तिङन्तनामधातुप्रकरणम्

Covers SK 2657–2677 (21 sūtras): all present. Every unit has its own kaumudī. Two units are long: 2665 (kyaṅ and the ācāra-kvip vārttikas) and 2677 (ṇic for muṇḍa… and the whole iṣṭhavat section). Each is outlined under its own number and then continued in the next group (2665 in C → D, 2677 in G → H). The spine is mainly an Aṣṭādhyāyī run, 3.1.8–3.1.21: 2657, 2663–2665, 2667–2668 and 2670–2677. The SK inserts the stem rules that kyac and kyaṅ need at the points where they are first required:

• 7.4.33 (2658), 1.4.15 (2659), 6.4.50 (2660), 7.4.34 (2661) and 7.1.51 (2662) come after 3.1.8;
• 6.4.15 (2666) comes after 3.1.11;
• 1.3.90 (2669) comes after 3.1.13.

Every AS-ref below is checked against references/ashtadhyayi/sutraani-data.json. All 21 match our source.json.

• Anuvṛtti from later in the SK. āj jaser asuk (7.1.50, SK 3572) gives asuk at 2662 (BM, TB). vṛttisargatāyaneṣu kramaḥ (1.3.38, SK 2711) is cited under 2670 for the ātmanepada of kramate.
• Anuvṛtti from p55. dhātoḥ karmaṇaḥ samānakartṛkād icchāyāṃ vā (3.1.7, SK 2608) gives karmaṇaḥ, icchāyām and vā at 2657, karmaṇaḥ and vā at 2664, and vā at 2665.
• Anuvṛtti from p53. The gaṇasūtras under ī ca gaṇaḥ (SK 2573) supply the ground for 2677:
• prātipadikād dhātvarthe bahulam iṣṭhavac ca;
• tat karoti tad ācaṣṭe.

tiṅantanāmadhātuprakaraṇam — outline (p58, SK 2657–2677, all 21 sūtras)

p57 (tiṅantayaṅlukprakaraṇam) closed at 2656 with na dhātulopa ārdhadhātuke and the colophon "iti yaṅluganta-prakriyā", which is supplied in brackets.

• Where the prakaraṇa begins. The print heads it "tiṅante nāmadhātuprakriyā 55" (NSP p. 445, scan page-0455). WS BM opens "atha nāmadhātuprakriyā nirūpyante". The name tiṅantanāmadhātuprakaraṇam and the number 58 are ours. sk2657 opens with supa ātmanaḥ kyac (3.1.8).
• Where it ends. At 2677 the list of iṣṭhavat substitutes closes with vṛndārakaṃ vṛndayati. The print closes "iti nāmadhātuprakriyā" (NSP p. 453, scan page-0463). Ours lacks it, so the English supplies it in brackets. p59 (tiṅantakaṇḍvādiprakaraṇam) opens with kaṇḍvādibhyo yak (3.1.27, SK 2678).
• Contents in outline.
• 2657–2660: kyac in the sense of wishing for oneself, and its stem.
• ī for a: putrīyati.
• Pada-status only for n-final stems: rājīyati, but vācyati and samidhyati.
• Optional kya-deletion after a consonant: samidhitā / samidhyitā.
• Not from m-final bases or indeclinables.
• 2661–2663: other forms in the wishing sense. The nipātanas aśanāyati, udanyati and dhanāyati; asuk (aśvasyati vaḍavā, kṣīrasyati); kāmyac (putrakāmyati).
• 2664–2665: behaving like. kyac after an object as standard (putrīyati chātram) and kyaṅ after an agent as standard (kṛṣṇāyate). s-deletion: ojāyate, yaśāyate / yaśasyate. puṃvat: kumārāyate, haritāyate.
• 2665 (vārttikas), 2666: behaving like, with kvip. avagalbhate; then kṛṣṇati, ati, mālāti, kavayati and so on. The nasal-final stems lengthen before kvi: idāmati, rājānati.
• 2667–2669: becoming. kyaṅ for bhṛśa etc. (bhṛśāyate, sumanāyate), with the saṃgrāma jñāpaka for separating an upasarga-like first member. kyaṣ for lohita and ḍāc-finals, optionally parasmaipada: lohitāyati / lohitāyate, paṭapaṭāyati.
• 2670–2674: kyaṅ in particular senses with an object. kaṣṭāyate, romanthāyate, tapasyati, bāṣpāyate, śabdāyate, sudināyate; sukha etc. in vedanā.
• 2675–2677: other affixes in the karaṇa sense. kyac for namas, varivas and citra (namasyati, citrīyate); ṇiṅ for puccha, bhāṇḍa and cīvara (utpucchayate); ṇic for muṇḍa… (muṇḍayati, satyāpayati).
• 2677 (continued): ṇic after any prātipadika, treated as before iṣṭhan. prathayati, svāpayati, tvādayati, śāvayati, udīcayati, tirāyayati, srajayati, sthavayati, jyāpayati and so on.

Threads that run through the prakaraṇa:

• The affixes that make a nominal root. Each forms a dhātu by sanādyantā dhātavaḥ (3.1.32, SK 2304), and the sup inside is deleted by luk, "being part of a dhātu" (2657).
• kyac (2657, 2664, 2675): parasmaipada; ī for a by kyaci ca (2658).
• kāmyac (2663).
• kyaṅ (2665, 2667, 2670–2674): ātmanepada because it is ṅ-it; lengthening by akṛtsārvadhātukayor dīrghaḥ (7.4.25, SK 2298).
• kyaṣ (2668): parasmaipada optional by 2669.
• kvip (the vārttikas under 2665).
• ṇiṅ (2676): ātmanepada always (TB).
• ṇic (2677, and the p53 gaṇasūtra).
• Pada-status of the stem. naḥ kye (2659) restricts it, before kyac and kyaṅ, to n-final stems. So:
• no lopaḥ śākalyasya in gavyati (2658);
• no kutva in vācyati and no jaśtva in samidhyati (2659);
• no ru in vidvasyate (TB, 2665) or tapasyati (BM, 2671).

