Outline · Prakaraṇa 56

तिङन्तयङ्प्रकरणम्

Covers SK 2629–2649 (21 sūtras): all present. Every unit has its own kaumudī. The spine is not an Aṣṭādhyāyī run. The SK gathers the rules that bear on the intensive yaṅ from AS books 3, 6, 7 and 8: 3.1.22 → 7.4.82 → 6.4.49 → 7.4.83 → 7.4.30 → 3.1.23–24 → 7.4.87–88 → 7.4.86 → 8.2.20 → 8.3.112 → 7.4.63 → 7.4.84–85 → 7.4.90 → 6.1.19–21 → 7.4.31 → 7.4.22. There are three short runs: 3.1.22–24 (when yaṅ comes, split by the abhyāsa rules), 6.1.19–21 (saṃprasāraṇa and kī), and the abhyāsa rules of 7.4.82–90, taken out of order (82, 83, then 87, 88, 86, then 84, 85, 90; 89 is not here). Every AS-ref below is checked against references/ashtadhyayi/sutraani-data.json, and all 21 agree with our sūtra fields.

Anuvṛtti from p55. atra lopo 'bhyāsasya (7.4.58, SK 2620) gives abhyāsasya to 2630, 2632, 2636, 2637 (recast as abhyāsāt), 2641, 2642, 2643 and 2644.
Anuvṛtti from later in this prakaraṇa. caraphaloś ca (2636) takes ataḥ nuk from nug ato 'nunāsikāntasya (7.4.85), which the SK teaches only at 2643. It comes before 2636 in the Aṣṭādhyāyī.
Anuvṛtti within it.
yaṅ from 2629 runs into 2634 and 2635, and nityam from 2634 into 2635.
yaṅlukoḥ from 2630 runs into 2632, 2636–2638, 2642, 2643 and 2644.
yaṅi comes from 2633 into 2648, and from 2645 into 2647.
Everything carried into 2636 is carried into 2638 as well (BM).

tiṅantayaṅprakaraṇam — outline (p56, SK 2629–2649, all 21 sūtras)

p55 (tiṅantasanprakaraṇam) closed at 2628 saḥ svidisvadisahīnāṃ ca with the verse on san after a sannanta. The print adds the colophon "iti sannantaprakriyā", which ours lacks.

Where the prakaraṇa begins. The print heads it "tiṅante yaṅantaprakriyā 53" (NSP p. 437, scan page-0447). Our name and number, tiṅantayaṅprakaraṇam 56, follow Wikisource style and our own numbering. sk2629 opens with the rule dhātor ekāco halādeḥ kriyāsamabhihāre yaṅ (3.1.22).
Where it ends. At 2649 ayaṅ yi kṅiti, with ḍoḍhaukyate and totraukyate. The print closes "iti yaṅantaprakriyā" (NSP p. 439, scan page-0449). Ours lacks it, so the English supplies it in brackets. The print then heads "tiṅante yaṅlugantaprakriyā 54", and p57 opens with yaṅo 'ci ca (2.4.74, SK 2650).
Contents in outline.
2629, 2634, 2635: when yaṅ comes.
Repetition or intensity of the action: bobhūyate.
With motion roots, only crookedness: vāvrajyate.
With lup and the rest, only censure: lolupyate.
2630–2632: the yaṅ-final stem. Guṇa of the abhyāsa (bobhūyate). ya is deleted after a consonant before an ārdhadhātuka (sosūcāṃcakre, sosūcitā). The abhyāsa is lengthened unless it has a k-marked augment (aṭāṭyate).
2633, 2648, 2649: the root's vowel before yaṅ. Guṇa for ṛ (arāryate), ī for ghrā and dhmā (jeghrīyate), ay for śī (śāśayyate). The tail of 2649 adds ḍoḍhaukyate.
2636–2638, 2643: nuk, as the anusvāra, after the a of the abhyāsa.
car and phal, with u for the following a: cañcūryate, pamphulyate.
jap and the rest: jañjapyate.
Every nasal-final root: yaṃyamyate.
2639, 2645–2647: root substitutes. l for the r of gṝ (jegilyate). saṃprasāraṇa for svap, syam and vyeñ (soṣupyate), but not for vaś (vāvaśyate). kī for cāy (cekīyate).
2640–2641: two blocked changes. No ṣ for the s of sic (nisesicyate). No palatal in the abhyāsa of kavati (kokūyate).
2642, 2644: the augments nīk and rīk. vanīvacyate, varīvṛtyate.

