तिङन्तरुधादिप्रकरणम्

Covers SK 2543–2546 (4 sūtras): all present. Every unit has its own kaumudī. The spine is not an Aṣṭādhyāyī run. The SK follows the Dhātupāṭha order of the rudhādi roots (1438 rudhir … 1462 pṛcī, our numbering) and teaches each rule at the root where it first bites: 3.1.78 → 6.4.23 → 7.3.92 → 7.2.71. Every AS-ref below is checked against references/ashtadhyayi/sutraani-data.json.

• Rules the prakaraṇa leans on but does not teach. The class-mark śnam needs three rules taught earlier: mid aco 'ntyāt paraḥ (1.1.47, SK 37) for its place, śnasor allopaḥ (6.4.111, SK 2469) for the loss of its a, and aniditām hala upadhāyāḥ kṅiti (6.4.24, SK 415) beside 2544. tṛṇaha im (2545) carries hali from uto vṛddhir luki hali (7.3.89, SK 2443) and piti from nābhyastasyāci piti sārvadhātuke (7.3.87, SK 2503). añjeḥ sici (2546) carries iṭ from iḍ atty artivyayatīnām (7.2.66, SK 2384).

tiṅantarudhādiprakaraṇam — outline (p50, SK 2543–2546, all 4 sūtras)

p49 (tudādi) closed at 2542 with "vṛt | mucādayo vṛttāḥ tudādayaś ca". Its last root is 1437 piśa.

• Where the prakaraṇa begins. sk2543 opens with the Dhātupāṭha entry 1438 rudhir āvaraṇe and the notice "nava svariteta iritaś ca": the nine roots from rudhir are ubhayapadin and ir-marked. The sūtra rudhādibhyaḥ śnam (3.1.78) follows.
• Where it ends. At 2546 with 1462 pṛcī saṃparke: pṛṇakti, paparca. The print closes "iti rudhādayaḥ" ("Here end the rudhādi roots"). Ours lacks it, so the English supplies it in brackets (scan page-0424). p51 opens with "atha sapta svaritetaḥ" and 1463 tanu (SK 2547).
• Contents in outline.
• 2543: rudh and the ir-marked roots. śnam for śap, placed after the last vowel and blocking guṇa: ruṇaddhi, rundhaḥ, arudhat / arautsīt; bhinatti, riṇakti, kṣuṇatti, chṛṇatti.
• 2544: indh, śiṣ, bhañj, bhuj. The root's own n is lost after śnam: indhe, inadhai, śinaṣṭi, śiṇḍhi, bhanakti, bhunakti.
• 2545: tṛh, hiṃs, und, añj. im for tṛṇah before a hal-initial pit ending: tṛṇeḍhi, tṛṇḍhaḥ; hinasti, unatti, anakti, aṅgdhi.
• 2546: añj in luṅ; tañc, vij, vṛj, pṛc. Obligatory iṭ for sic: āñjīt; tanakti, vinakti, vivijitha, vṛṇakti, pṛṇakti.

Threads that run through the prakaraṇa:

• śnam is the class-mark. It replaces śap (an apavāda). Being mit, it stands after the root's last vowel (mid aco 'ntyāt paraḥ), so the root is split: ru-na-dh, bhi-na-d, yu-na-j. Being nitya, it comes before the light-penultimate guṇa and blocks it: ruṇaddhi, not *roṇaddhi.
• The a of śnam depends on the ending. Before a kṅit sārvadhātuka it is lost (śnasor allopaḥ, SK 2469): rundhaḥ, bhinte, śiṃṣṭaḥ, tṛṇḍhaḥ. Before the pit endings, including those with āṭ, it stays: ruṇaddhi, ruṇadhāni, inadhai, śinaṣāṇi, anajāni.
• The n left standing. Once the a is lost, the n of śnam takes anusvāra (naś cāpadāntasya jhali, 8.3.24, SK 123) and then parasavarṇa (anusvārasya yayi parasavarṇaḥ, 8.4.58, SK 124): run-dh, bhin-t, aṅ-k, añ-j. ṇatva is asiddha to anusvāra, so anusvāra and then parasavarṇa come first. The n that parasavarṇa produces is asiddha to ṇatva, so rundhaḥ keeps its n. By na padānta… (1.1.58, SK 51) the a-loss is not sthānivat for anusvāra and parasavarṇa.
• Nasal-final roots. When the root already has a nasal (indh, bhañj, hiṃs, añj, tañc), that n is lost after śnam by 2544. So there is only ever one nasal: indhe, bhanakti, hinasti, anakti, tanakti.
• Single or double stop. The kaumudī writes rundhaḥ, rundhe, bhinte, abhinta, kṣunte and chṛnte with one stop. BM writes runddhaḥ, bhintte and so on. The translator takes the single stop as the optional jharo jhari savarṇe (8.4.65, SK 71), which BM invokes under 2544 (śiṇḍhi / śiṇḍḍhi). The English keeps the kaumudī's forms.
• ir-marked roots have two luṅ forms. aṅ is optional (irito vā, 3.1.57, SK 2269). Otherwise sic takes vṛddhi. arudhat / arautsīt, abhidat / abhaitsīt, aricat / araikṣīt, akṣudat / akṣautsīt, acchṛdat / acchardīt. In the ātmanepada, sic is lost by jhalo jhali (8.2.26, SK 2281): aruddha, abhitta, arikta, akṣutta.
• The kaumudī lists forms bare. Lakāra and person labels are the translator's, given in Pāṇinian terms. Derivation stages follow BM and TB as recorded in the unit notes.
• Dhātu numbers. The numbers in our markup run one lower than the print's (print 1439 rudhir … 1463 pṛcī). They are given as transmitted.
• Where the commentary lies. Only the kaumudī blocks are translated (SK-TRANSLATE §25).
• TB on 2543 is printed at the tail of sk.2541.t1, after TB on the tudādi roots. sk2543 has no t1 block.
• TB on 2544 is in sk.2544.t1. TB on 2545, including the glosses on undī and añjū, is in sk.2545.t1 (audit correction: it is not split into sk.2546.t1).
• sk.2546.t1 is a composite. It holds TB on 2546 and "iti rudhādayaḥ", then TB on the tanādi roots (SK 2547–2550, p51), then the head of TB on the kryādi roots. An NSP footnote on SK 2547 has been run into it by the OCR.
• BM: each unit has its own b1 block. BM's rudhādi colophon is at the end of sk.2546.b1.

Where the p47–53 audit (.work/sk-run/p47-53-audit-fixes.md and .work/sk-run/p47-53-audit/sk2543-2553.md) corrects a reading, this outline follows the corrected reading. The fixer is applying these to the en.yml in parallel.

• 2543: BM gives the sense (kartr-arthe sārvadhātuke … svārthe) but names no source rules. That kartari comes from kartari śap (3.1.68, SK 2167) and sārvadhātuke from sārvadhātuke yak (3.1.67, SK 2756) is the translator's identification.
• 2545 (jihiṃsa): jihiṃsa is the ṇal form, which is pit, not kit. BM's remark is general: the root being idit, its n is never lost, even before a kit affix.
• 2545 (TB location): TB on undī (kledana = ārdrībhāva) and añjū (vyakti = vivecana, mrakṣaṇa = snigdhatā) is in sk.2545.t1 itself. Only the print starts its p. 414 TB band mid-note.

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

• dvaidhīkaraṇe for chidir (2543; ours dvaidhikaraṇe);
• the colophon "iti rudhādayaḥ" (2546).

Still open for Ajai (mūla): none in this prakaraṇa.

Ordering: SK order throughout.