With kvip, prātipadika is named, so supaḥ is not connected and no pada operation applies (2665 k7): rājānati keeps its n (BM), tatati, tvacati (TB).

• The anuvṛtti chain (as the commentators state it):
• karmaṇaḥ / icchāyām / vā: 2608 → 2657 (TB; WS BM);
• supaḥ and kyac: 2657 → 2664;
• upamānād ācāre: 2664 → 2665;
• kyaṅ: 2665 → 2667 and 2670. It is svarita; kyaṣ is not (TB, BM).
• bhuvi: 2667 → 2668; acveḥ → 2668, for lohita only (kaumudī 2669);
• parasmaipada: śeṣāt kartari parasmaipadam (1.3.78, SK 2159) → 2669;
• karmaṇaḥ: 2671 → 2672, 2673, 2674;
• karaṇe: 2673 → 2675 (kaumudī), 2676 (TB), 2677 (BM).
• Treatment as masculine.
• Before kyaṅ by kyaṅmāninoś ca (6.3.36, SK 837), barred by na kopadhāyāḥ (6.3.37, SK 838): see 2665.
• Before ṇic through iṣṭhavat: etayati, dāradayati, śrayayati (2677).
• Separating a first member that looks like an upasarga. The saṃgrāma jñāpaka (2667) separates such a member when a dhātu-making affix is due. So aṭ / āṭ comes after it, and prakṛtyaikāc can apply to what remains. Examples:
• svamanāyata, udamanāyata, avāgalbhata (2667);
• samababhāṇḍata (2676, BM);
• praśasyayati, udaicicat, samyacicat, svāśaśvat (2677).

It is limited to cases where the upasarga's whole form is heard (oḍhāyitvā, 2667). It does not reach tiras or saha, which are not upasarga-like (TB, BM on 2677).

• "No vowel-substitute before ṇi while doubling is due." This is the oḥ puyaṇjy apare jñāpaka (7.4.80, SK 2577), traced in p54. Here it keeps the a of the abhyāsa in aujaḍhat (BM, TB), makes svar double in asasvat (BM, TB), and makes śrī double before vṛddhi in aśiśrayat (BM).
• aglopin stems in the caṅ-aorist. When the deletion removes an ak, there is no sanvadbhāva, lengthening or shortening:
• ajahalat and acakalat (2677 k9);
• apaprathat against apiprathat (the order of ṭi-lopa and vṛddhi);
• atitirāyat.
• āpuk. satya (from the satyāpa… list, SK 2563) and, by vārttika, artha and veda: satyāpayati, arthāpayati, vedāpayati (2677).
• Where the commentary lies. Only the kaumudī blocks are translated.
• TB printed under a neighbouring unit:
• 2657 at the tail of sk2656.t1, after p57's colophon;
• 2658 inside sk2659.t1, and 2660 at its tail;
• 2664 at the tail of sk2663.t1;
• sk2666.t1 carries TB on 2666, then all of TB on 2667 (through the avadhīr discussion), then TB on 2669.
• There is no TB on 2668 or 2672 (print).
• sk2677.t1 runs on past "iti nāmadhātuprakriyā" into TB on SK 2678 (p59's matter).
• BM in ours is garbled or short against Wikisource uttara-2/0154–0177:
• 2657: the anuvṛtti sentence is missing;
• 2658: "ātmanaḥ kim | rājñaḥ putram…" is lost;
• 2659: the sannipāta objection;
• 2662: "aāa" for aśva;
• 2663: anuvapadam;
• 2665: it breaks off at "sapatnīyate iti";
• 2666: the pathīnati argument (WS 0164);
• 2669: the paṭapaṭā join;
• 2675: the vārttika list is lost;
• 2677: two passages are duplicated.
• WS lacks a page between 0165 and 0166: the end of 2667, all of 2668, and 2669 up to the list.
• The kaumudī lists forms bare. Derivation stages and lakāra labels are the translator's, following BM and TB as recorded in the unit notes.

The p55–58 audit (.work/sk-run/p55-58-audit/sk2648-2657.md, sk2658-2667.md, sk2668-2677.md) found one MUST-FIX. A fixer is applying it and the MINOR items in parallel. This outline already follows the audit.

• 2665 (MUST-FIX): in hariṇīvācarati → haritāyate, hariṇī is the feminine of the colour-word harita ("tawny/green"), not "doe". Puṃvat removes the n of varṇād anudāttāt topadhāt to naḥ (4.1.39, SK 496) and the ṅīṣ. hariṇī "doe" (from hariṇa) would give hariṇāyate (BM).
• 2665: "sa ca vyavasthitaḥ": what is fixed in its distribution is the s-deletion, not "the option" (BM).
• 2665: sapatnīva is one analytic phrase for three different words (TB). Only the third, sapatnīyate, is "co-wife".
• 2665: "tadvati" is the kaumudī's own gloss of ojas within the vṛtti. It is not an anuvṛtti.
• 2662: BM reads sarvaprātipadikānām… as also giving kyac itself to every prātipadika in the lālasā sense, and suk or asuk before it.
• 2666: BM's case for pathenati (the inhan restriction, and the vṛtraghnī argument) is read in WS 0164. Our BM is lacunose.
• 2667: TB's Bhāṣya counter-example on usy apadāntāt is kosrā (print), not "kolā".
• 2671: tapas takes kyaṅ from the sūtra, and the vārttika adds parasmaipada in place of the ṅ-ātmanepada (BM).
• 2673: that the vārttika supplies kyaṅ in the karoti sense is BM's reading (śeṣa). The kaumudī does not say it.
• 2677: payo vratayati is "he eats milk" (BM: bhuṅkte, TB: aśnāti). There is no "alone, as a vow".
• 2677: the commentators are kept apart:
• trapayati is BM's form, and trāpayati is TB's and the print's;
• śrāyi is BM's stem, and śrayayati is TB's lemma and the print's;
• both give the same account of abubhravat and svāśaśvat.

