessive affix -mat / -vat. Bhaṭṭoji explains each clause.

So a desiderative san does not come after a form that already ends in a desiderative san. But it does come after a sannanta whose san is in the root's own sense (svārtha):

gup + san (SK 2393) → jugupsa "to censure" → + the desiderative san (SK 2608) with iṭ → jugupsiṣate, "he wishes to censure".

Likewise mīmāṃsiṣate, from mīmāṃsa (SK 2394).

One of Bhaṭṭoji's commentators explains why "same form" must mean sameness of meaning. Two *san*s cannot differ in sound at all. So only their senses can make them "same" or "different".

Here ends the sannanta-prakriyā, the section on the san-final.

Watch out. The restriction of SK 2627 gives ṣ to every ṇi-final stem. This rule removes three. sisvedayiṣati has s, though suṣvāpayiṣati has ṣ.

Terms used in this chapter

Summary

SKWhat it doesExample
2608desiderative san after a root whose action is wished, same agent; it is ārdhadhātuka and takes iṭpipaṭhiṣati, jighatsati, īrṣyiyiṣati / īrṣyiṣiṣati
2609san (and ktvā) kit after rud, vid, muṣ, grah, svap, pracchrurudiṣati, vividiṣati, mumuṣiṣati
2612jhal-initial san kit after an ik-final rootbubhūṣati, didīṣate
2613jhal-initial san kit after a consonant (class) near an ikbibhitsati, jughukṣati, titṛkṣati / titṛṃhiṣati
2617san with iṭ optionally kit after a consonant-initial, ral-final root with u/i penultimate; penultimate operations beat doublingdidyutiṣate / didyotiṣate; eṣiṣiṣati
2610no iṭ after grah, guh and uk-final rootsjighṛkṣati, suṣupsati, bubhūṣati
2611iṭ after kṝ, gṝ, dṛṅ, dhṛṅ, pracch; vārttika: no long iṭpipṛcchiṣati, cikariṣati, didariṣate; uddidhīrṣuḥ (not this rule: bhvādi dhṛ)
2618optional iṭ after iv-final roots, ṛdh, bhrasj, dambh, śri, svṛ, yu, ūrṇu, bhṛ, jñapi, sandudyūṣati / dideviṣati, śiśrīṣati / śiśrayiṣati, ūrṇunūṣati / ūrṇunuviṣati, ūrṇunaviṣati, siṣāsati / sisaniṣati
2625optional iṭ after vṛṅ, vṛñ and roots in ṝ (in upadeśa)titariṣati, titarīṣati, titīrṣati; dudhvūrṣati
2626iṭ after smi, pūṅ, ṛ, añjū, aś; ucch; recalls for ṇi-final stemssismayiṣate, pipaviṣate, aririṣati, ucicchiṣati; juhāvayiṣati, pipāvayiṣati
2614long vowel before jhal-initial san: final vowel, penultimate of han and of gam for a vowel-rootcicīṣati / cikīṣati, jigīṣati, jighāṃsati
2622optional long penultimate of tan; vārttika: san of apprehensiontitāṃsati / titaṃsati; śvā mumūrṣati
2615gam for iṇ before san, not in making known; vārttika: ik likewisejigamiṣati, pratīṣiṣati, adhijigamiṣati; passive jigāṃsyate
2616gam for iṅ before sanadhijigāṃsate
2619ī for the vowel of āp, jñap, ṛdh before s-initial sanīpsati, īrtsati, jñīpsati
2620the abhyāsa is deleted where SK 2623, 2619, 2621, 2624 applyīpsati; but ardidhiṣati
2621i or ī for the vowel of dambhdhipsati, dhīpsati / didambhiṣati
2623is for the vowel of mī, mā, ghu, rabh, labh, śak, pat, pad; vārttika: rādh in injuringpitsati, mitsati, ditsati, lipsate, śikṣati; ritsati
2624optional guṇa for intransitive mucmokṣate / mumukṣate; mumukṣati vatsam
2627ṣ after an abhyāsa's iṇ, before ṣaṇ, only in stu and ṇi-final stemstuṣṭūṣati, suṣvāpayiṣati; but sisikṣati (no ṣ); pariṣiṣikṣati (ṣ by SK 2277)
2628s stays for ṇi-final svid, svad, sah; Bhāṣya verse: no san after a desiderative sannantasisvedayiṣati; jugupsiṣate, mīmāṃsiṣate
← textbook contentsreading page (Technical) →