Vaiyākaraṇa-siddhānta-kaumudī — pūrvakṛdantaprakaraṇam

Learner edition · Prakaraṇa 67

67. Pūrvakṛdanta-prakaraṇa — The first kṛt affixes: agent nouns, upapada compounds, the past and present participles, and the affixes of habit

Bhaṭṭoji heads this section "now the section on words ending in a kṛt" (kṛdanta-prakaraṇa). The kṛtya affixes of chapter 66 came only in the bare action or the object. The kṛt affixes taught from here on are no longer kṛtyas, so SK 2832 (kartari kṛt) holds in full: unless a rule says otherwise, the affix denotes the agent. The chapter has 273 sūtras, SK 2895 to SK 3167, in four parts. The first part (SK 2895–2960) teaches the agent affixes that need no other word, such as ṇvul and tṛc (kārakaḥ, kartā). It then teaches the affixes that come only beside a named word, the upapada, from aṇ (kumbhakāraḥ) to khac (priyaṃvadaḥ). The second part runs from khac and ḍa to the past-participle affixes (the niṣṭhā) and the n that replaces their t. The third part continues the niṣṭhā and the senses of kta. The fourth teaches the participles of liṭ, the present participles with śatṛ and śānac, the affixes of habit (tācchīlika), and ṣṭran and itra. After the first part you will be able to build agent nouns such as kārakaḥ, nandanaḥ, grāhī, pacaḥ, kṣipaḥ, pibaḥ, jvālaḥ, nartakaḥ and gāyanī. You will build upapada compounds such as kumbhakāraḥ, godaḥ, sāmagaḥ, aṃśaharaḥ, kurucaraḥ, yaśaskarī, stambakariḥ, janamejayaḥ, stanandhayaḥ and dviṣantapaḥ. And you will know why bhārahāraḥ has ā but aṃśaharaḥ has a.

Before you start

From chapter 1 you know saṃjñā (a technical name), it (a marker sound), adhikāra (a heading), anuvṛtti (carrying a word forward), upadeśa, guṇa (a, e, o), vṛddhi (ā, ai, au), dhātu, upasarga, vibhāṣā (option), tapara, saṃprasāraṇa and niyama (a restricting rule). You know SK 26 (yena vidhis tadantasya): a rule stated for an item also applies to what ends in it.
From chapter 3 you know SK 54 (saṃyogāntasya lopaḥ): the last consonant of a cluster at the end of a pada is lost. You know SK 61 (e, o, ai, au → ay, av, āy, āv before a vowel) and SK 70 (an a, i or u that replaces ṛ takes an r after it).
From chapter 5 you know ṣṭutva (SK 113): a dental next to ṣ becomes retroflex. You know SK 118 (after ud, the s of sthā and stambh becomes th) and SK 128 (yathāsaṃkhya: items listed in two series pair off in order).
From chapter 6 you know SK 159 and SK 160, which give ṣ or s for a visarga inside a compound before k: dhanuṣkaraḥ, yaśaskarī.
From chapter 8 you know SK 254: a vowel-final stem takes vṛddhi before an affix marked with ñ or ṇ.
From chapter 11 you know the name kṛt (SK 374). From chapter 15 you know the feminine affixes ṭāp (ā), ṅīp and ṅīṣ (ī). You know SK 470: a stem in a made with a ṭ-marked affix, or with aṇ and the like, takes ṅīp. You know SK 463 (i for the a before ka in the feminine) and SK 504 (puṃyoga: ī for "the wife of").
From chapter 16 you know SK 623: beside a kṛt, the agent or object takes the sixth case. From chapter 18 you know the upapada, a word that a kṛt rule names in the seventh case (SK 781). You know that it is compounded with the kṛdanta (SK 782, upapadam atiṅ). You also know SK 735 (upamitaṃ vyāghrādibhiḥ…).
From chapter 24 you know aluk, a first member that keeps its case-ending (SK 958, SK 966, SK 972). From chapter 25 you know SK 999, which gives a short vowel, and SK 1008 (rātreḥ kṛti vibhāṣā: optional m after rātri before a kṛt).
From chapter 28 you know SK 1247 (yuvor anākau): the affix-parts yu and vu become ana and aka.
From chapter 44 you know sārvadhātuka (a verb-ending or a ś-marked affix) and ārdhadhātuka. You know śap, the class-sign a of the bhū class, and iṭ, the augment i. You know SK 2168 (guṇa before either kind of affix), SK 2189 (guṇa of a light penultimate), SK 2217 (no guṇa or vṛddhi before a k- or ṅ-marked affix) and SK 2234 (a sārvadhātuka not marked with p counts as ṅ-marked). You know SK 2282 (vṛddhi of a penultimate a) and SK 2313 (loss of ṇi).
From chapter 44 you also know SK 2360 (piba, jighra, dhama, paśya … for pā, ghrā, dhmā, dṛś … before a ś-marked affix), SK 2364, SK 2370 (ā for a root-final diphthong), SK 2372 (āto lopa iṭi ca: loss of root-final ā) and SK 2412 (saṃprasāraṇa of grah, jyā and others). You know SK 2179 (halādiḥ śeṣaḥ), SK 2197, SK 2302, SK 2340, SK 2348 and SK 2401.
From chapters 45–57 you know SK 2423 (śap lost after the ad class), SK 2430, SK 2461, SK 2512, SK 2516, SK 2517, SK 2536, SK 2574, SK 2581, SK 2582 and SK 2604. You know yaṅ, the intensive affix, and yaṅluk, the intensive whose yaṅ is deleted (SK 2650, SK 2656). You know ṇic, the causative affix, which leaves i (ṇi) on the root.
From chapter 61 you know SK 2716. From chapter 63 you know SK 2761 (āto yuk: y after root-final ā before a ñ- or ṇ-marked kṛt) and SK 2763 (no vṛddhi for certain m-final roots). From chapter 65 you know SK 2822 (arhe kṛtyatṛcaś ca).
From chapter 66 you know SK 2829 (the heading "after a dhātu"), SK 2830 (vāsarūpa: an exception of a different form blocks the general affix only optionally), SK 2831 (the kṛtyas run up to ṇvul), SK 2832 (kartari kṛt: a kṛt denotes the agent) and SK 2841 (bahulam). You know utsarga (the general rule) and apavāda (its exception), and nipātana (a finished form stated in a sūtra).

I.A. The general agent affixes ṇvul and tṛc — running example: `kṛ + ṇvul → kṛ + vu → kṛ + aka → (vṛddhi) kārakaḥ`; `kṛ + tṛc → (guṇa) kartṛ → kartā`

One rule opens the chapter. SK 2895 gives two affixes after any root, ṇvul and tṛc, both meaning the agent. They are the general affixes (utsarga) of agent nouns: every later affix of this part is an exception to them or to aṇ. Bhaṭṭoji then runs many rules taught for verbs through these two affixes. The running example is kṛ "to do". With ṇvul it gives kārakaḥ "maker"; with tṛc it gives kartā "maker".

SK 2895 · ण्वुल्तृचौ · ṇvultṛcau (A 3.1.133)

ण्वुल्तृचौ

ṇvultṛcau · A 3.1.133

धातोरेतौ स्तः । कर्तरि कृत् SK 2832 इति कर्त्रर्थे । युवोरनाकौ SK 1247 । कारकः । कर्ता । वोढुमर्हो वोढा । कारिका । कर्त्री । गाङ्कुटा - SK 2461 इति ङित्त्वम् । कुटिता । अञ्णिदित्युक्तेर्न ङित्त्वम् । कोटकः । विज इट् SK 2536 । विजिता । हनस्तोऽचिण्णलोः SK 2574 । घातकः । आतो युक् SK 2761 । दायकः । नोदात्तोपदेशस्य SK 2763 इति न वृद्धिः । शमकः । दमकः । अनिटस्तु नियामकः । जनिवध्योश्च SK 2512 । जनकः । वध हिंसायाम् । वधकः । रधिजभोरचि SK 2302 । रन्धकः । जम्भकः । नेट्यलिटि रधेः SK 2516 । रधिता । रद्धा । मस्जिनशोः SK 2517 इति नुम् । मङ्क्ता । नंष्टा । नशिता । रभेरशब्लिटोः SK 2581 । रम्भकः । रब्धा । लभेश्च SK 2582 । लम्भकः । लब्धा । तीषसह-SK 2340 इति वेट् । एषिता । एष्टा । सहिता । सोढा । दरिद्रातेरालोपः । दरिद्रिता । ण्वुलि न । दरिद्रायकः । कृत्यल्युट SK 2841 इत्येव सूत्रमस्तु । यत्र विहितास्ततोऽन्यत्रापि स्युरित्यर्थात् । एवं च बहुलग्रहणं योगविभागेन कृन्मात्रस्यार्थव्यभिचारार्थम् । पादाभ्यां ह्रियते पादहारकः । कर्मणि ण्वुल् ।

क्रमेः कर्तर्यात्मनेपदविषयात्कृत इण्निषेधो वाच्यः

(वार्तिकम्) ॥ प्रक्रन्ता ॥ कर्तरीति किम् । प्रक्रमितव्यम् । आत्मनेपदेति किम् । संक्रमिता । अनन्यभावे विषयशब्दः । तेन अनुपसर्गाद्वा SK 2716 इति विकल्पार्हस्य न निषेधः । क्रमिता । तदर्हत्वमेव तद्विषयत्वम् । तेन क्रन्तेत्यपीति केचित् । गमेरिट्-SK 2401इत्यत्र परस्मैपदग्रहणं तङानयोरभावं लक्षयति । संजिगमिषिता । एवं न वृद्भ्यश्चतुर्भ्यः SK 2348 । विवृत्सिता । यङन्तात् ण्वुल् । अल्लोपस्य स्थानिवत्त्वान्न वृद्धिः । पापचकः । यङ्लुगन्तात्तु पापाचकः ॥

Rule. After a dhātu, the affixes ṇvul and tṛc come. Bhaṭṭoji: "these two come after a dhātu, in the sense of the agent, by kartari kṛt (SK 2832)". The heading "after a dhātu" is SK 2829. In ṇvul the ṇ and l are markers, leaving vu. By SK 1247 (yuvor anākau), vu becomes aka. In tṛc the c is a marker, leaving tṛ.

Example.

kṛ + ṇvul → markers go → kṛ + vu

→ SK 1247 → kṛ + aka

→ the ṇ-marked affix brings vṛddhi (SK 254), with r (SK 70) → kār + aka → kārakaḥ.

kṛ + tṛc → kṛ + tṛ → guṇa (SK 2168) → kartṛ → kartā.

Feminines: kārikā, kartrī.

Applies when / blocked by. One of Bhaṭṭoji's commentators explains the markers. The ṇ of ṇvul is for vṛddhi. The l is for an accent rule. The c of tṛc keeps tṛc apart where a rule names only tṛ.

Bhaṭṭoji also gives voḍhum arhaḥ voḍhā "one fit to carry". Both commentators explain: this tṛc is in the sense "deserving to", by SK 2822 (arhe kṛtyatṛcaś ca).

Bhaṭṭoji then shows root changes taught for verbs, working here too:

By SK 2461 tṛc counts as ṅ-marked after kuṭ and its group, so there is no guṇa: kuṭitā. That rule excepts affixes marked with ñ or ṇ. So ṇvul is not ṅ-marked: koṭakaḥ.
By SK 2536 (vija iṭ): vijitā. One commentator explains that the affix with iṭ counts as ṅ-marked after vij, so there is no guṇa.
SK 2574 gives t for the n of han: ghātakaḥ. SK 2761 gives yuk after an ā-final root: dāyakaḥ.
By SK 2763 there is no vṛddhi for śam and dam: śamakaḥ, damakaḥ. But yam does not take iṭ, and it has vṛddhi: niyāmakaḥ.
SK 2512 bars vṛddhi for jan and vadh: janakaḥ, vadhakaḥ. Bhaṭṭoji cites vadha "to strike" as a root. One commentator adds that it is a separate root of the bhū class, not the substitute for han.
num comes in randhakaḥ, jambhakaḥ (SK 2302). By SK 2516 radh has radhitā and raddhā. By SK 2517, num: maṅktā, naṃṣṭā; with iṭ, no num: naśitā. By SK 2581: rambhakaḥ, rabdhā. By SK 2582: lambhakaḥ, labdhā.
By SK 2340, iṭ is optional: eṣitā / eṣṭā, sahitā / soḍhā.
daridrā loses its ā: daridritā. But not before ṇvul: daridrāyakaḥ. One commentator says both points rest on vārttikas.

Next, Bhaṭṭoji reads SK 2841 as kṛtyalyuṭaḥ alone. The word bahulam is split off as its own rule (yogavibhāga). It then lets any kṛt vary its sense. So ṇvul can denote the object: pādābhyāṃ hriyate pādahārakaḥ "one carried by the feet". One commentator reports that the source has only the statement kṛto bahulam here; scholars discuss this.

A vārttika: no iṭ for a kṛt in the agent sense after kram when it is in the domain of the ātmanepada endings: prakrantā. Not in the agent sense: prakramitavyam. Not ātmanepada: saṃkramitā. Bhaṭṭoji explains "domain" as "invariably". kram takes ātmanepada only optionally by SK 2716, so it keeps iṭ: kramitā. Some allow krantā too.

In SK 2401 and SK 2348, Bhaṭṭoji says, the word "parasmaipada" means "where there is no ātmanepada ending or āna". So these rules still work before a kṛt, where no verb-ending follows. saṃjigamiṣitā keeps the iṭ of SK 2401 in its desiderative san. vivṛtsitā lacks it there by SK 2348.

Last, ṇvul after a root with yaṅ. The lost a of yaṅ still counts (sthānivat), so there is no vṛddhi: pāpacakaḥ. After a yaṅluk stem nothing stands in the way: pāpācakaḥ.

Watch out. Do not think tṛc and ṇvul behave alike. ṇvul is ṇ-marked, so it brings vṛddhi and is never ṅ-marked: kārakaḥ, koṭakaḥ. tṛc brings guṇa: kartā, and after kuṭ none at all: kuṭitā.

I.B. lyu, ṇini and ac for the three gaṇas — running example: `nandi + lyu → (yu → ana) nandanaḥ`; `grah + ṇini → (vṛddhi) grāhī`; `pac + ac → pacaḥ`

Three lists of roots (gaṇa) take their own agent affixes in place of ṇvul and tṛc. The nandi-list takes lyu, the grahi-list ṇini, the paca-list ac. The paca-list is open: new members are recognised from usage. The running examples are nandanaḥ "one who gladdens", grāhī "one who takes", and pacaḥ "one who cooks".

SK 2896 · नन्दिग्रहिपचादिभ्यो ल्युणिन्यचः · nandigrahipacādibhyo lyuṇinyacaḥ (A 3.1.134)

नन्दिग्रहिपचादिभ्यो ल्युणिन्यचः

nandigrahipacādibhyo lyuṇinyacaḥ · A 3.1.134

नन्द्यादेर्ल्युर्ग्रह्यादेर्णिनिः पचादेरच् स्यात् । नन्दयतीति नन्दनः । जनमर्दयतीति जनार्दनः । मधुसूदनः । विशेषेण भीषयतीति विभीषणः । लवणः । नन्द्यादिगणे निपातनाण्णत्वम् । ग्राही । स्थायी । मन्त्री । विशयी । वृद्ध्याभावो निपातनात् । विषयी । इह षत्वमपि । परिभावी । परिभवी । पाक्षिको वृद्ध्यभावो निपात्यते । पचादिराकृतिगणः । शिवशमरिष्टस्य करे SK 3489 कर्मणि घटोऽठच् SK 1836 इति सूत्रयोः करोतेर्घटेश्चाच् प्रयोगात् । अच् प्रत्यये परे यङिलुग्विधानाच्च । केषांचिद्बाधकबाधनार्थः । पचतीति पचः । नदट् । चोरट् । देवट् । इत्यादयष्टितः । नदी । चोरी । देवी । दीव्यतेः । इगुपध - SK 2897 इति कः प्राप्तः । न्यङ्क्वादिषु पाठात् श्वापकोऽपि । यङोऽचि च SK 2650 इति लुक् । न धातुलोप - SK 2656 इति गुणवृद्धिनिषेधः । चेक्रियः । नेन्यः । लोलुवः । पोपुवः । मरीमृजः ।

चरिचलिपतिवदीनां वा द्वित्वमच्याक्चाभ्यासस्येति वक्तव्यम्

(वार्तिकम्) ॥ आगमस्य दीर्घत्वसामर्थ्यादभ्यासह्रस्वो हलादिः शेषः SK 2179 च न । चराचरः । चलाचलः । पतापतः । वदावदः ।

हन्तेर्घत्वं च

(वार्तिकम्) ॥ घत्वमभ्यासस्य उत्तरस्य तु अभ्यासाच्च SK 2430 इति कुत्वम् । घनाघनः ।

पाटेर्णिलुक्चोक्च दीर्घश्चाभ्यासस्य

(वार्तिकम्) ॥ पाटुपटः ॥ पक्षे चरः । चलः । पतः । वदः । हनः । पाटः । रात्रेः कृति - SK 1008 इति वा मुम् । रात्रिंचरः । रात्रचरः ॥

Rule. After the nandi-group comes lyu, after the grahi-group ṇini, after the paca-group ac, in that order. Both commentators say that "ādi" ("and the rest") goes with each of the three. One adds that nandi is the root with ṇi.

Example.

nandi + lyu → the marker l goes → yu becomes ana (SK 1247) → nandanaḥ.

grah + ṇini → markers go → grah + in → vṛddhi of the penultimate a → grāhin → grāhī.

pac + ac → pac + a → pacaḥ.

Applies when / blocked by.

lyu: nandanaḥ, janārdanaḥ "one who torments people", madhusūdanaḥ, vibhīṣaṇaḥ "one who terrifies greatly". One commentator notes that with an object word aṇ (SK 2913) would come in madhusūdana; the listing gives lyu. In lavaṇaḥ the ṇ is by nipātana in the list.

ṇini: grāhī, sthāyī, mantrī. In viśayī the lack of vṛddhi is by nipātana. In viṣayī the ṣ is too. paribhāvī / paribhavī: the lack of vṛddhi is optional.

ac: the paca-group is an ākṛtigaṇa, an open list. Bhaṭṭoji's proof: Pāṇini himself uses ac after kṛ and ghaṭ in SK 3489 and SK 1836. And the deletion of yaṅ before ac is taught at all. Bhaṭṭoji names three reasons for listing some words:

to attach a marker: nadaṭ, coraṭ, devaṭ are ṭ-marked, so nadī, corī, devī;
for fullness of statement;
to block a blocking rule. deva is listed because SK 2897 would give ka after div. jārabharā and śvapacā are listed because SK 2913 would give aṇ.

Because śvapāka is read in another list (nyaṅkvādi), śvapākaḥ is also correct. Both commentators take it as aṇ. One adds that it is the other option.

ac after a yaṅ-stem: yaṅ is lost (SK 2650), and guṇa and vṛddhi are barred (SK 2656). cekriyaḥ, nenyaḥ, loluvaḥ, popuvaḥ, marīmṛjaḥ.

Three vārttikas on ac:

car, cal, pat, vad may be doubled before ac, and the first copy (abhyāsa) takes the augment āk. Since the augment is long, it is not shortened, and SK 2179 does not remove the second consonant: carācaraḥ, calācalaḥ, patāpataḥ, vadāvadaḥ.
han takes these and gh in the first copy. The second copy gets a velar by SK 2430: ghanāghanaḥ.
pāṭi loses its ṇi, takes ūk, and the first copy is long: pāṭūpaṭaḥ.

In the other option: caraḥ, calaḥ, pataḥ, vadaḥ, hanaḥ, pāṭaḥ. By SK 1008, m is optional after rātri: rātriṃcaraḥ / rātricaraḥ.

Watch out. Under this rule the three affixes pair with the three lists in order: nandi takes lyu, not ṇini. And the paca-list is open, so a word not found in it may still take ac.

I.C. ka after an ik-penultimate root, jñā, prī and kṝ, and after ā-finals with an upasarga — running example: `kṣip + ka → (k: no guṇa) kṣipaḥ`; `pra + jñā + ka → (ā-loss) prajñaḥ`

The affix ka is a with the marker k. A k-marked affix blocks guṇa (SK 2217), and before it a root-final ā is lost (SK 2372). SK 2897 gives ka after roots with i, u, ṛ as penultimate, and after three named roots. SK 2898 gives it after ā-final roots with an upasarga. The running examples are kṣipaḥ "one who throws" and prajñaḥ "wise".

SK 2897 · इगुपधज्ञाप्रीकिरः कः · igupadhajñāprīkiraḥ kaḥ (A 3.1.135)

इगुपधज्ञाप्रीकिरः कः

igupadhajñāprīkiraḥ kaḥ · A 3.1.135

एभ्यः कः स्यात् । क्षिपः । लिखः । बुधः । कृशः । ज्ञः । प्रीणातीति प्रियः । किरतीति किरः । वासरूपविधिना ण्वुल्तृचावपि । क्षेपकः । क्षेप्ता ॥

Rule. The affix ka comes after a root whose penultimate is an ik (i, u, ṛ, ḷ), and after jñā "to know", prī "to please" and kṝ "to scatter". Such a root is igupadha. Both commentators explain kir: it is kṝ with i and r already made.

Example.

kṣip + ka → the k blocks guṇa → kṣipaḥ.

Also likhaḥ, budhaḥ, kṛśaḥ.

jñā + ka → ā lost (SK 2372) → jñaḥ.

prī + ka → priyaḥ (iy for ī, says one commentator) "one who pleases".

kṝ + ka → kiraḥ "one who scatters".

Applies when / blocked by. ka is an exception to ṇvul and tṛc. But by the vāsarūpa rule (SK 2830) they still come beside it: kṣepakaḥ, kṣeptā. Here the guṇa is back, since ṇvul and tṛc are not k-marked.

Watch out. "Penultimate" means the sound just before the last sound. In kṣip that is i. A root like pac, with a, is not igupadha.

SK 2898 · आतश्चोपसर्गे · ātaś copasarge (A 3.1.136)

आतश्चोपसर्गे

ātaścopasarge · A 3.1.136

कः स्यात् ।श्याद्व्यधा - SK 2903 इति णस्यापवादः । सुग्लः । प्रज्ञः ॥

Rule. After a root ending in ā, when an upasarga stands before it as upapada, ka comes. Bhaṭṭoji supplies "ka comes". He calls the rule the exception to the ṇa of SK 2903. One commentator explains: ṇa is taught for ā-final roots in general, with or without an upasarga.

Example.

su + glai + ka → ā for ai (SK 2370) → su + glā + a → ā lost (SK 2372) → suglaḥ.

pra + jñā + ka → ā lost → prajñaḥ.

Applies when / blocked by. Without an upasarga, the ā-final root takes ṇa by SK 2903. With an upasarga, this ka wins. SK 2903 then names śyā and avasā separately, so that they keep ṇa even with an upasarga.

Watch out. The ā of the root disappears before ka: write prajñaḥ, not prajñāḥ.

I.D. śa: agents built on the sārvadhātuka stem — running example: `pā + śa → (ś: sārvadhātuka; piba for pā, SK 2360; śap, pararūpa) pibaḥ`

The affix śa is a with the marker ś. An affix marked with ś is sārvadhātuka, like a present-tense ending. So the root takes its present stem before it: the class-sign śap and the special stems of SK 2360. SK 2899–2901 give śa after named roots. The running example is pā "to drink": pā + śa → piba for pā (SK 2360) → śap, and its a merges with the next → pibaḥ "one who drinks".

SK 2899 · पाघ्राध्माधेट्दृशः शः · pāghrādhmādheṭdṛśaḥ śaḥ (A 3.1.137)

पाघ्राध्माधेट्दृशः शः

pāghrādhmādheṭdṛśaḥ śaḥ · A 3.1.137

पिबतीति पिबः । जिघ्रः । धमः । धयः । धया कन्या । धेटष्टित्त्वात् स्तनन्धयीति खशीव ङीप् प्राप्तः । खशोऽन्यत्र नेष्यत इति हरदत्तः । पश्यतीति पश्यः ॥

घ्रः संज्ञायां न

(वार्तिकम्) ॥ व्याघ्रादिभिः - SK 735 इति निर्देशात् ॥

Rule. The affix śa comes after pā "to drink", ghrā "to smell", dhmā "to blow", dheṭ "to suck" and dṛś "to see". Both commentators say pā is the root "to drink", not pā "to protect". The latter has its class-sign deleted, and such roots are not taken here.

Example.

pā + śa → śa is sārvadhātuka → piba for pā (SK 2360) → śap and the a merges → pibaḥ.

Also jighraḥ (ghrā), dhamaḥ (dhmā), dhayaḥ (dheṭ), paśyaḥ "one who sees" (dṛś).

Applies when / blocked by. dhayā kanyā "a girl who sucks". dheṭ is ṭ-marked, so the feminine ṅīp (ī) might be expected, as in the khaś-form stanaṃdhayī (SK 2944). Haradatta holds that this ṅīp is not wanted except with khaś. (Haradatta wrote the Padamañjarī, a commentary on the Kāśikā, which explains the sūtras.) Bhaṭṭoji reports this as Haradatta's view. One of his commentators rejects the restriction; scholars discuss this.

A vārttika: śa does not come after ghrā in a name (saṃjñā). Bhaṭṭoji's proof is Pāṇini's own wording in SK 735, vyāghrādibhiḥ. One commentator explains: with śa Pāṇini would have written vyājighra.

Watch out. One commentator notes that "with an upasarga" is not carried into this rule. So paśyaḥ needs no upasarga.

SK 2900 · अनुपसर्गाल्लिम्पविन्दधारिपारिवेद्युदेजिचेतिसातिसाहिभ्यश्च · anupasargāl limpavindadhāripārivedyudejicetisātisāhibhyaś ca (A 3.1.138)

अनुपसर्गाल्लिम्पविन्दधारिपारिवेद्युदेजिचेतिसातिसाहिभ्यश्च

anupasargāllimpavindadhāripārivedyudejicetisātisāhibhyaśca · A 3.1.138

शः स्यात् । लिम्पः । विन्दः । धारयः । पारयः । वेदयः । उदेजयः । चेतयः । सातिः सुखार्थः । सौत्रो हेतुमण्ण्यन्तः । सातयः । वाऽसरूपन्यायेन क्विपि । सात् परमात्मा । सात्वन्तो भक्ताः । षह मर्षणे चुरादिः । हेतुमण्ण्यन्तो वा । साहयः । अनुपसर्गात्किम् । प्रलिपः ।

नौ लिम्पेर्वाच्यः

(वार्तिकम्) ॥ निलिम्पा देवाः ।

गवादिषु विदेः संज्ञायाम्

(वार्तिकम्) ॥ गोविन्दः । अरविन्दम् ॥

Rule. śa also comes after limp, vind, dhāri, pāri, vedi, udeji, ceti, sāti and sāhi, when no upasarga stands before them. Bhaṭṭoji supplies "śa comes". Several of these are causative stems with ṇi.

Example.

limpaḥ, vindaḥ. One commentator: the roots are named with their n already inserted, so only the tud-class lip and vid "to obtain" are meant.

dhāri + śa → śap → guṇa of i and ay → dhārayaḥ. So too pārayaḥ, vedayaḥ, udejayaḥ, cetayaḥ (the steps are one commentator's).

sāti "pleasure" is a root found only in the sūtras, here with causative ṇi: sātayaḥ. sāhi is ṣah "to endure" of the cur class, or its causative: sāhayaḥ.

Applies when / blocked by. By the vāsarūpa principle (SK 2830), sāti also takes kvip, an affix that leaves nothing: sāt, the supreme Self (paramātman), and sātvantaḥ, the devotees.

With an upasarga there is no śa: pralipaḥ. One commentator says this is ka by SK 2897.

Two vārttikas:

śa after limp with ni: nilimpāḥ, the gods.
after vind with go and the like, śa comes only in a name: govindaḥ, aravindam. One commentator calls this a restriction (niyama), and śa here an exception to aṇ.

Watch out. The nine stems are cited as they appear before śa. Do not take limpa and vinda as the bare roots.

SK 2901 · ददातिदधात्योर्विभाषा · dadātidadhātyor vibhāṣā (A 3.1.139)

ददातिदधात्योर्विभाषा

dadātidadhātyorvibhāṣā · A 3.1.139

शः स्यात् । ददः । दधः । पक्षे वक्ष्यमाणो णः । अनुपसर्गादित्येव । प्रदः । प्रधः ॥

Rule. After dā "to give" (dadāti) and dhā "to place" (dadhāti), śa comes optionally. Bhaṭṭoji adds that "without an upasarga" still holds.

Example.

dā + śa → class-sign deleted (ślu), the root doubled, ā lost → dadaḥ. Also dadhaḥ.

Both commentators give these steps.

Applies when / blocked by. In the other option comes ṇa, taught below (SK 2903): dāyaḥ, dhāyaḥ (one commentator). With an upasarga there is no śa: pradaḥ, pradhaḥ. Both commentators say these take ka by SK 2898.

Watch out. The option is between śa and ṇa, not between śa and nothing. dadaḥ and dāyaḥ are both correct for "giver".

I.E. ṇa, optional, invariable and vyavasthita, and ka for the house — running example: `jval + ṇa → (vṛddhi) jvālaḥ` / `jval + ac → jvalaḥ`; `dā + ṇa → (yuk) dāyaḥ`

The affix ṇa is a with the marker ṇ, so it brings vṛddhi. SK 2902 gives it optionally after a run of roots, with ac in the other option. SK 2903 gives it invariably after another set. SK 2904 gives ṇa after du and nī when no upasarga stands before them. SK 2905 makes it optional for grah, each form keeping its own meaning. SK 2906 then gives ka to grah for a house. The running examples are jval "to blaze", which gives jvālaḥ or jvalaḥ "flame", and dā "to give", which gives dāyaḥ "giver".

SK 2902 · ज्वलितिकसन्तेभ्यो णः · jvalitikasantebhyo ṇaḥ (A 3.1.140)

ज्वलितिकसन्तेभ्यो णः

jvalitikasantebhyo ṇaḥ · A 3.1.140

इति शब्द आद्यर्थः । ज्वलादिभ्यः कसन्तेभ्यो णः स्याद्वा । पक्षेऽच् । ज्वालः । ज्वलः । चालः । चलः । अनुपसर्गादित्येव । उज्ज्वलः ।

तनोतेरुपसंख्यानम्

(वार्तिकम्) ॥ इहानुपसर्गादिति विभाषेति च न संबध्यते । अवतनोतीत्यतानः ॥

Rule. After the roots from jval to kas in the root list, ṇa comes optionally. In the other option comes ac. Bhaṭṭoji explains the iti of the sūtra as "beginning with". "Optionally" runs on from SK 2901, says one commentator. "Without an upasarga" runs on too: Bhaṭṭoji says so.

Example.

jval + ṇa → marker ṇ goes → vṛddhi of the penultimate a → jvālaḥ.

jval + ac → jvalaḥ.

cal → cālaḥ / calaḥ.

With an upasarga, only ac: ujjvalaḥ.

Applies when / blocked by. One commentator adds that where a root qualifies, ka by SK 2897 also comes in the other option.

A vārttika adds tan "to spread" to the roots that take ṇa. Bhaṭṭoji says that here neither "without an upasarga" nor "optionally" applies: avatanoti iti avatānaḥ "one who spreads down".

Watch out. For tan the conditions are lifted. So avatānaḥ has both an upasarga and an invariable ṇa.

SK 2903 · श्याद्व्यधास्रुसंस्व्रतीणवसावहृलिहश्लिषश्वसश्च · śyādvyadhāsrusaṃsvratīṇavasāvahṛlihaśliṣaśvasaś ca (A 3.1.141)

श्याद्व्यधास्रुसंस्व्रतीणवसावहृलिहश्लिषश्वसश्च

śyādvyadhāsrusaṃsvratīṇavasāvahṛlihaśliṣaśvasaśca · A 3.1.141

श्यैङ्प्रभृतिभ्यो नित्यं णः स्यात् । श्यैङोऽवस्यतेश्चादन्तात्वात्सिद्धे पृथग्ग्रहणमुपसर्गे कं बाधितुम् । अवश्यायः । प्रतिश्यायः । आत् । दायः । धायः । व्याधः । स्रु गतौ । आङ्पूर्वः संपूर्वश्च । आस्रावः । संस्रावः । अत्यायः । अवसायः । अवहारः । लेहः । श्लेषः । श्वासः ॥

Rule. ṇa comes invariably after śyā, after any ā-final root, and after vyadh, āsru, saṃsru, atīṇ, avasā, avahṛ, lih, śliṣ and śvas. The sūtra is read here as naming āsru and saṃsru. Both commentators explain why the option has gone. The next rule says "without an upasarga" again, so that condition stops here. And "optionally", which went with it, stops too.

Example.

dā + ṇa → yuk after the ā (SK 2761) → dāy + a → dāyaḥ. Also dhāyaḥ. (The yuk step is both commentators'.)

vyadh → vyādhaḥ.

ā + sru "to go" → āsrāvaḥ; sam + sru → saṃsrāvaḥ.

atyāyaḥ, avasāyaḥ, avahāraḥ, lehaḥ, śleṣaḥ, śvāsaḥ.

Applies when / blocked by. śyā and avasā are ā-final already, so "ā-final" would cover them. Bhaṭṭoji says they are named to block ka when an upasarga is present. Without this, SK 2898 would give ka. So: avaśyāyaḥ, pratiśyāyaḥ.

One commentator adds that with ka by SK 2897, lehaḥ and śleṣaḥ would have no guṇa.

Watch out. This rule overrides SK 2898 only for śyā and avasā. Other ā-final roots with an upasarga still take ka: prajñaḥ.

SK 2904 · दुन्योरनुपसर्गे · dunyor anupasarge (A 3.1.142)

दुन्योरनुपसर्गे

dunyoranupasarge · A 3.1.142

णः स्यात् । दुनोतीति दावः । नीसाहचर्यात्सानुबन्धाद्दुनोतेरेव णः । दवतेस्तु पचाद्यच् । दवः । नयतीति नायः । उपसर्गे तु प्रदवः । प्रणयः ॥

Rule. After du and nī, without an upasarga, ṇa comes. Bhaṭṭoji supplies "ṇa comes".

Example.

du + ṇa → vṛddhi → dau + a → āv (SK 61) → dāvaḥ "that which torments".

nī + ṇa → nāyaḥ "leader".

Applies when / blocked by. There are two roots du. nī is cited in the root list with a marker. Bhaṭṭoji argues from this pairing (sāhacarya): du too means the marked root ṭudu "to torment" (dunoti). The other du "to go" (davati) takes the pacādi ac: davaḥ.

With an upasarga: pradavaḥ, praṇayaḥ.

Watch out. dāvaḥ and davaḥ come from different roots. Only dunoti takes this ṇa.

SK 2905 · विभाषा ग्रहः · vibhāṣā grahaḥ (A 3.1.143)

विभाषा ग्रहः

vibhāṣā grahaḥ · A 3.1.143

णो वा । पक्षेऽच् । व्यवस्थितविभाषेयम् । तेन जलचरे ग्राहः । ज्योतिषि ग्रहः ॥ भवतेश्चेति काशिका ॥ भवो देवः संसारश्च । भावाः पदार्थाः । भाष्यमते तु प्राप्त्यर्थाच्चुरादिण्यन्तादच् । भावः ॥

Rule. After grah, ṇa comes optionally; in the other option, ac. Bhaṭṭoji calls this an option with fixed domains (vyavasthita-vibhāṣā). Each form keeps to one meaning.

Example.

grah + ṇa → vṛddhi → grāhaḥ: an aquatic creature.

grah + ac → grahaḥ: a heavenly body.

Applies when / blocked by. The Kāśikā adds "and after bhū", Bhaṭṭoji says. Then: bhavaḥ, the god and also worldly existence; bhāvāḥ, things. Both commentators note that this addition is not in the Bhāṣya. On the Bhāṣya's view, says Bhaṭṭoji, bhāvaḥ comes another way: ac after the cur-class bhū "to obtain", with ṇi.

Watch out. A vyavasthita option is not a free choice. You cannot call a planet grāha.

SK 2906 · गेहे कः · gehe kaḥ (A 3.1.144)

गेहे कः

gehe kaḥ · A 3.1.144

गेहे कर्तरि ग्रहेः कः स्यात् । गृह्णाति धान्यादिकमिति गृहम् । तात्स्थ्याद्गृहादाराः ॥

Rule. After grah, the affix ka comes when the agent is a house (geha). Bhaṭṭoji: "when the agent is a house". One commentator stresses that geha qualifies the agent; it is not an upapada. The other calls this rule an exception to SK 2905.

Example.

grah + ka → saṃprasāraṇa of r (SK 2412) → gṛ-a-h + a → the a merges into ṛ → gṛh + a → gṛham "house".

The steps are one commentator's. Bhaṭṭoji's gloss: "it takes in grain and the like".

Applies when / blocked by. The plural gṛhāḥ also means "wife". Bhaṭṭoji: this is because she is located in the house (tātsthya). Both commentators explain it as a transferred use.

Watch out. geha is not a word placed before the root. The house is the agent itself: gṛham, not gehagṛha.

I.F. Agents with a stated quality: craft, rice and time, repetition, blessing — running example: `nṛt + ṣvun → (vu → aka, guṇa) nartakaḥ → (ṣ: ṅīṣ) nartakī`

The last rules of pāda 3.1 give agent affixes only when the agent has some quality. SK 2907–2909 are for a craftsman (śilpin). SK 2910 is for rice or time. SK 2911 is for one who repeats an action, and SK 2912 is for a blessing. The running example is nṛt "to dance". With ṣvun it gives nartakaḥ "dancer". The ṣ of the affix gives the feminine ī (ṅīṣ): nartakī.

SK 2907 · शिल्पिनि ष्वुन् · śilpini ṣvun (A 3.1.145)

शिल्पिनि ष्वुन्

śilpini ṣvun · A 3.1.145

क्रियाकौशलं शिल्पं तद्वति ष्वुन् स्यात् ।

नृतिखनिरञ्जिभ्य एव

(वार्तिकम्) ॥ नर्तकः । नर्तकी । खनकः । खनकी ।

असि अकेऽने च रञ्जेर्नलोपो वाच्यः

(वार्तिकम्) ॥ रजकः । रजकी । भाष्यमते तु नृतिखनिभ्यामेव ष्वुन् । रञ्जेस्तु क्वुन्शिल्पिसंज्ञयोः <{उ190}> इति क्वुन् । टाप् । रजिका । पुंयोगे तु रजकी ॥

Rule. The affix ṣvun comes when the agent has a craft (śilpa). Bhaṭṭoji defines craft as skill in an action. One commentator stresses that śilpini qualifies the agent, as gehe did in SK 2906.

Example.

nṛt + ṣvun → markers go → nṛt + vu → aka (SK 1247)

→ guṇa of the light penultimate (SK 2189) → nart + aka → nartakaḥ.

Feminine: the ṣ gives ṅīṣ (one commentator) → nartakī.

khan → khanakaḥ, khanakī "digger".

Applies when / blocked by. A vārttika: only after nṛt, khan and rañj. One commentator draws the result: the weaver's root veñ takes ṇa instead (SK 2903), giving vāya.

A second vārttika: the n of rañj is lost before as, aka and ana: rajakaḥ, rajakī "dyer, washerman". One commentator says Bhaṭṭoji has restated here, by its sense, a vārttika found under SK 2396.

Bhaṭṭoji then gives the Bhāṣya's view. ṣvun comes only after nṛt and khan. rañj takes kvun by Uṇādi 190, and its feminine has ṭāp: rajikā. On that view rajakī means only "the dyer's wife" (puṃyoga, SK 504).

Watch out. ṣvun is not for every craftsman. On the vārttika it covers three roots, and on the Bhāṣya's view only two.

SK 2908 · गस्थकन् · gas thakan (A 3.1.146)

गस्थकन्

gasthakan · A 3.1.146

गायतेस्थकन् स्यात् । शिल्पिनि कर्तरि । गाथकः ॥

Rule. After gai "to sing", the affix thakan comes, for an agent with a craft. Bhaṭṭoji glosses gaḥ as gāyateḥ. One commentator explains why: the affix can only mean a craftsman whose craft is singing. So gāṅ "to go" is not meant.

Example.

gai + thakan → ā for ai (SK 2370) → gā + thaka → gāthakaḥ "singer".

Watch out. "Craft" runs on from SK 2907. The agent must have the craft of singing.

SK 2909 · ण्युट् च · ṇyuṭ ca (A 3.1.147)

ण्युट् च

ṇyuṭ ca · A 3.1.147

गायनः । टित्त्वाद्गायनी ॥

Rule. The affix ṇyuṭ also comes after gai, for a craftsman. Both commentators say why this is a separate rule (yogavibhāga). One sūtra gas thakan ṇyuṭ ca would do. But then thakan would run on too. Split off, only ṇyuṭ runs on into SK 2910.

Example.

gai + ṇyuṭ → ā for ai (SK 2370) → gā + yu → ana (SK 1247) → yuk (SK 2761) → gāy + ana → gāyanaḥ "singer".

Being ṭ-marked, it gives the feminine gāyanī (Bhaṭṭoji).

Applies when / blocked by. Both commentators give the yuk by SK 2761. One explains the order: ā for ai is not caused by the affix, so it comes before vṛddhi.

Watch out. gāthakaḥ and gāyanaḥ both mean a singer by craft, from different affixes. Only ṇyuṭ runs on into SK 2910, and only its ṭ gives a feminine in ī: gāyanī.

SK 2910 · हश्च व्रीहिकालयोः · haś ca vrīhikālayoḥ (A 3.1.148)

हश्च व्रीहिकालयोः

haśca vrīhikālayoḥ · A 3.1.148

हाको हाङश्च ण्युट् स्यात् व्रीहौ काले च कर्तरि । जहात्युदकमिति हायनो व्रीहिः । जहाति भावानिति हायनो वर्षम् । जिहीते प्राप्नोति वा ॥

Rule. After hā, ṇyuṭ also comes when the agent is rice (vrīhi) or time (kāla). Bhaṭṭoji: "after hāk and hāṅ". hāk means "to abandon" and hāṅ "to go".

Example.

hā + ṇyuṭ → hāyana, as with gāyanaḥ.

hāyano vrīhiḥ: rice, which leaves the water behind.

hāyano varṣam: the year, which leaves things behind, or reaches them.

One commentator explains the rice: it grows up above the water.

Watch out. The affix comes only for these two agents. Rice and time are not upapadas; they are the agents themselves.

SK 2911 · प्रुसृल्वः समभिहारे वुन् · prusṛlvaḥ samabhihāre vun (A 3.1.149)

प्रुसृल्वः समभिहारे वुन्

prusṛlvaḥ samabhihāre vun · A 3.1.149

समभिहारग्रहणेन साधुकारित्वं लक्ष्यते । प्रवकः । सरकः । लवकः ॥

Rule. After pru, sṛ and lū, the affix vun comes when repetition (samabhihāra) is meant. Bhaṭṭoji: the mention of repetition indicates "doing it well" (sādhukāritva). One commentator says the Bhāṣya makes this explicit.

Example.

lū + vun → lū + aka → guṇa (SK 2168) and av (SK 61) → lavakaḥ "a skilled reaper".

Also pravakaḥ, sarakaḥ.

Applies when / blocked by. One commentator explains the logic. A man who does a thing often becomes skilled at it. So vun fits one who does it well even once. It does not fit one who does it badly many times. The case-form of prusṛlvaḥ is explained differently by the two commentators; scholars discuss this.

Watch out. "Repetition" here really means skill. Do not use lavakaḥ for a clumsy man who cuts all day.

SK 2912 · आशिषि च · āśiṣi ca (A 3.1.150)

आशिषि च

āśiṣi ca · A 3.1.150

आशीर्विषयार्थवृत्तेर्धातोर्वुन् स्यात्कर्तरि । जीवतात्-जीवकः । नन्दतात् - नन्दकः । आशीः प्रयोक्तुर्धर्मः । आशासितुः पित्रादेरियमुक्तिः ॥

Rule. vun also comes in a blessing (āśis), in the agent sense. Bhaṭṭoji: after a root whose meaning is the object of a blessing. The blessing belongs to the speaker, such as a father.

Example.

jīvatāt "may you live" → jīv + vun → jīvakaḥ.

nandatāt "may you rejoice" → nandakaḥ.

Applies when / blocked by. One commentator explains why Bhaṭṭoji words it so. The blessing is the speaker's property, not the meaning of the affix. The affix still means the agent, by kartari kṛt (SK 2832). The same commentator gives the feminine jīvakā: the i of SK 463 does not come, by a separate statement.

jīvatāt is an imperative form of blessing; its tāt is allowed by SK 2197.

With this rule the first pāda of the third book ends. The other commentator marks the change.

Watch out. A jīvaka is not "one who lives" in general. It is said by one who blesses, such as a father, of the one he wishes to live.

I.G. aṇ with an object upapada, and the upapada compound — running example: `kumbha (karman) + kṛ + aṇ → (vṛddhi) kumbha + kāra → (upapada compound) kumbhakāraḥ`

With SK 2913 the second pāda of the third book begins. From here on most affixes come only when a word stands beside the root as upapada. The general affix is aṇ, with the object as upapada. The upapada and the kṛdanta are then compounded by SK 782. The running example is kumbhaṃ karotīti kumbhakāraḥ "one who makes pots: a potter". SK 2914 keeps aṇ for three roots that the next rule would give ka.

SK 2913 · कर्मण्यण् · karmaṇy aṇ (A 3.2.1)

कर्मण्यण्

karmaṇyaṇ · A 3.2.1

कर्मण्युपपदे धातोरण् प्रत्ययः स्यात् । उपपदसमासः । कुम्भं करोतीति कुम्भकारः । आदित्यं पश्यतीत्यादावनभिधानान्न ।

शीलिकामिभक्ष्याचरिभ्यो णः

(वार्तिकम्) ॥ अणोऽपवादार्थं वार्तिकम् । मांसशीला । मांसकामा । मांसभक्षा । कल्याणाचारा ॥

ईक्षिक्षमिभ्यां च

(वार्तिकम्) ॥ सुखप्रतीक्षा । बहुक्षामा । कथं तर्हि गङ्गाधरभूधरादयः । कर्मणः शेषत्वविवक्षायां भविष्यन्ति ॥

Rule. When an object (karman) is the upapada, the affix aṇ comes after a root. One commentator stresses that the affix still means the agent. Bhaṭṭoji: "[there is] the upapada compound".

Example.

kumbha (object) + kṛ + aṇ → marker ṇ goes → vṛddhi (SK 254) with r (SK 70) → kumbha + kāra

→ upapada compound (SK 782) → kumbhakāraḥ "potter".

One commentator: kumbha stands in the sixth case by SK 623 and is compounded with kāra.

Applies when / blocked by. Bhaṭṭoji: the affix does not come in ādityaṃ paśyati "he sees the sun", because such a form would not express the sense (anabhidhāna). One commentator adds the unwanted forms ādityadarśa and the like, and says the Bhāṣya makes this explicit. The same commentator separates two analyses. kumbhaṃ karoti is the ordinary paraphrase. The technical analysis joins kāra to kumbha before any case-ending is added.

Two vārttikas give ṇa instead of aṇ after śīli, kāmi, bhakṣi and ācari, and after īkṣi and kṣami:

māṃsaśīlā "given to meat", māṃsakāmā, māṃsabhakṣā, kalyāṇācārā "who practises good", sukhapratīkṣā, bahukṣāmā "who endures much".

Bhaṭṭoji calls the first vārttika an exception to aṇ. One commentator explains its point. With aṇ the feminine would take ṅīp (SK 470). With ṇa it takes ṭāp. The masculine is the same either way.

Why then gaṅgādhara, bhūdhara, and not gaṅgādhāra with aṇ? Bhaṭṭoji: they arise when the object is meant as a mere relation (śeṣa). Both commentators: then there is no object upapada, so no aṇ. One adds that dhara is formed by itself and compounded as "the holder of the Gaṅgā".

Watch out. aṇ needs an object word as upapada. Without one, kṛ takes ṇvul or tṛc (SK 2895): kārakaḥ.

SK 2914 · ह्वावामश्च · hvāvāmaś ca (A 3.2.2)

ह्वावामश्च

hvāvāmaśca · A 3.2.2

अण् स्यात् । कापवादः । स्वर्गह्वायः । तन्तुवायः । धान्यमायः ॥

Rule. aṇ also comes after hvā "to challenge", vā "to weave" and mā "to measure", with an object upapada. Bhaṭṭoji: it is the exception to ka. The ka is the one SK 2915 would give, both commentators say.

Example.

svarga + hve + aṇ → ā for e (SK 2370) → yuk (SK 2761) → svarga + hvāy + a → svargahvāyaḥ "one who challenges heaven".

tantu + ve + aṇ → tantuvāyaḥ "weaver of threads".

dhānya + mā + aṇ → dhānyamāyaḥ "measurer of grain".

Both commentators give the order: ā first, then yuk.

Applies when / blocked by. All three roots end in ā after SK 2370. So SK 2915 would give ka and lose the ā. This rule restores aṇ. Both commentators say mā here is māṅ "to measure" or meṅ "to exchange", which take an object. The intransitive mā is not meant.

Watch out. It looks odd to restate aṇ, which SK 2913 already gave. The point is order: this rule is the exception to an exception.

I.H. ka (and ḍa) with an upapada, and the ṣ of -ṣṭha — running example: `go + dā + ka → (ā-loss, SK 2372) godaḥ`; `sama + sthā + ka → samasthaḥ`; `dvi + stha → (ṣ) dviṣṭhaḥ`

ka comes after ā-final roots with an upapada, against aṇ. A vārttika turns ka into ḍa for roots that would take saṃprasāraṇa. SK 2916 is split in two, and its second half gives ka to sthā in the bare action. Two rules from the eighth book then give the ṣ of praṣṭhaḥ and dviṣṭhaḥ. SK 2919–2921 give ka after more roots. The running examples are godaḥ "giver of cows", samasthaḥ "standing on even ground" and dviṣṭhaḥ.

SK 2915 · आतोऽनुपसर्गे कः · āto 'nupasarge kaḥ (A 3.2.3)

आतोऽनुपसर्गे कः

āto'nupasarge kaḥ · A 3.2.3

आदन्ताद्धातोरनुपसर्गात्कर्मण्युपपदे कः स्यान्नाऽण् । आतो लोपः । गोदः । पार्ष्णित्रम् । अनुपसर्गे किम् । गोसंदायः ।

कविधौ सर्वत्र संप्रसारणिभ्यो डः

(वार्तिकम्) ॥ ब्रह्म जिनाति ब्रह्मज्यः । सर्वत्रग्रहणात् आतश्चोपसर्गे SK 2898 । आह्वः । प्रह्वः ॥

Rule. After a root ending in ā, with no upasarga, ka comes and not aṇ, when an object is the upapada. Bhaṭṭoji adds the result: "the ā is lost" (SK 2372).

Example.

go + dā + ka → ā lost before k-marked ka (SK 2372) → go + d + a → godaḥ "giver of cows".

pārṣṇi + trai + ka → ā for ai → ā lost → pārṣṇitram "heel-guard" (both commentators).

Applies when / blocked by. With an upasarga, aṇ: gosaṃdāyaḥ, with yuk (both commentators).

A vārttika: wherever ka is taught, ḍa comes instead after roots liable to saṃprasāraṇa.

brahma + jyā + ḍa → the ḍ removes the final vowel and what follows (ṭi) → brahmajyaḥ "one who harms the brahman".

Both commentators explain. This loss normally needs a stem called bha, but the ḍ-marker forces it here. ka, being k-marked, would bring saṃprasāraṇa (SK 2412) and give brahmajiya.

Bhaṭṭoji: because of "everywhere" (sarvatra), ḍa also replaces the ka of SK 2898: āhvaḥ, prahvaḥ. With ka there would be saṃprasāraṇa, giving āhuva (both commentators).

Watch out. Two conditions: an object upapada and no upasarga. Lose either and the rule fails: gosaṃdāyaḥ has an upasarga, so aṇ returns.

SK 2916 · सुपि स्थः · supi sthaḥ (A 3.2.4)

सुपि स्थः

supi sthaḥ · A 3.2.4

सुपीति योगो विभज्यते । सुपि उपपदे आदन्तात्कः स्यात् । द्वाभ्यां पिबतीति द्विपः । समस्थः । विषमस्थः । ततः - स्थः ॥ सुपि तिष्ठतेः कः स्यात् । आरम्भसामर्थ्याद्भावे । आखूनामुत्थानमाखूत्थः ॥

Rule. Bhaṭṭoji splits the rule (yogavibhāga). First, supi: with any word ending in a case-ending (sup) as upapada, ka comes after an ā-final root. Then, sthaḥ: with a sup-word as upapada, ka comes after sthā. This ka is in the bare action (bhāva). One commentator says the split is in the Bhāṣya.

Example.

dvābhyāṃ pibati → dvi + pā + ka → ā lost → dvipaḥ "one who drinks with two [organs]: an elephant".

sama + sthā + ka → samasthaḥ "standing on even ground"; viṣamasthaḥ.

ākhūnām utthānam → ākhu + ud + sthā + ka → ākhūtthaḥ "the rising up of rats".

One commentator gives the steps of ākhūtthaḥ. The ā is lost. The s after ud becomes th (SK 118). The d becomes t.

Applies when / blocked by. Why is the second part bhāva? The first part already gives ka after sthā in the agent sense. A second rule for the same would be pointless. So, by "the force of its being undertaken" (ārambha-sāmarthya), it is taken in another sense. Both commentators add that it also makes the compound obligatory.

One commentator adds why the first part is shown with dvābhyām. An upasarga upapada is covered by SK 2898, an object upapada by SK 2915. So Bhaṭṭoji chose a third-case word.

