PREMABASESiddhānta-kaumudī
Prakaraṇa 56 · SK 2629–2649

तिङन्तयङ्प्रकरणम्

tiṅantayaṅprakaraṇam · 21 sūtras
📖 This chapter in the textbook →outline ↗
show
SK 2629 · A 3.1.22धात्वधिकारः (प्रथमः)
धातोरेकाचो हलादेः क्रियासमभिहारे यङ् ॥ २६२९ ॥
dhātorekāco halādeḥ kriyāsamabhihāre yaṅ
dhātor ekāco halādeḥ kriyāsamabhihāre yaṅ — After a dhātu that has a single vowel (ekāc) and begins with a consonant (hal), yaṅ comes when kriyāsamabhihāra [— repetition or intensity of the action — is to be made manifest].

Rule. The affix yaṅ comes after a dhātu that has a single vowel (ekāc) and begins with a consonant (halādi). It comes when kriyāsamabhihāra is to be made manifest. Bhaṭṭoji's gloss: "kriyāsamabhihāra is repetition (paunaḥpunya) and the sense of intensity (bhṛśa); when that is to be made manifest, let yaṅ come".

In yaṅ, the ṅ is a marker (it). What is added is ya.

The word is "made manifest" (dyotya), not "denoted" (vācya). One of Bhaṭṭoji's commentators gives the reason. What an affix denotes is the principal meaning, the thing qualified (viśeṣya). Desire is that in a desiderative. If yaṅ denoted the repetition, the repetition would become the principal meaning. So yaṅ only makes it manifest.

Example.

bhū "to become" has one vowel, ū, and begins with the consonant bh. "It becomes again and again, or intensely":

bhū + yaṅ (this rule); ṅ is a marker → bhū ya

→ yaṅ is marked with ṅ, so the root's ū takes no guṇa (SK 2217)

→ bhūya ends in yaṅ, so it is a dhātu (SK 2304)

→ the form ending in yaṅ is doubled (SK 2395) → bhū bhū ya

→ the abhyāsa's vowel is shortened (SK 2180) → bhu bhū ya

→ bh in the abhyāsa becomes b (SK 2182) → bu bhū ya

→ guṇa of the abhyāsa's u (SK 2630) → bo bhū ya

→ laṭ; the ātmanepada ending ta, because the stem ends in a ṅ-marked affix (SK 2157) → bobhūya + ta

→ śap (SK 2167) → bobhūya + a + ta

→ a + a → a (pararūpa, SK 191) → bobhūyata

→ the final a of the ending becomes e (SK 2233) → bobhūyate.

Applies when / blocked by. Bhaṭṭoji tests each word of the rule under the next sūtra, SK 2630.

  • "Single vowel": with jāgṛ "to wake", which has two vowels, there is no yaṅ. One says punaḥ punar jāgarti "he wakes again and again".
  • "Beginning with a consonant": with īkṣ "to look", which begins with a vowel, there is no yaṅ. One says bhṛśam īkṣate "he looks intensely".
  • "After a dhātu": this word makes yaṅ an ārdhadhātuka affix (SK 2187). The reason is given under SK 2630.

Can yaṅ come where a word like bhṛśam "intensely" already says the same? Yes, one of Bhaṭṭoji's commentators says. Yaṅ only makes the sense manifest, and what is already expressed can still be made manifest.

But Bhaṭṭoji reports a limit from the Bhāṣya. With ruc "to shine" and śubh "to be beautiful", yaṅ does not come in the sense of intensity. So bhṛśaṃ rocate, bhṛśaṃ śobhate. In repetition it does come: rorucyate, śośubhyate. One of Bhaṭṭoji's commentators gives the ground: the intensity is simply not expressed this way (anabhidhāna), as the Bhāṣya makes clear.

Vārttika. sūcisūtrimūtryaṭyartyaśūrṇotibhyo yaṅ vācyaḥ: "yaṅ should be stated after sūci, sūtri, mūtri, aṭ, ṛ, aśū and ūrṇu". These roots fail the sūtra's tests, one of Bhaṭṭoji's commentators explains.

  • sūci "to inform against", sūtri "to wind", mūtri "to pass urine" have more than one vowel.
  • aṭ "to go", ṛ "to go", aśū "to pervade" begin with a vowel.
  • ūrṇu fails both tests.

Their forms are taught under SK 2630–2633.

Watch out. Yaṅ adds no new meaning that becomes the main one. "He becomes" is still about becoming; yaṅ only colours it as repeated or intense. Contrast the desiderative of chapter 55, where the wish is the main meaning.

Siddhānta-kaumudī
पौनः पुन्यं भृशार्थश्च क्रियासमभिहारस्तस्मिन् द्योत्ये यङ् स्यात् ॥

kriyāsamabhihāra is repetition (paunaḥpunya) and the sense of intensity (bhṛśa). When that is to be made manifest (dyotya), let yaṅ come [after a dhātu of a single vowel that begins with a hal].

Tattvabodhinī
NSP p. 446–447
धातोरेकाचो—॥ क्रियासमभिहार इत्युक्तत्वाद्धातोरेव स्यान्न तु प्रातिपदिकादिति पृच्छति—धातोः किमिति ।—ब्रुवो वचिरित्यादीति । वावच्यते, चाख्यायत इत्यादिरूपसिद्धिः प्रयोजनमित्यर्थः । धातुग्रहणाभावे धिगित्यव्ययादपि स्यात्तस्य क्रियावाचित्वादेकाच्त्वाच्चेति तद्वारणार्थमपि धातुग्रहणमावश्यकमित्याहुः
translator notes
  • commentary BM (absent from our source.json; read in Wikisource uttara-2/0127) on 'tasmin dyotye': if kriyāsamabhihāra were denoted (vācya) by yaṅ, then, since what a pratyaya denotes is the principal [meaning], it would be the qualificand (viśeṣya), as desire is in a sannanta form. It is therefore only made manifest. Followed: 'to be made manifest (dyotya)', and the same in sutra_en.
  • commentary TB (sk.2629.t1, lemma 'dhātor ekāco—'; re-filed from the tail of sk.2628.t1 on 2026-10-05): because 'kriyāsamabhihāre' is said, [yaṅ] would come after a dhātu only and not after a prātipadika; hence the question 'dhātoḥ kim', which Bhaṭṭoji answers under SK 2630. Recorded; the English of this block is not affected.
  • textual Neighbour-filed commentary: this unit has no BM in our source.json. Its TB, together with the TB on SK 2630 ('dhātoḥ kim', 'bruvo vacir ity ādi'), stood at the tail of sk.2628.t1 and was re-filed as sk.2629.t1 on 2026-10-05. BM is missing from our source for the whole yaṅ-prakaraṇa: sk.2628.b1 goes from the end of BM on the sannanta-prakriyā to the heading of the śaiṣika section (the BM on the sarvasamāsaśeṣa that stood between them was re-filed to sk.p22.prose.b1). The BM was read in Wikisource uttara-2/0127–0132.
  • textual The kaumudī agrees with NSP p. 437 (scan page-0447, read), where the heading 'tiṅante yaṅantaprakriyā 53' stands before this sūtra.
SK 2630 · A 7.4.82V
गुणो यङ्लुकोः ॥ २६३० ॥
guṇo yaṅlukoḥ
guṇo yaṅlukoḥ — guṇa replaces [the ik vowel of the abhyāsa — abhyāsasya carried forward (anuvṛtti) from atra lopo 'bhyāsasya, SK 2620 —] when yaṅ follows and when there is luk [of yaṅ].

Rule. Guṇa replaces the ik vowel of the abhyāsa when yaṅ follows, and when there is luk of yaṅ. Luk is deletion of an affix; yaṅluk is the intensive whose yaṅ has been deleted, taught in chapter 57. Bhaṭṭoji's gloss: "let guṇa replace the abhyāsa when yaṅ follows and when there is yaṅluk".

One of Bhaṭṭoji's commentators gives the details. Yaṅlukoḥ is "when yaṅ or luk", and luk means luk of yaṅ, since yaṅ is in context. "Of the abhyāsa" is carried from SK 2620. Since guṇa replaces an ik (SK 34), the abhyāsa must end in i, u, ṛ or ḷ.

Example.

bhū + yaṅ (SK 2629) → bhū ya

→ doubling (SK 2395) → bhū bhū ya

→ shortening (SK 2180) and b for bh (SK 2182) in the abhyāsa → bu bhū ya

→ this rule: guṇa of the abhyāsa's u → bo bhū ya

→ the root's ū stays, because yaṅ is marked with ṅ (SK 2217)

→ laṭ: ātmanepada (SK 2157), śap (SK 2167), pararūpa (SK 191), e (SK 2233) → bobhūyate.

Bhaṭṭoji explains the endings. The form ending in yaṅ is a dhātu by sanādyantā dhātavaḥ (SK 2304), so laṭ and the other lakāras follow. It ends in a ṅ-marked affix, so it takes the ātmanepada. He gives two more forms: bobhūyāṃcakre (liṭ) and abobhūyiṣṭa (luṅ).

Is the form doubled again, for "again and again"? No, one of Bhaṭṭoji's commentators says. Yaṅ already conveys the repetition.

The vārttika roots. Bhaṭṭoji adds that sūci, sūtri and mūtri are roots of the curādi class ending in a. The curādi class is the tenth class of roots, headed by cur.

sūci + yaṅ (vārttika under SK 2629) → sūc i ya

→ ṇi is deleted before an ārdhadhātuka without iṭ (SK 2313) → sūc ya

→ doubling (SK 2395) → sūc sūc ya

→ only the first consonant stays (SK 2179), the vowel is shortened (SK 2180) → su sūc ya

→ this rule: guṇa → so sūc ya

→ laṭ → sosūcyate.

Likewise sosūtryate and momūtryate.

Why does the second s of sosūcyate not become ṣ? Bhaṭṭoji answers: these roots have more than one vowel (anekāc), so they are not taught with ṣ (aṣopadeśa). One of his commentators spells out the point. If sūc were taught with ṣ, its s would be a substitute by SK 2264. A substitute s after the abhyāsa's vowel would become ṣ.

Applies when / blocked by. Here Bhaṭṭoji answers "why 'after a dhātu' in SK 2629?" So that yaṅ may be ārdhadhātuka. Then bruvo vaciḥ (SK 2453) and the like apply before it. One of Bhaṭṭoji's commentators gives the forms: vāvacyate from brū "to speak", and cākhyāyate.

brū + yaṅ → yaṅ is ārdhadhātuka, so vac replaces brū (SK 2453) → vac ya

→ doubling, first consonant only (SK 2395, 2179) → va vac ya

→ lengthening (SK 2632) → vā vac ya → laṭ → vāvacyate.

Watch out. Guṇa falls on the abhyāsa, not on the root. In bobhūyate the first syllable has o, the second keeps ū.

Siddhānta-kaumudī
अभ्यासस्य गुणः स्याद्यङि यङ्लुकि च । सनाद्यन्ताः - SK 2304 इति धातुत्वाल्लडादयः । ङिदसन्तात्वादात्मनेपदम् । पुनः पुनरतिशयेन वा भवति बोभूयते । बोभूयांचक्रे । अबोभूयिष्ट । धातोः किम् । आर्धधातुकत्वं यथा स्यात् । तेन ब्रुवो वचिः - SK 2453 इत्यादि । एकाचः किम् । पुनः पुनर्जागर्ति । हलादेः किम् । भृशमीक्षते । भृशं रोचते शोभते इत्यत्र यङ्गेति भाष्यम् । पौनः पुन्ये तु स्यादेव । रोरुच्यते । शोशुभ्यते ॥

Let guṇa replace [the ik vowel of] the abhyāsa when yaṅ follows and when there is luk of yaṅ. Because [the form ending in yaṅ] is a dhātu by sanādyantāḥ… (SK 2304), laṭ and the other lakāras [come after it]. Because it ends in a ṅ-marked [pratyaya], the ātmanepada [endings come]. [It] comes into being again and again, or intensely: bobhūyate [bhū + yaṅ → doubled by sanyaṅoḥ (SK 2395) → the abhyāsa takes guṇa by this rule, while the ū of the root takes no guṇa because yaṅ is ṅ-marked → bobhūya (the other abhyāsa operations, such as b for bh by abhyāse carca, SK 2182, are left implicit) → laṭ, śap, pararūpa → bobhūyate]. bobhūyāṃcakre. abobhūyiṣṭa. Why [is] 'after a dhātu' [said in SK 2629]? So that [yaṅ] may be ārdhadhātuka. Thereby bruvo vaciḥ (SK 2453) and the like [operate before it: vāvacyate, cākhyāyate]. Why 'having a single vowel'? [No yaṅ in:] punaḥ punar jāgarti ('he wakes again and again'). Why 'beginning with a hal'? bhṛśam īkṣate ('he looks intensely'). In bhṛśaṃ rocate ('it shines intensely') and [bhṛśaṃ] śobhate ('it is intensely beautiful') yaṅ does not [come]: so the Bhāṣya. But in [the sense of] repetition (paunaḥpunya) it does come: rorucyate, śośubhyate.

vārttika
सूचिसूत्रिमूत्र्यट्यर्त्यशूर्णोतिभ्यो यङ् वाच्यः

(vārttika) yaṅ should be stated after sūci, sūtri, mūtri, aṭ, ṛ (arti), aśū and ūrṇu (ūrṇoti).

(वार्तिकम्) ॥ आद्यास्त्रयश्चुरादावदन्ताः । सोसूच्यते । सोसूत्र्यते । अनेकात्त्त्वेनाषोपदेशत्वात्षत्वं न । मोमूत्र्यते ॥

The first three are a-final [roots] in the curādi [class]. sosūcyate [sūci + yaṅ → ṇi elided → doubled → guṇa of the abhyāsa by this rule]. sosūtryate. Because, having more than one vowel (anekāc), [these roots] are not taught with ṣ (aṣopadeśa), there is no ṣ [for the s after the abhyāsa]. momūtryate.