A. rudh: śnam for śap, after the last vowel; the a of śnam lost before kṅit — running example: rudh + śap + ti → ru-na-dh + ti → ruṇaddhi; ru-na-dh + tas → run-dh + tas → rundhaḥ

• 2543 rudhādibhyaḥ śnam 3.1.78 — after rudh and the roots that follow it, śnam comes in its own sense when a sārvadhātuka denoting the agent follows. It is an apavāda of śap.
• BM gives the conditions (kartr-arthe sārvadhātuke, svārthe). Tracing them to kartari śap (SK 2167) and sārvadhātuke yak (SK 2756) is the translator's.
• Root entry 1438 rudhir āvaraṇe. Nine roots are svaritet and ir-marked.
• Place and priority. Being mit, śnam comes after the last vowel u: ru-na-dh. Being nitya, it blocks the light-penultimate guṇa. That guṇa is the later rule and would otherwise come first, giving *roṇaddhi; once śnam is in, the penultimate is no longer light (BM).
• ru-na-dh + ti: dh for t by jhaṣas tathor dho 'dhaḥ (8.2.40, SK 2280), and ṇ for n: ruṇaddhi.
• ru-na-dh + tas:
• tas is kṅit, so the a of śnam is lost (śnasor allopaḥ, SK 2469): run-dh + tas.
• ṇatva is asiddha, so anusvāra (SK 123) and parasavarṇa (SK 124) come first. The n they produce is asiddha to ṇatva, so it stays n.
• By na padānta… (SK 51) the a-loss is not sthānivat for anusvāra and parasavarṇa: rundhaḥ (BM: runddhaḥ).
• rundhanti; ātmanepada rundhe; roddhā, rotsyati.
• loṭ ruṇaddhu, rundhāt, rundhi; ruṇadhāni and ruṇadhai (āṭ is pit, so the a stays).
• laṅ aruṇat, arundhām; madhyama singular aruṇat / aruṇaḥ (after the loss of s, ru optional by daś ca, 8.2.75, SK 2468; TB, BM); aruṇadham.
• luṅ arudhat / arautsīt (aṅ optional, the root being irit; otherwise sic with vṛddhi); ātmanepada aruddha (sic lost by jhalo jhali).
• TB (tail of sk.2541.t1): śnam's ś is an it only to specify śnam in śnasor allopaḥ and śnān nalopaḥ. It does not give the sārvadhātuka designation. Nor is its point to bar guṇa, since the part before śnam is not an aṅga.
• 1439 bhidir vidāraṇe: bhinatti, bhinte, bhettā, bhetsyati / bhetsyate; abhinat, abhintām, abhinadam, abhinta; luṅ abhidat / abhaitsīt, abhitta.
• 1440 chidir dvaidhīkaraṇe (print): acchidat / acchaitsīt, acchitta.
• 1441 ricir virecane: riṇakti, riṅkte; rireca, ririce; rektā; ariṇak; aricat / araikṣīt, arikta.
• 1442 vicir pṛthagbhāve: vinakti, viṅkte.
• 1443 kṣudir saṃpeṣaṇe: kṣuṇatti, kṣunte, kṣottā; akṣudat / akṣautsīt, akṣutta.
• 1444 yujir yoge: yoktā.
• 1445 ucchṛdir dīptidevanayoḥ. The u is an it, for optional iṭ before ktvā (TB, BM).
• chṛṇatti, chṛnte; caccharda.
• liṭ ātmanepada cacchṛdiṣe / cacchṛtse: iṭ optional by se 'sici kṛtacṛtacchṛdatṛdanṛtaḥ (7.2.57, SK 2506).
• charditā; chardiṣyati / chartsyati; acchṛdat / acchardīt; acchardiṣṭa.
• 1446 utṛdir hiṃsānādarayoḥ: tṛṇatti and the rest, like chṛṇatti.
• 1447 kṛtī veṣṭane: parasmaipadin. kṛṇatti; its ārdhadhātuka forms are like the tudādi kṛtī (SK 2542).
• 1448 ñi-indhī dīptau: three ātmanepadin roots begin here (indh, khid, vid). Their forms follow under 2544.