Watch out. ākhūtthaḥ is not "a rat that rises". It names the act of rising. One commentator says it is masculine, not neuter.

SK 2917 · प्रष्ठोऽग्रगामिनि · praṣṭho 'gragāmini (A 8.3.92)

प्रष्ठोऽग्रगामिनि

praṣṭho'gragāmini · A 8.3.92

प्रतिष्ठत इति प्रष्ठो गौः । अग्रतो गच्छतीत्यर्थः । अग्रेति किम् । प्रस्थः ॥

Rule. The word praṣṭha, with ṣ, is given by nipātana when "one that goes in front" is meant. This is a rule of the eighth book, set here because the form needs it.

Example.

pratiṣṭhate "it sets forth" → pra + sthā + ka → ṣ by this rule → praṣṭhaḥ: an ox that goes in front.

One commentator: prastha comes by SK 2898; this rule only adds the ṣ.

Applies when / blocked by. Without the sense "going in front": prasthaḥ.

Watch out. One commentator explains why a special rule is needed. The general rule that turns s into ṣ needs a sound of the iṇ group before it (any vowel but a and ā, or h, y, v, r, l), or a velar. The s here follows a, so that rule could not apply.

SK 2918 · अम्बाम्बगोभूमिसव्यापद्वित्रिकुशेकुशङ्क्वङ्गुमञ्जिपुञ्जिपरमेबर्हिर्दिव्यग्निभ्यः स्थः · ambāmbagobhūmisavyāpadvitrikuśekuśaṅkvaṅgumañjipuñjiparamebarhirdivyagnibhyaḥ sthaḥ (A 8.3.97)

अम्बाम्बगोभूमिसव्यापद्वित्रिकुशेकुशङ्क्वङ्गुमञ्जिपुञ्जिपरमेबर्हिर्दिव्यग्निभ्यः स्थः

ambāmbagobhūmisavyāpadvitrikuśekuśaṅkvaṅgumañjipuñjiparamebarhirdivyagnibhyaḥ sthaḥ · A 8.3.97

स्थ इति कप्रत्ययान्तस्यानुकरणम् । षष्ठ्यर्थे प्रथमा । एभ्यः स्थस्य सस्य षः स्यात् । द्विष्ठः । त्रिष्ठः । इत ऊर्ध्वं कर्मणि सुपीति द्वयमप्यनुवर्तते । तत्राकर्मकेषु सुपीत्यस्य संबन्धः ॥

Rule. After amba, āmba, go, bhūmi, savya, apa, dvi, tri, kuśe, ku, śaṅku, aṅgu, mañji, puñji, parame, barhis, divi and agni, ṣ replaces the s of stha. Bhaṭṭoji: stha here cites the form made with ka. It is a first-case word with the sense of the sixth: "of stha".

Example.

dvi + sthā + ka → dvi + stha → this rule → dviṣṭhaḥ. Also triṣṭhaḥ.

Both commentators list the rest: ambaṣṭhaḥ, goṣṭhaḥ, bhūmiṣṭhaḥ, parameṣṭhaḥ, agniṣṭhaḥ and others. (The ṭh follows the ṣ by SK 113.) The two commentators analyse dviṣṭhaḥ differently; scholars discuss this. The word savya is read savya or savye.

Applies when / blocked by. Only the ka-form stha is meant. Both commentators: so there is no ṣ in bhūmisthitam.

Bhaṭṭoji adds a key note for all that follows. From here on both karmaṇi (SK 2913) and supi (SK 2916) run on. With intransitive roots, only supi is connected. One commentator says "from here" means from SK 2919.

Watch out. This ṣ is listed word by word. It does not come after every first member: bhūmisthitam keeps its s.

SK 2919 · तुन्दशोकयोः परिमृजापनुदोः · tundaśokayoḥ parimṛjāpanudoḥ (A 3.2.5)

तुन्दशोकयोः परिमृजापनुदोः

tundaśokayoḥ parimṛjāpanudoḥ · A 3.2.5

तुन्दशोकयोः कर्मणोरुपपदयोराभ्यां कः स्यात् ।

आलस्यसुखाहरणयोरिति वक्तव्यम्

(वार्तिकम्) ॥ तुन्दं परिमार्ष्टीति तुन्दपरिमृजोऽलसः । शोकापनुदः । सुखस्याहर्ता । अलसादन्यत्र तुन्दपरिमार्ज एव । यश्च संसारासारत्वोपदेशेन शोकमपनुदति स शोकापनोदः ।

कप्रकरणे मूलविभुजादिभ्य उपसंख्यानम्

(वार्तिकम्) ॥ मूलानि विभुजति मूलविभुजो रथः । आकृतिगणोऽयम् । महीध्रः । कुध्रः । गिलतीति गिलः ॥

Rule. With tunda "belly" and śoka "grief" as objects, ka comes after parimṛj "to rub" and apanud "to dispel". A vārttika adds: only in the senses of laziness and of bringing happiness.

Example.

tundaṃ parimārṣṭi → tunda + parimṛj + ka → tundaparimṛjaḥ "a lazy man (who rubs his belly)".

śoka + apanud + ka → śokāpanudaḥ "a bringer of happiness".

Applies when / blocked by. Bhaṭṭoji: one who is not lazy is only tundaparimārjaḥ. Its vṛddhi shows the general aṇ. And one who dispels grief by teaching that saṃsāra is worthless is śokāpanodaḥ. Whether tundaparimṛjaḥ may also take vṛddhi, the commentators report differently; scholars discuss this.

A vārttika: in the section on ka, add mūlavibhuja and the like. mūlāni vibhujati → mūlavibhujaḥ "a chariot (it crushes roots)". Bhaṭṭoji: this is an open list (ākṛtigaṇa): mahīdhraḥ, kudhraḥ, and gilaḥ "one who swallows". One commentator explains mahīdhraḥ: no guṇa before k-marked ka, and ṛ becomes r; with aṇ it would be mahīdhāra.

Watch out. The sense matters, not just the words. Only the lazy man is tundaparimṛja; anyone else is tundaparimārja.

SK 2920 · प्रे दाज्ञः · pre dājñaḥ (A 3.2.6)

प्रे दाज्ञः

pre dājñaḥ · A 3.2.6

दारूपाज्जानातेश्च प्रोपसृष्टात्कर्मण्युपपदे कः स्यात् । अणोऽपवादः । सर्वप्रदः । पथिप्रज्ञः । अनुपसर्ग इत्युक्तेः प्रादन्यस्मिन्सति न कः । गोसंप्रदायः ॥

Rule. After a root of the form dā, and after jñā, preceded by pra, ka comes with an object upapada. Bhaṭṭoji: "it is an exception to aṇ". One commentator explains "of the form dā": rules naming gā, mā and dā make no distinction among the roots of that form.

Example.

sarva + pra + dā + ka → ā lost → sarvapradaḥ "giver of all".

pathi + pra + jñā + ka → pathiprajñaḥ "one who knows the road well".

Applies when / blocked by. SK 2915 said "when there is no upasarga". So, Bhaṭṭoji says, when an upasarga other than pra is present there is no ka: gosaṃpradāyaḥ. One commentator: this rule covers only pra alone; with another upasarga even SK 2915 cannot give ka.

Watch out. pra opens the door that SK 2915 closed. Add sam as well, and the door shuts again: gosaṃpradāyaḥ takes aṇ.

SK 2921 · समि ख्यः · sami khyaḥ (A 3.2.7)

समि ख्यः

sami khyaḥ · A 3.2.7

गोसंख्यः ॥

Rule. After khyā preceded by sam, with an object upapada, the affix comes. Bhaṭṭoji names no affix here, and neither commentator does. It is taken as ka, running on from SK 2915 through SK 2920.

Example.

gāḥ saṃcaṣṭe "he watches over the cows" → go + sam + khyā + ka → ā lost → gosaṃkhyaḥ "one who watches over the cows".

Both commentators say khyā here is khyāñ, the substitute for cakṣ "to see".

Applies when / blocked by. Both commentators say the other khyā, "to tell", is not meant. It is not used with sam, and it takes only sārvadhātuka affixes.

Watch out. The sūtra says only "after khyā with sam". The object upapada comes from the note at SK 2918.

I.I. ṭak after gā and pā; ac after hṛ, grah, arh, ram and jap, śam, śī — running example: `sāman + gai + ṭak → (ā-loss) sāmagaḥ → (ṭ: ṅīp) sāmagī`; `aṃśa + hṛ + ac → (guṇa) aṃśaharaḥ` (against aṇ in `bhārahāraḥ`)

Two more affixes take the place of aṇ or ka with an upapada. ṭak is a with the markers ṭ and k: the k removes a root-final ā, and the ṭ gives the feminine ī. ac is plain a. It brings guṇa of a final vowel, but not vṛddhi, so its forms differ from those of aṇ. The running examples are sāmagaḥ "singer of the Sāman" with its feminine sāmagī, and aṃśaharaḥ "one who takes a share" against bhārahāraḥ "one who lifts a load".

SK 2922 · गापोष्टक् · gāpoṣ ṭak (A 3.2.8)

गापोष्टक्

gāpoṣṭak · A 3.2.8

अनुपसृष्टाभ्यामाभ्यां टक् स्यात्कर्मण्युपपदे । सामगः । सामगी । उपसर्गे तु सामसंगायः ॥

पिबतेः सुराशीध्वोरिति वाच्यम्

(वार्तिकम्) ॥ सुरापी । शीधुपी । अन्यत्र क्षीरपा ब्राह्मणी । सुरां पाति रक्षतीति सुरापा ॥

Rule. After gā "to sing" and pā "to drink", with no upasarga, ṭak comes when an object is the upapada. One commentator says only gai "to sing" is meant here, not the other roots of the form gā.

Example.

sāman + gai + ṭak → ā for ai (SK 2370) → sāma + gā + a → ā lost before the k (SK 2372) → sāmagaḥ.

Feminine: the ṭ gives ṅīp (SK 470) → sāmagī.

One commentator draws this from the two forms: the k is for the loss of ā, the ṭ for ṅīp.

Applies when / blocked by. With an upasarga: sāmasaṃgāyaḥ, with aṇ and yuk (both commentators).

A vārttika: after pibati "drinks", ṭak comes only with surā or śīdhu as upapada: surāpī, śīdhupī. Elsewhere: kṣīrapā brāhmaṇī "a brāhmaṇī who drinks milk". Both commentators give this ka by SK 2915, then ṭāp. And from pā "to protect": surāṃ pāti → surāpā "one who guards the surā".

Watch out. The feminine tells you the affix. surāpī has ṭak, with ī. kṣīrapā has ka, with ā.

SK 2923 · हरतेरनुद्यमनेऽच् · harater anudyamane 'c (A 3.2.9)

हरतेरनुद्यमनेऽच्

harateranudyamane'c · A 3.2.9

अंशहरः । अनुद्यमने किम् । भारहारः ।

शक्तिलाङ्गलाङ्कुशतोमरयष्टिघटघटीधनुष्षु ग्रहेरुपसंख्यानम्

(वार्तिकम्) ॥ शक्तिग्रहः । लाङ्गलग्रहः । सूत्रे च धार्येऽर्थे ॥ सूत्रग्रहः । यस्तु सूत्रं केवलमुपादत्ते न तु धारयति तत्राणेव । सूत्रग्राहः ॥

Rule. After hṛ "to take", when it does not mean lifting up (udyamana), ac comes with an object upapada.

Example.

aṃśa + hṛ + ac → guṇa (SK 2168) with r → aṃśa + har + a → aṃśaharaḥ "one who takes a share".

Lifting up: bhāra + hṛ + aṇ → vṛddhi → bhārahāraḥ "one who lifts a load".

Applies when / blocked by. A vārttika adds ac after grah with śakti, lāṅgala, aṅkuśa, tomara, yaṣṭi, ghaṭa, ghaṭī or dhanus: śaktigrahaḥ, lāṅgalagrahaḥ. One commentator: ac is not k-marked, so there is no saṃprasāraṇa.

A second vārttika adds sūtra "thread", in the sense of holding: sūtragrahaḥ. Bhaṭṭoji: one who merely takes up the thread and does not hold it takes aṇ: sūtragrāhaḥ.

One commentator notes why both ghaṭa and ghaṭī are listed. ghaṭa alone would have been enough. So the listing shows that a stem does not always include its feminine form. Bhaṭṭoji uses this at SK 2955.

Watch out. The difference is one vowel. ac gives -hara, -graha; aṇ gives -hāra, -grāha with vṛddhi.

SK 2924 · वयसि च · vayasi ca (A 3.2.10)

वयसि च

vayasi ca · A 3.2.10

उद्यमनार्थं सूत्रम् । कवचहरः कुमारः ॥

Rule. ac also comes after hṛ when age (vayas) is conveyed. Bhaṭṭoji: "the sūtra is for lifting up". That is, here ac comes even in the sense of lifting, which SK 2923 left out.

Example.

kavaca + hṛ + ac → kavacaharaḥ kumāraḥ "a boy old enough to bear armour".

Applies when / blocked by. One commentator raises the objection: SK 2923 already gives ac, so why this rule? Bhaṭṭoji's answer is that SK 2923 excludes lifting; this rule brings it back for age. The other commentator explains the age. The armour may be actually lifted, or only thought possible to lift. Either way the word conveys the age.

Watch out. Without age, lifting gets aṇ: bhārahāraḥ. With age, lifting gets ac: kavacaharaḥ.

SK 2925 · आङि ताच्छील्ये · āṅi tācchīlye (A 3.2.11)

आङि ताच्छील्ये

āṅi tācchīlye · A 3.2.11

पुष्पाण्याहरति तच्छीलः पुष्पाहरः । ताच्छील्ये किम् । भारहारः ॥

Rule. After hṛ preceded by āṅ (ā), ac comes when habit (tācchīlya) is meant. One commentator: habit means having that as one's nature.

Example.

puṣpāṇy āharati, being of that habit → puṣpa + ā + hṛ + ac → puṣpāharaḥ "one who brings flowers by habit".

One commentator: one who does it by nature, not for a reward.

Applies when / blocked by. Without habit: bhārahāraḥ. Bhaṭṭoji's counterexample has the same form as at SK 2923. Some copies read bhārāhāraḥ; scholars discuss this.

Watch out. Both āṅ and habit are needed for this ac. Bhaṭṭoji's counterexample, bhārahāraḥ, has the vṛddhi of aṇ.

SK 2926 · अर्हः · arhaḥ (A 3.2.12)

अर्हः

arhaḥ · A 3.2.12

अर्हतेरच् स्यात्कर्मण्युपपदे । अणोपवादः । पूजार्हा ब्राह्मणी ॥

Rule. After arh "to deserve", ac comes with an object upapada. Bhaṭṭoji: "it is an exception to aṇ".

Example.

pūjā + arh + ac → pūjārha → feminine with ṭāp → pūjārhā brāhmaṇī "a brāhmaṇī deserving honour".

Applies when / blocked by. One commentator explains why the rule is needed. arh has no penultimate a to take vṛddhi. So the masculine is the same with aṇ or ac. But an aṇ-stem would take ṅīp in the feminine. ac is taught so that the feminine is pūjārhā. That is why Bhaṭṭoji gives a feminine example.

Watch out. Do not write pūjārhī. The rule exists to give ā.

SK 2927 · स्तम्बकर्णयो रमिजपोः · stambakarṇayo ramijapoḥ (A 3.2.13)

स्तम्बकर्णयो रमिजपोः

stambakarṇayo ramijapoḥ · A 3.2.13

।

हस्तिसूचकयोरिति वक्तव्यम्

(वार्तिकम्) ॥ स्तम्बे रमते स्तम्बेरमो हस्ती । तत्पुरुषे कृति - SK 972 इति, 'हलदन्तात्' SK 966 इति वा ङेरलुक् । कर्णेजपः सूचकः ॥

Rule. With stamba "clump of grass" and karṇa "ear" as upapada, the affix comes after ram "to delight" and jap "to whisper". The affix is not named; it is taken as the ac of SK 2923. One commentator says ram and jap are intransitive, so only supi is connected, as SK 2918 directs. A vārttika: only when an elephant or an informer is meant.

Example.

stambe ramate → stambe + ram + ac → stamberamaḥ "an elephant (it delights in grass clumps)".

karṇe + jap + ac → karṇejapaḥ "an informer (he whispers in the ear)".

Bhaṭṭoji: the seventh-case ending -e is kept (aluk) by SK 972 or by SK 966.

Applies when / blocked by. Otherwise there is no such compound. Both commentators give stambe rantā and karṇe japitā, with tṛc. One commentator applies them to a cow, and to a teacher or a mosquito.

Watch out. The first member keeps its case-ending: stambe-, karṇe-. Do not write stambarama.

SK 2928 · शमि धातोः संज्ञायाम् · śami dhātoḥ saṃjñāyām (A 3.2.14)

शमि धातोः संज्ञायाम्

śami dhātoḥ saṃjñāyām · A 3.2.14

शम्भवः । शंवदः । पुनर्धातुग्रहणं बाधकविषयेऽपि प्रवृत्त्यर्थम् । कृञो हेत्वादिषु टो मा भूत् । शंकरा नाम परिव्राजिका तच्छीला ॥

Rule. With śam "happiness" as upapada, ac comes after any root, when a name (saṃjñā) is meant. Both commentators supply ac. One says "after any root whatever".

Example.

śam + bhū + ac → guṇa and av → śambhavaḥ.

śam + vad + ac → śaṃvadaḥ.

Applies when / blocked by. Why repeat "dhātu", when every affix here comes after a dhātu? Bhaṭṭoji: so that this rule applies even where a blocking rule would apply. The blocking rule is SK 2934, which gives ṭa after kṛ when cause, habit or compliance is conveyed. So śaṃkarā takes ac: a female wandering ascetic, whose habit is "making happiness". One commentator explains. Without "dhātu", SK 2934 would win in śaṃkara as the later rule. With it, ac alone comes.

Watch out. śaṃkarā looks like a habit-word, which would call for ṭa. The repeated "dhātu" keeps ac there.

SK 2929 · अधिकरणे शेतेः · adhikaraṇe śeteḥ (A 3.2.15)

अधिकरणे शेतेः

adhikaraṇe śeteḥ · A 3.2.15

खे शेते खशयः ।

पार्श्वादिषूपसंख्यानम्

(वार्तिकम्) ॥ पार्श्वाभ्यां शेते पार्श्वशयः । पृष्ठशयः । उदरेण शेते उदरशयः । उत्तानादिषु कर्तृषु ॥ उत्तानः शेते उत्तानशयः । अवमूर्धशयः । अवनतो मूर्धा यस्य सः अवमूर्धा । अधोमुखः शेते इत्यर्थः ।

गिरौ डश्छन्दसि

(वार्तिकम्) ॥ गिरौ शेते गिरिशः । लोके गिरिशयः । कथं तर्हि गिरिशमुपचचार प्रत्यहं सा सुकेशीति । गिरिरस्यास्तीति विग्रहे लोमादित्वाच्छः ॥

Rule. With a word for a locus (adhikaraṇa) as upapada, ac comes after śī "to lie". One commentator supplies ac.

Example.

khe śete "he lies in the sky" → kha + śī + ac → guṇa and ay → khaśayaḥ.

Applies when / blocked by. Three vārttikas follow.

Add pārśva and the like: pārśvābhyāṃ śete → pārśvaśayaḥ "one who lies on his sides"; pṛṣṭhaśayaḥ; udareṇa śete → udaraśayaḥ. Both commentators: these are not loci, as the third-case analyses show.
With uttāna and the like as agents: uttānaḥ śete → uttānaśayaḥ "one who lies on his back"; avamūrdhaśayaḥ "one who lies face down".
ḍa with giri, in the Veda: girau śete → giriśaḥ. Some copies of Bhaṭṭoji's text add that ordinary language has giriśayaḥ. One commentator says ordinary language has ac here, not ḍa.

How then the line giriśam upacacāra pratyahaṃ sā sukeśī "every day she, the fair-haired one, served Giriśa", outside the Veda? Bhaṭṭoji: there giriśa means "who has a mountain", with the taddhita śa, since giri is in the loma group.

Watch out. Two words giriśa come by different routes. The Vedic one is śī + ḍa; the poet's is giri + taddhita śa.

I.J. ṭa after car and sṛ — running example: `kuru (locus) + car + ṭa → kurucaraḥ → (ṭ: ṅīp) kurucarī`

The affix ṭa is a with the marker ṭ. Like ac it brings no vṛddhi. Its ṭ gives the feminine ī by SK 470. SK 2930–2933 give it after car "to move" and sṛ "to go". The running example is kurucaraḥ "one who roams among the Kurus", feminine kurucarī.

SK 2930 · चरेष्टः · careṣ ṭaḥ (A 3.2.16)

चरेष्टः

careṣṭaḥ · A 3.2.16

अधिकरणे उपपदे । कुरुचरः । कुरुचरी ॥

Rule. After car, ṭa comes when a locus is the upapada. Bhaṭṭoji supplies "when a locus is the upapada". One commentator: "locus" runs on from SK 2929.

Example.

kuruṣu carati → kuru + car + ṭa → kurucaraḥ "one who roams among the Kurus".

Feminine: the ṭ gives ṅīp (SK 470) → kurucarī.

Applies when / blocked by. One commentator warns against taking Kuru as the object. The rule that would make a place the object of an intransitive root is optional; and the Bhāṣya on this rule supports the locus reading.

Watch out. The feminine is kurucarī, with ī. That ī comes from the ṭ of the affix.

SK 2931 · भिक्षासेनादायेषु च · bhikṣāsenādāyeṣu ca (A 3.2.17)

भिक्षासेनादायेषु च

bhikṣāsenādāyeṣu ca · A 3.2.17

भिक्षां चरतीति भिक्षाचरः । सेनाचरः । आदायेति ल्यबन्तम् । आदायचरः । कथं प्रेक्ष्य स्थितां सहचरीमिति । पचादिषु चरडिति पाठात् ॥

Rule. ṭa also comes after car with bhikṣā "alms", senā "army" or ādāya "having taken" as upapada. One commentator: the rule is made for words that are not loci.

Example.

bhikṣāṃ carati → bhikṣā + car + ṭa → bhikṣācaraḥ "one who goes about for alms".

senācaraḥ. The two commentators gloss it differently; scholars discuss this.

ādāyacaraḥ. Bhaṭṭoji: ādāya is a form in -ya (lyap).

Applies when / blocked by. How then sahacarī "companion", in prekṣya sthitāṃ sahacarīm? saha is none of the listed words. Bhaṭṭoji: because caraṭ is read in the paca-list (SK 2896). Both commentators add a point. The pacādi ac would also give bhikṣācara. But ac would compound only optionally. The special ṭa makes the compound obligatory.

Watch out. The feminine sahacarī has ī because caraṭ is ṭ-marked, like ṭa.

SK 2932 · पुरोऽग्रतोऽग्रेषु सर्तेः · puro 'grato 'greṣu sarteḥ (A 3.2.18)

पुरोऽग्रतोऽग्रेषु सर्तेः

puro'grato'greṣu sarteḥ · A 3.2.18

पुरस्सरः । अग्रतसरः । अग्रमग्रेणाग्रे वा सरतीत्यग्रेसरः । सूत्रेऽग्रे इति एदन्तत्वमपि निपात्यते । कथं तर्हि यूथं तदग्रसरगर्वितकृष्णसारमिति । बाहुलकादिति हरदत्तः ॥

Rule. With puras, agratas or agre as upapada, ṭa comes after sṛ "to go". One commentator supplies ṭa.

Example.

puras + sṛ + ṭa → purassaraḥ "one who goes in front". Also agratassaraḥ.

agram agreṇa agre vā sarati → agresaraḥ.

Applies when / blocked by. Bhaṭṭoji: the e of agre in the sūtra is also given by nipātana. One commentator explains. In a compound the case-ending would be lost, giving agrasara. The nipātana keeps the e.

How then agrasara in a poet's line, yūthaṃ tad agrasaragarvitakṛṣṇasāram? Bhaṭṭoji quotes Haradatta: "by bāhulaka", the principle that rules apply variously.

Watch out. Write agresaraḥ with e. The form agrasara is accepted only through bāhulaka.

SK 2933 · पूर्वे कर्तरि · pūrve kartari (A 3.2.19)

पूर्वे कर्तरि

pūrve kartari · A 3.2.19

कर्तृवाचिनि पूर्वशब्दे उपपदे सर्तेष्टः स्यात् । पूर्वः सरतीति पूर्वसरः । कर्तरि किम् । पूर्वं देशं सरतीति पूर्वसारः ॥

Rule. When the word pūrva "first", denoting the agent, is the upapada, ṭa comes after sṛ. One commentator: kartari here means "a word denoting the agent".

Example.

pūrvaḥ sarati "he goes first" → pūrva + sṛ + ṭa → pūrvasaraḥ.

Applies when / blocked by. Why "agent"? Bhaṭṭoji: pūrvaṃ deśaṃ sarati "he goes to the eastern region" → pūrvasāraḥ. Here pūrva is the object. Bhaṭṭoji names no affix; the vṛddhi is that of the general aṇ.

Watch out. This upapada is the agent, not the object. The same word as object gives -sāra.

I.K. ṭa after kṛ — running example: `yaśas + kṛ + ṭa (hetu) → (s for visarga, SK 160) yaśaskara → (ṭ: ṅīp) yaśaskarī vidyā`

ṭa now comes after kṛ, against the aṇ of kumbhakāraḥ. SK 2934 gives it when cause, habit or compliance is conveyed. SK 2935 gives it with twenty-seven named words, whatever the sense. SK 2936 gives it with karman for a hireling. SK 2937 blocks it with nine words. The running example is yaśaskarī vidyā "learning that brings fame".

SK 2934 · कृञो हेतुताच्छील्यानुलोम्येषु · kṛño hetutācchīlyānulomyeṣu (A 3.2.20)

कृञो हेतुताच्छील्यानुलोम्येषु

kṛño hetutācchīlyānulomyeṣu · A 3.2.20

एषु द्योत्येषु करोतेष्टः स्यात् । अतः कृकमि SK 160 इति सः । यशस्करी विद्या । श्राद्धकरः । वचनकरः ॥

Rule. After kṛ "to do", ṭa comes when cause (hetu), habit (tācchīlya) or compliance (ānulomya) is conveyed. Bhaṭṭoji: "when these are conveyed (dyotya)". One commentator adds that an object upapada is understood.

Example.

yaśas + kṛ + ṭa → s for the visarga by SK 160 → yaśaskara → feminine ṅīp → yaśaskarī vidyā "learning that is a cause of fame".

śrāddhakaraḥ "one in the habit of performing the śrāddha".

vacanakaraḥ "one who complies with (his elders') words".

Applies when / blocked by. Both commentators explain "conveyed": the affix still means the agent, and the three senses are only signalled. One adds that if cause and the rest were the meaning of the affix, kartari kṛt (SK 2832) would be overridden. The same commentator gives the counterexample kumbhakāraḥ, with none of the three senses.

That commentator also says hetu here is an ordinary cause, not the technical "causer" of the causative. And the kṛ meant is the familiar karoti "to do", though its citation form has two markers. Another root kṛ "to injure" is not meant.

Watch out. These three senses are signalled, not named by the affix. yaśaskarī is still "she who makes fame", an agent.

SK 2935 · दिवाविभानिशाप्रभाभास्कारान्तानन्तादिबहुनान्दीकिम्लिपिलिबिबलिभक्तिकर्तृचित्रक्षेत्रसंख्याजङ्घाबाह्वयत्तद्धनुररुष्षु · divāvibhāniśāprabhābhāskārāntānantādibahunāndīkimlipilibibalibhaktikartṛcitrakṣetrasaṃkhyājaṅghābāhvayattaddhanuraruṣṣu (A 3.2.21)

दिवाविभानिशाप्रभाभास्कारान्तानन्तादिबहुनान्दीकिम्लिपिलिबिबलिभक्तिकर्तृचित्रक्षेत्रसंख्याजङ्घाबाह्वयत्तद्धनुररुष्षु

divāvibhāniśāprabhābhāskārāntānantādibahunāndīkimlipilibibalibhaktikartṛcitrakṣetrasaṃkhyājaṅghābāhvayattaddhanuraruṣṣu · A 3.2.21

एषु कृञष्टः स्याद् अहेत्वादावपि । दिवाकरः । विभाकरः । निशाकरः । कस्कादित्वात्सः । भास्करः । बहुकरः । बहुशब्दस्य वैपुल्यार्थं संख्यापेक्षया पृथग्ग्रहणम् । लिपिलिबिशब्दौ पर्यायौ । संख्या । एककरः । द्विकरः । कस्कादित्वादहस्करः । नित्यं समासेऽनुत्तरपदस्थस्य SK 159 इति षत्वम् । धनुष्करः । अरुष्करः ।

किंयत्तद्बहुषु कृञोऽज्विधानम्

(वार्तिकम्) ॥ इति वार्तिकम् । किंकरा । यत्करा । तत्करा । हेत्वादौ टं बाधित्वा परत्वादच् । पुंयोगे ङीष् । किंकरी ॥

Rule. With any of these as upapada, ṭa comes after kṛ, even when cause and the rest are not conveyed: divā, vibhā, niśā, prabhā, bhās, kāra, anta, ananta, ādi, bahu, nāndī, kim, lipi, libi, bali, bhakti, kartṛ, citra, kṣetra, a numeral, jaṅghā, bāhu, ahar, yad, tad, dhanus, arus. The word ahar is restored from the printed text; Bhaṭṭoji's own example ahaskaraḥ needs it. One commentator: the upapadas are construed as object or as plain sup-word, each as fits.

Example.

divākaraḥ, vibhākaraḥ, niśākaraḥ.

bhās + kṛ + ṭa → s for the visarga, because bhās is in the kaska list → bhāskaraḥ. Also ahaskaraḥ.

bahukaraḥ; numerals: ekakaraḥ, dvikaraḥ.

dhanus + kṛ + ṭa → ṣ by SK 159 → dhanuṣkaraḥ. Also aruṣkaraḥ.

Applies when / blocked by. Bhaṭṭoji explains two items. bahu is listed apart from "numeral" for its sense "abundance". lipi and libi are synonyms.

A vārttika: ac, not ṭa, after kṛ with kim, yad, tad or bahu: kiṃkarā, yatkarā, tatkarā. Bhaṭṭoji: even where cause and the rest are conveyed, ac sets aside ṭa, because it is later (paratva). So the feminine has ā. How then kiṃkarī? Bhaṭṭoji: it is ṅīṣ for "the wife of" (puṃyoga, SK 504). Other views on this vārttika exist; scholars discuss this.

The glosses of divākaraḥ differ between the two commentators; scholars discuss this.

Watch out. kiṃkarā has ac, so its feminine is in ā. kiṃkarī is "the wife of a kiṃkara".

SK 2936 · कर्मणि भृतौ · karmaṇi bhṛtau (A 3.2.22)

कर्मणि भृतौ

karmaṇi bhṛtau · A 3.2.22

कर्मशब्दे उपपदे करोतेष्टः स्यात् । कर्मकरो भृतकः । कर्मकारोऽन्यः ॥

Rule. When the word karman "work" itself is the upapada, ṭa comes after kṛ, when wages (bhṛti) are meant. Both commentators explain why the word is named. karmaṇi "object" already runs on. So naming karman again means the word karman itself.

Example.

karman + kṛ + ṭa → karmakaraḥ "a hired labourer".

Applies when / blocked by. Of another, not a hireling: karmakāraḥ. Bhaṭṭoji names no affix there; its ā is that of the general aṇ.

Watch out. karmaṇi here is not "with an object". It is the word karman.

SK 2937 · न शब्दश्लोककलहगाथावैरचाटुसूत्रमन्त्रपदेषु · na śabdaślokakalahagāthāvairacāṭusūtramantrapadeṣu (A 3.2.23)

न शब्दश्लोककलहगाथावैरचाटुसूत्रमन्त्रपदेषु

na śabdaślokakalahagāthāvairacāṭusūtramantrapadeṣu · A 3.2.23

एषु कृञष्टो न । हेत्वादिषु प्राप्तः प्रतिषिध्यते । शब्दकार इत्यादि ॥

Rule. With śabda, śloka, kalaha, gāthā, vaira, cāṭu, sūtra, mantra or pada as upapada, ṭa does not come after kṛ. Bhaṭṭoji: the ṭa that would come in cause and the rest is forbidden. One commentator names it: the ṭa of SK 2934.

Example.

śabda + kṛ → no ṭa → śabdakāraḥ "one who makes sounds", and so on.

Applies when / blocked by. Bhaṭṭoji does not name the affix of śabdakāraḥ. Its ā shows the general aṇ of SK 2913.

Watch out. Even a man who habitually makes verses is ślokakāraḥ, not ślokakaraḥ.

I.L. in after kṛ, hṛ, grah and bhṛ — running example: `stamba + kṛ + in → (guṇa) stambakariḥ vrīhiḥ`; `ātman + bhṛ + in → (mum by nipātana) ātmambhariḥ`

The affix in is i with the marker n. It brings guṇa, and its words end in -i. SK 2938–2939 give it with named objects and named agents. SK 2940 gives two finished words by nipātana. The running examples are stambakariḥ vrīhiḥ "rice, which makes clumps" and ātmambhariḥ "one who feeds only himself".

SK 2938 · स्तम्बशकृतोरिन् · stambaśakṛtor in (A 3.2.24)

स्तम्बशकृतोरिन्

stambaśakṛtorin · A 3.2.24

।

व्रीहिवत्सयोरिति वक्तव्यम्

(वार्तिकम्) ॥ स्तम्बकरिर्व्रीहिः । शकृत्करिर्वत्सः । व्रीहिवत्सयोः किम् । स्तम्बकारः । शकृत्कारः ॥

Rule. With stamba "clump" or śakṛt "dung" as object, in comes after kṛ. One commentator supplies "object" and "after kṛ". A vārttika: only when rice (vrīhi) or a calf (vatsa) is the agent. One commentator pairs them in order.

Example.

stamba + kṛ + in → guṇa with r → stamba + kari → stambakarir vrīhiḥ "rice, which makes clumps".

śakṛtkarir vatsaḥ "a calf, which makes dung".

Applies when / blocked by. Why "rice or calf"? Bhaṭṭoji: otherwise stambakāraḥ, śakṛtkāraḥ, with the vṛddhi of aṇ. One commentator adds that the n of in gives the second member its own first-syllable accent.

Watch out. Rice and calf are agents, not upapadas. The upapadas are stamba and śakṛt.

SK 2939 · हरतेर्दृतिनाथयोः पशौ · harater dṛtināthayoḥ paśau (A 3.2.25)

हरतेर्दृतिनाथयोः पशौ

haraterdṛtināthayoḥ paśau · A 3.2.25

दृतिनाथयोरुपपदयोर्हृञ इन् स्यात्पशौ कर्तरि । दृतिं हरतीति दृतिहरिः । नाथं नासारज्जुं हरतीति नाथहरिः । पशौ किम् । दृतिहारः । नाथहारः ॥

Rule. With dṛti "leather bag" or nātha "nose-rope" as upapada, in comes after hṛ when an animal (paśu) is the agent. Bhaṭṭoji names the affix in in his gloss.

Example.

dṛtiṃ harati → dṛti + hṛ + in → guṇa → dṛtihariḥ "an animal that carries off a leather bag".

nāthaṃ harati → nāthahariḥ "an animal that bears a nose-rope".

One commentator says nāthahari names a particular animal fitted with a nose-rope.

Applies when / blocked by. Not an animal: dṛtihāraḥ, nāthahāraḥ, with vṛddhi.

Watch out. Compare stambakariḥ (SK 2938). There the agent was rice or a calf; here it is any animal.

SK 2940 · फलेग्रहिरात्मम्भरिश्च · phalegrahir ātmambhariś ca (A 3.2.26)

फलेग्रहिरात्मम्भरिश्च

phalegrahirātmambhariśca · A 3.2.26

फलानि गृह्णाति फलेग्रहिः । उपपदस्य एदन्तत्वं ग्रहेरिन्प्रत्ययश्च निपात्यते । आत्मानं बिभर्तीति आत्मम्भरिः । आत्मनो मुमागमः । भृञ इन् । चात्कुक्षिम्भरिः । चान्द्रास्तु आत्मोदरकुक्षिष्विति पेठुः । ज्योत्स्नाकरम्भमुदरम्भरयश्चकोरा इति मुरारिः ॥

Rule. The words phalegrahi and ātmambhari are given by nipātana. "And" (ca) adds others not stated. Bhaṭṭoji names what each nipātana teaches.

Example.

phalāni gṛhṇāti → phala + grah + in → the upapada ends in e by nipātana → phalegrahiḥ "a tree that bears fruit".

ātmānaṃ bibharti → ātman + bhṛ + in → augment mum on ātman by nipātana → ātmambhariḥ "one who supports only himself".

One commentator: the n of ātman is lost first, and m comes after the a.

Applies when / blocked by. By ca, says Bhaṭṭoji: kukṣimbhariḥ "one who fills only his belly". One commentator notes that the Bhāṣya gives it as a separate statement. The Cāndra grammarians read the rule "with ātman, udara or kukṣi". Hence the poet Murāri's udarambhari, in a line on cakora birds feeding on moonlight. One commentator calls a current verse's kukṣimbharāḥ a mistake.

Watch out. The e of phale- and the m of ātmam- are not from any general rule. They are fixed by this sūtra alone.

I.M. khaś, with the short vowel and mum before a khit uttarapada — running example: `jana + ej-i + khaś → (śap, guṇa, ay, pararūpa) jana-ejaya → (sup-luk, mum) janamejayaḥ`; `śunī + dhaya → (hrasva) śuni-dhaya → (mum) śunindhayaḥ`

The affix khaś is a with two markers. Its ś makes it sārvadhātuka, so the root takes śap and its present stem, as with śa (group I.D). Its kh acts on the first member of the compound (the pūrvapada). Two rules from the sixth book are set here for that. SK 2943 shortens a long final vowel of the pūrvapada. SK 2942 then adds the augment m (mum) after it. The running examples are janamejayaḥ "one who makes people tremble" and śunindhayaḥ "a puppy sucking a bitch".

SK 2941 · एजेः खश् · ejeḥ khaś (A 3.2.28)

एजेः खश्

ejeḥ khaś · A 3.2.28

ण्यन्तादेजेः खश् स्यात् ॥

Rule. After eji, that is ej "to tremble" with causative ṇi, the affix khaś comes. Bhaṭṭoji: "after ej ending in ṇic". One commentator adds that an object upapada is understood.

Example. Bhaṭṭoji gives the example under the next rule: janamejayaḥ (SK 2942).

Applies when / blocked by. One commentator proves that the causative stem is meant. khaś is ś-marked so that śap may come. For the plain root ej, śap would make no difference to the form. So the marker shows that eji is meant.

Watch out. eji is not ej with a citation i. It is the causative stem, "to make tremble".

SK 2942 · अरुर्द्विषदजन्तस्य मुम् · arurdviṣadajantasya mum (A 6.3.67)

अरुर्द्विषदजन्तस्य मुम्

arurdviṣadajantasya mum · A 6.3.67

अरुषोद्विषतोऽजन्तस्य च मुमागमः स्यात् खिदन्ते उत्तरपदे न त्वव्ययस्य । शित्त्वाच्छबादि । जनमेजयतीति जनमेजयः ।

वातशुनीतिलशर्धेष्वजधेट्तुदजहातिभ्यः खश उपसंख्यानम्

(वार्तिकम्) ॥ वातमजा मृगाः ॥

Rule. The augment mum (m) comes to arus "wound", to dviṣat "enemy", and to a word ending in a vowel (ac), when an uttarapada ending in a kh-marked (khit) affix follows. It does not come to an indeclinable (avyaya). One commentator: "khit" runs on from SK 2943, and "uttarapada" from the heading SK 958. Bhaṭṭoji adds: "because khaś is ś-marked, śap and the rest follow".

Example.

janān ejayati → jana + eji + khaś

→ śap, since khaś is sārvadhātuka → guṇa of i and ay → jana + ejay + a + a

→ the two *a*s merge into one (pararūpa) → jana + ejaya

→ the case-ending of jana is lost → mum by this rule → janam-ejaya → janamejayaḥ.

The steps are one commentator's; Bhaṭṭoji says only "śap and the rest".

Applies when / blocked by. A vārttika adds khaś after aj, dheṭ, tud and jahāti, with vāta, śunī, tila and śardha in order: vātam ajanti → vātamajāḥ mṛgāḥ "deer that run against the wind".

One commentator explains why the rule says "ending in an ac" with the word "end". It means the near vowel left by the shortening of SK 2943. So shortening comes first, even where the vowel is short already.

Watch out. The m is an augment on the first member, not an affix. It comes only before a khit uttarapada, such as one made with khaś or khac.

SK 2943 · खित्यनव्ययस्य · khity anavyayasya (A 6.3.66)

खित्यनव्ययस्य

khityanavyayasya · A 6.3.66

खिदन्ते परे पूर्वपदस्य ह्रस्वः स्यात् । ततो मुम् । शुनिन्धयः । तिलन्तुदः । शर्धंजहा माषाः । शर्धोऽपानशब्दः तं जहतीति विग्रहः । जहातिरन्तर्भावितण्यर्थः ॥

Rule. When an uttarapada ending in a kh-marked affix follows, a short vowel replaces the final vowel of a pūrvapada that is not an indeclinable. One commentator: "short" runs on from SK 999. Bhaṭṭoji: "after that, mum". So this rule comes first, then SK 2942.

Example.

śunīṃ dhayati → śunī + dhaya (dheṭ + khaś, with śap)

→ this rule → śuni + dhaya → mum (SK 2942) → śunim + dhaya → śunindhayaḥ.

tilaṃ tudati → tilantudaḥ "one who crushes sesame".

śardhaṃjahā māṣāḥ "beans that make one break wind".

Applies when / blocked by. Both commentators fix the order. If mum came first, the pūrvapada would no longer end in a vowel, and the shortening could not apply.

Bhaṭṭoji explains śardhaṃjahāḥ. śardha is the sound of the downward breath. The analysis is "they leave it". But jahāti here carries a causative sense: "they make one release it".

Watch out. Shorten first, then add m: śuni- then śunim-. Never śunīm-.

SK 2944 · नासिकास्तनयोर्ध्माधेटोः · nāsikāstanayor dhmādheṭoḥ (A 3.2.29)

नासिकास्तनयोर्ध्माधेटोः

nāsikāstanayordhmādheṭoḥ · A 3.2.29

अत्र वार्तिकम् ।

स्तने धेटो नासिकायां ध्मश्चेति वाच्यम्

(वार्तिकम्) ॥ स्तनं धयतीति स्तनन्धयः । धेटष्टित्त्वात् स्तनन्धयी । नासिकन्धमः । नासिकन्धयः ॥

Rule. With nāsikā "nose" or stana "breast" as upapada, khaś comes after dhmā "to blow" and dheṭ "to suck". One commentator supplies khaś. Bhaṭṭoji then quotes a vārttika that sets aside the pairing in order (SK 128). dheṭ goes with stana; dhmā, and also dheṭ, go with nāsikā.

Example.

stanaṃ dhayati → stana + dheṭ + khaś → dhaya → mum → stanandhayaḥ "an infant".

Feminine: because dheṭ is ṭ-marked (Bhaṭṭoji) → stanandhayī.

nāsikā + dhmā + khaś → dhama for dhmā (SK 2360) → nāsika- (SK 2943) → mum → nāsikandhamaḥ "one who blows through the nose".

Also nāsikandhayaḥ. The steps are both commentators'.

Applies when / blocked by. Both commentators say "atra" ("on this") is Bhaṭṭoji's signal that the vārttika overrides the order of the sūtra.

On the feminine, one commentator reports a view of Haradatta and others: ṅīp from dheṭ comes only in the khaś-form. With other affixes the feminine takes ṭāp: dhayā (SK 2899). Scholars discuss this.

Watch out. The sūtra alone would pair nāsikā with dhmā and stana with dheṭ. The vārttika adds nāsikandhayaḥ and gives no stana + dhmā form.

SK 2945 · नाडीमुष्ट्योश्च · nāḍīmuṣṭyoś ca (A 3.2.30)

नाडीमुष्ट्योश्च

nāḍīmuṣṭyośca · A 3.2.30

एतयोरुपपदयोः कर्मणोर्ध्माधेटोः खश् स्यात् ॥ यथासंख्यंनेष्यते ॥ नाडिन्धमः । नाडिन्धयः । मुष्टिन्धमः । मुष्टिन्धयः ।

घटीखारीखरीषूपसंख्यानम्

(वार्तिकम्) ॥ घटिन्धमः । घटिन्धय इत्यादि ॥ खारी परिमाणविशेषः । खरी गर्दभी ॥

Rule. With nāḍī "tube" or muṣṭi "fist" as object, khaś comes after dhmā and dheṭ. khaś goes back to SK 2941; dhmā and dheṭ run on from SK 2944. A vārttika: pairing in order is not intended.

Example.

nāḍī + dhmā + khaś → nāḍi- (SK 2943) → mum → nāḍindhamaḥ; also nāḍindhayaḥ.

muṣṭindhamaḥ, muṣṭindhayaḥ.

Applies when / blocked by. A further vārttika adds ghaṭī, khārī and kharī: ghaṭindhamaḥ, ghaṭindhayaḥ and so on. Bhaṭṭoji: khārī is a measure, kharī a she-ass. Both commentators say the Bhāṣya makes the "no pairing" point explicit.

Watch out. All four combinations are correct here. Do not pair nāḍī only with dhmā.

SK 2946 · उदि कूले रुजिवहोः · udi kūle rujivahoḥ (A 3.2.31)

उदि कूले रुजिवहोः

udi kūle rujivahoḥ · A 3.2.31

उत्पूर्वाभ्यां रुजिवहिभ्यां कूले कर्मण्युपपदे खश् स्यात् । कूलमुद्रुजतीति कूलमुद्रुजः । कूलमुद्वहः ॥

Rule. After ruj "to break" and vah "to carry" preceded by ut, khaś comes when kūla "bank" is the object.

Example.

kūlam udrujati → kūla + ud + ruj + khaś → mum → kūlamudrujaḥ "that which breaks down the bank".

kūlamudvahaḥ "that which carries away the bank".

One commentator: the case-ending is lost, then mum.

Applies when / blocked by. One commentator explains why ut is not an upapada. It is stated as a separate word udi, not compounded with kūla. That blocks its being an upapada and pairing in order. Both commentators: ruj takes an object, and ut names no substance, so "object" qualifies kūla.

Watch out. ut stays with the root: -udruja, -udvaha. The m goes on kūla.

SK 2947 · वहाभ्रे लिहः · vahābhre lihaḥ (A 3.2.32)

वहाभ्रे लिहः

vahābhre lihaḥ · A 3.2.32

वहः स्कन्धस्तं लेढीति वहंलिहो गौः । अदादित्वाच्छपो लुक् । खशो ङित्वान्न गुणः । अभ्रंलिहो वायुः ॥

Rule. After lih "to lick", khaś comes with vaha "shoulder" or abhra "cloud" as object (one commentator).

Example.

vahaṃ leḍhi → vaha + lih + khaś → śap comes and is lost, since lih is of the ad class (SK 2423)

→ no guṇa → liha → mum → vahaṃlihaḥ "an ox (it licks its shoulder)".

abhraṃlihaḥ "the wind (it licks the clouds)".

Applies when / blocked by. Bhaṭṭoji: "because khaś is ṅit, there is no guṇa". khaś is not itself ṅ-marked. It is a sārvadhātuka without p, so it counts as ṅ-marked (SK 2234), and SK 2217 then blocks guṇa.

Watch out. Write -liha, not -leha. The ṅit status keeps the i.

SK 2948 · परिमाणे पचः · parimāṇe pacaḥ (A 3.2.33)

परिमाणे पचः

parimāṇe pacaḥ · A 3.2.33

प्रस्थम्पचा स्थाली । खारिम्पचः कटाहः ॥

Rule. After pac "to cook", khaś comes with a word for a measure (parimāṇa) as object. One commentator: measures such as prastha.

Example.

prastha + pac + khaś → śap → paca → mum → prasthampacā sthālī "a pot that cooks a prastha".

khārī + pac + khaś → khāri- (SK 2943) → mum → khārimpacaḥ kaṭāhaḥ "a cauldron that cooks a khārī". One commentator gives these steps.

Watch out. The measure-word is shortened first if long: khāri-, not khārī-.

SK 2949 · मितनखे च · mitanakhe ca (A 3.2.34)

मितनखे च

mitanakhe ca · A 3.2.34

मितम्पचा ब्राह्मणी । नखम्पचा यवागूः । पचिरत्र तापवाची ॥

Rule. khaś also comes after pac with mita "a measured amount" or nakha "nail" as object.

Example.

mitampacā brāhmaṇī "a brāhmaṇa woman who cooks a measured amount".

nakhampacā yavāgūḥ "gruel that scalds the nails".

Applies when / blocked by. Bhaṭṭoji: "pac here means heating". One commentator explains: nails cannot be softened by cooking.

Watch out. In nakhampacā the root does not mean "to cook". It means "to heat, scald".

SK 2950 · विध्वरुषोस्तुदः · vidhvaruṣos tudaḥ (A 3.2.35)

विध्वरुषोस्तुदः

vidhvaruṣostudaḥ · A 3.2.35

विधुन्तुदः । मुमि कृते संयोगान्तस्य लोपः । अरुन्तुदः ॥

Rule. After tud "to strike", khaś comes with vidhu "moon" or arus "vital spot" as object (one commentator).

Example.

vidhuṃ tudati → vidhu + tuda → mum → vidhuntudaḥ "Rāhu, tormentor of the moon".

arus + tuda → mum after the u → arums + tuda → the last consonant of the cluster is lost (SK 54) → aruntudaḥ "that which strikes the vital spot".

Applies when / blocked by. Bhaṭṭoji: "once mum has been added, the final of the conjunct is lost". One commentator places the m after the u of arus. The other cites a lexicon in which arus also means "a wound".

Watch out. arus is named in SK 2942 for this purpose. It ends in a consonant, so without that naming it could not take mum.

SK 2951 · असूर्यललाटयोर्दृशितपोः · asūryalalāṭayor dṛśitapoḥ (A 3.2.36)

असूर्यललाटयोर्दृशितपोः

asūryalalāṭayordṛśitapoḥ · A 3.2.36

असूर्यमित्यसमर्थसमासः । दृशिना नञः संबन्धात् । सूर्यं न पश्यन्तीत्यसूर्यम्पश्या राजदाराः । ललाटन्तपः सूर्यः ॥

Rule. After dṛś "to see" and tap "to burn", khaś comes with asūrya or lalāṭa "forehead" as object (one commentator).

Example.

sūryaṃ na paśyanti → asūrya + dṛś + khaś → paśya (SK 2360) → mum → asūryampaśyāḥ: the king's wives, "who do not see the sun".

lalāṭaṃ tapati → lalāṭantapaḥ: the sun, "which scorches the forehead".

Applies when / blocked by. Bhaṭṭoji: asūrya is a compound whose members are not connected in sense (asamartha-samāsa). The negative nañ belongs with "see", not with "sun". One commentator explains the word: the queens are so guarded that they do not see even the sun.

Watch out. asūryampaśyāḥ does not mean "seeing what is not the sun". The "not" goes with seeing.

SK 2952 · उग्रम्पश्येरम्मदपाणिन्धमाश्च · ugrampaśyerammadapāṇindhamāś ca (A 3.2.37)

उग्रम्पश्येरम्मदपाणिन्धमाश्च

ugrampaśyerammadapāṇindhamāśca · A 3.2.37

एते निपात्यन्ते । उग्रमिति क्रियाविशेषणं तस्मिन्नुपपदे दृशेः खश् । उग्रं पश्यतीत्युग्रंपश्यः । इरा उदकं तेन माद्यति दीप्यतेऽबिन्धनत्वादिति इरंमदो मेघज्योतिः । इह निपातनात् न श्यन्न । पाणयो ध्मायन्तेऽस्मिन्निति पाणिंधमोऽध्वा । अन्धकाराद्यावृत इत्यर्थः । तत्र हि सर्पाद्यपनोदनाय पाणयो ध्मायन्ते ॥

Rule. Three words are given by nipātana: ugrampaśya, irammada and pāṇindhama.

Example.

ugraṃ paśyati → ugraṃpaśyaḥ "one who looks fiercely". Bhaṭṭoji: ugram qualifies the action; khaś comes after dṛś.

irā "water" + mad → iraṃmadaḥ: lightning, which blazes on water as its fuel. Bhaṭṭoji: by nipātana, śyan does not come.

pāṇayo dhmāyante 'smin → pāṇiṃdhamaḥ: a road covered by darkness. Hands are sounded there to drive off snakes.

Applies when / blocked by. One commentator explains śyan. mad is of the div class, whose class-sign is śyan. So iraṃmada would be expected to have it. The same commentator shortens irā by SK 2943.

Watch out. In ugrampaśya the upapada is an adverb, not an object. And pāṇindhama denotes the locus, the road "in which" hands are sounded, not an agent. The nipātana covers both.

I.N. khac, and its short vowel — running example: `priya + vad + khac → (mum) priyaṃvadaḥ`; `dviṣat + tāpi + khac → (hrasva, 2955) dviṣat-tapi → (ṇi-loss; mum; t lost) dviṣantapaḥ`

The affix khac is a with the markers kh and c. Its kh brings mum, as with khaś. But it has no ś, so it is ārdhadhātuka. Before it the ṇi of a causative can be lost, and SK 2955 shortens the vowel before that ṇi. SK 2957 fixes two words by nipātana. SK 2959–2960 give khac after kaṣ and kṛ. The running examples are priyaṃvadaḥ "one who speaks what is pleasant" and dviṣantapaḥ "tormentor of enemies".