Threads that run through the prakaraṇa:

yaṅ makes a dhātu. The yaṅ-final is a dhātu by sanādyantā dhātavaḥ (3.1.32, SK 2304), so the lakāras follow. Because yaṅ is ṅ-marked, the endings are ātmanepada (anudāttaṅita…, 1.3.12, SK 2157), and the root's own vowel takes no guṇa: the ū of bobhūyate stays (BM).
yaṅ is made manifest (dyotya), not denoted. If it were denoted it would be the principal meaning, the qualificand, as desire is in a sannanta (BM, 2629). So yaṅ can come even where bhṛśa or punaḥ punaḥ already says the same (BM, 2630).
"dhātoḥ" in 2629 makes yaṅ ārdhadhātuka. So bruvo vaciḥ (2.4.53, SK 2453) and the like apply before it: vāvacyate, cākhyāyate (BM; the forms are TB's).
The abhyāsa frame. The steps are:
doubling by sanyaṅoḥ (6.1.9, SK 2395), or of the second syllable for a vowel-initial root (ajāder dvitīyasya, 6.1.2, SK 2176);
halādiḥ śeṣaḥ (7.4.60, SK 2179);
guṇa of the abhyāsa (2630);
lengthening of the abhyāsa (2632).

The prakaraṇa adjusts the frame at these points:

augments to the abhyāsa: nuk (2636, 2638, 2643), nīk (2642), rīk (2644). They are k-marked, so the abhyāsa is not lengthened: vanīvacyate (2642), and yaṃyamyate is what akitaḥ guards (BM, TB on 2632);
the a after the abhyāsa: u for car and phal (2637);
what is doubled: the root is changed first by ay (2649, TB: para and antaraṅga), by rīṅ (cekrīyate, 2639: para), by saṃprasāraṇa (śośūyate, soṣupyate), and by ī and kī (2647–2648).
"Only" (nityam = eva). In 2634 nitya restricts yaṅ after motion roots to crookedness, not kriyāsamabhihāra, and it runs into 2635 for censure (BM).
TB adds that "nitya" shows the takrakauṇḍinya maxim is not invariable.
BM notes that the Śabdenduśekhara argues for yaṅ in kriyāsamabhihāra after motion roots too. The English keeps Bhaṭṭoji's "na tu kriyāsamabhihāre".
nuk stands for the anusvāra. Bhaṭṭoji says so first under 2636 (lakṣyate). The vārttika sa ca padāntavad vācyaḥ lets vā padāntasya (8.4.59, SK 125) make it homorganic, or leave it.
Under 2643 he refers back with "ity uktam". BM on 2636 also credits the Bhāṣya with this, so the "said" could point to either.
So each nuk form has a pair: cañcūryate / caṃcūryate, pamphulyate / paṃphulyate, yay̐yamyate / yaṃyamyate.
Recalled rules (smāryate). Bhaṭṭoji re-cites earlier rules where the yaṅ forms need them:
ghumāsthā… (6.4.66, SK 2462): dedīyate, pepīyate, seṣīyate, under 2639;
vibhāṣā śveḥ (6.1.30, SK 2420): śośūyate / śeśvīyate, under 2639;
yaṅi ca (SK 2633): sāsmaryate, under 2639;
rīṅ ṛtaḥ (7.4.27, SK 1234) and suṭ: cekrīyate, saṃceskrīyate, under 2639;
ye vibhāṣā (6.4.43, SK 2319): jājāyate / jañjanyate, under 2643;
na vaśaḥ (2646) is itself presented by BM as a recall: Bhaṭṭoji "recalls its prohibition".
The kaumudī lists forms bare. Derivation stages are the translator's, following BM and TB as recorded in the unit notes.
Where the commentary lies. Only the kaumudī blocks are translated (SK-TRANSLATE §25).
No BM in our source for any p56 unit. sk.2628.b1 runs from BM on the sannanta straight into BM on the sarvasamāsaśeṣa. The notes cite BM as read in Wikisource uttara-2/0127–0134.
TB printed under a neighbouring unit:
2629 and 2630 at the tail of sk.2628.t1;
2633 under sk.2632.t1;
2638 at the tail of sk.2634.t1;
2642 and the 2643 vārttika (ghnī) under sk.2641.t1.
There is no TB on 2635, 2636, 2637, 2643's sūtra or 2646.
sk.2645.t1 is garbled (repeats, loses the vevīyate lemma). The print has it whole.
sk.2649.t1 runs on into TB on SK 2650. That is p57's matter.

