तिङन्तपरस्मैपदप्रकरणम्

Covers SK 2745–2755 (11 sūtras): all present. Every unit has its own kaumudī. The spine is an unbroken Aṣṭādhyāyī run, 1.3.79–1.3.89, the last eleven pada-restriction rules of the first adhyāya. Every AS-ref below is checked against references/ashtadhyayi/sutraani-data.json, and all 11 match our source.json.

• What the prakaraṇa does. Every rule here is a restriction to parasmaipada. Each overrides a rule that would give ātmanepada: svaritañitaḥ kartrabhiprāye kriyāphale (1.3.72, SK 2158), anudāttaṅita ātmanepadam (1.3.12, SK 2157), gandhanāvakṣepaṇa… (SK 2705) or ṇicaś ca (1.3.74, SK 2564). The last sūtra (2755) withdraws some of these parasmaipadas.
• What is supplied. "parasmaipadam" is supplied in every sūtra (BM: "parasmaipadam iti śeṣaḥ"). The translator names śeṣāt kartari parasmaipadam (1.3.78, SK 2159) as its source; BM names no rule (audit MINOR). "kartari" is carried forward (BM: "kartarīty asyānuvṛttiḥ"), as the kaumudī's "kartari ity eva" at 2745 shows.

tiṅantaparasmaipadaprakaraṇam — outline (p62, SK 2745–2755, all 11 sūtras)

p61 (tiṅantātmanepadaprakaraṇam) closed at 2744 with vibhāṣopapadena pratīyamāne (1.3.77). Its colophon "ity ātmanepadaprakriyā" is supplied in brackets there.

• Where the prakaraṇa begins. The print heads it "tiṅante parasmaipadaprakriyā 59" (NSP p. 462, scan page-0472). In kaumudī type it opens with "śeṣāt kartari parasmaipadam | 2159 | atti |". Ours lacks that line, and the English supplies it in brackets at the head of sk.2745.k1. The scan reads atti; collation batch04's "asti" is a misreading. The name tiṅantaparasmaipadaprakaraṇam and the number 62 are ours.
• Where it ends. 2755 closes with "samāptā padavyavasthā" ("the arrangement of the padas is complete"). The print then has "iti parasmaipadaprakriyā" (NSP p. 463, scan page-0473). Ours lacks it, so the English supplies it in brackets. p63 (bhāvakarmatiṅprakaraṇam) opens with sārvadhātuke yak (3.1.67, SK 2756).
• Contents in outline.
• 2745–2748: upasarga + root, where the fruit goes to the agent. anukaroti, abhikṣipati, pravahati, parimṛṣyati.
• 2749–2751: ram with an upasarga, against its anudātta-it ātmanepada. viramati; transitive uparamati; intransitive uparamati / uparamate.
• 2752–2754: ṇic-final roots, against ṇicaś ca. bodhayati padmam; nigārayati, calayati (but ādayate for ad); śāyayati.
• 2755: the counter-rule. For pā, dam, āṅ-yam, āṅ-yas, pari-muh, ruc, nṛt, vad and vas (and, by vārttika, dheṭ), the ṇic-final takes ātmanepada again: pāyayate, dhāpayete. With akartrabhiprāya, śeṣāt still gives vatsān pāyayati payaḥ. The four optional-pada rules are then recalled, and the section is closed.

Threads that run through the prakaraṇa:

• "kartari ity eva" and the karmakartṛ (2745). The parasmaipada is only for the agent-lakāra, so in bhāva and karman ātmanepada stands (anukriyate). With an object-agent (karmakartṛ), karmavat karmaṇā tulyakriyaḥ (3.1.87, SK 2766) gives ātmanepada. There are two ways to explain this:
• Kāryātideśa (the principal view). That rule is later and blocks this one (TB, BM).
• Śāstrātideśa. What is extended is bhāvakarmaṇoḥ (SK 2679). That rule is earlier and cannot block this one, so "kartṛ" has to be read twice: "an agent that is only an agent". The kaumudī takes the two mentions from kartari śap (SK 2167) and SK 2159. BM names kartari karmavyatihāre (SK 2680) instead of SK 2167. This is recorded; the English keeps the kaumudī.
• ṇeḥ = the hetumat ṇic. It runs from 2752 to 2753, 2754 and 2755. TB holds that the ṇi carried forward from the budhādi sūtra is the hetumat ṇic, and that "aṇau" in 2754 refers to it too. TB rejects as un-Bhāṣya the "kecit" view that "aṇau" excludes curādi ṇic-finals (audit MINOR on 2754; a fixer may write "[causative (hetumat)] ṇi").
• "anantarasya" (2753, 2755). A prohibition touches only the injunction just before it (BM, TB). So the ad vārttika (2753) and na pādamy… (2755) cancel only the parasmaipada of 2753/2754. When the fruit is not meant for the agent (akartrabhiprāya), parasmaipada still comes by śeṣāt kartari (SK 2159): ādayaty annaṃ baṭunā, vatsān pāyayati payaḥ. BM (Wikisource; lost in ours) extends the same limit to the dheṭ upasaṃkhyāna.
• Where the commentary lies. Only the kaumudī blocks are translated.
• TB on the opening line, on 2745 ("asya bādhāt", śāstrātideśa) and on 2746 ("svarited iti") is at the tail of sk.2744.t1, after p61's colophon.
• BM on 2749 is at the end of sk.2748.b1.
• TB on 2753, 2754 and 2755 (with "iti padavyavasthā samāptā | iti parasmaipadaprakriyā") is inside sk.2753.t1. That block then runs on into TB on SK 2756–2758 (p63).
• No TB on 2747, 2749 or 2751 (print).
• 2751. Ours has no BM, but Wikisource BM (uttara-2/0210) has it: "upād iti rama iti parasmaipadam iti cānuvartate" (audit MINOR).
• BM in ours drops lines against Wikisource:
• 2752: the glosses of drāvayati and srāvayati. The "slip" the unit note attributes to BM is an OCR drop (audit MINOR).
• 2753: BM's point that drinking counts as nigaraṇa, which is why pā is named at 2755.
• 2755: the dheṭ clause.

The p62–66 audit (.work/sk-run/p62-66-audit/a-sk2745-2755.md) found no MUST-FIX here. A fixer is applying the MINOR items in parallel. This outline already follows the audit:

• 2752: the TB's "calanārthānām anenātmanepade siddhe 'pi" is a print slip for "parasmaipade" (ours = print, scan 0473). The English carries the corrected sense.
• 2753: "devadattena, a third-case word" is the translator's gloss, not Bhaṭṭoji's.
• 2755 k1: "the others have a sentient agent and are intransitive" holds when not in ṇi (aṇau) (BM). That is the condition of 2754.
• 2755 k3: BM cites the example as bhikṣā vāsayati, with bhikṣā in the first case ("alms make [one] stay"). The kaumudī (ours = print = Wikisource) has bhikṣāṃ.

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

• the opening line with atti;
• 2745: anuvartya (ours "uvartya");
• 2755:
• vartsyati (ours "vartsayati");
• luṭi ca kḷpaḥ = SK 2351 (ours cross-refers SK 2341);
• the colophon.
• 2753: only ādikhādyor na (the vārttika of SK 540) is a vārttika. "neti pratiṣiddham" and k5 are Bhaṭṭoji's prose. Ours tags all of k4, and the "(vārtikam)" tags at k3/k5 and 2755 k3 are editorial.

Still open for Ajai (mūla): every item under Print readings stands in ours unedited (source.json has not been changed).

Ordering: strict SK order.

A. Parasmaipada after upasarga + root, even when the fruit goes to the agent — running example: anu + kṛñ + laṭ → (ñit, SK 2158 would give anukurute) → anukaroti