SK 2953 · प्रियवशे वदः खच् · priyavaśe vadaḥ khac (A 3.2.38)

प्रियवशे वदः खच्

priyavaśe vadaḥ khac · A 3.2.38

प्रियंवदः । वशंवदः ॥

गमेः सुपि वाच्यः

(वार्तिकम्) ॥ असंज्ञार्थमिदम् । मितङ्गमो हस्ती ।

विहायसो विह इति वाच्यम्

(वार्तिकम्) ।

खच्च डिद्वा वाच्यः

(वार्तिकम्) ॥ विहंगमः । विहङ्गः । भुजंगमः । भुजंगः ॥

Rule. After vad "to speak", khac comes with priya "pleasant" or vaśa "will" as object.

Example.

priya + vad + khac → mum (SK 2942) → priyaṃvadaḥ.

vaśaṃvadaḥ "obedient".

Applies when / blocked by. Why khac, when khaś runs on? One commentator: for the short vowel of SK 2955. The other reports the Vṛtti's view that the c lets SK 2955 name khac alone. Both commentators answer an objection: "then put khac in the next rule". It stands here, they answer, to show that khac comes after other roots too. One reports this answer as what "they say".

A vārttika: khac after gam with a sup-word as upapada. Bhaṭṭoji: this is for words that are not names: mitaṅgamaḥ "an elephant (going at a measured pace)". Names come by SK 2964.

A vārttika: viha replaces vihāyas "sky". Another: khac is optionally ḍ-marked.

vihāyas + gam + khac → vihaṃgamaḥ; with ḍ-marked khac, the ṭi of gam is lost → vihaṅgaḥ "bird".

Also bhujaṃgamaḥ / bhujaṃgaḥ "snake".

Whether the ḍ-marking holds only with viha or more widely, the two commentators disagree; scholars discuss this.

Watch out. priyaṃvada has mum because khac is khit. Do not write priyavada.

SK 2954 · द्विषत्परयोस्तापेः · dviṣatparayos tāpeḥ (A 3.2.39)

द्विषत्परयोस्तापेः

dviṣatparayostāpeḥ · A 3.2.39

खच्स्यात् ॥

Rule. After tāpi (tap with causative ṇi), khac comes with dviṣat "enemy" or para "foe" as object. Bhaṭṭoji supplies khac. One commentator: tāpi covers both roots tap, of the cur class and of the bhū class.

Example. The forms are completed by the next rule: dviṣantapaḥ, paraṃtapaḥ (SK 2955).

Watch out. The root is the causative tāpi, with long ā. The short a of -tapa comes from SK 2955.

SK 2955 · खचि ह्रस्वः · khaci hrasvaḥ (A 6.4.94)

खचि ह्रस्वः

khaci hrasvaḥ · A 6.4.94

खच्परे णौ उपधाया ह्रस्वः स्यात् । द्विषन्तं परं वा तापयतीति द्विषन्तपः । परंतपः । घटघटीग्रहणाल्लिङ्गविशिष्टपरिभाषा अनित्या । तेनेह न । द्विषतीं तापयतीति द्विषतीतापः ॥

Rule. When ṇi followed by khac comes next, a short vowel replaces the penultimate. One commentator: "before ṇi" runs on from SK 2604, "penultimate" from SK 2364.

Example.

dviṣantaṃ tāpayati → dviṣat + tāpi + khac

→ this rule: ā → a → dviṣat + tapi + a

→ the ṇi is lost before the ārdhadhātuka khac → dviṣat + tapa

→ mum (SK 2942) → dviṣa-m-t + tapa → the cluster-final t is lost (one commentator) → dviṣam + tapa

→ the m appears as n before t → dviṣantapaḥ.

paraṃtapaḥ.

The steps after the shortening are both commentators'.

Applies when / blocked by. One commentator: with khaś neither the short vowel nor the loss of ṇi could happen. khaś is sārvadhātuka, and ṇi is not lost before it.

Bhaṭṭoji adds a limit. Because the vārttika on SK 2923 lists both ghaṭa and ghaṭī, the rule that a stem includes its feminine form is not universal. So dviṣatī "a hostile woman" does not count as dviṣat. Then: dviṣatīṃ tāpayati → dviṣatītāpaḥ. Here khac does not come; one commentator calls dviṣatītāpaḥ the aṇ-form.

Watch out. dviṣantapaḥ has short a in -tapa; dviṣatītāpaḥ has long ā. The first has khac, the second aṇ.

SK 2956 · वाचि यमो व्रते · vāci yamo vrate (A 3.2.40)

वाचि यमो व्रते

vāci yamo vrate · A 3.2.40

वाक् शब्दे उपपदे यमेः खच् स्याद्व्रते गम्ये ॥

Rule. With the word vāc "speech" as upapada, khac comes after yam "to restrain", when a vow (vrata) is understood.

Example. The form comes with the next rule: vācaṃyamaḥ (SK 2957).

Applies when / blocked by. One commentator explains why this rule is needed. SK 2957 could not supply the khac itself. If it did, khac would come outside a vow too.

Watch out. The vow is the condition. Restraint of speech for any other reason does not take khac.

SK 2957 · वाचंयमपुरंदरौ च · vācaṃyamapuraṃdarau ca (A 6.3.69)

वाचंयमपुरंदरौ च

vācaṃyamapuraṃdarau ca · A 6.3.69

वाक्पुरोरमन्तत्वं निपात्यते । वाचंयमो मौनव्रती । व्रते किम् । अशक्त्यादिना वाचं यच्छतीति वाग्यामः ॥

Rule. The words vācaṃyama and puraṃdara are given by nipātana. Bhaṭṭoji: what the nipātana fixes is the -am ending of vāc and pur.

Example.

vācaṃ yacchati → vāc + yam + khac → case-ending lost → am by this rule → vācaṃyamaḥ "one under a vow of silence".

For puraṃdara see SK 2958.

Applies when / blocked by. Why "vow" in SK 2956? Bhaṭṭoji: one who restrains speech through inability or the like is vāgyāmaḥ. Its vṛddhi shows the general aṇ. One commentator adds that mum (SK 2942) has no scope here. (Both vāc and pur end in a consonant.)

Watch out. The m of vācaṃ- is not mum. It is the am given by this nipātana.

SK 2958 · पूःसर्वयोर्दारिसहोः · pūḥsarvayor dārisahoḥ (A 3.2.41)

पूःसर्वयोर्दारिसहोः

pūḥsarvayordārisahoḥ · A 3.2.41

पुरं दारयतीति पुरन्दरः । सर्वंसहः । सहिग्रहणमसंज्ञार्थम् ॥

भगे च दारेरिति काशिका

(वार्तिकम्) ॥ बाहुलकेन लब्धमिदमित्याहुः । भगं दारयतीति भगन्दरः ॥

Rule. With pur "city" or sarva "all" as object, khac comes after dāri and sah respectively. dāri is dṝ "to split" with causative ṇi; both commentators identify it. One gives the pairing in order.

Example.

puraṃ dārayati → pur + dāri + khac → short vowel before ṇi + khac (SK 2955) → ṇi lost → case-ending lost → am (SK 2957) → puraṃdaraḥ "shatterer of cities".

sarva + sah + khac → mum → sarvaṃsahaḥ "enduring all".

The operations of puraṃdaraḥ are one commentator's; they are set out here in the order of SK 2955.

Applies when / blocked by. Bhaṭṭoji: sahi is named for words that are not names. Both commentators say names come by SK 2963.

The Kāśikā adds dāri with bhaga, Bhaṭṭoji says. "They say" it is got by bāhulaka: bhagaṃ dārayati → bhagaṃdaraḥ "fistula".

Watch out. Two different rules give the nasal here. In sarvaṃsahaḥ it is mum; in puraṃdaraḥ it is the am of SK 2957.

SK 2959 · सर्वकूलाभ्रकरीषेषु कषः · sarvakūlābhrakarīṣeṣu kaṣaḥ (A 3.2.42)

सर्वकूलाभ्रकरीषेषु कषः

sarvakūlābhrakarīṣeṣu kaṣaḥ · A 3.2.42

सर्वङ्कषः खलः । कूलङ्कषा नदी । अभ्रङ्कषो वायुः । करीङ्कषा वात्या ॥

Rule. After kaṣ "to rub, scrape", with sarva, kūla, abhra or karīṣa as object, the affix comes. Neither Bhaṭṭoji nor his commentators names it. It is taken as khac, running on from SK 2953.

Example.

sarva + kaṣ + khac → mum → sarvaṅkaṣaḥ "a villain (grinding down all)".

kūlaṅkaṣā "a river (scraping the bank)".

abhraṅkaṣaḥ "the wind (grazing the clouds)".

karīṣaṅkaṣā "a whirlwind (sweeping up dry dung)".

Applies when / blocked by. The forms alone cannot decide between khac and khaś. Both are khit, and both bring mum.

Watch out. Bhaṭṭoji writes the nasal before k as ṅ here (sarvaṅkaṣaḥ) and as ṃ elsewhere (bhayaṃkaraḥ, SK 2960).

SK 2960 · मेघर्तिभयेषु कृञः · meghartibhayeṣu kṛñaḥ (A 3.2.43)

मेघर्तिभयेषु कृञः

meghartibhayeṣu kṛñaḥ · A 3.2.43

मेघङ्करः । भयंकरः । भयशब्देन तदन्तविधिः । अभयङ्करः ॥

Rule. After kṛ, with megha "cloud", ṛti or bhaya "fear" as object, khac comes. As at SK 2959, the affix is not named but taken from the run since SK 2953.

Example.

megha + kṛ + khac → guṇa with r → mum → meghaṅkaraḥ "cloud-maker".

bhayaṃkaraḥ "fear-causing".

Applies when / blocked by. Bhaṭṭoji: with the word bhaya, the rule also applies to a word ending in it (tadantavidhi, SK 26): abhayaṅkaraḥ "causing fearlessness". One commentator says the Bhāṣya makes this explicit. Bhaṭṭoji's printed text gives no example with ṛti. One commentator glosses ṛtiṅkara, with ṛti meaning "going".

Watch out. Compare kumbhakāraḥ (SK 2913, aṇ) and meghaṅkaraḥ (khac). khac brings no vṛddhi, and it brings mum.

II.A. khac (and aṇ) after kṛñ, bhū, the "name" roots and gam — running example: `kṣema + kṛ + khac → kṣema-m-kara → kṣemaṅkaraḥ`; `kṣema + kṛ + aṇ → kṣemakāraḥ`

The run of khac rules from Part I goes on for four more sūtras. khac is marked with kh and c, so only a is added. Its kh gives the upapada the augment m (mum, SK 2942), as in priyaṃvadaḥ. SK 2961 lets kṛñ take aṇ as well as khac with three upapadas. SK 2962 lets bhū take khac in two senses other than the agent. SK 2963–2964 give khac in names. The running example is kṣema "welfare" with kṛñ "to make". With khac the upapada gets m: kṣemaṅkaraḥ. With aṇ the root gets vṛddhi: kṣemakāraḥ. Both mean "maker of welfare".

SK 2961 · क्षेमप्रियमद्रेऽण् च · kṣemapriyamadre 'ṇ ca (A 3.2.44)

क्षेमप्रियमद्रेऽण् च

kṣemapriyamadre'ṇ ca · A 3.2.44

एषु कृञोऽण् स्यात् । चात् खच् । क्षेमङ्करः । क्षेमकारः । प्रियङ्करः । प्रियकारः । मद्रङ्करः । मद्रकारः । वेति वाच्येऽण्ग्रहणं हेत्वादिषु टो मा भूदिति । कथं तर्हि अल्पारम्भाः क्षेमकरा इति । कर्मणः शेषत्वविवक्षायां पचाद्यच् ॥

Rule. With kṣema, priya or madra as upapada, kṛñ takes the affix aṇ. By the word ca "and", khac also comes, carried forward from SK 2960. So each upapada gives two words.

Example.

kṣema + kṛ + khac → kh and c are markers → kṣema + kṛ + a

→ the root vowel takes guṇa (SK 2168) → kṣema + kara

→ khac is kh-marked, so the upapada takes m (SK 2942) → kṣema-m-kara

→ the m is sounded ṅ before k → kṣemaṅkaraḥ.

kṣema + kṛ + aṇ → ṇ is a marker → kṣema + kṛ + a

→ the affix is ṇ-marked, so the root vowel takes vṛddhi (SK 254) → kṣema + kāra

→ kṣemakāraḥ.

Bhaṭṭoji gives the other pairs: priyaṅkaraḥ, priyakāraḥ "doer of what is dear"; madraṅkaraḥ, madrakāraḥ.

Applies when / blocked by. Why name aṇ at all? Pāṇini could have said "optionally (vā) khac". Bhaṭṭoji answers: aṇ is named so that ṭa does not come. ṭa is the affix taught after kṛñ for cause and the like (SK 2934). One of his commentators sets out the objection. With "optionally", aṇ would follow anyway where khac is absent, by karmaṇy aṇ (SK 2913). So naming aṇ must have a further point: it keeps ṭa out.

Watch out. Bhaṭṭoji asks about the phrase alpārambhāḥ kṣemakarāḥ "small undertakings bring welfare". It has neither m nor vṛddhi. His answer: here the object is meant as a mere relation (śeṣa), not as an object. So the affix is ac of the pacādi class (SK 2896). One of his commentators adds the reason. When the object counts only as śeṣa, there is no object-upapada, so neither aṇ nor khac comes.

SK 2962 · आशिते भुवः करणभावयोः · āśite bhuvaḥ karaṇabhāvayoḥ (A 3.2.45)

आशिते भुवः करणभावयोः

āśite bhuvaḥ karaṇabhāvayoḥ · A 3.2.45

आशितशब्दे उपपदे भवतेः खच् । आशितो भवत्यनेनाशितंभवः ओदनः । आशितस्य भवनं आशितम्भवः ॥

Rule. With āśita "sated" as upapada, bhū "to become" takes khac. Bhaṭṭoji supplies khac. The affix here does not mean the agent. It means the instrument (karaṇa) or the bare action (bhāva).

Example.

Instrument: āśito bhavaty anena "by this one becomes sated".

āśita + bhū + khac → āśita + bhū + a

→ guṇa (SK 2168), and o becomes av before a vowel (SK 61) → āśita + bhava

→ m for the upapada (SK 2942) → āśitaṃbhavaḥ odanaḥ "rice that sates".

Bare action: āśitasya bhavanam "the becoming sated" → āśitambhavaḥ.

One of Bhaṭṭoji's commentators labels the two analyses as instrument and bare action.

Applies when / blocked by. One of Bhaṭṭoji's commentators adds a third form. By vāsarūpa (SK 2830), lyuṭ may also come: āśitabhavanam. ghañ cannot come, because it is of the same form as khac. The same commentator says Jayāditya also cites āśitabhavanam.

Watch out. This rule names its own senses. So the default agent sense of kartari kṛt (SK 2832) does not hold here.

SK 2963 · संज्ञायां भृतॄवृजिधारिसहितपिदमः · saṃjñāyāṃ bhṛtṝvṛjidhārisahitapidamaḥ (A 3.2.46)

संज्ञायां भृतॄवृजिधारिसहितपिदमः

saṃjñāyāṃ bhṛtṝvṛjidhārisahitapidamaḥ · A 3.2.46

विश्वं बिभर्तीति विश्वम्भरः । विश्वंभरा । रथंतरं साम । इह रथेन तरतीति व्युत्पत्तिमात्रं न त्ववयवार्थानुगमः । पतिंवरा कन्या । शत्रुंजयो हस्ती । युगंधरः पर्वतः । शत्रुंसहः । शत्रुंतपः । अरिंदमः । दमिः शमनायां तेन सकर्मक इत्युक्तम् । मतान्तरे तु अन्तर्भावितण्यर्थोऽत्र दमिः ॥

Rule. In a name (saṃjñā), khac comes after eight roots. They are bhṛ "to bear", tṝ "to cross", vṛ "to choose", ji "to conquer", dhāri "to hold", sah "to endure", tap "to burn" and dam "to tame". One of Bhaṭṭoji's commentators supplies khac. In the examples, except the mere etymology of rathaṃtara, an object stands as upapada.

Example.

viśvaṃ bibharti "he sustains the universe".

viśva + bhṛ + khac → guṇa (SK 2168) → viśva + bhara

→ m for the upapada (SK 2942) → viśvambharaḥ.

One commentator: viśvambhara is a name of Viṣṇu, viśvaṃbharā of the earth.

Bhaṭṭoji's other names: patiṃvarā kanyā "a maiden choosing a husband" (vṛ); śatruṃjayo hastī, an elephant named "Foe-conquering" (ji); yugaṃdharaḥ parvataḥ, a mountain named "Yoke-bearer" (dhāri); śatruṃsahaḥ, śatruṃtapaḥ, ariṃdamaḥ.

For yugaṃdhara one commentator gives two further steps. The causative stem is shortened by khaci hrasvaḥ (SK 2955), and its ṇi is lost.

Applies when / blocked by. Bhaṭṭoji also cites rathaṃtaram, the name of a sāman, a chant of the Sāmaveda. "It crosses by chariot" (rathena tarati) is, he says, a mere etymology (vyutpattimātra). The meaning of the parts is not carried through into the word. One of his commentators gives the reason: a sāman cannot cross anything by chariot. The other says why the etymology is kept at all. It serves the accent and the avagraha, the pause that splits a compound in recitation.

Bhaṭṭoji discusses dam. It is listed in the sense of "quelling", so it was called transitive (sakarmaka). In another view, dam here has the causative sense taken into it (antarbhāvita-ṇyartha). Both of his commentators name the holders. The first view is Mādhava's and others'; the second is Haradatta's and others'.

Watch out. The sūtra's vṛji is two roots, vṛ and ji. One commentator divides it so, and Bhaṭṭoji gives a separate example for each. Why "in a name"? One commentator answers with kuṭumbaṃ bibharti "he supports a family" → kuṭumbaṃbhāraḥ. It is not a name, so khac does not come. The form has aṇ, with vṛddhi in bhāra.

SK 2964 · गमश्च · gamaś ca (A 3.2.47)

गमश्च

gamaśca · A 3.2.47

सुतंगमः ॥

Rule. gam "to go" also takes khac in a name. One of Bhaṭṭoji's commentators supplies "in a name, khac".

Example.

suta + gam + khac → suta + gam + a

→ m for the upapada (SK 2942) → sutaṃgamaḥ, a name.

Applies when / blocked by. Why not add gam to the list of SK 2963? One of Bhaṭṭoji's commentators answers. gam stands alone so that only gam runs on into SK 2965. The roots bhṛ, tṝ and the rest do not.

II.B. ḍa after gam and han: the ṭi lost by the ḍ-marker — running example: `anta + gam + ḍa → anta-g-a → antagaḥ`

From SK 2965 the affix is ḍa. Its ḍ is a marker, so only a is added. The ḍ also takes away the ṭi of the root. The ṭi is the last vowel of a form together with anything after it. In gam the ṭi is am. Normally SK 316 (ṭeḥ) drops a ṭi only in a bha stem. bha is a name a stem takes before certain affixes. A root before ḍa is not bha. Yet the ṭi goes "by the force of the ḍ-marking": the marker would be pointless otherwise. SK 2965 teaches ḍa after gam, SK 2966–2967 after han "to strike". The running example: antaṃ gacchati "it goes to the end" gives antagaḥ.

SK 2965 · अन्तात्यन्ताध्वदूरपारसर्वानन्तेषु डः · antātyantādhvadūrapārasarvānanteṣu ḍaḥ (A 3.2.48)

अन्तात्यन्ताध्वदूरपारसर्वानन्तेषु डः

antātyantādhvadūrapārasarvānanteṣu ḍaḥ · A 3.2.48

संज्ञयामिति निवृत्तम् । एषु गमेर्डः स्यात् । डित्त्वासामर्थ्यादभस्यापि टेर्लोपः । अन्तं गच्छतीत्यन्तग इत्यादि ।

सर्वत्रपन्नयोरुपसंख्यानम्

(वार्तिकम्) ॥ सर्वत्रगः । पन्नं पतितं गच्छतीति पन्नगः । पन्नमिति पद्यतेः क्तान्तं क्रियाविशेषणम् । उरसो लोपश्च ॥ उरसा गच्छतीत्युरगः ।

सुदूरोरधिकरणे

(वार्तिकम्) ॥ सुखेन गच्छत्यत्र सुगः । दुर्गः ।

अन्यत्रापि दृश्यत इति वक्तव्यम्

(वार्तिकम्) ॥ ग्रामगः ॥

डे च विहायसो विहादेशो वक्तव्यः

(वार्तिकम्) ॥ विहगः ॥

Rule. With anta, atyanta, adhvan, dūra, pāra, sarva or ananta as upapada, gam takes ḍa. gam runs on from SK 2964. "In a name" has ceased, says Bhaṭṭoji.

Example.

antaṃ gacchati "it goes to the end".

anta + gam + ḍa → ḍ is a marker → anta + gam + a

→ the ṭi am of gam is lost by the force of the ḍ-marking → anta + g + a

→ antagaḥ. Bhaṭṭoji: "and so on" for the other upapadas.

Applies when / blocked by. Bhaṭṭoji gives four vārttikas.

sarvatrapannayor upasaṃkhyānam: also with sarvatra and panna. sarvatragaḥ "going everywhere". pannaṃ patitaṃ gacchati "it goes low" → pannagaḥ, a serpent. Bhaṭṭoji: panna is the kta-form of pad, used to qualify the action (kriyāviśeṣaṇa).
uraso lopaś ca: with uras "breast", and its final is lost. urasā gacchati "it goes by its breast" → uragaḥ.
suduror adhikaraṇe: with su or dur, in the sense of a locus (adhikaraṇa). sukhena gacchaty atra "one goes easily here" → sugaḥ; durgaḥ.
anyatrāpi dṛśyate: ḍa is seen elsewhere too. grāmagaḥ "going to the village".

A fifth vārttika makes viha the substitute for vihāyas "sky" before ḍa: vihagaḥ, a bird.

One of Bhaṭṭoji's commentators says "seen elsewhere too" would give all these forms by itself. The other vārttikas only elaborate it (prapañca). The same commentator explains the viha vārttika. The earlier viha substitution belonged to khac, giving vihaṅgamaḥ. So ḍa needs its own statement. The other commentator adds a limit for su and dur. In the object sense they take khal (SK 3305): sugamaḥ panthāḥ "an easy road", durgamaḥ.

Watch out. Do not look for a bha stem to explain the loss of am. The ḍ-marker alone does it. The same holds for every ḍa form in this part.

SK 2966 · आशिषि हनः · āśiṣi hanaḥ (A 3.2.49)

आशिषि हनः

āśiṣi hanaḥ · A 3.2.49

शत्रुं वध्याच्छत्रुहः । आशिषि किम् । शत्रुघातः ।

दारावाहनोऽणन्तस्य च टः संज्ञायाम्

(वार्तिकम्) ॥ दारुशब्दे उपपदे आङ्पूर्वाद्धन्तरेण् टकारश्चान्तादेशो वक्तव्य इत्यर्थः । दार्वाघाटः ।

चारौ वा

(वार्तिकम्) ॥ चार्वाघाटः ।

कर्मणि समि च

(वार्तिकम्) ॥ कर्मण्युपदे संपूर्वाद्धन्तेरुक्तं वेत्यर्थः । वर्णान्संहन्तीति वर्णसङ्घाटः । पदसङ्घाटः । वर्णसङ्घातः । पदसङ्घातः ॥

Rule. han takes ḍa when a wish (āśīḥ) is understood. One of Bhaṭṭoji's commentators glosses: with an object as upapada.

Example.

śatruṃ vadhyāt "may he slay the enemy".

śatru + han + ḍa → śatru + han + a

→ the ṭi an is lost by the ḍ-marking → śatru + h + a

→ śatruhaḥ.

Applies when / blocked by. Why "when a wish is understood"? Without a wish there is śatrughātaḥ. One of Bhaṭṭoji's commentators explains it. With no wish there is no ḍa, so aṇ comes, and t replaces the n of han by SK 2574. He stops there. The h also becomes gh before the ṇ-marked affix, by SK 358.

Bhaṭṭoji adds three vārttikas.

dārāv āhano 'ṇ antasya ca ṭaḥ saṃjñāyām: with dāru "wood" as upapada, āṅ + han takes aṇ, and ṭ replaces its final, in a name. dārvāghāṭaḥ.
cārau vā: with cāru, the same optionally. cārvāghāṭaḥ.
karmaṇi sami ca: with sam and an object upapada, the same optionally. varṇān saṃhanti "he combines the sounds" → varṇasaṅghāṭaḥ or varṇasaṅghātaḥ; likewise padasaṅghāṭaḥ, padasaṅghātaḥ.

One of Bhaṭṭoji's commentators: the first vārttika teaches ṭ together with aṇ. The other: aṇ is already available by SK 2913, so the vārttika's real work is the ṭ. One commentator adds that the word karmaṇi in the third vārttika stops cārau from running on.

Watch out. The forms in -ghāta show the t of SK 2574, and those in -ghāṭa the ṭ of the vārttika. Keep the two apart.

SK 2967 · अपे क्लेशतमसोः · ape kleśatamasoḥ (A 3.2.50)

अपे क्लेशतमसोः

ape kleśatamasoḥ · A 3.2.50

अपपूर्वाद्धन्तेर्डः स्यात् । अनाशीरर्थमिदम् । क्लेशापहः पुत्रः । तमोपहः सूर्यः ॥

Rule. With kleśa "distress" or tamas "darkness" as upapada, han preceded by apa takes ḍa. Bhaṭṭoji supplies ḍa and han. He adds: this rule is for cases where no wish is meant.

Example.

kleśa + apa + han + ḍa → the ṭi an is lost → kleśa-apa-h-a

→ like vowels merge → kleśāpahaḥ putraḥ "a son who removes distress".

In the same way: tamopahaḥ sūryaḥ "the sun, dispeller of darkness".

Applies when / blocked by. Why this rule, if SK 2966 already gives ḍa after han? One of Bhaṭṭoji's commentators answers. SK 2966 needs a wish. This rule gives ḍa where there is none. That is the point of Bhaṭṭoji's words "for no wish" (anāśīrartham).

II.C. han with ṇini and ṭak; the nipātanas pāṇigha, tāḍagha — running example: `jāyā + han + ṭak → jāyā-ghn-a → jāyāghnaḥ`

Five more rules teach han with an object upapada. SK 2968 gives ṇini with kumāra and śīrṣa. SK 2969–2971 give ṭak, marked with ṭ and k. Before it the a of han is lost, and h becomes gh before n. SK 2972 states two fixed forms (nipātana) for artisans. The running example is jāyāghnaḥ, a man whose marks show he will kill his wife. Its feminine counterpart is patighnī.

SK 2968 · कुमारशीर्षयोर्णिनिः · kumāraśīrṣayor ṇiniḥ (A 3.2.51)

कुमारशीर्षयोर्णिनिः

kumāraśīrṣayorṇiniḥ · A 3.2.51

कुमारघाती । शिरसः शीर्षभावो निपात्यते । शीर्षघाती ॥

Rule. With kumāra "child" or śīrṣa "head" as object-upapada, han takes ṇini. Both of Bhaṭṭoji's commentators supply han. ṇini leaves in: its ṇ is a marker, and its last i is for pronunciation, one commentator says.

Example.

kumāra + han + ṇini → kumāra + han + in

→ t for n (SK 2574), gh for h (SK 358), vṛddhi of the penultimate a (SK 2282) → kumāra + ghātin

→ kumāraghātī "child-slayer".

Bhaṭṭoji: śīrṣa in place of śiras is given by nipātana → śīrṣaghātī "head-smiter".

Applies when / blocked by. One of Bhaṭṭoji's commentators asks why the rule is needed. ṇini would come anyway by SK 2988 and the rules after it. But those need habit (tācchīlya), necessity or indebtedness. This rule gives ṇini without those senses. It also secures śīrṣa for śiras.

SK 2969 · लक्षणे जायापत्योष्टक् · lakṣaṇe jāyāpatyoṣ ṭak (A 3.2.52)

लक्षणे जायापत्योष्टक्

lakṣaṇe jāyāpatyoṣṭak · A 3.2.52

हन्तेष्टक् स्याल्लक्षणवति कर्तरि । जायाघ्नो ना । पतिघ्नी स्त्री ॥

Rule. With jāyā "wife" or pati "husband" as object-upapada, han takes ṭak. The condition: the agent bears a mark (lakṣaṇa). Bhaṭṭoji glosses: "when the agent possesses a mark". Both of his commentators explain the word. lakṣaṇa in the sūtra ends in the ac of the arśaādi class (SK 1933), so it means "having a mark".

Example.

jāyā + han + ṭak → ṭ and k are markers → jāyā + han + a

→ the penultimate a of han is lost (SK 2363) → jāyā + hn + a

→ h becomes gh before n (SK 358) → jāyāghna

→ jāyāghno nā "a man marked as wife-killer".

pati + han + ṭak → patighna → the affix is ṭ-marked, so the feminine takes ṅīp → patighnī strī.

One of Bhaṭṭoji's commentators gives the loss of a and the ṅīp.

Watch out. The man has not killed anyone. One of Bhaṭṭoji's commentators says the mark, such as a palm-line, only presages the death. Calling him a "killer" is a figurative use (gauṇa).

SK 2970 · अमनुष्यकर्तृके च · amanuṣyakartṛke ca (A 3.2.53)

अमनुष्यकर्तृके च

amanuṣyakartṛke ca · A 3.2.53

जायाघ्नस्तिलकालकः । पतिघ्नी पाणिरेखा । पित्तघ्नं घृतम् । अमनुष्येति किम् । आखुघातः शूद्रः । अथ कथं बलभद्रः प्रलम्बघ्नः । शत्रुघ्नः । कृतघ्न इत्यादि । मूलविभुजादित्वात्सिद्धम् ॥ चोरघातो नगरघातो हस्तीति तु बाहुलकादणि ॥

Rule. han also takes ṭak when the action has a non-human agent. Both of Bhaṭṭoji's commentators gloss it with an object upapada.

Example.

jāyāghnas tilakālakaḥ "a black mole presaging the wife's death".

patighnī pāṇirekhā "a palm-line presaging the husband's death".

pitta + han + ṭak → pittaghnaṃ ghṛtam "ghee that destroys bile".

Not non-human: ākhughātaḥ śūdraḥ "a śūdra who kills rats".

Applies when / blocked by. Bhaṭṭoji asks about pralambaghnaḥ "slayer of Pralamba" (Balabhadra), śatrughnaḥ and kṛtaghnaḥ. Their agents are human, so ṭak is impossible. He answers: they belong to mūlavibhuja and the like. One commentator adds that these take ka, by the vārttika under SK 2919. And coraghāto hastī "an elephant that kills thieves"? The agent is non-human, yet the form has aṇ. Bhaṭṭoji: aṇ comes by bahulaka (SK 2841).

Watch out. Why bahulaka and not vāsarūpa (SK 2830)? One of Bhaṭṭoji's commentators answers. Markers make no difference of form (nānubandhakṛtam asārūpyam). So ṭak and aṇ count as the same form, and vāsarūpa cannot let aṇ in. The same commentator notes that amanuṣya usually means a demon. Here the attested forms require "non-human".

SK 2971 · शक्तौ हस्तिकपाटयोः · śaktau hastikapāṭayoḥ (A 3.2.54)

शक्तौ हस्तिकपाटयोः

śaktau hastikapāṭayoḥ · A 3.2.54

हन्तेष्टक् स्यात् शक्तौ द्योत्यायाम् । मनुष्यकर्तृकार्थमिदम् । हस्तिघ्नो ना । कपाटघ्नश्चोरः । कवाटेति पाठान्तरम् ॥

Rule. With hastin "elephant" or kapāṭa "door" as upapada, han takes ṭak when capability (śakti) is conveyed. Bhaṭṭoji: "this rule is for a human agent".

Example.

hastin + han + ṭak → hastighno nā "a man able to kill an elephant".

kapāṭaghnaś coraḥ "a thief able to break down a door".

Bhaṭṭoji: kavāṭa is a variant reading for kapāṭa.

Applies when / blocked by. One of Bhaṭṭoji's commentators gives the counter-example. viṣeṇa hastinaṃ hanti "he kills an elephant with poison" → hastighātaḥ. Anyone who acts has some capability. So the commentator takes the word to mean a high degree of it: killing by one's own strength.

SK 2972 · पाणिघताडघौ शिल्पिनि · pāṇighatāḍaghau śilpini (A 3.2.55)

पाणिघताडघौ शिल्पिनि

pāṇighatāḍaghau śilpini · A 3.2.55

हन्तेष्टक् टिलोपो घत्वं च निपात्यते पाणिताडयोरुपपदयोः । पाणिघः । ताडघः । शिल्पिनि किम् । पाणिघातः । ताडघातः ।

राजघ उपसंख्यानम्

(वार्तिकम्) ॥ राजनं हन्ति राजघः ॥

Rule. For an artisan (śilpin), pāṇigha and tāḍagha are given by nipātana. Bhaṭṭoji lists what the nipātana teaches. With pāṇi or tāḍa as upapada, han takes ṭak. The ṭi is lost, and gh replaces the h.

Example.

pāṇi + han + ṭak → ṭi an lost; gh for h → pāṇighaḥ "one who strikes with the hand".

tāḍa + han + ṭak → tāḍaghaḥ "one who strikes by beating".

One of Bhaṭṭoji's commentators glosses tāḍa as "beating", and the agent as a wrestler or the like.

Not an artisan: pāṇighātaḥ, tāḍaghātaḥ.

Applies when / blocked by. A vārttika adds rājagha to the nipātana. rājānaṃ hanti "he kills a king" → rājaghaḥ.

Watch out. These forms do not follow SK 2969. There the a of han is lost and ghn results. Here the whole ṭi goes, so the forms end in -gha.

II.D. Making or becoming what one was not: khyun, khiṣṇuc, khukañ — running example: `anāḍhyam āḍhyaṃ kurvanty anena → āḍhya + kṛ + khyun → āḍhya-m-kar-ana → āḍhyaṃkaraṇam`

Two rules teach three kh-marked affixes for a change of state. The sense is that of cvi: "becoming what one was not" (abhūtatadbhāva). cvi is an affix that turns āḍhya "rich" into āḍhyī in āḍhyīkaroti "makes rich". These rules apply when the upapada has that sense but does not end in cvi. SK 2973 gives khyun after kṛñ, in the instrument sense. SK 2974 gives khiṣṇuc and khukañ after bhū, in the agent sense. Bhaṭṭoji then recalls three rules taught in an earlier chapter (SK 432, 373, 429): on spṛś, on ṛtvij and similar words, and on dṛś. The running example is āḍhyaṃkaraṇam "the means by which they make a poor man rich".

SK 2973 · आढ्यसुभगस्थूलपलितनग्नान्धप्रियेषु च्व्यर्थेष्वच्वौ कृञः करणे ख्युन् · āḍhyasubhagasthūlapalitanagnāndhapriyeṣu cvyartheṣv acvau kṛñaḥ karaṇe khyun (A 3.2.56)

आढ्यसुभगस्थूलपलितनग्नान्धप्रियेषु च्व्यर्थेष्वच्वौ कृञः करणे ख्युन्

āḍhyasubhagasthūlapalitanagnāndhapriyeṣu cvyartheṣvacvau kṛñaḥ karaṇe khyun · A 3.2.56

एषु च्व्यर्थेष्वच्व्यन्तेषु कर्मसूपपदेषु कृञः ख्युन् स्यात् । आढ्यामाढ्यं कुर्वन्त्यनेन आढ्यकरणम् । अच्वौ किम् । आढ्यीकुर्वन्त्यनेन । प्रतिषेधसामर्थ्यात् इह ल्युडपि नेति काशिका । भाष्यमते तु ल्युट् स्यादेव । अच्वावित्युत्तरार्थम् ॥

Rule. Seven words may stand as object-upapada: āḍhya "rich", subhaga "fortunate", sthūla "stout", palita "grey", nagna "naked", andha "blind", priya "dear". They must have the cvi sense and must not end in cvi. Then kṛñ takes khyun in the sense of the instrument (karaṇa).

Example.

anāḍhyam āḍhyaṃ kurvanty anena "that by which they make one who is not rich rich".

āḍhya + kṛ + khyun → kh and n are markers → āḍhya + kṛ + yu

→ ana replaces yu (SK 1247) → āḍhya + kṛ + ana

→ guṇa (SK 2168) → āḍhya + karana → ṇ for n after r (SK 235) → āḍhya + karaṇa

→ m for the upapada (SK 2942) → āḍhyaṃkaraṇam.

One of Bhaṭṭoji's commentators gives the ana and the m. He adds subhagaṃkaraṇam and the like.

Applies when / blocked by. The same commentator calls khyun an exception (apavāda) to lyuṭ. lyuṭ is the general affix of the instrument. Why "not ending in cvi"? Bhaṭṭoji: āḍhyīkurvanty anena — here the upapada ends in cvi, so there is no khyun. Does lyuṭ come there instead? The Kāśikā says no: by the force of the prohibition, not even lyuṭ comes. On the Bhāṣya's view lyuṭ does come. Then the word acvau serves the next rule. Bhaṭṭoji reports both views.

Watch out. Both of Bhaṭṭoji's commentators explain the Kāśikā's reasoning. With a cvi-final upapada, lyuṭ and khyun would give the same form, āḍhyīkaraṇam. So the prohibition would do nothing unless it barred lyuṭ too. One commentator reports Kaiyaṭa: the Vṛtti's view conflicts with the Bhāṣya and is unsound.

SK 2974 · कर्तरि भुवः खिष्णुच्खुकञौ · kartari bhuvaḥ khiṣṇuckhukañau (A 3.2.57)

कर्तरि भुवः खिष्णुच्खुकञौ

kartari bhuvaḥ khiṣṇuckhukañau · A 3.2.57

आढ्यादिषु च्व्यर्थेष्वच्व्यन्तेषु भवतेरेतौ स्तः । अनाढ्य आड्यो भवतीति आढ्यंभविष्णु । आढ्यंभावुकः । स्पृशोऽनुदके क्विन् SK 432 । घृतस्पृक् । कर्मणीति निवृत्तम् । मन्त्रेण स्पृशतीति मन्त्रस्पृक् । ऋत्विद्गधृक्स्रद्गिगुष्णिगञ्चुयुजिक्रुञ्चां च SK 373 । व्याख्यातम् । त्यदादिषु दृशोऽनालोचने कञ्च SK 429 ।

समानान्ययोश्चेति वाच्यम्

(वार्तिकम्) ॥ सदृक् । सदृशः । अन्यादृक् । अन्यादृशः ।

क्सोऽपि वाच्यः

(वार्तिकम्) ॥ तादृक्षः । सदृक्षः । अन्यादृक्षः ॥

Rule. Under the same conditions, bhū takes khiṣṇuc and khukañ in the agent sense. The upapada is āḍhya and the rest, in the cvi sense, not ending in cvi.

Example.

anāḍhya āḍhyo bhavati "one who is not rich becomes rich".

āḍhya + bhū + khiṣṇuc → kh and c are markers → āḍhya + bhū + iṣṇu

→ guṇa, and o becomes av (SK 2168, SK 61) → āḍhya + bhaviṣṇu → m → āḍhyaṃbhaviṣṇuḥ.

āḍhya + bhū + khukañ → kh and ñ are markers → āḍhya + bhū + uka

→ the affix is ñ-marked, so vṛddhi; au becomes āv (SK 61) → āḍhya + bhāvuka → āḍhyaṃbhāvukaḥ.

One of Bhaṭṭoji's commentators names the markers and the vṛddhi.

Applies when / blocked by. One commentator says acvau runs on from SK 2973. Otherwise āḍhyībhaviṣṇuḥ and āḍhyībhāvukaḥ would result. The other says the word kartari stops karaṇe from running on. Whether the i of khiṣṇuc is needed, scholars discuss this.

Bhaṭṭoji then recalls three sūtras of this stretch taught earlier, in the section on consonant-final stems.

spṛśo 'nudake kvin (SK 432): after spṛś "to touch", kvin, with an upapada other than udaka "water". ghṛtaspṛk "touching ghee". Bhaṭṭoji: "with an object" has stopped here. So also mantreṇa spṛśati "touches with a mantra" → mantraspṛk.
ṛtvigdadhṛksragdiguṣṇigañcuyujikruñcāṃ ca (SK 373): "explained there".
tyadādiṣu dṛśo 'nālocane kañ ca (SK 429): after dṛś not meaning "seeing", with tyad and the rest, kañ too.

Two vārttikas extend the last rule. samānānyayoś ceti vācyam: with samāna and anya too. sadṛk, sadṛśaḥ "similar"; anyādṛk, anyādṛśaḥ "of another kind". kso 'pi vācyaḥ: the affix ksa too. tādṛkṣaḥ, sadṛkṣaḥ, anyādṛkṣaḥ.

Watch out. sadṛk does not mean "one who sees similarly". Both of Bhaṭṭoji's commentators put kvin and kañ here in the karmakartṛ sense. karmakartṛ is an object that acts as its own agent: "it presents itself to view as similar".

II.E. Zero and near-zero affixes with a sup upapada: kvip, ṇvi, viṭ, kap — running example: `mitra + dviṣ + kvip → mitra-dviṣ → mitradviṭ`

This group teaches affixes that leave little or nothing behind. kvip, ṇvi and viṭ consist of markers, a v, and an i for pronunciation. The markers go. The v stands alone as an affix of one sound (apṛkta) and is lost by ver apṛktasya (SK 375). What remains is the bare root, ending a pada. So the root's final shows pada-final sandhi. The upapada here is any word with a case-ending (sup). SK 2975 gives kvip after twelve roots, SK 2976 ṇvi after bhaj, SK 2977–2978 viṭ after ad, and SK 2979 kap after duh. The running example: mitra + dviṣ + kvip gives mitradviṭ "hater of a friend".

SK 2975 · सत्सूद्विषद्रुहदुहयुजविदभिदच्छिदजिनीराजामुपसर्गेऽपि क्विप् · satsūdviṣadruhaduhayujavidabhidacchidajinīrājām upasarge 'pi kvip (A 3.2.61)

सत्सूद्विषद्रुहदुहयुजविदभिदच्छिदजिनीराजामुपसर्गेऽपि क्विप्

satsūdviṣadruhaduhayujavidabhidacchidajinīrājāmupasarge'pi kvip · A 3.2.61

एभ्यः क्विप्स्यादुपरसर्गे सत्यसति च सुप्युपपदे । द्युसत् । उपनिषत् । अण्डसूः । प्रसूः । मित्रद्विट् । प्रद्विट् । मित्रध्रुक् । प्रध्रुक् । गोधुक् । प्रधुक् । अश्वयुक् । प्रयुक् । वेदवित् । निविदित्यादि ।

अग्रग्रामाभ्यां नयतेर्णो वाच्यः

(वार्तिकम्) ॥ अग्रणीः । ग्रामणीः ॥

Rule. Twelve roots take kvip when a sup-ending word is upapada, with or without an upasarga. The roots are sad, sū, dviṣ, druh, duh, yuj, vid, bhid, chid, ji, nī and rāj.

Example.

mitra + dviṣ + kvip → the whole of kvip goes → mitra-dviṣ

→ the final ṣ, ending the pada, appears as ṭ → mitradviṭ "hater of a friend".

With an upasarga: pradviṭ.

Bhaṭṭoji's other pairs: dyusat "dwelling in heaven", upaniṣat; aṇḍasūḥ "egg-bearing", prasūḥ; mitradhruk, pradhruk; godhuk "milker of cows", pradhuk; aśvayuk "yoker of horses", prayuk; vedavit "knower of the Veda", nivit.

In the second form of each pair, the upasarga itself stands in the place of the upapada.

Applies when / blocked by. Both of Bhaṭṭoji's commentators explain upaniṣat: its ṣ comes by sadir aprateḥ. Both say dyusat keeps its s. One commentator adds more forms: kāṣṭhabhit, rajjucchit, śatrujit (with tuk), senānīḥ, virāṭ. The same commentator draws a point from upasarge 'pi "even with an upasarga". It shows that elsewhere a rule saying "with a sup" excludes an upasarga. The other commentator fixes the roots. sū is the adādi sūte. Every vid is included except vidḷ lābhe "to find".

A vārttika: agragrāmābhyāṃ nayater ṇo vācyaḥ. After agra and grāma, ṇ replaces the n of nī. agraṇīḥ "leader", grāmaṇīḥ "village headman". One commentator calls this a fresh injunction: the r is not in the same pada, so ṇatva would not otherwise apply. The other says usage already makes the point known. Scholars discuss this. That commentator adds that with aṇ there is no ṇ: grāmanāyaḥ.

Watch out. kvip leaves nothing to see. What you see is its effect: the root ends a pada. So pada-final sandhi applies to the root's last sound.

SK 2976 · भजो ण्विः · bhajo ṇviḥ (A 3.2.62)

भजो ण्विः

bhajo ṇviḥ · A 3.2.62

सुप्युपसर्गे चोपपदे भजेर्ण्विः स्यात् । अंशभाक् । प्रभाक् ॥

Rule. bhaj "to share" takes ṇvi when a sup-ending word or an upasarga is upapada. ṇvi vanishes like kvip, but its ṇ still works.

Example.

aṃśa + bhaj + ṇvi → the affix goes; it was ṇ-marked → vṛddhi of the penultimate a (SK 2282) → aṃśa-bhāj

→ the final j, ending the pada, appears as k → aṃśabhāk "sharer of a portion".

With an upasarga: prabhāk.

One of Bhaṭṭoji's commentators gives the vṛddhi.

Watch out. The affix is invisible, yet the root shows vṛddhi. That is the trace of the ṇ-marker.

SK 2977 · अदोऽनन्ने · ado 'nanne (A 3.2.68)

अदोऽनन्ने

ado'nanne · A 3.2.68

विट् स्यात् । आममत्ति आमात् । सस्यात् । अनन्ने किम् । अन्नादः ॥

Rule. ad "to eat" takes viṭ when a sup-ending word other than anna "food" is upapada. Both of Bhaṭṭoji's commentators say viṭ runs on from a Vedic sūtra. It is janasanakhanakramagamo viṭ (A 3.2.67), given later as SK 3413. One explains why the SK skips it here: "in the Veda" runs on into it.

Example.

āmam atti "he eats raw [food]".

āma + ad + viṭ → viṭ leaves nothing → āma-ad

→ like vowels merge → āmād → the final d appears as t → āmāt.

Likewise sasyāt "eater of grain".

Applies when / blocked by. With anna the rule does not apply: annādaḥ "eater of food". One of Bhaṭṭoji's commentators says this has aṇ by karmaṇy aṇ (SK 2913).

SK 2978 · क्रव्ये च · kravye ca (A 3.2.69)

क्रव्ये च

kravye ca · A 3.2.69

अदेर्विट् । पूर्वेण सिद्धे वचनमण्बाधनार्थम् । क्रव्यात् आममांसभक्षकः । कथं तर्हि क्रव्यादोऽस्रप आशर इति । पक्वमांसशब्दे उपपदेऽण् । उपपदस्य क्रव्यादेशः पृषोदरादित्वात् ॥

Rule. With kravya "raw flesh" as upapada too, ad takes viṭ. Bhaṭṭoji: "after ad, viṭ".

Example.

kravya + ad + viṭ → kravya-ad → kravyāt "an eater of raw flesh".

Applies when / blocked by. The form would already come by SK 2977. So why this rule? Bhaṭṭoji: it blocks aṇ. One of his commentators calls it a restriction (niyama). With kravya as upapada, ad takes only viṭ, never aṇ.

Watch out. A line of the Amarakośa, a Sanskrit thesaurus, has kravyādo 'srapa āśaraḥ. How can kravyāda, with aṇ, stand? Bhaṭṭoji answers. The aṇ comes with pakvamāṃsa "cooked flesh" as upapada. Then kravya replaces that upapada, because the word belongs to the pṛṣodarādi class. That class admits irregular forms as found in use.

SK 2979 · दुहः कब् घश्च · duhaḥ kab ghaś ca (A 3.2.70)

दुहः कब् घश्च

duhaḥ kab ghaśca · A 3.2.70

कामदुघा ॥

Rule. duh "to milk" takes kap when a sup-ending word is upapada. And gh replaces the final of the root. The kab of the sūtra is kap, with p voiced before gh.

Example.

kāmaṃ dogdhi "she yields what is desired".

kāma + duh + kap → k and p are markers → kāma + duh + a

→ gh replaces the h → kāma + dugh + a

→ feminine → kāmadughā.

Both of Bhaṭṭoji's commentators supply dhenuḥ "a cow".

Watch out. The sūtra names two operations at once. kap is the affix, and gh is a substitute for the root's final. One of Bhaṭṭoji's commentators reads the gh as replacing the final of the base.

II.F. manin, kvanip, vanip, vic and kvip after any root — running example: `su + śṝ + manin → su-śar-man → suśarmā`; `vi + jan + vanip → vi-jā-van → vijāvā`

Several affixes are taught first for the Veda, in rules the SK gives in a later chapter. SK 2980 frees four of them for ordinary usage, after any root. SK 2983 does the same for kvip. Between them come two rules the new affixes need. SK 2981 bars iṭ before a kṛt that begins with certain consonants. SK 2982 turns a nasal into ā before viṭ and van. The running examples: suśarmā "one who crushes well", with manin; and vijāvā, with vanip after jan "to be born".

SK 2980 · अन्येभ्योऽपि दृश्यन्ते · anyebhyo 'pi dṛśyante (A 3.2.75)

अन्येभ्योऽपि दृश्यन्ते

anyebhyo'pi dṛśyante · A 3.2.75

छन्दसीति निवृत्तम् । मनिन् क्वनिप् वनिप् विच् एते प्रत्यया धातोः स्युः ॥

Rule. The affixes manin, kvanip, vanip and vic are seen after other roots too. Bhaṭṭoji: "in the Veda" has stopped. These affixes come after a root. One of his commentators names their source. manin, kvanip and vanip run on from āto manin-kvanib-vanipaś ca (A 3.2.74, SK 3418). vic runs on from vij upe chandasi (A 3.2.73, SK 3417).

Example. The other commentator names the markers.

manin → final n a marker, i for pronunciation → man.

kvanip, vanip → k and p markers, i for pronunciation → van.

vic → c a marker, i for pronunciation → v, which is then lost.

Bhaṭṭoji's examples come under the next two sūtras: suśarmā and prātaritvā at SK 2981, vijāvā, avāvā, roṭ, reṭ and sugaṇ at SK 2982.

Watch out. The sūtra stands within the scope of the two Vedic rules it draws on. One commentator says so. Only "in the Veda" is dropped.

SK 2981 · नेड् वशि कृति · neḍ vaśi kṛti (A 7.2.8)

नेड् वशि कृति

neḍ vaśi kṛti · A 7.2.8

वशादेः कृत इण्न स्यात् । शॄ । सुशर्मा । प्रातरित्वा ॥

Rule. iṭ does not come before a kṛt affix that begins with a vaś sound. vaś is a pratyāhāra; it includes m and v. So man and van take no iṭ. One of Bhaṭṭoji's commentators says vaś qualifies kṛt. By the reading "beginning with" (tadādi-vidhi), it means a kṛt that begins with vaś.

Example.

suṣṭhu śṛṇāti "he crushes well".

su + śṝ + manin → su + śṝ + man → m is a vaś sound, so no iṭ

→ guṇa: ar for ṝ → su + śar + man → suśarmā.

prātar eti "he goes early".

prātar + iṇ + kvanip → prātar + i + van → no iṭ

→ tuk after the short vowel (SK 2858) → prātar-i-t-van → prātaritvā.

Both commentators name the affixes; one gives the tuk.

Watch out. The i of prātaritvā is the root iṇ "to go" itself, not the augment iṭ. The rule has just barred iṭ.

SK 2982 · विड्वनोरनुनासिकस्यात् · viḍvanor anunāsikasyāt (A 6.4.41)

विड्वनोरनुनासिकस्यात्

viḍvanoranunāsikasyāt · A 6.4.41

अनुनासिकस्य आत्स्यात् । विजायत इति विजावा । ओणृ । अवावा । विच् । रोट् । रेट् । सुगण् ॥

Rule. Before viṭ and van, ā replaces a nasal.

Example.

vijāyate "is born".

vi + jan + vanip → vi + jan + van

→ ā replaces the nasal n → vi + ja-ā + van

→ like vowels merge (savarṇadīrgha) → vijāvā.

oṇṛ: oṇ + vanip → oṇ + van → ā replaces the nasal ṇ → o + ā + van

→ av for o (SK 61) → avāvan → avāvā.

One of Bhaṭṭoji's commentators gives these steps.

Bhaṭṭoji then illustrates vic from SK 2980: roṭ, reṭ, sugaṇ. One commentator: roṭ, reṭ are vic after ruṣ, riṣ "to injure"; the v is lost by SK 375. The other commentator adds guṇa of the penultimate vowel. sugaṇ is vic after gaṇ.

Watch out. vijāvā and avāvā are also the vanip examples promised at SK 2980. Only these two show the ā of this rule. roṭ, reṭ and sugaṇ simply illustrate vic.

SK 2983 · क्विप् च · kvip ca (A 3.2.76)

क्विप् च

kvip ca · A 3.2.76

अयमपि दृश्यते । सत्सूद्विष - SK 2975 इति त्वस्यैव प्रपञ्चः । उखास्रत् । पर्णध्वत् । वाहभ्रट् ॥

Rule. kvip too is seen. One of Bhaṭṭoji's commentators widens it: after any root, with or without an upapada, in the Veda and in ordinary usage.

Example.

ukhāyāḥ sraṃsate "it slips from the pot".

ukhā + sraṃs + kvip → the nasal is lost; s becomes d, then t → ukhāsrat.

Also parṇadhvat "falling from the leaf" and vāhabhraṭ "falling from the horse".