The p55–58 audit (.work/sk-run/p55-58-audit/sk2628-2637.md, sk2638-2647.md, sk2648-2657.md) found no MUST-FIX in p56. Where its MINOR items bear on a reading, this outline follows the audit. A fixer is applying them to the notes in parallel.

2630: the bobhūyate bracket skips a step. b for bh in the abhyāsa is by abhyāse carca (8.4.54, SK 2182). BM leaves that step implicit too.
2631: TB's second form is aharyāt in the print and OCR, not "ahāryāt".
2632: "ḍoḍhaukyata" in TB is the print's own reading, not an OCR slip. The intended wrong form is presumably ḍauḍhaukyata. "TB also says ṭya is doubled" belongs in the TB note.
2634: the whole kaumudī is on NSP p. 437. Only its TB runs on to p. 438.
2636: "the preceding rule" nug ato 'nunāsikāntasya means preceding in Aṣṭādhyāyī order (7.4.85 before 7.4.87). In the SK it comes later, at 2643, and nuk is carried forward from it.
2637: hali ca (8.2.77, SK 354) lengthens the u of car only (cūr before y). phul has no r or v, so pamphulyate keeps the short u.
2639:
in śośūyate, "guṇa of the abhyāsa" is the translator's step; BM ends at the lengthening;
"[than rīṅ ṛtaḥ]" after sāsmaryate's "para" is the translator's inference, and it is correct (7.4.30 comes after 7.4.27).
2640: s and na are carried forward. ṣ is not: it comes from the mūrdhanyaḥ adhikāra.
2641: there is no root "kav". kavateḥ cites the bhvādi root kuṅ with śap. BM's objection is that the designation with śtip is pointless, and the answer is that kavati is cited with śap.
2642: under this sūtra BM glosses yaṅlukoḥ only as "yaṅi yaṅluki ca". Naming guṇo yaṅlukoḥ (SK 2630) as its source is the translator's inference, made from BM on 2643–2644.
2643: "it has been said [at SK 2636]" is the translator's inference. See the nuk thread.
2646: the yaṅ context is BM's. Its lead-in reads "vaśadhātor yaṅi vāvaśyate…". It is not the translator's alone.

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

2630: ṅidantatvāt, anekāctvena, and yaṅ neti (the print's conjunct is unclear; WS and BM read yaṅ neti).
2631: ato lopaḥ is SK 2308 (WS 2608 is wrong).
2633: sthānivattvāt.
2637: hali ca is SK 354 (WS 654 is wrong).
2639: yaṅi ca is SK 2633.
2644: calīkḷpyate, in the order varīvṛścyate, parīpṛcchyate.
2648: jeghrīyate (WS reads jeghnīyate, a slip).
2649: the colophon "iti yaṅantaprakriyā".
2630: ours gives sanādyantāḥ as SK 2304, which is right. The print's 2604 is a misprint.

Still open for Ajai (mūla):

2639: ours reads "yaṅi ca <SK2623>". The print and WS have 2633. The English gives SK 2633.
2630: "yaṅgeti" in ours. The print's conjunct is unclear, and the sense needs yaṅ neti. Check the printed copy.
2649: the colophon is missing in ours.
The other mūla corrections above (2630 ṅidantatvāt and anekāctvena, 2633 sthānivattvāt) are for Ajai's printed copy.
Variants, not errors: WS adds "yaṅluki ca" at 2633. The "(vārttikam)" tags on 2630.k3, 2636.k3, 2643.k3 and 2644.k3 are editorial; the print marks the vārttika (k2) instead.

Ordering: grouped by outcome, which departs from SK order in these places:

2634–2635 are pulled forward into group A, with 2629.
2648–2649 are pulled forward into group C, with 2633.
2643 is moved into group D, after 2636–2638. Its nuk is the one they carry.
2645–2647 are pulled forward into group E, with 2639.
2644 joins 2642 in group G.
Groups are placed by their first sūtra. Within each group, SK order is kept.