• 2745 anuparābhyāṃ kṛñaḥ 1.3.79 — parasmaipada after kṛñ preceded by anu or parā.
• kṛñ is ñit, so svaritañitaḥ… (SK 2158) would give ātmanepada whenever the fruit goes to the agent (kartṛga phala). This rule gives parasmaipada even then. It also overrides the ātmanepada of gandhanāvakṣepaṇa… (SK 2705) in those senses, which BM lists: gandhana, avakṣepaṇa, sevana, sāhasikya, pratiyatna, prakathana, upayoga.
• Forms: anukaroti ("imitates"), parākaroti ("rejects").
• kartari ity eva: no parasmaipada for an l in bhāva or karman.
• Karmakartṛ. In BM's anukriyate śabdaḥ svayam eva, 2766 (later, kāryātideśa) gives ātmanepada. On the śāstrātideśa view, "kartṛ" is read twice. See the threads above.
• 2746 abhipratyatibhyaḥ kṣipaḥ 1.3.80 — parasmaipada after kṣip ("to throw") preceded by abhi, prati or ati.
• kṣip is svaritet, so SK 2158 would give ātmanepada for a fruit going to the agent. This rule gives parasmaipada even there (BM, TB).
• Forms: abhikṣipati ("throws at"); BM adds pratikṣipati and atikṣipati.
• 2747 prād vahaḥ 1.3.81 — parasmaipada after vah preceded by pra. vah is svaritet, so the reason is the same (BM).
• Form: pravahati ("flows, carries forward").
• No TB (print).
• 2748 parer mṛṣaḥ 1.3.82 — parasmaipada after mṛṣ preceded by pari.
• The divādi mṛṣ (titikṣā) is svaritet and would take both padas, so this rule gives parimṛṣyati ("endures").
• The bhvādi mṛṣ is parasmaipadin anyway: parimarṣati. TB instead derives parimarṣati from the curādi mṛṣ without ṇic (ā dhṛṣād vā); BM covers both. The English follows the kaumudī's bhvādi.
• Yogavibhāga (kecit). "pareḥ" is split off as its own rule, with "vahaḥ" carried forward into it: parivahati. BM: this split is not in the Bhāṣya.

B. Parasmaipada for ram with an upasarga, against its anudātta-it ātmanepada — running example: vi + ram + laṭ → (anudāttet, SK 2157 would give viramate) → viramati

• 2749 vyāṅparibhyo ramaḥ 1.3.83 — parasmaipada after ram preceded by vi, āṅ or pari.
• ram is anudāttet, so anudāttaṅita… (SK 2157) would make it ātmanepadin (BM, filed under sk.2748.b1).
• Forms: viramati ("stops"); BM adds āramati and pariramati.
• 2750 upāc ca 1.3.84 — parasmaipada after ram preceded by upa too. "ramaḥ" is carried forward from 2749.
• The example is transitive: yajñadattam uparamati = uparamayati ("makes Yajñadatta stop"). ram here has the ṇic-sense incorporated (antarbhāvitaṇyartha); roots have many senses (BM).
• Why not one rule, vyāṅparyupebhyaḥ? So that only upa carries into 2751 (BM, TB).
• Intransitive uparamati is not the example here, because 2751 makes it optional (BM).
• 2751 vibhāṣā 'karmakāt 1.3.85 — after upa + ram used intransitively, parasmaipada is optional. upāt and ramaḥ are carried forward. BM (Wikisource) also carries parasmaipadam forward.
• Forms: uparamati / uparamate = nivartate ("he desists").

C. Parasmaipada for ṇic-final roots, against ṇicaś ca — running example: budh + (hetumat) ṇic → bodhi- + laṭ → (ṇicaś ca, SK 2564 would give bodhayate) → bodhayati padmam