translator notes
  • commentary BM (absent from our source.json; read in Wikisource uttara-2/0127): 'yaṅlukoḥ' is the seventh case (saptamī) of the dvandva 'yaṅ ca luk ca'; luk here means luk of yaṅ only, since yaṅ is what is present in context; 'abhyāsasya' is carried forward (anuvartate) from 'atra lopo 'bhyāsasya'; by the ik-paribhāṣā [iko guṇavṛddhī] 'of [an abhyāsa] ending in ik' is obtained. Followed: the anuvṛtti bracket in sutra_en, 'luk [of yaṅ]', and '[the ik vowel of]' in sutra_en and k1.
  • commentary BM (Wikisource uttara-2/0127–0128) on bobhūyate: doubling by 'sanyaṅoḥ'; the ū takes no guṇa because yaṅ is ṅ-marked; from bobhūya, laṭ, 'tip', śap, pararūpa. There is no further doubling by 'kriyāsamabhihāre dve vācye', which concerns only the loṭ-forms prescribed by 'kriyāsamabhihāre loṭ' (the Bhāṣya there reads it of the loṭ second-person singular), nor by 'nityavīpsayoḥ' for repetition, since the vārttika's doubling of loṭ forms in the sense of repetition shows (jñāpana) that the sūtra-doubling does not operate where repetition is otherwise conveyed. Followed: the bracketed derivation. BM's 'tipi' names a parasmaipada ending, which does not fit the ātmanepada form; it is not carried into the English. Recorded.
  • commentary BM (Wikisource uttara-2/0128) on 'dhātoḥ kim': since 'kriyāsamabhihāre' already yields a dhātu, why 'dhātoḥ'? Without it yaṅ would not be ārdhadhātuka, because only a pratyaya prescribed after 'dhātoḥ' is ārdhadhātuka; only when yaṅ is ārdhadhātuka do 'bruvo vaciḥ', 'veño vayiḥ' and the like, prescribed under the heading 'ārdhadhātuke', take effect before it. Followed: '[said in SK 2629]' and '[operate before it]'.
  • commentary TB, filed in sk.2629.t1: on 'bruvo vacir ity ādi' — the purpose is the derivation of forms such as vāvacyate and cākhyāyate. Others say (āhuḥ) that 'dhātu' is needed also to exclude the indeclinable dhik, which would otherwise take yaṅ because it denotes an action and has one vowel. Followed: the bracketed forms vāvacyate, cākhyāyate; the alternative purpose is recorded.
  • commentary BM (Wikisource uttara-2/0128) on 'bhṛśam īkṣate': it should not be objected that yaṅ would not come here anyway because bhṛśa itself conveys intensity; intensity is made manifest (dyotya) by yaṅ, not denoted, and what is already expressed can still be made manifest. On 'yaṅ neti': [it does not come] because of non-expression (anabhidhāna), as the Bhāṣya makes clear. On 'paunaḥpunye tu': yaṅ comes even with the words punaḥ punaḥ alongside; non-expression is stated for the sense of intensity only. Followed: 'yaṅ does not [come]'.
  • commentary BM (Wikisource uttara-2/0128): the vārttika's compound is a dvandva of sūci, sūtri, mūtri, aṭi, arti, aśū, ūrṇoti. sūca paiśunye, sūtra veṣṭane, mūtra prasravaṇe are a-final in the kathādi subgroup of curādi; [yaṅ] would not obtain for them because they have more than one vowel, hence the statement. aṭa gatau, ṛ gatau, aśa bhojane, aśū vyāptau need it because they do not begin with a hal; ūrṇu because it neither begins with a hal nor has a single vowel. Followed: the root identifications in k2. BM names aśa bhojane as well as aśū vyāptau; the English keeps the vārttika's 'aśū'. Recorded.
  • commentary BM (Wikisource uttara-2/0128) on sosūcyate: from the ṇi-final [root], yaṅ, elision of ṇi, doubling, and guṇa of the abhyāsa by 'guṇo yaṅlukoḥ'. Bhaṭṭoji is warding off the mistaken view that sūc is taught with ṣ: if it were, then after s replaced ṣ by 'dhātvādeḥ ṣaḥ saḥ' (SK 2264), the s of the second part, being a substitute s, would become ṣ. Followed: the bracket after sosūcyate and '[for the s after the abhyāsa]'.
  • textual Collation batch08 and scan, NSP p. 437 (page-0447, read): ours 'ṅidasantātvād ātmanepadam' → print 'ṅidantatvād ātmanepadam' (so Wikisource). The English follows the print. Mūla: for Ajai's printed copy.
  • textual Ours reads 'yaṅgeti bhāṣyam'. At NSP p. 437 (page-0447) the conjunct under ṅ is unclear (collation batch08, class C). Wikisource reads 'yaṅ neti', and BM's lemma is 'yaṅ neti'. The sense requires 'yaṅ na iti', and the English follows that. Check the printed copy.
  • textual Cross-reference: ours gives 'sanādyantāḥ' as SK 2304, and in our numbering SK 2304 is sanādyantā dhātavaḥ (3.1.32). The print ('sanādyantāḥ–2604', NSP p. 437, page-0447, read) and Wikisource ('sū 2604') both have 2604, which in our numbering is doṣo ṇau. Ours is correct and the print number is a misprint. Collation batch08 lists this as an error in ours; it is not.
  • textual Ours 'anekāttttvenāṣopadeśatvāt' is garbled; print (NSP p. 437, page-0447) 'anekāctvenāṣopadeśatvāt'; Wikisource 'anekāckatvena', which is also BM's lemma. The English follows the print. Mūla: for Ajai's printed copy.
  • textual The tag '(vārttikam)' at the head of k3 is editorial. In the print the vārttika (k2) is marked by a sign placed before it, and k3 is Bhaṭṭoji's own prose.
SK 2631 · A 6.4.49V
यस्य हलः ॥ २६३१ ॥
yasya halaḥ
yasya halaḥ — [Elision (lopa) — carried forward (anuvṛtti) from ato lopaḥ, SK 2308 —] of the cluster 'ya' after a hal [— halaḥ a fifth-case word (pañcamyanta) —] [when an ārdhadhātuka follows — ārdhadhātuke, the governing rule (adhikāra), SK 2307].

Rule. The cluster ya is deleted (lopa) after a consonant, when an ārdhadhātuka affix follows. Bhaṭṭoji's gloss: "'yasya' takes the cluster (saṃghāta). Let there be deletion of the item ya that follows a hal, when an ārdhadhātuka follows".

One of Bhaṭṭoji's commentators gives the carried words. "Deletion" comes from ato lopaḥ (SK 2308). "When an ārdhadhātuka follows" is the heading of SK 2307. Halaḥ "after a consonant" is in the fifth case.

How does a deletion of the cluster work? Bhaṭṭoji cites two rules.

  • By ādeḥ parasya (SK 44), the deletion falls on the first sound of ya, the y.
  • The a that is left is deleted by ato lopaḥ (SK 2308).

Example.

liṭ of sosūcya (SK 2630). The stem has many vowels, so liṭ is formed with ām (SK 2306), as in chapter 44:

sosūcya + ām

→ ām is an ārdhadhātuka affix; ya follows the consonant c

→ this rule: y deleted (SK 44) → sosūca + ām

→ the a deleted (SK 2308) → sosūc + ām → sosūcām

→ with liṭ of kṛ after it → sosūcāṃcakre.

Likewise, in luṭ with iṭ: sosūcitā, sosūtritā, momūtritā.

Applies when / blocked by. Why take the cluster ya, and not y alone? Bhaṭṭoji's commentators answer with īrṣyitā, from īrṣy "to envy". There the y follows a consonant, ṣ, but no a follows the y. So the y is kept. One of the commentators adds that taking a cluster brings in the maxim that only a meaningful item (arthavat) is meant. So the ya of putrakāmyā is not deleted.

Why "after a consonant"? The same commentator gives lolūyitā and popūyitā. There ya follows a vowel, ū, and stays.

Watch out. This rule needs a further ārdhadhātuka affix after the yaṅ-stem. In laṭ the affix is śap, a sārvadhātuka, so sosūcyate keeps its ya.

Siddhānta-kaumudī
यस्येति संघातग्रहणम् । हलः परस्य यशब्दस्य लोपः स्यादार्धधातुके । आदेः परस्य SK 44 । अतो लोपः SK 2308 । सोसूचांचक्रे । सोसूचिता । सोसूत्रिता । मोमूत्रिता ॥

'yasya' takes the cluster (saṃghāta) [y + a]. Let there be elision (lopa) of the item 'ya' that follows a hal, when an ārdhadhātuka follows. [The elision falls on its first sound, y, by] ādeḥ parasya (SK 44); [the remaining a is elided by] ato lopaḥ (SK 2308). sosūcāṃcakre [liṭ: sosūcya + ām → ya elided → sosūcām → sosūcāṃcakre]. sosūcitā. sosūtritā. momūtritā.

Tattvabodhinī
NSP p. 447
—यस्य हलः ॥—संघातग्रहणमिति । तेन ईर्ष्यिता अहर्यादित्यादौ यलोपो न । संघातग्रहणेन चार्थवद्ग्रहणपरिभाषोपस्थित्या पुत्रकाम्येत्यादौ नातिप्रसङ्गः । वर्णग्रहणे तु उक्तपरिभाषाया अनुपस्थित्या स्यादेवात्र यलोप इति भावः । हलः किम् । लोलूयिता । पोपूयिता
translator notes
  • commentary BM (absent from our source.json; read in Wikisource uttara-2/0128): 'yasya' is to ward off the mistaken view that the a after y is only for pronunciation; 'saṃghātagrahaṇa' means the cluster of y and a; 'halaḥ' is a fifth-case word (pañcamī); 'ārdhadhātuke' is in force as the governing heading (adhikṛta); 'lopa' is carried forward (anuvartate) from 'ato lopaḥ'. Followed: all three brackets in sutra_en.
  • commentary BM (Wikisource uttara-2/0128–0129): objection — by 'alo 'ntyasya' the elision would fall on the final a of 'ya'; hence 'ādeḥ parasya', which is its exception (apavāda). 'ato lopaḥ': when the y of the cluster is elided, the remaining a goes by 'ato lopaḥ'. Objection: then do not take the cluster; prescribe elision of y alone and let 'ato lopaḥ' give the rest. No: [taking the cluster] serves to stop elision of the y in īrṣyitā (īrṣy + tṛc with iṭ). The derivation starts from 'in liṭ, sosūci ām'. Followed: '[The elision falls on its first sound, y]', '[the remaining a]', and the bracket after sosūcāṃcakre.
  • commentary TB on 'saṃghātagrahaṇam': hence there is no elision of y in forms such as īrṣyitā [OCR 'īrdhvatā'] and 'aharyāt' [so the OCR and the print, NSP p. 437, scan page-0447]. Taking the cluster also brings in the paribhāṣā that a meaningful item is taken (arthavadgrahaṇa), so there is no over-application in putrakāmyā and the like; if the single sound were taken, that paribhāṣā would not apply and y would be elided there. halaḥ kim: lolūyitā, popūyitā. Agrees with BM on 'cluster'; followed: 'takes the cluster (saṃghāta) [y + a]'.
  • textual The kaumudī agrees with NSP p. 437 (scan page-0447, read), including the cross-reference 'ato lopaḥ 2308'. Wikisource has 'sū 2608' there, which is wrong. The OCR page-0447 drops the line with the sūtra and the opening 'yasyeti saṃghātagrah-', but the scan has it.
  • textual Our sk.2631.b-layer is empty: BM on the whole yaṅ-prakaraṇa is missing from our source (see sk.2629). BM was read in Wikisource uttara-2/0128–0129.
SK 2632 · A 7.4.83V
दीर्घोऽकितः ॥ २६३२ ॥
dīrgho'kitaḥ
dīrgho 'kitaḥ — A long vowel replaces [the vowel of] an abhyāsa [— abhyāsasya carried forward (anuvṛtti) from atra lopo 'bhyāsasya, SK 2620 —] that is not k-marked (akit) [when yaṅ follows and when there is luk of yaṅ — yaṅlukoḥ carried forward (anuvṛtti) from guṇo yaṅlukoḥ, SK 2630].

Rule. A long vowel replaces the vowel of an abhyāsa that has no k-marked augment (akit). This happens when yaṅ follows or there is yaṅluk. Bhaṭṭoji's gloss: "let a long vowel replace an abhyāsa that is not k-marked, when yaṅ follows and when there is yaṅluk".

One of Bhaṭṭoji's commentators gives the carried words: "of the abhyāsa" from SK 2620, and "when yaṅ or yaṅluk" from SK 2630.

Example.

aṭ "to go", a root of the vārttika under SK 2629:

aṭ + yaṅ → aṭ ya

→ the root begins with a vowel, so the second part, ṭya, is doubled (SK 2395, 2176) → a ṭya ṭya

→ only the first consonant stays in the abhyāsa: the y goes (SK 2179) → a ṭa ṭya

→ this rule: the abhyāsa's a is lengthened → a ṭā ṭya

→ laṭ → aṭāṭyate.

Bhaṭṭoji's commentators both note that ṭya, with the y of the affix, is what is doubled. Chapter 44 used this form to show that sanyaṅoḥ speaks of the whole form ending in yaṅ.

Applies when / blocked by. Why "with no k-marked augment"? Bhaṭṭoji's commentators point ahead to yaṃyamyate and raṃramyate (SK 2643). There the abhyāsa has the augment nuk, marked with k. So it is not lengthened.

One of Bhaṭṭoji's commentators argues further. The word akitaḥ also shows (it is a jñāpaka) that, among operations on the abhyāsa, the special rules do not block the general ones. His example is ḍoḍhaukyate (SK 2649): there the lengthening, being the later rule, would otherwise block the shortening of the abhyāsa.

Watch out. An a in the abhyāsa becomes ā: vāvrajyate, aṭāṭyate.

Siddhānta-kaumudī
अकितोऽभ्यासस्य दीर्घः स्याद्यङि यङ्लुकि च । अटाट्यते ॥

Let a long vowel replace [the vowel of] an abhyāsa that is not k-marked, when yaṅ follows and when there is luk of yaṅ. aṭāṭyate [aṭ + yaṅ → 'ṭya' doubled by ajāder dvitīyasya (SK 2176) → y of the abhyāsa removed by halādiḥ śeṣaḥ (SK 2179) → aṭaṭya → lengthening of the abhyāsa by this rule → aṭāṭyate].

Tattvabodhinī
NSP p. 447
—दीर्घोऽकितः । अकितः किम् । यंयम्यते । रंरम्यते । ननु दीर्घश्रुत्योपस्थितेनाचा ‘अत्र लोपोऽभ्यासस्य’ इत्यतोऽनुवृत्तमभ्यासस्येत्येतद्विशिष्येत । ततश्चाजन्ताभ्यासस्य दीर्घो भवतीत्यर्थात् पापच्यत इत्यादावेव दीर्घो भवेन्न तु यंयम्यते जङ्घनीतीत्यादौ । तत्र हि परत्वान्नुकि कृतेऽभ्यासस्याजन्तत्वाभावाद्दीर्घस्य प्रसक्तिरेव नास्तीति किमनेनाकित इत्यनेन । न चाभ्यासावयवस्याचो दीर्घ इति वैयधिकरण्यान्वये स्वीकृतेऽकिद्ग्रहणं प्रयोजनवदिति वाच्यम् । संभवति सामानाधिकरण्ये वैयधिकरण्यस्यान्याय्यत्वात् । येन नाप्राप्तिन्यायेन नुको दीर्घापवादत्वस्वीकाराच्च । तस्मादकित इति व्यर्थमिति चेत् । अत्राहुः । डोढौक्यते इत्यत्राभ्यासह्रस्वं बाधित्वा परत्वात् ‘दीर्घोऽकितः’ इति दीर्घे कृते डोढौक्यत इति स्यात् । तस्माद् अभ्यासविकारेषु उत्सर्गविधीन् वाधका न वाधन्त इति ज्ञापनायाकित इति ग्रहणं कर्तव्यमिति । तथा चोक्तज्ञापकनिर्वाहाय ‘दीर्घोऽकितः’ इत्यत्राभ्यासस्याच इति वैयधिकरण्यपक्षस्याश्रयणात् यंयम्यत इत्यादौ दीर्घात्प्रागेव नुकिं कृतेऽपि दीर्घः स्यात्, तद्वारणार्थमकित इत्यावश्यकमिति दिक् ॥—अटाट्यत इति । इह ट्यशब्दस्य द्वित्वम् ॥—यङि च । इह ‘गुणोऽर्तिसंयोगाद्योः’ इत्येतत्सर्वमप्यनुवर्तते ॥—यकारपरस्येति । यदि भाष्योदाहरणं सामान्यापेक्षं ज्ञापकं स्यात्तदाऽर्यमाख्यत् आरर्यदित्यादि सिध्यति
translator notes
  • commentary BM (absent from our source.json; read in Wikisource uttara-2/0129): 'abhyāsasya' because it is carried forward (anuvṛtti) from 'atra lopo 'bhyāsasya'; 'yaṅi yaṅluki ca' because it is carried forward from 'guṇo yaṅlukoḥ'. The purpose of 'akitaḥ' will be stated at yaṃyamyate and the like (SK 2643). Followed: both anuvṛtti brackets in sutra_en.
  • commentary BM (Wikisource uttara-2/0129), lead-in to this sūtra: from aṭ with yaṅ, 'ṭya' is doubled by 'ajāder dvitīyasya'; halādiśeṣa removes the y in the abhyāsa, giving aṭaṭyate. Followed: the bracketed derivation of aṭāṭyate.
  • commentary TB (sk.2632.t1) on 'aṭāṭyata iti': here 'ṭya' is doubled (OCR 'vya', print 'ṭya'). Followed: the doubled element in the bracketed derivation of aṭāṭyate.
  • commentary TB on 'akitaḥ': the counter-examples are yaṃyamyate, raṃramyate. Objection: the long vowel, which brings in 'ac', qualifies 'abhyāsasya', so only a vowel-final abhyāsa is lengthened; after nuk the abhyāsa no longer ends in a vowel, so 'akitaḥ' seems pointless. Answer: in ḍoḍhaukyate the lengthening, being later (para), would block the shortening of the abhyāsa, and a wrong form would result (the print itself reads 'ḍoḍhaukyata iti syāt' (NSP p. 437, scan page-0447, read), and the OCR agrees; this is the print's reading, not an OCR slip, and it does not show the intended wrong form);'akitaḥ' is therefore an indication (jñāpaka) that among operations on the abhyāsa the exceptions (bādhaka) do not block the general rules (utsarga). And for that indication to work, the long vowel is taken as replacing the ac belonging to the abhyāsa, construed in a different case-relation (vaiyadhikaraṇya), so that a long vowel would follow even after nuk in yaṃyamyate; 'akitaḥ' is needed to stop it. Recorded: 'not k-marked' is kept literal in the English, with the counter-examples left to SK 2643.
  • textual The kaumudī agrees with NSP p. 437 (scan page-0447, read). The TB block filed under this unit (sk.2632.t1) also carries the TB on SK 2633 (lemmas 'yaṅi ca' [OCR 'yaṅki ca'] and 'yakāraparasyeti'). BM on this unit is missing from our source (see sk.2629).
SK 2633 · A 7.4.30V
यङि च ॥ २६३३ ॥
yaṅi ca
yaṅi ca — Also when yaṅ follows, [guṇa replaces the ṛ (ṛtaḥ — carried forward (anuvṛtti) from rīṅ ṛtaḥ, SK 1234) of ṛ (arti) and of a root beginning with a conjunct (saṃyogādi) — guṇo 'rtisaṃyogādyoḥ, SK 2380, carried forward (anuvṛtti) as a whole].

