तद्धिकाधिकारप्रकरणे अपत्यादिविकारान्तार्थसाधारणप्रत्ययाः

Covers SK 1072–1078 (7 sūtras), Aṣṭādhyāyī 4.1.82–86 (the opening of the taddhita section: the three adhikāras and the affixes common to all senses from apatya to vikāra) with the two ādivṛddhi rules 7.2.117–118 spliced in where the SK reads them. All numbers present.

taddhitādhikāraprakaraṇe apatyādivikārāntārthasādhāraṇapratyayāḥ — outline (p26, SK 1072–1078, all 7 sūtras)

The taddhita section opens. Having finished the compounds (samāsa), Bhaṭṭoji sets up the three headings under which every taddhita rule as far as prāgdiśo vibhaktiḥ (5.3.1, SK 1947) is read — after the first of the samartha words, optionally (1072) — and the default affix aṇ that stands in for any sense-rule that names no affix of its own, as far as tena dīvyati (4.4.2, SK 1550) (1073). The rest of the prakaraṇa is the small set of affixes common (sādhāraṇa) to all the prāgdīvyatīya senses — apatya, samūha, jāta, devatā … vikāra — taught for particular bases: aṇ kept for the aśvapati class (1074), ṇya for diti/aditi/āditya and pati-final words (1077), añ for the utsa class (1078), plus a cluster of vārttikas assigning special affixes to yama, pṛthivī, deva, bahis, sthāman, loman, go, agni and kali. Because every one of these forms needs vṛddhi of its first vowel, the SK reads the two ādivṛddhi rules of 7.2 here (1075–1076). Only the kaumudī blocks are translated; Bālamanoramā (BM) and Tattvabodhinī (TB) supply the derivations cited. Ordering: SK order throughout; 1077's nine vārttikas are split off from the sūtra proper into their own group D so that each special base gets a line.

A. The three headings — after which word, optionally, and which affix by default — running example: aśvapater apatyam → āśvapatam (the vākya survives beside the derivative)

• 1072 samarthānāṃ prathamād vā 4.1.82 — an adhikāra, not a vidhi (no vidheya appears — BM): three words are held over as far as prāgdiśo vibhaktiḥ (5.3.1, SK 1947). samarthānām — nirdhāraṇa genitive, "among the samartha words"; samartha = pariniṣṭhita, a word on which the sandhi operations have already been performed (so not the ekārthībhāva of samarthaḥ padavidhiḥ 2.1.1, which the paribhāṣā would give for nothing); prathamāt — the word named first in each sense-rule (the tasya of tasyāpatyam); vā — optionally, so the analytic phrase remains correct. On the running form: in aśvapateḥ apatyam the sandhi-finished pada aśvapateḥ is the samartha, it is the tasya (first-named) of tasyāpatyam, so the affix goes after it; by vā the phrase aśvapater apatyam is also good.
• 1073 prāgdīvyato 'ṇ 4.1.83 — an adhikāra: aṇ governs up to but not including tena dīvyati [khanati jayati jitam] (4.4.2, SK 1550) — 'dīvyat' being an imitation of that sūtra's word (BM; TB records the alternative that it is a mere boundary-fragment, as īśvara in prāg rīśvarāt 1.4.56). In sense-rules that name no affix (tasyāpatyam etc.), aṇ answers "which affix?" and 1072 answers "after what?"; where a specific affix is stated (iñ etc.), aṇ, the utsarga, yields. On the running form: aśvapati + aṇ (drawn in here by 1074's 'ca') → āśvapatam.
• 1074 aśvapatyādibhyaś ca 4.1.84 — after aśvapati and the rest (TB's list: aśvapati, gaṇapati, rāṣṭrapati, kulapati, gṛhapati, paśupati, kṣetrapati) aṇ occurs in the prāgdīvyatīya senses — an apavāda of the ṇya that 1077 will teach after every pati-final word; without this rule aśvapati would take ṇya. (prāgdīvyatīya = "belonging to [the section] before dīvyati": avyayībhāva prāgdīvyat by 2.1.12, then cha by vṛddhāc chaḥ 4.2.114 — BM/TB.) On the running form: aśvapati + aṇ → the form is finished by group B: āśvapatam.