One commentator gives the analyses and the steps.

Applies when / blocked by. Bhaṭṭoji: SK 2975 is only an elaboration (prapañca) of this very rule. One commentator explains. kvip after any root already covers the twelve roots of SK 2975. So that rule adds nothing new; it spells out cases.

Watch out. The other commentator reports a variant. The Vṛtti reads vahābhraṭ, with vaha "shoulder", according to Haradatta. Bhaṭṭoji's reading is vāhabhraṭ.

II.G. What kvip does to the root, and ka/kvip after sthā — running example: `aṅga + gam + kvip → aṅga-ga → aṅga-ga-t → aṅgagat`; `he prāṇ`; `śam + sthā + ka → śaṃsthaḥ`

kvip leaves a bare root at the end of a pada. Several rules then reshape that root. SK 2984 gives ṇ in he prāṇ. SK 2985 shortens chādi. SK 2986 drops the nasal of gam. Under the first two, Bhaṭṭoji runs through many other kvip forms. SK 2987 returns to affixes: ka and kvip after sthā "to stand". The running example aṅgagat "going to Aṅga" shows the nasal lost and t added.

SK 2984 · अन्तः · antaḥ (A 8.4.20)

अन्तः

antaḥ · A 8.4.20

पदान्तस्यानितेर्नस्य णत्वं स्यादुपसर्गस्थान्निमित्तात्परश्चेत् । हे प्राण् । शास इत् - SK 2486 इतीत्वम् । मित्राणि शास्ति मित्रशीः ।

आशासः क्वावुपाधाया इत्वं वाच्यम्

(वार्तिकम्) ॥ आशीः इत्वोत्वे । गीः । पूः ॥

Rule. The n of the root an "to breathe" becomes ṇ at the end of a pada. The condition: a cause of ṇatva stands before it, in an upasarga. One of Bhaṭṭoji's commentators lists what runs on. aniteḥ (SK 2478) gives the root an. raṣābhyāṃ no ṇaḥ (SK 235) gives the change. upasargāt (SK 2287) gives the upasarga.

Example.

pra + an + kvip → the affix goes → pra-an → prān

→ ṇ for the pada-final n after the r of pra → prāṇ

→ vocative singular → he prāṇ.

One commentator reads an here as an prāṇane "to breathe" with kvip. Both say why the example is a vocative. Elsewhere the final n would be lost. In the vocative na ṅisaṃbuddhyoḥ (SK 352) bars that loss.

Applies when / blocked by. Bhaṭṭoji glosses antaḥ as "at the end of a pada". One commentator explains: the end is reckoned from the following word. The n of an is always final in the root itself, so that reading would be empty.

Bhaṭṭoji then gives more kvip forms.

mitrāṇi śāsti "he instructs friends". śās + kvip → i for the ā by śāsa id aṅhaloḥ (SK 2486) → mitraśīḥ. One commentator adds the lengthening by SK 433.

A vārttika: āśāsaḥ kvāv upadhāyā itvaṃ vācyam. Before kvi, i replaces the penultimate of āśās "to wish" → āśīḥ.

With i and u replacing ṝ: gīḥ (from gṝ), pūḥ (from pṝ). One commentator gives SK 2390 for the i and SK 2494 for the u.

Watch out. Is the āśās vārttika a restriction or a fresh rule? One of Bhaṭṭoji's commentators reports both views; the other takes it as a fresh rule. Scholars discuss this.

SK 2985 · इस्मन्त्रन्क्विषु च · ismantrankviṣu ca (A 6.4.97)

इस्मन्त्रन्क्विषु च

ismantrankviṣu ca · A 6.4.97

एषु छादेर्ह्रस्वः स्यात् । तनुच्छत् । अनुनासिकस्य क्वि -SK 2666 इति दीर्घः । मो नो धातोः SK 341 । प्रतान् । प्रशान् । च्छ्वोः -SK 2561 इत्यूठ् । अक्षद्यूः । ज्वरत्वरः -SK 2654 इत्यूठ् । जूः । जूरौ । जूरः । तूः । स्रूः । ऊः । वृद्धिः । जनानवतीति जनौः । जनावौ । जनावः । मूः । मुवौ । मुवः । सुमूः । सुम्वौ । सुम्वः । राल्लोपः SK 2655 । मूर्छा, मूः । मुरौ । मुरः । धुर्वी, धूः ॥

Rule. Before is, man, tran and kvi, a short vowel replaces the ā of chādi "to cover". One of Bhaṭṭoji's commentators gives the sources. chādeḥ runs on from SK 3297, hrasvaḥ from SK 2955.

Example.

tanuṃ chādayati "it covers the body".

tanu + chādi + kvip → short a for ā → tanu + chad → tanucchat.

The same commentator gives the other three: chadiḥ with is, chadma with man, chatram with tran.

Applies when / blocked by. Bhaṭṭoji now runs through other kvip forms, naming only the rules.

Lengthening by anunāsikasya kvijhaloḥ kṅiti (SK 2666), and n for m by mo no dhātoḥ (SK 341): pratān, praśān. One commentator gives the roots tan and śam.
The substitute ū (ūṭh) by cchvoḥ (SK 2561): akṣadyūḥ "one who plays with dice" (akṣair dīvyati). One commentator gives ūṭh for the v of div.
ūṭh by jvaratvara… (SK 2654): jūḥ, jūrau, jūraḥ; tūḥ; srūḥ; ūḥ.
vṛddhi: janān avati "he protects people" → janauḥ, janāvau, janāvaḥ. One commentator gives the rule, SK 73.
mūḥ, muvau, muvaḥ; sumūḥ, sumvau, sumvaḥ. One commentator gives oḥ supi (SK 281) for the v of sumvau.
rāl lopaḥ (SK 2655): from mūrchā, mūḥ, murau, muraḥ; from dhurvī, dhūḥ.

Watch out. Only tanucchat belongs to this sūtra. The rest of the list is a tour of kvip forms. Each line names its own rule.

SK 2986 · गमः क्वौ · gamaḥ kvau (A 6.4.40)

गमः क्वौ

gamaḥ kvau · A 6.4.40

अनुनासिकलोपः स्यात् । अङ्गत् ।

गमादीनामिति वक्तव्यम्

(वार्तिकम्) ॥ परीतत् । संयत् । सुनुत् ।

ऊङ् च गमादीनामिति वक्तव्यम्

(वार्तिकम्) ॥ लोपश्च ॥ अग्रेगूः । अग्रेभ्रूः ॥

Rule. Before kvi, the nasal of gam is lost. Bhaṭṭoji supplies "the nasal is lost". One of his commentators says it comes from SK 2428. That rule works only before an affix beginning with a jhal consonant. kvi is not one, so this rule is needed.

Example.

aṅgaṃ gacchati "goes to Aṅga".

aṅga + gam + kvip → the affix goes → m is lost → aṅga-ga

→ tuk after the short vowel (SK 2858) → aṅga-ga-t → aṅgagat.

One commentator gives the m-loss and tuk, and adds vaṅgagat, kaliṅgagat.

Applies when / blocked by. A vārttika: gamādīnām iti vaktavyam, "of gam and the rest". parītat, saṃyat, sunut. One of Bhaṭṭoji's commentators reads and explains purītat and sunat; scholars discuss this.

A second vārttika: ūṅ ca gamādīnām iti vaktavyam. ū replaces the penultimate of gam and the rest, and the nasal is lost. agre gacchati "goes in front" → agregūḥ; likewise agrebhrūḥ from bhram. One commentator says the locative of agre stays by SK 972.

SK 2987 · स्थः क च · sthaḥ ka ca (A 3.2.77)

स्थः क च

sthaḥ ka ca · A 3.2.77

चात् क्विप् । शंस्थः । शंस्थाः । शमि धातोः - SK 2928 इत्यचं बाधितुं सूत्रम् ॥

Rule. sthā takes ka, and by ca, kvip. One of Bhaṭṭoji's commentators gives the upapada: a sup-ending word, with or without an upasarga. Under SK 2988 he says supi runs on from SK 2916.

Example.

śam + sthā + ka → k is a marker → the ā is lost (SK 2372) → śaṃsthaḥ.

śam + sthā + kvip → the affix goes → śaṃsthāḥ.

Applies when / blocked by. ka already comes by supi sthaḥ (SK 2916), and kvip by kvip ca (SK 2983). Bhaṭṭoji: this sūtra blocks the ac of śami dhātoḥ (SK 2928). Both of his commentators explain. That rule names "a root", so its ac would come after every root after śam. It would set aside ka and kvip after sthā, as it sets aside ṭa after kṛñ.

Watch out. In śaṃsthāḥ the ā does not become ī by SK 2462. Bhaṭṭoji's commentators give different grounds for this; scholars discuss this.

II.H. ṇini: habit, comparison, vow, repetition, and man — running example: `uṣṇa + bhuj + ṇini → uṣṇa-bhoj-in → uṣṇabhojī`

Five rules teach ṇini with a sup-ending upapada. ṇini leaves in, and its ṇ brings vṛddhi to the root. A light penultimate vowel takes guṇa instead, by SK 2189. SK 2988 is the general rule: ṇini for habitual disposition (tācchīlya), when the upapada is not a word for a genus. SK 2989–2992 add cases that SK 2988 would not reach. They cover comparison, a vow, repetition, and the root man "to think". The running example is uṣṇabhojī "one who habitually eats his food hot".

SK 2988 · सुप्यजातौ णिनिस्ताच्छील्ये · supy ajātau ṇinis tācchīlye (A 3.2.78)

सुप्यजातौ णिनिस्ताच्छील्ये

supyajātau ṇinistācchīlye · A 3.2.78

अजात्यर्थे सुपि धातोर्णिनिः स्यात्ताच्छील्ये द्योत्ये । उष्णभोजी । शीतभोजी । अजातौ किम् । ब्राह्मणानाममन्त्रयिता । ताच्छील्ये किम् । उष्णं भुङ्क्ते कदाचित् । इह वृत्तिकारेणोपसर्गभिन्न एव सुपि णिनिरिति व्याख्याय उत्प्रतिभ्यामाङि सर्तेरुपसंख्यानमिति पठितम् । हरदत्तमाधवीदिभिश्च तदेवानुसृतम् । तच्च भाष्यविरोधादुपेक्ष्यम् । प्रसिद्धश्चोपसर्गेऽपि णिनिः । स बभूवोपजीविनाम् । अनुयायिवर्गः । पतत्यधो धाम विसारि । न वञ्चनीयाः प्रभवोऽनुजीविभिरित्यादौ ।

साधुकारिण्युपसंख्यानम्

(वार्तिकम्) ।

ब्रह्मणि वदः

(वार्तिकम्) ॥ अताच्छील्यार्थं वार्तिकद्वयम् । साधुकारी । ब्रह्मवादी ॥

Rule. A root takes ṇini when a habitual disposition (tācchīlya) is conveyed. The upapada is a sup-ending word whose meaning is not a genus (jāti).

Example.

uṣṇa + bhuj + ṇini → uṣṇa + bhuj + in

→ the light penultimate u takes guṇa (SK 2189) → uṣṇa + bhojin → uṣṇabhojī.

Likewise śītabhojī "one who habitually eats it cold".

Not a genus? Bhaṭṭoji: brāhmaṇān āmantrayitā "one who habitually invites brāhmaṇas". brāhmaṇa is a genus word, so there is no ṇini. One of his commentators says tṛn comes instead, because habit is meant.

Not a habit? uṣṇaṃ bhuṅkte kadācit "he eats his food hot now and then": here there is no ṇini.

Applies when / blocked by. Bhaṭṭoji reports the Vṛtti. Its author held that ṇini comes only with a sup-word other than an upasarga. So he read a vārttika to add sṛ after ut or prati with āṅ. Haradatta, Mādhava and others followed him. Bhaṭṭoji rejects this "because it conflicts with the Bhāṣya". ṇini with an upasarga is well established in usage. He quotes four phrases:

sa babhūvopajīvinām "he became foremost of the dependants";
anuyāyivargaḥ "the band of followers";
patatyadho dhāma visāri "the spreading radiance falls downward";
na vañcanīyāḥ prabhavo 'nujīvibhiḥ "masters are not to be deceived by dependants".

Each has ṇini after a root with an upasarga: upa-jīvin, anu-yāyin, vi-sārin, anu-jīvin. Both of his commentators side with the Bhāṣya. One calls the vārttika's place in the Bhāṣya an interpolation.

Two vārttikas give ṇini without habit, Bhaṭṭoji says.

sādhukāriṇy upasaṃkhyānam: in the sense "one who does it well" → sādhukārī.
brahmaṇi vadaḥ: after vad with brahman "the Veda" → brahmavādī.

One commentator takes kṛ in the first as standing for any root: sādhudāyī too.

Watch out. Is brahman not a genus word? One of Bhaṭṭoji's commentators reports the answer of the moderns (navya). SK 2988 excludes only a genus of living beings. That fits the counter-example brāhmaṇān āmantrayitā.

SK 2989 · कर्तर्युपमाने · kartary upamāne (A 3.2.79)

कर्तर्युपमाने

kartaryupamāne · A 3.2.79

णिनिः स्यात् । उपदार्थः कर्ता प्रत्ययार्थस्य कर्तुरुपमानम् । उष्ट्र इव क्रोशति उष्ट्रक्रोशी । ध्वाङ्क्षरावी । अताच्छील्यार्थं जात्यर्थं च सूत्रम् । कर्तरि किम् । अपूपानिव भक्षयति माषान् । उपमाने किम् । उष्ट्रः क्रोशति ॥

Rule. ṇini comes when the upapada names an agent that is the standard of comparison (upamāna). Bhaṭṭoji: "the agent which is the upapada's meaning is the standard for the agent which is the affix's meaning".

Example.

uṣṭra iva krośati "he cries out like a camel".

uṣṭra + kruś + ṇini → guṇa (SK 2189) → uṣṭra + krośin → uṣṭrakrośī.

Likewise dhvāṅkṣarāvī "cawing like a crow".

Applies when / blocked by. Bhaṭṭoji: the sūtra gives ṇini without habit, and with a genus word. uṣṭra "camel" and dhvāṅkṣa "crow" are genus words, which SK 2988 would exclude.

Why "agent"? apūpān iva bhakṣayati māṣān "he eats the beans as if they were cakes". The cakes are compared, but they are objects, not agents. No ṇini.

Why "standard of comparison"? uṣṭraḥ krośati "the camel cries out". Nothing is compared. No ṇini.

SK 2990 · व्रते · vrate (A 3.2.80)

व्रते

vrate · A 3.2.80

णिनिः स्यात् । स्थण्डिलशायी ॥

Rule. ṇini comes when a vow (vrata) is understood. One of Bhaṭṭoji's commentators gives the upapada: a sup-ending word.

Example.

sthaṇḍila + śī + ṇini → vṛddhi of the final vowel (SK 254), and ai becomes āy (SK 61) → sthaṇḍila + śāyin

→ sthaṇḍilaśāyī "one who, by vow, lies on the bare ground".

Watch out. The condition here is the vow, not habit. The commentator's gloss is "when a vow is understood" (vrate gamye).

SK 2991 · बहुलमाभीक्ष्ण्ये · bahulam ābhīkṣṇye (A 3.2.81)

बहुलमाभीक्ष्ण्ये

bahulamābhīkṣṇye · A 3.2.81

पौनःपुन्ये द्योत्ये सुप्युपपदे णिनिः । क्षीरपायिण उशीनराः ॥

Rule. Variously (bahulam), ṇini comes when repetition (ābhīkṣṇya) is conveyed. Bhaṭṭoji glosses it as "again and again" (paunaḥpunya). The upapada is a sup-ending word.

Example.

kṣīra + pā + ṇini → pā ends in ā, so it takes yuk before the ṇ-marked kṛt (SK 2761) → kṣīra + pāyin

→ ṇ for the stem-final n, optionally, after the r in the first member (SK 1055, with SK 197 for the sounds between) → kṣīrapāyiṇaḥ.

kṣīrapāyiṇa uśīnarāḥ "the Uśīnaras drink milk again and again".

Neither Bhaṭṭoji nor his commentator gives the ṇ step here.

Applies when / blocked by. One of Bhaṭṭoji's commentators says this rule lets a genus word be the upapada. kṣīra "milk" is one, and Bhaṭṭoji chose it to suggest this.

Watch out. Repetition (ābhīkṣṇya) is not the same as habit (tācchīlya, SK 2988). Read the example as "again and again", not "by habit".

SK 2992 · मनः · manaḥ (A 3.2.82)

मनः

manaḥ · A 3.2.82

सुपि मन्यतेर्णिनिः स्यात् । दर्शनीयमानी ॥

Rule. man takes ṇini when a sup-ending word is upapada. Bhaṭṭoji writes manyateḥ: the man of the divādi class, manyate "thinks".

Example.

darśanīya + man + ṇini → vṛddhi of the penultimate a (SK 2282) → darśanīya + mānin

→ darśanīyamānī "one who thinks another handsome".

One commentator: darśanīyamānī devadatto yajñadattasya, Devadatta who thinks Yajñadatta handsome.

Applies when / blocked by. Why only the divādi man? Both of Bhaṭṭoji's commentators answer: bahulam runs on from SK 2991. So the tanādi manu "to understand" is not taken. One adds a consequence for the next rule. With khaś the divādi man takes śyan, not the tanādi affix u.

II.I. Thinking of oneself: khaś, and am for an ic-final monosyllable — running example: `paṇḍita + man + śyan + khaś → paṇḍita-m-manya → paṇḍitaṃmanyaḥ`; `go + am + manyaḥ → gāṃmanyaḥ`

SK 2993 gives khaś after man when the thinker thinks of himself. khaś has two markers that both act. Its ś makes it sārvadhātuka, a name for verb-endings and ś-marked affixes. So the class sign śyan (ya) comes between root and affix. Its kh gives the upapada m (SK 2942). SK 2994 then puts am in place of that m after a one-vowel word ending in an ic vowel. ic is the pratyāhāra of the vowels other than a and ā. The running examples: paṇḍitaṃmanyaḥ "who thinks himself learned", and gāṃmanyaḥ "who thinks himself a cow".

SK 2993 · आत्ममाने खश्च · ātmamāne khaś ca (A 3.2.83)

आत्ममाने खश्च

ātmamāne khaśca · A 3.2.83

स्वकर्मके मनने वर्तमानान्मन्यतेः सुपि खश् स्यात् चाण्णिनिः । पण्डितमात्मानं मन्यते पण्डितंमन्यः । पण्डितमानी । खित्यनव्ययस्य SK 2943 । कालिंमन्या । अनव्ययस्य किम् । दिवामन्या ॥

Rule. manyate takes khaś when the thinking has oneself as its object (ātmamāna). The upapada is a sup-ending word. By ca, ṇini also comes.

Example.

paṇḍitam ātmānaṃ manyate "he thinks himself learned".

paṇḍita + man + khaś → ś makes khaś sārvadhātuka → śyan → paṇḍita + man-ya + a → paṇḍita + manya

→ kh gives m to the upapada (SK 2942) → paṇḍitaṃmanyaḥ.

With ṇini: paṇḍitamānī.

One of Bhaṭṭoji's commentators gives the śyan and the m.

Applies when / blocked by. A feminine upapada is shortened by khity anavyayasya (SK 2943). kālīm ātmānaṃ manyate → kāliṃmanyā "she who thinks herself Kālī". Both commentators add a point. This shortening is the later rule. So it blocks the masculine form of the stem (puṃvadbhāva) by striyāḥ puṃvat… (SK 831).

Why "not an indeclinable" in SK 2943? Bhaṭṭoji: divāmanyā. divā "by day" is indeclinable, so it keeps its ā and takes no m.

Watch out. The ca has a further job. ṇini would come anyway beside khaś, by vāsarūpa. One commentator says ca keeps ṇini alive, so that ṇini, not khaś, runs on into SK 2996.

SK 2994 · इच एकाचोऽम्प्रत्ययवच्च · ica ekāco 'm pratyayavac ca (A 6.3.68)

इच एकाचोऽम्प्रत्ययवच्च

ica ekāco'mpratyayavacca · A 6.3.68

इजन्तादेकाचोऽम् स्यात्स च स्वाद्यम्वत्खिदन्ते परे । औतोऽम्शसोः SK 285 । गांमन्यः । वाम्शसोः SK 302 । स्त्रियंमन्यः । स्त्रींमन्यः । नृ । नरंमन्यः । भुवंमन्यः । श्रियमात्मानं मन्यते श्रिमन्यं कुलम् । भाष्यकारवचनात् श्रीशब्दस्य ह्रस्वो मुममोरभावश्च ॥

Rule. A first member that ends in an ic vowel and has one vowel (ekāc) takes am. This happens before a second member ending in a kh-marked affix. And this am is treated like the case-affix am. One of Bhaṭṭoji's commentators gives the sources. khiti runs on from SK 2943, and "second member" from SK 958. The same commentator calls am an exception (apavāda) to mum.

Example.

gām ātmānaṃ manyate "he thinks himself a cow".

go + man + khaś → go + am + manya

→ am counts as the case-affix, so auto 'mśasoḥ (SK 285) applies: ā for o and the vowel of am → gām + manya

→ gāṃmanyaḥ.

With strī "woman": iyaṅ optionally by vām-śasoḥ (SK 302) → striyaṃmanyaḥ, strīṃmanyaḥ.

With nṛ "man": guṇa (SK 275) → naraṃmanyaḥ.

With bhū "earth": bhuvaṃmanyaḥ.

One commentator names the rules for nara- (SK 275) and bhuva- (SK 271).

Applies when / blocked by. Why "like the case-affix"? So that the case-affix rules (SK 285, SK 302, SK 275) apply to it. One commentator says bhuvaṃmanyaḥ too needs this: its uvaṅ is taught before a vowel-initial affix. The other commentator adds two more effects. The m of am is not a marker, by na vibhaktau tusmāḥ (SK 190). And am comes after the word, like an affix.

Watch out. Bhaṭṭoji gives a special case. śriyam ātmānaṃ manyate "it thinks itself Śrī" → śrimanyaṃ kulam "a family that thinks itself Śrī". On the Bhāṣya's authority, śrī is shortened, and neither mum nor am appears. One commentator supplies the rule: hrasvo napuṃsake prātipadikasya (SK 318). The family is neuter, so śrī takes the neuter's short vowel. Both commentators say the am is then removed by svamor napuṃsakāt (SK 319).

II.J. The bhūte heading; ṇini in the past; kvip of han under restriction — running example: `brahman + han + kvip → brahma-han → brahmahā`

SK 2995 opens a heading: "in the past" (bhūte). Every affix from here to SK 2151 comes after a root that denotes a past action. SK 2996–2997 give ṇini in the past after yaj and han. SK 2997 also opens a second heading, karmaṇi "with an object upapada", running to SK 3006. SK 2998 gives kvip after han with three objects. It is the first of several restrictions (niyama). A rule that only repeats what another rule gives must be read as limiting it. The running example is brahmahā "slayer of a brāhmaṇa".

SK 2995 · भूते · bhūte (A 3.2.84)

भूते

bhūte · A 3.2.84

अधिकारोऽयम् । वर्तमाने लट् SK 2151 । इति यावत् ॥

Rule. This is a heading (adhikāra): "in the past". It runs up to vartamāne laṭ (SK 2151), not including it. Both of Bhaṭṭoji's commentators say it qualifies the root. dhātoḥ (SK 2829) is in force, so the sense is "after a root that denotes the past".

Example.

soma + yaj + ṇini (SK 2996) → somayājī "one who has sacrificed with soma".

The past sense comes from this heading.

Watch out. The heading stops before SK 2151. One of Bhaṭṭoji's commentators says that limit rests on explanation (vyākhyāna), not on the wording.

SK 2996 · करणे यजः · karaṇe yajaḥ (A 3.2.85)

करणे यजः

karaṇe yajaḥ · A 3.2.85

करणे उपपदे भूतार्थाद्यजेर्णिनिः स्यात्कर्तरि । सोमेनेष्टवान् सोमयाजी । अग्निष्टोमयाजी ॥

Rule. With an instrument-word (karaṇa) as upapada, yaj "to sacrifice" in the past takes ṇini, in the agent sense. One of Bhaṭṭoji's commentators says ṇini alone runs on, because it is marked svarita. khaś of SK 2993 does not.

Example.

somena iṣṭavān "he has sacrificed with soma".

soma + yaj + ṇini → vṛddhi of the penultimate a (SK 2282) → soma + yājin → somayājī.

Likewise agniṣṭomayājī "one who has sacrificed with the agniṣṭoma".

Watch out. soma here is the instrument, not the object. Both commentators take somena as "with the soma creeper", the means of the sacrifice.

SK 2997 · कर्मणि हनः · karmaṇi hanaḥ (A 3.2.86)

कर्मणि हनः

karmaṇi hanaḥ · A 3.2.86

पितृव्यघाती । कर्मणीत्येतत् सहे च SK 3006 इति यावदधिक्रियते ॥

Rule. With an object-word (karman) as upapada, han in the past takes ṇini. One of Bhaṭṭoji's commentators supplies ṇini. Bhaṭṭoji adds: karmaṇi is a heading up to sahe ca (SK 3006).

Example.

pitṛvyaṃ hatavān "he has killed his paternal uncle".

pitṛvya + han + ṇini → t for n (SK 2574) → gh for h (SK 358) → vṛddhi → pitṛvya + ghātin

→ pitṛvyaghātī.

One commentator gives SK 2574 and SK 358.

Watch out. The other commentator reports the Kāśikā. It requires the object to be contemptible (kutsita), so that coraṃ hatavān "he killed a thief" gives no ṇini. Bhaṭṭoji says nothing of this.

SK 2998 · ब्रह्मभ्रूणवृत्रेषु क्विप् · brahmabhrūṇavṛtreṣu kvip (A 3.2.87)

ब्रह्मभ्रूणवृत्रेषु क्विप्

brahmabhrūṇavṛtreṣu kvip · A 3.2.87

एषु कर्मसूपपदेषु हन्तेर्भूते क्विप्स्यात् । ब्रह्महा । भ्रूणहा । वृत्रहा । क्विप् च SK 2983 इत्येव सिद्धे नियमार्थमिदम् । ब्रह्मादिष्वेव हन्तेरेव भूते एव क्विबेवेति चतुर्विधोऽत्र नियम इति काशिका । ब्रह्मादिष्वेव क्विबेवेति द्विविधो नियम इति भाष्यम् ॥

Rule. With brahman, bhrūṇa "embryo" or vṛtra as object-upapada, han in the past takes kvip.

Example.

brahman + han + kvip → the affix goes → brahma-han → brahmahā "slayer of a brāhmaṇa".

Likewise bhrūṇahā "slayer of an embryo", vṛtrahā "slayer of Vṛtra".

Applies when / blocked by. kvip ca (SK 2983) would already give these forms. So, Bhaṭṭoji says, this rule is for restriction (niyamārtha). He reports two readings.

The Kāśikā: fourfold. Only with brahman etc.; only after han; only in the past; only kvip.
The Bhāṣya: twofold. Only with brahman etc.; only kvip.

One of Bhaṭṭoji's commentators gives an example for each of the four. "Only with brahman": no kvip in puruṣaṃ hatavān. "Only after han": no kvip in brahma adhītavān. "Only in the past": no kvip in brahma hanti. "Only kvip": no aṇ in brahma hatavān. The other commentator assigns the four restrictions differently; scholars discuss this.

Watch out. "Only kvip" does not shut out the past participle. Both commentators say the restriction excludes only affixes taught with an upapada. So brahma hatavān "he killed a brāhmaṇa" stands.

II.K. kvip of kṛñ, su, ci, and ini of vi-krī — running example: `su + kṛ + kvip → su-kṛ-t → sukṛt`

Four more rules give kvip in the past with fixed upapadas, after kṛñ, suñ "to press" and ci "to pile". Like SK 2998 they repeat what kvip ca (SK 2983) gives, so they too restrict. Bhaṭṭoji reports how far each restriction reaches. SK 3003 gives ini after vi-krī "to sell". A short root vowel before kvip takes the augment t (tuk, SK 2858). So the running example sukṛt "one who has done well" ends in t.

SK 2999 · सुकर्मपापमन्त्रपुण्येषु कृञः · sukarmapāpamantrapuṇyeṣu kṛñaḥ (A 3.2.89)

सुकर्मपापमन्त्रपुण्येषु कृञः

sukarmapāpamantrapuṇyeṣu kṛñaḥ · A 3.2.89

सौ कर्मादिषु च कृञः क्विप्स्यात् । त्रिविधोऽत्र नियम इति काशिका । सुकृत् । कर्मकृत् । पापकृत् । मन्त्रकृत् । पुण्यकृत् । क्विबेवेति नियमात्कर्म कृतवानित्यत्राण्न । कृञ एवेति नियमान्मन्त्रमधीतवान्मन्त्राध्यायः । अत्र न क्विप् । भूत एवेति नियमान्मन्त्रं करोति करिष्यति वेति विवक्षायां न क्विप् । स्वादिष्वेवेति नियमाभावादन्यस्मिन्नप्युपपदे क्विप् । शास्त्रकृत् । भाष्यकृत् ॥

Rule. With su, karman, pāpa, mantra or puṇya as upapada, kṛñ in the past takes kvip. Bhaṭṭoji supplies kvip.

Example.

su + kṛ + kvip → the affix goes → tuk after the short ṛ (SK 2858) → su-kṛ-t → sukṛt "one who has done well".

Likewise karmakṛt "one who has done the work", pāpakṛt, mantrakṛt, puṇyakṛt.

Applies when / blocked by. "The restriction here is threefold", says the Kāśikā, and Bhaṭṭoji works it out.

"Only kvip": in karma kṛtavān "he has done the work", aṇ does not come.
"Only after kṛñ": mantram adhītavān "he has studied the mantra" gives mantrādhyāyaḥ, not kvip.
"Only in the past": for mantraṃ karoti or kariṣyati "makes / will make", no kvip.

There is no restriction "only with su etc.". So kvip comes with other upapadas too: śāstrakṛt "author of a śāstra", bhāṣyakṛt "author of a Bhāṣya".

Watch out. One of Bhaṭṭoji's commentators repeats the point of SK 2998. ktavatu still comes: karma kṛtavān.

SK 3000 · सोमे सुञः · some suñaḥ (A 3.2.90)

सोमे सुञः

some suñaḥ · A 3.2.90

सोमसुत् । चतुर्विधोऽत्र नियम इति काशिका । एवमुत्तरसूत्रेऽपि ॥

Rule. With soma as object-upapada, suñ "to press" in the past takes kvip. One of Bhaṭṭoji's commentators supplies "object", "past" and kvip.

Example.

soma + su + kvip → tuk (SK 2858) → somasut "one who has pressed soma".

Applies when / blocked by. "The restriction here is fourfold", says the Kāśikā, and Bhaṭṭoji adds: so too in the next sūtra. One commentator gives the effects. With aṇ: no somasāva for "has pressed soma". Other roots with soma take aṇ: somakrāya. Other objects with su take aṇ: surāsāva. And no kvip for "presses" or "will press".

SK 3001 · अग्नौ चेः · agnau ceḥ (A 3.2.91)

अग्नौ चेः

agnau ceḥ · A 3.2.91

अग्निचित् ॥

Rule. With agni as object-upapada, ci in the past takes kvip.

Example.

agni + ci + kvip → tuk (SK 2858) → agnicit "one who has piled the fire-altar".

One of Bhaṭṭoji's commentators: one who has piled with bricks the altar-ground called agni.

Applies when / blocked by. The restriction is fourfold, as at SK 3000. One commentator: agnicāya with aṇ does not occur for "has piled the fire"; with other objects aṇ comes: iṣṭakācāya.

SK 3002 · कर्मण्यग्न्याख्यायाम् · karmaṇy agnyākhyāyām (A 3.2.92)

कर्मण्यग्न्याख्यायाम्

karmaṇyagnyākhyāyām · A 3.2.92

कर्मण्युपपदे कर्मण्येव कारके चिनोतेः क्विप्स्यात् अग्न्याधारस्थलविशेषस्याख्यायाम् । श्येन इव चितः श्येनचित् ॥

Rule. With an object-word as upapada, ci takes kvip in the sense of the object itself. The condition: the whole names a site that holds the fire.

Example.

śyena iva citaḥ "piled like a falcon".

śyena + ci + kvip → tuk → śyenacit.

One commentator: śyena "falcon" here means falcon-shaped. The other: the compound is the fixed name of a brick site for the āhavanīya fire.

Applies when / blocked by. The sūtra's karmaṇi and the heading's karmaṇi do two jobs. One of Bhaṭṭoji's commentators keeps them apart. The heading's karmaṇi gives the upapada. The sūtra's karmaṇi gives the affix its sense. So kartari kṛt (SK 2832) does not apply here.

Watch out. śyenacit is not "one who piled a falcon". The affix means the thing piled.

SK 3003 · कर्मणीनिविक्रियः · karmaṇīni vikriyaḥ (A 3.2.93)

कर्मणीनिविक्रियः

karmaṇīnivikriyaḥ · A 3.2.93

कर्मण्युपपदे विपूर्वात्क्रीणातेरिनिः स्यात् ।

कुत्सितग्रहणं कर्तव्यम्

(वार्तिकम्) ॥ सोमविक्रयी । घृतविक्रयी ॥

Rule. With an object-word as upapada, krī "to buy" with vi "to sell" takes ini in the past. ini leaves in; unlike ṇini it has no ṇ.

Example.

soma + vi + krī + ini → guṇa (SK 2168) → vi-kre-in → ay for e (SK 61) → vikrayin

→ somavikrayī "one who has sold soma". Likewise ghṛtavikrayī.

Applies when / blocked by. A vārttika: kutsitagrahaṇaṃ kartavyam. The word "contemptible" is to be added. The rule holds only when the object brings contempt on the seller. Both of Bhaṭṭoji's commentators explain. Selling soma is forbidden, so its seller is contemned. One commentator gives the counter-example dhānyavikrāyaḥ "grain-seller", with aṇ.

II.L. kvanip after dṛś, yudh and kṛñ; karmaṇi ends — running example: `pāraṃ dṛṣṭavān → pāra + dṛś + kvanip → pāradṛśvā`

Three rules give kvanip in the past. kvanip leaves van, and its k keeps the root from guṇa. SK 3004 gives it after dṛś "to see", SK 3005 after yudh and kṛñ with rājan, SK 3006 with saha. With SK 3006 the heading karmaṇi ends. The running example pāradṛśvā means "one who has seen the far shore".

SK 3004 · दृशेः क्वनिप् · dṛśeḥ kvanip (A 3.2.94)

दृशेः क्वनिप्

dṛśeḥ kvanip · A 3.2.94

कर्मणि भूत इत्येव । पारं दृष्टवान् पारदृश्वा ॥

Rule. dṛś takes kvanip. Bhaṭṭoji: "with an object" and "in the past" continue.

Example.

pāraṃ dṛṣṭavān "he has seen the far shore".

pāra + dṛś + kvanip → pāra + dṛś + van → no guṇa before the k-marked affix (SK 2217)

→ pāradṛśvan → pāradṛśvā.

Applies when / blocked by. kvanip is already available by SK 3418 and SK 2980. Both of Bhaṭṭoji's commentators therefore read this rule as a restriction. It shuts out manin and vanip, and aṇ and the other upapada affixes. It does not shut out the niṣṭhā (dṛṣṭavān).

Watch out. One commentator says other times are not excluded. So viśvadṛśvan "seeing the universe", analysed in the present, is no fault.

SK 3005 · राजनि युधिकृञः · rājani yudhikṛñaḥ (A 3.2.95)

राजनि युधिकृञः

rājani yudhikṛñaḥ · A 3.2.95

क्वनिप्स्यात् । युधिरन्तर्भावितण्यर्थः । राजानं योधितवान् राजयुध्वा । राजकृत्वा ॥

Rule. With rājan "king" as object-upapada, yudh "to fight" and kṛñ take kvanip, in the past.

Example.

rājānaṃ yodhitavān "he made the king fight".

rājan + yudh + kvanip → rāja + yudh + van → rājayudhvā.

rājan + kṛ + kvanip → tuk after the short ṛ (SK 2858) → rāja-kṛ-t-van → rājakṛtvā "one who made the king".

Applies when / blocked by. yudh takes no object, so how can an object be its upapada? Bhaṭṭoji: yudh here has the causative sense taken into it (antarbhāvita-ṇyartha). It means "make fight". One commentator raises and answers the objection in the same way.

Watch out. Is this a restriction like SK 3004, or a fresh rule? One commentator cites the Śabdakaustubha. These forms could come through SK 2980 only as a last resort. So SK 3004–3006 may as well be called prescriptive. The older works are silent; scholars discuss this.

SK 3006 · सहे च · sahe ca (A 3.2.96)

सहे च

sahe ca · A 3.2.96

कर्मणीति निवृत्तम् । सहयुध्वा । सहकृत्वा ॥

Rule. With saha "together" as upapada, yudh and kṛñ also take kvanip. Bhaṭṭoji: karmaṇi has stopped here.

Example.

saha + yudh + kvanip → sahayudhvā "one who fought together with another".

saha + kṛ + kvanip → tuk → sahakṛtvā "one who did something together with another".

Watch out. The heading karmaṇi began at SK 2997. Bhaṭṭoji's remark marks its end, so saha need not be an object.

II.M. ḍa after jan — running example: `sarasi + jan + ḍa → sarasi-j-a → sarasijam`

Five rules give ḍa after jan "to be born", in the past. As at SK 2965, the ḍ-marker removes the ṭi an. SK 3007 takes a seventh-case upapada, SK 3008 a fifth-case one, SK 3009 an upasarga in a name, and SK 3010 anu with an object. SK 3011 lets ḍa be seen elsewhere too, and lifts every condition. The running example sarasijam "lotus" means "born in a lake".

SK 3007 · सप्तम्यां जनेर्डः · saptamyāṃ janer ḍaḥ (A 3.2.97)

सप्तम्यां जनेर्डः

saptamyāṃ janerḍaḥ · A 3.2.97

सरसिजम् । मन्दुरायां जातो मन्दुरजः । ङ्यापोः - SK 1001 इति ह्रस्वः ॥

Rule. With a seventh-case word as upapada, jan in the past takes ḍa.

Example.

sarasi + jan + ḍa → the ṭi an is lost by the ḍ-marking → sarasi + j + a → sarasijam "lotus".

mandurāyāṃ jātaḥ "born in a stable" → mandurajaḥ. Bhaṭṭoji: the shortening of mandurā is by ṅyāpoḥ… (SK 1001).

One of Bhaṭṭoji's commentators gives the ṭi-loss for sarasijam.

Applies when / blocked by. In sarasijam the seventh-case ending -i stays. One commentator says it stays by SK 966 or SK 972. Both rules let the ending stand inside the compound.

SK 3008 · पञ्चम्यामजातौ · pañcamyām ajātau (A 3.2.98)

पञ्चम्यामजातौ

pañcamyāmajātau · A 3.2.98

जातिशब्दवर्जिते पञ्चम्यन्ते उपपदे जनेर्डः स्यात् । संस्कारजः । अदृष्टजः ॥

Rule. With a fifth-case word that is not a genus-word as upapada, jan takes ḍa. One of Bhaṭṭoji's commentators adds "denoting the past".

Example.

saṃskārāj jātaḥ "born of a saṃskāra" → saṃskāra + jan + ḍa → saṃskārajaḥ.

Likewise adṛṣṭajaḥ "born of the unseen".

Watch out. The fifth-case ending does not stay here: the form is saṃskārajaḥ.

SK 3009 · उपसर्गे च संज्ञायाम् · upasarge ca saṃjñāyām (A 3.2.99)

उपसर्गे च संज्ञायाम्

upasarge ca saṃjñāyām · A 3.2.99

प्रजा स्यात्सन्ततौ जने ॥

Rule. With an upasarga as upapada, jan takes ḍa when the whole is a name. One of Bhaṭṭoji's commentators says "name" is a condition on the whole derived word.

Example.

pra + jan + ḍa → ṭi lost → pra + j + a → feminine → prajā.

Bhaṭṭoji quotes a line of the Amarakośa: prajā syāt santatau jane. prajā means offspring or people.

SK 3010 · अनौ कर्मणि · anau karmaṇi (A 3.2.100)

अनौ कर्मणि

anau karmaṇi · A 3.2.100

अनुपूर्वाज्जनेः कर्मण्युपपदे डः स्यात् । पुमांसमनुरुध्य जाता पुमनुजा ॥

Rule. jan with anu takes ḍa when an object is upapada.

Example.

pumāṃsam anurudhya jātā "she who was born following upon a male".

pums + anu + jan + ḍa → ṭi lost → pum-anu-j-a → pumanujā.

One of Bhaṭṭoji's commentators says the final of the cluster in pums is lost.

Applies when / blocked by. jan takes no object. How can an object be its upapada? Both of Bhaṭṭoji's commentators answer. With anu, jan means "to be born following upon", and that takes an object. The male is the object.

SK 3011 · अन्येष्वपि दृश्यते · anyeṣv api dṛśyate (A 3.2.101)

अन्येष्वपि दृश्यते

anyeṣvapi dṛśyate · A 3.2.101

अन्येष्वप्युपपदेषु जनेर्डः स्यात् । अजः । द्विजः । ब्राह्मणजः । अपिशब्दः सर्वोपाधिव्यभिचारार्थः । तेन धात्वन्तरादपि कारकान्तेरेष्वपि क्वचित् । परितः खाता परिखा ॥

Rule. ḍa after jan is seen with other upapadas too.

Example.

a + jan + ḍa → ajaḥ "unborn".

dvi + jan + ḍa → dvijaḥ "twice-born".

brāhmaṇa + jan + ḍa → brāhmaṇajaḥ "born of a brāhmaṇa".

One commentator says aja and dvija have no seventh-case word, and brāhmaṇaja has a genus word. So neither SK 3007 nor SK 3008 could give them.

Applies when / blocked by. Bhaṭṭoji: the word api "also" lets every condition of the ḍa rules lapse (sarvopādhivyabhicāra). So ḍa sometimes comes after another root, and in the sense of another kāraka. paritaḥ khātā "dug all around" → parikhā "a moat". Here the root is khan "to dig", not jan. And the affix means the object, not the agent.

Watch out. One of Bhaṭṭoji's commentators says this api also covers the vārttika "seen elsewhere too" at SK 2965. He finds the same point in the Bhāṣya.

II.N. The niṣṭhā: name and affixes — running example: `snā + kta → snātam (bhāva)`; `stu + kta → stutaḥ (karman)`; `kṛ + ktavatu → kṛtavān (agent)`

The SK now turns from upapada affixes to the past participle. Two rules set it up. SK 3012 gives the name niṣṭhā to the two affixes kta and ktavatu. SK 3013 adds a niṣṭhā after a root denoting the past. kta leaves ta; ktavatu leaves tavat. Their k keeps the root from guṇa. The running examples show the three senses. snātaṃ mayā "bathing has been done by me" is the bare action. stutas tvayā viṣṇuḥ "Viṣṇu has been praised by you" is the object. viṣṇur viśvaṃ kṛtavān "Viṣṇu made the universe" is the agent.

SK 3012 · क्तक्तवतू निष्ठा · ktaktavatū niṣṭhā (A 1.1.26)

क्तक्तवतू निष्ठा

ktaktavatū niṣṭhā · A 1.1.26

एतौ निष्ठासंज्ञौ स्तः ॥

Rule. The affixes kta and ktavatu are each called niṣṭhā. This is a name (saṃjñā). One of Bhaṭṭoji's commentators says the singular niṣṭhā applies to each of the two separately.

Example.

snā + kta → kta bears the name niṣṭhā → snā + ta → snātam.

kṛ + ktavatu → ktavatu bears the name niṣṭhā → kṛ + tavat → kṛtavān.

Applies when / blocked by. The other commentator answers an objection of mutual dependence (anyonyāśraya). The name needs the affixes to exist, and SK 3013 adds the affixes by the name. His answer: the name is "one that will be" (bhāvinī saṃjñā). SK 3013 means "that affix comes which, once it has come, will be called niṣṭhā". So this sūtra in effect prescribes kta and ktavatu.

Watch out. From here to the end of this part, "the niṣṭhā-t" means the t of ta or tavat. Most rules of II.O–II.Q replace it with n.

SK 3013 · निष्ठा · niṣṭhā (A 3.2.102)

निष्ठा

niṣṭhā · A 3.2.102

भूतार्थवृत्तेर्धातोर्निष्ठा स्यात् । तत्र तयोरेव - SK 2833 इति भावकर्मणोः क्तः कर्तरि कृत् SK 2832 इति कर्तरि क्तवतुः । उकावितौ । स्नातं मया । स्तुतस्त्वया विष्णुः । विष्णुर्विश्वं कृतवान् ॥

Rule. A niṣṭhā comes after a root whose meaning lies in the past. Bhaṭṭoji divides the senses. kta comes in bhāva and karman, by tayor eva… (SK 2833). ktavatu comes in the agent sense, by kartari kṛt (SK 2832). The u and the k are markers: the u of ktavatu, and the k of both.

Example.

snā + kta → snā + ta → snātaṃ mayā "bathing has been done by me" (bhāva).

stu + kta → stu + ta → no guṇa before the k-marked affix (SK 2217) → stutas tvayā viṣṇuḥ (karman).

kṛ + ktavatu → kṛ + tavat → no guṇa → kṛtavat → viṣṇur viśvaṃ kṛtavān (agent).

Applies when / blocked by. Why bhāva for snā but karman for stu? One of Bhaṭṭoji's commentators explains. SK 2152 runs on into SK 2833. After a root with no object, kta is in bhāva. After a root with an object, it is in karman. snā "to bathe" takes no object; stu "to praise" takes one.

Watch out. kta does not mean the agent here. The agent sense belongs to ktavatu. Exceptions where kta means the agent come in a later part of the chapter.

II.O. n for the t of the niṣṭhā — running example: `bhid + kta → bhid-ta → bhin-na → bhinnaḥ`; `kṣi + ktavatu → kṣī-tavat → kṣīṇavān`

Most rules in II.O–II.Q replace the niṣṭhā-t with n. They come from the Tripādī, the last three quarter-chapters (*pāda*s) of the eighth book, A 8.2–8.4 (here A 8.2.42 ff.). So the n is asiddha to every earlier rule. Earlier rules apply as if the t were still there: the change to a velar (kutva, bhugnaḥ) and saṃprasāraṇa (jīnaḥ) come first. SK 3014, a rule of the sixth book, is placed first to prepare kṣīṇavān. SK 3015–3019 give n after long kṣī, after r or d, after certain ā-final roots, after the lvādi, and after o-marked roots. The running example: bhid + kta gives bhinnaḥ "split". Both t and d become n.

SK 3014 · निष्ठायामण्यदर्थे · niṣṭhāyām aṇyadarthe (A 6.4.60)

निष्ठायामण्यदर्थे

niṣṭhāyāmaṇyadarthe · A 6.4.60

ण्यदर्थो भावकर्मणी ततोऽन्यत्र निष्ठायां क्षियो दीर्घः स्यात् ॥

Rule. Before a niṣṭhā outside the sense of ṇyat, kṣi takes a long vowel. Bhaṭṭoji: the sense of ṇyat is bhāva and karman. "Elsewhere" is therefore the agent. One of his commentators gives the sources. kṣiyaḥ runs on from SK 3338, dīrghaḥ from SK 3544. In effect, he says, the rule works before ktavatu.

Example.

kṣi + ktavatu → kṣi + tavat → agent sense, so the vowel is long → kṣī + tavat.

SK 3015 then gives kṣīṇavān.

Applies when / blocked by. One of Bhaṭṭoji's commentators raises a doubt. By bahula, ṇyat also comes in the instrument and other senses. Still, bhāva and karman are the senses SK 2833 names for it. Only those are meant here.

SK 3015 · क्षियो दीर्घात् · kṣiyo dīrghāt (A 8.2.46)

क्षियो दीर्घात्

kṣiyo dīrghāt · A 8.2.46

दीर्घात् क्षियो निष्ठातस्य नः स्यात् । क्षीणवान् । भावकर्मणोस्तु क्षितः कामो मया । श्र्युकः किति SK 2381 । श्रितः । श्रितवान् । भूतः । भूतवान् । क्षुतः ।

ऊर्णोतेर्णुवद्भावो वाच्यः

(वार्तिकम्) ॥ तेन एकाच्त्वान्नेट् । ऊर्णुतः । नुतः । वृतः ॥

Rule. After kṣī with a long vowel, n replaces the niṣṭhā-t. One of Bhaṭṭoji's commentators says "niṣṭhā-t" and "n" run on from SK 3016.

Example.

kṣī + tavat (from SK 3014) → n for t → kṣī + navat

→ ṇ for n after ṣ → kṣīṇavān "he has wasted away".

In bhāva or karman there is no lengthening, so no n: kṣitaḥ kāmo mayā "desire has been made to waste away by me".

Both commentators say kṣi here has the causative sense taken in.

Applies when / blocked by. Bhaṭṭoji then recalls śryukaḥ kiti (SK 2381). It bars iṭ before a k-marked affix after śri and the uk-final roots. śritaḥ, śritavān; bhūtaḥ, bhūtavān; kṣutaḥ. Both commentators derive kṣutaḥ from ṭukṣu "to sound".

A vārttika: ūrṇoter ṇuvadbhāvo vācyaḥ. ūrṇu "to cover" is treated like ṇu. So it counts as single-vowelled and takes no iṭ: ūrṇutaḥ. Bhaṭṭoji adds nutaḥ and vṛtaḥ. One commentator states the problem the vārttika solves. ūrṇu has two vowels, so śryukaḥ kiti would not reach it.

Watch out. The long vowel is the trigger. kṣitaḥ keeps its t because SK 3014 gave no long vowel in the object sense.

SK 3016 · रदाभ्यां निष्ठातो नः पूर्वस्य च दः · radābhyāṃ niṣṭhāto naḥ pūrvasya ca daḥ (A 8.2.42)

रदाभ्यां निष्ठातो नः पूर्वस्य च दः

radābhyāṃ niṣṭhāto naḥ pūrvasya ca daḥ · A 8.2.42

रेफदकाराभ्यां परस्य निष्ठातस्य नः स्यात् । निष्ठापेक्षया पूर्वस्य धातोर्दकारस्य च । शॄ । ऋत इत् SK 2390 । रपरः । णत्वम् । शीर्णः । बहिरङ्गत्वेन वृद्धेरसिद्धत्वान्नेह । कृतस्यापत्यं कार्तिः । भिन्नः । छिन्नः ॥

Rule. After r or d, n replaces the niṣṭhā-t. And n also replaces the root's d that stands before the niṣṭhā.

Example.

bhid + kta → bhid + ta

→ t after d becomes n; the d becomes n too → bhin + na → bhinnaḥ "split".

Likewise chinnaḥ "cut".

śṝ + kta → i for ṝ by ṛta id dhātoḥ (SK 2390), followed by r → śir + ta

→ n for t → śir + na

→ the i is lengthened before the consonant by hali ca (SK 354), a step Bhaṭṭoji leaves unstated → śīrna

→ ṇ for n → śīrṇaḥ "crushed".

Applies when / blocked by. Bhaṭṭoji gives a case where the rule fails: kārtiḥ "the offspring of Kṛta". kṛta + the taddhita iñ gives initial vṛddhi, kār-. Now the t follows r. But the vṛddhi is bahiraṅga, conditioned by something outside the niṣṭhā. So it is asiddha to this rule, and the t stays. One commentator says the vṛddhi depends on the taddhita affix, which lies outside the niṣṭhā-t.

Watch out. The r or d must stand right before the t. Both of Bhaṭṭoji's commentators give caritam and uditam. There iṭ stands between, so the t stays.

SK 3017 · संयोगादेरातो धातोर्यण्वतः · saṃyogāder āto dhātor yaṇvataḥ (A 8.2.43)

संयोगादेरातो धातोर्यण्वतः

saṃyogāderāto dhātoryaṇvataḥ · A 8.2.43

निष्ठातस्य नः स्यात् । द्राणः । स्त्यानः । ग्लानः । म्लानः ॥

Rule. The root must begin with a consonant cluster (saṃyoga) and contain a yaṇ, a semivowel y, v, r or l. After its ā, n replaces the niṣṭhā-t. Bhaṭṭoji supplies "n for the niṣṭhā-t"; one of his commentators calls this a completion of the sūtra.

Example.

drā + kta → drā + ta → n for t → drā + na → ṇ after r → drāṇaḥ.

glai + kta → ā for the final diphthong (SK 2370) → glā + ta → glānaḥ "weary".

Likewise styānaḥ, mlānaḥ.

The commentators give the roots: drā "to go basely" and glai "to lose joy" (both), and mlai, also "to lose joy" (one).

Watch out. All three conditions must hold. The root begins with a cluster (dr, st, gl, ml), contains a semivowel, and has ā just before the t.

SK 3018 · ल्वादिभ्यः · lvādibhyaḥ (A 8.2.44)

ल्वादिभ्यः

lvādibhyaḥ · A 8.2.44

एकविंशतेर्लूञादिभ्यः प्राग्वत् । लूनः । ज्या । ग्रहिज्या -SK 2412 । जीनः ॥

दुग्वोर्दीर्घश्च

(वार्तिकम्) ॥ दु गतौ दूनः । टुदु उपताप इत्ययं तु न गृह्यते सानुबन्धकत्वात् । मृदुतया दुतयेति माघः । गूनः ।

पूञो विनाशे

(वार्तिकम्) ॥ पूना यवाः । विनष्टा इत्यर्थः । पूतमन्यत् ।

सिनोतेर्ग्रासकर्मकर्तृकस्य

(वार्तिकम्) ॥ सिनो ग्रासः । ग्रासेति किम् । सिता पाशेन सूकरी । कर्मकर्तृकेति किम् । सितो ग्रासो देवदत्तेन ॥

Rule. After the twenty-one roots beginning with lūñ, n replaces the niṣṭhā-t, as before. Both of Bhaṭṭoji's commentators identify the list. The kryādi class has twenty-two roots beginning with pūñ, the pvādi. Leave out pūñ, and the rest are the lvādi.