Print readings (NSP, scan-confirmed) that the outline follows:

• 2661: gardheṣu (orthographic).
• 2662: madhvasyati.
• 2664: prāsāde.
• 2665: paṭvīmṛdvyāv, avagalbhate (×2), anunāsikatvam, anubandhāsañjana-, prātipadika-, dhātor, and the sentence "mādhavas tu nāmadhātor api vṛddhim icchati | akavāyīt", which is missing in ours.
• 2666: svāṃcakāra.
• 2667: "evaṃ ca"; the ❀ on usyomāṅkṣv āṭaḥ pratiṣedhaḥ.
• 2671: vartanāyāṃ; ity artha iti.
• 2677:
• kṛñarthe; payasayati; tṛpra → trāpayati;
• the restorations: [amimradat], [parivraḍhayati; paryavavraḍhat], [aujiḍhat … hti], [samyacicat], [sthiraṃ sthāpayati];
• the colophon.
• Against the print, with WS and BM: payasvinīm (print payasvinīt), and samababhāṇḍata (2676; print b/v unclear).

Still open for Ajai (mūla):

• 2669: "karuṇā nityacarmaṇī". The print's letter is ध or घ, not च. The English keeps carman.
• 2659: ours and the print order "gārgīyati | vācyati | … samidhyati". WS and BM's lemma give vātsīyati. The English keeps the transmitted order.
• 2677:
• k8 (halikalyor adantatvaṃ ca nipātyate) and k12 (avyayānāṃ bhamātre ṭilopaḥ) are filed as vārttikas, but the print sets them in plain type.
• adadadrāyat and āmumuāyayat are as transmitted (WS and BM differ).
• trāpayati is as transmitted (BM: trapayati).
• 2667: "ṣāṣṭe sthitam" is ours and the print's; ṣāṣṭhe is meant.
• 2674: the print gives no form after sukhaṃ vedayate. WS adds sukhāyate, and the outline does not supply it.
• Every mūla correction listed under Print readings stands in ours unedited (source.json has not been changed). The '(वार्तिकम्)' tags at the head of the prose blocks are editorial.

Ordering: strict SK order. The two long units are split across groups: 2665 is in C, and its kvip vārttikas are in D; 2677 is in G, and its iṣṭhavat section (k11–k13) is in H.

A. kyac in the sense of wishing for oneself, and the stem before it — running example: putram ātmana icchati: putra-am + kyac → (sup luk) putra-ya → (ī for a, 2658) putrīya → putrīyati; rājānam icchati: rājan-ya → (pada, n deleted) rāja-ya → rājīyati

• 2657 supa ātmanaḥ kyac 3.1.8 — kyac comes optionally, in the sense of wishing, after a sup-final word that is the object of the wishing and is wished for the wisher himself. karmaṇaḥ, icchāyām and vā are carried from SK 2608 (TB; WS BM).
• In putrīyati, putram is the sup-final object, and the son is wanted for oneself (ātmanaḥ). After kyac, the sup am undergoes luk, being now part of a dhātu (sanādyantā dhātavaḥ, SK 2304; BM): putra-ya.
• ātmanaḥ excludes parasya putram icchati (TB). By the samartha paribhāṣā there is no kyac for mahāntaṃ putram icchati (WS BM).
• 2658 kyaci ca 7.4.33 — before kyac, ī replaces a. asya is carried from asya cvau (7.4.32, SK 2118), and ī from ī ghrādhmoḥ (7.4.31, SK 2648) (BM).
• The a of putra before kyac → ī: putrīyati.
• With o and au: vānto yi pratyaye (6.1.79, SK 63) gives gavyati and nāvyati. lopaḥ śākalyasya (8.3.19, SK 67) does not delete the v, which is not pada-final. "For it is as follows" leads into 2659 (BM).
• 2659 naḥ kye 1.4.15 — before kyac and kyaṅ, only an n-final sup-final word is a pada. It is a niyama on suptiṅantaṃ padam (1.4.14, SK 29). kya does not include kyaṣ, which never follows a consonant-final (BM, TB).
• rājan-ya: being a pada, the stem loses its n, and 2658 gives ī: rājīyati. The n-deletion is not asiddha for the ī, by the restriction in nalopaḥ supsvara… (8.2.2, SK 353) (BM, TB).
• Not a pada:
• vācyati: no kutva, and no saṃprasāraṇa by vacisvapi… (6.1.15, SK 2409; BM);
• samidhyati: no jaśtva;
• adasyati: no tyadādi a, no u/m, no ru (BM).
• By the sannipāta paribhāṣā, the y of kyac is not deleted by yasya halaḥ (6.4.49, SK 2631) after the v it caused: gavyāṃcakāra, gavyitā; nāvyāṃcakāra, nāvyitā.
• pratyayottarapadayoś ca (7.2.98, SK 1373) gives tvadyati, madyati for a single referent; with more than one, yuṣmadyati, asmadyati. ekavacane is carried from tvamāv ekavacane (7.2.97, SK 389; BM).
• hali ca (8.2.77, SK 354) applies only after a dhātu: gīryati, pūryati, but divyati (div is an underived word).
• Mādhava's puryati is cintya, since pur and gir are alike.
• The prācaḥ's dīvyati is a blunder.
• Also:
• kartrīyati (rīṅ ṛtaḥ, 7.4.27, SK 1234);
• gārgīyati (kyacvyoś ca, 6.4.152, SK 2119);
• kavīyati (akṛtsārva-, SK 2298).
• 2660 kyasya vibhāṣā 6.4.50 — after a consonant, kyac and kyaṅ are optionally deleted before an ārdhadhātuka. halaḥ is carried from SK 2631, and ārdhadhātuke is the adhikāra of 6.4.46 (SK 2307) (BM).
• samidh-ya-itā:
• ādeḥ parasya (1.1.54, SK 44) puts the deletion on the y;
• ato lopaḥ (6.4.48, SK 2308) deletes the a;
• that deletion is sthānivat, so there is no laghūpadha guṇa: samidhitā.
• Without the deletion: samidhyitā.
• vārttika (sk.2660.k2) māntaprakṛtikasubantād avyayāc ca kyac na: no kyac after a sup-final whose base ends in m, or after an indeclinable. So kim icchati, idam icchati and svar icchati stay phrases.
• Bhaṭṭoji rewords māntāvyayebhyaḥ pratiṣedhaḥ so that putram icchati still gives putrīyati (BM, TB).

