तद्धिताधिकारे प्राग्दीव्यतीयप्रकरणम्

Covers SK 1514–1547 (34 sūtras): all present. The spine is Aṣṭādhyāyī 4.3.134–168, minus 4.3.150–151 (dvyacaś chandasi, notvadvardhrabilvāt), which the SK reads in the Vaidika section (SK 3453–3454). One rule from another pāda is spliced in where the SK needs it: 7.3.5 (nyagrodhasya ca kevalasya, SK 1543), read after plakṣādibhyo 'ṇ.

taddhitādhikāre prāgdīvyatīyaprakaraṇam — outline (p30, SK 1514–1547, all 34 sūtras)

This prakaraṇa closes the prāgdīvyatīya affixes (the range of prāg dīvyato 'ṇ, 4.1.83, SK 1073) with the vikāra–avayava section. tasya vikāraḥ (1514) sets up the sense "a product (vikāra) of that": what the original material has become in another state. avayave ca… (1515) adds the sense "a part (avayava) of that", but only for words for living beings, herbs and trees. Every later rule works in both senses for those words and in vikāra alone for the rest. TB marks the boundary at 1514: the śaiṣika affixes stop here (samāptāḥ śaiṣikāḥ).

The default affix is the general aṇ. Two rival defaults compete with it:

• añ after u-final and anudāttādi words (1519–1520), and after words for living beings and the rajatādi list (1532).
• mayaṭ, optional after any base (1523) and obligatory after vṛddha words, monosyllables and śarādi (1524).

Most of the section consists of apavādas that set aside one or both of these defaults. The commentators' interest is mostly in accent: whether a base is ādyudātta (so the añ of 1520 does not reach it) or anudāttādi (so it does). The section ends with the fruit rules (1541–1546). There, luk or lup of the vikāra/avayava affix names the fruit, and yuktavadbhāva (1.2.51, SK 1294) decides its gender and number.

Only the kaumudī blocks are translated (SK-TRANSLATE §25). The readings cited as BM (Bālamanoramā) and TB (Tattvabodhinī) are the ones recorded in the unit notes. Where the p30 audit corrects a reading, the corrected reading is followed. This applies to 1514, 1519, 1521, 1522, 1523, 1528, 1533, 1534, 1541, 1542, 1543, 1545, 1546 and 1547. Ordering: SK order throughout; nothing is reordered.

Affix shapes used below (7.1.1–2): vu → aka, ḍh → ey, ṭh → ik. A ñ- or ṇ-marker brings ādivṛddhi (7.2.117, SK 1075). A stem-final a/ā/i/ī is deleted before the affix by yasyeti ca (6.4.148, SK 311). Stems in -an keep their ṭi by an (6.4.167, SK 1155).

A. The two senses: "product of that" and "part of that" — running example: aśmano vikāraḥ → āśmaḥ; mayūrasyāvayavaḥ → māyūraḥ

• 1514 tasya vikāraḥ 4.3.134 — after a sixth-case word, the general affixes aṇ etc., and the special ones still to be taught, in the sense "a vikāra of that". Vārttika aśmano vikāre ṭilopo vaktavyaḥ: before a vikāra affix, aśman loses its ṭi an. This is an apavāda of the prakṛtibhāva of an (BM). aśman + aṇ → the ṭi goes, ā by ādivṛddhi: āśmaḥ. bhasman keeps its ṭi by an: bhāsmanaḥ. mṛttikā gives mārtikaḥ. All three bases are ādyudātta (by the nit accent of manin or tikan), so no añ-apavāda reaches them and they show the plain aṇ (BM). TB (printed under sk.1513.t1, after samāptāḥ śaiṣikāḥ): the śaiṣika affixes do not operate here, so there is no ṭhak of halasīrāṭ ṭhak (SK 1504). Hence hālaḥ and sairaḥ take aṇ in the vikāra sense.
• 1515 avayave ca prāṇyoṣadhivṛkṣebhyaḥ 4.3.135 — after words for living beings, herbs and trees, the same affixes also in the sense "a part of that". The ca keeps vikāra: these words take both senses, all other words vikāra only (TB, BM). mayūra, a living being, takes the añ of 1532: māyūraḥ (part or product of a peacock). mūrvā, a herb, is ādyudātta, so it gets the plain aṇ: maurvam (its stalk or ash). pippala, a tree, also gets aṇ: paippalam. BM and TB differ on the accent of pippala; only TB's ādyudātta reading explains the aṇ.