A. When yaṅ comes: repetition or intensity, and the two "only" senses — running example: `bhū + yaṅ (kriyāsamabhihāra) → bobhūyate`; `vraj + yaṅ (crookedness only) → vāvrajyate`; `lup + yaṅ (censure only) → lolupyate` (2634–2635 pulled forward)

2629 dhātor ekāco halādeḥ kriyāsamabhihāre yaṅ 3.1.22 — yaṅ comes after a dhātu that has one vowel and begins with a consonant, when kriyāsamabhihāra is to be made manifest. kriyāsamabhihāra is repetition (paunaḥpunya) or intensity (bhṛśa). It is dyotya, not vācya (BM).
bhū (one vowel, initial bh) + yaṅ: bobhūyate, "it comes into being again and again, or intensely" (formed by 2630).
The tests are given under 2630:
dhātoḥ kim? To make yaṅ ārdhadhātuka, so that vāvacyate and cākhyāyate are formed (bruvo vaciḥ, SK 2453).
ekācaḥ kim? punaḥ punar jāgarti, with no yaṅ.
halādeḥ kim? bhṛśam īkṣate.
Not after ruc and śubh in the sense of intensity. This is the Bhāṣya, and it rests on non-expression (anabhidhāna, BM). In repetition they do take it: rorucyate, śośubhyate.
vārttika (sk.2630.k2): yaṅ also after sūci, sūtri, mūtri, aṭ, ṛ, aśū and ūrṇu. These fail ekāc or halādi (BM). Their forms are given under 2630–2633.
2634 nityaṃ kauṭilye gatau 3.1.23 — after a dhātu of motion, yaṅ comes only when crookedness is meant, not in kriyāsamabhihāra. nitya = eva, and yaṅ is carried from 2629 (BM). TB: nitya shows the takrakauṇḍinya maxim is not invariable.
vāvrajyate, "he goes crookedly".
BM reports that the Śabdenduśekhara argues against this restriction. The English keeps Bhaṭṭoji's "na tu kriyāsamabhihāre".
2635 lupasadacarajapajabhadahadaśagṝbhyo bhāvagarhāyām 3.1.24 — after lup, sad, car, jap, jabh, dah, daś and gṝ, yaṅ comes only when the action (bhāva = dhātvartha) is censured. yaṅ comes from 2629, and nityam (eva) from 2634 (BM).
lolupyate, "he cuts in a censurable way". Also sāsadyate.
The augments these roots need follow: car (2636–2637), jap, jabh, dah, daś (2638), gṝ (2639).

B. The yaṅ-final stem: guṇa and lengthening of the abhyāsa, ya lost before an ārdhadhātuka — running example: `bhū + yaṅ → bhū-bhū-ya (sanyaṅoḥ) → (guṇa of the abhyāsa, 2630; b for bh) bobhūya → bobhūyate`; `aṭ-ya → a-ṭya-ṭya → aṭaṭya → aṭāṭyate` (2632); `sosūcya + ām → sosūcām → sosūcāṃcakre` (2631)