B. Other forms in the wishing sense: nipātana, asuk, kāmyac — running example: aśva + asuk + kyac → aśvas-ya → aśvasyati (vaḍavā); putram ātmana icchati: putra + kāmyac → putrakāmyati

• 2661 aśanāyodanyadhanāyā bubhukṣāpipāsāgarddheṣu 7.4.34 — aśanāya, udanya and dhanāya are kyac-final nipātanas in the senses of hunger, thirst and greed, respectively (krameṇa, BM).
• The forms:
• aśanāyati (no ī; ā by akṛtsārva-);
• udanyati (udan for udaka, and the n is not deleted);
• dhanāyati.
• Outside those senses: aśanīyati, udakīyati, dhanīyati. That is food for later use, water for bathing, and wealth wanted by a poor man (BM, TB).
• The print reads gardheṣu (an orthographic variant).
• 2662 aśvakṣīravṛṣalavaṇānām ātmaprītau kyaci 7.1.51 — aśva, kṣīra, vṛṣa and lavaṇa take the augment asuk before kyac, in the sense of one's own gratification. asuk is carried from āj jaser asuk (7.1.50, SK 3572) (BM, TB). Being k-it, it attaches as the final part.
• vārttika (k2) aśvavṛṣayor maithunecchāyām: aśva-as-ya → aśvasyati vaḍavā, vṛṣasyati gauḥ.
• vārttika (k4) kṣīralavaṇayor lālasāyām: kṣīrasyati bālaḥ, lavaṇasyati uṣṭraḥ.
• vārttika (k6) sarvaprātipadikānāṃ kyaci lālasāyāṃ sugasukau: suk or asuk for every prātipadika before kyac, in the sense of craving. BM reads it as giving kyac itself as well (audit).
• Forms: dadhisyati, dadhyasyati; madhusyati, madhvasyati (print; ours madhyavasyati).
• 2663 kāmyac ca 3.1.9 — kāmyac comes in the same domain as kyac (2657). Its k is pronounced, so it is not an it (TB).
• putrakāmyati. The y is not deleted by yasya halaḥ, because:
• it is a meaningless part of kāmyac;
• yasya there takes the sequence y + a, as in bebhiditā, where the ya of yaṅ is meaningful (BM).
• yaśaskāmyati (s by so 'padādau, 8.3.38, SK 152; BM), sarpiṣkāmyati.
• The m-final and indeclinable bar of 2660 does not continue: kiṃkāmyati, svaḥkāmyati (TB).

C. Behaving like: kyac after an object as standard, kyaṅ after an agent as standard — running example: putram ivācarati chātram: putra + kyac → putrīyati chātram; kṛṣṇa ivācarati: kṛṣṇa + kyaṅ → (lengthening, SK 2298) kṛṣṇāya → kṛṣṇāyate

• 2664 upamānād ācāre 3.1.10 — kyac comes after a sup-final object that is a standard of comparison, in the sense of behaving. supaḥ and kyac are carried from 2657, and karmaṇaḥ and vā from SK 2608 (BM).
• putrīyati chātram ("treats the pupil like a son"): putra is the upamāna, and chātra is the one so treated. Also viṣṇūyati dvijam.
• vārttika (k2) adhikaraṇāc ca: the same from a locus as standard (a seventh-case word, TB). prāsādīyati kuṭyāṃ bhikṣuḥ, kuṭīyati prāsāde (print; ours prasāde).
• 2665 kartuḥ kyaṅ salopaś ca 3.1.11 — kyaṅ comes optionally after an agent that is a standard of comparison, in the sense of behaving. An s-final agent-word loses its s. upamānād ācāre is carried from 2664, and vā from SK 2608.
• Construal. BM reads kartuḥ twice. The first is a fifth case: "kartuḥ kyaṅ". The second is a sixth case, qualified by "sa" (its sixth-case ending elided, with tadantavidhi): "kartuḥ salopaś ca". BM takes the ca as standing for tu; TB takes it as anvācaya.
• The option. Without kyaṅ the analytic phrase stands. The s-deletion comes only together with kyaṅ (kyaṅsanniyogaśiṣṭa), and that s-deletion is fixed in its distribution (vyavasthita, BM).
• kṛṣṇāyate: kṛṣṇa is the agent-standard. kyaṅ follows, a is lengthened by akṛtsārva- (SK 2298), and kyaṅ's ṅ gives ātmanepada (BM).
• vārttika (k2) ojaso 'psaraso nityam itareṣāṃ vibhāṣayā:
• s-deletion is obligatory in ojāyate (within the vṛtti, ojas = "one possessing ojas") and apsarāyate;
• it is optional in yaśāyate / yaśasyate and vidvāyate / vidvasyate. vidvasyate has no ru, not being a pada by 2659 (TB).
• tvadyate, madyate; with more than one referent, yuṣmadyate, asmadyate.
• puṃvat by kyaṅmāninoś ca (6.3.36, SK 837):
• kumārāyate (ṅīṣ removed).
• hariṇīvācarati → haritāyate. Here hariṇī is the feminine of the colour-word harita. Puṃvat removes the n of varṇād anudāttāt… (4.1.39, SK 496) and the ṅīṣ (BM).
• gurūyate.
• sapatnīva stands for three words (TB):
• sapatnī "female enemy" (ṅīn by śārṅgaravādi, 4.1.73, SK 527) gives sapatnāyate;
• sapatnī "having the same master" (nityaṃ sapatnyādiṣu, 4.1.35, SK 492; ṅīp) gives sapatīyate;
• sapatnī "co-wife" is always feminine, so there is no puṃvat: sapatnīyate.
• yuvāyate. Age is not a jāti, so jāteś ca does not block puṃvat (BM, TB).
• paṭvīmṛdūyate (print paṭvīmṛdvyāv iva): the first member is not followed by kyaṅ, so it gets no puṃvat.
• na kopadhāyāḥ (6.3.37, SK 838) bars puṃvat for a penultimate k: pācikāyate, not pācakāyate.
• The unit's kvip vārttikas (k4–k7) are taken in D.