B. aṇ where añ or mayaṭ would come — running example: bilvasya vikāro 'vayavo vā → bailvam

• 1516 bilvādibhyo 'ṇ 4.3.136 — aṇ after bilva etc. It is an apavāda of añ and mayaṭ (TB). bilva + aṇ → ādivṛddhi i → ai, final a deleted: bailvam. BM, following the Kaustubha, makes it narrower. gavedhuka would already get aṇ by 1517, so it is listed here only to block mayaṭ. The other members are listed to block the añ of 1520.
• 1517 kopadhāc ca 4.3.137 — aṇ also after a base whose penultimate is k; an apavāda of añ. tarku + aṇ → vṛddhi a → ā, and the u takes guṇa → av: tārkavam. Without this rule, orañ (1519) would come. tittiḍīka is madhyodātta, so the añ of 1520 is set aside: taittiḍīkam.
• 1518 trapujatunoḥ ṣuk 4.3.138 — after trapu ("tin") and jatu ("lac"): aṇ, carried from 1516, and the base takes the augment ṣuk. The rule is an apavāda of orañ. It applies in the vikāra sense only, because tin and lac are not living beings etc. (TB). trapu + ṣ + a → trāpuṣam; likewise jātuṣam.

C. añ: u-final, anudāttādi, optionally palāśādi, and ṣlañ for śamī — running example: devadāror avayavo vikāro vā → daivadāravam

• 1519 orañ 4.3.139 — añ after a base ending in u or ū, in vikāra and avayava as applicable. devadāru + añ → ādivṛddhi e → ai; the final u takes guṇa → av before the a: daivadāravam; also bhādradāravam. Both bases are ādyudātta by a phiṭ rule, so 1520 would not cover them and this rule is not redundant (BM, TB).
• 1520 anudāttādeś ca 4.3.140 — añ also after a base whose first vowel is anudātta. dadhittha, antodātta and so anudātta-initial, gives dāghittham. The source reads gh; dādhittham is intended, as the kaumudī itself reads at 1533. Likewise kāpittham. BM and TB differ on why the bases are antodātta: BM says they are underived words with the phiṭ accent, TB derives dadhittha as a kṛt compound.
• 1521 palāśādibhyo vā 4.3.141 — añ optionally after palāśa etc. This is an ubhayatra-vibhāṣā (TB): for the anudāttādi members (palāśa, khadira, śiṃśapā, syandana) añ was already obligatory by 1520, and for the rest (karīra etc.) it was not available at all. palāśa is antodātta as one of ghṛtādi (phiṭ ghṛtādīnāṃ ca; TB's OCR reads "vṛtādi"): pālāśam. karīra is ādyudātta by the nit accent: kārīram. The optional aṇ produces the same written forms.
• 1522 śamyāḥ ṣlañ 4.3.142 — ṣlañ (→ -la, with ādivṛddhi) after śamī, in vikāra and avayava. śamī + la → śāmīlaṃ bhasma. Because ṣ is a marker, the feminine takes ṅīṣ (4.1.41): śāmīlī sruk. TB records that Mādhava reads the affix as ṭit (ṭitvān ṅīp). The Aṣṭādhyāyī reference text also reads śamyāṣ ṭlañ. BM has no block on this sūtra in the neighbouring units. TB on it is printed under sk.1521.t1.

D. mayaṭ: optional after any base, obligatory after some, and its special cases — running example: aśman + mayaṭ → aśmamayam (optionally āśmanam)