Rule. Guṇa replaces the ṛ of ṛ (arti) and of a root beginning with a consonant cluster (saṃyogādi), when yaṅ follows. Bhaṭṭoji's gloss: "let guṇa replace the ṛ of ṛ and of [a root] beginning with a conjunct, when yaṅ follows".

The whole of the preceding Aṣṭādhyāyī rule, guṇo 'rtisaṃyogādyoḥ (SK 2380), is carried forward; Bhaṭṭoji's commentators agree on this. "Of ṛ" (ṛtaḥ) comes from rīṅ ṛtaḥ (SK 1234), one of them adds. Why is this rule needed? Because yaṅ is marked with ṅ, guṇa would otherwise be barred.

Example.

ṛ "to go", a root of the vārttika under SK 2629:

ṛ + yaṅ → ṛ ya

→ this rule: guṇa, a for ṛ, with r after it (raparatva) → ar ya

→ the root begins with a vowel, so the second part, rya, is doubled (SK 2395, 2176) → a rya rya

→ only the first consonant stays in the abhyāsa (SK 2179) → a ra rya

→ the abhyāsa has no k-marked augment: its a is lengthened (SK 2632) → a rā rya

→ laṭ (SK 2157, 2167, 191, 2233) → arāryate.

Why is the r doubled, when na ndrāḥ (SK 2446) forbids doubling an r that begins a cluster? Bhaṭṭoji answers: for an r followed by y there is no such ban. His ground is that arāryate is the Bhāṣya's own example. One of his commentators places that example in the Bhāṣya on SK 2629. He also shows what the ban would give: ya alone would be doubled, and the wrong form *\aryāyate would result. How widely the Bhāṣya's example should be extended, scholars discuss this.

More forms. Bhaṭṭoji gives these.

  • arāritā (luṭ): arārya + i + tā → ya follows the consonant r and an ārdhadhātuka follows, so ya is deleted (SK 2631) → arāritā.
  • aśāśitā, from aś, a root of the vārttika: aś + yaṅ → śya doubled (SK 2176) → a śya śya → y of the abhyāsa goes (SK 2179) → a śa śya → lengthening (SK 2632) → a śā śya → luṭ, ya deleted (SK 2631) → aśāśitā.
  • ūrṇonūyate, from ūrṇu: by na ndrāḥ (SK 2446) the r is not doubled, and nu is doubled instead. Bhaṭṭoji gives the finished form without the remaining steps.
  • bebhidyate, from bhid "to split": bhid ya → doubling → bhid bhid ya → first consonant only (SK 2179) → bhi bhid ya → b for bh (SK 2182) → bi bhid ya → guṇa of the abhyāsa (SK 2630) → be bhid ya → laṭ → bebhidyate.

Applies when / blocked by. In luṭ, bebhidya + i + tā: the ya is deleted (SK 2631). Now the light i of bhid is the penultimate before the affix. Should it take guṇa (SK 2189)? Bhaṭṭoji says no: "because the deletion of a is treated as the original (sthānivat), there is no guṇa of the penultimate". The deleted a still counts as standing there. So bebhiditā.

The conjunct-initial case of this rule, smṛ → sāsmaryate, is recalled by Bhaṭṭoji under SK 2639.

Watch out. This rule does not touch every ṛ. Only ṛ itself and a ṛ-root beginning with a cluster take guṇa. For kṛ, Bhaṭṭoji gives cekrīyate (SK 2639).

Siddhānta-kaumudī
अर्तेः संयोगादेश्च ऋतो गुणः स्याद्यङि । यकारपररेफस्य न द्वित्वनिषेधः । अरार्यते इति भाष्योदाहरणात् । अरारिता । अशाशिता । ऊर्णोनूयते । बेभिद्यते । अल्लोपस्य स्थानिवत्तवान्नोपधागुणः । बेभिदिता ॥

Let guṇa replace the ṛ of ṛ (arti) and of [a root] beginning with a conjunct consonant, when yaṅ follows. For an r followed by y there is no prohibition of doubling, because of the Bhāṣya's example arāryate [ṛ + yaṅ → guṇa by this rule, with r added (raparatva) → ar-ya → 'rya' doubled, the prohibition na ndrāḥ (SK 2446) not applying to this r → halādiḥ śeṣaḥ → a-ra-rya → lengthening of the abhyāsa by dīrgho 'kitaḥ (SK 2632) → arāryate]. arāritā [ya elided by yasya halaḥ (SK 2631)]. aśāśitā [aś + yaṅ → 'śya' doubled → y of the abhyāsa removed by halādiḥ śeṣaḥ → lengthening by dīrgho 'kitaḥ → ya elided by yasya halaḥ]. ūrṇonūyate [ūrṇu-ya → by na ndrāḥ, 'nu' is doubled]. bebhidyate [bhid-ya → bhid doubled → halādiḥ śeṣaḥ → guṇa of the abhyāsa]. Because the elision of a [of yaṅ] is treated as the original (sthānivat), there is no guṇa of the penultimate [i of bhid before tās]: bebhiditā.

translator notes
  • commentary BM (absent from our source.json; read in Wikisource uttara-2/0129): when yaṅ is added to ṛ, guṇa is prohibited because yaṅ is ṅ-marked, so this sūtra is stated; 'guṇo 'rtisaṃyogādyoḥ' is carried forward (anuvartate), and 'ṛtaḥ' from 'rīṅ ṛtaḥ'; hence 'arteḥ…'. Followed: both anuvṛtti brackets in sutra_en.
  • commentary TB, printed under sk.2632.t1 (lemma 'yaṅki ca' in the OCR): here the whole of 'guṇo 'rtisaṃyogādyoḥ' is carried forward (anuvartate). Agrees with BM; followed ('carried forward as a whole').
  • commentary BM (Wikisource uttara-2/0129) on 'yakāraparasya': from ṛ + ya, guṇa a with r added gives ar ya; if na ndrāḥ prohibited doubling of the r, ya alone would be doubled, and with the abhyāsa lengthened by 'dīrgho 'kitaḥ' *aryāyate would result; hence Bhaṭṭoji's statement. 'bhāṣyodāharaṇāt': supply 'under the sūtra dhātor ekācaḥ' (SK 2629). Followed: the bracketed derivation of arāryate.
  • commentary TB (sk.2632.t1) on 'yakāraparasya': if the Bhāṣya's example were an indication (jñāpaka) with general scope, then forms such as āraryat from 'aryam ākhyat' would also be established. Recorded; the English does not decide the scope of the indication.
  • commentary BM (Wikisource uttara-2/0129): arāritā — ya elided by 'yasya halaḥ'; aśāśitā — from aś with yaṅ, 'śya' doubled, y of the abhyāsa removed by halādiśeṣa, lengthening by 'dīrgho 'kitaḥ', ya elided by 'yasya halaḥ'; ūrṇonūyate — from ūrṇu ya, 'nu' is doubled by na ndrāḥ, halādiśeṣa, 'abhyāsadīrgha'; bebhidyate — bhid doubled, halādiśeṣa, guṇa of the abhyāsa. Followed: the brackets after these forms. BM's 'abhyāsadīrgha' for ūrṇonūyate does not match the form's abhyāsa (no, with guṇa) and is not carried into the English. Recorded.
  • commentary BM (Wikisource uttara-2/0129) on 'allopasya sthānivattvāt': from the yaṅ-ending bebhidya, luṭ, tās, iṭ, elision of y and elision of a give bebhiditā; the objection is that guṇa of the light penultimate would apply before tās; the answer is that the elision of a is sthānivat. Followed: '[of yaṅ]' and '[i of bhid before tās]'.
  • textual Collation batch08 and scan, NSP p. 437 (page-0447, read): ours 'sthānivattavān' → print 'sthānivattvān' (so Wikisource). The English follows the print. Mūla: for Ajai's printed copy.
  • textual Ours and the print (NSP p. 437, page-0447) read 'ṛto guṇaḥ syād yaṅi'. Wikisource adds 'yaṅluki ca'. The English follows ours and the print.
  • textual This unit has no TB block in our source.json. Its TB is printed under sk.2632.t1. BM is missing from our source (see sk.2629).
SK 2634 · A 3.1.23V
नित्यं कौटिल्ये गतौ ॥ २६३४ ॥
nityaṃ kauṭilye gatau
nityaṃ kauṭilye gatau — After [a dhātu] in the sense of motion (gati), [yaṅ — carried forward (anuvṛtti) from dhātor ekāco halādeḥ kriyāsamabhihāre yaṅ, SK 2629 —] comes invariably (nityam, i.e. only) in the sense of crookedness (kauṭilya).

Rule. After a root meaning motion (gati), yaṅ comes only when crookedness (kauṭilya) is meant. It does not come in the sense of kriyāsamabhihāra. Bhaṭṭoji's gloss: "after [a root] meaning motion, let yaṅ come only [when] crookedness [is to be made manifest], and not in kriyāsamabhihāra".

"Yaṅ" is carried forward from SK 2629. The word nityam "invariably" here means eva "only", one of Bhaṭṭoji's commentators explains. It is to be read with kauṭilye: "only in crookedness". Bhaṭṭoji's "not in kriyāsamabhihāra" states what "only" shuts out.

Example.

vraj "to go"; "he goes crookedly":

vraj + yaṅ (SK 2629, limited by this rule) → vraj ya

→ doubling (SK 2395) → vraj vraj ya

→ only the first consonant stays in the abhyāsa (SK 2179) → va vraj ya

→ the abhyāsa has no k-marked augment, so its vowel is lengthened (SK 2632) → vā vraj ya

→ laṭ, as for bobhūyate (SK 2157, 2167, 191, 2233) → vāvrajyate.

Applies when / blocked by. Does yaṅ never come after a motion root for "again and again"? Bhaṭṭoji says it does not. One of his commentators reports that the Śabdenduśekhara argues at length that it does; scholars discuss this. In this book, follow Bhaṭṭoji.

Watch out. With a motion root, vāvrajyate does not mean "he goes again and again". It means "he goes crookedly".

Siddhānta-kaumudī
गत्यर्थात्कौटिल्य एव यङ् स्यान्न तु क्रियासमभिहारे । कुटिलं व्रजति वाव्रज्यते ॥

After [a dhātu] meaning motion, let yaṅ come only [when] crookedness (kauṭilya) [is to be made manifest], and not in [the sense of] kriyāsamabhihāra. 'He goes crookedly': vāvrajyate.

Tattvabodhinī
NSP p. 447–448
—नित्यं कौटिल्ये । नित्यशब्दोऽवधारणार्थक इत्याह—कौटिल्य एवेति । तक्रकौण्डिन्यन्यायेनैवेष्टे सिद्धे नित्यग्रहणं तस्य न्यायस्यानित्यत्वज्ञापनार्थम् । तेन ‘मतिबुद्धिपूजार्थेभ्यश्च’ इति वर्तमाने क्तेनावाधनात्पूजार्थेभ्यो भूते क्तः सिध्यति । ततश्च ‘क्तस्य च वर्तमाने’ इति षष्ठी न प्रवर्तत इति ‘पूजितो यः सुरासुरैः’ इति तृतीया साधुरेव ॥—इत्यादीति । जंजभ्यते । दंदश्यते । दंदह्यते । बंभज्यते । पस धातुर्दन्त्यान्तः सौत्रो गत्यर्थ इति स्पश बाधनस्पर्शनयोरित्यत्र माधवः । पंपस्यते । काशिकायां तु तालव्यान्तः क्वचिद्दृश्यते । एवं जञ्जपीति । जञ्जभीति । दन्दहीतीत्यादि यङ्लुकि उदाहर्तव्यम्
translator notes
  • commentary BM (absent from our source.json; read in Wikisource uttara-2/0129–0130): 'nitya' is in the sense of eva ('only') and is to be construed with 'kauṭilye'; 'yaṅ' is carried forward (anuvartate) from 'dhātor ekācaḥ'; hence 'gatyarthāt kauṭilya eva': yaṅ comes after a dhātu used in a sense of motion only when crookedness is to be made manifest. 'na tu kriyāsamabhihāre' states what eva excludes. Followed: the anuvṛtti bracket and 'only' in sutra_en, and '[is to be made manifest]' in k1.
  • commentary TB on 'kauṭilya eva': the word nitya expresses restriction (avadhāraṇa). Since the desired result would already follow by the maxim of buttermilk and Kauṇḍinya (takrakauṇḍinya-nyāya), 'nitya' indicates that this maxim is not invariable. Hence, while 'matibuddhipūjārthebhyaś ca' (SK 3089) is in force in the present, the past kta after roots meaning worship is not blocked; then the sixth case by 'ktasya ca vartamāne' (SK 625) does not apply, and the third case in 'pūjito yaḥ surāsuraiḥ' is correct. Agrees with BM on 'only'; followed.
  • commentary BM (Wikisource uttara-2/0130): in the Śabdenduśekhara it is argued at length that after roots meaning motion yaṅ does come in the sense of kriyāsamabhihāra too. Recorded; the English keeps Bhaṭṭoji's 'na tu kriyāsamabhihāre'.
  • textual The kaumudī agrees with NSP p. 437 (scan page-0447, read), where all of SK 2634 is printed; only its TB runs on to p. 438. The TB block filed under this unit (sk.2634.t1) also carries, after its lemma 'ityādīti', the TB on SK 2638 (jaṃjabhyate…). BM is missing from our source (see sk.2629).
SK 2635 · A 3.1.24V
लुपसदचरजपजभदहदशगॄभ्यो भावगर्हायाम् ॥ २६३५ ॥
lupasadacarajapajabhadahadaśagṝbhyo bhāvagarhāyām
lupasadacarajapajabhadahadaśagṝbhyo bhāvagarhāyām — After lup, sad, car, jap, jabh, dah, daś and gṝ, [yaṅ — carried forward (anuvṛtti) from SK 2629 — comes only (nityam, carried forward (anuvṛtti) from nityaṃ kauṭilye gatau, SK 2634, in the sense of eva)] when the action (bhāva) is censured.

Rule. After lup "to cut", sad, car "to go about", jap "to mutter", jabh, dah, daś and gṝ "to swallow", yaṅ comes only when the action is censured (bhāvagarhā). Bhaṭṭoji's gloss: "after these, let yaṅ come only when the meaning of the dhātu is censured".

One of Bhaṭṭoji's commentators gives the carried words. "Yaṅ" comes from SK 2629. "Nityam", meaning "only", comes from SK 2634. Bhāva here is the meaning of the dhātu, the action itself.

Example.

lup "to cut"; "he cuts in a censurable way":

lup + yaṅ (SK 2629, limited by this rule) → lup ya

→ yaṅ is marked with ṅ, so the root's light u takes no guṇa (SK 2217)

→ doubling (SK 2395) → lup lup ya

→ only the first consonant stays in the abhyāsa (SK 2179) → lu lup ya

→ guṇa of the abhyāsa's u (SK 2630) → lo lup ya

→ laṭ, as for bobhūyate → lolupyate.

Likewise from sad: sāsadyate.

Applies when / blocked by. Several of these roots need more than the plain frame. car takes an anusvāra and u (SK 2636, 2637). jap, jabh, dah and daś take an anusvāra (SK 2638). gṝ takes l for r (SK 2639).

Watch out. lolupyate does not mean "he cuts again and again". With these eight roots, yaṅ carries censure only.

Siddhānta-kaumudī
एभ्यो धात्वर्थगर्हायामेव यङ् स्यात् । गर्हितं लुम्पति लोलुप्यते । सासद्यते ॥

After these, let yaṅ come only when the meaning of the dhātu (dhātvartha) is censured (garhā). 'He cuts in a censurable way': lolupyate. sāsadyate.

translator notes
  • commentary BM (absent from our source.json; read in Wikisource uttara-2/0130): the compound is a dvandva of lup, sad, car, jap, jabh, dah, daś, gṝ; 'yaṅ' is carried forward (anuvartate); 'bhāva' is the meaning of the dhātu, and censure belonging to it is 'bhāvagarhā'; 'nityam' is carried forward, meaning eva. Hence 'ebhyaḥ…'. Followed: the anuvṛtti brackets and 'only' in sutra_en and k1, and 'meaning of the dhātu' for bhāva.
  • textual The kaumudī agrees with NSP p. 437 (scan page-0447, read). No TB on this sūtra: on NSP pp. 437–438 the TB passes from nityaṃ kauṭilye (SK 2634) to 'ityādīti' on SK 2638. BM is missing from our source (see sk.2629).
SK 2636 · A 7.4.87V
चरफलोश्च ॥ २६३६ ॥
caraphalośca
caraphaloś ca — Also for car and phal, [nuk is added to the a (ataḥ nuk — carried forward (anuvṛtti) from nug ato 'nunāsikāntasya, SK 2643) of the abhyāsa (abhyāsasya — carried forward (anuvṛtti) from atra lopo 'bhyāsasya, SK 2620) when yaṅ follows or there is luk of yaṅ (yaṅlukoḥ — carried forward (anuvṛtti) from guṇo yaṅlukoḥ, SK 2630)].