B. indh, śiṣ, bhañj, bhuj: the root's own n is lost after śnam — running example: i-na-ndh + te → i-na-dh + te → in-dh + te → indhe; bha-na-ñj + ti → bha-na-j + ti → bhanakti; i-na-ndh + ai → inadhai

• 2544 śnān nalopaḥ 6.4.23 — the n that follows śnam is lost. śnāt is śnam with its m-anubandha dropped. TB gives anakti and bhanakti as the main examples.
• i-na-n-dh + te: this rule removes the root's n, and śnasor allopaḥ removes the a: indhe (BM: inddhe). intse, indhitā.
• TB: the n of śnam that remains takes anusvāra and parasavarṇa. It is not lost by aniditām…, because the a-loss is asiddha by asiddhavad atrābhāt, or else sthānivat.
• inadhai (loṭ, uttama singular) shows why the rule is needed. With āṭ the ending is pit, not ṅit. So the a of śnam stays, and aniditām… (SK 415), which needs a kṅit ending, cannot remove the n. Only this rule can (BM).
• BM: for indhe aniditām… alone would have sufficed, but the sūtra is cited because it is the proper rule, and it is needed for anakti.
• laṅ aindha, aindhāḥ (BM: ainddha, ainddhāḥ).
• 1449 khida dainye: khinte, khettā. 1450 vida vicāraṇe: vinte, vettā.
• The parasmaipadin roots follow. 1451 śiṣḷ viśeṣaṇe, aniṭ:
• śinaṣṭi, śiṃṣṭaḥ (a lost; anusvāra), śiṃṣanti.
• liṭ śiśeṣitha: iṭ is obligatory by the krādi restriction, the root being neither vowel-final nor a-containing (BM).
• śeṣṭā, śekṣyati.
• loṭ madhyama singular śin-ṣ + hi:
• hi → dhi (hujhalbhyo her dhiḥ, 6.4.101, SK 2425);
• jaśtva ṣ → ḍ, then ṣṭutva dh → ḍh;
• the ḍ optionally lost by jharo jhari savarṇe (SK 71);
• anusvāra and parasavarṇa ṇ.
• Result: śiṇḍhi / śiṇḍḍhi.
• BM dissents, following the Śabdenduśekhara: parasavarṇa precedes jaśtva and jharo jhari in the Aṣṭādhyāyī, so they are asiddha to it, and the form with anusvāra alone is proper. The English keeps Bhaṭṭoji's forms.
• śinaṣāṇi (āṭ is pit, so the a stays); aśinaṭ; luṅ aśiṣat (aṅ, the root being ḷdit).
• 1452 piṣḷ saṃcūrṇane: like śiṣ, pinaṣṭi.
• 1453 bhañjo āmardane. TB: the o is for n in the niṣṭhā (bhagnaḥ).
• bhanakti: the root's n lost by this rule.
• babhañjitha / babhaṅktha: iṭ optional by the Bhāradvāja restriction (BM).
• bhaṅktā.
• 1454 bhuja pālanābhyavahārayoḥ: bhunakti, bhoktā, bhokṣyati, abhunak. TB: the ātmanepada (bhuṅkte) comes by bhujo 'navane.
• 1455 tṛha and 1456 hisi hiṃsāyām are named here and conjugated under 2545.

C. tṛh (and hiṃs, und, añj): im for tṛṇah before a hal-initial pit ending — running example: tṛ-na-h + ti → tṛ-na-i-h + ti → tṛṇeh + ti → tṛṇeḍhi; tṛ-na-h + tas → tṛṇḍhaḥ (no im)