2630 guṇo yaṅlukoḥ 7.4.82 — guṇa replaces the ik vowel of the abhyāsa when yaṅ follows or when there is luk of yaṅ. abhyāsasya comes from SK 2620. "luk" means luk of yaṅ, since yaṅ is in context (BM).
bhū-bhū-ya: the ū of the abhyāsa → o. b for bh is by abhyāse carca (SK 2182), a step BM leaves implicit (audit). The root's ū stays, because yaṅ is ṅ-marked. With laṭ, śap and pararūpa: bobhūyate.
Also bobhūyāṃcakre, abobhūyiṣṭa. The yaṅ-final is a dhātu (SK 2304) and ṅ-marked, hence the ātmanepada.
No further doubling by kriyāsamabhihāre dve vācye or by nityavīpsayoḥ (BM).
The vārttika roots: sūci → ṇi deleted → doubled → guṇa: sosūcyate, sosūtryate, momūtryate.
These roots have more than one vowel, so they are not taught with ṣ (aṣopadeśa). So the second s does not become ṣ (BM).
2631 yasya halaḥ 6.4.49 — the cluster ya (saṃghāta) is deleted after a consonant when an ārdhadhātuka follows. lopa comes from ato lopaḥ (6.4.48, SK 2308), under the ārdhadhātuke adhikāra (SK 2307) (BM).
The deletion falls on the first sound y by ādeḥ parasya (1.1.54, SK 44). The a left over goes by ato lopaḥ.
sosūcya + ām (liṭ) → ya deleted → sosūcāṃcakre. Also sosūcitā, sosūtritā, momūtritā.
Why take the cluster and not y alone? So that the y of īrṣyitā is kept (BM, TB). TB adds aharyāt, and arthavadgrahaṇa: no deletion in putrakāmyā.
halaḥ kim? lolūyitā, popūyitā (TB).
2632 dīrgho 'kitaḥ 7.4.83 — a long vowel replaces the vowel of an abhyāsa that has no k-marked augment, when yaṅ follows or there is luk of yaṅ. abhyāsasya comes from SK 2620, and yaṅlukoḥ from 2630 (BM).
aṭ + yaṅ: ṭya is doubled by ajāder dvitīyasya (SK 2176), and the y of the abhyāsa goes by halādiḥ śeṣaḥ (SK 2179), giving aṭaṭya. The abhyāsa a → ā: aṭāṭyate.
akitaḥ kim? yaṃyamyate, raṃramyate (TB; BM points ahead to SK 2643).
TB's argument: without akitaḥ the long vowel would come after nuk too. akitaḥ is also a jñāpaka that among operations on the abhyāsa the exceptions do not block the general rules. The example is ḍoḍhaukyate, where the later lengthening would otherwise block the shortening. The wrong form is not printed (see the audit item).

C. The root's vowel before yaṅ: guṇa for ṛ, ī for ghrā and dhmā, ay for śī — running example: `ṛ + yaṅ → ar-ya (guṇa, 2633) → 'rya' doubled → a-ra-rya → (lengthening, 2632) arāryate` (2648–2649 pulled forward)

2633 yaṅi ca 7.4.30 — guṇa also replaces the ṛ of ṛ (arti) and of a root beginning with a conjunct when yaṅ follows. The whole of guṇo 'rtisaṃyogādyoḥ (7.4.29, SK 2380) is carried, and ṛtaḥ from rīṅ ṛtaḥ (SK 1234) (BM, TB). Without it, guṇa would be barred because yaṅ is ṅ-marked (BM).
ṛ + yaṅ → guṇa with r (raparatva) → ar-ya.
'rya' is doubled. na ndrāḥ (SK 2446) does not stop an r that is followed by y, because of the Bhāṣya's example. Otherwise *aryāyate would result (BM).
Then halādiḥ śeṣaḥ and lengthening (2632): arāryate.
TB weighs how far the Bhāṣya's indication reaches (āraryat). The English does not decide.
Before an ārdhadhātuka (2631): arāritā; aś → śya doubled → lengthened → ya deleted: aśāśitā.
ūrṇonūyate: by na ndrāḥ, 'nu' is doubled.
bebhidyate: bhid doubled, then guṇa of the abhyāsa.
bebhiditā: the deletion of the a of yaṅ is sthānivat, so the i of bhid takes no laghūpadhaguṇa before tās.
The conjunct-initial case: sāsmaryate (smṛ), recalled under 2639.
2648 ī ghrādhmoḥ 7.4.31 — ī replaces the final of ghrā and dhmā when yaṅ follows. yaṅi comes from 2633 (BM). "[The final of]" is the translator's.
ghrā-ya → ghrī-ya → doubling and the rest: jeghrīyate. Likewise dedhmīyate.
Why a long ī, when akṛtsārvadhātukayoḥ (SK 2298) would lengthen a short i anyway?
BM: for the sake of asya cvau (SK 2118).
TB: first for yaṅluk (jeghrītaḥ), or equally for asya cvau.
The English is neutral.
2649 ayaṅ yi kṅiti 7.4.22 — ay replaces the final of śīṅ before an affix that begins with y and is k- or ṅ-marked. śīṅaḥ comes from śīṅaḥ sārvadhātuke guṇaḥ (7.4.21, SK 2441) (BM).
śī-ya → śay-ya. ay comes first, being later and antaraṅga (TB). Then doubling and lengthening of the abhyāsa: śāśayyate.
The kaumudī tail, on ḍhauk: after doubling, the abhyāsa is shortened to u, and that takes guṇa by 2630: ḍoḍhaukyate. Likewise trauk: totraukyate (BM).
Close: "[iti yaṅantaprakriyā]", supplied from the print. TB closes alike. p57 opens with the yaṅluk, SK 2650.