Rule. For car and phal, nuk is added to the a of the abhyāsa, when yaṅ follows or there is yaṅluk. Bhaṭṭoji's gloss: "let nuk be added to the a of the abhyāsa of these two, when yaṅ follows or there is yaṅluk".

One of Bhaṭṭoji's commentators gives the carried words. "nuk to the a" (ataḥ nuk) comes from nug ato 'nunāsikāntasya (SK 2643). That rule precedes this one in the Aṣṭādhyāyī, though the Kaumudī teaches it after. "Of the abhyāsa" comes from SK 2620, and "when yaṅ or yaṅluk" from SK 2630.

Bhaṭṭoji adds: "by nuk the anusvāra is meant (lakṣyate)". One of his commentators traces this to the Bhāṣya on SK 2643's sūtra. Why is it needed? In yaṃyamyate the n of nuk is not pada-final, and the next sound, y, is not a jhal. So SK 123 could give no anusvāra, and a plain n would be heard. Taking nuk to mean the anusvāra gives yaṃyamyate. The same nuk is carried into this rule, so here too it means the anusvāra.

Vārttika. sa ca padāntavad vācyaḥ: "and it [the anusvāra] should be stated to be as if at the end of a pada (padāntavat)". Bhaṭṭoji gives the purpose: "so that vā padāntasya (SK 125) may apply". The homorganic nasal of SK 124 then becomes optional. One of Bhaṭṭoji's commentators adds that this vārttika belongs to SK 2643's sūtra and is carried here.

Example.

car "to go about", with censure (SK 2635):

car + yaṅ → car ya

→ doubling (SK 2395) → car car ya

→ first consonant only in the abhyāsa (SK 2179) → ca car ya

→ this rule: nuk, that is the anusvāra, after the a of the abhyāsa → caṃ car ya

→ u for the a of the root (SK 2637) → caṃ cur ya

→ u lengthened by hali ca (SK 354) → caṃ cūr ya

→ the abhyāsa has a k-marked augment, so no lengthening (SK 2632)

→ by the vārttika, vā padāntasya (SK 125): the anusvāra optionally becomes ñ, homorganic with c → cañ cūr ya or caṃ cūr ya

→ laṭ → cañcūryate or caṃcūryate.

Watch out. Nuk goes after the a of the abhyāsa, at its end (SK 36). It is not put into the root.

Siddhānta-kaumudī
अनयोरभ्यासस्यातो नुक् स्याद्यङ्यङ्लुकोः । नुगित्यनेनानुस्वारो लक्ष्यते ॥

Let nuk be added to the a of the abhyāsa of these two when yaṅ follows or there is luk of yaṅ. By 'nuk' the anusvāra is meant (lakṣyate) [as in the rule that precedes this one in the Aṣṭādhyāyī, nug ato 'nunāsikāntasya (7.4.85), given below at SK 2643, from which nuk is carried forward].

vārttika
स च पदान्तवद्वाच्यः

(vārttika) And it [the anusvāra] should be stated to be as if at the end of a pada (padāntavat).

(वार्तिकम्) ॥ वा पदान्तस्य SK 125 इति यथा स्यात् ॥

So that vā padāntasya (SK 125) may apply.

translator notes
  • commentary BM (absent from our source.json; read in Wikisource uttara-2/0130): 'abhyāsasya' is carried forward (anuvartate) from 'atra lopaḥ…', 'ataḥ nuk' from 'nug ato 'nunāsikāntasya', and 'yaṅlukoḥ' from 'guṇo yaṅlukoḥ'; hence 'anayoḥ…'. In nuk, k is the it and u is for pronunciation. Followed: the three anuvṛtti brackets in sutra_en.
  • commentary BM (Wikisource uttara-2/0130) on 'lakṣyate': in the preceding sūtra (pūrvasūtre) 'nug ato 'nunāsikāntasya', the Bhāṣya says that nuk must be taken as standing for the anusvāra, so that the anusvāra is heard in yaṃyamyate, raṃramyate; otherwise, since no jhal follows, 'naś ca' [naś cāpadāntasya jhali, SK 123] could give no anusvāra and n itself would be heard. Because that nuk is carried forward here, it stands for the anusvāra here too. Followed: the bracket at the end of k1.
  • commentary BM (Wikisource uttara-2/0130) on 'padāntavat': this vārttika belongs to the preceding sūtra 'nug ato 'nunāsikāntasya' and is carried forward (anuvartate) here as well. Recorded.
  • textual The tag '(vārttikam)' at the head of k3 is editorial. In the print (NSP p. 437, scan page-0447, read) the vārttika 'sa ca padāntavad vācyaḥ' (k2) is marked by a sign placed before it, and k3 is Bhaṭṭoji's prose. The kaumudī otherwise agrees with the print.
  • textual No TB on this sūtra (NSP pp. 437–438: the TB passes from SK 2634 to SK 2638). BM is missing from our source (see sk.2629).
  • translator BM's 'pūrvasūtre' means preceding in Aṣṭādhyāyī order (7.4.85 before 7.4.87). In the Siddhānta-kaumudī the order is reversed: Bhaṭṭoji first states the lakṣaṇā here, and at SK 2643 ('nukānusvāro lakṣyata ity uktam') refers back to it. The bracket in k1 is worded to keep both orders visible.
SK 2637 · A 7.4.88V
उत् परस्यातः ॥ २६३७ ॥
ut parasyātaḥ
ut parasyātaḥ — u replaces the a [of car and phal (caraphaloḥ — carried forward (anuvṛtti) from SK 2636)] that follows [the abhyāsa (abhyāsasya — carried forward (anuvṛtti) from atra lopo 'bhyāsasya, SK 2620, and recast as a fifth-case word (pañcamyanta), abhyāsāt, because of 'para')], [when yaṅ follows or there is luk of yaṅ — yaṅlukoḥ carried forward (anuvṛtti) from guṇo yaṅlukoḥ, SK 2630].

Rule. u replaces the a of car and phal that comes after the abhyāsa, when yaṅ follows or there is yaṅluk. Bhaṭṭoji's gloss: "let u replace the a of car and phal that follows the abhyāsa, when yaṅ follows or there is yaṅluk".

One of Bhaṭṭoji's commentators explains the carried words. "Of the abhyāsa" comes from SK 2620. Because it is joined with para "following", it is recast in the fifth case: "after the abhyāsa". "Of car and phal" comes from SK 2636, and "when yaṅ or yaṅluk" from SK 2630.

Example.

caṃ car ya (SK 2636)

→ this rule: u for the a that follows the abhyāsa → caṃ cur ya

→ Bhaṭṭoji: "lengthening by hali ca (SK 354)": the u is the penult of a root ending in r, before the consonant y → caṃ cūr ya

→ ñ optionally for the anusvāra (SK 125) → cañcūryate, caṃcūryate.

phal: phal ya → doubling, first consonant only (SK 2179) → pha phal ya → p for ph (SK 2182) → pa phal ya

→ the anusvāra (SK 2636) → paṃ phal ya

→ this rule → paṃ phul ya

→ optionally m, homorganic with ph (SK 125) → pamphulyate or paṃphulyate.

Applies when / blocked by. Bhaṭṭoji names hali ca for the lengthening without restriction. But hali ca lengthens only the penult of a root ending in r or v. cur ends in r; phul does not. So pamphulyate keeps a short u, as Bhaṭṭoji's own form shows.

Watch out. The u replaces the root's a, not the abhyāsa's. The abhyāsa keeps a and its anusvāra: caṃ-, paṃ-.

Siddhānta-kaumudī
चरफलोरभ्यासात्परस्यात उत्स्याद्यङ्यङ्लुकोः । हलि च SK 354 इति दीर्घः । चञ्चूर्यते । चंचूर्यते । पम्फुल्यते । पंफुल्यते ॥

Let u replace the a of car and phal that follows the abhyāsa, when yaṅ follows or there is luk of yaṅ. [The u is] lengthened by hali ca (SK 354). cañcūryate, caṃcūryate ['he goes about in a censurable way': car + yaṅ → doubled → halādiḥ śeṣaḥ → anusvāra added after the a of the abhyāsa (SK 2636) → by vā padāntasya (SK 125) the anusvāra optionally becomes ñ, homorganic with what follows → a of the second part → u by this rule → u lengthened by hali ca → cañcūryate; without the homorganic nasal, caṃcūryate]. pamphulyate, paṃphulyate [here the homorganic replacement of the anusvāra is m].

translator notes
  • commentary BM (absent from our source.json; read in Wikisource uttara-2/0130): 'abhyāsasya', carried forward (anuvṛtta) from 'atra lopaḥ…', is changed (vipariṇamyate) into the fifth case because it is joined with the word 'para'; 'yaṅlukoḥ' is carried forward from 'guṇo yaṅlukoḥ' and 'caraphaloḥ' as well; hence 'caraphaloḥ…'. Followed: the anuvṛtti brackets in sutra_en.
  • commentary BM (Wikisource uttara-2/0130): 'dīrghaḥ' — supply 'of the u'. cañcūryate means 'he goes in a censurable way': yaṅ, doubling, halādiśeṣa, the anusvāra augment after the a of the abhyāsa, its homorganic ñ by 'vā padāntasya', u for the a in the second part, and its lengthening; caṃcūryate in the alternative without the homorganic nasal; in pamphulyate the homorganic replacement of the anusvāra is m. Followed: '[The u is]' and both bracketed derivations.
  • textual The kaumudī agrees with NSP p. 437 (scan page-0447, read), including the cross-reference 'hali ca 354'. Wikisource has 'sū 654' there, which is wrong (SK 654 is upasarjanaṃ pūrvam). No TB on this sūtra. BM is missing from our source (see sk.2629).
  • translator '[The u is] lengthened by hali ca (SK 354)' is as general in the kaumudī (and in BM, 'ukārasyeti śeṣaḥ') as in the English, but hali ca lengthens the vowel only before r or v, so it operates for car (cūr-ya), not for phal: phul has no r or v, and the kaumudī's own pamphulyate and paṃphulyate keep the short u.
SK 2638 · A 7.4.86V
जपजभदहदशभञ्जपशां च ॥ २६३८ ॥
japajabhadahadaśabhañjapaśāṃ ca
japajabhadahadaśabhañjapaśāṃ ca — Also for jap, jabh, dah, daś, bhañj and paś, [nuk is added to the a of the abhyāsa when yaṅ follows or there is luk of yaṅ — everything carried forward (anuvṛtti) into caraphaloś ca, SK 2636, being carried forward here too].

Rule. For jap, jabh, dah, daś, bhañj and paś, too, nuk is added to the abhyāsa, when yaṅ follows or there is yaṅluk. Bhaṭṭoji's gloss: "let nuk be added to the abhyāsa of these, when yaṅ follows or there is yaṅluk". Everything carried into SK 2636 is carried here too, one of Bhaṭṭoji's commentators says.

Example.

jap "to mutter"; "he mutters in a censurable way" (SK 2635):

jap + yaṅ → jap ya

→ doubling (SK 2395) → jap jap ya

→ first consonant only (SK 2179) → ja jap ya

→ this rule: the anusvāra after the a of the abhyāsa → jaṃ jap ya

→ k-marked augment, so no lengthening (SK 2632)

→ ñ for the anusvāra, homorganic with j (SK 125) → jañ jap ya

→ laṭ → jañjapyate.

Applies when / blocked by. Bhaṭṭoji ends with "and so on". Bhaṭṭoji's commentators fill in the forms: jañjabhyate, dandahyate, dandaśyate, bambhajyate. In bambhajyate the root appears without its own nasal; compare vanīvacyate under SK 2642, where Bhaṭṭoji names that deletion.

What is the last root? Bhaṭṭoji's commentators report two readings. Mādhava takes it as pas, a root found only in the sūtras (sautra), with a dental s, meaning motion. The Kāśikā has it with a palatal ś. One of the commentators gives pampasyate.

Watch out. These roots have no u for their a; that belongs only to car and phal (SK 2637).

Siddhānta-kaumudī
एषामभ्यासस्य नुक् स्याद्यङ्यङ्लुकोः । गर्हितं जपति । जञ्जप्यते इत्यादि ॥

Let nuk be added to the abhyāsa of these when yaṅ follows or there is luk of yaṅ. 'He mutters in a censurable way': jañjapyate, and so on [: jañjabhyate, dandahyate, dandaśyate, bambhajyate, pampasyate].

translator notes
  • commentary BM (absent from our source.json; read in Wikisource uttara-2/0130): everything that is carried forward (anuvṛtta) into 'caraphaloś ca' is carried forward here too; hence 'eṣām…'. Followed: the anuvṛtti bracket in sutra_en.
  • commentary BM (Wikisource uttara-2/0131) on 'ityādi': jañjabhyate, dandahyate, dandaśyate, bambhajyate. Mādhava says that pas is a sūtra-root (sautra) ending in a dental and meaning motion; the Kāśikā says it ends in a palatal. Followed: the first four bracketed forms.
  • commentary TB, printed under sk.2634.t1 (lemma 'ityādīti'): jaṃjabhyate, daṃdaśyate, daṃdahyate, vaṃbhajyate [OCR, for baṃbhajyate]. Mādhava takes pas as a dental-final sautra root meaning motion; pampasyate. In the Kāśikā it is sometimes found as palatal-final. In yaṅluk the examples would be jañjapīti, jañjabhīti, dandahīti and so on. Followed: the bracketed pampasyate, from TB (BM gives no pas form).
  • textual The kaumudī agrees with NSP p. 438 (scan page-0448, read). This unit has no TB block in our source.json; its TB is at the tail of sk.2634.t1. BM is missing from our source (see sk.2629).
SK 2639 · A 8.2.20V
ग्रो यङि ॥ २६३९ ॥
gro yaṅi
gro yaṅi — [l replaces the r (ro laḥ — carried forward (anuvṛtti) from kṛpo ro laḥ, SK 2350)] of gṝ when yaṅ follows.

Rule. l replaces the r of gṝ (girati) when yaṅ follows. Bhaṭṭoji's gloss: "let l replace the r of gṝ when yaṅ follows". Bhaṭṭoji's commentators agree that "l for r" (ro laḥ) is carried from kṛpo ro laḥ (SK 2350). Graḥ is the sixth case of gṝ, one of them adds.

Example.

gṝ "to swallow"; "he swallows in a censurable way" (SK 2635). Bhaṭṭoji names the root by its present form girati, in which the root appears as gir. The form jegilyate shows the same gir before yaṅ; Bhaṭṭoji does not spell out that step.

gir ya

→ this rule: l for r → gil ya

→ doubling (SK 2395) → gil gil ya

→ first consonant only (SK 2179) → gi gil ya

→ g becomes j (SK 2245) → ji gil ya

→ guṇa of the abhyāsa (SK 2630) → je gil ya

→ laṭ → jegilyate.

Applies when / blocked by. The rest of Bhaṭṭoji's text under this sūtra recalls earlier rules. The first four act on the root when yaṅ follows, before doubling. The fifth, suṭ, comes only after doubling.

Ghumāsthā… (SK 2462): ī for the ā of the ghu roots (dā and dhā, so named by SK 2373) and the rest, before a consonant-initial k- or ṅ-marked ārdhadhātuka. Yaṅ is one.

dā + yaṅ → this rule: ī for ā → dī ya

→ doubling, shortening (SK 2395, 2180) → di dī ya

→ Bhaṭṭoji: "guṇa" (SK 2630) → de dī ya

→ laṭ → dedīyate.

Likewise pepīyate (pā) and seṣīyate. One commentator's lead-in names sthā, but sthā would not give this form; its root is the sā (so) of that list.

Vibhāṣā śveḥ (SK 2420): optional saṃprasāraṇa for śvi before yaṅ.