• 1523 mayaḍ vaitayor bhāṣāyām abhakṣyācchādanayoḥ 4.3.143 — in ordinary language, mayaṭ optionally, in vikāra and avayava, after any base whatever (prakṛtimātrāt), unless the thing is food or clothing. The word etayoḥ lets mayaṭ reach even the domains of the special affixes, so bilvamayam stands beside bailvam (BM, TB). aśman + mayaṭ: the base has pada status before mayaṭ, so its n is deleted: aśmamayam. In the other option the form is āśmanam. BM and TB derive āśmanam from Aśman, the personal name (the son of Vasiṣṭha, also called Aśmaka), with prakṛtibhāva by an. The ṭi-lopa of 1514 is for aśman "stone" only. With food or clothing there is no mayaṭ: maudgaḥ sūpaḥ, kārpāsam ācchādanam.
• 1524 nityaṃ vṛddhaśarādibhyaḥ 4.3.144 — mayaṭ obligatorily after a vṛddha word and after śara etc.; an apavāda of the option in 1523. āmra is vṛddha: āmramayam; also śaramayam. Vārttika ekāco nityam: obligatory mayaṭ after a monosyllable. TB says it comes from splitting off "nityam" (yogavibhāga). tvac → tvaṅmayam, vāc → vāṅmayam, ap → ammayam. The usage āpyam has another derivation: āpa (aṇ, by tasyedam) + svārthika ṣyañ.
• 1525 goś ca purīṣe 4.3.145 — mayaṭ after go when its dung is meant: gomayam. TB: dung is neither product nor part, so the affix is in the tasyedam sense; the go-vikāra takes yat (1538; cf. gavyaṃ payaḥ). The NSP footnote instead treats dung as a transformation of the cow's fodder, so the rule stays within the section. The kaumudī names no affix; mayaṭ is the translator's supplement.
• 1526 piṣṭāc ca 4.3.146 — mayaṭ after piṣṭa ("flour"), in the vikāra sense: piṣṭamayaṃ bhasma. paiṣṭī surā has aṇ by tasyedam (SK 1500), with the general "this is its" relation intended (sāmānyavivakṣā).
• 1527 saṃjñāyāṃ kan 4.3.147 — with piṣṭa carried forward, kan when a name is meant: piṣṭa + ka → piṣṭakaḥ, a particular flour product, the cake (apūpa). The Amarakośa line "pūpo 'pūpaḥ piṣṭakaḥ syāt" is quoted in support. A puroḍāśa is not a piṣṭaka (BM).
• 1528 vrīheḥ puroḍāśe 4.3.148 — mayaṭ necessarily after vrīhi when the product is a puroḍāśa; an apavāda of the bilvādi aṇ (1516): brīhimayaḥ puroḍāśaḥ (so the source, with b). Other rice products keep aṇ: vraiham. This is why vrīhi is in the bilvādi list (BM).
• 1529 asaṃjñāyāṃ tilayavābhyām 4.3.149 — mayaṭ necessarily after tila and yava when no name is meant: tilamayam, yavamayam. As names: tailam, yāvakaḥ (yava + aṇ, then svārthika kan by yāvādibhyaḥ; BM, TB).

E. aṇ and añ that set aside mayaṭ (and añ) — running example: rajatasya vikāraḥ → rājatam (not rajatamayam)

• 1530 tālādibhyo 'ṇ 4.3.152 — aṇ after tāla etc.; an apavāda of añ and mayaṭ. Gaṇasūtra tālād dhanuṣi: for tāla, only when the product is a bow. tāla is vṛddha, so the aṇ displaces the mayaṭ of 1524: tālaṃ dhanuḥ. For other products the mayaṭ stays: tālamayam. indrāyudha is antodātta, so the aṇ displaces the añ of 1520: aindrāyudham (BM).
• 1531 jātarūpebhyaḥ parimāṇe 4.3.153 — aṇ after words for gold when the product is a measure. The plural takes in the synonyms: hāṭakaḥ / tāpanīyaḥ / sauvarṇo niṣkaḥ. When no measure is meant: hāṭakamayī yaṣṭiḥ. BM and TB differ on which base's aṇ displaces mayaṭ and which displaces añ.
• 1532 prāṇirajatādibhyo 'ñ 4.3.154 — añ after words for living beings and after rajata etc. śuka and baka are ādyudātta (prāṇināṃ kupūrvam), so only this rule gives them añ: śaukam, bākam. rajata is anudāttādi and would already have añ; it is listed again to block mayaṭ, which would otherwise win as the later rule (paratva): rājatam (BM, TB). TB: the rule also blocks mayaṭ after vṛddha words for living beings.
• 1533 ñitaś ca tatpratyayāt 4.3.155 — añ after a word that already ends in a ñit vikāra/avayava affix; an apavāda of mayaṭ.
• śāmīla (from ṣlañ, 1522) is vṛddha, so mayaṭ would come by 1524; this rule gives śāmīlam instead. dādhittha (from añ, 1520) gives dādhittham; likewise kāpittham.
• Counter-example: bailva has the aṇ of 1516, which is not ñit, so mayaṭ comes: bailvamayam.
• The Bhāṣya rejects this sūtra, on the ground that a product of a product is not expressed by a second affix (BM).
• The Śekhara (NSP footnote) holds that the affix of 1522 is ṭit. If it were ṣit, rejecting this rule would give ṅīṣ and keeping it would give ṅīp, a difference in result that is not wanted.