D. Behaving like, with kvip; the nasal-final stem lengthened before kvi — running example: rājeva ācarati: rājan + kvip → (2666) rājān → (no pada, n kept) rājānati; kṛṣṇa ivācarati: kṛṣṇa + kvip → kṛṣṇa-a-ti → (ato guṇe) kṛṣṇati

• 2665 (continued: vārttikas k4–k7)
• vārttika (k4) ācāre 'vagalbhaklībahoḍebhyaḥ kvib vā vaktavyaḥ:
• Optional kvip after avagalbha, klība and hoḍa. By the vā, kyaṅ comes too (BM): avagalbhāyate (TB).
• The three end in pacādi ac. With kvip, that ac is stipulated to be anudātta and nasal, so its a is an it. That gives taṅ: avagalbhate (print; ours agalbhate), klībate, hoḍate.
• The vārttika's very undertaking shows that ām comes even after a stem that was once polysyllabic: avagalbhāṃcakre, klībāṃcakre, hoḍāṃcakre.
• From plain galbha, or galbha with another upasarga, Mādhava allows only kyaṅ. Bhaṭṭoji holds it more proper to say that taṅ does not arise there.
• vārttika (k6) sarvaprātipadikebhyaḥ kvib vā vaktavyaḥ: optional kvip after every prātipadika. The previous vārttika only attaches the anubandha, and kvip is restated in it.
• Because prātipadika is named, supaḥ is not connected, and no pada operation applies: rājānati (BM), tatati, tvacati (TB).
• The forms:
• kṛṣṇati: pararūpa with śap by ato guṇe (6.1.97, SK 191).
• a (= Viṣṇu): ati, ataḥ, anti.
• liṭ without ām, since the vārttika on kāspratyayāt (3.1.35, SK 2306) reads anekācaḥ: au, atuḥ, uḥ.
• The steps: doubling → ato guṇe → ata ādeḥ (7.4.70, SK 2248) → au for ṇal (āta au ṇalaḥ, 7.1.34, SK 2371, BM) → vṛddhi.
• In atus etc., ā is deleted by āto lopa iṭi ca (6.4.64, SK 2372).
• mālāti: by the liṅgaviśiṣṭa paribhāṣā, or with the single substitute taken as the end of the preceding part (pūrvānta).
• Also mālāṃcakāra, and laṅ amālāt. hal-ṅy-āb… (6.1.68, SK 252) does not delete the t: by association with ṅī, it deletes only su after āp.
• With iṭ and sak: amālāsīt.
• kavayati, āśīrliṅ kavīyāt.
• dhātoḥ carried into sici vṛddhiḥ… (7.2.1, SK 2297) means "purely a dhātu", so a nāmadhātu gets no vṛddhi: akavayīt (Kaiyaṭa).
• [Mādhava: akavāyīt] (print; missing in ours).
• vayati (vi, a bird), vivāya, vivyatuḥ, avayīt / avāyīt.
• śrayati, śiśrāya, śiśriyatuḥ.
• pitarati, and in the benedictive with rīṅ, pitriyāt.
• bhavati (bhūr iva): bubhāva, abhāvīt. By the abhivyakta maxim these rules take only the root of the dhātupāṭha, so none applies:
• gātisthā… (2.4.77, SK 2223);
• bhuvo vuk (6.4.88, SK 2174);
• bhavater aḥ (7.4.73, SK 2181).
• dravati: no caṅ by ṇiśri… (3.1.48, SK 2312), so adrāvīt.
• 2666 anunāsikasya kvijhaloḥ kṅiti 6.4.15 — a long vowel replaces the penultimate of an aṅga ending in a nasal, before kvi and before a jhal-initial k-it or ṅ-it affix. upadhāyāḥ is carried from nopadhāyāḥ (6.4.7, SK 370), and dīrghaḥ from ḍhralope… (6.3.111, SK 174) (BM, TB). TB's jhal cases: śāntaḥ, dāntaḥ.
• idam + kvip: the a before m is lengthened, giving idāmati. Likewise rājānati, pathīnati, mathīnati and ṛbhukṣīṇati.
• On pathīnati, TB records three views and does not decide. BM prefers pathenati, because inhan… (6.4.12, SK 356) restricts the lengthening to before śi.
• BM adds that in fact there is no ācāra-kvip after consonant-finals (Śabdenduśekhara).
• The kvip forms continue:
• dyaur iva: devati (Mādhava; no u by diva ut, 6.1.131, SK 337). Bhaṭṭoji holds dyavati proper, with ūṭh by cchvoḥ… (6.4.19, SK 2561).
• ka iva: kati. In liṭ, cakau (Haradatta), or caka (Mādhava): by ṇyallopau…, the a-deletion sets aside vṛddhi.
• sva iva: svati, sasvau, sasva. Svāmī's svāṃcakāra (print) is anākara. BM and TB note that the Bhāṣya never says "pratyayagrahaṇam apanīya", so ām after an affix-final stem is available and the form is in fact defensible.