śvi + yaṅ → saṃprasāraṇa, v → u, comes first → śui ya

→ pūrvarūpa (SK 330): u + i → u → śu ya

→ doubling (SK 2395) → śu śu ya

→ the root's u is lengthened before y (SK 2298) → śu śū ya

→ guṇa of the abhyāsa (SK 2630) → śo śū ya

→ laṭ → śośūyate.

Why does saṃprasāraṇa come before doubling? One of Bhaṭṭoji's commentators cites the maxim saṃprasāraṇaṃ tadāśrayaṃ ca kāryaṃ balavat: saṃprasāraṇa, and what depends on it, is the stronger. Without saṃprasāraṇa, the other form is śeśvīyate.

Yaṅi ca (SK 2633): guṇa for a ṛ-root beginning with a cluster.

smṛ + yaṅ → smṛ ya

→ smṛ begins with the cluster sm. Guṇa by SK 2633 competes with rī by SK 1234. SK 2633 is the later rule (para), so it wins → smar ya

→ doubling (SK 2395) → smar smar ya

→ first consonant only (SK 2179) → sa smar ya

→ lengthening (SK 2632) → sā smar ya

→ laṭ → sāsmaryate.

Rīṅ ṛtaḥ (SK 1234): rī for a final ṛ before y.

kṛ + yaṅ → kṛ ya

→ kṛ does not begin with a cluster, so SK 2633 does not apply

→ rī for ṛ, before doubling, because it is the later rule → krī ya

→ doubling (SK 2395) → krī krī ya

→ first consonant only, shortening (SK 2179, 2180) → ki krī ya

→ c for k (SK 2245) → ci krī ya

→ guṇa of the abhyāsa (SK 2630) → ce krī ya

→ laṭ → cekrīyate.

With sam, Bhaṭṭoji adds "suṭ": saṃceskrīyate. The augment suṭ (s) comes before the k of kṛ by SK 2550–2551, one of Bhaṭṭoji's commentators says. It comes only after rī and doubling are done: sam + ce + s + krī ya. The other commentator defends this order at length. If suṭ came first, skṛ would begin with a cluster, guṇa by SK 2633 would win, and rī would be lost. He rests the order on a vārttika: aḍabhyāsavyavāye 'pi suṭ kāt pūrvaḥ, "suṭ comes before the k even when the augment aṭ or the abhyāsa stands between". He discusses how this fits the Bhāṣya's treatment of saṃcaskaratuḥ; scholars discuss this.

Watch out. The first four recalled operations happen to the root before doubling, so both copies show the change: dī/di, śū/śu, smar/sa, krī/ki. suṭ alone comes after doubling.

Siddhānta-kaumudī
गिरते रेफस्य लत्वं स्याद्यङि । गर्हितं गिलति । जेगिल्यते । घुमास्था - SK 2462 इतीत्वम् । गुणः । देदीयते । पेपीयते । सेषीयते । विभाषा श्वेः SK 2420 । शोशूयते । शेश्वीयते । यङि च SK 2623 । सास्मर्यते । रीङृतः SK 1234 । चेक्रीयते । सुट् । संचेस्क्रीयते ॥

Let l replace the r of gṝ (girati) when yaṅ follows. 'He swallows in a censurable way': jegilyate. [For dā, pā and the like:] ī by ghumāsthā… (SK 2462); guṇa [of the abhyāsa]: dedīyate, pepīyate, seṣīyate. [For śvi:] vibhāṣā śveḥ (SK 2420) [optional saṃprasāraṇa]: śośūyate [śvi + yaṅ → saṃprasāraṇa and pūrvarūpa first, saṃprasāraṇa and what depends on it being the stronger → śu doubled → lengthened by akṛtsārvadhātukayor dīrghaḥ (SK 2298) → guṇa of the abhyāsa → śośūyate], śeśvīyate [without saṃprasāraṇa]. [For smṛ:] yaṅi ca (SK 2633): sāsmaryate [smṛ-ya → since the root begins with a conjunct, guṇa by yaṅi ca, which is later (para) [than rīṅ ṛtaḥ], with r added → smar-ya → doubled → lengthening of the abhyāsa by dīrgho 'kitaḥ (SK 2632) → sāsmaryate]. [For kṛ:] rīṅ ṛtaḥ (SK 1234): cekrīyate [kṛ-ya → rīṅ for the ṛ, done first because it is later (para) [than doubling] → krī-ya → krī doubled → guṇa of the abhyāsa → cekrīyate]. [With sam,] suṭ: saṃceskrīyate [suṭ is added after rīṅ and doubling have been done].

Tattvabodhinī
NSP p. 448
—ग्रो यङि । ‘कृपो रो लः’ इत्यतो रो ल इति वर्तते तदाह—रेफस्य लत्वमिति ॥—चेक्रीयते इति । परत्वाद्रीङि कृते द्विर्वचनम् ॥—संचेस्क्रीयते इति । रीङि द्वित्वे च सुट् । न च संकृ इत्यस्यामवस्थायां द्वित्वात्परत्वात्सुटि कृते संयोगादित्वेन रीङ्प्रवृत्तेः प्राक् गुणः स्यादिति वाच्यम् । अडभ्यासव्यवायेऽपीत्यस्यारम्भे अडभ्यासयोः कृतयोरेव सुट् न तु ततः प्रागिति सिद्धान्तात् । एतेन सुटो बहिरङ्गत्वान्न संयोगादित्वेन गुण इति केषांचित्समाधानं परास्तम् । अभ्यासात्परस्य सुटि कृते साभ्यासस्याङ्गस्य संयोगादित्वमृदन्तत्वं च नास्तीति गुणप्राप्तेरभावात् । ननु ‘अडभ्यासव्यवायेऽपि सुट् कात्पूर्वः’ इति वार्तिकप्रत्याख्यानाय ‘पूर्वं धातुरुपसर्गेण युज्यते’ इति पक्षं स्वीकृत्य भाष्यकृता संचस्करतुरित्यादावन्तरङ्गत्वात्सुटि ससुट्कस्यैव द्वित्वाभ्युगमात् संचेस्क्रियते इति न सिध्येत् । तत्र हि सुटि कृते रीङं बाधित्वा परत्वात् ‘यङि च’ इति गुणे संचस्कर्यत इत्यनिष्टरूपप्रसङ्गात् । यदि तु इष्टानुरोधेन पूर्वं धातुः साधनेन युज्यते पश्चादुपसर्गेणेति मताश्रयणेन रीङादेशं द्वित्वं च कृत्वा पश्चात्सुटि इष्टं रूपं सिध्यतीत्युच्यते, तर्हि ‘अडभ्यासव्यवायेऽपि’ इति वचनं स्वीकर्तव्यमेव स्यादिति चेत् । अत्राहुः । अडभ्यासे-त्यादिवार्तिकं विनापि संचस्करतुरित्यादिरूपसिद्धये भाष्यकृता ‘पूर्वं धातुरुपसर्गेण युज्यते’ इति पक्ष आश्रितः । संचेस्क्रीयते समचेस्क्रियत इत्यादिसिद्धिस्तु वार्तिकायत्तैवेति
translator notes
  • commentary BM (absent from our source.json; read in Wikisource uttara-2/0131): 'graḥ' is the sixth-case singular of gṝ; 'ro laḥ' is carried forward (anuvartate) from 'kṛpo ro laḥ'; hence 'girateḥ…'. Followed: the anuvṛtti bracket in sutra_en.
  • commentary TB: 'ro laḥ' continues (vartate) from 'kṛpo ro laḥ'; hence 'rephasya latvam'. Agrees with BM; followed.
  • commentary BM (Wikisource uttara-2/0131): 'dā, pā, sthā ebhyo yaṅi viśeṣam āha — ghumāsthā…' introduces dedīyate, pepīyate, seṣīyate. śośūyate: from śvi with yaṅ, by 'saṃprasāraṇaṃ tadāśrayaṃ ca kāryaṃ balavat' saṃprasāraṇa and pūrvarūpa, then 'śu' is doubled and lengthened by 'akṛtsārvadhātukayoḥ'. sāsmaryate: smṛ ya; since it begins with a conjunct, guṇa by 'yaṅi ca', because that is later (paratvāt); r added; doubling; lengthening of the abhyāsa by 'dīrgho 'kitaḥ'. cekrīyate: kṛ ya; rīṅ for the ṛ because it is later (paratvāt); krī doubled; guṇa of the abhyāsa. suṭ: supply 'by saṃparibhyāṃ karotau bhūṣaṇe and samavāye ca' (SK 2550, 2551). Followed: the bracketed derivations of śośūyate (except its guṇa step, see the translator's note), sāsmaryate and cekrīyate.
  • commentary TB on cekrīyate: 'rīṅ is done first because it is later (para); then doubling'. Followed: '[than doubling]' in the cekrīyate bracket. TB on saṃceskrīyate: suṭ comes after rīṅ and doubling. It should not be argued that at the stage saṃ-kṛ suṭ, being later than doubling, would be made first, so that the root would begin with a conjunct and get guṇa before rīṅ could operate. For once the vārttika 'aḍabhyāsavyavāye 'pi suṭ kāt pūrvaḥ' is admitted, the settled view (siddhānta) is that suṭ comes only after aṭ and the abhyāsa have been made, not before. This also refutes some people's answer that there is no guṇa [from beginning with a conjunct] because suṭ is bahiraṅga: once suṭ comes after the abhyāsa, the aṅga with its abhyāsa neither begins with a conjunct nor ends in ṛ, so guṇa cannot arise. Objection: the Bhāṣyakāra, to do without that vārttika, adopted the view that the dhātu is first joined with the upasarga, and in saṃcaskaratuḥ took the doubling to be of the form with suṭ, suṭ being antaraṅga; then saṃceskrīyate could not be derived, since after suṭ, guṇa by 'yaṅi ca', being later, would block rīṅ and give the unwanted *saṃcaskaryate. If one instead adopts the view that the dhātu is joined first with the means (sādhana) and afterwards with the upasarga, then that vārttika must be accepted after all. Answer: the Bhāṣyakāra adopted the view that the dhātu is first joined with the upasarga to derive saṃcaskaratuḥ and the like without the vārttika, but the derivation of saṃceskrīyate, samaceskrīyata and the like does depend on the vārttika. Followed: '[suṭ is added after rīṅ and doubling have been done]'.
  • translator BM introduces dedīyate, pepīyate, seṣīyate as from dā, pā and sthā. seṣīyate is evidently from so (sā), which ghumāsthā… lists as 'sā'; sthā would not give this form. The English therefore says '[For dā, pā and the like]' and does not name the third root.
  • textual Cross-reference: ours reads 'yaṅi ca <SK2623>' (SK 2623 is sani mīmāghu…). The print reads 'yaṅi ca 2633' (NSP p. 438, scan page-0448, read), and so does Wikisource; yaṅi ca is SK 2633. The English gives SK 2633. Mūla: for Ajai's printed copy. The rest of the kaumudī agrees with the print.
  • textual BM is missing from our source (see sk.2629). It was read in Wikisource uttara-2/0131.
  • translator Two steps in the k1 brackets are the translator's, not BM's. (1) In the śośūyate derivation, BM ends with the lengthening by akṛtsārvadhātukayor dīrghaḥ; the following '→ guṇa of the abhyāsa' is supplied from guṇo yaṅlukoḥ (SK 2630), which the kaumudī's 'guṇaḥ' in the preceding line covers. (2) BM says only 'paratvāt' for the guṇa of smṛ by yaṅi ca; '[than rīṅ ṛtaḥ]' names the rule it is later than (yaṅi ca, 7.4.30, comes after rīṅ ṛtaḥ, 7.4.27).
SK 2640 · A 8.3.112V
सिचो यङि ॥ २६४० ॥
sico yaṅi
sico yaṅi — [A retroflex (mūrdhanya) — mūrdhanyaḥ, the governing rule (adhikāra) —] does not [— na carried forward (anuvṛtti) from na raparasṛpi…, SK 3168 —] replace the s [— saḥ, a sixth-case word (ṣaṣṭhyanta), carried forward (anuvṛtti) from saheḥ sāḍaḥ saḥ, SK 335 —] of sic when yaṅ follows.

Rule. ṣ does not replace the s of sic when yaṅ follows. Bhaṭṭoji's gloss: "let ṣ not replace the s of sic when yaṅ follows".

One of Bhaṭṭoji's commentators gives the sources. "Of s" (saḥ, a sixth case) is carried from saheḥ sāḍaḥ saḥ (SK 335). "Not" (na) is carried from na raparasṛpi… (SK 3168), taught later. The ṣ itself is not carried: it comes from the heading "a retroflex" (mūrdhanyaḥ), under which these rules stand.

Example.

ni + sic + yaṅ → ni sic ya

→ doubling (SK 2395) → ni sic sic ya

→ first consonant only (SK 2179) → ni si sic ya

→ guṇa of the abhyāsa (SK 2630) → ni se sic ya

→ two ṣ would now be due. Bhaṭṭoji's commentators name them:

  • for the abhyāsa's s, after the prefix ni, by upasargāt sunoti… (SK 2270);
  • for the s after the abhyāsa, by sthādiṣv abhyāsena ca… (SK 2277)

→ this rule blocks both → ni se sic ya

→ laṭ → nisesicyate.

Applies when / blocked by. Why "when yaṅ follows"? One of Bhaṭṭoji's commentators gives the desiderative abhiṣiṣikṣati. There san follows, and the ṣ does come.

Watch out. The rule blocks ṣ in both places, the abhyāsa and the root. nisesicyate has no ṣ at all.

Siddhānta-kaumudī
सिचः सस्य षो न स्याद्यङि । निसेसिच्यते ॥

Let ṣ not replace the s of sic when yaṅ follows [— 's' and the negation 'na' being carried forward (anuvṛtti), and 'ṣ' (mūrdhanyaḥ) coming under the governing rule (adhikāra), as stated under the sūtra]: nisesicyate [ni + sic; neither s becomes ṣ].

Tattvabodhinī
NSP p. 448
—सिचो यङि । अभ्याससकारस्य ‘उपसर्गात्सुनोति—’ इति, ततः परस्य तु ‘स्थादिष्वभ्यासेन च—’ इति षत्वे प्राप्ते निषेधोऽयम् । यङि किम् । अभिषिषिक्षति
translator notes
  • commentary BM (absent from our source.json; read in Wikisource uttara-2/0131): 'saḥ', a sixth-case word, is carried forward (anuvartate) from 'saheḥ sāḍaḥ saḥ' [SK 335]; 'mūrdhanyaḥ' is the governing rule (adhikṛta); 'na' is carried forward from 'na raparasṛpi…' [SK 3168]; hence Bhaṭṭoji's gloss. In nisesicyate this rule prohibits the ṣ that would come by 'upasargāt sunoti…' [SK 2270] and by 'sthādiṣv abhyāsena ca…' [SK 2277]. Followed: the three anuvṛtti brackets in sutra_en.
  • commentary TB (sk.2640.t1): ṣ would obtain for the s of the abhyāsa by 'upasargāt sunoti…' [SK 2270], and for the s after it by 'sthādiṣv abhyāsena ca…' [SK 2277]; this is the prohibition of that. Why 'when yaṅ follows'? abhiṣiṣikṣati [desiderative, where ṣ does come]. Followed: '[neither s becomes ṣ]'.
  • textual Our kaumudī matches NSP p. 438 (scan page-0448, read); the OCR misreads the sūtra as 'siccho' and ṣa as pa. No BM block is filed under this unit (none of the p56 units has one in our source.json); BM was read in Wikisource uttara-2/0131.
SK 2641 · A 7.4.63V
न कवतेर्यङि ॥ २६४१ ॥
na kavateryaṅi
na kavater yaṅi — The palatal substitute (cutva) [— cuḥ carried forward (anuvṛtti) from kuhoś cuḥ, SK 2245 —] does not replace [the velar] of the abhyāsa [— abhyāsasya carried forward (anuvṛtti) from atra lopo 'bhyāsasya, SK 2620 —] of ku as named 'kavati' [the bhvādi root kuṅ] when yaṅ follows.

Rule. The palatal substitute (cutva) does not replace the velar of the abhyāsa of kavati, when yaṅ follows. Bhaṭṭoji's gloss: "let the palatal substitution not take place in the abhyāsa of kavati, when yaṅ follows".

One of Bhaṭṭoji's commentators gives the carried words. "The palatal" (cuḥ) comes from kuhoś cuḥ (SK 2245). "Of the abhyāsa" comes from SK 2620.

Kavati is not a root "kav". It is the root kuṅ "to sound" of the bhvādi class, named by its form with śap: kav-a-ti.