F. Affixes after measures and single bases: krīta-affixes, vuñ, ḍhañ — running example: uṣṭrasya vikāro 'vayavo vā → auṣṭrakaḥ

• 1534 krītavat parimāṇāt 4.3.156 — after a measure-word, the affixes taught for "bought" (krīta), with the same conditions, are extended to the vikāra sense; an apavāda of aṇ etc. niṣkeṇa krītam → naiṣkikam (ṭhak by asamāse niṣkādibhyaḥ, 5.1.20), so niṣkasya vikāraḥ → naiṣkikaḥ. Likewise śatyaḥ / śatikaḥ (yat and ṭhan by śatāc ca ṭhanyatāv aśate, 5.1.21). The kaumudī says these affixes begin from "prāg vahateṣ ṭhak (SK 1548)", which is 4.4.1. The krīta-affixes actually stand under prāg vateṣ ṭhañ (5.1.18, SK 1680), so the reference is probably corrupt; the SK 1548 cross-reference comes from the source.
• 1535 uṣṭrād vuñ 4.3.157 — vuñ (→ aka) after uṣṭra; an apavāda of the animal-añ (1532). uṣṭra + aka → ādivṛddhi u → au: auṣṭrakaḥ.
• 1536 umorṇayor vā 4.3.158 — vuñ optionally after umā ("flax") and ūrṇā ("wool"): aumakam, aurṇakam. Without vuñ, the affixes are respectively:
• umā is ādyudātta, so it takes the plain aṇ: aumam.
• ūrṇā is antodātta, so it takes the añ of 1520: orṇam. The source prints this form without vṛddhi; aurṇam is expected.
• 1537 eṇyā ḍhañ 4.3.159 — ḍhañ (→ eya) after the feminine eṇī ("doe"); an apavāda of the animal-añ. The ī is deleted by yasyeti ca: aiṇeyam. The feminine form in the sūtra shows its purpose: the masculine eṇa takes the añ: aiṇam.

G. yat for go, payas, dru, and vaya for a wooden measure — running example: dror vikāro 'vayavo vā → dravyam

• 1538 gopayasor yat 4.3.160 — yat after go and payas. go + ya → av by vānto yi pratyaye (6.1.79, SK 63): gavyam; payas gives payasyam. yat after go is already available by gor ajādiprasaṅge yat (under SK 1077); the rule is repeated to block the optional mayaṭ of 1523 (BM, TB).
• 1539 droś ca 4.3.161 — yat also after dru ("tree"); an apavāda of the ekāco nityam mayaṭ (1524) and of orañ (1519). dru + ya → the u takes guṇa by or guṇaḥ (6.4.146) → dro, then o → av before y: dravyam. The dravya of "dravya-guṇa-karma" is a different word, formed from the root dru with yat by aco yat (TB).
• 1540 māne vayaḥ 4.3.162 — with dru carried forward, vaya when the product is a measure; an apavāda of yat (TB): druvayam, a wooden measure such as a prastha. A line of unnamed source: yautavaṃ druvayaṃ pāyyam iti mānārthakaṃ trayam ("these three denote a measure").