E. Becoming: kyaṅ after bhṛśa and the rest, kyaṣ after lohita and ḍāc-finals — running example: abhṛśo bhṛśo bhavati: bhṛśa + kyaṅ → bhṛśāya → bhṛśāyate; lohita + kyaṣ → lohitāya → lohitāyati / lohitāyate

• 2667 bhṛśādibhyo bhuvy acver lopaś ca halaḥ 3.1.12 — kyaṅ (from 2665, BM) comes after bhṛśa etc., when they are not cvi-final, in the sense of becoming. A consonant-final member loses that consonant.
• acveḥ is a paryudāsa, which gives the restriction to abhūtatadbhāva.
• bhṛśāyate. The Bhāṣya's counter-example: kva divā bhṛśā bhavanti ("where are the bright-at-night [stars] by day?"; TB).
• sumanas: the s is deleted, giving sumanāyate. sumanas here is the bahuvrīhi "good-minded" (BM).
• The saṃgrāma jñāpaka.
• The curādi list reads "saṃgrāma yuddhe". "grāma" alone would have implied sam, so reading sam shows that a first member with the form of an upasarga is separated when a dhātu-making affix is due.
• Hence aṭ before manas: svamanāyata; also unmanāyate, udamanāyata. The same separation of ava gives avāgalbhata, avāgalbhiṣṭa.
• The jñāpaka covers only like cases (sajātīyaviṣaya): only where the whole upasarga is heard.
• ā + ūḍhaḥ = oḍhaḥ. The ā is not separated, so there is no compound and no lyap: oḍhāyitvā.
• Contrast unmanāyya and avagalbhya, which take lyap by samāse 'nañpūrve ktvo lyap (7.1.37, SK 3332), the compound coming from kugatiprādayaḥ (2.2.18, SK 761) (BM).
• vārttika (print ❀) usyomāṅkṣv āṭaḥ pratiṣedhaḥ: āṭ takes no pararūpa before us, om or āṅ.
• ausrīyat (against usy apadāntāt, 6.1.96, SK 2214).
• auṅkārīyat and auḍhīyat (against omāṅoś ca, 6.1.95, SK 80).
• That this is already established by the ca of āṭaś ca (6.1.90, SK 269) is "laid down in the sixth [adhyāya]". The Bhāṣya's counter-example on usy apadāntāt is kosrā (TB, print).
• 2668 lohitādiḍājbhyaḥ kyaṣ 3.1.13 — kyaṣ comes after lohita etc. and after ḍāc-final stems, in the sense of becoming (bhuvi from 2667, BM).
• acveḥ also continues, but for lohita only; the kaumudī states this under 2669.
• There is no TB on 2668 (print). Its forms come under 2669.
• 2669 vā kyaṣaḥ 1.3.90 — after a kyaṣ-final, parasmaipada is optional. parasmaipada is carried from śeṣāt kartari parasmaipadam (1.3.78, SK 2159; BM).
• lohitāyati / lohitāyate ("what was not red becomes red").
• abhūtatadbhāva qualifies lohita alone.
• It cannot qualify ḍāc, which imitates a sound (BM).
• The "ādi" is rejected. vārttika (k2) lohitaḍājbhyaḥ kyaṣvacanaṃ bhṛśādiṣv itarāṇi.
• The k of kyaṣ is also rejected in the Bhāṣya, so 2659 and 2660 do not take kyaṣ (TB).
• paṭapaṭāyati, paṭapaṭāyate. ḍāc comes without kṛ, bhū or as, because kyaṣ is prescribed after ḍāc-finals.
• BM's derivation: paṭat is doubled for ḍāc, the t + p take a single pararūpa, and the ṭi is deleted.
• Some read a list (lohita, śyāma, duḥkha, … carman) and give kyaṣ forms in both padas. That conflicts with the Bhāṣya, so these take only kyaṅ: śyāmāyate. Within the vṛtti, sukha, duḥkha and the like mean "possessing that".
• liṅgaviśiṣṭa: lohinīyati, lohinīyate. TB: with kyaṅ, the form would be taṅ only, with puṃvat.