Example.

ku + yaṅ → the root's u is lengthened before y (SK 2298) → kū ya

→ doubling (SK 2395) → kū kū ya

→ shortening (SK 2180) → ku kū ya

→ kuhoś cuḥ (SK 2245) would give c for k; this rule blocks it → ku kū ya

→ guṇa of the abhyāsa (SK 2630) → ko kū ya

→ laṭ → kokūyate.

Applies when / blocked by. Bhaṭṭoji: "but for kauti and kuvati, cokūyate". These are two other roots ku: one of the adādi class (kauti) and one of the tudādi class (kuvati). For them the palatal comes: ku kū ya → cu kū ya → co kū ya → cokūyate.

Why name the root with śap at all, when "of ku" might do? Bhaṭṭoji's commentators agree that the naming picks out the bhvādi root. Which of the other two it chiefly serves to exclude, scholars discuss this. One commentator appeals to the maxim that a root whose class-sign is lost (luk-vikaraṇa) is not taken when another could be.

Watch out. Only the bhvādi root is protected. The same sounds from another class give cokūyate.

Siddhānta-kaumudī
कवतेरभ्यासस्य चुत्वं न स्याद्यङि । कोकूयते । कौतिकुवत्योस्तु चोकूयते ॥

Let the palatal substitution (cutva) [— carried forward (anuvṛtti) from SK 2245] not take place in the abhyāsa [— carried forward (anuvṛtti) from SK 2620] of kavati [: the bhvādi root kuṅ, named with śap] when yaṅ follows: kokūyate. But for kauti and kuvati [the cutva does take place]: cokūyate.

Tattvabodhinī
NSP p. 448
—न कवतेर्यङि । इह कोरित्येव वाच्ये कवतेरिति शपा निर्देशात्कुङ् शब्द इति भ्वादिरेव गृह्यते न तु कु शब्द इति अदादिः कूङ् शब्द इति तुदादिश्चेत्याशयेनाह—कौतिकुवत्योस्त्विति । ननु कोरित्युक्तेऽपि कूङः प्रसङ्गो नास्ति दीर्घान्तत्वाल्लुग्विकरणपरिभाषया च कौतेरपि प्रसङ्गो नास्तीति चेत् । अत्राहुः । कोरित्युक्ते तु निरनुबन्धपरिभाषाया अपि जागरूकत्वादुभयोर्ग्रहणं स्यादिति कौतिव्यावृत्त्यर्थं शपा निर्देशः कर्तव्यः । कृते च तस्मिन्कुवतिमपि व्यावर्तयतीत्यत्र तात्पर्यमिति कुङ इत्ये-वात्र सुवचम् ॥—इत्यादीति । दनीध्वस्यते । बनीभ्रस्यते । चनीकस्यते । पनीपत्यते । पनीपद्यते । चनीस्कद्यते । एवं वनीवञ्चीति । सनीस्रंसीति । दनीध्वंसीतीत्यादि यङ्लुक्युदाहर्तव्यम् ॥—घ्नीभाव इति । ह्नीभावे कृतेऽपि ‘अभ्यासाच्च’ इति कुत्वेन जेघ्नीयत इति सिध्यति, तथापि प्रक्रियालाघवाय घ्नी इत्युक्तमिति मनोरमा । अन्ये तु ‘सेर्हिः’ ‘मेर्निः’ इति इकारोच्चारणसामर्थ्याद्यथा ‘एहः’ इत्युत्वं न भवति तथा ह्नीभावे हि कृते ‘अभ्यासाच्च’ इति कुत्वं न स्यादित्याशयेन घ्नीभावः कृत इत्याहुः । अत्र च दिग्यादेशेनैव घ्नीभावेन द्वित्वं न बाध्यते पुनःप्रसङ्गविज्ञानादिति केचित् । वस्तुतस्तु यङन्तावयवस्य द्वितीयाजवधिकस्य द्वित्वं घ्नीभावस्तु प्रकृतिमात्रस्येति विषयभेदाद्वित्वं निर्वाधमेव
translator notes
  • commentary BM (absent from our source.json; read in Wikisource uttara-2/0131): Bhaṭṭoji glosses 'kavater abhyāsasya' on the view that 'abhyāsasya' is carried forward (anuvartate) from 'atra lopo [’bhyāsasya]' [SK 2620] and 'cuḥ' from 'kuhoś cuḥ' [SK 2245]. Followed: both anuvṛtti brackets.
  • commentary BM (Wikisource uttara-2/0131–0132) on 'kautikuvatyos tu': objection — 'koḥ yaṅi' would suffice, so the designation with śtip (śtipā nirdeśaḥ, 'kavati') is pointless. Reply: 'kavati' is taken in order to designate the root with śap (śapā nirdeśārthaṃ kavatigrahaṇam), so 'ku śabde' with luk of the vikaraṇa [adādi] and 'kuṅ śabde' with śa [tudādi] are not taken. Nor can it be said that with 'koḥ' the long-final śa-root would not be included anyway: the tudādi root is read with short final ('kuṅ'), as the Nyāsa accepts. In fact, on the Nyāsa's reading, the only result is that the tudādi root does not undergo [the rule]; the luk-vikaraṇa root is excluded already by the paribhāṣā that of a luk-vikaraṇa and a non-luk-vikaraṇa root only the latter is taken. Followed: '[: the bhvādi root kuṅ, named with śap]'.
  • commentary TB (sk.2641.t1): though 'koḥ' would have done, 'kavateḥ' names the root with śap, so that only bhvādi 'kuṅ śabde' is taken, not adādi 'ku śabde' nor tudādi 'kūṅ śabde' [so the OCR; the scan has the same lemma]; hence 'kautikuvatyos tu'. Objection: with 'koḥ' the long-final kūṅ would not arise, and kauti would be excluded by the luk-vikaraṇa paribhāṣā. Reply: with 'koḥ' the paribhāṣā about roots without anubandha would also be in force, so both would be included; the śap-designation must be made to exclude kauti, and once made it excludes kuvati too — that is the intent; 'kuṅaḥ' would equally well have been said. TB thus differs from BM on which root the designation is chiefly meant to exclude. Recorded; the English names neither purpose.
  • textual Our kaumudī matches NSP p. 438 (scan page-0448, read); the OCR misreads cutvam as 'sutvam' and kokūyate as 'kokryate'. Neighbour-filed commentary: our sk.2641.t1 carries, after TB on this sūtra, TB on SK 2642 (from 'ity ādīti', the further examples) and on the vārttika of SK 2643 (from the lemma 'ghnībhāva iti', OCR 'grībhāva'), all printed in sequence on NSP p. 438.
SK 2642 · A 7.4.84V
नीग्वञ्चुस्रंसुध्वंसुभ्रंसुकसपतपदस्कन्दाम् ॥ २६४२ ॥
nīgvañcusraṃsudhvaṃsubhraṃsukasapatapadaskandām
nīg vañcusraṃsudhvaṃsubhraṃsukasapatapadaskandām — The augment nīk comes [to the abhyāsa — abhyāsasya carried forward (anuvṛtti) from atra lopo 'bhyāsasya, SK 2620 —] of vañc, sraṃs, dhvaṃs, bhraṃs, kas, pat, pad and skand [when yaṅ or yaṅ-luk follows — yaṅlukoḥ carried forward (anuvṛtti) from guṇo yaṅlukoḥ, SK 2630].

Rule. The augment nīk comes to the abhyāsa of vañc, sraṃs, dhvaṃs, bhraṃs, kas, pat, pad and skand, when yaṅ follows or there is yaṅluk. Bhaṭṭoji's gloss: "let the augment nīk come to the abhyāsa of these, when yaṅ or yaṅluk follows".

"Of the abhyāsa" comes from SK 2620, one of Bhaṭṭoji's commentators says. "When yaṅ or yaṅluk" is the phrase of SK 2630. The same commentator names that rule as its source under SK 2643 and SK 2644, though not here.

Example.

vañc + yaṅ → vañc ya

→ doubling (SK 2395) → vañc vañc ya

→ first consonant only (SK 2179) → va vañc ya

→ this rule: nīk, marked with k, goes at the end of the abhyāsa (SK 36) → va nī vañc ya

→ Bhaṭṭoji: "because [SK 2632] says akitaḥ, there is no lengthening"

→ Bhaṭṭoji: "deletion of n": the root's penultimate nasal goes before the ṅ-marked yaṅ (SK 415) → va nī vac ya

→ laṭ → vanīvacyate.

Likewise sanīsrasyate, from sraṃs.

Applies when / blocked by. One of Bhaṭṭoji's commentators adds a point on timing. The lengthening would be due before the augment has come. Still the k-marking blocks it, though the augment is yet to come.

For "and so on", Bhaṭṭoji's commentators give danīdhvasyate, banībhrasyate, canīkasyate, panīpatyate, panīpadyate, and from skand canīskadyate (so one commentator spells it; the other's text keeps the n). In canīkasyate, note c for k in the abhyāsa (SK 2245): ka- becomes ca-. canīskadyate is given as the commentators print it. One of the commentators adds forms with yaṅluk, such as vanīvañcīti.

Watch out. The vowel of the abhyāsa stays short a: va-, sa-, not vā-. The k of nīk sees to that.

Siddhānta-kaumudī
एषामभ्यासस्य नीगागमः स्याद्यङ्यङ्लुकोः । अकित इत्युक्तेर्न दीर्घः । नलोपः । वनीवच्यते । सनीस्रस्यते इत्यादि ॥

Let the augment nīk come to the abhyāsa [— carried forward (anuvṛtti) from SK 2620] of these when yaṅ or yaṅ-luk follows [— carried forward (anuvṛtti) from SK 2630]. Because [dīrgho 'kitaḥ, SK 2632] says 'of [an abhyāsa] without k (akitaḥ)', there is no lengthening [of the abhyāsa vowel]. [Then] deletion of the n [of the root, by aniditāṃ hala upadhāyāḥ kṅiti, SK 415]: vanīvacyate [vañc]. sanīsrasyate [sraṃs], and so on.

translator notes
  • commentary BM (absent from our source.json; read in Wikisource uttara-2/0132–0133): 'abhyāsasya' because it is carried forward (anuvṛtti) from 'atra lopaḥ' [SK 2620]; Bhaṭṭoji's 'yaṅlukoḥ' means 'when yaṅ follows and when yaṅ-luk follows' (yaṅi yaṅluki ca). Under this sūtra BM names no source for 'yaṅlukoḥ'. Followed: the abhyāsasya anuvṛtti bracket in sutra_en and k1.
  • commentary BM (Wikisource uttara-2/0132): 'na dīrghaḥ' means the lengthening of the abhyāsa that would obtain before the augment [BM writes 'rīgāgamāt prāk', evidently for nīk] does not come; the k-marking, though it is still to come, blocks the lengthening. 'nalopaḥ': the n is deleted by 'aniditām' [SK 415]. 'ity ādi': sanīsrasyate, danīdhvasyate, banībhrasyate, canīkasyate, panīpatyate, panīpadyate, canīskandyate (so Wikisource, with the n of skand kept; TB has canīskadyate). Followed: '[dīrgho 'kitaḥ, SK 2632]', '[of the abhyāsa vowel]' and the rule for the n-deletion.
  • commentary TB, printed under sk.2641.t1, on 'ity ādīti': danīdhvasyate, banībhrasyate, canīkasyate, panīpatyate, panīpadyate, canīskadyate (read from the scan, NSP p. 438; the OCR has 'panīpatyate' twice and 'canīskathate'), and in yaṅ-luk vanīvañcīti, sanīsraṃsīti, danīdhvaṃsīti and so on are to be cited. Recorded; agrees with BM on the yaṅ forms.
  • textual Our kaumudī matches NSP p. 438 (scan page-0448, read). This unit has no TB block of its own: TB on it is printed under sk.2641.t1 (lemma 'ity ādīti'). No BM in our source; read in Wikisource.
  • translator The source of 'yaṅlukoḥ' given in sutra_en and k1 (carried forward from guṇo yaṅlukoḥ, SK 2630) is the translator's inference: BM does not state it under this sūtra, but names guṇo yaṅlukoḥ as the source of the same anuvṛtti under SK 2643 and SK 2644.
SK 2643 · A 7.4.85V
नुगतोऽनुनासिकान्तस्य ॥ २६४३ ॥
nugato'nunāsikāntasya
nug ato 'nunāsikāntasya — The augment nuk comes to [the abhyāsa — abhyāsasya carried forward (anuvṛtti) from atra lopo 'bhyāsasya, SK 2620 —] ending in short a, of [an aṅga — aṅgasya, the governing rule (adhikāra) —] ending in a nasal [when yaṅ or yaṅ-luk follows — yaṅlukoḥ carried forward (anuvṛtti) from guṇo yaṅlukoḥ, SK 2630].

Rule. The augment nuk comes to an abhyāsa ending in short a, which belongs to an aṅga ending in a nasal. An aṅga is the base before an affix. This happens when yaṅ follows or there is yaṅluk. Bhaṭṭoji's gloss: "let nuk come to the abhyāsa ending in a which belongs to an aṅga ending in a nasal".

One of Bhaṭṭoji's commentators gives the details. "Of the aṅga" is the heading. "Of the abhyāsa" comes from SK 2620, and "when yaṅ or yaṅluk" from SK 2630.

Bhaṭṭoji adds: "it has been said that by nuk the anusvāra is meant". He said so under SK 2636. One of his commentators credits the same point to the Bhāṣya, so the words "it has been said" may point to either.

Example.

yam "to restrain":

yam + yaṅ → yam ya

→ doubling (SK 2395) → yam yam ya

→ first consonant only (SK 2179) → ya yam ya

→ this rule: the abhyāsa ends in a; the aṅga yam ends in the nasal m → the anusvāra → yaṃ yam ya

→ k-marked augment, so no lengthening (SK 2632)

→ by the vārttika under SK 2636, vā padāntasya (SK 125): the anusvāra optionally becomes a nasalised y → yay̐ yam ya or yaṃ yam ya

→ laṭ → yay̐yamyate or yaṃyamyate.

Applies when / blocked by. The rule says ataḥ, "of short a", with the marker t. By the force (sāmarthya) of this t, nuk does not come even to a short a that was long before (bhūtapūrvadīrgha). One of Bhaṭṭoji's commentators explains why the t is needed. The abhyāsa is always shortened, so a long vowel could never be there. The t would have no work unless it shut out a short a that stands for an original ā.

bhām "to be angry":

bhām + yaṅ → bhām ya

→ doubling (SK 2395) → bhām bhām ya

→ first consonant only, the vowel shortened (SK 2179, 2180) → bha bhām ya

→ this a was long before, so no nuk

→ b for bh (SK 2182) → ba bhām ya

→ no k-marked augment, so the abhyāsa is lengthened (SK 2632) → bā bhām ya

→ laṭ → bābhāmyate.

Bhaṭṭoji recalls ye vibhāṣā (SK 2319): ā optionally replaces the n of jan before y.

  • With the ā: jā ya → doubling (SK 2395) → jā jā ya → shortening (SK 2180) → ja jā ya → lengthening (SK 2632) → jā jā ya → laṭ → jājāyate. No nuk comes, because the aṅga jā does not end in a nasal.
  • Without it, the root ends in a nasal: ja jan ya → this rule → jaṃ jan ya → ñ (SK 125) → jañjanyate.

Vārttika. hanter hiṃsāyāṃ yaṅi ghnībhāvo vācyaḥ: "for han in the sense of injuring, ghnī should be stated [as its substitute] when yaṅ follows".

han + yaṅ, in the sense of injuring → by the vārttika, ghnī replaces han → ghnī ya

→ doubling (SK 2395) → ghnī ghnī ya

→ first consonant only, the vowel shortened (SK 2179, 2180) → ghi ghnī ya

→ gh becomes jh (SK 2245), then j (SK 2182) → ji ghnī ya

→ guṇa of the abhyāsa (SK 2630) → je ghnī ya

→ laṭ → jeghnīyate "he kills again and again, or intensely".

Why does ghnī come before doubling? One of Bhaṭṭoji's commentators reads yaṅi as a locative of domain (viṣayasaptamī): "when yaṅ is intended". So ghnī is put in before the doubling, though the doubling is the later rule.

Why "in the sense of injuring"? Bhaṭṭoji gives jaṅghanyate, from han in its other sense, going.

han ya → doubling, first consonant only → ha han ya

→ this rule: the root ends in the nasal n → haṃ han ya

→ the abhyāsa's h becomes jh (SK 2245), then j (SK 2182) → jaṃ han ya

→ after an abhyāsa, the h of han becomes gh (SK 2430) → jaṃ ghan ya

→ ṅ for the anusvāra, homorganic with gh (SK 125) → jaṅghanyate.