Example.

lū + kta → lū + ta → n for t → lūnaḥ "cut".

jyā + kta → saṃprasāraṇa by grahijyā… (SK 2412) → ji + ta

→ lengthening by halaḥ (SK 2559) → jī + ta → n for t → jīnaḥ "aged".

Applies when / blocked by. Why not n by SK 3017 for jyā? Both commentators answer. One says the n of SK 3017 is asiddha, so saṃprasāraṇa comes first. The other says saṃprasāraṇa is siddha and obligatory (nitya). Either way, ā is gone before the t. So n comes by this rule.

Three vārttikas follow.

dugvor dīrghaś ca: after du and gu, n, and the vowel is lengthened. du "to go" → dūnaḥ. ṭudu "to afflict" is not included, since it carries a marker. Hence Māgha's mṛdutayā dutayā. gu → gūnaḥ.
pūño vināśe: after pūñ in the sense of perishing. pūnā yavāḥ "the barley has spoiled". In another sense: pūtam "purified".
sinoter grāsakarmakartṛkasya: after si "to bind", when the mouthful is an object acting as agent. sino grāsaḥ "the mouthful has bound itself" (into a ball). Not the mouthful: sitā pāśena sūkarī "the sow bound with a noose". Not an object-agent: sito grāso devadattena "the mouthful bound by Devadatta".

Watch out. The lvādi start from lūñ, not pūñ. pūñ gets n only by the vārttika, and only when it means perishing.

SK 3019 · ओदितश्च · oditaś ca (A 8.2.45)

ओदितश्च

oditaśca · A 8.2.45

भुजो । भुग्नः । टुओश्वि । उच्छूनः । ओहाक् । प्रहीणः । (गणसूत्रम् -) स्वादय ओदितः इत्युक्तम् । सूनः । सूनवान् । दूनः । दूनवान् । ओदिन्मध्ये डीङ पाठमर्थ्यान्नेट् । उड्डीनः ॥

Rule. After a root marked with o (odit), n replaces the niṣṭhā-t. Bhaṭṭoji cites the roots with their markers, since the o is what brings the rule in.

Example.

bhujo "to bend": bhuj + ta → the n is asiddha, so g for j first → bhug + ta → bhugnaḥ "bent".

ṭuośvi "to swell": ut + śvi + ta → saṃprasāraṇa, u for v → ut + śu + ta → lengthening → ucchūnaḥ "swollen".

ohāk "to leave": pra + hā + ta → ī for ā (SK 2462) → pra + hī + na → ṇ (SK 2835) → prahīṇaḥ.

Both of Bhaṭṭoji's commentators give these steps; one cites SK 3039 for the absence of iṭ in ucchūnaḥ.

Applies when / blocked by. Bhaṭṭoji recalls a gaṇasūtra, a rule in the root list: svādaya oditaḥ "sū and the rest are o-marked". So sūnaḥ, sūnavān; dūnaḥ, dūnavān. One commentator: these are nine roots of the divādi class, beginning with ṣūṅ.

ḍīṅ "to fly" is read among the o-marked roots. Bhaṭṭoji: by the force of that reading (pāṭhasāmarthya), it takes no iṭ → uḍḍīnaḥ. Both commentators explain. With iṭ, the t would never follow the root, and the reading would be pointless. One reports another view, by which iṭ does come and merges; scholars discuss this.

Watch out. In bhugnaḥ the velar comes before the n. That order follows from the n being asiddha.

II.P. śyaiṅ before the niṣṭhā: saṃprasāraṇa and n by sense — running example: `śyai + kta → śyā-ta → śi-ta → śī-ta → śīnam / śītam`

Four rules handle śyaiṅ "to go", which appears as śyā (SK 2370). Two things depend on the sense. Saṃprasāraṇa turns śyā into śi (SK 3020, 3022, 3023). And n replaces the niṣṭhā-t except in the sense of touch (SK 3021). The running example: congealed ghee is śīnaṃ ghṛtam, but cold water is śītaṃ jalam.

SK 3020 · द्रवमूर्तिस्पर्शयोः श्यः · dravamūrtisparśayoḥ śyaḥ (A 6.1.24)

द्रवमूर्तिस्पर्शयोः श्यः

dravamūrtisparśayoḥ śyaḥ · A 6.1.24

द्रवस्य मूर्तौ काठिन्ये स्पर्शे चार्थे श्यैङः संप्रसारणं स्यान्निष्ठायाम् ॥

Rule. śyaiṅ takes saṃprasāraṇa before a niṣṭhā in two senses. One is the solidifying (mūrti), or hardening, of a liquid. The other is touch (sparśa). One of Bhaṭṭoji's commentators says saṃprasāraṇa runs on from SK 1003. Bhaṭṭoji himself supplies "before a niṣṭhā".

Example.

śyai + kta → ā for the final diphthong (SK 2370) → śyā + ta

→ saṃprasāraṇa: i for y → śi-ā + ta → the ā merges into the i → śi + ta.

One commentator gives these steps. SK 3021 continues the derivation.

SK 3021 · श्योऽस्पर्शे · śyo 'sparśe (A 8.2.47)

श्योऽस्पर्शे

śyo'sparśe · A 8.2.47

श्यैङो निष्ठातस्य नः स्यादस्पर्शेऽर्थे । हलः - SK 2559 इति दीर्घः । शीनं घृतम् । अस्पर्शे किम् । शीतं जलम् । द्रवमूर्तिस्पर्शयोः किम् । संश्यानो वृश्चिकः । शीतात्संकुचित इत्यर्थः ॥

Rule. After śyaiṅ, n replaces the niṣṭhā-t, except when the sense is touch. One of Bhaṭṭoji's commentators says "touch" here means the particular touch called cold.

Example.

śi + ta (from SK 3020) → lengthening by halaḥ (SK 2559) → śī + ta

→ not touch, so n for t → śīnaṃ ghṛtam "congealed ghee".

Touch: śītaṃ jalam "cold water". Saṃprasāraṇa comes by SK 3020, but no n.

Applies when / blocked by. Bhaṭṭoji asks why SK 3020 names its senses. saṃśyāno vṛścikaḥ "the scorpion has drawn itself together from cold". Neither sense is meant, so there is no saṃprasāraṇa. One commentator says the n here comes by SK 3017.

Watch out. One of Bhaṭṭoji's commentators notes the strict order. The lengthening by halaḥ comes before the n, because the n is a Tripādī rule.

SK 3022 · प्रतेश्च · prateś ca (A 6.1.25)

प्रतेश्च

prateśca · A 6.1.25

प्रतिपूर्वस्य श्यः संप्रसारणं स्यान्निष्ठायाम् । प्रतिशीनः ॥

Rule. After prati, śyaiṅ takes saṃprasāraṇa before a niṣṭhā. One of Bhaṭṭoji's commentators says this holds even outside the two senses of SK 3020.

Example.

prati + śyā + kta → saṃprasāraṇa → prati + śi + ta → lengthening → n (SK 3021) → pratiśīnaḥ.

The same commentator glosses it as pratigata and gives SK 3021 for the n.

SK 3023 · विभाषाऽभ्यवपूर्वस्य · vibhāṣā 'bhyavapūrvasya (A 6.1.26)

विभाषाऽभ्यवपूर्वस्य

vibhāṣā'bhyavapūrvasya · A 6.1.26

श्यः संप्रसारणं वा स्यात् । अभिश्यानं घृतम् । अभिशीनम् । अवश्यानोऽवशीनो वृश्चिकः । व्यवस्थितविभाषेयम् । तेनेह न । समवश्यानः ॥

Rule. After abhi or ava, śyaiṅ optionally takes saṃprasāraṇa before a niṣṭhā.

Example.

abhiśyānaṃ ghṛtam or abhiśīnam "congealed ghee".

avaśyānaḥ or avaśīnaḥ vṛścikaḥ "the scorpion drawn together".

Applies when / blocked by. Bhaṭṭoji: this is a fixed option (vyavasthitavibhāṣā). Such an option holds only in fixed cases. So it does not reach samavaśyānaḥ, which has no saṃprasāraṇa. One of his commentators says the option covers both kinds of case. With ghee it is in the sense of hardening. With the scorpion neither sense is present.

II.Q. añc and the iṭ-ban for optionally-seṭ roots; the niṣṭhā substitute siddha for ṣatva — running example: `sam + añc + kta → sam-ac-ta → sam-ak-ta → samak-na → samaknaḥ`

Two rules and a vārttika finish the n rules of this part. SK 3024 gives n after añc "to bend, go". SK 3025 bars iṭ before the niṣṭhā after a root that takes iṭ only optionally elsewhere. Bhaṭṭoji gives the examples of both under SK 3025. A vārttika then makes the niṣṭhā substitute count as having taken effect for four operations, as in vṛkṇaḥ. The running example samaknaḥ "joined" uses both rules; vṛkṇaḥ shows the vārttika.

SK 3024 · अञ्चोऽनपादाने · añco 'napādāne (A 8.2.48)

अञ्चोऽनपादाने

añco'napādāne · A 8.2.48

अञ्चो निष्ठातस्य नः स्यान्न त्वपादाने ॥

Rule. After añc, n replaces the niṣṭhā-t. But not in connection with an apādāna, the point of departure. One of Bhaṭṭoji's commentators glosses: when no apādāna stands in the same sentence.

Example. Bhaṭṭoji gives the examples under SK 3025.

sam + añc + kta → … → samaknaḥ "joined".

With an apādāna: udaktam udakaṃ kūpāt "water drawn up from the well". kūpāt "from the well" is the apādāna, so the t stays.

SK 3025 · यस्य विभाषा · yasya vibhāṣā (A 7.2.15)

यस्य विभाषा

yasya vibhāṣā · A 7.2.15

यस्य क्वचिद्विभाषयेड्विहितस्ततो निष्ठाया इण्न स्यात् । उदितो वा SK 3328 इति क्त्वायां वेट्त्वादिह नेट् । समक्नः । अनपादाने किम् । उदक्तमुदकं कूपात् । नत्वस्यासिद्धत्वात् व्रश्च - SK 294 इति षत्वे प्राप्ते ।

निष्ठादेशः षत्वस्वरप्रत्ययेड्विधिषु सिद्धो वाच्यः

(वार्तिकम्) ॥ वृक्णः । वृक्णवान् ॥

Rule. Take a root for which iṭ is optional in some context. Before the niṣṭhā it takes no iṭ. One of Bhaṭṭoji's commentators gives the sources. niṣṭhāyām runs on from SK 3039, and "no iṭ" from SK 2981.

Example.

añc takes iṭ optionally before ktvā by udito vā (SK 3328). So before the niṣṭhā: no iṭ.

sam + añc + kta → sam + añc + ta → no iṭ

→ the nasal is lost (SK 415) → sam + ac + ta

→ the n to come is asiddha, so k for c before the t → sam + ak + ta

→ n for t by SK 3024 → samaknaḥ "joined".

One commentator gives the nasal-loss; both give the velar.

Applies when / blocked by. Bhaṭṭoji then raises a problem. Since the n is asiddha, ṣ-substitution by vraśca… (SK 294) would apply. So a vārttika is given: niṣṭhādeśaḥ ṣatvasvarapratyayeḍvidhiṣu siddho vācyaḥ. The niṣṭhā substitute counts as siddha for four operations: ṣ-substitution (ṣatva), accent (svara), affixes (pratyaya) and iṭ. So vṛkṇaḥ, vṛkṇavān "cut".

Both commentators trace vṛkṇa from o vraścū "to cut". Saṃprasāraṇa (SK 2412), no iṭ by this rule, k for c, n by SK 3019, the s lost (SK 380), and ṇ. With the n siddha, the c is no longer before a jhal consonant. So SK 294 cannot turn it into ṣ.

Watch out. What does yasya "of which" mean? One commentator reads it as "after which". The other takes it with the ārdhadhātuka: "the affix belonging to which". Scholars discuss this.

II.R. ṣ withheld: pariskanda, pratistabdha, nistabdha — running example: `pari + skand → pariskandaḥ (eastern), pariṣkandaḥ / pariskandaḥ (elsewhere)`

Part II ends with two rules that withhold ṣ for s. Both stand in the eighth book, among the rules that turn s into ṣ. SK 3026 keeps the s of skand after pari among the eastern Bharatas. SK 3027 keeps the s of stanbh after prati and ni. The running example: the eastern form is pariskandaḥ; elsewhere both pariṣkandaḥ and pariskandaḥ stand.

SK 3026 · परिस्कन्दः प्राच्यभरतेषु · pariskandaḥ prācyabharateṣu (A 8.3.75)

परिस्कन्दः प्राच्यभरतेषु

pariskandaḥ prācyabharateṣu · A 8.3.75

पूर्वेण मूर्धन्ये प्राप्ते तदभावो निपात्यते । परिस्कन्दः । प्राच्येति किम् । परिष्कन्दः । परिस्कन्दः ।परेश्च SK 2399 इति विकल्पः । स्तम्भेः SK 2272 इति षत्वे प्राप्ते ॥

Rule. Among the eastern Bharatas, pariskanda is given by nipātana without ṣ. Bhaṭṭoji: ṣ would be due by the preceding rule, and its absence is stated by nipātana.

Example.

pari + skand → ṣ would be due by pareś ca (SK 2399) → withheld by nipātana → pariskandaḥ.

Why "eastern"? Elsewhere pariṣkandaḥ or pariskandaḥ. The option is by SK 2399.

Applies when / blocked by. One of Bhaṭṭoji's commentators places the rule under apadāntasya mūrdhanyaḥ (SK 210). He forms the word from skand with the niṣṭhā-t elided. The other forms it with ac, or with the t elided as the Kāśikā says. Scholars discuss this.

Watch out. Bhaṭṭoji closes by setting up the next rule. ṣ would be due by stambheḥ (SK 2272).

SK 3027 · प्रतिस्तब्धनिस्तब्धौ च · pratistabdhanistabdhau ca (A 8.3.114)

प्रतिस्तब्धनिस्तब्धौ च

pratistabdhanistabdhau ca · A 8.3.114

अत्र षत्वं न स्यात् ॥

Rule. In pratistabdha and nistabdha, ṣ does not replace s. Bhaṭṭoji: "here there is no ṣatva".

Example.

prati + stanbh + kta → ṣ would be due by SK 2272 → withheld → pratistabdhaḥ.

ni + stanbh + kta → nistabdhaḥ.

Watch out. Neither commentator says anything on this sūtra. The rule only stops ṣ; the rest of the form is built as usual.

III.A. More substitutes for the niṣṭhā-t: n, k, v, m — running example: `div + kta → dyū + ta → dyūnaḥ; śuṣ + ta → śuṣkaḥ; pac + ta → pakvaḥ; kṣai → kṣāmaḥ; pra-styā → prastīmaḥ`

From SK 3016 on, earlier rules turned the t of a niṣṭhā into n. A niṣṭhā is the name of the two past-participle affixes kta and ktavatu (SK 3012). This group carries the same run of the Tripādī forward. After a few more roots the niṣṭhā-t becomes n. After others it becomes k, v or m. Bhaṭṭoji often supplies the words "of the t of a niṣṭhā" himself; they continue from SK 3016. The running example is a row of single roots. div gives dyūnaḥ, śuṣ gives śuṣkaḥ "dried", and pac gives pakvaḥ "cooked". kṣai gives kṣāmaḥ, and styā after pra gives prastīmaḥ "congealed".

SK 3028 · दिवोऽविजिगीषायाम् · divo 'vijigīṣāyām (A 8.2.49)

दिवोऽविजिगीषायाम्

divo'vijigīṣāyām · A 8.2.49

दिवो निष्ठातस्य नः स्यादविजिगीषायाम् । द्यूनः । विजिगीषायां तु द्यूतम् ॥

Rule. After the root div, n replaces the t of a niṣṭhā. It does not happen when the sense is "desire to conquer" (vijigīṣā). Bhaṭṭoji: "after div, n replaces the niṣṭhā-t, when not in desire to conquer".

Example.

div + kta → the marker k goes → div + ta

→ the v becomes ū (ūṭh) by cchvoḥ śūḍ anunāsike ca (SK 2561) → di + ū + ta

→ the i becomes y before the ū → dyū + ta

→ this rule: the t becomes n → dyūnaḥ.

Both of Bhaṭṭoji's commentators name SK 2561 for the ū.

Applies when / blocked by. In the sense of desire to conquer the t stays: dyūtam "gambling". Both commentators give the reason. Dice are thrown with the wish to win, so desire to conquer is present.

Watch out. What dyūnaḥ means is left open. One commentator glosses it "praised", the other "wasted, exhausted"; scholars discuss this. The words "of the t of a niṣṭhā" are Bhaṭṭoji's gloss. They continue from SK 3016, but neither commentator names the source.

SK 3029 · निर्वाणोऽवाते · nirvāṇo 'vāte (A 8.2.50)

निर्वाणोऽवाते

nirvāṇo'vāte · A 8.2.50

अवाते इति छेदः । निर्पूर्वाद्वातेर्निष्ठातस्य नत्वं स्याद्वातस्चेत्कर्ता न । निर्वाणोऽग्निर्मुनिश्च । वाते तु निर्वातो वातः ॥

Rule. After the root vā preceded by nir, n replaces the niṣṭhā-t. The condition is that wind is not the agent. Bhaṭṭoji first says how to divide the words: avāte, "not wind".

Example.

nir + vā + kta → nir + vā + ta

→ this rule: the t becomes n → nir + vā + na

→ the n after r becomes ṇ (ṇatva) → nirvāṇaḥ.

Bhaṭṭoji: nirvāṇo 'gniḥ "the fire is extinguished"; nirvāṇo muniḥ "the sage has come to rest".

One of his commentators says the kta here denotes the agent, by SK 3086 (group III.M below).

Applies when / blocked by. When wind is the agent, the t stays: nirvāto vātaḥ "the wind has blown itself out". One of Bhaṭṭoji's commentators adds a bhāva form, also with t: nirvātaṃ vātena.

Watch out. In nirvāṇo dīpo vātena "the lamp is put out by the wind" the n stands. The wind there is not the agent. One commentator calls it the instrument (karaṇa), the other the cause (hetu); scholars discuss this.

SK 3030 · शुषः कः · śuṣaḥ kaḥ (A 8.2.51)

शुषः कः

śuṣaḥ kaḥ · A 8.2.51

निष्ठात इत्येव । शुष्कः ॥

Rule. After the root śuṣ "to dry", k replaces the niṣṭhā-t. Bhaṭṭoji says only "niṣṭhātaḥ ity eva": the words "the t of a niṣṭhā" still hold here. They first stand in SK 3016.

Example.

śuṣ + kta → śuṣ + ta

→ this rule: the t becomes k → śuṣ + ka

→ śuṣkaḥ "dried".

One of Bhaṭṭoji's commentators names the root as śuṣa śoṣaṇe.

SK 3031 · पचो वः · paco vaḥ (A 8.2.52)

पचो वः

paco vaḥ · A 8.2.52

पक्वः ॥

Rule. After the root pac "to cook", v replaces the niṣṭhā-t. Bhaṭṭoji gives only the form.

Example.

pac + kta → pac + ta

→ this rule: the t becomes v → pac + va

→ the c becomes k → pakvaḥ "cooked".

Applies when / blocked by. Why does the c still become k? One of Bhaṭṭoji's commentators answers: the v is asiddha, treated as not yet there. This rule stands late in the Tripādī. The rule that makes c into k (kutva) is coḥ kuḥ (SK 378). It stands earlier in the Tripādī, so it does not see the v. It still finds c before the consonant t.

SK 3032 · क्षायो मः · kṣāyo maḥ (A 8.2.53)

क्षायो मः

kṣāyo maḥ · A 8.2.53

क्षामः ॥

Rule. After the root kṣai "to waste away", m replaces the niṣṭhā-t. Bhaṭṭoji gives only the form.

Example.

kṣai + kta → kṣai + ta

→ the final ai becomes ā by ādeca upadeśe 'śiti (SK 2370) → kṣā + ta

→ this rule: the t becomes m → kṣā + ma → kṣāmaḥ.

Both of Bhaṭṭoji's commentators name SK 2370 for the ā.

Watch out. One commentator gives kṣāmaḥ two senses. With kta in the agent (SK 3086) it means "wasted away". If a causative sense is read into the root, it means "made to waste away". So the form is best learned without a fixed gloss.

SK 3033 · स्त्यः प्रपूर्वस्य · styaḥ prapūrvasya (A 6.1.23)

स्त्यः प्रपूर्वस्य

styaḥ prapūrvasya · A 6.1.23

प्रात् स्त्यः संप्रसारणं स्यान्निष्ठायाम् ॥

Rule. The root styā takes saṃprasāraṇa when it follows pra and a niṣṭhā follows. styā is styai after its ai has become ā. Saṃprasāraṇa here turns the y into i. Bhaṭṭoji: "saṃprasāraṇa for styā after pra, before a niṣṭhā".

Example.

pra + styā + kta → pra + styā + ta

→ this rule: y becomes i → pra + sti-ā + ta

→ the ā merges into the i (pūrvarūpa) → pra + sti + ta

→ the i is lengthened by halaḥ (SK 2559) → prastī + ta.

Both of Bhaṭṭoji's commentators give these steps. Bhaṭṭoji gives no form here; the forms come under the next rule.

Applies when / blocked by. One commentator draws "saṃprasāraṇa" from SK 1003. He draws "before a niṣṭhā" from SK 3044, which the SK teaches later. In the Aṣṭādhyāyī that rule stands just before this one. The other commentator says why saṃprasāraṇa comes first. The n of SK 3017 is asiddha at this point, so it cannot get in ahead.

SK 3034 · प्रस्त्योऽन्यतरस्याम् · prastyo 'nyatarasyām (A 8.2.54)

प्रस्त्योऽन्यतरस्याम्

prastyo'nyatarasyām · A 8.2.54

निष्ठातस्य मो वा स्यात् । प्रस्तीमः । प्रस्तीतः । प्रात्किम् । स्त्यानः ॥

Rule. After styā preceded by pra, m optionally replaces the niṣṭhā-t. Bhaṭṭoji supplies the words "m for the t of a niṣṭhā". One of Bhaṭṭoji's commentators calls them a śeṣa, words to be supplied.

Example.

prastī + ta (from SK 3033)

→ this rule: the t becomes m → prastīmaḥ "congealed".

In the other option the t stays → prastītaḥ.

One commentator glosses prastīmaḥ "become compact, congealed".

Applies when / blocked by. Why no n in prastītaḥ? One commentator explains. After saṃprasāraṇa, prastī- has no y and no final ā. So SK 3017 finds nothing to work on. Bhaṭṭoji asks "why after pra?" and gives only the form: styānaḥ. The reason is this. Without pra there is no saṃprasāraṇa. styā keeps its y and its ā, and SK 3017 gives n. One commentator notes the same n in saṃstyānaḥ.

III.B. phulla by nipātana, and praphulta with an upasarga — running example: `ñiphalā + kta → phal + la → phullaḥ; pra + phal + ta → praphultaḥ`

Some niṣṭhā forms are not built step by step. A rule states them ready-made. This is nipātana: the irregular parts of the stated form are simply accepted. SK 3035 states four such forms, but only when no upasarga comes first. SK 3036 and SK 3037 then explain the forms with an upasarga. The running example is the root ñiphalā viśaraṇe. Alone it gives phullaḥ, with l for the niṣṭhā-t. After pra it gives praphultaḥ, with the t kept.

SK 3035 · अनुपसर्गात् फुल्लक्षीबकृशोल्लाघाः · anupasargāt phullakṣībakṛśollāghāḥ (A 8.2.55)

अनुपसर्गात् फुल्लक्षीबकृशोल्लाघाः

anupasargāt phullakṣībakṛśollāghāḥ · A 8.2.55

ञिफाला । फुल्लः । निष्ठातस्य लत्वं निपात्यते । क्तवत्वेकदेशस्यापीदं निपातनमिष्यते । फुल्लवान् । क्षीबादिषु तु क्तप्रत्ययस्यैव तलोपः । तस्यासिद्धत्वात्प्राप्तस्येटोऽभावश्च निपात्यते । क्षीबो मत्तः । कृशस्तनुः । उल्लाघो नीरोगः । अनुपसर्गात्किम् ॥

Rule. When no upasarga precedes, phulla, kṣīva, kṛśa and ullāgha are given by nipātana. In each, something irregular happens to the niṣṭhā. Bhaṭṭoji says exactly what. Our text spells the second word kṣība, as in the heading. The print reads kṣīva, and this chapter follows the print.

Example.

For phulla only the l is nipātana:

ñiphalā + kta → the markers ñi and ā go → phal + ta

→ by nipātana, l for the niṣṭhā-t → phal + la

→ the a becomes u (SK 3037) → phullaḥ.

One of Bhaṭṭoji's commentators: there is no iṭ by SK 3036, and the u comes by SK 3037. So only the l needs the nipātana.

Bhaṭṭoji: the nipātana also reaches the kta inside ktavatu → phullavān.

For the other three, the nipātana is the loss of the t of kta. It is not the t of ktavatu. That loss is asiddha, so the iṭ that would follow is also withheld by nipātana.

kṣīv + kta → t lost, no iṭ → kṣīvaḥ "intoxicated".

Likewise kṛśaḥ "thin" and ullāghaḥ "free of disease".

One commentator names the roots: kṣīvṛ "to be intoxicated", kṛśa "to be thin", and lāghṛ "to be able" after ut.

Applies when / blocked by. Bhaṭṭoji asks "why when not after an upasarga?" One commentator says he means praphultaḥ. Bhaṭṭoji first teaches SK 3036 and SK 3037, which that form needs. The answer then comes under SK 3037.

Watch out. One commentator says the root is ñiphalā viśaraṇe, not phala niṣpattau. With the second, the lack of iṭ too would need a nipātana. The senses are fixed by the nipātana as well, he adds: kṣīva means only "intoxicated", and so on.

SK 3036 · आदितश्च · āditaś ca (A 7.2.16)

आदितश्च

āditaśca · A 7.2.16

आकारेतो निष्ठाया इण्न स्यात् ॥

Rule. After a root marked with ā (an ādit root), iṭ does not come to a niṣṭhā. iṭ is the augment i that an ārdhadhātuka affix takes when it begins with a consonant other than y (SK 2184). Bhaṭṭoji: "after a root with ā as its it, iṭ does not come to a niṣṭhā".

Example.

ñiphalā has the marker ā, so it is ādit:

phal + ta → no iṭ → stays phal + ta.

This gives phullaḥ (SK 3035) and praphultaḥ (SK 3037).

Bhaṭṭoji gives no example of his own here.

Applies when / blocked by. One of Bhaṭṭoji's commentators supplies the missing words. "iṭ does not come" continues from neḍ vaśi kṛti (SK 2981). "To a niṣṭhā" continues from śvīdito niṣṭhāyām (SK 3039). The SK teaches that rule a little later, though the Aṣṭādhyāyī has it first.

SK 3037 · ति च · ti ca (A 7.4.89)

ति च

ti ca · A 7.4.89

चरफलोपत उत्स्यात्तादौ किति । प्रफुल्तः । प्रक्षीबितः । प्रकृशितः । प्रोल्लाघितः । कथं तर्हि लोध्रद्रुमं सानुमतः प्रफुल्लमिति । फुल्ल विकसने पचाद्यच् । सूत्रं तु फुल्तादिनिवृत्त्यर्थम् । उत्फुल्लसंफुल्लयोरुपसंख्यानम् ॥

Rule. u replaces the a of car and phal before a t-initial affix that is kit. A kit affix is one marked with k, like the niṣṭhā. One of Bhaṭṭoji's commentators draws car and phal from SK 2636 and "u for a" from SK 2637.

Example.

With an upasarga there is no nipātana, so the niṣṭhā-t stays:

pra + phal + kta → pra + phal + ta (no iṭ, SK 3036)

→ this rule: a becomes u → pra + phul + ta → praphultaḥ.

The other three roots of SK 3035 take iṭ after pra: prakṣīvitaḥ, prakṛśitaḥ, prollāghitaḥ.

Applies when / blocked by. Bhaṭṭoji raises a verse: lodhradrumaṃ sānumataḥ praphullam "the lodhra tree of the mountain, in full bloom". How is praphullam right, with pra? His answer: it comes from the root phulla vikasane "to blossom" with the affix ac of the pacādi list (SK 2896). Then why state phulla in SK 3035 at all? Bhaṭṭoji: "the sūtra is for excluding phulta and the like". One commentator calls this wording illustrative. The rule also secures phullavān and the accent of phulla (SK 3693).

A vārttika adds two words: utphulla and saṃphulla. One commentator says it serves to exclude the forms utphulta and saṃphulta.

Watch out. Bhaṭṭoji states the condition "kit". One commentator carries it from SK 2456 by "frog's leap" (maṇḍūkapluti), skipping the rules between. He then calls that carrying idle, since both roots are seṭ; scholars discuss this.

III.C. Optional n, no iṭ after śvi and the īdit, and roots that refuse n — running example: `nud + kta → nunnaḥ / nuttaḥ; undī → unnaḥ / uttaḥ; but dhyai → dhyātaḥ, mad → mattaḥ`

Three rules adjust the n of SK 3016 and SK 3017. SK 3038 makes it optional after six roots. SK 3039 explains why one of them, undī, takes no iṭ. SK 3040 refuses the n altogether after five roots. The running example shows both sides. nud gives nunnaḥ or nuttaḥ. undī "to wet" gives unnaḥ or uttaḥ. But dhyai "to think" gives only dhyātaḥ, and mad "to rejoice" only mattaḥ.

SK 3038 · नुदविदोन्दत्राघ्राह्रीभ्योऽन्यतरस्याम् · nudavidondatrāghrāhrībhyo 'nyatarasyām (A 8.2.56)

नुदविदोन्दत्राघ्राह्रीभ्योऽन्यतरस्याम्

nudavidondatrāghrāhrībhyo'nyatarasyām · A 8.2.56

एभ्यो निष्ठातस्य नो वा । नुन्नः । नुत्तः । विदविचारणे रौधादिक एव गृह्यते उन्दिना परेण साहचर्यात् । विन्नः । वित्तः । वेत्तेस्तु विदितः । विद्यतेर्विन्नः । उन्दी ॥

Rule. After nud, vid, und, trā, ghrā and hrī, n optionally replaces the niṣṭhā-t. Bhaṭṭoji: "after these, n optionally replaces the t of a niṣṭhā".

Example.

nud + kta → nud + ta

→ this rule: the t becomes n, and the d before it becomes n too (SK 3016) → nunnaḥ.

In the other option → nuttaḥ.

Applies when / blocked by. Both of Bhaṭṭoji's commentators say what the option does. For hrī no n was due at all. For the other five it was due without option, by SK 3016 or SK 3017. This rule makes it optional in both cases.

Which vid? Bhaṭṭoji: only vid vicāraṇe "to discern", of the rudh class. The reason is association (sāhacarya): a word is taken as the one of the same kind as the word it is paired with. Here the neighbour is und, which follows. und is a root of the rudh class, so vid too is taken as the rudh-class root. So vinnaḥ, vittaḥ. The other vid roots differ. From vetti (vid jñāne "to know") comes viditaḥ. One commentator: that root takes iṭ, and iṭ stands between the d and the t, so no n. From vidyate (vid sattāyām "to be") comes only vinnaḥ. That root takes no iṭ, so SK 3016 gives n without option.

Watch out. Bhaṭṭoji's list breaks off at the word undī. One commentator calls it a pointer to the examples. They follow under the next rule.

SK 3039 · श्वीदितो निष्ठायाम् · śvīdito niṣṭhāyām (A 7.2.14)

श्वीदितो निष्ठायाम्

śvīdito niṣṭhāyām · A 7.2.14

श्वयतेरीदितश्च निष्ठाया इण्न । उन्नः । उत्तः । त्राणः । त्रातः । घ्राणः । घ्रातः । ह्रीणः । ह्रीतः ॥

Rule. After śvi and after a root marked with ī (an īdit root), iṭ does not come to a niṣṭhā. One of Bhaṭṭoji's commentators draws "iṭ does not come" from SK 2981.

Example.

undī "to wet" is marked with ī:

und + kta → this rule: no iṭ → und + ta

→ the root's nasal is lost (SK 415) → ud + ta

→ by SK 3038, optionally n for the t and the d → unnaḥ; otherwise uttaḥ.

Then Bhaṭṭoji resumes the examples of SK 3038: trāṇaḥ / trātaḥ, ghrāṇaḥ / ghrātaḥ, hrīṇaḥ / hrītaḥ.

Watch out. Only unnaḥ and uttaḥ show this rule. The other six forms belong to SK 3038. No form from śvi is given here; śūnaḥ comes at SK 3057.

SK 3040 · न ध्याख्यापॄमूर्छिमदाम् · na dhyākhyāpṝmūrchimadām (A 8.2.57)

न ध्याख्यापॄमूर्छिमदाम्

na dhyākhyāpṝmūrchimadām · A 8.2.57

एभ्यो निष्ठातस्य नत्वं न । ध्यातः । ख्यातः । पूर्तः । राल्लोपः SK 2655 । मूर्तः । मत्तः ॥

Rule. After dhyā, khyā, pṝ, mūrch and mad, the niṣṭhā-t does not become n. Bhaṭṭoji: "after these there is no n for the t of a niṣṭhā".

Example.

dhyai + kta → ai becomes ā (SK 2370) → dhyā + ta

→ SK 3017 would give n; this rule blocks it → dhyātaḥ.

Likewise khyātaḥ.

pṝ + kta → no iṭ by śryukaḥ kiti (SK 2381) → ṝ becomes ur (SK 2494) → pur + ta → the u is lengthened → pūrtaḥ.

mūrch + kta → rāl lopaḥ (SK 2655) drops the ch → mūr + ta → mūrtaḥ.

mad + kta → no iṭ, since madī is īdit (SK 3039) → mattaḥ.

In pūrtaḥ, mūrtaḥ and mattaḥ, SK 3016 would have given n. This rule stops it.

Applies when / blocked by. One of Bhaṭṭoji's commentators adds that mūrchā is ādit, so it takes no iṭ by SK 3036. Its u is lengthened by hali ca (SK 354).

Watch out. Which khyā is meant is disputed. One commentator takes it as khyā prakathane "to tell". The other holds that this root has no use before an ārdhadhātuka affix, so naming it adds nothing; scholars discuss this.

III.D. Fixed forms in restricted senses — vitta, bhitta, ṛṇa — and sphī — running example: `vid (lābhe) + kta → vittam (dhanam) ~ vinnaḥ`

Three more nipātanas close the run of niṣṭhā-t rules from book 8. Each fixes a form for one sense only. vitta keeps its t where vinna would have n. bhitta keeps its t where bhinna has n. ṛṇa gets an n that no rule gives. A fourth rule, from book 6, gives sphī for sphāy. The running example is vid lābhe "to obtain". It gives vittam "wealth", but vinnaḥ "obtained".

SK 3041 · वित्तो भोगप्रत्यययोः · vitto bhogapratyayayoḥ (A 8.2.58)

वित्तो भोगप्रत्यययोः

vitto bhogapratyayayoḥ · A 8.2.58

विन्दतेर्निष्ठातस्य निपातोऽयं भोग्ये प्रतीते चार्थे । वित्तं धनम् । वित्तः पुरुषः । अनयोः किम् । विन्न । विभाषा गमहन -SK 3099 इति क्वसौ वेट्त्वादिह नेट् ॥

Rule. vitta is a nipātana: the niṣṭhā form of vindati (vid lābhe "to obtain"). It holds in two senses: "what is enjoyed" (bhoga) and "what is known" (pratyaya). One of Bhaṭṭoji's commentators explains pratyaya as "well known".

Example.

vid + kta → vid + ta → kept by nipātana → vittaṃ dhanam "wealth", what is enjoyed.

vittaḥ puruṣaḥ "a well-known man".

In other senses → vinnaḥ "obtained".

Applies when / blocked by. Why no iṭ? Bhaṭṭoji: vid lābhe takes iṭ optionally before kvasu, by SK 3099. kvasu is a past participle affix taught later in this chapter (SK 3095). A root that is optionally seṭ anywhere takes no iṭ before a niṣṭhā (yasya vibhāṣā, SK 3025). One commentator adds that SK 2246 gives the same result.

Watch out. Which n does the nipātana set aside? One commentator says the optional n of SK 3038. The other frames the rule against all the vid roots: vid jñāne must take iṭ, and the other two are not used in these senses. That is why Bhaṭṭoji says vindateḥ. Scholars discuss this.

SK 3042 · भित्तं शकलम् · bhittaṃ śakalam (A 8.2.59)

भित्तं शकलम्

bhittaṃ śakalam · A 8.2.59

भिन्नमन्यत् ॥

Rule. bhitta, with no n for the niṣṭhā-t of bhid, is a nipātana meaning "a fragment" (śakala). Bhaṭṭoji: "elsewhere, bhinnam".

Example.

bhid + kta → bhid + ta → no n, by this nipātana → bhittam "a fragment".

Elsewhere → n for both t and d (SK 3016) → bhinnam "split".

Watch out. One of Bhaṭṭoji's commentators says bhitta is rūḍha, a word fixed by usage. It means a fragment without regard to the sense "split". So bhitta and śakala are synonyms and are not used together. When the action of splitting is meant, the form is bhinnam.

SK 3043 · ऋणमाधमर्ण्ये · ṛṇam ādhamarṇye (A 8.2.60)

ऋणमाधमर्ण्ये

ṛṇamādhamarṇye · A 8.2.60

ऋधातोः क्ते तकारस्य नत्वं निपात्यते अधमर्णव्यवहारे । ऋतमन्यत् ॥

Rule. After the root ṛ, n replaces the t of kta in the context of debt. Bhaṭṭoji glosses it "in a debtor's transaction". Elsewhere the form is ṛtam.

Example.

ṛ + kta → ṛ + ta → by nipātana, n → ṛṇam "debt".

Elsewhere → ṛtam, which one of Bhaṭṭoji's commentators glosses "true".

Applies when / blocked by. One commentator describes the transaction. Another's property is taken and is to be repaid within a time, with interest. The other commentator explains Bhaṭṭoji's gloss. "Debt" here stands by secondary signification (lakṣaṇā) for the whole transaction. A transaction links giver and taker, so the word for the creditor, uttamarṇa (SK 573), is also accounted for.

SK 3044 · स्फायः स्फी निष्ठायाम् · sphāyaḥ sphī niṣṭhāyām (A 6.1.22)

स्फायः स्फी निष्ठायाम्

sphāyaḥ sphī niṣṭhāyām · A 6.1.22

स्फीतः ॥

Rule. sphī replaces sphāy before a niṣṭhā. One of Bhaṭṭoji's commentators names the root sphāyī vṛddhau "to grow".

Example.

sphāy + kta → this rule: sphī for sphāy → sphī + ta → sphītaḥ "grown".

Applies when / blocked by. One commentator meets an objection: how then is sphītikāma formed? From the ṇic-stem of "sphītam ācaṣṭe" ("he calls it grown"), with the uṇādi affix i (by aca iḥ).

Watch out. This rule's word niṣṭhāyām "before a niṣṭhā" is carried elsewhere. One commentator carries it back to SK 3033 and on to SK 3072.

III.E. iṭ restored to the niṣṭhā (and ktvā) against yasya vibhāṣā — running example: `nir + kuṣ + kta → niṣ + kuṣ + i + ta → niṣkuṣitaḥ`

SK 3025, yasya vibhāṣā, was taught near the end of Part II (group II.Q). A root that takes iṭ optionally anywhere takes no iṭ before a niṣṭhā. A root that may or may not take iṭ is called veṭ. The six rules here answer that ban for particular roots. Some give iṭ back without option, some make it optional. Most also cover ktvā, the kṛt affix -tvā taught in a later chapter. Bhaṭṭoji calls the first of them a renewed prescription (punarvidhi). It says again what a general rule had given, so that the ban cannot take it away. The running example is the root kuṣ after nir. It takes iṭ: niṣkuṣitaḥ.

SK 3045 · इण्निष्ठायाम् · iṇ niṣṭhāyām (A 7.2.47)

इण्निष्ठायाम्

iṇniṣṭhāyām · A 7.2.47

निरः कुषो निष्ठाया इट् स्यात् । यस्य विभाषा SK 3025 इति निषेधे प्राप्ते पुनर्विधिः । निष्कुषितः ॥

Rule. After kuṣ preceded by nir, iṭ comes to a niṣṭhā. Bhaṭṭoji: "after kuṣ preceded by nir, iṭ comes to a niṣṭhā". One of Bhaṭṭoji's commentators says the words "after nir-kuṣ" carry forward from niraḥ kuṣaḥ (SK 2560).

Example.

nir + kuṣ + kta → nir + kuṣ + ta

→ this rule: iṭ → nir + kuṣ + i + ta

→ niṣkuṣitaḥ.

Applies when / blocked by. Why is the rule needed? iṭ already comes by SK 2184. But SK 2560 makes nir-kuṣ take iṭ only optionally before tṛc and the like. Then yasya vibhāṣā (SK 3025) would ban iṭ before the niṣṭhā. Bhaṭṭoji: "a renewed prescription, since the ban by yasya vibhāṣā would apply". One commentator adds that the iṭ is obligatory. Restating the word iṭ stops the word "optionally" from carrying on from SK 2279.

Watch out. Bhaṭṭoji himself does not say "obligatorily" here. The point comes from his commentator.

SK 3046 · वसतिक्षुधोरिट् · vasatikṣudhor iṭ (A 7.2.52)

वसतिक्षुधोरिट्

vasatikṣudhoriṭ · A 7.2.52

आभ्यां क्त्वानिष्ठयोर्नित्यमिट् स्यात् । उषितः । क्षुधितः ॥

Rule. After vas "to dwell" (vasati) and kṣudh, iṭ comes to ktvā and to a niṣṭhā. Bhaṭṭoji says the iṭ is obligatory (nityam). One of Bhaṭṭoji's commentators draws "ktvā and a niṣṭhā" from SK 3049, which the SK teaches later.

Example.

vas + kta → this rule: iṭ → vas + i + ta

→ saṃprasāraṇa, since vas is in the yaj group (SK 2409) → us + i + ta

→ ṣ for s by SK 2410 → uṣitaḥ.

kṣudh + kta → kṣudh + i + ta → kṣudhitaḥ.

Both commentators give the steps of uṣitaḥ.

Applies when / blocked by. One commentator says what this rule overrides. Both roots would be aniṭ by ekāca upadeśe 'nudāttāt (SK 2246). The other explains the name vasati, with its class sign. The vas of the ad class (vaste) is seṭ and takes iṭ anyway.

Watch out. One commentator defends kṣudhitaḥ as a plain kta form. He rejects a derivation with the affix itac, given in a work on the roots.

SK 3047 · अञ्चेः पूजायाम् · añceḥ pūjāyām (A 7.2.53)

अञ्चेः पूजायाम्

añceḥ pūjāyām · A 7.2.53

पूजार्थादञ्चेः क्त्वानिष्ठयोरिट् स्यात् । अञ्चितः । गतौ तु अक्तः ॥

Rule. After añc in the sense "to honour" (pūjā), iṭ comes to ktvā and to a niṣṭhā. Bhaṭṭoji: "after añc meaning to honour, iṭ comes to ktvā and a niṣṭhā".

Example.

añc + kta (honour) → this rule: iṭ → añc + i + ta → añcitaḥ.

In the sense of going → no iṭ → the nasal is lost (SK 415) → ac + ta → k for c (coḥ kuḥ, SK 378) → aktaḥ.

Applies when / blocked by. Both of Bhaṭṭoji's commentators say why the rule is needed. añc is marked with u. So udito vā (SK 3328) would make iṭ optional before ktvā. And yasya vibhāṣā would ban it before a niṣṭhā. Both also say why añcitaḥ keeps its nasal: nāñceḥ pūjāyām (SK 424) forbids its loss.

Watch out. Unlike SK 3046 and SK 3048, Bhaṭṭoji does not say "obligatorily" here.

SK 3048 · लुभो विमोहने · lubho vimohane (A 7.2.54)

लुभो विमोहने

lubho vimohane · A 7.2.54

लुभः क्त्वानिष्ठयोर्नित्यमिट् स्यान्नतु गार्ध्ये । लुभितः । गार्ध्ये तु लुब्धः ॥

Rule. After lubh in the sense "to bewilder" (vimohana), iṭ comes to ktvā and to a niṣṭhā. Bhaṭṭoji: the iṭ is obligatory, "but not in the sense of greed".

Example.

lubh + kta (bewilder) → this rule: iṭ → lubh + i + ta → lubhitaḥ "bewildered".

In the sense of greed → lubdhaḥ "greedy".

Applies when / blocked by. One of Bhaṭṭoji's commentators separates two roots. lubh vimohane is of the tud class; lubh gārdhye "to be greedy" is of the div class. Only the first is meant here. Both commentators say why the rule is needed. tīṣasaha… (SK 2340) would make iṭ optional before ktvā. yasya vibhāṣā would then ban it before a niṣṭhā.

Watch out. One commentator adds that lubdhaḥ "desirous" has kta in the agent. He cites SK 3089 (group III.M) for this.

SK 3049 · क्लिशः क्त्वानिष्ठयोः · kliśaḥ ktvāniṣṭhayoḥ (A 7.2.50)

क्लिशः क्त्वानिष्ठयोः

kliśaḥ ktvāniṣṭhayoḥ · A 7.2.50

इड्वा स्यात् । क्लिश उपतापे नित्यं प्राप्ते । क्लिशू विबाधने । अस्य क्त्वायां विकल्पे सिद्धेऽपि निष्ठायां निषेधे प्राप्ते विकल्पः । क्लिशितः । क्लिष्टः ॥

Rule. After kliś, iṭ is optional before ktvā and a niṣṭhā. Bhaṭṭoji supplies "iṭ is optional". One of Bhaṭṭoji's commentators draws "optionally" from SK 2279.

Example.

kliś + kta → with iṭ → kliśitaḥ.

Without iṭ → kliṣṭaḥ.

Applies when / blocked by. Bhaṭṭoji says what the option does for each of two roots. For kliśa upatāpe "to be afflicted", iṭ would be obligatory. One commentator names SK 2184 for that. For kliśū vibādhane "to torment", the option before ktvā already exists. The root is marked with ū, and SK 2279 gives it. But before a niṣṭhā, yasya vibhāṣā would ban iṭ. This rule gives the option there too.

SK 3050 · पूङश्च · pūṅaś ca (A 7.2.51)

पूङश्च

pūṅaśca · A 7.2.51

पूङः क्त्वानिष्ठयोरिड्वा स्यात् ॥

Rule. After pūṅ too, iṭ is optional before ktvā and a niṣṭhā. One commentator lists the words carried forward. "ktvā and a niṣṭhā" come from SK 3049. "iṭ" comes from SK 3045. "Optionally" comes from SK 2279.

Example.

pū + kta → with iṭ → pū + i + ta.

Without iṭ → pū + ta.

The finished forms, pavitaḥ and pūtaḥ, come under the next rule.

Applies when / blocked by. Both commentators say what this overrides. śryukaḥ kiti (SK 2381) would ban iṭ before a kit affix after pū.

III.F. A seṭ niṣṭhā that is not kit, and kta at the start of an action — running example: `śī + kta → śī + i + ta (not kit) → śe + i + ta → śayitaḥ`

A niṣṭhā is marked with k, so it is kit. Before a kit affix the root's vowel takes no guṇa (SK 2217). Four rules of this group (SK 3051, 3052, 3055, 3056) take away that kit-hood from a niṣṭhā that has iṭ (a seṭ niṣṭhā). The root then takes guṇa. One of Bhaṭṭoji's commentators draws "not, when seṭ" from na ktvā seṭ (SK 3322). He draws "kit" from asaṃyogāl liṭ kit (SK 2242). Both are taught elsewhere in the SK. Three rules here (SK 3053, 3054, 3056), and a vārttika under SK 3052, speak of ādikarman, the beginning of an action. SK 3053 says which sense a kta has there. SK 3054 is an iṭ rule. The running example is śīṅ "to lie down". It gives śayitaḥ, with guṇa.

SK 3051 · पूङः क्त्वा च · pūṅaḥ ktvā ca (A 1.2.22)

पूङः क्त्वा च

pūṅaḥ ktvā ca · A 1.2.22

पूङः क्त्वा निष्ठा च सेट् किन्न स्यात् । पवितः । पूतः । क्त्वाग्रहणमुत्तरार्थम् । नोपधात् - SK 3324 इत्यत्र हि क्त्वैव संबध्यते ॥

Rule. After pūṅ, a ktvā and a niṣṭhā that have iṭ are not kit. Bhaṭṭoji: "after pūṅ, a seṭ ktvā and niṣṭhā are not kit". One of Bhaṭṭoji's commentators draws "niṣṭhā" from SK 3052, the next rule in the SK.

Example.

pū + kta → with iṭ (SK 3050) → pū + i + ta

→ this rule: not kit, so guṇa → po + i + ta

→ o becomes av before a vowel (SK 61) → pavitaḥ.

Without iṭ the affix stays kit, and there is no guṇa → pūtaḥ.

Applies when / blocked by. A seṭ ktvā is already not kit by na ktvā seṭ. So why name ktvā here? Bhaṭṭoji: "for the following rule". In nopadhāt… (SK 3324) only ktvā is to carry on. The two commentators give different reasons why the niṣṭhā does not carry on; scholars discuss this.

SK 3052 · निष्ठा शीङ्स्विदिमिदिक्ष्विदिधृषः · niṣṭhā śīṅsvidimidikṣvididhṛṣaḥ (A 1.2.19)

निष्ठा शीङ्स्विदिमिदिक्ष्विदिधृषः

niṣṭhā śīṅsvidimidikṣvididhṛṣaḥ · A 1.2.19

एभ्यः सेण्निष्ठा किन्न स्यात् । शयितः । शयितवान् । अनुबन्धनिर्देशो यङ्लुङ्निवृत्त्यर्थः । शेश्यितः । शेश्यितवान् ।

आदिकर्मणि निष्ठा वक्तव्या

(वार्तिकम्) ॥

Rule. After śīṅ, svid, mid, kṣvid and dhṛṣ, a niṣṭhā that has iṭ is not kit. Bhaṭṭoji: "after these, a seṭ niṣṭhā is not kit".

Example.

śī + kta → iṭ → śī + i + ta

→ this rule: not kit, so guṇa → śe + i + ta

→ e becomes ay (SK 61) → śayitaḥ.

With ktavatu → śayitavān.

Applies when / blocked by. Bhaṭṭoji: naming śīṅ with its marker ṅ excludes the yaṅluk form. yaṅluk is the intensive stem whose affix yaṅ has been elided. There the niṣṭhā stays kit, so there is no guṇa: śeśyitaḥ, śeśyitavān. One of Bhaṭṭoji's commentators gives the y by er anekācaḥ…. The other cites the maxim behind the marker: a root named with its marker is not taken in yaṅluk.

A vārttika: ādikarmaṇi niṣṭhā vaktavyā "a niṣṭhā should be stated in ādikarman". ādikarman is the beginning of an action. Both commentators explain why it is needed. A long action, like making a mat, may have begun. Its first moments are past, but the action as a whole is not. A niṣṭhā taught for the past would not reach it.

Watch out. One commentator reports others who say more. For them the vārttika also gives kta in the agent sense in ādikarman.

SK 3053 · आदिकर्मणि क्तः कर्तरि च · ādikarmaṇi ktaḥ kartari ca (A 3.4.71)

आदिकर्मणि क्तः कर्तरि च

ādikarmaṇi ktaḥ kartari ca · A 3.4.71

आदिकर्मणि यः क्तः स कर्तरि स्यात् चाद्भावकर्मणोः ॥

Rule. A kta used for the beginning of an action (ādikarman) denotes the agent. By the word ca it also denotes bhāva (the bare action) and karman (the object). Bhaṭṭoji: "the kta in ādikarman is in the agent; by ca, in bhāva and karman".

Example. Bhaṭṭoji gives none here. He returns to the rule under SK 3080:

pra + kṛ + kta (agent) → prakṛtaḥ kaṭaṃ saḥ "he has begun to make the mat".

pra + kṛ + kta (object) → prakṛtaḥ kaṭas tena "the mat has been begun by him".

Applies when / blocked by. One of Bhaṭṭoji's commentators draws bhāva and karman from tayor eva kṛtyaktakhalarthāḥ (SK 2833). The other says why ca is needed. Without it, the maxim of the buttermilk and Kauṇḍinya (takrakauṇḍinya-nyāya) would apply. Naming the agent would then shut out the other two senses.

SK 3054 · विभाषा भावादिकर्मणोः · vibhāṣā bhāvādikarmaṇoḥ (A 7.2.17)

विभाषा भावादिकर्मणोः

vibhāṣā bhāvādikarmaṇoḥ · A 7.2.17

भावे आदिकर्मणि चादितो निष्ठाया इड्वा स्यात् । प्रस्वेदितश्चैत्रः । प्रस्वेदितं तेन । ञिष्विदेति भ्वादिरत्र गृह्यते । ञीद्भिः साहचर्यात् । स्विद्यतेस्तु स्विदित इत्येव । ञिमिदा । ञिक्ष्विदा । दिवादी भ्वादी च । प्रमेदितः । प्रमेदितवान् । प्रक्ष्वेदितः । प्रक्ष्वेदितवान् । प्रधर्षितः । प्रधर्षितवान् । प्रधर्षितं तेन । सेट् किम् । प्रस्विन्नः । प्रस्विन्नं तेनेत्यादि ॥

Rule. In bhāva and in ādikarman, iṭ comes optionally to a niṣṭhā after an ādit root. Normally SK 3036 bans it. One of Bhaṭṭoji's commentators lists the words carried: "after an ādit" from SK 3036, "niṣṭhā" from SK 3039, "not" from SK 2981. So the ban becomes optional.