F. kyaṅ in particular senses, with an object — running example: śabdaṃ karoti: śabda + kyaṅ → śabdāya → śabdāyate

• 2670 kaṣṭāya kramaṇe 3.1.14 — kyaṅ comes after kaṣṭa in the fourth case, in the sense of zeal (utsāha). It is carried from 2665 as svarita; kyaṣ is not (TB, BM).
• kaṣṭāya kramate → kaṣṭāyate: "he is eager to commit sin".
• The fourth case comes by kriyārthopapadasya… (2.3.14, SK 581; BM).
• kramate takes taṅ by vṛttisarga… (1.3.38, SK 2711).
• vārttika (k2) satrakakṣakaṣṭakṛcchragahanebhyaḥ kaṇvacikīrṣāyām: kaṇva is sin. Within the vṛtti, satra etc. mean sin, and kyaṅ follows them as second-case words: satrāyate, kakṣāyate.
• "pāpaṃ cikīrṣati" is an analysis not in the derivative's own words (asvapadavigraha).
• 2671 karmaṇo romanthatapobhyāṃ varticaroḥ 3.1.15 — kyaṅ comes after romantha and tapas as objects, in the senses of turning over and practising, respectively.
• romanthaṃ vartayati → romanthāyate.
• vārttika (k2) hanucalana iti vaktavyam: only with movement of the jaw. So not kīṭo romanthaṃ vartayati.
• Kaiyaṭa reads that as "eats it"; the Nyāsa and Haradatta read it as "makes it round".
• vārttika (k4) tapasaḥ parasmaipadaṃ ca: tapaś carati → tapasyati. Here kyaṅ comes from the sūtra, and parasmaipada sets aside the ṅ-ātmanepada (BM).
• The print reads vartanāyāṃ (ours vartanīyāṃ).
• 2672 bāṣpoṣmabhyām udvamane 3.1.16 — kyaṅ comes after bāṣpa and ūṣman as objects (karmaṇaḥ from 2671, BM), in the sense of emitting.
• bāṣpam udvamati → bāṣpāyate, ūṣmāyate.
• vārttika (k2) phenāc ceti vācyam: phenāyate.
• There is no TB on 2672.
• 2673 śabdavairakalahābhrakaṇvameghebhyaḥ karaṇe 3.1.17 — kyaṅ comes after śabda, vaira, kalaha, abhra, kaṇva and megha as objects (TB), in the sense of karoti. It is an apavāda of the ṇic of tat karoti (BM).
• śabdaṃ karoti → śabdāyate.
• The Nyāsa also allows ṇic: śabdayati. BM reads the attribution as aruci, since this is not in the Bhāṣya.
• vārttika (k2) sudinadurdinanīhārebhyaś ca (BM supplies kyaṅ in the karoti sense): sudināyate.
• 2674 sukhādibhyaḥ kartṛvedanāyām 3.1.18 — kyaṅ comes after sukha etc. as objects (BM), in the sense of awareness, when sukha etc. belong to the agent of the awareness. kartṛ is a separate word with its sixth-case ending elided (TB).
• The kaumudī gives only the phrase sukhaṃ vedayate. The print has no derived form; WS adds sukhāyate.
• Why kartṛ? parasya sukhaṃ vedayate: only the phrase is used. sukhādi includes sukha, duḥkha, tṛpra and kṛcchra (TB).

G. Other affixes in the karaṇa sense: kyac for namas…, ṇiṅ for puccha…, ṇic for muṇḍa… — running example: namas + kyac → namasya → namasyati devān (= pūjayati); muṇḍaṃ karoti: muṇḍa + ṇic → muṇḍayati

• 2675 namovarivaścitraṅaḥ kyac 3.1.19 — kyac comes after namas, varivas and citra(ṅ), in particular actions (karaṇe from 2673): worship, service and wonder.
• namasyati devān ("worships the gods"). The second case stands because a kāraka-vibhakti is stronger (BM).
• varivasyati gurūn ("serves the teachers").
• citrīyate ("is amazed"; others: "causes amazement"). kyac rather than kyaṅ is prescribed for citra so that ī comes (2658), and its ṅ is for taṅ (TB).
• 2676 pucchabhāṇḍacīvarāṇ ṇiṅ 3.1.20 — ṇiṅ comes after puccha, bhāṇḍa and cīvara, for particular actions (karaṇe, TB).
• The ṅ serves two ends. It lets ṇer aniṭi (6.4.51, SK 2313) take ṇiṅ together with ṇic, and it makes the ātmanepada invariable (TB).
• The print has no gloss of the sūtra (our k1 is a bare daṇḍa).
• vārttika (k2) pucchād udasanavyasanaparyasaneṣu: utpucchayate, vipucchayate, paripucchayate.
• vārttika (k4) bhāṇḍāt samācayane: saṃbhāṇḍayate. The luṅ is samababhāṇḍata, with aṭ after sam by the separation of 2667 (BM). Ours reads samavabhāṇḍata.
• vārttika (k6) cīvarād arjane paridhāne ca: saṃcīvarayate bhikṣuḥ.
• 2677 muṇḍamiśraślakṣṇalavaṇavratavastrahalakalakṛtatūstebhyo ṇic 3.1.21 — ṇic comes after muṇḍa, miśra, ślakṣṇa, lavaṇa, vrata, vastra, hala, kala, kṛta and tūsta, in the sense of kṛ (karaṇe from 2673, BM; print kṛñarthe).
• muṇḍaṃ karoti → muṇḍayati.
• vārttika (k2) vratād bhojanatannivṛttyoḥ: payo vratayati ("he eats milk"; BM, TB) and śūdrānnaṃ vratayati ("he avoids a śūdra's food").
• vārttika (k4) vastrāt samācchādane: saṃvastrayati.
• vārttika (k6) halyādibhyo grahaṇe. By nipātana (k8; plain type in the print), hali and kali are made a-final: halayati, kalayati.
• Vṛddhi comes first (paratva), and iṣṭhavat still deletes an ak, the ā.
• So the stem is aglopin, with no sanvadbhāva and no lengthening: ajahalat, acakalat.
• kṛtayati ("accepts a service done"), and vitūstayati (tūsta means hair, matted hair, or sin).
• Why name them, when prātipadikād dhātvarthe (gaṇasūtra under SK 2573) already gives ṇic?
• muṇḍa etc. are named so that ṇic comes even when the word depends on another word (sāpekṣa): muṇḍayati māṇavakam, miśrayaty annam, ślakṣṇayati vastram, lavaṇayati vyañjanam.
• hali and kali are named for their a-final form.
• satya (in satyāpapāśa…, 3.1.25, SK 2563) is named for āpuk: satyāpayati.
• The others are named for elaboration only.
• vārttika (k10) arthavedayor apy āpug vaktavyaḥ: arthāpayati, vedāpayati.
• The rest of the unit (k11–k13) is in H.