Why is the substitute ghnī and not hnī? With hnī, the h would become gh anyway by SK 2430. Bhaṭṭoji's commentators say ghnī is stated for lightness of the derivation (prakriyālāghava). One of them also records an opposed view; scholars discuss this.

Watch out. Nuk needs a short a in the abhyāsa that was short from the start. bābhāmyate has none, and jeghnīyate has i, not a.

Siddhānta-kaumudī
अनुनासिकान्तस्याङ्गस्य योऽभ्यासोऽदन्तस्तस्य नुक् स्यात् । नुकानुस्वारो लक्ष्यत इत्युक्तम् । यँय्यम्यते । यंयम्यते । तपरत्वसामर्थ्याद्भूतपूर्वदीर्घस्यापि न । भाम क्रोधे । बाभाम्यते । ये विभाषा SK 2319 । जाजायते । जञ्जन्यते ।

Let nuk come to the abhyāsa ending in [short] a which belongs to an aṅga ending in a nasal [— 'abhyāsa' and 'when yaṅ or yaṅ-luk follows' carried forward (anuvṛtti) from SK 2620 and SK 2630]. It has been said [at SK 2636] that by nuk the anusvāra is meant (lakṣyate): yay̐yamyate [with the nasalized y], yaṃyamyate [with anusvāra]. By the force (sāmarthya) of the t-marking (taparatva) [in 'ataḥ'], [nuk] does not come even to [a short a] that was formerly long (bhūtapūrvadīrgha). bhāma — in [the sense of] anger: bābhāmyate. ye vibhāṣā (SK 2319) [— ā optionally replaces the n of jan before y]: jājāyate; jañjanyate.

vārttika
हन्तेर्हिंसायां यङि घ्नीभावो वाच्यः

(vārttika) For han in the sense of injuring, ghnī should be stated [as its substitute] when yaṅ follows.

(वार्तिकम्) ॥ जेघ्नीयते । हिंसायां किम् । जङ्घन्यते ॥

jeghnīyate. Why 'in the sense of injuring'? jaṅghanyate [han in its sense of going].

translator notes
  • commentary BM (absent from our source.json; read in Wikisource uttara-2/0132): 'aṅgasya' is the governing rule (adhikṛta); 'abhyāsasya' is carried forward (anuvartate) from 'atra lopaḥ' [SK 2620], and 'yaṅlukoḥ' from 'guṇo yaṅlukoḥ' [SK 2630]; hence Bhaṭṭoji's gloss. Followed: the bracketed adhikāra and anuvṛttis in sutra_en and k1.
  • commentary BM (Wikisource uttara-2/0132) on 'nukā': objection — in yaṃyamyate the n of nuk is not pada-final and is not followed by a jhal, so how does anusvāra come by 'naś ca' [naś cāpadāntasya jhali, SK 123]? Bhaṭṭoji answers with 'nukā [anusvāro lakṣyate]'. Followed: 'by nuk the anusvāra is meant'.
  • commentary BM (Wikisource uttara-2/0132) on 'taparatvasāmarthyāt': objection — since shortening is prescribed for the abhyāsa, no long vowel is there, so the t in 'ataḥ' is pointless. Reply: the t-marking serves to take only an a that is short by nature, not a short a that is the substitute of a long one. In bābhāmyate the a that arose by the shortening rule with ā as its original (sthānin) does not get nuk. On jan, Bhaṭṭoji recalls the optional ā for n before yaṅ by 'ye vibhāṣā'. Followed: '[a short a]', '[in ataḥ]' and the bracket after ye vibhāṣā.
  • commentary BM (Wikisource uttara-2/0132) on 'ghnībhāvaḥ': even with hnī, the h would become gh by kutva through 'abhyāsāc ca' [SK 2430]; still ghnī is pronounced with gh for lightness of the derivation (prakriyālāghava), as there is no saving of effort otherwise. jeghnīyate means 'kills repeatedly or intensely'. 'yaṅi' here is a seventh case of the domain (viṣayasaptamī), 'when yaṅ is intended', so ghnībhāva comes before doubling although doubling is the later (para) rule. jaṅghanyate means 'goes in a reprehensible way' (garhitaṃ gacchati). Followed: '[han in its sense of going]'; the ordering of ghnī before doubling is not stated by the kaumudī and is not added to the English.
  • commentary TB, printed under sk.2641.t1 (lemma 'ghnībhāva iti', OCR 'grībhāva'): even if hnī were made, jeghnīyate would result by kutva through 'abhyāsāc ca'; ghnī is said for lightness of derivation — 'so the Manoramā'. Others say: just as by the force of pronouncing i in 'serhiḥ' and 'mer niḥ' the u of 'eruḥ' does not come, so with hnī the kutva by 'abhyāsāc ca' would not come; hence ghnī. Some say the ghnī substitute does not block doubling, by the principle of re-application (punaḥprasaṅgavijñāna). In fact doubling is of the part of the yaṅ-final up to the second vowel, while ghnī replaces the base alone; the domains differ, so doubling is unobstructed. TB records a view on the purpose of 'gh' opposed to the one it cites from the Manoramā (which BM shares). Recorded.
  • textual The '(vārtikam)' tag at the head of our k3 is editorial. NSP p. 438 (scan page-0448, read) marks the vārttika 'hanter hiṃsāyāṃ yaṅi ghnībhāvo vācyaḥ' with its vārttika sign before it; the English labels k2 only. Otherwise our kaumudī matches the print (yay̐yamyate, jañjanyate, jeghnīyate; the OCR misreads ghnī as 'mrī').
  • textual This unit has no TB block of its own: TB on its vārttika is printed under sk.2641.t1. There is no TB on the sūtra itself in the print (NSP p. 438). No BM in our source; read in Wikisource.
  • translator yay̐yamyate (print: ya with candrabindu, then yyamyate) and yaṃyamyate are the two outcomes of the anusvāra before y: with optional replacement by the homorganic nasal (by 'vā padāntasya', SK 125, which SK 2636's vārttika makes applicable) and without it. The kaumudī states only the two forms; the brackets name the difference.
  • translator The bracket '[at SK 2636]' is the translator's: Bhaṭṭoji makes the statement there ('nug ity anenānusvāro lakṣyate', under caraphaloś ca). 'ity uktam' could equally point to the Bhāṣya, since BM on SK 2636 (Wikisource uttara-2/0130) attributes the anusvāra-upalakṣaṇa for this very sūtra to the Bhāṣya ('… ity uktaṃ bhāṣye'). The kaumudī does not say which is meant.
SK 2644 · A 7.4.90V
रीगृदुपधस्य च ॥ २६४४ ॥
rīgṛdupadhasya ca
rīg ṛdupadhasya ca — And the augment rīk comes [to the abhyāsa — abhyāsasya carried forward (anuvṛtti) from atra lopo 'bhyāsasya, SK 2620 —] of [a dhātu] whose penultimate is [short] ṛ [when yaṅ or yaṅ-luk follows — yaṅlukoḥ carried forward (anuvṛtti) from guṇo yaṅlukoḥ, SK 2630].

Rule. The augment rīk comes to the abhyāsa of a dhātu whose penultimate (upadhā) is ṛ, when yaṅ follows or there is yaṅluk. Bhaṭṭoji's gloss: "let the augment rīk come to the abhyāsa of a dhātu whose penultimate is ṛ, when yaṅ or yaṅluk follows". One of Bhaṭṭoji's commentators gives the carried words: "of the abhyāsa" from SK 2620, and "when yaṅ or yaṅluk" from SK 2630.

The ṛ must be short. One of Bhaṭṭoji's commentators gives the counter-example cekīrtyate, from kṝt.

Example.

vṛt "to exist":

vṛt + yaṅ → vṛt ya

→ doubling (SK 2395) → vṛt vṛt ya

→ a for the abhyāsa's ṛ, with r after it (SK 2244) → vart vṛt ya

→ first consonant only (SK 2179) → va vṛt ya

→ this rule: rīk at the end of the abhyāsa (SK 36) → va rī vṛt ya

→ k-marked augment, so no lengthening (SK 2632)

→ laṭ → varīvṛtyate.

These steps are as one of Bhaṭṭoji's commentators gives them.

Applies when / blocked by. Bhaṭṭoji adds three forms.

  • narīnṛtyate, from nṛt "to dance". After the r of rīk, the n of the root might become ṇ (SK 235). Bhaṭṭoji: "because it is in the kṣubhnādi list, there is no ṇ" (SK 792).
  • jarīgṛhyate, from grah "to seize". One of Bhaṭṭoji's commentators gives the order: saṃprasāraṇa by SK 2412 first, grah → gṛh; then doubling, a for ṛ (SK 2244), first consonant only, and rīk. The g of the abhyāsa becomes j (SK 2245).
  • calīkḷpyate, from kṛp "to be able". Bhaṭṭoji: "l in both places", by kṛpo ro laḥ (SK 2350). One of his commentators explains the two places. The r of rīk becomes l. The ṛ of the root becomes ḷ.

kṛp ya → doubling → a + r for the abhyāsa's ṛ (SK 2244) → first consonant only → ka kṛp ya

→ c for k (SK 2245) → ca kṛp ya

→ rīk → ca rī kṛp ya

→ l for the r of rīk, ḷ for the root's ṛ (SK 2350) → ca lī kḷp ya

→ laṭ → calīkḷpyate.

Vārttika. rīg ṛtvata iti vaktavyam: "it should be stated: rīk [comes to the abhyāsa of a dhātu] containing ṛ (ṛtvat)". One of Bhaṭṭoji's commentators gives the reason. In vraśc, after saṃprasāraṇa, the ṛ is not the penultimate: ś is. So the sūtra's "whose penultimate is ṛ" would fail. The vārttika replaces it with "containing ṛ". He adds that ṛtvat is a bha, not a pada, so its t takes no jaś.

Bhaṭṭoji's forms:

  • varīvṛścyate, from vraśc "to cut": saṃprasāraṇa (SK 2412) → vṛśc; doubling, a for ṛ, first consonant only, rīk.
  • parīpṛcchyate, from pracch "to ask": saṃprasāraṇa (SK 2412) → pṛcch; then the same steps.

Watch out. rīk goes on the abhyāsa, which by then has a, not ṛ: va- + rī → varī-. The ṛ stays in the second copy: varīvṛtyate.

Siddhānta-kaumudī
ऋदुपधस्य धातोरभ्यासस्य रीगागमः स्याद्यङ्यङ्लुकोः । वरीवृत्यते । क्षुभ्नादित्वान्न णः । नरीनृत्यते । जरीगृह्यते । उभयत्र लत्वम् । चलीकॢप्यते ।

Let the augment rīk come to the abhyāsa [— carried forward (anuvṛtti) from SK 2620] of a dhātu whose penultimate is ṛ, when yaṅ or yaṅ-luk follows [— carried forward (anuvṛtti) from SK 2630]: varīvṛtyate [vṛt]. Because [nṛt] is in the kṣubhnādi list [kṣubhnādiṣu ca, SK 792], ṇ does not [replace the n]: narīnṛtyate. jarīgṛhyate [grah]. In both places l is substituted (latva) [by kṛpo ro laḥ, SK 2350 — for the r of rīk in the abhyāsa, and in the ṛ of the root]: calīkḷpyate [kṛp].

vārttika
रीगृत्वत इति वक्तव्यम्

(vārttika) It should be stated: rīk [comes to the abhyāsa of a dhātu] containing ṛ (ṛtvataḥ).

(वार्तिकम्) ॥ वरीवृश्च्यते । परीपृच्छ्यते ॥

varīvṛścyate [vraśc]. parīpṛcchyate [pracch].

Tattvabodhinī
NSP p. 448
—रीगृदुपधस्य च । ऋदिति किम् । चेकीर्त्यते । णिजभावपक्षे एकाच्त्वाद्यङ्
translator notes
  • commentary BM (absent from our source.json; read in Wikisource uttara-2/0132–0133): 'abhyāsasya' because it is carried forward (anuvṛtti) from 'atra lopaḥ' [SK 2620]; 'yaṅlukoḥ' because it is carried forward from 'guṇo yaṅlukoḥ' [SK 2630]. Followed: both anuvṛtti brackets.
  • commentary BM (Wikisource uttara-2/0133): varīvṛtyate — from 'vṛtu vartane' with yaṅ: doubling, a for ṛ (urat), deletion of the non-initial consonants (halādiḥ śeṣaḥ), rīk to the abhyāsa. On narīnṛtyate, anticipating ṇ for n, Bhaṭṭoji says 'kṣubhnāditvān na ṇaḥ'. jarīgṛhyate — from 'graha upādāne' with yaṅ: saṃprasāraṇa by 'grahijyā…' [SK 2412], doubling, urat, halādiḥ śeṣaḥ, rīk. Followed: the bracketed roots; the steps are BM's and are not added to the kaumudī.
  • commentary BM (Wikisource uttara-2/0133) on 'ubhayatra': kṛp with yaṅ — doubling, urat, r-insertion (raparatva), halādiḥ śeṣaḥ, rīk; in calīkḷpyate the r in the abhyāsa becomes l by 'kṛpo ro laḥ', and in the second part the r that is a component of ṛ is replaced by what resembles l (ḷ). Followed: the bracket naming the two places and the rule.
  • commentary BM (Wikisource uttara-2/0133): objection — in varīvṛścyate from vraśc, how does the abhyāsa get rīk, since the root has ś as penultimate and so is not ṛ-penultimate? Hence the vārttika: in 'rīg ṛdupadhasya', 'ṛdupadhasya' is to be removed and 'ṛtvataḥ' stated; 'ṛtvat' = 'that which has ṛ'; by 'tasau matvarthe' it is bha, not pada, so there is no jaś-substitution. parīpṛcchyate: pracch with yaṅ, saṃprasāraṇa ṛ for r by 'grahijyā…', pūrvarūpa, doubling, raparatva, halādiḥ śeṣaḥ, rīk; varīvṛścyate likewise from vraśc. Followed: 'containing ṛ (ṛtvataḥ)' and the bracketed roots.
  • commentary TB (sk.2644.t1; OCR 'kad iti', scan 'ṛd iti'): why short ṛ (ṛt)? cekīrtyate — [from kṝt], in the option without ṇic it is single-vowelled, so yaṅ comes. Followed: '[short] ṛ' in sutra_en.
  • textual The '(vārtikam)' tag at the head of our k3 is editorial. NSP p. 438 (scan page-0448, read) marks the vārttika 'rīg ṛtvata iti vaktavyam' with its vārttika sign before it; the English labels k2 only. Ours matches the print, including 'calīkḷpyate' (print ligature klṛ) and the order varīvṛścyate, parīpṛcchyate (Wikisource reverses the order and reads 'calīklṛpyate').
SK 2645 · A 6.1.19V
स्वपिस्यमिव्येञां यङि ॥ २६४५ ॥
svapisyamivyeñāṃ yaṅi
svapisyamivyeñāṃ yaṅi — [saṃprasāraṇa — carried forward (anuvṛtti) from ṣyaṅaḥ saṃprasāraṇam…, SK 1003 —] takes place in svap, syam and vyeñ when yaṅ follows.

Rule. Saṃprasāraṇa takes place in svap "to sleep", syam and vyeñ, when yaṅ follows. Bhaṭṭoji's gloss: "let saṃprasāraṇa take place when yaṅ follows". The word saṃprasāraṇa is carried forward from SK 1003, one of Bhaṭṭoji's commentators says.

Example.

svap + yaṅ → svap ya

→ this rule: saṃprasāraṇa, v → u → suap ya

→ pūrvarūpa (SK 330): u + a → u → sup ya

→ doubling (SK 2395) → sup sup ya

→ first consonant only (SK 2179) → su sup ya

→ guṇa of the abhyāsa (SK 2630) → so sup ya

→ ṣ for the s of the second part (SK 212) → so ṣup ya

→ laṭ → soṣupyate.

These steps are as one of Bhaṭṭoji's commentators gives them.

syam + yaṅ → saṃprasāraṇa, y → i, and pūrvarūpa → sim ya

→ doubling, first consonant only → si sim ya

→ guṇa → se sim ya

→ no ṣ, since syam is not taught with ṣ (it is not ṣopadeśa) → sesimyate.

Bhaṭṭoji's commentators both give this reason.

vyeñ + yaṅ → saṃprasāraṇa, y → i, and pūrvarūpa → vi ya

→ the i is lengthened → vī ya

→ doubling, first consonant only, shortening → vi vī ya

→ guṇa → ve vī ya

→ laṭ → vevīyate.

Bhaṭṭoji's commentators say the lengthening is by halaḥ (SK 2559) or by SK 2298.

Applies when / blocked by. Why "when yaṅ follows"? One of Bhaṭṭoji's commentators gives svapnak, where svap keeps its v.