• 2545 tṛṇaha im 7.3.92 — to tṛh once śnam is added (tṛṇah), the augment im is added before a hal-initial pit sārvadhātuka.
• tṛṇahaḥ is a sixth case naming the root with śnam made. hali comes from uto vṛddhir luki hali (SK 2443) and piti from nābhyastasyāci piti sārvadhātuke (SK 2503) (BM, TB).
• TB adds that on the Bhāṣya's conclusion under pratyayalope pratyayalakṣaṇam only piti is carried forward. tṛṇahāni is still safe, because "aci na" runs on.
• Why "tṛṇaha", not "tṛhaḥ". The rule names the root with śnam so that im, being specially enjoined, does not block śnam. It is not merely to exclude the tudādi tṛh (TB).
• BM: im would otherwise block śnam by the yena nāprāpti maxim. Blocking holds even where both could apply, as with "brāhmaṇebhyo dadhi dīyatāṃ takraṃ kauṇḍinyāya".
• tṛṇah + ti: im (mit, so after the a of śnam), guṇa a + i → e, then h → ḍh, t → dh, ṣṭutva and loss of the first ḍh: tṛṇeḍhi.
• tṛṇḍhaḥ: tas is not pit, so no im. The a of śnam is lost by SK 2469.
• tatarha, tarhitā; laṅ atṛṇeṭ (tip lost by halṅyādi, then ḍhatva and jaśtva; BM).
• hisi:
• hinasti: the root is idit, so it has num. After śnam, that num is lost by śnān nalopaḥ (2544).
• jihiṃsa (liṭ, ṇal): the root is idit, so its n is never lost, even before a kit affix (BM; audit correction).
• hiṃsitā.
• 1457 undī kledane (kledana = ārdrībhāva, TB):
• unatti, untaḥ, undanti.
• undāñcakāra: ām by ijādeś ca (3.1.36, SK 2237; BM).
• laṅ aunat, auntām, aundan, aunaḥ / aunat, aunadam.
• 1458 añjū vyaktimrakṣaṇakāntigatiṣu (vyakti = vivecana, mrakṣaṇa = snigdhatā, TB):
• a-na-ñj + ti → anakti: the root's n lost by śnān nalopaḥ, j → g by kutva, g → k by cartva. Parasavarṇa is asiddha, so the n is still there to be lost (BM).
• aṅktaḥ: the n and the a are lost, and the n of śnam becomes ṅ by parasavarṇa. añjanti.
• ānañja; ānañjitha / ānaṅktha.
• aṅktā / añjitā: iṭ optional from the ū-marking (svaratisūtisūyatidhūñūdito vā, 7.2.44, SK 2279). The kaumudī does not say this; the translator does.
• aṅgdhi (dhi for hi); anajāni (āṭ is pit, so the a stays); laṅ ānak.
• Its luṅ leads into 2546.

D. añj in luṅ: obligatory iṭ for sic; tañc, vij, vṛj, pṛc — running example: añj + sic + t (luṅ) → āñj + iṭ + sic + t → āñjīt

• 2546 añjeḥ sici 7.2.71 — after añj, iṭ is added to sic obligatorily.
• iṭ comes from iḍ atty artivyayatīnām (SK 2384). The ū-marking would give only an optional iṭ, so this rule exists to make it obligatory (BM). vibhāṣā is not carried forward (TB).
• āñjīt. TB asks why "sici", and answers aṅktā: in luṭ the iṭ stays optional (2545).
• 1459 tañcū saṅkocane: tanakti; taṅktā / tañcitā. BM: the penultimate is n, the root being cited with parasavarṇa made. Its forms are worked like añjū's.
• 1460 ovijī bhayacalanayoḥ. The o is an it. It is seṭ, since only the irit vij is among the aniṭ roots (BM).
• vinakti, viṅktaḥ.
• By vija iṭ (1.2.2, SK 2536) an iṭ-initial affix is ṅit, so there is no guṇa: vivijitha, vijitā.
• avinak; luṅ avijīt.
• 1461 vṛjī varjane: vṛṇakti, varjitā.
• 1462 pṛcī saṃparke: pṛṇakti, paparca. TB: being ī-marked, it takes no iṭ in the niṣṭhā (pṛktaḥ). In apṛkta ekāl… the word is pṛkta with nañ.
• Close: "iti rudhādayaḥ" (Here end the rudhādi roots). It is supplied from the print; BM and TB close alike. p51 opens with the tanādi roots, 1463 tanu (SK 2547).