Example.

pra + svid + kta (ādikarman, agent) → this rule: iṭ → pra + svid + i + ta

→ seṭ, so not kit (SK 3052) → guṇa of the light penultimate (SK 2189) → pra + sved + i + ta

→ prasveditaś caitraḥ "Caitra has begun to sweat".

In bhāva → prasveditaṃ tena.

One commentator says the pra marks the beginning of the action.

Applies when / blocked by. Which svid does SK 3052 name? Bhaṭṭoji: the bhū-class ñiṣvidā, by association with the roots marked ñī. From the div-class svid (svidyati) the form is only sviditaḥ. It stays kit and takes no guṇa. ñimidā and ñikṣvidā belong to both classes. So: prameditaḥ, prameditavān; prakṣveditaḥ, prakṣveditavān. And from dhṛṣ: pradharṣitaḥ, pradharṣitavān, pradharṣitaṃ tena.

Why "seṭ" in SK 3052? Bhaṭṭoji: prasvinnaḥ, prasvinnaṃ tena. Without iṭ the niṣṭhā stays kit, and the t becomes n.

Watch out. Haradatta takes only one kṣvid here, of the div class. One commentator rejects this, since midi too is read in both classes. That is why Bhaṭṭoji says "divādi and also bhvādi".

SK 3055 · मृषस्तितिक्षायाम् · mṛṣas titikṣāyām (A 1.2.20)

मृषस्तितिक्षायाम्

mṛṣastitikṣāyām · A 1.2.20

सेण्निष्ठाकिन्न स्यात् । मर्षितः । मर्षितवान् । क्षमायां किम् । अपमृषितं वाक्यम् । अविस्पष्टमित्यर्थः ॥

Rule. After mṛṣ in the sense of forbearance (titikṣā), a niṣṭhā with iṭ is not kit. Bhaṭṭoji: "a seṭ niṣṭhā is not kit". One of Bhaṭṭoji's commentators draws "niṣṭhā" from SK 3052.

Example.

mṛṣ + kta → iṭ → mṛṣ + i + ta

→ this rule: not kit, so guṇa → marṣ + i + ta → marṣitaḥ.

With ktavatu → marṣitavān.

Applies when / blocked by. In another sense the niṣṭhā stays kit, and there is no guṇa. Bhaṭṭoji: apamṛṣitaṃ vākyam "an indistinct utterance".

Watch out. One commentator notes that the list of roots gives mṛṣ only in the sense of forbearance. The sense named in this rule shows that roots have more than one meaning.

SK 3056 · उदुपधाद्भावादिकर्मणोरन्यतरस्याम् · udupadhād bhāvādikarmaṇor anyatarasyām (A 1.2.21)

उदुपधाद्भावादिकर्मणोरन्यतरस्याम्

udupadhādbhāvādikarmaṇoranyatarasyām · A 1.2.21

उदुपधात्परा भावादिकर्मणोः सेण्निष्ठा वा किन्न स्यात् । द्युतितम् । द्योतितम् । मुदितम् । मोदितं साधुना । प्रद्युतितः । प्रद्योतितः । प्रमुदितः । प्रमोदितः साधुः । उदुपधात्किम् । विदितम् । भावेत्यादि किम् । रुचितं कार्षापणम् । सेट्किम् । क्रुष्टम् ।

शब्विकरणेभ्य एवेष्यते

(वार्तिकम्) ॥ नेह । गुध्यतेर्गुधितम् ॥

Rule. After a root whose penultimate (upadhā) is u, in bhāva and ādikarman, a seṭ niṣṭhā is optionally not kit. Bhaṭṭoji: "a seṭ niṣṭhā after a root with penultimate u is optionally not kit in bhāva and ādikarman".

Example.

dyut + kta (bhāva) → iṭ → dyut + i + ta

→ kit: no guṇa → dyutitam; not kit: guṇa → dyotitam.

Likewise muditam / moditaṃ sādhunā.

In ādikarman: pradyutitaḥ / pradyotitaḥ, pramuditaḥ / pramoditaḥ sādhuḥ.

One of Bhaṭṭoji's commentators assigns the examples: the plain forms are bhāva, the pra- forms ādikarman.

Applies when / blocked by. Bhaṭṭoji gives a counterexample for each condition.

No penultimate u: viditam.

Not bhāva or ādikarman: rucitaṃ kārṣāpaṇam.

Not seṭ: kruṣṭam.

One commentator says rucitam has kta in the agent, by SK 3086.

A vārttika: śabvikaraṇebhya eveṣyate "this is wanted only for roots with the class sign śap". A class sign (vikaraṇa) is the affix between root and ending. Bhaṭṭoji: not in gudhitam, from gudhyati. Both commentators say gudh "to wrap" is of the div class.

III.G. ṇi lost before a seṭ niṣṭhā, and where the niṣṭhā keeps or lacks iṭ — running example: `bhū + ṇic → bhāvi + i + ta → bhāv + ita → bhāvitaḥ`

A causative stem ends in ṇi, the affix ṇic without its markers. Before an ārdhadhātuka affix without iṭ, ṇi is lost by SK 2313. Before a seṭ niṣṭhā that rule cannot act. SK 3057 fills the gap. Bhaṭṭoji then uses the rule as a place to collect niṣṭhā forms. Some lack iṭ, some keep it against yasya vibhāṣā. The running example is bhāvi, the causative stem of bhū "to be". It gives bhāvitaḥ.

SK 3057 · निष्ठायां सेटि · niṣṭhāyāṃ seṭi (A 6.4.52)

निष्ठायां सेटि

niṣṭhāyāṃ seṭi · A 6.4.52

णेर्लोपः स्यात् । भावितः । भावितवान् । श्वीदितः - SK 3039 इति नेट् । संप्रसारणम् । शूनः । दीप्तः । गुहू, गूढः । वनु, वतः । तनु, ततः । पतेः सनि वेट्कत्वादिडभावे प्राप्ते द्वितीयाश्रित - SK 686 इति सूत्रे वेट्कत्वात्सिद्धे कृन्तत्यादीनामीदित्त्वेनानित्यत्वज्ञापनाद्वा । तेन धावितमिभराजधियेत्यादि । यस्य विभाषा SK 3025 इत्यत्रैकाच इत्येव । दरिद्रतः ॥

Rule. Before a niṣṭhā that has iṭ, ṇi is deleted. Bhaṭṭoji: "ṇi is deleted". One of Bhaṭṭoji's commentators draws "ṇi" from ṇer aniṭi (SK 2313) and "deleted" from SK 2372. SK 2313 works only before an affix without iṭ, so a new rule is needed.

Example.

bhāvi + kta → iṭ → bhāvi + i + ta

→ this rule: ṇi lost → bhāv + ita → bhāvitaḥ.

With ktavatu → bhāvitavān.

Applies when / blocked by. Bhaṭṭoji then lists more niṣṭhā forms.

No iṭ, by śvīditaḥ (SK 3039).

śvi + kta → no iṭ → saṃprasāraṇa → śūnaḥ.

Both commentators give the steps. Saṃprasāraṇa, since śvi is in the yaj group (SK 2409), gives śu. halaḥ (SK 2559) lengthens it to śū. oditaś ca (SK 3019) gives n for the t.

Also dīptaḥ, from the īdit dīpī "to shine".

No iṭ, by yasya vibhāṣā. These roots are veṭ.

guhū → gūḍhaḥ. Both commentators give the steps: ḍh for h, dh for t, retroflexion (ṣṭutva), loss of the first ḍh, and lengthening.

vanu → vataḥ; tanu → tataḥ. Their n is lost by SK 2428.

iṭ kept against yasya vibhāṣā.

pat is veṭ before san, so the ban should apply.

But Bhaṭṭoji says the iṭ is given by nipātana in dvitīyā śrita… (SK 686) → patitaḥ.

He gives a second reason. kṛt, cṛt and nṛt are veṭ by se 'sici (SK 2506). Their lack of iṭ in the niṣṭhā would follow anyway. Yet they are marked with ī. That marking would be pointless, so it shows that yasya vibhāṣā is not binding (anitya). Such a hint is a jñāpaka: a feature of the grammar that makes a principle known.

Bhaṭṭoji: "hence dhāvitam in ibharājadhiyā". One commentator: dhūñ is veṭ, and yet the ban does not apply.

Only for one-vowelled roots.

Bhaṭṭoji: in yasya vibhāṣā, "having one vowel" (ekāc) is understood.

So daridrā, with more than one vowel, keeps its iṭ → daridritaḥ.

Watch out. daridrā is veṭ before san, like pat. One commentator raises this point. It still keeps its iṭ in the niṣṭhā, because it has more than one vowel.

III.H. Aniṭ niṣṭhā forms fixed for particular senses — running example: `kṣubh + kta → kṣubh + ta → kṣubdhaḥ (mantha) ~ kṣubhitam`

These eight rules continue the iṭ rules of book 7. Most take up a root that would take iṭ in the niṣṭhā. In a stated sense the form is aniṭ, without iṭ. Several forms are nipātanas, with further irregular parts. SK 3059 works the other way: it restricts an aniṭ form to one sense. The running example is kṣubh "to shake". For mantha, barley-meal mixed with a liquid, it gives kṣubdhaḥ, with no iṭ. Elsewhere it gives kṣubhitam.

SK 3058 · क्षुब्धस्वान्तध्वान्तलग्नम्लिष्टविरिब्धफाण्टबाढानि मन्थमनस्तमःसक्ताविस्पष्टस्वरानायासभृशेषु · kṣubdhasvāntadhvāntalagnamliṣṭaviribdhaphāṇṭabāḍhāni manthamanastamaḥsaktāvispaṣṭasvarānāyāsabhṛśeṣu (A 7.2.18)

क्षुब्धस्वान्तध्वान्तलग्नम्लिष्टविरिब्धफाण्टबाढानि मन्थमनस्तमःसक्ताविस्पष्टस्वरानायासभृशेषु

kṣubdhasvāntadhvāntalagnamliṣṭaviribdhaphāṇṭabāḍhāni manthamanastamaḥsaktāvispaṣṭasvarānāyāsabhṛśeṣu · A 7.2.18

क्षुब्धादीन्यष्टावनिट्कानि निपात्यन्ते समुदायेन मन्थादिषु वाच्येषु । द्रवद्रव्यसंपृक्ताः सक्तवो मन्थो मन्थनदण्डश्च । क्षुब्धो मन्थश्चेत् । स्वान्तं मनः । ध्वान्तं तमः । लग्नं सक्तम् । निष्ठानत्वमपि निपातनात् । म्लिष्टविस्पष्टम् । विरिब्धः स्वरः । म्लेच्छ रेभृ अनयोरुपधाया इत्वमपि निपात्यते । फाण्टमनायाससाध्यः कषायविशेषः । माधवस्तु नवनीतभावात्प्रागवस्थापन्नं द्रव्यं फाण्टमिति वेदभाष्ये आह । बाढं भृशम् । अन्यत्र तु क्षुभितम् । क्षुब्धो राजेति त्वागमशास्त्रस्यानित्यत्वात् । स्वनितम् । ध्वनितम् । लगितम् । म्लेच्छितम् । विरेभितम् । फणितम् । बाहितम् ॥

Rule. Eight aniṭ niṣṭhā forms are given by nipātana for eight senses, in order. The words are paired one to one (yathāsaṃkhya), says one of Bhaṭṭoji's commentators. Bhaṭṭoji: they are given "as whole words" (samudāyena). One commentator explains: they are fixed by usage (rūḍha), and the sense of the parts should not be pressed.

Example. Bhaṭṭoji lists the eight with their senses. One commentator gives the steps.

kṣubh + ta → no iṭ → dh for t (SK 2280), b for bh → kṣubdhaḥ, of mantha: barley-meal mixed with a liquid, or the churning-stick.

svan + ta, dhvan + ta → no iṭ → lengthening (SK 2666) → svāntam "mind", dhvāntam "darkness".

lag + ta → lagnam "attached". Bhaṭṭoji: its n for the niṣṭhā-t is also nipātana.

mlecch + ta → mliṣṭam "indistinct".

vi + rebh + ta → no iṭ → i for e (nipātana) → dh for t (SK 2280), b for bh → viribdhaḥ "a tone". Bhaṭṭoji: the i of mlecch and rebhṛ is nipātana too.

phaṇ + ta → phāṇṭam, a decoction made without effort.

bāh + ta → bāḍham "intensely".

Applies when / blocked by. In other senses iṭ comes. Bhaṭṭoji: kṣubhitam, svanitam, dhvanitam, lagitam, mlecchitam, virebhitam, phaṇitam, bāhitam. But kṣubdho rājā "the king is agitated" is also right. Bhaṭṭoji's reason: the teaching of augments (āgama-śāstra) is not binding (anitya).

Watch out. Bhaṭṭoji records a second view of phāṇṭa. Mādhava, in his Veda-commentary, calls it the substance just before it becomes fresh butter.

SK 3059 · धृषिशसी वैयात्ये · dhṛṣiśasī vaiyātye (A 7.2.19)

धृषिशसी वैयात्ये

dhṛṣiśasī vaiyātye · A 7.2.19

एतौ निष्ठायामविनये एवानिटौ स्तः । धृष्टः । विशस्तः । अन्यत्र धर्षितः । विशसितः । भावादिकर्मणोस्तु वैयात्ये धृषिर्नास्ति । अत एव नियमार्थमिदं सूत्रमिति वृत्तिः । धृषेरादित्त्वे फलं चिन्त्यमिति हरदत्तः । माधवस्तु भावादिकर्मणोरवैयात्ये विकल्पमाह । धृष्टम् । धर्षितम् । प्रधृष्टः । प्रधर्षितः ॥

Rule. dhṛṣ and śas are aniṭ before a niṣṭhā only in the sense of impudence (vaiyātya). Bhaṭṭoji: "these two are aniṭ before a niṣṭhā only in immodesty".

Example.

dhṛṣ + kta → no iṭ → dhṛṣṭaḥ; vi + śas + kta → viśastaḥ. One of Bhaṭṭoji's commentators glosses both "ill-behaved".

Elsewhere → iṭ → dharṣitaḥ "treated with violence", viśasitaḥ "injured" (his glosses).

Applies when / blocked by. Both commentators say why the rule is a restriction (niyama). ñidhṛṣā is ādit, so SK 3036 already bans iṭ. śasu is marked with u, so it is veṭ, and yasya vibhāṣā bans iṭ. The rule adds nothing new. It restricts: aniṭ only in impudence.

Bhaṭṭoji then reports three views. The Vṛtti: in bhāva and ādikarman dhṛṣ does not occur in the sense of impudence; so the rule is restrictive. Haradatta: the purpose of dhṛṣ's marker ā "is to be considered". Mādhava: in bhāva and ādikarman, outside impudence, there is an option: dhṛṣṭam / dharṣitam, pradhṛṣṭaḥ / pradharṣitaḥ.

Watch out. One commentator turns Haradatta's doubt back on him. On Mādhava's view the marker ā has a purpose after all. Scholars discuss this.

SK 3060 · दृढः स्थूलबलयोः · dṛḍhaḥ sthūlabalayoḥ (A 7.2.20)

दृढः स्थूलबलयोः

dṛḍhaḥ sthūlabalayoḥ · A 7.2.20

स्थूले बलवति न निपात्यते । दृह दृहि वृद्धौ । क्तस्येडभावः तस्य ढत्वम् । हस्य लोपः । इदितो नलोपश्च । दृहितः । दृंहितोऽन्यः ॥

Rule. dṛḍha is a nipātana in the senses "stout" (sthūla) and "strong" (bala). Both of Bhaṭṭoji's commentators read bala as "possessing strength", with the affix ac of the arśaādi list. One adds that the two senses differ: a stout man may be weak, a strong man may be lean.

Example. The roots are dṛha and dṛhi "to grow". Bhaṭṭoji lists what the nipātana gives.

dṛh + kta → no iṭ → ḍh for the t → the h is deleted → dṛḍhaḥ.

For dṛhi, marked with i, the inserted nasal is also deleted → dṛḍhaḥ.

Applies when / blocked by. In other senses the forms take iṭ: dṛhitaḥ, dṛṃhitaḥ.

Watch out. Why must the loss of h be stated by nipātana? Both commentators answer. The regular route would be ḍh for h, then ḍh for t, then loss of the first ḍh (SK 2335). But that loss stands in the Tripādī, so it is asiddha. Then the ṛ would seem to stand before a conjunct. Later forms like draḍhimā would go wrong. A deletion stated in book 7 is not asiddha, so it gives the right result.

SK 3061 · प्रभौ परिवृढः · prabhau parivṛḍhaḥ (A 7.2.21)

प्रभौ परिवृढः

prabhau parivṛḍhaḥ · A 7.2.21

वृह वृहि वृद्धौ । निपातनं प्राग्वत् । परिवृहितः । परिवृंहितोऽन्यः ॥

Rule. parivṛḍha is a nipātana in the sense "master" (prabhu). Bhaṭṭoji: the roots are vṛha and vṛhi "to grow", and "the nipātana is as before".

Example.

pari + vṛh + kta → as in SK 3060: no iṭ, ḍh for t, h lost → parivṛḍhaḥ "master".

Applies when / blocked by. In other senses: parivṛhitaḥ, parivṛṃhitaḥ. One of Bhaṭṭoji's commentators says the point is the same as before. It secures later forms like vraḍhimā and parivraḍhayya.

SK 3062 · कृच्छ्रगहनयोः कषः · kṛcchragahanayoḥ kaṣaḥ (A 7.2.22)

कृच्छ्रगहनयोः कषः

kṛcchragahanayoḥ kaṣaḥ · A 7.2.22

कषो निष्ठाय इण्न स्यादेतयोरर्थयोः । कष्टं दुःखं तत्कारणं च । स्तात्कष्टं कृच्छ्रमाभीलम् । कष्टो मोहः । कष्टं शास्त्रम् । दुरवगाहमित्यर्थः । कषितमन्यत् ॥

Rule. The niṣṭhā after kaṣ takes no iṭ in two senses: "painful" (kṛcchra) and "hard to penetrate" (gahana). One of Bhaṭṭoji's commentators glosses kaṣ "to harm".

Example.

kaṣ + kta → no iṭ → ṭ for t after ṣ (ṣṭutva) → kaṣṭa.

Bhaṭṭoji: kaṣṭa means pain and also its cause. He quotes the line syāt kaṣṭaṃ kṛcchram ābhīlam, from Amara's lexicon (one commentator says).

kaṣṭo mohaḥ "delusion is painful", a cause of pain.

kaṣṭaṃ śāstram "a difficult śāstra", one hard to penetrate.

Applies when / blocked by. In other senses: kaṣitam.

Watch out. How does kṛcchra reach "cause of pain"? One commentator says the word means both. The other says the second sense comes by secondary signification (lakṣaṇā). Scholars discuss this.

SK 3063 · घुषिरविशब्दने · ghuṣir aviśabdane (A 7.2.23)

घुषिरविशब्दने

ghuṣiraviśabdane · A 7.2.23

घुषिर्निष्ठायामनिट् स्यात् । घुष्टा रज्जुः । अविशब्दने किम् । घुषितं वाक्यम् । शब्देन प्रकटीकृताभिप्रायमित्यर्थः ॥

Rule. ghuṣ(ir) takes no iṭ in the niṣṭhā, except in the sense "making known by words" (viśabdana). The ir is the root's marker.

Example.

ghuṣ + kta → no iṭ → ghuṣṭā rajjuḥ. One of Bhaṭṭoji's commentators glosses it "a rope that has been made" or "stretched".

In the sense of making known → iṭ → ghuṣitaṃ vākyam. Bhaṭṭoji: "a statement whose intent has been made plain by words".

Watch out. One commentator raises a difficulty. The root in the sense of making known takes the curādi affix ṇic. Then ṇic stands between root and niṣṭhā, so the ban could not reach it anyway. His answer: the very exclusion is a jñāpaka. It shows that this root's ṇic is optional.

SK 3064 · अर्देः संनिविभ्यः · ardeḥ saṃnivibhyaḥ (A 7.2.24)

अर्देः संनिविभ्यः

ardeḥ saṃnivibhyaḥ · A 7.2.24

एतत्पूर्वादर्देर्निष्ठाया इण्न स्यात् । समर्णः । न्यर्णः । व्यर्णः । अर्दितोऽन्यः ॥

Rule. After the root ard preceded by sam, ni or vi, the niṣṭhā takes no iṭ. Bhaṭṭoji: "the niṣṭhā after ard preceded by these does not take iṭ".

Example.

sam + ard + kta → no iṭ → sam + ard + ta

→ n for the t and for the d before it (SK 3016) → sam + arn + na

→ ṇatva after r → samarṇaḥ.

Likewise nyarṇaḥ, vyarṇaḥ.

These steps are one commentator's (in a fuller copy of his text); Bhaṭṭoji gives the forms only.

Applies when / blocked by. Without these upasargas: arditaḥ.

SK 3065 · अभेश्चाविदूर्ये · abheś cāvidūrye (A 7.2.25)

अभेश्चाविदूर्ये

abheścāvidūrye · A 7.2.25

अभ्यर्णम् । नातिदूरं नासन्नं वा । अभ्यर्दितमन्यत् ॥

Rule. ard after abhi also takes no iṭ in the niṣṭhā, when the sense is "not being far" (āvidūrya).

Example.

abhi + ard + kta → no iṭ → n for t and d, then ṇ → abhyarṇam.

Bhaṭṭoji gives only the form. The steps shown are those of SK 3064, carried over; no commentator gives them here.

Bhaṭṭoji glosses it "not very far", or "not near".

Applies when / blocked by. Otherwise: abhyarditam.

Watch out. Why does the rule not simply say "near" (sāmīpya)? Both of Bhaṭṭoji's commentators answer. "Near" would leave out "not very far". Bhaṭṭoji signals this with his double gloss. One commentator keeps the two glosses apart.

III.I. The niṣṭhā from ṇic-stems, and the options of iṭ — running example: `dami + kta → dāntaḥ / damitaḥ`

These six rules go on with the iṭ rules of book 7, with one rule from book 6. The first three deal with causative stems in ṇi. SK 3066 fixes vṛtta. SK 3067 fixes śṛta for śrā and its causative. SK 3068 gives seven forms that are optional nipātanas. The rest give iṭ as an option where it was banned or required. The running example is dami, the causative stem of damu upaśame. It gives dāntaḥ or damitaḥ.

SK 3066 · णेरध्ययने वृत्तम् · ṇer adhyayane vṛttam (A 7.2.26)

णेरध्ययने वृत्तम्

ṇeradhyayane vṛttam · A 7.2.26

ण्यन्ताद्वृत्तेः क्तस्येडभावो णिलुक्चाधीयमानेऽर्थे । वृत्तं छन्दश्छात्रेण । संपादितम् । अधीतमिति यावत् । अन्यत्र तु वर्तिताः रज्जुः ॥

Rule. From vṛt ending in ṇic, in the sense "studied" (adhyayana), vṛtta is a nipātana. Bhaṭṭoji: the nipātana gives two things. The kta takes no iṭ, and ṇic is removed by luk.

Example.

varti + kta (the causative of vṛt) → no iṭ → ṇic removed by luk → vṛt + ta

→ vṛttaṃ chandaś chātreṇa "the metre has been gone through by the student".

Bhaṭṭoji glosses it "accomplished", that is, learned.

Applies when / blocked by. In other senses: vartitā rajjuḥ "a twisted rope".

Watch out. It is luk, not lopa. Both of Bhaṭṭoji's commentators say why it matters. If ṇic were removed by lopa, the stem would still count it present (pratyayalakṣaṇa). Then the root would take guṇa. With luk it keeps its ṛ.

SK 3067 · शृतं पाके · śṛtaṃ pāke (A 6.1.27)

शृतं पाके

śṛtaṃ pāke · A 6.1.27

श्रातिश्रपयत्योः क्ते शृभावो निपात्यते । क्षीरहविषोः पाके । शृतं क्षीरम् । स्वयमेव विक्लिन्नं पक्वं वेत्यर्थः । क्षीरहविर्भ्यामन्यत्तु श्राणं श्रपितं वा ॥

Rule. Before kta, śṛ is a nipātana in place of śrā and of its causative śrapi. It holds when milk or an oblation (havis) is cooked. Bhaṭṭoji: "in the cooking of milk or of an oblation".

Example.

śrā + kta or śrapi + kta → śṛ by nipātana → śṛ + ta

→ śṛtaṃ kṣīram.

Bhaṭṭoji glosses it two ways. Milk "softened of itself", or milk "cooked".

One of Bhaṭṭoji's commentators matches the glosses to the two roots. Plain śrā is intransitive, and its kta is in the agent (SK 3086). The causative śrapi means the same as pac, and its kta is in the object.

Applies when / blocked by. For things other than milk and oblation: śrāṇam or śrapitam.

Watch out. Where does the limit to milk and oblation come from? One commentator: from a vārttika. The other: from the "optionally" of SK 3023, carried here as a fixed option. Scholars discuss this.

SK 3068 · वा दान्तशान्तपूर्णदस्तस्पष्टच्छन्नज्ञप्ताः · vā dāntaśāntapūrṇadastaspaṣṭacchannajñaptāḥ (A 7.2.27)

वा दान्तशान्तपूर्णदस्तस्पष्टच्छन्नज्ञप्ताः

vā dāntaśāntapūrṇadastaspaṣṭacchannajñaptāḥ · A 7.2.27

एते णिचि निष्ठान्ता वा निपात्यते । पक्षे । दमितः । शमितः । पूरितः । दासितः । स्पाशितः । छादितः । ज्ञपितः ॥

Rule. Seven niṣṭhā forms from causative stems are optional nipātanas. They are dānta, śānta, pūrṇa, dasta, spaṣṭa, channa and jñapta. One of Bhaṭṭoji's commentators draws "from ṇic-stems" from SK 3066.

Example. One commentator gives the steps.

dami + kta → no iṭ (nipātana) → ṇi lost → lengthening (SK 2666) → dāntaḥ.

śami + kta → in the same way → śāntaḥ.

pūri + kta → no iṭ → ṇi lost → n after r (SK 3016) → ṇ → pūrṇaḥ.

chādi + kta → no iṭ, no vṛddhi (nipātana) → ṇi lost → n (SK 3016) → channaḥ.

Likewise dastaḥ, spaṣṭaḥ, jñaptaḥ.

Applies when / blocked by. In the other option the forms have iṭ. Bhaṭṭoji: damitaḥ, śamitaḥ, pūritaḥ, dāsitaḥ, spāśitaḥ, chāditaḥ, jñapitaḥ. One commentator: damitaḥ and śamitaḥ have a short vowel because the roots are mit. Such roots shorten their vowel before ṇi (mitāṃ hrasvaḥ, SK 2568).

Watch out. Some held that pūrṇa is redundant here, given SK 3040, which already blocks n after pṝ. One commentator answers that pūrṇa here comes from the causative of a different root, pūrī. Some texts read jñāpitaḥ for jñapitaḥ; scholars discuss this.

SK 3069 · रुष्यमत्वरसंघुषास्वनाम् · ruṣyamatvarasaṃghuṣāsvanām (A 7.2.28)

रुष्यमत्वरसंघुषास्वनाम्

ruṣyamatvarasaṃghuṣāsvanām · A 7.2.28

एभ्यो निष्ठाया इड्वा । रुषितः । रुष्टः । आन्तः । अमितः । तूर्णः । त्वरितः । अस्याऽऽत्त्वे फलं मन्दम् । संघुष्टः । संघुषितः । आस्वान्तः । आस्वनितः ॥

Rule. After ruṣ, am, tvar, saṃ-ghuṣ and ā-svan, the niṣṭhā optionally takes iṭ. One of Bhaṭṭoji's commentators: "optionally" carries from SK 3068, but "from ṇic-stems" stops.

Example.

ruṣ + kta → ruṣitaḥ / ruṣṭaḥ.

am + kta → amitaḥ; or no iṭ → lengthening (SK 2666) → āntaḥ.

tvar + kta → tvaritaḥ; or no iṭ → ū for v (SK 2654) → tūr + ta → n (SK 3016), then ṇ → tūrṇaḥ.

sam + ghuṣ + kta → saṃghuṣṭaḥ / saṃghuṣitaḥ.

ā + svan + kta → āsvāntaḥ / āsvanitaḥ.

The steps of āntaḥ and tūrṇaḥ are given by both commentators.

Applies when / blocked by. Both commentators say what each option overrides. ruṣ is veṭ by SK 2340, so yasya vibhāṣā would ban iṭ. ñitvarā is ādit, so SK 3036 would ban it. Here both become optional.

Watch out. Bhaṭṭoji remarks that tvar's marker ā "has little purpose". Both commentators explain: it only shows the middle voice (ātmanepada), which a short marker would show as well. For saṃ-ghuṣ, the two commentators account for the option differently; scholars discuss this.

SK 3070 · हृषेर्लोमसु · hṛṣer lomasu (A 7.2.29)

हृषेर्लोमसु

hṛṣerlomasu · A 7.2.29

हृषेर्निष्ठाया इड्वा स्यात् लोमसु विषये । हृष्टं हृषितं लोम ।

विस्मितप्रतिघातयोश्च

(वार्तिकम्) ॥ हृषो हृषितो मैत्रः । विस्मितः प्रतिहतो वेत्यर्थः । अन्यत्र तु । हृषु अलीके । उदित्वान्निष्ठायां नेट् । हृष तुष्टौ इट् ॥

Rule. After hṛṣ, the niṣṭhā optionally takes iṭ when the hair (loman) is concerned. One of Bhaṭṭoji's commentators: the hairs are the agent.

Example.

hṛṣ + kta → hṛṣṭaṃ loma / hṛṣitaṃ loma "the hair has bristled".

One commentator: the kta is in the agent, by SK 3086.

Applies when / blocked by. A vārttika: vismitapratighātayoś ca "also in astonishment and in being thwarted". Bhaṭṭoji: hṛṣṭo / hṛṣito maitraḥ "Maitra is astonished", or "thwarted".

Elsewhere there is no option. Bhaṭṭoji names two roots. hṛṣu alīke "to speak falsely" is marked with u, so it takes no iṭ in the niṣṭhā. hṛṣa tuṣṭau "to be pleased" takes iṭ. One commentator: so hṛṣṭa means "one who has lied", and hṛṣita "pleased".

SK 3071 · अपचितश्च · apacitaś ca (A 7.2.30)

अपचितश्च

apacitaśca · A 7.2.30

चायतेर्निपातोऽयं वा । अपचितः । अपचायितः ॥

Rule. apacita is an optional nipātana from the root cāy. Bhaṭṭoji: "this is a nipātana of cāy, optionally".

Example.

apa + cāy + kta → by nipātana → apacitaḥ "honoured".

In the other option → apacāyitaḥ.

Watch out. One of Bhaṭṭoji's commentators quotes the Bhāṣya differently. There the nipātana is ci for the causative of ciñ after apa.

III.J. pī for pyāy, and a short vowel for hlād — running example: `pyāy + kta → pī + na → pīnaḥ; pra + hlād + ta → prahlannaḥ`

Two short rules change the root itself before a niṣṭhā. SK 3072 gives pī for pyāy, with a fixed domain for the option. SK 3073 shortens the vowel of hlād "to be glad". The running example shows both. pyāy gives pīnaḥ. hlād after pra gives prahlannaḥ "delighted".

SK 3072 · प्यायः पी · pyāyaḥ pī (A 6.1.28)

प्यायः पी

pyāyaḥ pī · A 6.1.28

वा स्यान्निष्ठायाम् । व्यवस्थितविभाषेयम् । तेन स्वाङ्गे नित्यम् । पीनं सुखम् । अन्यत्र प्यानः पीनः स्वेदः ॥

सोपसर्गस्य न

(वार्तिकम्) ॥ प्रप्यानः ॥

आङ्पूर्वस्यान्धूधसोः स्यादेव

(वार्तिकम्) ॥ आपीनोन्धुः । आपीनमूधः ॥

Rule. pī optionally replaces pyāy before a niṣṭhā. One of Bhaṭṭoji's commentators draws "optionally" from SK 3023 and "before a niṣṭhā" from SK 3044. Bhaṭṭoji: this is a fixed-domain option (vyavasthita-vibhāṣā). The working of such an option is fixed by context: obligatory in some places, optional in others, absent in others.

Example.

pyāy + kta → this rule: pī → pī + ta

→ n for t by oditaś ca (SK 3019), since the root is opyāyī → pīnaḥ.

In the other option → no iṭ (SK 3039), the y lost → n → pyānaḥ.

One commentator gives these steps.

Applies when / blocked by. Bhaṭṭoji sets out the domains.

For a part of one's own body (svāṅga) the pī is obligatory: pīnaṃ sukham. Scholars discuss this reading, since the body-part condition suggests mukham "face".

Elsewhere it is optional: pyānaḥ / pīnaḥ svedaḥ "sweat that has increased".

Not with an upasarga: prapyānaḥ.

But after āṅ, with andhu "well" and ūdhas "udder", it does occur: āpīno 'ndhuḥ "a well that has filled", āpīnam ūdhaḥ "a swollen udder".

One commentator says these limits are clear in the Bhāṣya.

SK 3073 · ह्लादो निष्ठायाम् · hlādo niṣṭhāyām (A 6.4.95)

ह्लादो निष्ठायाम्

hlādo niṣṭhāyām · A 6.4.95

ह्रस्वः स्यात् । प्रह्लन्नः ॥

Rule. Before a niṣṭhā, a short vowel (hrasva) replaces the vowel of hlād. One of Bhaṭṭoji's commentators draws "short" from khaci hrasvaḥ (SK 2955).

Example.

pra + hlād + kta → this rule → pra + hlad + ta

→ no iṭ, since hlādī is īdit (SK 3039)

→ n for the t and for the d before it (SK 3016) → prahlannaḥ "delighted".

Both commentators give these steps.

III.K. Root changes before a kit t-initial affix: i, hi, dath, t, the long upasarga, jagdhi, long kṣī — running example: `sthā + kta → sthi + ta → sthitaḥ; dā → dath + ta → dattaḥ; pra + dā + ta → pra + t + ta → prattaḥ`

Several roots ending in a vowel change shape before a t-initial affix that is kit. The niṣṭhā is such an affix. SK 3074 sets the condition "before a kit affix beginning with t" (ti kiti). The next rules carry it. Some roots take i, dhā takes hi, and dā takes dath. After a vowel-final upasarga, dā shrinks to t. SK 3079 then lengthens the upasarga before it. SK 3080 gives jagdhi for ad, and SK 3081 an optional long ī for kṣi. The running example has three forms. sthā "to stand" gives sthitaḥ. dā "to give" gives dattaḥ. After pra it gives prattaḥ.

SK 3074 · द्यतिस्यतिमास्थामित्ति किति · dyatisyatimāsthām it ti kiti (A 7.4.40)

द्यतिस्यतिमास्थामित्ति किति

dyatisyatimāsthāmitti kiti · A 7.4.40

एषामिकारोऽन्तादेशः स्यात्तादौ किति । ईत्वदद्भावयोरपवादः । दितः । सितः । मा माङ् मेङ् । मितः । स्थितः ॥

Rule. i replaces the final of do, so, mā and sthā before a t-initial affix that is kit. One of Bhaṭṭoji's commentators splits the words as it – ti – kiti. Bhaṭṭoji: the rule is an exception (apavāda) to two others. They are ī by SK 2462 and dad by SK 3077.

Example.

do + kta (do "to cut") → this rule: i for the final → di + ta → ditaḥ.

so + kta (ṣo "to end") → sitaḥ.

mā + kta → mitaḥ.

sthā + kta → sthi + ta → sthitaḥ.

Bhaṭṭoji: mā, māṅ and meṅ are all included. One commentator names the maxim that includes them.

Watch out. do and so are named with the ending -ti, as dyati and syati (citation with śtip). One commentator says this fixes which root is meant. He rejects a view that it shuts out the yaṅluk forms. In yaṅluk these roots take iṭ, so this rule never gets a chance there.

SK 3075 · शाच्छोरन्यतरस्याम् · śācchor anyatarasyām (A 7.4.41)

शाच्छोरन्यतरस्याम्

śācchoranyatarasyām · A 7.4.41

शितः । शातः । छितः । छातः । व्यवस्थितविभाषात्वाद्व्रतविषये श्यतेर्नित्यम् । संशितं व्रतम् । सम्यक्संपादितमित्यर्थः । संशितो ब्राह्मणः । व्रतविषयकयत्नवानित्यर्थः ॥

Rule. For śā and chā, i optionally replaces the final before a t-initial kit affix. One of Bhaṭṭoji's commentators: these are śo "to sharpen" and cho "to cut", named with their ā already made.

Example.

śo + kta → śitaḥ / śātaḥ "sharpened".

cho + kta → chitaḥ / chātaḥ "cut".

Applies when / blocked by. Bhaṭṭoji: this is a fixed-domain option. So for śo the i is obligatory where a vow (vrata) is concerned.

saṃśitaṃ vratam "a vow thoroughly accomplished".

saṃśito brāhmaṇaḥ "a brāhmaṇa who exerts himself in a vow".

One commentator says the domain is clear in the Bhāṣya.

SK 3076 · दधातेर्हिः · dadhāter hiḥ (A 7.4.42)

दधातेर्हिः

dadhāterhiḥ · A 7.4.42

तादौ किति । अभिहितम् । निहितम् ॥

Rule. hi replaces dhā (dadhāti) before a t-initial kit affix. One of Bhaṭṭoji's commentators draws ti kiti from SK 3074.

Example.

abhi + dhā + kta → this rule: hi → abhi + hi + ta → abhihitam "said".

ni + dhā + kta → nihitam "deposited".

Watch out. One commentator says the name dadhāti shuts out dheṭ "to drink". He adds that the yaṅluk form gets no hi, since it takes iṭ: dādhitaḥ. For dheṭ see SK 3079.

SK 3077 · दो दद् घोः · do dad ghoḥ (A 7.4.46)

दो दद् घोः

do dad ghoḥ · A 7.4.46

घुसंज्ञकस्य दा इत्यस्य दथ् स्यात्तादौ किति । चर्त्वम् । दत्तः । घोः किम् । दातः । तान्तो वायमादेशः न चैवं विदत्तमित्यादावुपसर्गस्य दस्ति SK 3079 इति दीर्घापत्तिः । तकारादौ तद्विधानात् । दान्तो वा धान्तो वा । न च दान्तत्वे निष्ठानत्वं धान्तत्वे झषस्तथोः - SK 2280 इति धत्वं शङ्क्यम् । सन्निपातपरिभाषाविरोधात् ॥

Rule. dath replaces dā when it is ghu, before a t-initial kit affix. ghu is the name of the roots dā and dhā; dāp and daip are excluded from it. Bhaṭṭoji: "for dā that has the name ghu, dath is substituted".

Example.

dā + kta → this rule: dath → dath + ta

→ th becomes t before t (cartva, by khari ca, SK 121) → dattaḥ "given".

Bhaṭṭoji asks "why ghu?" and answers: dātaḥ. One of Bhaṭṭoji's commentators: that is dāp "to cut", which is not ghu.

Applies when / blocked by. Bhaṭṭoji then weighs other shapes for the substitute.

It could end in t. Would dasti (SK 3079) then lengthen the upasarga in vidattam? No. dasti needs a second member beginning with t.

It could end in d or in dh. Would the niṣṭhā then take n (SK 3016), or dh (SK 2280)? No. That would conflict with the sannipāta maxim. A change made because of a combination cannot cause the destruction of that combination.

Watch out. One commentator reports the earlier view: the substitute ends only in th. Bhaṭṭoji follows it in saying dath. With "or it ends in t" he shows that the other views also work.

SK 3078 · अच उपसर्गात्तः · aca upasargāt taḥ (A 7.4.47)

अच उपसर्गात्तः

aca upasargāttaḥ · A 7.4.47

अजन्तादुपसर्गात्परस्य दा इत्यस्य घोरचस्तः स्यात्तादौ किति । चर्त्वम् । प्रत्तः । अवत्तः । अवदत्तं विदत्तं च प्रदत्तं चादिकर्मणि । सुदत्तमवुदत्तं च निदत्तमिति चेष्यते ॥ चशब्दाद्यथाप्राप्तम् ॥

Rule. After an upasarga ending in a vowel, t replaces the vowel of the ghu dā, before a t-initial kit affix. Both of Bhaṭṭoji's commentators say the word acaḥ serves twice. Once it describes the upasarga ("ending in a vowel"). Once it names what is replaced ("the vowel").

Example.

pra + dā + kta → this rule: the ā becomes t → pra + dt + ta

→ cartva: the d becomes t → prattaḥ.

Likewise avattaḥ, from ava.

One commentator says this rule is an exception to dad (SK 3077).

Applies when / blocked by. Bhaṭṭoji quotes a verse; one commentator says it is in the Bhāṣya. It accepts avadattam, vidattam, and pradattam in ādikarman. It also accepts sudattam, anudattam and nidattam. Bhaṭṭoji: by ca, the form otherwise obtained also holds. One commentator explains: dad sometimes still comes after these words.

Watch out. "In ādikarman" goes only with pradattam, say both commentators. One glosses it "begun to give".

SK 3079 · दस्ति · dasti (A 6.3.124)

दस्ति

dasti · A 6.3.124

इगन्तोपसर्गस्य दीर्घः स्याद्दादेशो यस्तकारस्तदादावुत्तरपदे । खरिच SK 121 इति चर्त्वमाश्रयात्सिद्धम् । नीत्तम् । सूत्तम् । घुमास्था SK 2462 इतीत्वम् । धेट् । धीतम् । गीतम् । पीतम् । जनसन - SK 2504 इत्यात्वम् । जातम् । सातम् । खातम् ॥

Rule. The final of an upasarga ending in ik (i, u, ṛ or ḷ; here i or u) is lengthened. This happens before a second member that begins with the t that replaced dā. Both of Bhaṭṭoji's commentators list the words carried: "ik" from SK 1045, "upasarga" from SK 1044, "long" from SK 174, and "second member" from SK 958.

Example.

ni + dā + kta → t for the ā (SK 3078) → ni + dt + ta

→ cartva: d becomes t

→ this rule: ni becomes nī → nīttam.

Likewise sūttam.

These steps are one commentator's.

Applies when / blocked by. Cartva stands in the Tripādī, so it should be asiddha here. Bhaṭṭoji: it counts as done, "because of the basis" (āśrayāt siddham). This rule needs the t that cartva makes. So for this rule cartva is not asiddha.

Bhaṭṭoji then lists more niṣṭhā forms.

ī by ghumāsthā… (SK 2462): dhītam from dheṭ, gītam, pītam.

ā by janasana… (SK 2504): jātam, sātam, khātam.

Both commentators: dheṭ gets no hi (SK 3076), since that rule names dhā by its class sign ślu. ślu is the elision of the class sign śap after juhoti and the roots of its class (SK 2489); dadhāti is one of them.

Watch out. One commentator doubts sūttam. The Bhāṣya argues that su is no upasarga for SK 3078.

SK 3080 · अदो जग्धिर्ल्यप्ति किति · ado jagdhir lyap ti kiti (A 2.4.36)

अदो जग्धिर्ल्यप्ति किति

ado jagdhirlyapti kiti · A 2.4.36

ल्यबिति लुप्तसप्तमीकम् । अदो जग्धिः स्यात् ल्यपि तादौ किति च । इकार उच्चारणार्थः । धत्वम् । झरो झरि SK 71 । जग्धः । आदिकर्मणि क्तः कर्तरि च SK 3053 । प्रकृतः कटं सः । प्रकृतः कटस्तेन । निष्ठायामण्यदर्थे SK 3014 इति दीर्घः । क्षियो दीर्घात् SK 3015 इति णत्वम् । प्रक्षीणः सः ॥

Rule. jagdhi replaces ad "to eat" before lyap and before a t-initial kit affix. lyap is a kṛt affix taught in a later chapter. Bhaṭṭoji: in the rule, lyap has lost its seventh-case ending. The i of jagdhi is for pronunciation only. One of Bhaṭṭoji's commentators: were it a marker, num would come.

Example.

ad + kta → this rule → jagdh + ta

→ dh for the t (SK 2280) → jagdh + dha

→ the first dh is optionally dropped by jharo jhari (SK 71) → jagdhaḥ "eaten".

Both commentators give these steps.

Applies when / blocked by. Bhaṭṭoji then returns to SK 3053, for more examples. One commentator says so.

pra + kṛ + kta (agent) → prakṛtaḥ kaṭaṃ saḥ "he has begun to make the mat".

Both commentators: here kta names the agent, so saḥ is in the first case and the object kaṭam in the second.

pra + kṛ + kta (object) → prakṛtaḥ kaṭas tena "the mat has been begun by him".

pra + kṣi + kta (agent) → long ī (SK 3014) → n, then ṇ (SK 3015) → prakṣīṇaḥ saḥ "he has begun to waste away".

SK 3081 · वाक्रोशदैन्ययोः · vā''krośadainyayoḥ (A 6.4.61)

वाक्रोशदैन्ययोः

vākrośadainyayoḥ · A 6.4.61

क्षियो निष्ठायां दीर्घो वा स्यादाक्रोशे दैन्ये च । क्षीणायुर्भव । क्षितायुर्वा । क्षीणोऽयं तपस्वी । क्षितो वा ॥

Rule. For kṣi "to waste", a long vowel is optional before a niṣṭhā, in imprecation (ākrośa) and in wretchedness (dainya). One of Bhaṭṭoji's commentators: "not in the sense of ṇyat" carries from SK 3014. That sense is bhāva and karman.

Example.

kṣi + kta → long ī → n, then ṇ (SK 3015) → kṣīṇāyur bhava "may your life waste away"; or kṣitāyuḥ.

kṣīṇo 'yaṃ tapasvī "this ascetic is wasted"; or kṣitaḥ.

One commentator: the first shows imprecation, with kta in the agent (SK 3086). The second shows wretchedness and means "emaciated".

Watch out. SK 3015 (kṣiyo dīrghāt) works only "after long kṣī" (dīrghāt). So the short forms keep their t.

III.L. ṣ for s in niṣṇāta, pratiṣṇāta, kapiṣṭhala, viṣṭhala — running example: `ni + snā + ta → ni-ṣnā-ta → niṣṇātaḥ`

These four rules come from book 8. They replace an s with the retroflex ṣ after certain first members. In two of them the root is snā "to bathe" (mostly in the niṣṭhā). In the other two the word is sthala "place". The running example is niṣṇātaḥ "expert". The n after ṣ becomes ṇ by ṣṭutva (SK 113), which Bhaṭṭoji leaves unstated.

SK 3082 · निनदीभ्यां स्नातेः कौशले · ninadībhyāṃ snāteḥ kauśale (A 8.3.89)

निनदीभ्यां स्नातेः कौशले

ninadībhyāṃ snāteḥ kauśale · A 8.3.89

आभ्यां स्नातेः सस्य षः स्यात्कौशले गम्ये । निष्णातः । शास्त्रेषु । नद्यां स्नातीति नदीष्णः । सुपि SK 2916 इति कः ॥

Rule. After ni and nadī, ṣ replaces the s of snā, when skill (kauśala) is meant. One of Bhaṭṭoji's commentators: "s" carries from SK 335, under the heading apadāntasya mūrdhanyaḥ (SK 210).

Example.

ni + snā + kta → ni + snā + ta

→ this rule: ṣ for s → ni + ṣnā + ta

→ ṣṭutva: n becomes ṇ → niṣṇātaḥ śāstreṣu "expert in the śāstras".

nadyāṃ snāti "he bathes in a river", that is, skilfully → nadīṣṇaḥ.

Bhaṭṭoji: the affix is ka, by supi (SK 2916).

Watch out. One commentator explains "by supi". It is the rule supi sthaḥ split in two (yogavibhāga), so that supi stands as a rule of its own.

SK 3083 · सूत्रं प्रतिष्णातम् · sūtraṃ pratiṣṇātam (A 8.3.90)

सूत्रं प्रतिष्णातम्

sūtraṃ pratiṣṇātam · A 8.3.90

प्रतेः स्नातेः षत्वम् । प्रतिष्णातं सूत्रम् । शुद्धमित्यर्थः । अन्यत्र प्रतिस्नातम् ॥

Rule. After prati, ṣ replaces the s of snā when the word qualifies sūtra. Bhaṭṭoji: "ṣ for snā after prati".

Example.

prati + snā + kta → ṣ → pratiṣṇātaṃ sūtram. Bhaṭṭoji glosses it "clean" (śuddha).

Elsewhere → pratisnātam.

SK 3084 · कपिष्ठलो गोत्रे · kapiṣṭhalo gotre (A 8.3.91)

कपिष्ठलो गोत्रे

kapiṣṭhalo gotre · A 8.3.91

कपिष्ठलो नाम यस्य कापिष्ठलिः पुत्रः । गोत्रे किम् । कपीनां स्थलं कपिस्थलम् ॥

Rule. kapiṣṭhala, with ṣ, is the form when it names a gotra. A gotra is a line of descent. Bhaṭṭoji: Kapiṣṭhala is the one whose son, in the gotra sense, is Kāpiṣṭhali.

Example.

kapi + sthala (a gotra name) → ṣ, then ṭh → kapiṣṭhalaḥ.

Bhaṭṭoji: "why gotra?" kapīnāṃ sthalam "the place of the monkeys" → kapisthalam.

SK 3085 · विकुशमिपरिभ्यः स्थलम् · vikuśamiparibhyaḥ sthalam (A 8.3.96)

विकुशमिपरिभ्यः स्थलम्

vikuśamiparibhyaḥ sthalam · A 8.3.96

एभ्यः स्थलस्य सस्य षः स्यात् । विष्ठलम् । कुष्ठलम् । शमिष्ठलम् । परिष्ठलम् ॥

Rule. After vi, ku, śami and pari, ṣ replaces the s of sthala.

Example.

vi + sthala → this rule: ṣ → vi + ṣthala

→ ṣṭutva: th becomes ṭh → viṣṭhalam.

Likewise kuṣṭhalam, śamiṣṭhalam, pariṣṭhalam.

The ṭh is left unstated by Bhaṭṭoji.

III.M. Which sense the kta expresses: agent, locus, present, neuter bhāva — running example: `gam + kta (kartari) → gaṅgāṃ gataḥ; ās + kta (adhikaraṇe) → mukundasyāsitam idam`

By SK 2833 a kta denotes only bhāva and karman. It comes in the past (SK 3013). These five rules widen both. SK 3086 gives kta in the agent after certain roots. SK 3087 gives it in the sense of the locus (adhikaraṇa). SK 3088 and SK 3089 give it in the present. SK 3090 gives a neuter kta for the bare action. The running example has two forms. gam "to go" with kta in the agent gives gaṅgāṃ gataḥ "he has gone to the Gaṅgā". ās "to sit" with kta in the locus gives mukundasyāsitam idam "this is where Mukunda sat".

SK 3086 · गत्यर्थाकर्मकश्लिषशीङ्स्थाऽऽसवसजनरुहजीर्यतिभ्यश्च · gatyarthākarmakaśliṣaśīṅsthā''savasajanaruhajīryatibhyaś ca (A 3.4.72)

गत्यर्थाकर्मकश्लिषशीङ्स्थाऽऽसवसजनरुहजीर्यतिभ्यश्च

gatyarthākarmakaśliṣaśīṅsthā''savasajanaruhajīryatibhyaśca · A 3.4.72

एभ्यः कर्तरि क्तः स्यात् भावकर्मणोश्च । गङ्गां गतः । गङ्गां प्राप्तः । ग्लानः सः । लक्ष्मीमाश्लिष्टो हरिः । शेषमधिशयितः । वैकुण्ठमधिष्ठितः । शिवमुपासितः । हरिदिनमुपोषितः । राममनुजातः । गरुडमारूढः । विश्वमनुजीर्णः । पक्षे प्राप्ता गङ्गा येनेत्यादि ॥

Rule. kta comes in the agent sense after three kinds of root. They are roots meaning going, roots that take no object (akarmaka), and śliṣ, śīṅ, sthā, ās, vas, jan, ruh and jṝ. It also comes in bhāva and karman. One of Bhaṭṭoji's commentators draws "agent" from SK 3053 and "bhāva and karman" from SK 2152.

Example. Bhaṭṭoji's examples:

gam + kta → nasal lost (SK 2428) → gaṅgāṃ gataḥ "he has gone to the Gaṅgā".

pra + āp + kta → gaṅgāṃ prāptaḥ "he has reached the Gaṅgā".

glai + kta → ā (SK 2370), n (SK 3017) → glānaḥ saḥ "he is wearied".

lakṣmīm āśliṣṭo hariḥ "Hari has embraced Lakṣmī".

śeṣam adhiśayitaḥ "he has lain on Śeṣa".

vaikuṇṭham adhiṣṭhitaḥ "he has taken his stand in Vaikuṇṭha"; the i is by SK 3074.

śivam upāsitaḥ "he has attended on Śiva".

haridinam upoṣitaḥ "he has fasted on Hari's day"; the iṭ is by SK 3046.

rāmam anujātaḥ "he has followed Rāma".

garuḍam ārūḍhaḥ "he has mounted Garuḍa".

viśvam anujīrṇaḥ "he has worn away the universe".

One commentator gives the rules for gataḥ, glānaḥ, adhiṣṭhitaḥ and upoṣitaḥ.

Applies when / blocked by. In the other option the kta is in the object. Bhaṭṭoji: prāptā gaṅgā yena "he by whom the Gaṅgā has been reached".