Watch out. svap and syam both begin with s, but only svap is taught with ṣ. So only soṣupyate has ṣ. Compare sosūcyate (SK 2630).

Siddhānta-kaumudī
संप्रसारणं स्याद्यङि । सोषुप्यते । सेसिम्यते । वेवीयते ॥

Let saṃprasāraṇa [— carried forward (anuvṛtti) from SK 1003] take place [in these] when yaṅ follows: soṣupyate [svap]. sesimyate [syam]. vevīyate [vyeñ].

Tattvabodhinī
NSP p. 448–449
—स्वपिस्यमि—। यङि किम् । स्वप्नक् ॥—सेसिम्यते इति । स्यमु शब्दे दन्त्याजन्तसादित्वाभावेनाषोपदेशत्वान्न षः ॥—वेवीयते इति । हलः इति ‘अकृत्सार्व—’ इति वा दीर्घः ॥
translator notes
  • commentary BM (absent from our source.json; read in Wikisource uttara-2/0133): 'saṃprasāraṇaṃ syāt' supplies the remainder (śeṣapūraṇa), because it is carried forward (anuvṛtti) from 'ṣyaṅaḥ saṃprasāraṇam' [SK 1003]. Followed: the anuvṛtti brackets.
  • commentary BM (Wikisource uttara-2/0133): soṣupyate — svap with yaṅ, u for v by saṃprasāraṇa, pūrvarūpa, doubling, guṇa of the abhyāsa; ṣ for s in the second part. sesimyate — syam with yaṅ, saṃprasāraṇa of y, pūrvarūpa, doubling and so on; no ṣ, since it is not ṣopadeśa. vevīyate — vyeñ with yaṅ, saṃprasāraṇa of y, pūrvarūpa, doubling and so on; the lengthening is by 'halaḥ' [SK 2559] or by 'akṛtsārvadhātukayoḥ' [SK 2298]. Recorded; the kaumudī gives the forms only and the English adds only the roots.
  • commentary TB (sk.2645.t1, read from the scan, NSP pp. 438–439): why 'when yaṅ follows'? svapnak. sesimyate — 'syamu śabde': it does not begin with s followed by a dental or a vowel, so it is not ṣopadeśa, and there is no ṣ. vevīyate — lengthening by 'halaḥ' or by 'akṛtsārva…'. Agrees with BM. Recorded.
  • textual Our kaumudī matches NSP p. 438 (scan page-0448, read). Our TB block sk.2645.t1 is garbled: it repeats 'svapisyami— yaṅi kim | svapnak ||—sesi-' and loses the lemma 'vevīyate iti | halaḥ iti'. The print (scan page-0448 foot, read; partly under the library watermark) has '…syamu śabde dantyājantasāditvābhāvenāṣopadeśatvān na ṣaḥ ||—vevīyate iti | halaḥ iti 'akṛtsārva—' iti vā dīrghaḥ ||'. Commentary layer only; source.json not edited.
SK 2646 · A 6.1.20V
न वशः ॥ २६४६ ॥
na vaśaḥ
na vaśaḥ — [saṃprasāraṇa] does not take place in vaś [when yaṅ follows].

Rule. Saṃprasāraṇa does not take place in vaś, when yaṅ follows. Bhaṭṭoji gives only the form, vāvaśyate.

One of Bhaṭṭoji's commentators supplies the context. With yaṅ after vaś, saṃprasāraṇa would come by SK 2412, since yaṅ is marked with ṅ. Bhaṭṭoji recalls its prohibition with this sūtra. "Saṃprasāraṇa" and "when yaṅ follows" also run on from SK 2645.

Example.

vaś + yaṅ → vaś ya

→ SK 2412 would give saṃprasāraṇa of the v; this rule forbids it

→ doubling (SK 2395) → vaś vaś ya

→ first consonant only (SK 2179) → va vaś ya

→ lengthening (SK 2632) → vā vaś ya

→ laṭ → vāvaśyate.

Applies when / blocked by. vaś is called a Vedic root only "for the most part" (prāyika), as one commentator notes.

Watch out. This is a prohibition, not a new operation. Without it, vaś would follow the grah group of SK 2412.

Siddhānta-kaumudī
वावश्यते ॥

vāvaśyate [vaś: no saṃprasāraṇa].

translator notes
  • commentary BM (absent from our source.json; read in Wikisource uttara-2/0133): with yaṅ after the root vaś, in vāvaśyate, saṃprasāraṇa would obtain by 'grahijyā…' [SK 2412]; Bhaṭṭoji recalls its prohibition with 'na vaśaḥ'. That vaś is Vedic holds only for the most part (prāyika), as was said earlier. Followed: '[saṃprasāraṇa]' in sutra_en and k1, and the yaṅ context '[when yaṅ follows]' in sutra_en, which BM's lead-in ('atha vaśadhātor yaṅi vāvaśyate ity atra… tanniṣedhaṃ smārayati') sets explicitly.
  • textual Our kaumudī matches NSP p. 438 (scan page-0448, read). There is no TB on this sūtra: in the print the TB goes from SK 2645 ('…iti vā dīrghaḥ', top of p. 439) straight to SK 2647 ('cāyaḥ kī'). No BM in our source; read in Wikisource.
  • translator The kaumudī gives no gloss, only the example. Both brackets in sutra_en ('[saṃprasāraṇa]' and '[when yaṅ follows]') are supplied from BM's lead-in to this sūtra (see the BM note); the yaṅ context also continues from SK 2645 ('yaṅi').
SK 2647 · A 6.1.21V
चायः की ॥ २६४७ ॥
cāyaḥ kī
cāyaḥ kī — kī is substituted for cāy [when yaṅ follows — yaṅi carried forward (anuvṛtti) from svapisyamivyeñāṃ yaṅi, SK 2645].

Rule. kī replaces cāy when yaṅ follows. Bhaṭṭoji gives only the form, cekīyate. "When yaṅ follows" is carried from SK 2645, one of his commentators says. The root is cāyṛ "to honour, to perceive".

Example.

cāy + yaṅ → this rule: kī replaces the whole base → kī ya

→ doubling (SK 2395) → kī kī ya

→ shortening (SK 2180) → ki kī ya

→ c for k (SK 2245) → ci kī ya

→ guṇa of the abhyāsa (SK 2630) → ce kī ya

→ laṭ → cekīyate.

Applies when / blocked by. Why a long ī? One of Bhaṭṭoji's commentators says it is for the intensive with yaṅluk, where kī also comes: cekītaḥ.

Watch out. kī has more than one sound, so it replaces the whole of cāy (SK 45), not just its last sound.

Siddhānta-kaumudī
चेकीयते ॥

[kī replaces cāy when yaṅ follows — carried forward (anuvṛtti) from SK 2645:] cekīyate.

Tattvabodhinī
NSP p. 449
—चायः की । यङि परे चायः की स्यात् । दीर्घोच्चारणं यङ्लुगर्थम् । अत एव ‘न लुमता—’ इति निषेधोऽपि न प्रवर्तत इति यङ्लुक्यपि चायः की स्यादेव । चेकीतः
translator notes
  • commentary BM (absent from our source.json; read in Wikisource uttara-2/0133–0134): 'yaṅi' is the remainder (śeṣa), because it is carried forward (anuvṛtti) from 'svapisyamivyeñāṃ yaṅi' [SK 2645]. cekīyate: from 'cāyṛ pūjāniśāmanayoḥ' with yaṅ, kī replaces the base, then doubling and so on. Followed: the anuvṛtti brackets.
  • commentary TB (sk.2647.t1): when yaṅ follows, kī replaces cāy. The long ī is pronounced for the sake of yaṅ-luk; for that very reason the prohibition 'na lumatā…' does not operate either, and kī replaces cāy in yaṅ-luk too: cekītaḥ. Followed: 'when yaṅ follows'; the yaṅ-luk extension is TB's and is not added to the English.
  • textual Our kaumudī matches NSP p. 439 (scan page-0449, read).
SK 2648 · A 7.4.31V
ई घ्राध्मोः ॥ २६४८ ॥
ī ghrādhmoḥ
ī ghrādhmoḥ — ī is substituted for [the final of] ghrā and dhmā [when yaṅ follows — yaṅi carried forward (anuvṛtti) from yaṅi ca, SK 2633].

Rule. ī replaces the final of ghrā "to smell" and dhmā "to blow", when yaṅ follows. Bhaṭṭoji gives only the forms. One of his commentators says yaṅi "when yaṅ follows" is carried from SK 2633. The ī replaces the final ā, by SK 42.

Example.

ghrā + yaṅ → ghrā ya

→ this rule: ī for the final ā → ghrī ya

→ doubling (SK 2395) → ghrī ghrī ya

→ first consonant only, the vowel shortened (SK 2179, 2180) → ghi ghrī ya

→ gh becomes a ca-series sound, jh (SK 2245), and then j (SK 2182) → ji ghrī ya

→ guṇa of the abhyāsa (SK 2630) → je ghrī ya

→ laṭ → jeghrīyate.

Likewise from dhmā: dhmī ya → di dhmī ya → de dhmī ya → dedhmīyate.

Applies when / blocked by. Why a long ī? A short i would be lengthened anyway before y (SK 2298). Bhaṭṭoji's commentators give different purposes for the long vowel. One points to asya cvau (SK 2118), which carries the ī from here. The other points first to the form with yaṅluk, jeghrītaḥ. Scholars discuss this.

Watch out. The ī comes before doubling, so both copies have it: ghrī, then ghi after shortening.

Siddhānta-kaumudī
जेघ्रीयते । देध्मीयते ॥

[ī replaces the final of ghrā and dhmā when yaṅ follows — carried forward (anuvṛtti) from SK 2633:] jeghrīyate. dedhmīyate.

Tattvabodhinī
NSP p. 449
—ई घ्राध्मोः । दीर्घोच्चारणं प्राग्वत् । जेघ्रीतः । ‘अस्य च्चौ’ इत्याद्यर्थं दीर्घग्रहणमिति तु सुवचम्
translator notes
  • commentary BM (absent from our source.json; read in Wikisource uttara-2/0134): 'ghrādhmoḥ' is the sixth-case dual of the dvandva of ghrā and dhmā; 'yaṅi' is the remainder (śeṣa), carried forward (anuvṛtti) from 'yaṅi ca' [SK 2633]. jeghrīyate, dedhmīyate: after ī is made, doubling and so on. Even if short i were prescribed, the long would come by 'akṛtsārvadhātukayoḥ' [SK 2298]; the long is pronounced for the sake of the following 'asya cvau' [SK 2118]. Followed: the anuvṛtti brackets.
  • commentary TB (sk.2648.t1): the long is pronounced as before [under SK 2647, i.e. for yaṅ-luk]: jeghrītaḥ. It may equally well be said that the long is taken for the sake of 'asya cvau' and the like. Recorded; the English is neutral.
  • textual Our kaumudī matches NSP p. 439 (scan page-0449, read). Wikisource's kaumudī reads 'jeghnīyate' (evidently a slip; its BM lemma has 'jeghrīyate').
  • translator '[the final of]' is the translator's: ī replaces the last sound of ghrā and dhmā (by alo 'ntyasya), which neither the kaumudī nor the commentators state here.
  • translator The commentators differ on the purpose of the long ī: TB's first reason is yaṅ-luk (jeghrītaḥ), with 'asya cvau' as an alternative; BM gives 'asya cvau' only. The English states neither purpose (§16).
SK 2649 · A 7.4.22V
अयङ् यि क्ङिति ॥ २६४९ ॥
ayaṅ yi kṅiti
ayaṅ yi kṅiti — ay (ayaṅ) is substituted for [the final of] śīṅ [— śīṅaḥ carried forward (anuvṛtti) from śīṅaḥ sārvadhātuke guṇaḥ, SK 2441 —] when [an affix] beginning with y and marked with k or ṅ follows.

Rule. ay replaces the final of śī (śīṅ "to lie down") before an affix that begins with y and is marked with k or ṅ. Bhaṭṭoji's gloss: "let the substitute ayaṅ replace śīṅ when [an affix] beginning with y and marked with k or ṅ follows".

One of Bhaṭṭoji's commentators gives the details. "Of śīṅ" is carried from SK 2441. In ayaṅ, the ṅ is a marker. Being ṅ-marked, the substitute replaces only the final sound (SK 43). Yaṅ begins with y and is marked with ṅ, so it qualifies.

Example.

śī + yaṅ → śī ya

→ this rule: ay for ī → śay ya

→ doubling (SK 2395) → śay śay ya

→ first consonant only (SK 2179) → śa śay ya

→ lengthening of the abhyāsa (SK 2632) → śā śay ya

→ laṭ → śāśayyate.

Why does ay come before doubling? One of Bhaṭṭoji's commentators answers: it is the later rule (para) and the inner one (antaraṅga).

Applies when / blocked by. Bhaṭṭoji ends the section with ḍhauk "to approach" and trauk:

ḍhauk + yaṅ → ḍhauk ya

→ doubling (SK 2395) → ḍhauk ḍhauk ya

→ first consonant only (SK 2179) → ḍhau ḍhauk ya

→ the abhyāsa is shortened (SK 2180): the short of au is u → ḍhu ḍhauk ya

→ ḍ for ḍh (SK 2182) → ḍu ḍhauk ya

→ "then guṇa", by SK 2630 → ḍo ḍhauk ya

→ laṭ → ḍoḍhaukyate.

Likewise totraukyate.

Here both the shortening and the guṇa reach the abhyāsa. One of Bhaṭṭoji's commentators ties this to SK 2632, as you saw there.

The printed text closes here: iti yaṅantaprakriyā, "here ends the working of the yaṅ-final". The next chapter takes up the intensive with luk of yaṅ.

Watch out. ḍoḍhaukyate has o in the abhyāsa, not au. The abhyāsa's au is first shortened to u, and only then takes guṇa.

Siddhānta-kaumudī
शीङोऽयङादेशः स्याद्यादौ क्ङिति परे । शाशय्यते । अभ्यासस्य ह्रस्वः । ततो गुणः । डोढौक्यते । तोत्रौक्यते ॥

Let ayaṅ be substituted for śīṅ [— carried forward (anuvṛtti) from SK 2441] when [an affix] beginning with y and marked with k or ṅ follows: śāśayyate. [In ḍhauk,] the abhyāsa is shortened, then guṇa [by guṇo yaṅlukoḥ, SK 2630]: ḍoḍhaukyate. totraukyate [trauk]. [Here ends the yaṅanta-prakriyā (the section on the yaṅ-final).]

Tattvabodhinī
alignment inferredNSP p. 449
—शाशय्यते इति । परत्वादन्तरङ्गत्वाच्चायङादेशे कृते द्वित्वम् ॥ इति यङन्तप्रक्रिया ॥
translator notes
  • commentary BM (absent from our source.json; read in Wikisource uttara-2/0134): 'yi' is a seventh-case word; 'kṅiti' is analysed as 'that of which k and ṅ are the it'. 'śīṅaḥ' because it is carried forward (anuvṛtti) from 'śīṅaḥ sārvadhātuke guṇaḥ' [SK 2441]. In ayaṅ, ṅ is it and the a after y is for pronunciation; being ṅ-marked, it replaces the final. Followed: the anuvṛtti brackets, 'ay (ayaṅ)' and '[the final of]'.
  • commentary BM (Wikisource uttara-2/0134): śāśayyate — with śī ya standing, ay replaces ī, then doubling and lengthening of the abhyāsa. For ḍhauk with yaṅ Bhaṭṭoji states the particular: after doubling, the abhyāsa is shortened to u, and that takes guṇa by 'guṇo yaṅlukoḥ'. totraukyate is the form of trauk. Followed: '[In ḍhauk,]', '[by guṇo yaṅlukoḥ, SK 2630]' and '[trauk]'.
  • commentary TB (sk.2649.t1) on 'śāśayyate': ayaṅ is made first, because it is later (para) and antaraṅga, and then doubling. Recorded.
  • textual Our source lacks the colophon. The print ends the unit '…totraukyate || iti yaṅantaprakriyā ||' (NSP p. 439, scan page-0449, read) and then gives the heading 'tiṅante yaṅlugantaprakriyā 54'. The English supplies the colophon in brackets. TB closes with the same 'iti yaṅantaprakriyā' (inside sk.2649.t1; the TB on SK 2650, 'yaṅo 'ci ca', that followed it was re-filed to sk.2650.t1 on 2026-10-05); Wikisource has 'iti tiṅantayaṅprakaraṇam'. Mūla omission, for Ajai. Otherwise our kaumudī matches the print.
← all prakaraṇasoutline →