D. nuk (the anusvāra) after the a of the abhyāsa; u for the a that follows — running example: `car + yaṅ → ca-car-ya → caṃ-car-ya (nuk, 2636) → caṃ-cur-ya (u, 2637) → caṃ-cūr-ya (hali ca) → cañcūryate / caṃcūryate` (2643 moved here, after 2638)

2636 caraphaloś ca 7.4.87 — for car and phal, nuk is added to the a of the abhyāsa when yaṅ follows or there is luk of yaṅ. ataḥ nuk comes from 7.4.85 (taught at SK 2643, the rule before this one in Aṣṭādhyāyī order), abhyāsasya from SK 2620, and yaṅlukoḥ from SK 2630 (BM).
By nuk the anusvāra is meant (lakṣyate). Otherwise, with no jhal following, naś cāpadāntasya jhali (SK 123) could not give an anusvāra (BM, from the Bhāṣya on 7.4.85). In ca-car-ya this gives caṃ-car-ya.
vārttika (sk.2636.k2) sa ca padāntavad vācyaḥ: the anusvāra counts as pada-final. That brings in vā padāntasya (SK 125), which gives an optional homorganic nasal. BM: the vārttika belongs to 7.4.85 and is carried here.
2637 ut parasyātaḥ 7.4.88 — u replaces the a of car and phal that follows the abhyāsa, when yaṅ follows or there is luk of yaṅ. abhyāsasya becomes abhyāsāt, a fifth case, because it is joined with para (BM).
caṃ-car-ya → caṃ-cur-ya. Then hali ca (SK 354) lengthens the u, but only for car: cañcūryate (ñ by vā padāntasya) / caṃcūryate, "he goes about in a censurable way".
phal: pamphulyate / paṃphulyate. The homorganic nasal is m, and the u stays short, since hali ca needs an r or v (audit).
2638 japajabhadahadaśabhañjapaśāṃ ca 7.4.86 — nuk likewise for jap, jabh, dah, daś, bhañj and paś. Everything carried into 2636 runs on here (BM).
jap-ya → ja-jap-ya → jaṃ-jap-ya → jañjapyate, "he mutters in a censurable way".
"And so on" (BM, TB): jañjabhyate, dandahyate, dandaśyate, bambhajyate; pampasyate (TB only).
Mādhava takes the root as the dental-final sautra pas. The Kāśikā has it palatal-final. TB's yaṅluk forms: jañjapīti and the like.
2643 nug ato 'nunāsikāntasya 7.4.85 — nuk is added to an abhyāsa ending in short a belonging to an aṅga that ends in a nasal, when yaṅ follows or there is luk of yaṅ. aṅgasya is the adhikāra; abhyāsasya and yaṅlukoḥ are carried from SK 2620 and SK 2630 (BM).
yam-ya → ya-yam-ya → yaṃ-yam-ya: yay̐yamyate (homorganic nasalised y) / yaṃyamyate. The anusvāra reading was "said" at 2636, or in the Bhāṣya (BM on 2636); see the audit item.
taparatva: ataḥ takes only an a that is short by nature, not one that was formerly long. In bhām (anger), the abhyāsa bhā is shortened to bha, so it gets no nuk (BM), and the abhyāsa is lengthened (2632): bābhāmyate.
Recalled: ye vibhāṣā (SK 2319). ā optionally replaces the n of jan before y: jājāyate. Without it, nuk: jañjanyate.
vārttika (sk.2643.k2) hanter hiṃsāyāṃ yaṅi ghnībhāvo vācyaḥ: for han meaning injuring, ghnī comes before yaṅ: jeghnīyate.
Not injuring (going): nuk, jaṅghanyate.
yaṅi is a viṣayasaptamī, so ghnī comes before doubling (BM; not added to the English).
Why gh and not h? For lightness of derivation (BM, and TB citing the Manoramā). TB also records an opposed view.