Watch out. Why name śīṅ, sthā, ās and vas, which take no object? One commentator answers. With upasargas they do take objects. After adhi, śīṅ and sthā take one by SK 542. upa with vas takes one by SK 544. upa with ās means "to attend on". So "akarmaka" would not reach them. One commentator reads mlānaḥ for glānaḥ; scholars discuss this.

SK 3087 · क्तोऽधिकरणे च ध्रौव्यगतिप्रत्यवसानार्थेभ्यः · kto 'dhikaraṇe ca dhrauvyagatipratyavasānārthebhyaḥ (A 3.4.76)

क्तोऽधिकरणे च ध्रौव्यगतिप्रत्यवसानार्थेभ्यः

kto'dhikaraṇe ca dhrauvyagatipratyavasānārthebhyaḥ · A 3.4.76

एभ्योऽधिकरणे क्तः स्यात् चाद्याथाप्राप्तम् । ध्रौव्यं स्थैर्यम् ॥ मुकुन्दस्यासितामिदमिदं यातं रमापतेः । भुक्तमेतदनन्तस्येत्यूचुर्गोप्यो दिदृक्षवः ॥ आसेरकर्मकत्वात्कर्तरि भावे च । आसितो मुकुन्दः । आसितं तेन । गत्यर्थेभ्यः कर्तरि कर्मणि । अनन्तेनेदं भुक्तम् । कथं भुक्ता ब्राह्मणा इति । भुक्तमिति एषामिति मत्वर्थीयोऽच् । वर्तमाने इत्यधिकृत्य ॥

Rule. kta comes in the sense of the locus (adhikaraṇa) after roots meaning remaining still, going and eating. By ca, it also comes in the senses otherwise obtained. Bhaṭṭoji glosses dhrauvya "remaining still" as sthairya, steadiness.

Example. Bhaṭṭoji quotes a verse:

mukundasyāsitam idam idaṃ yātaṃ ramāpateḥ / bhuktam etad anantasyety ūcur gopyo didṛkṣavaḥ.

"'This is where Mukunda sat; this is where Ramā's lord went; this is where Ananta ate' — so said the gopīs, longing to see him."

Both of Bhaṭṭoji's commentators read the three forms as places. āsitam is the seat, yātam the path, bhuktam the place of eating. The agent stands in the sixth case, by SK 626.

Applies when / blocked by. Bhaṭṭoji then gives the other senses, root by root.

ās takes no object, so also the agent and bhāva: āsito mukundaḥ "Mukunda sat"; āsitaṃ tena "there was sitting by him".

Roots of going take the agent and the object: ramāpatir idaṃ yātaḥ "Ramā's lord has gone to this place"; tenedaṃ yātam "this has been gone to by him".

bhuj takes the object: anantenedaṃ bhuktam "this has been eaten by Ananta".

How then is bhuktā brāhmaṇāḥ "the brāhmaṇas have eaten" right? Bhaṭṭoji: bhuktam asti eṣām "they have eating". The word is bhukta with an affix ac in the sense "having it" (matvarthīya).

The rules that follow are taught "under the heading vartamāne", in the present. One commentator: the word is carried by frog's leap from vartamāne laṭ (SK 2151).

Watch out. One commentator adds that the Bhāṣya also allows bhāva after roots of going. Bhaṭṭoji names only the agent and the object; scholars discuss this.

SK 3088 · ञीतः क्तः · ñītaḥ ktaḥ (A 3.2.187)

ञीतः क्तः

ñītaḥ ktaḥ · A 3.2.187

ञिक्ष्विदा । क्ष्विण्णः । ञि इन्धी । इद्धः ॥

Rule. After a root marked with ñi (a ñīt root), kta comes in the present. One of Bhaṭṭoji's commentators: the rule is needed because kta otherwise denotes the past.

Example.

ñikṣvidā: kṣvid + kta → no iṭ, since it is ādit (SK 3036) → kṣvid + ta

→ n for the t and for the d before it (SK 3016) → ṇ after ṣ (ṇatva) → kṣviṇṇaḥ.

ñiindhī: indh + kta → no iṭ, since it is īdit (SK 3039) → nasal lost → idh + ta → dh for t (SK 2280) → d for the first dh → iddhaḥ.

Both commentators name the iṭ rules; one names the loss of the nasal. Bhaṭṭoji gives only the finished forms, and no one spells out the other steps.

Watch out. These forms look like past participles, but here they denote the present. The heading vartamāne, announced at SK 3087, gives that sense.

SK 3089 · मतिबुद्धिपूजार्थेभ्यश्च · matibuddhipūjārthebhyaś ca (A 3.2.188)

मतिबुद्धिपूजार्थेभ्यश्च

matibuddhipūjārthebhyaśca · A 3.2.188

मतिरिहेच्छा । बुद्धेः पृथगुपादानात् । राज्ञां मतः इष्टः । तैरिष्यमाण इत्यर्थः । बुद्धः । विदितः । पूजितः । अर्चितः । चकारोऽनुक्तसमुच्चयार्थः । शीलितो रक्षितः क्षान्त आकृष्टो जुष्ट इत्यादि ॥

Rule. kta also comes in the present after roots meaning desire (mati), understanding (buddhi) and honouring (pūjā). One of Bhaṭṭoji's commentators: it is in bhāva and karman, by SK 2833.

Example.

rājñāṃ mataḥ, iṣṭaḥ "desired by kings", that is, being desired by them.

buddhaḥ, viditaḥ "understood, known".

pūjitaḥ, arcitaḥ "honoured, worshipped".

One commentator: rājñām is in the sixth case by SK 625. Both: iṣṭaḥ has no iṭ by yasya vibhāṣā, since iṣ is veṭ by SK 2340.

Applies when / blocked by. Bhaṭṭoji: mati here means desire, since buddhi is named separately. The ca gathers in what is not stated: śīlitaḥ, rakṣitaḥ, kṣāntaḥ, ākruṣṭaḥ, juṣṭaḥ and so on. One commentator says this list is a verse in the Bhāṣya.

Watch out. Some texts read ākṛṣṭaḥ for ākruṣṭaḥ. One commentator's list of roots has kruś "to call"; scholars discuss this reading.

SK 3090 · नपुंसके भावे क्तः · napuṃsake bhāve ktaḥ (A 3.3.114)

नपुंसके भावे क्तः

napuṃsake bhāve ktaḥ · A 3.3.114

क्लीबत्व विशिष्टे भावे कालसामान्ये क्तः स्यात् । जल्पितम् । शयितम् । हसितम् ॥

Rule. kta comes in bhāva, in the neuter. Bhaṭṭoji adds: in time in general (kālasāmānye). One of Bhaṭṭoji's commentators: this rule stands outside the headings "in the present" and the rest.

Example.

jalp + kta → iṭ → jalpitam "talk".

śī + kta → śayitam "lying down".

has + kta → hasitam "laughter".

Watch out. One commentator, citing the Bhāṣya, limits the rule to roots that take no object. Bhaṭṭoji's examples are all of that kind. In gataṃ haṃsasya "the goose's going" the object is simply not intended.

III.N. Other past participles: ṅvanip and atṛn — running example: `su + ṅvanip → su + t + van → sutvā; jṝ + atṛn → jar + at → jaran (beside jīrṇaḥ)`

The niṣṭhā is not the only past affix. Two rules of the bhūte "in the past" heading (SK 2995) give others. SK 3091 gives ṅvanip after su and yaj. SK 3092 gives atṛn after jṝ. The running example has both. su "to press out" gives sutvā. jṝ "to grow old" gives jaran, beside the niṣṭhā jīrṇaḥ.

SK 3091 · सुयजोर्ङ्वनिप् · suyajor ṅvanip (A 3.2.103)

सुयजोर्ङ्वनिप्

suyajorṅvanip · A 3.2.103

सुनोतेर्यजेश्चङ्वनिप्स्याद्भूते । सुत्वा । सुत्वानौ । यज्वा । यज्वानौ ॥

Rule. ṅvanip comes in the past after su (sunoti, "to press out") and yaj "to sacrifice". Bhaṭṭoji: "ṅvanip after sunoti and yaj, in the past".

Example.

su + ṅvanip → the markers go → su + van

→ no guṇa, since the affix is ṅ-marked

→ the augment t (tuk) → su + t + van → sutvā, dual sutvānau.

yaj + ṅvanip → yaj + van → yajvā, dual yajvānau.

Applies when / blocked by. One of Bhaṭṭoji's commentators explains the markers. The ṅ blocks guṇa. The p serves the accent and the augment t. Why ṅ and not k? A k-marked vanip would bring saṃprasāraṇa to yaj by SK 2409.

Watch out. One commentator says sunoti is ṣuñ "to press out". Two other roots su are not taken. Either they do not express the sense (anabhidhāna), or association with yaj rules them out.

SK 3092 · जीर्यतेरतृन् · jīryater atṛn (A 3.2.104)

जीर्यतेरतृन्

jīryateratṛn · A 3.2.104

भूत इत्येव । जरन् । जरन्तौ । जरन्तः । वासरूपन्यायेन निष्ठापि । जीर्णो जीर्णवान् ॥

Rule. atṛn comes after jṝ (jīryati, "to grow old"). Bhaṭṭoji: "in the past" still holds (bhūte ity eva).

Example.

jṝ + atṛn → the markers ṛ and n go → jṝ + at

→ guṇa: ṝ becomes ar → jar + at

→ the affix is ugit, so num comes (SK 361) → jarant

→ jaran, jarantau, jarantaḥ.

One of Bhaṭṭoji's commentators: the a of jaran is not lengthened. The loss of the final t is asiddha.

Applies when / blocked by. Bhaṭṭoji: by the vāsarūpa principle the niṣṭhā comes too. That principle is SK 2830: an exception of a different form blocks the general affix only optionally.

jṝ + kta → i for ṝ (SK 2390), with r → lengthening (SK 354) → n (SK 3016), then ṇ → jīrṇaḥ.

With ktavatu → jīrṇavān.

IV.A. liṭ in the Veda and its participles kānac and kvasu; kvasu in the bhāṣā — running example: `ni + sad + liṭ (bhāṣā, past) → kvasu → (reduplication; e + abhyāsa-loss) ni-sed-vas → (ṅīp; vas → us; ṣ) niṣeduṣīm`

The last rule, SK 3092, gave the past participle jaran, still under the heading bhūte "in the past" (SK 2995). That heading runs on into this group. SK 3093 gives the lakāra liṭ for the past in the Veda. SK 3094 and SK 3095 let two participle affixes, kānac and kvasu, stand in its place. SK 3097 gives the same liṭ, always replaced by kvasu, in ordinary speech, after three roots. SK 3098 fixes three more such forms by nipātana. SK 3096, which stands between SK 3095 and SK 3097, is about the iṭ of vas; it is taken in IV.B with SK 3099. The running example is niṣeduṣīm "her who had sat down". Here sad "to sit" with ni takes liṭ, and kvasu replaces it. The reduplicated root shrinks to sed, the feminine ī is added, vas becomes us, and two s sounds become ṣ.

SK 3093 · छन्दसि लिट् · chandasi liṭ (A 3.2.105)

छन्दसि लिट्

chandasi liṭ · A 3.2.105

Rule. In the Veda (chandas), the lakāra liṭ comes in the sense of the past. The words "in the past" (bhūte) carry over from SK 2995. Bhaṭṭoji's paraphrase is "in the past in general" (bhūtasāmānye).

This sūtra has no gloss of its own. Bhaṭṭoji explains SK 3093, SK 3094 and SK 3095 together, under SK 3095. One of his commentators does the same.

Example. Bhaṭṭoji gives no Vedic example. The line tasthivāṃsam (SK 3095) is a poet's use. He cites it to show that poets use these forms widely.

This rule supplies the liṭ; SK 3094–3095 supply what stands in its place.

Applies when / blocked by. Elsewhere liṭ is taught by parokṣe liṭ (SK 2171). There it needs a past that is not of today (anadyatana) and out of the speaker's sight (parokṣa). One of Bhaṭṭoji's commentators says these conditions are not intended in the Veda. Any past will do.

The other commentator meets a doubt. Is this rule not already covered by chandasi luṅlaṅliṭaḥ (SK 3423)? No. That rule stands under a heading that needs a connection between verbs (dhātusaṃbandha). This one applies without it.

SK 3094 · लिटः कानज्वा · liṭaḥ kānaj vā (A 3.2.106)

लिटः कानज्वा

liṭaḥ kānajvā · A 3.2.106

Rule. The affix kānac optionally replaces liṭ, in the Veda, in the past. "In the Veda" comes from SK 3093, "in the past" from SK 2995. kānac is a substitute of a lakāra (lādeśa), like the verb-endings that replace a lakāra.

Example. Bhaṭṭoji gives none here. One of his commentators cites cakrāṇa:

kṛ + liṭ (Veda) → kānac replaces liṭ → markers k and c go → kṛ + āna

→ reduplication and the usual sandhi → cakrāṇa.

Applies when / blocked by. Why is the word liṭ stated again, when it already carries over? One of Bhaṭṭoji's commentators answers. So that every liṭ is taken, including that of parokṣe liṭ (SK 2171). Otherwise the maxim "the nearest" (anantarasya) would take only the liṭ of SK 3093.

The same commentator gives two uses for vā "optionally". It lets liṭ itself be heard in the other option. Or it shows that the option between unlike affixes (vāsarūpa, SK 2830) does not hold for substitutes of a lakāra. Its use comes up at SK 3097.

Watch out. What is the k of kānac for? The other commentator answers. In tistirāṇa it blocks the guṇa that ṛcchatyṝtām (SK 2383) would give. In cakrāṇa the affix already counts as k-marked, by asaṃyogāl liṭ kit (SK 2242).

SK 3095 · क्वसुश्च · kvasuś ca (A 3.2.107)

क्वसुश्च

kvasuśca · A 3.2.107

इह भूतसामान्ये छन्दसि लिट् तस्य विधीयमानौ क्वसुकानचावपि छान्दसाविति त्रिमुनिमतम् । कवयस्तु बहुलं प्रयुञ्जते । तं तस्थिवांसं नगरोपकण्ठे इति । श्रेयांसि सर्वाण्यधिजग्मुषस्ते इत्यादि ॥

Rule. The affix kvasu also replaces liṭ, in the Veda, in the past. Its markers are k and u; vas remains. Bhaṭṭoji's gloss covers SK 3093–3095. In the Veda liṭ comes in the past in general. kvasu and kānac, its substitutes, are also Vedic only. "This is the view of the three sages" (trimunimatam). "Poets, however, use them widely."

Example. Bhaṭṭoji quotes two poets' verse-quarters, to show that use.

taṃ tasthivāṃsaṃ nagaropakaṇṭhe "him who had halted near the city":

sthā + kvasu → tasthivāṃsam, second case.

One commentator explains the long vowel of -vāṃs and the num, which comes because kvasu is marked with u. He names a rule for the lengthening; scholars discuss which rule applies.

śreyāṃsi sarvāṇy adhijagmuṣas te "of you, who had attained all excellences":

adhi + gam + kvasu → reduplication → jagam + vas

→ before śas (as), vas becomes us (vasoḥ saṃprasāraṇam, SK 435) → jagam + us + as

→ before this vowel-initial affix the penultimate a is lost (SK 2363) → jagm + us + as

→ s becomes ṣ → adhijagmuṣaḥ.

These steps are one commentator's; Bhaṭṭoji gives only the forms.

Applies when / blocked by. Why is kvasu a separate sūtra and not part of SK 3094? One of Bhaṭṭoji's commentators answers. So that kvasu alone carries on into SK 3097, without kānac.

Watch out. Are kvasu and kānac correct outside the Veda? Bhaṭṭoji gives the sages' view, and adds that poets use them. One commentator supports the poets from a form used in the Bhāṣya. The other says poets simply follow one another, and rejects the claim that the forms belong to ordinary speech. Scholars discuss this.

SK 3097 · भाषायां सदवसश्रुवः · bhāṣāyāṃ sadavasaśruvaḥ (A 3.2.108)

भाषायां सदवसश्रुवः

bhāṣāyāṃ sadavasaśruvaḥ · A 3.2.108

सदादिभ्यो भूतसामान्ये भाषायां लिड्वा स्यात् तस्य च नित्यं क्वसुः । निषेदुषीमासनबन्धधीरः । अध्यूषुषस्तामभवञ्जनस्य । शुश्रुवान् ॥

Rule. In ordinary speech (bhāṣā), after sad "to sit", vas "to dwell" and śru "to hear", liṭ comes optionally in the past in general. kvasu then necessarily replaces it. liṭ carries over from SK 3093, vā from SK 3094 and kvasu from SK 3095.

Example. The running example, from Bhaṭṭoji's quotation niṣeduṣīm āsanabandhadhīraḥ:

ni + sad + liṭ → kvasu replaces liṭ (this rule) → ni + sad + vas

→ reduplication; then e for the a and loss of the reduplicated syllable (ata ekahalmadhye…) → ni-sed-vas

→ feminine ṅīp (ī), because kvasu is marked with u → ni-sed-vas-ī

→ vas becomes us (SK 435) → ni-sed-us-ī

→ s becomes ṣ, after ni (sadir aprateḥ, SK 2271) and in the affix → niṣeduṣīm.

One commentator adds: there is no iṭ, since no consonant-initial affix remains. The masculine is niṣedivān, with iṭ (SK 3096, in IV.B).

adhi + vas → saṃprasāraṇa of the v (SK 2409) → reduplication; halādiśeṣa and savarṇadīrgha

→ vas becomes us (SK 435); ṣ → adhyūṣuṣaḥ "of the people who had dwelt in it".

The commentators read the ending differently; scholars discuss this.

śru + kvasu → reduplication → śuśruvān "one who has heard".

Applies when / blocked by. The "optionally" goes with liṭ alone, says one of Bhaṭṭoji's commentators. In the other option luṅ comes, the lakāra taught for the past in general. kvasu is obligatory once liṭ comes. So this liṭ never takes a verb-ending.

Watch out. Why must "optionally" be carried over at all? Without it, this liṭ would always set luṅ aside, as its exception. The same commentator explains why SK 2830, the option between unlike affixes, cannot help. liṭ and luṅ count as of the same form. And that option does not hold for substitutes of a lakāra, as noted at SK 3094.

SK 3098 · उपेयिवाननाश्वाननूचानश्च · upeyivān anāśvān anūcānaś ca (A 3.2.109)

उपेयिवाननाश्वाननूचानश्च

upeyivānanāśvānanūcānaśca · A 3.2.109

एते निपात्यन्ते । उपपूर्वादिणो भाषायामपि भूतमात्रे लिड्वा तस्य नित्यं क्वसुः । इट् । उपेयिवान् । उपेयुषः स्वामपि मूर्तिमग्र्याम् । उपेयुषी । उपेत्यविवक्षितम् । ईयिवान् । समीयिवान् । नञ्पूर्वादश्नातेः क्वसुरिडभावश्च । धृतजयधृतेरनाशुष इति भारविः । अनुपूर्वाद्वचेः कर्तरि कानच् । वेदस्यानुवचनं कृतवाननूचानः ॥

Rule. Three forms are given by nipātana: upeyivān, anāśvān and anūcāna. Bhaṭṭoji names what each one fixes.

Example.

upa + iṇ "to go" → liṭ, even in ordinary speech, in the past as such; kvasu necessarily replaces it

→ reduplication → the augment iṭ (by the nipātana) → upeyivān "one who has approached".

Bhaṭṭoji quotes upeyuṣaḥ svām api mūrtim agryām "of one who had attained even his own foremost form". He also gives the feminine upeyuṣī.

The upa is not intended as a condition: īyivān, samīyivān.

nañ + aś (aśnāti, "eats") → kvasu, and no iṭ → anāśvān "one who has not eaten".

Bhaṭṭoji cites Bhāravi: dhṛtajayadhṛter anāśuṣaḥ "of the one who had not eaten, whose resolve was fixed on victory".

anu + vac "to speak" → kānac in the sense of the agent; saṃprasāraṇa → anūcānaḥ.

Bhaṭṭoji: one who has completed the recitation of the Veda is anūcāna.

Applies when / blocked by. Which iṭ is this? Both commentators say it is the ordinary one, before a consonant-initial ārdhadhātuka (SK 2184). One traces its history. SK 2981 had barred it, a restriction brought it back, and SK 3096 (in IV.B) barred it again. This nipātana restores that same iṭ.

Watch out. upeyuṣī has no iṭ. The v of vas becomes u, so the consonant that called for the iṭ is gone. One commentator uses this to show that no new iṭ is being taught. The other makes the same point with upeyuṣaḥ.

IV.B. The iṭ of vasu — running example: `ad + kvasu → (reduplication, halādiśeṣa, lengthening, savarṇadīrgha) ād-vas → (still ekāc) ād-i-vas → ādivān`; `gam → jagmivān / jaganvān`

The affix vas left by kvasu begins with a consonant. Does it take the augment iṭ (i)? Two rules from the seventh book answer, and Bhaṭṭoji places them among the kvasu rules. SK 3096 stands in the SK between SK 3095 and SK 3097. It is taken here, beside SK 3099, the other iṭ rule for vas. The running example is ādivān "one who has eaten", from ad. After reduplication the root is still one syllable, so iṭ comes. Beside it stand jagmivān and jaganvān from gam, where iṭ is optional.

SK 3096 · वस्वेकाजाद्घसाम् · vasv ekājādghasām (A 7.2.67)

वस्वेकाजाद्घसाम्

vasvekājādghasām · A 7.2.67

कृतद्विर्वचनानामेकाचामादन्तानां घसेश्च वसोरिट् नान्येषाम् । एकाच् । आदिवान् । आरिवान् । आत् । ददिवान् । जक्षिवान् । एषां किम् । बभूवान् ॥

Rule. The affix vasu (here the vas of kvasu) takes iṭ only after three kinds of root. First, roots that still have one vowel (ekāc) once reduplication is done. Second, roots ending in ā. Third, ghas. "Not after others," says Bhaṭṭoji.

Example. The running example:

ad + kvasu → ad + vas

→ reduplication → halādiśeṣa (only the first consonant of the reduplicated syllable stays) → lengthening of that syllable → savarṇadīrgha

→ ād + vas: still one vowel

→ iṭ (this rule) → ād-i-vas → ādivān.

ṛ "to go" → guṇa (SK 2383) and the same steps → ārivān.

An ā-final root: dā → reduplication → dadā + vas → iṭ (this rule) → the ā is lost before iṭ → dadivān.

ghas, which replaces ad by SK 2424: jakṣivān.

Bhaṭṭoji's counterexample: babhūvān, with no iṭ.

Applies when / blocked by. Reduplication is obligatory, and after it no root has one vowel. So how can iṭ ever come? One of Bhaṭṭoji's commentators answers: hence "once reduplication has been done". The rule means roots that are still ekāc even after reduplication.

The same commentator traces the iṭ. neḍ vaśi kṛti (SK 2981) bars iṭ before vas. Another restriction, the kṛādi-niyama, brings iṭ back everywhere. This sūtra then restricts (niyama) that restored iṭ to the three kinds of root.

Watch out. Why name ā-final roots separately? dadā is not ekāc. And its ā is lost only before iṭ, so it cannot first shrink to one vowel. Both commentators explain this. One of them also explains ghas at length; the root would lose its vowel before reduplication could apply.

SK 3099 · विभाषा गमहनविदविशाम् · vibhāṣā gamahanavidaviśām (A 7.2.68)

विभाषा गमहनविदविशाम्

vibhāṣā gamahanavidaviśām · A 7.2.68

एभ्यो वसोरिड्वा । जग्मिवान् । जगन्वान् । जघ्निवान् । जघन्वान् । विविदिवान् । विविद्वान् । विविशिवान् । विविश्वान् । विशिना साहचर्याद्विन्दतेर्ग्रहणम् । वेत्तेस्तु विविद्वान् । नेड्वशि कृति SK 2981 इतीण्निषेधः ॥

दृशेश्च

(वार्तिकम्) ॥ ददृशिवान् । ददृश्वान् ॥

Rule. After gam "to go", han "to strike", vid "to find" and viś "to enter", vasu takes iṭ optionally.

Example. The running example: gam + kvasu → jagmivān (with iṭ) or jaganvān (without).

One of Bhaṭṭoji's commentators: with iṭ, the penultimate a is lost by SK 2363.

han → jaghnivān / jaghanvān.

vid → vividivān / vividvān.

viś → viviśivān / viviśvān.

Applies when / blocked by. Which vid? Bhaṭṭoji says it is the vid of vindati "finds". It is named together with viś, and the association (sāhacarya) decides. From vetti "knows" there is only vividvān. neḍ vaśi kṛti (SK 2981) bars iṭ there.

A vārttika, dṛśeś ca: "and after dṛś" "to see": dadṛśivān / dadṛśvān.

IV.C. śatṛ and śānac for laṭ: the present participles — running example: `pac + laṭ → śatṛ → pac + śap + at → pacat (num) → pacantaṃ caitraṃ paśya`; `→ śānac → paca + āna → (muk) pacamānam`

The heading "in the past" is left behind. These rules replace the present lakāra laṭ, taught by vartamāne laṭ (SK 2151). SK 3100 gives śatṛ and śānac in its place, when the participle refers to the same thing as a word not in the first case. SK 3101 gives the augment m before āna. SK 3102 and SK 3103 add address and the senses "mark" and "cause". SK 3104 gives āsīnaḥ. The running example is pac "to cook". With śatṛ: pacantaṃ caitraṃ paśya "see Caitra, who is cooking". With śānac: pacamānaṃ caitraṃ paśya, the same.

SK 3100 · लटः शतृशानचावप्रथमासमानाधिकरणे · laṭaḥ śatṛśānacāv aprathamāsamānādhikaraṇe (A 3.2.124)

लटः शतृशानचावप्रथमासमानाधिकरणे

laṭaḥ śatṛśānacāvaprathamāsamānādhikaraṇe · A 3.2.124

अप्रथमान्तेन सामानाधिकरण्ये सतीत्यर्थः । शबादि । पचन्तं चैत्रं पश्य ॥

Rule. śatṛ and śānac replace laṭ. The condition is co-reference (sāmānādhikaraṇya) with a word that does not end in the first case. Co-reference means that the two words refer to the same thing.

One of Bhaṭṭoji's commentators: only the laṭ of SK 2151, by the maxim "the nearest". In śatṛ the markers are ś and ṛ, leaving at. In śānac they are ś and c, leaving āna.

Example. The running example:

pac + laṭ → śatṛ replaces laṭ (this rule) → pac + at

→ śatṛ is marked with ś, so it is sārvadhātuka (SK 2166) → the class-sign śap (a) comes → pac + a + at

→ the two a merge as the second (pararūpa) → pacat

→ num, because śatṛ has ṛ as a marker; a marker u, ṛ or ḷ (ugit) brings num → pacant- → pacantam.

pacantaṃ caitraṃ paśya "see Caitra, who is cooking". pacantam and caitram are both second case.

Bhaṭṭoji writes only "śap and the rest". Both commentators name the reason: the ś-marking makes the affix sārvadhātuka.

Applies when / blocked by. One commentator meets a doubt. laṭ is never used without a substitute. How can it be co-referent with anything? Answer: the co-reference seen in śatṛ and śānac is assumed for the laṭ they replace.

Watch out. What does pacantam mean? Both commentators gloss it as "the substratum of the activity of cooking". The participle names the person who does it.

SK 3101 · आने मुक् · āne muk (A 7.2.82)

आने मुक्

āne muk · A 7.2.82

अङ्गस्यातो मुगागमः स्यादाने परे । पचमानं चैत्रं पश्य । लडित्यनुवर्तमाने पुनर्लड्ग्रहणमधिकविधानार्थम् । तेन प्रथमासामानाधिकरण्येऽपि क्वचित् । सन् ब्राह्मणः ।

माड्याक्रोशे इति वाच्यम्

(वार्तिकम्) ॥ मा जीवन् यः परावज्ञादुःखदग्धोऽपि जीवति । माङि लुङ् SK 2219 इति प्राप्ते एतद्वचनसामर्थ्यालुक् ॥

Rule. When āna follows, the augment muk (m) comes to the a of an aṅga. "Of an a" carries over from ato yeyaḥ (SK 2212). One of Bhaṭṭoji's commentators: the k is a marker and the u is for pronunciation. An augment marked with k is added at the end.

Example. The running example:

pac + laṭ → śānac (SK 3100) → pac + āna

→ śap → paca + āna

→ muk (this rule) → paca-m-āna → pacamānaṃ caitraṃ paśya.

Applies when / blocked by. Bhaṭṭoji adds a point about SK 3100. laṭ was already carried over into it, yet it is named again. This is "to prescribe more" than the wording states (adhikavidhāna). So śatṛ and śānac sometimes come even with a first-case word:

as + śatṛ → san brāhmaṇaḥ "a brāhmaṇa who is".

One commentator gives the steps: śap is elided, and the a of as is lost by śnasor allopaḥ.

A vārttika, māṅy ākrośe: they also replace laṭ when the particle māṅ is used in reviling (ākrośa).

mā jīvan yaḥ parāvajñāduḥkhadagdho 'pi jīvati "he is not living who lives on, though burnt by the pain of others' contempt".

After māṅ, luṅ would come by māṅi luṅ (SK 2219). Bhaṭṭoji says the vārttika's statement overrides this. One commentator spells it out: laṭ, and with it śatṛ, comes instead.

Watch out. Bhaṭṭoji's commentators construe "to the a of an aṅga" in slightly different ways; scholars discuss this. The form pacamāna is the same either way; one commentator argues that the accent is not.

SK 3102 · सम्बोधने च · sambodhane ca (A 3.2.125)

सम्बोधने च

sambodhane ca · A 3.2.125

हे पचन् । हे पचमान ॥

Rule. śatṛ and śānac also replace laṭ in address (sambodhana). One of Bhaṭṭoji's commentators says "śatṛ and śānac" is to be supplied.

Example.

pac + śatṛ → pacat → vocative → he pacan "O you who are cooking!"

pac + śānac → pacamāna → he pacamāna.

Watch out. A vocative is a first-case form. The same commentator notes that the restated laṭ of SK 3100 (Bhaṭṭoji's remark under SK 3101) already allows the first case. So this rule only spells the point out (prapañcārtha).

SK 3103 · लक्षणहेत्वोः क्रियायाः · lakṣaṇahetvoḥ kriyāyāḥ (A 3.2.126)

लक्षणहेत्वोः क्रियायाः

lakṣaṇahetvoḥ kriyāyāḥ · A 3.2.126

क्रियायाः परिचायके हेतौ चार्थे वर्तमानाद्धातोर्लटः शतृशानचौ स्तः । शयाना भुञ्जते यवनाः । अर्जयन्वसति । हरिं पश्यन्मुच्यते । हेतुः फलं कारणं च । कृत्यचः SK 2835 । प्रपीयमाणः सोमः ॥

Rule. śatṛ and śānac replace laṭ when the root's meaning is a mark (lakṣaṇa) or a cause (hetu) of another action. Bhaṭṭoji glosses the mark as "what makes an action known". "Cause" covers both the result aimed at (phala) and the producing cause (kāraṇa).

Example. Bhaṭṭoji gives three.

A mark: śī + śānac → śayānā bhuñjate yavanāḥ "the Yavanas eat lying down".

A cause, as result: arj + ṇic + śatṛ → arjayan vasati "he dwells there in order to earn".

A cause, as producer: dṛś + śatṛ → hariṃ paśyan mucyate "seeing Hari, one is released".

Applies when / blocked by. In śayānāḥ, what does lying down mark? One commentator says it marks the eating. The other says it shows that the eaters are Yavanas. Scholars discuss this.

How is earning a cause of dwelling, when dwelling produces it? One commentator answers. Knowing that dwelling leads to what one wants is what makes one dwell. So the wanted earning is also a cause.

Watch out. Bhaṭṭoji adds a form with ṇ: prapīyamāṇaḥ somaḥ "soma being drunk". pra and āna are not in one pada. The ṇ comes by kṛtyacaḥ (SK 2835), which gives ṇ to the n of a kṛt after an upasarga with r.

SK 3104 · ईदासः · īdāsaḥ (A 7.2.83)

ईदासः

īdāsaḥ · A 7.2.83

आसः परस्यानस्य ईत्स्यात् । आदेः परस्य SK 44 । आसीनः ॥

Rule. After ās "to sit", ī replaces the āna of SK 3101. By ādeḥ parasya (SK 44) it replaces only the first sound of āna.

Example.

ās + śānac → ās + āna

→ ī for the first sound of āna (this rule, with SK 44) → ās + īna

→ āsīnaḥ "seated".

Applies when / blocked by. One of Bhaṭṭoji's commentators: āsaḥ is a fifth case, "after ās". āna carries over from SK 3101.

IV.D. vasu for vid's śatṛ, the name sat, and sat for lṛṭ — running example: `kṛ + lṛṭ → sya + śatṛ / śānac → kariṣya-t / kariṣya-m-āṇa → kariṣyantaṃ kariṣyamāṇaṃ paśya`; `vid + śatṛ → vasu → vidvān`

Three short rules finish the participles. SK 3105 lets vasu replace the śatṛ of vid "to know": vidvān. SK 3106 gives śatṛ and śānac the joint name sat. SK 3107 then lets sat replace the future lakāra lṛṭ. The running example is kṛ "to do" in the future. With śatṛ and śānac it gives kariṣyantaṃ kariṣyamāṇaṃ paśya "see him who is going to act".

SK 3105 · विदेः शतुर्वसुः · videḥ śatur vasuḥ (A 7.1.36)

विदेः शतुर्वसुः

videḥ śaturvasuḥ · A 7.1.36

वेत्तेः परस्य शतुर्वसुरादेशो वा स्यात् । विदन् । विद्वान् । विदुषी ॥

Rule. After vid, the vid of vetti "knows", vasu optionally replaces śatṛ. "Optionally" (anyatarasyām) carries over from SK 2197, says one of Bhaṭṭoji's commentators.

Example.

vid + śatṛ → no substitute → vidan.

vid + śatṛ → vasu replaces śatṛ (this rule) → vid + vas → vidvān.

Feminine: vid + vas → ṅīp, because vasu is marked with u → vas becomes us (SK 435) → ṣ → viduṣī.

The feminine steps are the commentator's; Bhaṭṭoji gives the forms.

Applies when / blocked by. Why only the vid of vetti? The same commentator answers. śatṛ is a parasmaipada affix, and two other roots vid are ātmanepadin. The vid of vindati has a marker, and a root named bare does not take in one named with a marker.

Watch out. vasu would be ugit anyway, as a substitute of śatṛ, whose marker is ṛ (it is treated like the original, sthānivat). The other commentator says the explicit u has a purpose. It lets vasoḥ saṃprasāraṇam (SK 435) cover kvasu as well.

SK 3106 · तौ सत् · tau sat (A 3.2.127)

तौ सत्

tau sat · A 3.2.127

तौ शतृशानचौ सत्संज्ञौ स्तः ॥

Rule. "Those two", śatṛ and śānac, have the name (saṃjñā) sat. One of Bhaṭṭoji's commentators: tau points back to SK 3100.

Example.

pacat (with śatṛ) and pacamāna (with śānac): both affixes are called sat.

The name is used at once in SK 3107: kariṣyat, kariṣyamāṇa.

SK 3107 · लृटः सद् वा · lṛṭaḥ sad vā (A 3.3.14)

लृटः सद् वा

lṛṭaḥ sad vā · A 3.3.14

व्यवस्थितविभाषेयम् । तेनाप्रथमासामानाधिकरण्ये प्रत्ययोत्तरपदयोः संबोधने लक्षणहेत्वोश्च नित्यम् । करिष्यन्तं करिष्यमाणं पश्य । करिष्यतोऽपत्यं कारिष्यतः । करिष्यद्भक्तिः । हे करिष्यन् । अर्जयिष्यन्वसति । प्रथमासमानाधिकरणेऽपि क्वचित् । करिष्यतीति करिष्यन् ॥

Rule. sat — that is, śatṛ and śānac — optionally replaces lṛṭ, the lakāra of the future. Bhaṭṭoji: this is a fixed option (vyavasthita-vibhāṣā). The option is not free. It is decided by the context.

Example. The running example:

kṛ + lṛṭ → the future sign sya, with iṭ and guṇa → kariṣya

→ śatṛ for lṛṭ (this rule) → kariṣya-t → num, since śatṛ is ugit → kariṣyantam

→ or śānac → muk (SK 3101) → kariṣya-m-āṇa → kariṣyamāṇam.

kariṣyantaṃ kariṣyamāṇaṃ paśya "see him who is going to act".

Applies when / blocked by. Bhaṭṭoji names where sat is obligatory (nitya):

with a word not in the first case: kariṣyantaṃ kariṣyamāṇaṃ paśya;
before an affix: kariṣyato 'patyaṃ kāriṣyataḥ "the offspring of one who is going to act";
before a later member of a compound (uttarapada): kariṣyadbhaktiḥ;
in address: he kariṣyan;
in the senses mark and cause: arjayiṣyan vasati "he lives here to earn".

Sometimes it comes even with a first-case word: kariṣyati iti kariṣyan.

One of Bhaṭṭoji's commentators explains "obligatory". In these contexts the verb-endings are excluded. He adds that only the lṛṭ of lṛṭ śeṣe ca (SK 2193) is meant.

Watch out. The commentators analyse the compound kariṣyadbhaktiḥ differently; scholars discuss this. Bhaṭṭoji's point is only that sat stands before the later member.

IV.E. Independent present affixes: śānan, cānaś, and śatṛ in restricted senses — running example: `yaj + śānan → yaja + āna → (muk) yajamānaḥ`; `adhi + iṅ + śatṛ → adhīyan`

The affixes of this group also mean the present, but most of them are not substitutes of laṭ. Both commentators call śānan and cānaś independent affixes (one says "like ṇvul"), not substitutes of laṭ; one adds the śatṛ of SK 3110. So they come only in the agent sense, by kartari kṛt (SK 2832). SK 3108 gives śānan after pū and yaj. SK 3109 gives cānaś for disposition, age and capacity. SK 3110–3113 give śatṛ after iṅ, dhāri, dviṣ, su and arh, each in a narrow sense. The running example is yajamānaḥ "the sacrificer", from yaj with śānan, and adhīyan "one who studies with ease", from adhi + iṅ with śatṛ.

SK 3108 · पूङ्यजोः शानन् · pūṅyajoḥ śānan (A 3.2.128)

पूङ्यजोः शानन्

pūṅyajoḥ śānan · A 3.2.128

वर्तमाने । पवमानः । यजमानः ॥

Rule. After pūṅ "to purify" and yaj "to sacrifice", the affix śānan comes, in the present. Bhaṭṭoji supplies only "in the present" (vartamāne). One of his commentators: the word comes from vartamāne laṭ (SK 2151), but "laṭ" does not come with it.

Example. The running example:

yaj + śānan → markers ś and n go → yaj + āna

→ ś-marked, so sārvadhātuka → śap → yaja + āna

→ muk (SK 3101) → yaja-m-āna → yajamānaḥ "sacrificing, the sacrificer".

pū + śānan → śap, guṇa, av → pava + āna → muk → pavamānaḥ "purifying".

Applies when / blocked by. Why is śānan not just a substitute of laṭ, like śānac? One of Bhaṭṭoji's commentators answers. As an independent affix it comes only in the agent, not in the bare action or the object. The other adds two points. As a substitute it would also need a "vā", so that pavate, yajate could stand beside it. And it would come in the bare action and the object, by SK 2152.

Watch out. The other commentator names the markers. The ś gives sārvadhātuka status; the n is for the accent.

Both commentators meet one more objection. If śānan is not a substitute of laṭ, would its object take the sixth case, as beside other kṛts? No. In somaṃ pavamānaḥ the object stays in the second case. In na lokāvyaya… (SK 627), "tṛn" is a pratyāhāra that takes in this affix (see SK 3115).

SK 3109 · ताच्छील्यवयोवचनशक्तिषु चानश् · tācchīlyavayovacanaśaktiṣu cānaś (A 3.2.129)

ताच्छील्यवयोवचनशक्तिषु चानश्

tācchīlyavayovacanaśaktiṣu cānaś · A 3.2.129

एषु द्योत्येषु कर्तरि चानश् । भोगं भुञ्जानः । कवचं बिभ्राणं । शत्रून्निघ्नानः ॥

Rule. The affix cānaś comes in the agent when one of three things is conveyed. They are habitual disposition (tācchīlya), age (vayas) and capacity (śakti). One of Bhaṭṭoji's commentators: the word vacana in the sūtra is only for clarity. The markers are c and ś, leaving āna.

Example. Bhaṭṭoji gives one for each sense.

Disposition: bhuj + cānaś → bhogaṃ bhuñjānaḥ "given to enjoying pleasures".

Age: bhṛ + cānaś → kavacaṃ bibhrāṇaḥ "wearing armour", that is, able to by force of youth.

Capacity: han + cānaś → the penultimate a is lost (SK 2363) → śatrūn nighnānaḥ "able to strike down enemies".

Applies when / blocked by. cānaś is not a substitute of laṭ. So it does not count as an ātmanepada ending. That is why it follows han, a parasmaipada root. Both commentators make this point.

Watch out. There is no muk in bhuñjānaḥ. SK 3101 gives muk only to an a before āna. bhuñj has no such a. One commentator explains this.

SK 3110 · इङ्धार्योः शत्रकृच्छ्रिणि · iṅdhāryoḥ śatr akṛcchriṇi (A 3.2.130)

इङ्धार्योः शत्रकृच्छ्रिणि

iṅdhāryoḥ śatrakṛcchriṇi · A 3.2.130

आभ्यां शतृ स्यादकृच्छ्रिणि कर्तरि । अधीयन् । धारयन् । अकृच्छ्रिणि किम् । कृच्छ्रेणाधीते । धारयति ॥

Rule. After iṅ "to study" (with adhi) and dhāri "to hold" (the causative of dhṛ), śatṛ comes. The agent must be one for whom the action is without difficulty (akṛcchrin). One of Bhaṭṭoji's commentators: akṛcchra is the absence of pain, and akṛcchrin the one who has it.

Example. The running example:

adhi + iṅ + śatṛ → adhīyan "one who studies with ease".

dhāri + śatṛ → śap → dhāraya + at → dhārayan "one who holds with ease".

Applies when / blocked by. Why "without difficulty"? Bhaṭṭoji: kṛcchreṇādhīte "he studies with difficulty"; dhārayati, "he holds", with difficulty. The finite verb comes, not śatṛ.

Watch out. Why is a rule needed? iṅ is ātmanepadin, so śatṛ could not come to it by SK 3100. One commentator gives this reason. He adds that this śatṛ too is an independent affix.

SK 3111 · द्विषोऽमित्रे · dviṣo 'mitre (A 3.2.131)

द्विषोऽमित्रे

dviṣo'mitre · A 3.2.131

द्विषन् शत्रुः ॥

Rule. After dviṣ "to hate", śatṛ comes when the agent is an enemy (amitra). One of Bhaṭṭoji's commentators supplies śatṛ and "the agent".

Example.

dviṣ + śatṛ → dviṣ + at → dviṣan "an enemy" (śatru).

SK 3112 · सुञो यज्ञसंयोगे · suño yajñasaṃyoge (A 3.2.132)

सुञो यज्ञसंयोगे

suño yajñasaṃyoge · A 3.2.132

सर्वे सुन्वन्तः सर्वे यजमानाः सत्रिणः ॥

Rule. After suñ "to press" (sunoti), śatṛ comes when the pressing is connected with a sacrifice (yajña).

Example.

su + śatṛ → sunvantaḥ.

Bhaṭṭoji: sarve sunvantaḥ sarve yajamānāḥ satriṇaḥ "the performers of a sattra are all pressers, all sacrificers".

sunvantaḥ has śatṛ by this rule; yajamānāḥ has śānan by SK 3108.

Applies when / blocked by. Why "connection" (saṃyoga) with a sacrifice, and not just "sacrifice"? Both commentators answer. The word points to the master of the sacrifice (yajamāna). So the affix does not come for the officiating priests.

Watch out. Why the plural? A rite with one master has one presser: sunvan. Both commentators explain the plural. One says it fits only a sattra. The other: in a sattra the many masters do the priests' pressing themselves.

SK 3113 · अर्हः प्रशंसायाम् · arhaḥ praśaṃsāyām (A 3.2.133)

अर्हः प्रशंसायाम्

arhaḥ praśaṃsāyām · A 3.2.133

अर्हन् ॥

Rule. After arh "to deserve", śatṛ comes when praise (praśaṃsā) is meant. One of Bhaṭṭoji's commentators supplies "the affix śatṛ".

Example.

arh + śatṛ → śap → arha + at → arhan "worthy of honour".

The same commentator: only the praiseworthy is fit for honour, so praise is understood.

Applies when / blocked by. Without praise, the finite verb. The other commentator gives arhati cauro vadham "the thief deserves death".

IV.F. The tācchīlika heading; tṛn, iṣṇuc, gsnu, knu — running example: `kṛ + tṛn → (guṇa) kar-tṛ → kartā (kaṭān)`; `alam + kṛ + iṣṇuc → alaṃkariṣṇuḥ`; `ji + gsnu → (g-it: no guṇa) jiṣṇuḥ`

SK 3114 opens a long heading. Up to the kvip of SK 3157, every affix comes in an agent whose habit, duty or skill is the action. These are the affixes of habit (tācchīlika). SK 3115 gives the general one, tṛn. SK 3116–3118 give iṣṇuc, SK 3119 gsnu and SK 3120 knu. The running example has three forms. kartā kaṭān is "a habitual maker of mats", from kṛ with tṛn. alaṃkariṣṇuḥ is "one given to adorning", with iṣṇuc. jiṣṇuḥ is "one given to conquering", from ji with gsnu; the g blocks guṇa.

SK 3114 · आक्वेस्तच्छीलतद्धर्मतत्साधुकारिषु · ā kves tacchīlataddharmatatsādhukāriṣu (A 3.2.134)

आक्वेस्तच्छीलतद्धर्मतत्साधुकारिषु

ākvestacchīlataddharmatatsādhukāriṣu · A 3.2.134

क्विपमभिव्याप्य वक्ष्यमाणाः प्रत्ययास्तच्छीलतद्धर्मतत्साधुकारिषु कर्तृषु बोध्याः ॥

Rule. This is a heading (adhikāra). The affixes taught from here, up to and including kvip, come in three kinds of agent. One whose habit (śīla) is that action (tacchīla). One whose duty (dharma) it is (taddharman). One who does it well (tatsādhukārin). The kvip is that of SK 3157.

Example.

kṛ + tṛn (SK 3115) → kartā kaṭān "one whose habit is making mats".

The affix comes only because the agent has the action as his habit.

Applies when / blocked by. One of Bhaṭṭoji's commentators defines the three. The tacchīla acts by nature, with no regard to a result. The taddharman acts in the thought "this is my duty". The tatsādhukārin does the action well. The other commentator: each "tad" "that" means the root's own action.

Watch out. The ā of ā kveḥ includes the limit (abhividhi), both commentators say. So kvip itself is inside the heading. One of them adds a point from the Bhāṣya. In this range the affixes generally do not come after a root with an upasarga. The sūtras below name the exceptions.

SK 3115 · तृन् · tṛn (A 3.2.135)

तृन्

tṛn · A 3.2.135

कर्ता कटान् ॥

Rule. The affix tṛn comes after a root, in the agents of SK 3114. Its marker is n; tṛ remains.

Example. The running example:

kṛ + tṛn → kṛ + tṛ

→ tṛ is ārdhadhātuka; guṇa (SK 2168) → kar + tṛ → kartṛ

→ first case → kartā.

Bhaṭṭoji: kartā kaṭān "a habitual maker of mats".

Applies when / blocked by. Why is kaṭān in the second case and not the sixth? Beside a kṛt, SK 623 would give the object the sixth case. But na lokāvyaya… (SK 627) bars it after "tṛn". Both commentators explain. There "tṛn" is a pratyāhāra, from the t of śatṛ (SK 3100) to the n of this tṛn. So it covers this affix too.

Watch out. One commentator gives examples of the other two senses. Duty: muṇḍayitāraḥ śrāviṣṭhāyanāḥ "the Śrāviṣṭhāyanas shave [the bride]", their own custom. Skill: kartā kaṭam, one who makes a mat well.

SK 3116 · अलंकृञ्निराकृञ्प्रजनोत्पचोत्पतोन्मदरुच्यपत्रपवृतुवृधुसहचर इष्णुच् · alaṃkṛñnirākṛñprajanotpacotpatonmadarucyapatrapavṛtuvṛdhusahacara iṣṇuc (A 3.2.136)

अलंकृञ्निराकृञ्प्रजनोत्पचोत्पतोन्मदरुच्यपत्रपवृतुवृधुसहचर इष्णुच्

alaṃkṛñnirākṛñprajanotpacotpatonmadarucyapatrapavṛtuvṛdhusahacara iṣṇuc · A 3.2.136

अलंकरिष्णुः । निराकरिष्णुः । प्रजनिष्णुः । उत्पतिष्णु । उत्पचिष्णुः । उन्मदिष्णुः । रोचिष्णुः । अपत्रपिष्णुः । वर्तिष्णुः । वर्धिष्णुः । सहिष्णुः । चरिष्णुः ॥

Rule. After twelve bases, the affix iṣṇuc comes, in the agents of SK 3114. They are alaṃkṛñ, nirākṛñ, prajan, utpac, utpat, unmad, ruc, apatrap, vṛt, vṛdh, sah and car. Its marker is c; iṣṇu remains.

Example. The running example:

alam + kṛ + iṣṇuc → alam + kṛ + iṣṇu

→ guṇa (SK 2168) → alam + kar + iṣṇu → alaṃkariṣṇuḥ "one given to adorning".

Bhaṭṭoji lists all twelve: alaṃkariṣṇuḥ, nirākariṣṇuḥ, prajaniṣṇuḥ, utpatiṣṇuḥ, utpaciṣṇuḥ, unmadiṣṇuḥ, rociṣṇuḥ, apatrapiṣṇuḥ, vartiṣṇuḥ, vardhiṣṇuḥ, sahiṣṇuḥ, cariṣṇuḥ.

Applies when / blocked by. One of Bhaṭṭoji's commentators names the roots. kṛñ is only ḍukṛñ "to do", the better known root. prajan is jan "to be born" with pra. apatrap is trap "to be ashamed" with apa. He reports a view on ut. It is read with pac, pat and mad one by one. That excludes other upasargas.

Watch out. Several of these bases carry upasargas, against the general bar noted at SK 3114. The sūtra names them, so the affix comes.

SK 3117 · णेश्छन्दसि · ṇeś chandasi (A 3.2.137)

णेश्छन्दसि

ṇeśchandasi · A 3.2.137

वीरुधः पारयिष्णवः ॥

Rule. In the Veda, iṣṇuc comes after a root ending in ṇi, the causative affix. iṣṇuc carries over from SK 3116.

Example.

pṛ + ṇic → pāri → pāri + iṣṇu

→ ay replaces ṇi (ayāmantālvāyyetnviṣṇuṣu, SK 2311) → pāray + iṣṇu

→ pārayiṣṇavaḥ.

Bhaṭṭoji: vīrudhaḥ pārayiṣṇavaḥ "plants that habitually bring one across".

Applies when / blocked by. One of Bhaṭṭoji's commentators: SK 2311 sets aside the loss of ṇi (SK 2313), which would otherwise apply.

SK 3118 · भुवश्च · bhuvaś ca (A 3.2.138)

भुवश्च

bhuvaśca · A 3.2.138

छन्दसीत्येव । भविष्णुः । कथं तर्हि जगत्प्रभोरप्रभविष्णुः वैष्णवमिति । निरङ्कुशाः कवयः । चकारोऽनुक्तसमुच्चयार्थः । भ्राजिष्णुरिति वृत्तिः । एवं क्षयिष्णुः । नैतद्भाष्ये दृष्टम् ॥

Rule. In the Veda, iṣṇuc also comes after bhū "to become". Bhaṭṭoji: "in the Veda" still holds. One of his commentators: the rule is for bhū that is not causative.

Example.

bhū + iṣṇuc → guṇa → bho + iṣṇu → av (SK 61) → bhaviṣṇuḥ.

Applies when / blocked by. How then is jagatprabhor aprabhaviṣṇu vaiṣṇavam correct, outside the Veda? Bhaṭṭoji: "poets are unrestrained" (niraṅkuśāḥ kavayaḥ).

Watch out. The Vṛtti takes the ca "and" to gather in forms not stated: bhrājiṣṇuḥ, and likewise kṣayiṣṇuḥ. Bhaṭṭoji: "this is not seen in the Bhāṣya." One commentator says "this" is the Vṛtti's reading of ca.

SK 3119 · ग्लाजिस्थश्च ग्स्नुः · glājisthaś ca gsnuḥ (A 3.2.139)

ग्लाजिस्थश्च ग्स्नुः

glājisthaśca gsnuḥ · A 3.2.139

छन्दसीति निवृत्तम् । गिदयं न तु कित् । तेन स्थः ईत्वं न । ग्लास्नुः । गित्त्वान्न गुणः । जिष्णुः । स्थास्नुः । चाद्भुवः । श्र्युकः किति SK 2381 इत्यत्र गकारप्रश्लेषान्नेट् । भूष्णुः ॥

दंशेश्छन्दस्युपसंख्यानम्

(वार्तिकम्) ॥ दङ्क्ष्णवः पशवः ॥

Rule. After glā "to be weary", ji "to conquer" and sthā "to stand", the affix gsnu comes. By the ca it also comes after bhū, drawn in from SK 3118. Bhaṭṭoji: "in the Veda" has ceased. Its marker is g; snu remains.

Example. The running example:

ji + gsnu → ji + snu

→ the affix is g-marked: no guṇa → the s becomes ṣ, the n becomes ṇ

→ jiṣṇuḥ "one given to conquering".

glā + gsnu → glāsnuḥ.

sthā + gsnu → sthāsnuḥ, with no ī.

bhū + gsnu → no iṭ → bhūṣṇuḥ.