• 2752 budhayudhanaśajaneṅprudrusrubhyo ṇeḥ 1.3.86 — after budh, yudh, naś, jan, iṅ, pru, dru and sru ending in ṇi, parasmaipada comes. It is an apavāda of ṇicaś ca (1.3.74, SK 2564).
• bodhayati padmam ("[the sun] makes the lotus open"): the non-ṇi agent, the lotus, is not sentient, so 2754 would not reach it. Likewise yodhayati kāṣṭhāni, nāśayati duḥkham and janayati sukham (BM).
• adhyāpayati vedam (iṅ + ṇic).
• prāvayati = prāpayati, drāvayati = vilāpayati, srāvayati = syandayati. pru, dru and sru are named here for senses other than motion, since motion is covered by 2753 (TB). Hence Bhaṭṭoji's glosses.
• BM: the ātmanepada in "prītiṃ bhaktajanasya yo janayate" is an error, or else the ṇic-final of a ṇic-final.
• 2753 nigaraṇacalanārthebhyaś ca 1.3.87 — parasmaipada also after ṇi-final roots meaning swallowing (nigaraṇa = bhakṣaṇa) or moving (calana = kampana). ṇeḥ is carried forward from 2752.
• Forms: nigārayati, āśayati, bhojayati, calayati, kampayati.
• vārttika (k2) adeḥ pratiṣedhaḥ: not for ad. So ādayate devadattena ("makes Devadatta eat"). devadattena is in the third case because the karma-designation of the instigated agent by gatibuddhi… (1.4.52, SK 540) is barred by that rule's vārttika ādikhādyor na (k4).
• k5. The prohibition cancels only what this sūtra gives (anantarasya, BM). When the fruit is not intended for the agent, śeṣāt (SK 2159) still gives ādayaty annaṃ baṭunā.
• 2754 aṇāv akarmakāc cittavatkartṛkāt 1.3.88 — parasmaipada after the ṇi-final of a root that, when not in ṇi, is intransitive and has a sentient agent.
• śete kṛṣṇaḥ → taṃ gopī śāyayati ("the cowherdess makes him sleep"). śī is intransitive and Kṛṣṇa is sentient, so the result is parasmaipada, not ṇicaś ca's śāyayate.
• Counter-examples (BM, TB):
• the agent is not sentient: vrīhayaḥ śuṣyanti → śoṣayate;
• the root is transitive: kaṭaṃ karoti → kaṭaṃ kārayati, which this rule does not restrict, so ṇicaś ca still governs it.
• "aṇau". It bars a second ṇic on an already hetumat-ṇic-final intransitive: ārohayate (TB).

D. The counter-rule: ātmanepada restored for nine ṇic-final roots — running example: pā ("drink" = nigaraṇa, so 2753 would give pāyayati) + ṇic → pā-y-i (yuk) → pāyayate

• 2755 na pādamyāṅyamāṅyasaparimuharucinṛtivadavasaḥ 1.3.89 — no parasmaipada after the ṇi-finals of pā, dam, āṅ-yam, āṅ-yas, pari-muh, ruc, nṛt, vad and vas.
• Why the rule is needed. A prohibition needs something to prohibit:
• pā means swallowing, so 2753 would give parasmaipada;
• the others are intransitive with a sentient agent when not in ṇi, so 2754 would (audit: supply "[aṇau]");
• nṛt also means moving, so it falls under both rules.
• Forms:
• pāyayate: yuk by śācchāsāhvāvyāvepāṃ yuk;
• damayate: shortening by mittva;
• āyāmayate: no shortening, mittva being denied by yamo 'pariveṣaṇe;
• also āyāsayate, parimohayate, rocayate, nartayate, vādayate;
• vāsayate: from bhvādi vasa nivāse. The adādi vasa ācchādane is transitive and never reached here (BM, TB).
• dam was called transitive under divādi. This very prohibition shows it is also intransitive (BM). TB: dam is intransitive when its object is intended as agent.
• vārttika (k2) dheṭa upasaṃkhyānam: dheṭ is added to the prohibition. dhāpayete śiśum ekaṃ samīcī ("the two, united, suckle one child"). It is a mantra; samīcī is a first-case dual (TB).
• k3, akartrabhiprāya. Here śeṣāt (SK 2159) still gives parasmaipada (anantarasya): vatsān pāyayati payaḥ, damayantī kamanīyatāmadam, bhikṣāṃ vāsayati (BM reads bhikṣā; see audit above).
• k3, resumption. In sūtra order, Bhaṭṭoji recalls the four rules that make the pada optional:
• vā kyaṣaḥ (SK 2669): lohitāyati / lohitāyate;
• dyudbhyo luṅi (SK 2345): adyutat / adyotiṣṭa;
• vṛdbhyaḥ syasanoḥ (SK 2347): vartsyati / vartiṣyate, vivṛtsati / vivartiṣate;
• luṭi ca kḷpaḥ (SK 2351): kalptā, kalptāsi / kalpitāse, kalptāse, kalpsyati / kalpiṣyate / kalpsyate, cikḷpsati / cikalpiṣate / cikḷpsate.

Then "samāptā padavyavasthā" and [iti parasmaipadaprakriyā].