E. Root substitutes before yaṅ: l for r, saṃprasāraṇa (but not for vaś), kī for cāy — running example: `svap + yaṅ → sup-ya (saṃprasāraṇa, pūrvarūpa) → su-sup-ya → so-ṣup-ya (guṇa of the abhyāsa; ṣ) → soṣupyate` (2645–2647 pulled forward)

2639 gro yaṅi 8.2.20 — l replaces the r of gṝ (girati) when yaṅ follows. ro laḥ comes from kṛpo ro laḥ (8.2.18, SK 2350) (BM, TB).
jegilyate, "he swallows in a censurable way".
The rest of the kaumudī recalls root rules for yaṅ:
ī by ghumāsthā… (SK 2462), with guṇa: dedīyate, pepīyate, seṣīyate. seṣīyate is from so/sā, not sthā, against BM's "dā pā sthā" (translator).
vibhāṣā śveḥ (SK 2420): saṃprasāraṇa and pūrvarūpa first, then śu doubled and lengthened (SK 2298) → śośūyate. Without it: śeśvīyate.
yaṅi ca (SK 2633): smṛ begins with a conjunct, so guṇa, which is para, comes with r → doubling → lengthening (2632): sāsmaryate.
rīṅ ṛtaḥ (SK 1234): rīṅ comes first (para), then krī is doubled with guṇa: cekrīyate.
With sam, suṭ comes after rīṅ and doubling: saṃceskrīyate. TB's long defence rests on aḍabhyāsavyavāye 'pi suṭ kāt pūrvaḥ.
2645 svapisyamivyeñāṃ yaṅi 6.1.19 — saṃprasāraṇa for svap, syam and vyeñ when yaṅ follows. It is carried from ṣyaṅaḥ saṃprasāraṇam… (6.1.13, SK 1003) (BM).
soṣupyate: svap → sup, doubled, guṇa of the abhyāsa, ṣ in the second part (BM).
sesimyate: syam → sim. It gets no ṣ, because syamu is not ṣopadeśa (BM, TB).
vevīyate: vyeñ → vī. The length is by halaḥ (SK 2559) or akṛtsārva… (SK 2298) (BM, TB).
yaṅi kim? svapnak (TB).
2646 na vaśaḥ 6.1.20 — no saṃprasāraṇa for vaś (when yaṅ follows). It would otherwise come by grahijyā… (6.1.16, SK 2412).
vāvaśyate. The yaṅ context is in BM's lead-in (audit).
2647 cāyaḥ kī 6.1.21 — kī replaces cāy when yaṅ follows. yaṅi comes from 2645 (BM).
cāy-ya → kī-ya → doubling and the rest: cekīyate.
TB: the long ī is pronounced for yaṅluk (cekītaḥ). The English does not add it.

F. Two changes blocked before yaṅ: no ṣ for the s of sic, no palatal in the abhyāsa of kavati — running example: `ni + sic + yaṅ → ni-se-sic-ya → nisesicyate` (neither s → ṣ); `ku(ṅ) + yaṅ → ko-kū-ya → kokūyate` (k kept)

2640 sico yaṅi 8.3.112 — ṣ does not replace the s of sic when yaṅ follows. saḥ (sixth case) comes from saheḥ sāḍaḥ saḥ (8.3.56, SK 335), and na from na raparasṛpi… (8.3.110, SK 3168). ṣ is the mūrdhanyaḥ adhikāra, not an anuvṛtti (BM; audit).
In ni-se-sic-ya, the ṣ that is blocked would come:
for the abhyāsa's s, by upasargāt sunoti… (8.3.65, SK 2270);
for the s after it, by sthādiṣv abhyāsena ca… (8.3.64, SK 2277) (BM, TB).
Result: nisesicyate.
yaṅi kim? abhiṣiṣikṣati (TB).
2641 na kavater yaṅi 7.4.63 — the palatal substitute (cutva) does not replace the velar of the abhyāsa of kavati when yaṅ follows. cuḥ comes from kuhoś cuḥ (7.4.62, SK 2245), and abhyāsasya from SK 2620 (BM). kavati is the bhvādi kuṅ, cited with śap; there is no root "kav" (audit).
kokūyate.
adādi ku (kauti) and tudādi kuṅ (kuvati) do take the palatal: cokūyate.
What the śap-citation chiefly excludes is disputed:
BM: only the tudādi root, since the luk-vikaraṇa paribhāṣā already excludes kauti.
TB: kauti, and kuvati along with it.
The English names neither purpose.