H. ṇic after any prātipadika, treated as before iṣṭhan — running example: pṛthum ācaṣṭe: pṛthu + ṇic → (iṣṭhavat: ra for ṛ, ṭi-lopa) prath-i → prathayati; luṅ apiprathat (ṭi-lopa after vṛddhi) / apaprathat (before)

• 2677 (continued: k11–k13) — the forms of the gaṇasūtra prātipadikād dhātvarthe bahulam iṣṭhavac ca (SK 2573):
• The satyāpa… list (SK 2563):
• vipāśayati, rūpayati, upavīṇayati, anutūlayati, upaślokayati;
• abhiṣeṇayati (ṣ by upasargāt sunoti…, 8.3.65, SK 2270);
• abhyaṣeṇayat (prāk sitāt…, 8.3.63, SK 2276);
• abhiṣiṣeṇayiṣati (sthādiṣv abhyāsena ca…, 8.3.64, SK 2277);
• anulomayati;
• tvacayati: tvaca is from tvaca saṃvaraṇe with gha. tvac would give tvayati (BM).
• saṃvarmayati, varṇayati, avacūrṇayati.
• puṃvadbhāva and the rest, by iṣṭhavat: etayati (enī), dāradayati (daradam).
• ra for ṛ: prathayati.
• When ṭi-lopa follows vṛddhi, the au deleted is not an ak: sanvadbhāva gives apiprathat.
• When it precedes, the stem is aglopin: apaprathat (BM, TB).
• Also mradayati; [amimradat,] amamradat, bhraśayati, kraśayati, draḍhayati; ababhraśat, acakraśat, adadraḍhat, and [parivraḍhayati; paryavavraḍhat].
• ūḍhi: [aujiḍhat]. hti is doubled because ḍhatva etc. are asiddha, even though pūrvatrāsiddhīyam advitve is anitya. Others double ḍhi: auḍiḍhat.
• ūḍha: aujaḍhat, auḍaḍhat. The abhyāsa keeps a by the oḥ puyaṇji jñāpaka (SK 2577): no vowel-substitute before ṇi while doubling is due.
• prakṛtyaikāc (6.4.163, SK 2010) blocks ṭi-lopa: svāpayati (vṛddhi, puk).
• For tvam / aham, the older view makes ṭi-lopa first as nitya: tvāpayati, māpayati.
• Bhaṭṭoji's view, tvādayati, mādayati, is the right one: pararūpa comes first as antaraṅga, then prakṛtibhāva.
• The Bhāṣya does not reject prakṛtibhāva.
• Dual: yuṣmayati, asmayati.
• śvan: śāvayati.
• ṭi-lopa is by nas taddhite (6.4.144, SK 679). By yena nāprāpti, prakṛtyaikāc blocks only ṭeḥ (6.4.155, SK 1786).
• The stem is bha, so saṃprasāraṇa follows.
• Others hold that nas taddhite is not extended: śunayati.
• vidvas: vidvayati (some; aṅgavṛtta), vidāvayati (others), vidayati (yet others: saṃprasāraṇa as nitya, pūrvarūpa, ṭi-lopa). BM keeps the first.
• The añc stems (the upasarga separated, so prakṛtyaikāc):
• udīcayati; udaicicat;
• pratīcayati; pratyacicat, and with prakṛtibhāva by ikaḥ 'savarṇe… (6.1.127, SK 91), pratiacicat;
• samīcayati; [samyacicat,] samiacicat.
• tiryañc: tirāyayati.
• tiras gets tiri: the ṭi-lopa is bahiraṅga and so asiddha.
• By asiddhavad atrābhāt (6.4.22, SK 2183), on the ciṇo luk analogy, there is no second ṭi-lopa. The aṅgavṛtta paribhāṣā gives the same result.
• luṅ atitirāyat: aglopin, so no shortening.
• The other -añc stems:
• sadhrāyayati; asasadhrāyat;
• aviviṣvadrāyat;
• devadrāyayati; adidevadrāyat;
• adadadrāyat;
• adamuāyayati; ādadamuāyat;
• amumuāyayati; āmumuāyayat (transmitted, not reconstructed).
• Further luṅ forms: bhāvayati; abībhavat, abubhravat, aśiśrayat, ajūgavat, arīrayat, anūnavat, svāśaśvat (sva separated).
• (k12; plain type in the print) avyayānāṃ bhamātre ṭilopaḥ: svayati.
• asasvat: svar is doubled, the ac-substitute being barred.
• asisvat: ṭi-lopa first.
• bahu: bhāvayati; others give bahayati, which TB rejects.
• luk by vinmatoḥ:
• srajayati: no vṛddhi, being saṃjñāpūrvaka;
• śrayayati; aśiśrayat;
• payasayati: no ṭi-lopa, since luk, its exception, has applied.
• yaṇādi-lopa: sthavayati, davayati. "dūrayaty avanate vivasvati" is explained from a stem dūrāt.
• Substitutes:
• yavayati, kanayati (yuvālpayoḥ…, 5.3.64, SK 2019);
• nedayati, sādhayati;
• praśasyayati: no śra or jya, the upasarga being separated;
• jyāpayati, varṣayati, prāpayati, [sthāpayati], sphāpayati;
• varayati, vārayati, baṃhayati, garayati;
• trāpayati (print, TB; BM trapayati);
• drāghayati, vṛndayati.
• Close: "[iti nāmadhātuprakriyā]", supplied from the print. p59 opens with SK 2678.