H. Naming the fruit: luk, lup and their exceptions — running example: āmalakyāḥ phalam → āmalakam

• 1541 phale luk 4.3.163 — luk of the vikāra/avayava affix when the fruit is meant; a fruit counts as both product and part (TB, BM). The commentators start from āmalakī + mayaṭ, the vikāra/avayava affix.
• Luk of mayaṭ by this rule, then luk of the feminine ṅīṣ by luk taddhitaluki (1.2.49, SK 1408). āmalakī is gaurādi; TB's OCR "ḍīp" is probably the ṅīṣ that BM gives.
• The result is āmalaka, neuter: āmalakam.
• An element removed by luk is not sthānivat, so yasyeti ca does not apply (TB).
• 1542 plakṣādibhyo 'ṇ 4.3.164 — aṇ after plakṣa etc.; an apavāda of añ (orañ for śigru and karkandhū; 1520 for the rest). There is no luk in the fruit sense, because the rule was taught expressly (vidhānasāmarthya): plākṣam, "the plakṣa fruit". TB's Amara quotation lists plākṣa and naiyagrodha among fruit-names.
• 1543 nyagrodhasya ca kevalasya 7.3.5 — the SK reads it here, with nyagrodha a member of plakṣādi. BM identifies it only as the sūtra that follows na yvābhyām… (7.3.3, SK 1098). That the kaumudī's "no vṛddhi" and "augment aic" carry over from 7.3.3 is the translator's inference. For nyagrodha standing alone, vṛddhi does not fall on the y, and aic is inserted before it: nyagrodha + aṇ → naiyagrodham. The rule is a niyama if nyagrodha is derived (nyak rohati), and a vidhi if it is underived. Counter-example (not kevala): nyāgrodhamūlāḥ śālayaḥ.
• 1544 jambvā vā 4.3.165 — after jambū, aṇ optionally for the fruit: jāmbavam. Otherwise orañ (1519) applies and phale luk removes it. The word is then neuter in agreement with phala, so the vowel is shortened: jambu.
• 1545 lup ca 4.3.166 — lup of the fruit-affix after jambū, also optional. Under lup, by lupi yuktavad vyaktivacane (1.2.51, SK 1294), the form keeps the base's gender and number: jambvāḥ phalam → jambūḥ (feminine singular, even for several fruits). Vārttikas:
• phalapākaśuṣām upasaṃkhyānam — lup after plants that wither when their fruit ripens; an apavāda of phale luk. BM: lup of the bilvādi aṇ gives masculine plural by yuktavat: vrīhayaḥ, mudgāḥ. TB's OCR reads "luk" here, but only lup gives vrīhayaḥ.
• puṣpamūleṣu bahulam — lup is variable (bahula) for flowers and roots:
• with lup, feminine by yuktavat: mallikā, jātī, vidārī.
• no lup: pāṭālāni puṣpāṇi (so the source; BM and TB read pāṭalāni), sālvāni mūlāni.
• by bāhulaka, luk instead, so the neuter follows the qualified word: aśokam, karavīram.
• 1546 harītakyādibhyaś ca 4.3.167 — lup of the fruit-affix after harītakī etc. harītakī is antodātta, so it takes the añ of 1520. Lup replaces the luk of 1541. BM identifies the next line as a vārttika: harītakyādīnāṃ liṅgam eva prakṛtivat (only the gender follows the base, not the number). The result is feminine and plural: harītakyāḥ phalāni → harītakyaḥ. TB: the lup also covers the mayaṭ of 1524 after drākṣā etc., the añ after anudāttādi words, and the aṇ after the rest.

I. Products of two -hita derivatives — running example: kaṃsīyasya vikāraḥ → kāṃsyam

• 1547 kaṃsīyaparaśavyayor yañañau luk ca 4.3.168 — after kaṃsīya and paraśavya, yañ and añ respectively, with luk of their cha and yat. By elimination (pariśeṣa), the luk falls on the cha and yat inside the bases (BM). The luk is not of yañ and añ, which would make teaching them pointless, and not of the bases.
• kaṃsa + cha (tasmai hitam, under prāk krītāc chaḥ 5.1.1) → kaṃsīyam. Its vikāra: + yañ, luk of cha, ādivṛddhi, a deleted → kāṃsyam.
• paraśu + yat (ugavādibhyo yat 5.1.2), with guṇa and av → paraśavyam. Its vikāra: + añ, luk of yat, and the u again → av → pāraśavaḥ.
• The y is not removed by halas taddhitasya (6.4.150), because "īti" is carried forward there.
• The tail of sk.1547.t1 (from iti prāgdīvyatīyāḥ samāptāḥ) and sk.1547.n1 belong to SK 1548ff. (p31).