B. The vṛddhi every taddhita form needs — ādivṛddhi before ñit/ṇit and kit affixes — running example: aśvapati + aṇ → āśvapatam (kit case: bahis + īkak → bāhīkaḥ, agni + ḍhak → āgneyam)

• 1075 taddhiteṣv acām ādeḥ 7.2.117 — when a taddhita marked with ñ or ṇ follows ('aco ñṇiti' 7.2.115 and 'vṛddhiḥ' from mṛjer vṛddhiḥ 7.2.114 carried — BM), vṛddhi replaces the first of the vowels of the base ('acām' nirdhāraṇa genitive). Because the sthānin is named expressly, iko guṇavṛddhī (1.1.3) does not confine it to ik (TB). On the running form: aṇ is ṇit; the first ac of aśvapati is a → ā: āśvapati + a; then the final i drops by yasyeti ca (6.4.148) → āśvapata- → āśvapatam (the neuter covers apatya and the rest — samūha etc.).
• 1076 kiti ca 7.2.118 — and when a taddhita marked with k follows, likewise ādivṛddhi. Read here, though off-topic, for convenience (BM) and for bāhīkaḥ below (TB). The SK's own examples close the aṇ derivation: aśvapater apatyādi → āśvapatam; gaṇapati (in the aśvapatyādi list) → gāṇapatam; hence gāṇapatyo mantraḥ (with ṇya) is "simply an error" (TB defends it as ṣyañ on the devatā-word gāṇapata, cāturvarṇyādi being an ākṛtigaṇa — not harmonized). The kit case proper: bahis + īkak → bāhīkaḥ (D below); agni + ḍhak → āgneyam (E below) — a → ā by this rule.

C. ṇya after diti, aditi, āditya, pati-final words (and yama) — running example: diter apatyam → daityaḥ

• 1077 dityadityādityapatyuttarapadāṇ ṇyaḥ 4.1.85 — after diti, aditi, āditya, and after a word whose uttarapada is pati, ṇya in the prāgdīvyatīya senses ('prāgdīvyataḥ' carried — BM) — an apavāda of aṇ (1073). On the running form: diti + ṇya, ṇ an it by cuṭū; ādivṛddhi by 1075 (ṇit) on the first ac i → ai; final i of diti dropped by yasyeti ca (6.4.148): daitya- → daityaḥ (BM: daiteyāḥ is wrong, defensible only as pṛṣodarādi). aditi + ṇya → ādityaḥ ("of aditi"); or āditya + ṇya → ādityaḥ ("of āditya", in jāta etc.): ādivṛddhi is vacuous, a-lopa by yasyeti ca, then optionally doubling of y (vārttika on anaci ca 8.4.47) and y-lopa by halo yamāṃ yami lopaḥ (8.4.64) give a one-y or two-y form (the y-lopa of 6.4.151 is barred, the inner ṇya not being an apatya affix — BM). prajāpati + ṇya (pati-final) → prājāpatyaḥ.
• vārttika yamāc ca (k2–k3) — "and after yama, ṇya" — read so in the Kāśikā (i.e. not in the Bhāṣya — BM/TB): yama + ṇya → yāmyaḥ (TB: also derivable as svārthika ṣyañ on yāma).

D. Special affixes for special bases — the vārttika cluster under 1077 (pṛthivī, deva, bahis, sthāman, loman, go) — running example: bahir bhavaḥ → bāhyaḥ / bāhīkaḥ