G. The augments nīk and rīk to the abhyāsa — running example: `vañc + yaṅ → va-vañc-ya → vanī-vañc-ya (nīk; k-marked, so no lengthening) → vanī-vac-ya (n deleted) → vanīvacyate`; `vṛt → va-vṛt-ya → varī-vṛt-ya → varīvṛtyate` (2644 joins 2642)

2642 nīg vañcusraṃsudhvaṃsubhraṃsukasapatapadaskandām 7.4.84 — the augment nīk comes to the abhyāsa of vañc, sraṃs, dhvaṃs, bhraṃs, kas, pat, pad and skand when yaṅ follows or there is luk of yaṅ. abhyāsasya comes from SK 2620 (BM). yaṅlukoḥ is from SK 2630, a translator's inference (audit).
nīk is k-marked, so dīrgho 'kitaḥ (2632) gives no lengthening. The k-marking blocks it even before the augment comes (BM). The root's n is deleted by aniditāṃ hala upadhāyāḥ kṅiti (6.4.24, SK 415): vanīvacyate, sanīsrasyate.
"And so on" (BM, TB): danīdhvasyate, banībhrasyate, canīkasyate, panīpatyate, panīpadyate, canīskadyate. That is TB's spelling; WS BM has canīskandyate.
TB's yaṅluk forms: vanīvañcīti, sanīsraṃsīti.
2644 rīg ṛdupadhasya ca 7.4.90 — the augment rīk comes to the abhyāsa of a dhātu whose penultimate is short ṛ, when yaṅ follows or there is luk of yaṅ. abhyāsasya and yaṅlukoḥ come from SK 2620 and SK 2630 (BM). "Short" follows TB: ṛt kim? cekīrtyate (kṝt).
vṛt + yaṅ → doubled → urat (7.4.66, SK 2244) gives a for the abhyāsa's ṛ → halādiḥ śeṣaḥ → rīk: varīvṛtyate (BM).
narīnṛtyate: nṛt is in the kṣubhnādi list (kṣubhnādiṣu ca, SK 792), so there is no ṇ.
jarīgṛhyate: grah takes saṃprasāraṇa by grahijyā… (SK 2412) first (BM).
latva in both places by kṛpo ro laḥ (SK 2350): the r of rīk becomes l, and the ṛ of the root becomes ḷ: calīkḷpyate.
vārttika (sk.2644.k2) rīg ṛtvata iti vaktavyam: rīk for a dhātu containing ṛ, not only one with ṛ as its penultimate. ṛtvat is bha, so there is no jaś (BM).
vraśc → saṃprasāraṇa → varīvṛścyate.
pracch → parīpṛcchyate.

← reading all prakaraṇas →

तिङन्तयङ्प्रकरणम्

tiṅantayaṅprakaraṇam — outline (p56, SK 2629–2649, all 21 sūtras)

A. When yaṅ comes: repetition or intensity, and the two "only" senses — running example: bhū + yaṅ (kriyāsamabhihāra) → bobhūyate; vraj + yaṅ (crookedness only) → vāvrajyate; lup + yaṅ (censure only) → lolupyate (2634–2635 pulled forward)

C. The root's vowel before yaṅ: guṇa for ṛ, ī for ghrā and dhmā, ay for śī — running example: ṛ + yaṅ → ar-ya (guṇa, 2633) → 'rya' doubled → a-ra-rya → (lengthening, 2632) arāryate (2648–2649 pulled forward)

D. nuk (the anusvāra) after the a of the abhyāsa; u for the a that follows — running example: car + yaṅ → ca-car-ya → caṃ-car-ya (nuk, 2636) → caṃ-cur-ya (u, 2637) → caṃ-cūr-ya (hali ca) → cañcūryate / caṃcūryate (2643 moved here, after 2638)

E. Root substitutes before yaṅ: l for r, saṃprasāraṇa (but not for vaś), kī for cāy — running example: svap + yaṅ → sup-ya (saṃprasāraṇa, pūrvarūpa) → su-sup-ya → so-ṣup-ya (guṇa of the abhyāsa; ṣ) → soṣupyate (2645–2647 pulled forward)