• 1077 (cont.) vārttika pṛthivyā ñāñau (k4–k5) — after pṛthivī, ñ and añ: pṛthivī + ñ → ādivṛddhi ṛ → ār, ī dropped by yasyeti ca, ṭāp → pārthivā; with añ → ṅīp by ṭiḍḍhāṇañ… (4.1.15) → pārthivī. Both affixes are taught precisely so that both feminines come out (BM/TB).
• 1077 (cont.) vārttika devād yañañau (k6–k7) — after deva, yañ and añ: deva + yañ → ādivṛddhi e → ai, a-lopa → daivyam; with añ → daivam.
• 1077 (cont.) vārttika bahiṣaṣ ṭilopo yañ ca (k8–k9) — after the avyaya bahis, yañ with ṭilopa of the base: on the running form bahis + yañ → ṭi (-is) dropped, ādivṛddhi a → ā → bāhyaḥ ("existing outside" etc.). The express ṭilopa is a jñāpaka that avyayānāṃ bhamātre ṭilopaḥ is anitya (whence ārād bhavaḥ → ārātīyaḥ — BM/TB).
• 1077 (cont.) vārttika īkak ca (k10–k11) — and īkak after bahis, with ṭilopa: bahis + īkak → ṭilopa, ādivṛddhi by 1076 kiti ca (īkak is kit) → bāhīkaḥ.
• 1077 (cont.) vārttika sthāmno 'kāraḥ (k12–k13) — after sthāman, the affix a: aśvatthāman (bahuvrīhi "whose sthāman (stance) is like a horse's" — tadanta-vidhi on the Bhāṣya's example) + a → ṭilopa of -an by nas taddhite (6.4.144), no ādivṛddhi (a is not ṇit — with aṇ there would have been) → aśvatthāmaḥ; s → t by pṛṣodarādīni yathopadiṣṭam (6.3.109).
• 1077 (cont.) vārttika bhavārthe tu lug vaktavyaḥ (k14–k15) — but in the sense bhava ("existing in"), luk of that affix a: aśvatthāman + a → luk → aśvatthāmā.
• 1077 (cont.) vārttika lomno 'patyeṣu bahuṣv akāraḥ (k16–k17) — after loman, when the offspring are many, the affix a — an apavāda of the iñ that bāhvādi membership would give (bāhvādibhyaś ca 4.1.96): uḍuloman ("whose hairs are like stars", tadanta-vidhi) + a → ṭilopa by nas taddhite → uḍulomāḥ, acc. uḍulomān. Why "many"? one offspring: uḍuloman + iñ → auḍulomiḥ (ādivṛddhi, ṭilopa).
• 1077 (cont.) vārttika gor ajādiprasaṅge yat (k18–k19) — after go, where an ac-initial affix would apply, yat instead (an apavāda of aṇ; BM/TB read 'sarvatra' — in all prāgdīvyatīya senses, not apatya only): go + yat → gavyam (gavi bhavam, gor āgatam, gaur devatā asya …). Why "ac-initial"? hal-initial affixes are untouched: go + rūpya (hetumanuṣyebhyo 'nyatarasyāṃ rūpyaḥ 4.3.81) → gorūpyam "come from cows as the cause"; go + mayaṭ (mayaṭ ca 4.3.82) → gomayam.

E. añ after the utsa class; ḍhak after agni and kali — running example: utsa + añ → autsaḥ (and agner apatyādi → āgneyam)

• 1078 utsādibhyo 'ñ 4.1.86 — after utsa and the rest, añ in the prāgdīvyatīya senses (supplied — BM) — an apavāda of aṇ and of aṇ's own apavādas iñ etc. (TB). On the running form: utsa + añ → ñit, so ādivṛddhi by 1075 u → au, a-lopa by yasyeti ca → autsaḥ. (TB adds the gaṇa-sūtra baṣkayāse "not in compound": bāṣkayaḥ but gaurbaṣkayiḥ; and ādhenavam = dhenūnāṃ samūhaḥ by tadanta-vidhi.)
• vārttika agnikalibhyāṃ ḍhak vaktavyaḥ (k2–k3) — after agni and kali, ḍhak — read here (dropping the Bhāṣya's 'tatra sarvatra' from the dṛṣṭaṃ sāma 4.2.7 vārttika) so that one statement covers all the prāgdīvyatīya senses, superseding agner ḍhak (4.2.33) and the kaler-ḍhak vārttika (BM/TB): agni + ḍhak → ḍh → eya by āyaneyīnīyiyaḥ… (7.1.2), ādivṛddhi by 1076 kiti ca (ḍhak is kit) a → ā, i-lopa by yasyeti ca → āgneyam (agner apatyam, agninā dṛṣṭaṃ sāma, agner idam, agnau bhavam …); kali + ḍhak → kāleyam.