Applies when / blocked by. Bhaṭṭoji explains each form.

The affix is g-marked, not k-marked. So sthā does not get ī by ghumāsthā… (SK 2462). That rule works only before a k- or ṅ-marked affix.
Why no guṇa in jiṣṇuḥ? Both commentators: a g is read into kṅiti ca (SK 2217) by coalescence (praśleṣa). It then bars guṇa before a g-marked affix too.
Why no iṭ in bhūṣṇuḥ? A g is read in the same way into śryukaḥ kiti (SK 2381). That rule denies iṭ before a k-marked affix.

A vārttika, daṃśeś chandasy upasaṃkhyānam: "in the Veda, daṃś 'to bite' should be added".

daṃś + gsnu → daṅkṣṇavaḥ paśavaḥ "beasts given to biting".

One commentator gives the steps: ṣ, then k, then ṣ for the affix's s (SK 294, SK 295, SK 212). The nasal is not lost by SK 415, since gsnu is not k- or ṅ-marked.

Watch out. One of Bhaṭṭoji's commentators reads the affix as ksnu, with the g written as k. The sense is the same.

SK 3120 · त्रसिगृधिधृषिक्षिपेः क्नुः · trasigṛdhidhṛṣikṣipeḥ knuḥ (A 3.2.140)

त्रसिगृधिधृषिक्षिपेः क्नुः

trasigṛdhidhṛṣikṣipeḥ knuḥ · A 3.2.140

त्रस्नुः । गृध्नुः । धृष्णुः । क्षिप्नुः ॥

Rule. After tras "to tremble", gṛdh "to covet", dhṛṣ "to dare" and kṣip "to throw", the affix knu comes, in the agents of SK 3114. Its marker is k; nu remains.

Example.

tras + knu → no iṭ (neḍ vaśi kṛti, SK 2981) → trasnuḥ.

gṛdh + knu → k-marked: no guṇa (SK 2217) → gṛdhnuḥ.

dhṛṣ + knu → no guṇa → dhṛṣṇuḥ.

kṣip + knu → kṣipnuḥ.

The reasons are one commentator's; the other gives the iṭ point only. Bhaṭṭoji gives only the forms.

IV.G. ghinuṇ and vuñ — running example: `śam + ghinuṇ → śam-in → (upadhā-vṛddhi barred, SK 2763) śamī`; `nind + vuñ → nind-aka → nindakaḥ`

Two more affixes of habit follow. ghinuṇ (SK 3121–3125) leaves in. Its ṇ calls for vṛddhi and its gh for the velar substitute. vuñ (SK 3126–3127) leaves vu, which becomes aka. SK 3126 also gives the first of three indications about the option between unlike affixes in this heading. The running example: śamī "one given to calm", from śam with ghinuṇ, with no vṛddhi. And nindakaḥ "one given to blaming", from nind with vuñ.

SK 3121 · शमित्यष्टाभ्यो घिनुण् · śamity aṣṭābhyo ghinuṇ (A 3.2.141)

शमित्यष्टाभ्यो घिनुण्

śamityaṣṭābhyo ghinuṇ · A 3.2.141

उकार उच्चारणार्थ इति काशिका । अनुबन्ध इति भाष्यम् । तेन शमिनितरा शमिनीतरेत्यत्र गितश्च SK 987 इति ह्रस्वविकल्पः । न चैवं शमिनावित्यादौ नुम्प्रसङ्गः । झल्ग्रहणमपकृष्य झन्तानामेव तद्विधानात् । नोदात्तोपदेशस्य - SK 2763 इति वृद्धिनिषेधः । शमी । तमी । दमी । श्रमी । भ्रमी । क्षमी । क्लमी । प्रमादी । उत्पूर्वान्मदेः अलंकृञादिसूत्रेणेष्णुजुक्तो वासरूपविधिना घिनुणपि । उन्मादी । ताच्छीलिकेषु वासरूपविधिर्नास्तीति तु प्रायिकम् ॥

Rule. After the eight roots śam and the rest, the affix ghinuṇ comes, in the agents of SK 3114. Both commentators: iti here means "and the rest" (ādi). The roots stand in the div class. The markers gh and ṇ go, and in remains.

Example. The running example:

śam + ghinuṇ → śam + in

→ the ṇ would bring vṛddhi of the penultimate a (SK 2282)

→ barred by nodāttopadeśasya… (SK 2763) → śamin → śamī.

Likewise tamī, damī, śramī, bhramī, kṣamī, klamī.

But pra + mad → vṛddhi → pramādī. One commentator: mad is not m-final, so SK 2763 does not reach it.

Applies when / blocked by. Is the u of ghinuṇ a marker? The Kāśikā says it is for pronunciation only. The Bhāṣya says it is a marker. Bhaṭṭoji draws a consequence from the Bhāṣya's view. In śaminitarā / śaminītarā, the shortening is optional, by ugitaś ca (SK 987). The feminine śaminī may shorten its ī before the comparative affix tara.

Bhaṭṭoji meets an objection. If the affix is marked with u, would śamī, śaminau take num? No. The word jhal is drawn back (apakarṣa) from napuṃsakasya jhalacaḥ (SK 314). So that num is taught only for stems ending in a jhal consonant. śamin ends in n and is not such a stem.

Watch out. ut + mad takes iṣṇuc by SK 3116: unmadiṣṇuḥ. Yet ghinuṇ also comes, by the option between unlike affixes: unmādī. Bhaṭṭoji: the maxim that this option does not hold among the affixes of habit is only generally valid (prāyika). One commentator traces the maxim to SK 3126 and its limit to SK 3133.

On the optional shortening, one commentator reports a different view; scholars discuss this.

SK 3122 · संपृचानुरुधाङ्यमाङ्यसपरिसृसंसृजपरिदेविसंज्वरपरिक्षिपपरिरटपरिवदपरिदहपरिमुहदुषद्विषद्रुहदुहयुजाक्रीडविविचत्यजरजभजातिचरापचरामुषाभ्याहनश्च · saṃpṛcānurudhāṅyamāṅyasaparisṛsaṃsṛjaparidevisaṃjvaraparikṣipapariraṭaparivadaparidahaparimuhaduṣadviṣadruhaduhayujākrīḍavivicatyajarajabhajāticarāpacarāmuṣābhyāhanaś ca (A 3.2.142)

संपृचानुरुधाङ्यमाङ्यसपरिसृसंसृजपरिदेविसंज्वरपरिक्षिपपरिरटपरिवदपरिदहपरिमुहदुषद्विषद्रुहदुहयुजाक्रीडविविचत्यजरजभजातिचरापचरामुषाभ्याहनश्च

saṃpṛcānurudhāṅyamāṅyasaparisṛsaṃsṛjaparidevisaṃjvaraparikṣipapariraṭaparivadaparidahaparimuhaduṣadviṣadruhaduhayujākrīḍavivicatyajarajabhajāticarāpacarāmuṣābhyāhanaśca · A 3.2.142

घिनुण् स्यात् । संपर्की । अनुरोधी । आयामी । आयासी । परिसारी । संसर्गी । परिदेवी । संज्वारी । परिक्षेपी । परिराटी । परिवादी । परिदाही । परिमोही । दोषी । द्वेषी । द्रोही । दोही । योगी । आक्रीडी । विवेकी । त्यागी । रागी । भागी । अतिचारी । अपचारी । आमोषी । अभ्यघाती ॥

Rule. ghinuṇ also comes after twenty-seven bases, most of them with an upasarga. They are saṃpṛc, anurudh, āyam, āyas, parisṛ, saṃsṛj, paridevi, saṃjvar, parikṣip, pariraṭ, parivad, paridah, parimuh, duṣ, dviṣ, druh, duh, yuj, ākrīḍ, vivic, tyaj, raj, bhaj, aticar, apacar, āmuṣ and abhyāhan.

Example.

sam + pṛc + ghinuṇ → sam + pṛc + in

→ guṇa of the light penultimate (SK 2189) → sam + parc + in

→ the gh-marked affix brings k for c (cajoḥ ku ghiṇṇyatoḥ, SK 2863) → saṃparkin → saṃparkī.

abhi + ā + han + ghinuṇ → vṛddhi (SK 2282) → t for n (SK 2574) → gh for h (SK 358) → abhyāghātī.

The rest, as Bhaṭṭoji lists them: anurodhī, āyāmī, āyāsī, parisārī, saṃsargī, paridevī, saṃjvārī, parikṣepī, parirāṭī, parivādī, paridāhī, parimohī, doṣī, dveṣī, drohī, dohī, yogī, ākrīḍī, vivekī, tyāgī, rāgī, bhāgī, aticārī, apacārī, āmoṣī.

Applies when / blocked by. One of Bhaṭṭoji's commentators sorts the forms. Roots with a light penultimate i, u or ṛ take guṇa: anurodhī. Roots with penultimate a take vṛddhi: tyāgī.

Watch out. paridevi is the bhū-class root devṛ "to lament", not the causative of dīv. Both commentators say so. A causative is secondary here, and the other bases are not causatives.

SK 3123 · वौ कषलसकत्थस्रम्भः · vau kaṣalasakatthasrambhaḥ (A 3.2.143)

वौ कषलसकत्थस्रम्भः

vau kaṣalasakatthasrambhaḥ · A 3.2.143

विकाषी । विलासी । विकत्थी । विस्रम्भी ॥

Rule. With vi, ghinuṇ comes after kaṣ "to injure", las "to sport", katth "to boast" and srambh "to trust".

Example.

vi + kaṣ + ghinuṇ → vṛddhi of the penultimate a (SK 2282) → vikāṣī "given to injuring".

vi + las → vilāsī "given to sport".

vi + katth → vikatthī "given to boasting".

vi + srambh → visrambhī "given to trusting".

The glosses of the roots are one commentator's.

Watch out. This rule's ghinuṇ does not always block other affixes. One commentator gives vikatthanaḥ, with yuc, beside vikatthī (see SK 3133).

SK 3124 · अपे च लषः · ape ca laṣaḥ (A 3.2.144)

अपे च लषः

ape ca laṣaḥ · A 3.2.144

चाद्वौ । अपलाषी । विलाषी ॥

Rule. With apa, ghinuṇ comes after laṣ "to desire". Bhaṭṭoji: by the ca, also with vi.

Example.

apa + laṣ + ghinuṇ → vṛddhi → apalāṣī.

vi + laṣ → vilāṣī.

Applies when / blocked by. One of Bhaṭṭoji's commentators: with apa or vi, and only in the senses of SK 3114.

SK 3125 · प्रे लपसृद्रुमथवदवसः · pre lapasṛdrumathavadavasaḥ (A 3.2.145)

प्रे लपसृद्रुमथवदवसः

pre lapasṛdrumathavadavasaḥ · A 3.2.145

प्रलापी । प्रसारी । प्रद्रावी । प्रमाथी । प्रवादी । प्रवासी ॥

Rule. With pra, ghinuṇ comes after lap "to speak", sṛ, dru, math "to churn", vad and vas "to dwell".

Example.

pra + lap + ghinuṇ → vṛddhi → pralāpī "given to prattle".

pra + sṛ → vṛddhi with r → prasārī.

pra + dru → vṛddhi → pradrau + in → āv (SK 61) → pradrāvī.

pramāthī, pravādī, pravāsī.

SK 3126 · निन्दहिंसक्लिशखादविनाशपरिक्षिपपरिरटपरिवादिव्याभाषासूयो वुञ् · nindahiṃsakliśakhādavināśaparikṣipapariraṭaparivādivyābhāṣāsūyo vuñ (A 3.2.146)

निन्दहिंसक्लिशखादविनाशपरिक्षिपपरिरटपरिवादिव्याभाषासूयो वुञ्

nindahiṃsakliśakhādavināśaparikṣipapariraṭaparivādivyābhāṣāsūyo vuñ · A 3.2.146

पञ्चम्यर्थे प्रथमा । एभ्यो वुञ् स्यात् । निन्दकः । हिंसक इत्यादि । ण्वुला सिद्धे वुञ्वचनं ज्ञापकं तच्छीलादिषु वासपरूपन्यायेन तृजादयो नेति ॥

Rule. After ten bases, the affix vuñ comes, in the agents of SK 3114. The bases are nind "to blame", hiṃs "to injure", kliś, khād, vināśi, parikṣip, pariraṭ, parivādi, vyābhāṣ and asūy. Bhaṭṭoji: the first case in the sūtra stands for the fifth. The marker ñ goes, and vu becomes aka (SK 1247).

Example. The running example:

nind + vuñ → nind + vu → nind + aka → nindakaḥ "one given to blaming".

hiṃs + vuñ → hiṃsakaḥ.

One of Bhaṭṭoji's commentators completes the list: kleśakaḥ, khādakaḥ, vināśakaḥ, parikṣepakaḥ, parirāṭakaḥ, parivādakaḥ, vyābhāṣakaḥ, asūyakaḥ.

Applies when / blocked by. Both commentators identify the bases. vināśa is the causative of naś with vi. parivādi is the causative of vad with pari. asūy is a root of the kaṇḍu list, formed with yak.

Watch out. This is an indication (jñāpaka). ṇvul (SK 2895) would already give nindakaḥ. So why teach vuñ? Bhaṭṭoji: it shows that in the senses of SK 3114, tṛc and the other general affixes do not come by the option between unlike affixes (SK 2830).

One of Bhaṭṭoji's commentators spells it out. Suppose the option held in these senses. Then tṛc and the others would come there, though ṇvul, being of the same form as vuñ, is always blocked by it. Suppose it does not hold. Then tṛn (SK 3115) is what would come in these senses, and vuñ has a purpose: it sets tṛn aside. He adds that the maxim is only generally valid (SK 3121).

SK 3127 · देविक्रुशोश्चोपसर्गे · devikruśoś copasarge (A 3.2.147)

देविक्रुशोश्चोपसर्गे

devikruśoścopasarge · A 3.2.147

आदेवकः । आक्रोशकः । उपसर्गे किम् । देवयिता । क्रोष्टा ॥

Rule. With an upasarga, vuñ also comes after devi and kruś "to cry out". devi is div ending in ṇi. Both commentators: it takes in the causative of div "to play" and a cur-class root, which also ends in ṇi.

Example.

ā + devi + vuñ → ā + devi + aka → the ṇi is lost (SK 2313) → ādevakaḥ.

ā + kruś + vuñ → guṇa (SK 2189) → ākrośakaḥ "one given to reviling".

Applies when / blocked by. Why "with an upasarga"? Bhaṭṭoji: devayitā, kroṣṭā. Without one, the general tṛn comes (SK 3115).

IV.H. yuc and its limits — running example: `cal + yuc → (yu → ana, SK 1247) cal-ana → calanaḥ`; `vṛt (anudātteta) + yuc → vartanaḥ`

The affix yuc leaves yu, which becomes ana by yuvor anākau (SK 1247). SK 3128–3131 give it after four kinds of root. SK 3132–3133 then take it away from some of them. Along the way Bhaṭṭoji draws two more indications about the option between unlike affixes. The running example is calanaḥ "one given to moving", from cal "to move", an intransitive root of motion. Beside it stands vartanaḥ, from vṛt "to turn". That root is marked with an anudātta vowel and begins with a consonant.

SK 3128 · चलनशब्दार्थादकर्मकाद्युच् · calanaśabdārthād akarmakād yuc (A 3.2.148)

चलनशब्दार्थादकर्मकाद्युच्

calanaśabdārthādakarmakādyuc · A 3.2.148

चलनार्थाच्छब्दार्थाच्च युच् स्यात् । चलनः । चोपनः । कम्पनः । शब्दनः । रवणः । अकर्मकात्किम् । पठिता विद्याम् ॥

Rule. After an intransitive (akarmaka) root meaning motion (calana) or sound (śabda), the affix yuc comes, in the agents of SK 3114. Its marker is c; yu remains.

Example. The running example:

cal + yuc → cal + yu → ana for yu (SK 1247) → calanaḥ "one given to moving".

Motion: copanaḥ (cup), kampanaḥ (kamp).

Sound: śabdanaḥ (śabd), ravaṇaḥ (ru, with guṇa).

One of Bhaṭṭoji's commentators names the roots.

Applies when / blocked by. Why "intransitive"? Bhaṭṭoji: paṭhitā vidyām "one who habitually recites learning". paṭh takes an object here, so yuc does not come. The general tṛn (SK 3115) does.

Watch out. Is śabd "to sound" intransitive? One of Bhaṭṭoji's commentators says yes. Its object, the sound, is included in the root's meaning.

SK 3129 · अनुदात्तेतश्च हलादेः · anudāttetaś ca halādeḥ (A 3.2.149)

अनुदात्तेतश्च हलादेः

anudāttetaśca halādeḥ · A 3.2.149

अकर्मकाद्युच् स्यात् । वर्तनः । वर्धनः । अनुदात्तेतः किम् । भविता । हलादेः किम् । एधिता । अकर्मकात्किम् । वसिता वस्त्रम् ॥

Rule. yuc also comes after an intransitive root that has an anudātta vowel as its marker (anudātteta) and begins with a consonant (halādi). "Intransitive" and yuc carry over from SK 3128.

Example. The running example:

vṛt + yuc → vṛt + ana → guṇa of the light penultimate (SK 2189) → vartanaḥ.

vṛdh + yuc → vardhanaḥ.

Applies when / blocked by. Bhaṭṭoji gives a counterexample for each condition. The general tṛn comes in each.

Not marked with an anudātta vowel: bhavitā.
Not beginning with a consonant: edhitā.
Not intransitive: vasitā vastram "one who habitually wears a garment".

Watch out. Why "beginning with" (ādi) a consonant, and not just "consonant"? Both commentators answer. Without ādi the rule would mean "ending in a consonant". Then the desiderative of gup would miss it, since that stem ends in a vowel. With ādi, jugupsanaḥ is formed. The anudātta marker of gup serves the desiderative stem.

SK 3130 · जुचङ्क्रम्यदन्द्रम्यसृगृधिज्वलशुचलषपतपदः · jucaṅkramyadandramyasṛgṛdhijvalaśucalaṣapatapadaḥ (A 3.2.150)

जुचङ्क्रम्यदन्द्रम्यसृगृधिज्वलशुचलषपतपदः

jucaṅkramyadandramyasṛgṛdhijvalaśucalaṣapatapadaḥ · A 3.2.150

जु इति सौत्रो धातुर्गतौ वेगे च । जवनः । चङ्क्रमणः । दन्द्रमणः । सरणः । पूर्वेण सिद्धे पदिग्रहणं लषपतपत SK 3134 इत्युकञा बाधा माभूदिति । तेन ताच्छीलिकेषु परस्परं वासरूपविधिर्नास्तीति । तेनालंकृञस्तृन्न ॥

Rule. yuc also comes after ten bases: ju, caṅkramya, dandramya, sṛ, gṛdh, jval, śuc, laṣ, pat and pad. Bhaṭṭoji: ju is a root found only in the sūtras (sautra), meaning "to go" and "speed". One of his commentators: it is not in the Dhātupāṭha.

Example.

ju + yuc → guṇa, av → javanaḥ.

kram + yaṅ → caṅkramya → + ana

→ the a of yaṅ is lost (ato lopaḥ, SK 2308) → the y is lost (yasya halaḥ, SK 2631)

→ caṅkramaṇaḥ "one given to roaming".

Likewise dandramaṇaḥ; and sṛ → saraṇaḥ.

One commentator continues the list: gardhanaḥ, jvalanaḥ, śocanaḥ, laṣaṇaḥ, patanaḥ, padanaḥ.

Applies when / blocked by. pad already takes yuc by SK 3129. Why name it again? Bhaṭṭoji: so that its yuc is not blocked by the ukañ of SK 3134, which also names pad.

Watch out. This is a second indication (jñāpaka). The naming of pad would be pointless if the affixes of habit allowed the option between unlike affixes among themselves. So, Bhaṭṭoji says, they do not allow it "mutually" (parasparam). Hence no tṛn after alaṃkṛñ: only alaṃkariṣṇuḥ (SK 3116).

Both commentators keep the two indications apart. SK 3126 shows that the general affixes, tṛc and the rest, do not come in these senses. This one shows that the affixes of habit do not alternate with one another.

SK 3131 · क्रुधमण्डार्थेभ्यश्च · krudhamaṇḍārthebhyaś ca (A 3.2.151)

क्रुधमण्डार्थेभ्यश्च

krudhamaṇḍārthebhyaśca · A 3.2.151

क्रोधनः । रोषणः । मण्डनः । भूषणः ॥

Rule. yuc also comes after roots that mean what krudh "to be angry" or maṇḍ "to adorn" means. So synonyms count too.

Example.

krudh + yuc → guṇa → krodhanaḥ "irascible".

ruṣ "to be angry" → roṣaṇaḥ.

maṇḍ → maṇḍanaḥ; bhūṣ "to adorn" → bhūṣaṇaḥ.

One of Bhaṭṭoji's commentators names the four roots.

SK 3132 · न यः · na yaḥ (A 3.2.152)

न यः

na yaḥ · A 3.2.152

यकारान्ताद्युच् न स्यात् । क्नूयिता । क्ष्मायिता ॥

Rule. yuc does not come after a root ending in y. One of Bhaṭṭoji's commentators: yaḥ is a fifth case, and by tadanta it means "ending in y".

Example.

knūy "to sound" → no yuc → tṛn → knūyitā.

kṣmāy "to shake" → kṣmāyitā.

Both roots are marked with an anudātta vowel, so SK 3129 would give yuc. This rule stops it.

Watch out. Is na yaḥ the root nay "to go"? No, says one commentator. That root would take yuc anyway, by SK 3129, and a rule for it would be pointless. na is the negative.

SK 3133 · सूददीपदीक्षश्च · sūdadīpadīkṣaś ca (A 3.2.153)

सूददीपदीक्षश्च

sūdadīpadīkṣaśca · A 3.2.153

युच् न स्यात् । सूदिता । दीपिता । दीक्षिता । नमिकम्पि - SK 3147 इति रेण युचो बाधे सिद्धे दीपेर्ग्रहणं ज्ञापयति ताच्छीलिकेषु वासरूपविधिर्नास्तीति प्रायिकमिति । तेन कम्रा कमना युवतिः । कम्प्रा कम्पना शाखा । यदि सूदेर्युज् न । कथं मधुसूदनः । नन्द्यादिः ॥

Rule. yuc also does not come after sūd "to pour", dīp "to shine" and dīkṣ "to consecrate". Bhaṭṭoji supplies "yuc does not come".

Example.

sūd → sūditā; dīp → dīpitā; dīkṣ → dīkṣitā.

These are tṛn forms. The yuc of SK 3129 is barred.

Applies when / blocked by. How then is madhusūdanaḥ formed, with ana after sūd? Bhaṭṭoji: sūd is in the nandi list. One commentator adds: so lyu comes by SK 2896.

Watch out. This is the third indication. dīp already takes ra by SK 3147, and ra would block yuc. Yet dīp is named here. Bhaṭṭoji: this shows that the maxim of SK 3130 is only generally valid (prāyika). So ra and yuc may stand side by side:

kamrā / kamanā yuvatiḥ "a young woman given to desire";

kamprā / kampanā śākhā "a bough given to trembling".

One commentator adds that the Bhāṣya means "generally valid", not "only for ra and yuc". He gives gantā beside gāmukaḥ, and vikatthanaḥ beside vikatthī.

IV.I. ukañ, ṣākan, ini, āluc — running example: `pat + ukañ → (vṛddhi) pāt-uka → pātukaḥ`; `jalp + ṣākan → jalpākaḥ` (fem. `varākī`); `ji + ini → (guṇa, ay) jayī`

Four more affixes of habit, each with its own list. ukañ (SK 3134) brings vṛddhi by its ñ. ṣākan (SK 3135) gives a feminine in ī by its ṣ. ini (SK 3136–3137) leaves in, like ghinuṇ, but brings guṇa. āluc (SK 3138) gives spṛhayāluḥ "eager". The running example has three parts. pātukaḥ "given to falling" is from pat. jalpākaḥ "talkative" is from jalp, with feminine varākī from vṛ. jayī "victorious" is from ji.

SK 3134 · लषपतपदस्थाभूवृषहनकमगमशॄभ्य उकञ् · laṣapatapadasthābhūvṛṣahanakamagamaśṝbhya ukañ (A 3.2.154)

लषपतपदस्थाभूवृषहनकमगमशॄभ्य उकञ्

laṣapatapadasthābhūvṛṣahanakamagamaśṝbhya ukañ · A 3.2.154

लाषुकः । पातुक इत्यादि ॥

Rule. After ten roots, the affix ukañ comes, in the agents of SK 3114. They are laṣ, pat, pad, sthā, bhū, vṛṣ, han, kam, gam and śṝ. Its marker is ñ; uka remains.

Example. The running example:

pat + ukañ → pat + uka → the ñ brings vṛddhi of the penultimate a (SK 2282) → pātukaḥ.

laṣ → lāṣukaḥ. Bhaṭṭoji then says "and so on".

One of his commentators fills in the rest:

pādukaḥ;
sthā → yuk (SK 2761) → sthāyukaḥ;
bhāvukaḥ, varṣukaḥ;
han → t for n (SK 2574), gh for h (SK 358) → ghātukaḥ;
kāmukaḥ, gāmukaḥ, śārukaḥ.

Watch out. ukañ does not shut out every other affix of habit. One commentator, at SK 3133, gives gantā beside gāmukaḥ.

SK 3135 · जल्पभिक्षकुट्टलुण्टवृङः षाकन् · jalpabhikṣakuṭṭaluṇṭavṛṅaḥ ṣākan (A 3.2.155)

जल्पभिक्षकुट्टलुण्टवृङः षाकन्

jalpabhikṣakuṭṭaluṇṭavṛṅaḥ ṣākan · A 3.2.155

जल्पाकः । भिक्षाकः । कुट्टाकः । लुण्टाकः । वराकः । वराकी ॥

Rule. After jalp "to talk", bhikṣ "to beg", kuṭṭ "to cut", luṇṭ "to rob" and vṛṅ, the affix ṣākan comes. Its markers are ṣ and n; āka remains.

Example. The running example:

jalp + ṣākan → jalp + āka → jalpākaḥ "talkative".

bhikṣākaḥ, kuṭṭākaḥ, luṇṭākaḥ.

vṛ + āka → guṇa → varākaḥ.

Feminine: the ṣ-marked affix brings ṅīṣ (SK 498) → varākī.

Both commentators say the ṣ is for ṅīṣ.

Applies when / blocked by. One commentator reports an explanation. ṣākan is taught where other affixes would come. jalp and bhikṣ would take yuc (SK 3128, SK 3129). vṛ would take ki or kin (SK 3151).

SK 3136 · प्रजोरिनिः · prajor iniḥ (A 3.2.156)

प्रजोरिनिः

prajoriniḥ · A 3.2.156

प्रजवी । प्रजविनौ । प्रजविनः ॥

Rule. After ju with pra, the affix ini comes. ju is the sautra root of SK 3130, "to go". One of Bhaṭṭoji's commentators: the i after n is for pronunciation; in remains.

Example.

pra + ju + ini → pra + ju + in

→ guṇa (SK 2168) → pra + jo + in → av (SK 61) → prajavin

→ prajavī, prajavinau, prajavinaḥ.

Watch out. One commentator criticizes the division of the sūtras. praju could have been read in the next sūtra, and this one dropped. A case-ending and a ca would then be saved.

SK 3137 · जिदृक्षिविश्रीण्वमाव्यथाभ्यमपरिभूप्रसूभ्यश्च · jidṛkṣiviśrīṇvamāvyathābhyamaparibhūprasūbhyaś ca (A 3.2.157)

जिदृक्षिविश्रीण्वमाव्यथाभ्यमपरिभूप्रसूभ्यश्च

jidṛkṣiviśrīṇvamāvyathābhyamaparibhūprasūbhyaśca · A 3.2.157

जयी । दरी । क्षयी । विश्रयी । अत्ययी । वमी । अव्यथी । अभ्यमी । परिभवी । प्रसवी ॥

Rule. ini also comes after ten bases: ji, dṛ, kṣi, viśri, iṇ, vam, avyath, abhyam, paribhū and prasū.

Example. The running example:

ji + ini → ji + in → guṇa → je + in → ay (SK 61) → jayin → jayī "victorious".

The rest, as Bhaṭṭoji lists them: darī, kṣayī, viśrayī, atyayī (iṇ with ati), vamī, avyathī, abhyamī, paribhavī, prasavī.

Applies when / blocked by. Both commentators name the roots. dṛ is dṛṅ "to respect". viśri is śriñ "to serve" with vi. prasū is ṣū "to urge" with pra. The ṣū roots that carry another marker are not taken.

Watch out. In avyatha, the negative nañ is compounded with the root by nipātana. Both commentators say so. Its n is then lost (SK 757): avyathī "fearless".

SK 3138 · स्पृहिगृहिपतिदयिनिद्रातन्द्राश्रद्धाभ्य आलुच् · spṛhigṛhipatidayinidrātandrāśraddhābhya āluc (A 3.2.158)

स्पृहिगृहिपतिदयिनिद्रातन्द्राश्रद्धाभ्य आलुच्

spṛhigṛhipatidayinidrātandrāśraddhābhya āluc · A 3.2.158

आद्यास्त्रयश्चुरादावदन्ताः । स्पृहयालुः । गृहयालुः । पतयालुः । दयालुः । निद्रालुः । तत्पूर्वो द्रा । तदो नान्तत्वं निपात्यते । तन्द्रालुः ।

शीङो वाच्यः

(वार्तिकम्) ॥ शयालुः ॥

Rule. After spṛhi, gṛhi, pati, dayi, nidrā, tandrā and śraddhā, the affix āluc comes. Its marker is c; ālu remains.

Example. Bhaṭṭoji: the first three are cur-class roots ending in a (adanta).

spṛha + ṇic → spṛhi → spṛhi + ālu

→ ay for ṇi (SK 2311) → spṛhay + ālu → spṛhayāluḥ "eager".

Likewise gṛhayāluḥ, patayāluḥ.

dayāluḥ "compassionate"; ni + drā → nidrāluḥ "sleepy".

In tandrā, drā is preceded by tad. Bhaṭṭoji: the n-final form of tad is given by nipātana → tandrāluḥ "drowsy".

A vārttika, śīṅo vācyaḥ: āluc should be stated after śīṅ "to lie" too.

śī + āluc → guṇa → śe + ālu → ay → śayāluḥ "given to lying down".

Applies when / blocked by. Why does Bhaṭṭoji add "ending in a"? Both commentators answer. With ṇic, guṇa or vṛddhi of the penultimate would follow. But the lost a is treated as still present (sthānivat). So neither applies.

Watch out. Bhaṭṭoji gives no form for śraddhā. Both commentators explain śraddhāluḥ, from dhā after the indeclinable śrat. Scholars discuss whether the form stood in Bhaṭṭoji's text.

IV.J. ru, kmarac, ghurac, kurac, kvarap, and gatvara — running example: `naś + kvarap → naś-vara → naśvaraḥ`; `iṇ + kvarap → (k-it: no guṇa; p-it: tuk) i-t-vara → itvaraḥ`

Six short rules give five more affixes and one nipātana. ru leaves ru. kmarac leaves mara; being k-marked, it takes no guṇa, the plain effect of SK 2217. ghurac leaves ura, and its gh brings the velar. kurac leaves ura and blocks guṇa. kvarap leaves vara; its k blocks guṇa and its p brings tuk. SK 3144 fixes gatvaraḥ by nipātana. The running example: naśvaraḥ "perishable", from naś. And itvaraḥ "given to going", from iṇ, with no guṇa and the augment t.

SK 3139 · दाधेट्सिशदसदो रुः · dādheṭsiśadasado ruḥ (A 3.2.159)

दाधेट्सिशदसदो रुः

dādheṭsiśadasado ruḥ · A 3.2.159

दारुः । धारुः । सेरुः । शद्रुः । सद्रुः ॥

Rule. After dā, dheṭ "to suck", si, śad and sad "to sit", the affix ru comes, in the agents of SK 3114.

Example.

dā + ru → dāruḥ.

dheṭ → ā for the final diphthong (SK 2370) → dhā + ru → dhāruḥ.

si + ru → guṇa → seruḥ.

śad + ru → śadruḥ; sad + ru → sadruḥ.

Applies when / blocked by. Which dā? One of Bhaṭṭoji's commentators reports a view. Only the three roots ḍudāñ "to give", do "to cut" and deṅ "to protect".

SK 3140 · सृघस्यदः क्मरच् · sṛghasyadaḥ kmarac (A 3.2.160)

सृघस्यदः क्मरच्

sṛghasyadaḥ kmarac · A 3.2.160

सृमरः । घस्मरः । अद्मरः ॥

Rule. After sṛ "to go", ghas "to eat" and ad "to eat", the affix kmarac comes. Its markers are k and c; mara remains.

Example.

sṛ + kmarac → k-marked: no guṇa, the plain effect of kṅiti ca (SK 2217) → sṛmaraḥ.

Bhaṭṭoji gives only the forms.

ghas + mara → ghasmaraḥ "voracious".

ad + mara → admaraḥ.

Watch out. ghas and ad both mean "to eat". Is ghas here only the substitute of ad (SK 2424)? No. One of Bhaṭṭoji's commentators calls it a separate base.

SK 3141 · भञ्जभासमिदो घुरच् · bhañjabhāsamido ghurac (A 3.2.161)

भञ्जभासमिदो घुरच्

bhañjabhāsamido ghurac · A 3.2.161

भङ्गुरः । भासुरः । मेदुरः ॥

Rule. After bhañj "to break", bhās "to shine" and mid, the affix ghurac comes. Its markers are gh and c; ura remains.

Example.

bhañj + ghurac → bhañj + ura

→ the gh-marked affix brings the velar (SK 2863): g for j → bhaṅguraḥ "fragile".

bhās + ura → bhāsuraḥ "shining".

mid + ura → guṇa → meduraḥ.

Watch out. One of Bhaṭṭoji's commentators reports a restriction. After bhañj the affix comes only when the object acts as its own agent (karmakartṛ), by the nature of the word's meaning.

SK 3142 · विदिभिदिच्छिदेः कुरच् · vidibhidicchideḥ kurac (A 3.2.162)

विदिभिदिच्छिदेः कुरच्

vidibhidicchideḥ kurac · A 3.2.162

विदुरः । भिदुरम् । छिदुरम् ॥

Rule. After vid "to know", bhid "to split" and chid "to cut", the affix kurac comes. Its markers are k and c; ura remains. Both commentators: only the vid of knowing, not the one of gaining.

Example.

vid + kurac → vid + ura → k-marked: no guṇa → viduraḥ "wise".

bhid + ura → bhiduram; chid + ura → chiduram.

Watch out. The Vṛtti, one commentator reports, limits the affix to an object acting as its own agent. This is not in the Bhāṣya. He cites Māgha, who uses chidura of an ordinary agent.

SK 3143 · इण्नश्जिसर्तिभ्यः क्वरप् · iṇnaśjisartibhyaḥ kvarap (A 3.2.163)

इण्नश्जिसर्तिभ्यः क्वरप्

iṇnaśjisartibhyaḥ kvarap · A 3.2.163

इत्वरः । इत्वरी । नश्वरः । जित्वरः । सृत्वरः ॥

Rule. After iṇ "to go", naś "to perish", ji "to conquer" and sṛ "to go", the affix kvarap comes. Its markers are k and p; vara remains.

Example. The running example:

iṇ + kvarap → i + vara

→ k-marked: no guṇa

→ p-marked kṛt after a short vowel: tuk (hrasvasya piti kṛti tuk, SK 2858) → i-t-vara → itvaraḥ.

Feminine: ṅīp (SK 470) → itvarī.

naś + vara → naśvaraḥ "perishable".

ji → tuk → jitvaraḥ; sṛ → sṛtvaraḥ.

Applies when / blocked by. Both commentators name the work of the markers. The k bars guṇa; the p brings tuk.

SK 3144 · गत्वरश्च · gatvaraś ca (A 3.2.164)

गत्वरश्च

gatvaraśca · A 3.2.164

गमेरनुनासिकलोपोऽपि निपात्यते । गत्वरी ॥

Rule. gatvara is also given by nipātana. One of Bhaṭṭoji's commentators: both kvarap after gam and the loss of its nasal are fixed by it. Bhaṭṭoji: "the loss of the nasal of gam is also given by nipātana".

Example.

gam + kvarap → gam + vara

→ the m is lost (by the nipātana) → ga + vara

→ tuk (SK 2858) → ga-t-vara → gatvaraḥ "transient".

Feminine: ṅīp → gatvarī.

Watch out. Why must the loss of m be fixed by nipātana? The same commentator answers. The ordinary rule would not apply, because vara does not begin with a jhal consonant.

IV.K. ūka and ra — running example: `jāgṛ + ūka → (guṇa + r) jāgar-ūka → jāgarūkaḥ`; `yaj + yaṅ → yāyaj(ya) + ūka → yāyajūkaḥ`

Two more affixes. ūka comes after jāgṛ (SK 3145) and after three intensive stems (SK 3146). ra comes after seven roots (SK 3147). The running example is jāgarūkaḥ "wakeful", from jāgṛ with guṇa and r. Beside it stands yāyajūkaḥ "one who sacrifices often", from the intensive of yaj.

SK 3145 · जागुरूकः · jāgurūkaḥ (A 3.2.165)

जागुरूकः

jāgurūkaḥ · A 3.2.165

जागर्तेरूकः स्यात् । जागरूकः ॥

Rule. After jāgṛ "to be awake", the affix ūka comes. jāguḥ is the fifth-case form of jāgṛ: jāguḥ + ūkaḥ. Bhaṭṭoji: "after jāgṛ, ūka comes".

Example. The running example:

jāgṛ + ūka → guṇa of the ṛ, with r (SK 70) → jāgar + ūka → jāgarūkaḥ "wakeful".

Watch out. Is jāgarūkaḥ a nipātana? No, say both commentators. If the whole form were fixed, ūka could not carry over into SK 3146.

SK 3146 · यजजपदशां यङः · yajajapadaśāṃ yaṅaḥ (A 3.2.166)

यजजपदशां यङः

yajajapadaśāṃ yaṅaḥ · A 3.2.166

एभ्यो यङ्न्तेभ्यः ऊकः स्यात् । दशामिति भाविना नलोपेन निर्देशः । यायजूकः । जञ्जपूकः । दन्दशूकः ॥

Rule. After yaj "to sacrifice", jap "to mutter" and daś "to bite", ending in the intensive affix yaṅ, ūka comes. ūka carries over from SK 3145. Bhaṭṭoji: daś names daṃś in the form it will have after the loss of its n. That loss is still to come.

Example. The running example:

yaj + yaṅ → reduplication → the reduplicated syllable is lengthened (dīrgho 'kitaḥ, SK 2632) → yāyajya

→ + ūka → the a of yaṅ is lost (SK 2308) → the y is lost (SK 2631)

→ yāyaj + ūka → yāyajūkaḥ "one who sacrifices often".

jap → the reduplicated syllable takes nuk (SK 2638) → jañjapūkaḥ.

daṃś → likewise → dandaśūkaḥ "given to biting".

The steps are both commentators'; Bhaṭṭoji gives only the forms.

SK 3147 · नमिकम्पिस्म्यजसकमहिंसदीपो रः · namikampismyajasakamahiṃsadīpo raḥ (A 3.2.167)

नमिकम्पिस्म्यजसकमहिंसदीपो रः

namikampismyajasakamahiṃsadīpo raḥ · A 3.2.167

नम्रः । कम्प्रः । स्मेरः । जसिर्नञ्पूर्वः क्रियासातत्ये वर्तते । अजस्रम् । सन्ततमित्यर्थः । कम्रः । हिंस्रः । दीप्रः ॥

Rule. After seven bases, the affix ra comes, in the agents of SK 3114. They are nam, kamp "to tremble", smi, ajas, kam "to desire", hiṃs and dīp "to shine".

Example.

nam + ra → namraḥ "bowing".

kamp + ra → kampraḥ.

smi + ra → guṇa → smeraḥ "smiling".

kamraḥ, hiṃsraḥ, dīpraḥ.

Applies when / blocked by. What is ajas? Bhaṭṭoji: it is the root jas with the negative nañ, meaning continuity of action. So ajasram means "continuous". One of his commentators: jas is jasu "to free". No root ajas is in the Dhātupāṭha. nañ is compounded with the root by nipātana, and its n is lost (SK 757).

Watch out. ra does not always block yuc. By SK 3133, kamanā stands beside kamrā and kampanā beside kamprā.

IV.L. u, and ki/kin treated like liṭ — running example: `kṛ + san → cikīrṣa + u → (a lost) cikīrṣuḥ`; `pā + ki (liḍvat: reduplication) → pa-pā-i → (ā lost) papiḥ`

The affix u comes after desiderative stems and a few roots (SK 3148), in two nipātanas (SK 3149) and in the Veda after kya-stems (SK 3150). SK 3151 gives ki and kin, which behave as if they were liṭ: the root is reduplicated. A vārttika extends them to ordinary speech. The running example has two forms. cikīrṣuḥ "wishing to do" is the desiderative of kṛ with u. papiḥ "drinker" is pā with ki, reduplicated as before liṭ.

SK 3148 · सनाशंसभिक्ष उः · sanāśaṃsabhikṣa uḥ (A 3.2.168)

सनाशंसभिक्ष उः

sanāśaṃsabhikṣa uḥ · A 3.2.168

चिकीर्षुः । आशंसुः । भिक्षुः ॥

Rule. The affix u comes after a stem ending in the desiderative affix san, after āśaṃs "to wish" and after bhikṣ "to beg". It comes in the agents of SK 3114.

Example. The running example:

kṛ + san → the desiderative stem cikīrṣa

→ + u → the final a of the stem is lost (ato lopaḥ, SK 2308) → cikīrṣ + u → cikīrṣuḥ "wishing to do".

ā + śaṃs + u → āśaṃsuḥ.

bhikṣ + u → bhikṣuḥ "mendicant".

Applies when / blocked by. Does san mean the affix or the roots ṣaṇ "to give" and "to share"? Both commentators say the affix. Their proof: vijigīṣu is read in the garga list.

Watch out. āśaṃs is śaṃs with ā, meaning "to wish". It is not śaṃs "to praise". One commentator explains this.

SK 3149 · विन्दुरिच्छुः · vindur icchuḥ (A 3.2.169)

विन्दुरिच्छुः

vinduricchuḥ · A 3.2.169

वेत्तेर्नुम् एषेश्छत्वं च निपात्यते । वेत्ति तच्छीलो विन्दः । इच्छति इच्छुः ॥

Rule. vindu and icchu are given by nipātana, with the affix u. Bhaṭṭoji: the augment num for vid (vetti), and ch for the ṣ of iṣ, are fixed by nipātana.

Example.

vid + u → num (by the nipātana) → vind + u → vinduḥ.

Bhaṭṭoji: one whose habit is to know (vetti tacchīlaḥ) is vinduḥ.

iṣ + u → ch for ṣ (by the nipātana) → icch + u → icchuḥ.

Bhaṭṭoji: one who wishes (icchati) is icchuḥ.

Watch out. bindu "drop", with b, is a different word. One commentator derives it from another root, among the Uṇādi words.

SK 3150 · क्याच्छन्दसि · kyāc chandasi (A 3.2.170)

क्याच्छन्दसि

kyācchandasi · A 3.2.170

देवाञ्जिगाति सुम्नयुः ॥

Rule. In the Veda, u comes after a stem ending in kya. One commentator: kya covers the denominative affixes kyac, kyaṅ and kyaṣ.

Example.

sumna + kyac "desires well-being for himself" (SK 2657) → sumnaya

→ + u → the a is lost (SK 2308) → sumnayuḥ.

Bhaṭṭoji: devāñ jigāti sumnayuḥ "desiring well-being for himself, he goes to the gods".

Applies when / blocked by. Why no ī before ya? Both commentators answer. In the Veda na cchandasy aputrasya (SK 3588) bars the ī of kyaci ca (SK 2658). It also bars the lengthening of SK 2298.

SK 3151 · आदृगमहनजनः किकिनौ लिट् च · ādṛgamahanajanaḥ kikinau liṭ ca (A 3.2.171)

आदृगमहनजनः किकिनौ लिट् च

ādṛgamahanajanaḥ kikinau liṭ ca · A 3.2.171

आदन्तादृदन्ताद्गमादिभ्यश्च किकिनौ स्तश्छन्दसि तौ च लिड्वत् । पपिः सोमं ददिर्गाः । बभ्रिर्वज्रम् । जग्मिर्युवा । जघ्निर्वृत्रममित्रियम् । जज्ञिः ।

भाषायां धाञ्कृसृगमिजनिनमिभ्यः

(वार्तिकम्) ॥ दधिः । चक्रिः । सस्रिः । जग्मिः । जज्ञिः । नेमिः ।

सासहिवावहिचाचलिपापतीनामुपसंख्यानम्

(वार्तिकम्) ॥ यङ्न्तेभ्यः सहेत्यादिभ्यः किकिनौ पतेर्नीगभावश्च निपात्यते ॥

Rule. In the Veda, the affixes ki and kin come after roots ending in ā or ṛ, and after gam, han and jan. They are treated like liṭ (liḍvat). So what happens to a root before liṭ happens before them: reduplication and the rest. One of Bhaṭṭoji's commentators: ki and kin differ only in accent.

Example. The running example:

pā + ki → like liṭ: reduplication → pa-pā + i → the ā is lost → papiḥ.

Bhaṭṭoji: papiḥ somam "drinker of soma", dadir gāḥ "giver of cows", babhrir vajram "bearer of the thunderbolt", jagmir yuvā, jaghnir vṛtram amitriyam "slayer of the hostile Vṛtra", jajñiḥ.

han + ki → reduplication → the penultimate a is lost (SK 2363) → jaghniḥ.

Applies when / blocked by. Why is somam in the second case, not the sixth? One commentator answers: na lokāvyaya… (SK 627) bars the sixth. He adds that "like liṭ" transfers the operations only. ki and kin are not replaced by verb-endings.

A vārttika, bhāṣāyāṃ dhāñkṛsṛgamijaninamibhyaḥ: in ordinary speech, ki and kin come after dhā, kṛ, sṛ, gam, jan and nam.

dadhiḥ, cakriḥ, sasriḥ, jagmiḥ, jajñiḥ.

nam + ki → reduplication → e and loss of the reduplicated syllable → nemiḥ.

A second vārttika, sāsahivāvahicācalipāpatīnām upasaṃkhyānam: sāsahi, vāvahi, cācali and pāpati should be added. Bhaṭṭoji: ki and kin after sah and the others ending in yaṅ, and, for pat, no augment nīk, are fixed by nipātana.

Watch out. Why are two Vedic sūtras taught here, in a book of ordinary usage? Both commentators answer. They are set out to make clear the vārttika for ordinary speech.

The commentators name different rules for the gh of jaghniḥ; scholars discuss this.

IV.M. najiṅ, āru, kru/klukan, varac — running example: `sthā + varac → sthāvaraḥ`; `yā + yaṅ → yāyāya + vara → (a lost; y lost) yāyāvaraḥ`

Five rules give five affixes of habit, and a vārttika adds krukan. najiṅ (SK 3152) gives svapnak "sleepy". āru (SK 3153) gives vandāruḥ. kru and klukan (SK 3154) give bhīruḥ and bhīlukaḥ "timid". varac (SK 3155–3156) gives sthāvaraḥ and yāyāvaraḥ. The running example is sthāvaraḥ "stationary", from sthā with varac. Beside it stands yāyāvaraḥ "a wanderer", from the intensive of yā "to reach". SK 3156 explains why its ā survives.

SK 3152 · स्वपितृषोर्नजिङ् · svapitṛṣor najiṅ (A 3.2.172)

स्वपितृषोर्नजिङ्

svapitṛṣornajiṅ · A 3.2.172

स्वप्नक् । तृष्णक् । तृष्णजौ । तृष्णजः । धृषेश्चेति वाच्यमिति काशिकादौ । धृष्णक् ॥

Rule. After svap "to sleep" and tṛṣ "to thirst", the affix najiṅ comes. One of Bhaṭṭoji's commentators: the i is for pronunciation and the ṅ is a marker; naj remains.

Example.

svap + najiṅ → svap + naj → svapnaj → first case svapnak "given to sleeping".

tṛṣ + naj → ṅ-marked: no guṇa → ṇ for n → tṛṣṇaj

→ tṛṣṇak, tṛṣṇajau, tṛṣṇajaḥ "given to thirsting".

Applies when / blocked by. Bhaṭṭoji: "it should be stated: and after dhṛṣ", so say the Kāśikā and other works: dhṛṣṇak. One commentator takes the phrase "the Kāśikā and other works" to mean it is not in the Bhāṣya.

SK 3153 · शॄवन्द्योरारुः · śṝvandyor āruḥ (A 3.2.173)

शॄवन्द्योरारुः

śṝvandyorāruḥ · A 3.2.173

शरारुः । वन्दारुः ॥

Rule. After śṝ "to injure" and vand "to salute, to praise", the affix āru comes. One of Bhaṭṭoji's commentators names the roots.

Example.

śṝ + āru → guṇa, with r → śar + āru → śarāruḥ "harmful".

vand + āru → vandāruḥ "given to praising".

SK 3154 · भियः क्रुक्लुकनौ · bhiyaḥ kruklukanau (A 3.2.174)

भियः क्रुक्लुकनौ

bhiyaḥ kruklukanau · A 3.2.174

भीरुः । भीलुकः ।

क्रुकन्नपि वाच्यः

(वार्तिकम्) ॥ भीरुकः ॥

Rule. After bhī "to fear", the affixes kru and klukan come. One of Bhaṭṭoji's commentators: they are k-marked, so there is no guṇa.

Example.

bhī + kru → bhī + ru → bhīruḥ "timid".

bhī + klukan → bhī + luka → bhīlukaḥ.

Applies when / blocked by. A vārttika, krukann api vācyaḥ: "krukan too should be stated" → bhīrukaḥ. One commentator: this is why the Amarakośa lists bhīru, bhīruka and bhīluka.

SK 3155 · स्थेशभासपिसकसो वरच् · stheśabhāsapisakaso varac (A 3.2.175)

स्थेशभासपिसकसो वरच्

stheśabhāsapisakaso varac · A 3.2.175

स्थावरः । ईश्वरः । भास्वरः । पेस्वरः । कस्वरः ॥

Rule. After sthā "to stand", īś "to rule", bhās "to shine", pis "to go" and kas "to go", the affix varac comes. Its marker is c; vara remains.

Example. The running example:

sthā + varac → sthā + vara → sthāvaraḥ "stationary".

īś + vara → no iṭ (neḍ vaśi kṛti, SK 2981) → īśvaraḥ "lord".

bhāsvaraḥ "shining".

pis → pesvaraḥ.

kasvaraḥ.

Watch out. The feminine īśvarī "the wife of a lord" has ṅīṣ by puṃyoga. One of Bhaṭṭoji's commentators notes this.

SK 3156 · यश्च यङः · yaś ca yaṅaḥ (A 3.2.176)

यश्च यङः

yaśca yaṅaḥ · A 3.2.176

यातेर्यङन्ताद्वरच् स्यात् । अतो लोपः SK 2380 । तस्य अचः परस्मिन् SK 50 इति स्थानिवद्भावे प्राप्ते पदस्य चरमावयवे द्विर्वचनादौ च कर्तव्ये परनिमित्तोऽजादेशे न स्थानिवत् तस्य यलोपं प्रति स्थानिवद्भावनिषेधात् लोपो व्योः - SK 873 इति यलोपः । अल्लोपस्य स्थानिवत्त्वामाश्रित्य आतो लोपे प्राप्ते । वरे लुप्तं न स्थानिवत् । यायावरः ॥

Rule. varac also comes after yā "to reach" ending in the intensive affix yaṅ. Bhaṭṭoji: "after yā ending in yaṅ, varac comes". One of his commentators: yaḥ is the sixth-case form of yā.

Example. The running example, with Bhaṭṭoji's reasoning:

yā + yaṅ → reduplication → lengthening (SK 2632) → yāyāya

→ + vara → the a of yaṅ is lost (ato lopaḥ, SK 2308) → yāyāy + vara

→ the y is lost before v (lopo vyor vali, SK 873) → yāyā + vara

→ the ā of yā stays → yāyāvaraḥ "a wanderer".

Applies when / blocked by. Bhaṭṭoji meets two difficulties.

First, the lost a would count as still present (sthānivat) by acaḥ parasmin (SK 50). Then the y would not stand before v, and could not be lost. But SK 51 denies this. A vowel-substitute does not count as the original when the loss of y is to be done. So SK 873 applies.

Second, relying on that same sthānivat a, the ā of yā might be lost (SK 2372). The answer is a clause of SK 51: "a loss before vara is not treated as the original" (vare luptaṃ na sthānivat). So the ā stays.

Watch out. yasya halaḥ (SK 2631) does not delete this y. One of Bhaṭṭoji's commentators explains: the y does not follow a consonant. lopo vyor vali does the work.

IV.N. kvip, which closes the tācchīlika range; kvip after other roots; bhū with kvip or ḍu — running example: `vi + bhrāj + kvip → (ṣ, SK 294; jaś, car) vibhrāṭ`; `vac + kvip (lengthening, no saṃprasāraṇa) → vāk`; `pra + bhū + ḍu → (ṭi lost) prabhuḥ`

The affix kvip leaves nothing behind: the root itself serves as the stem. SK 3157 gives it after eight bases and closes the heading of SK 3114. SK 3158 gives it after other roots too, with four vārttikas. SK 3159 restricts the kvip of bhū. SK 3160 gives ḍu after bhū with three upasargas. The running example has three forms. vibhrāṭ "shining" is bhrāj with vi and kvip. vāk "speech" is vac with kvip, lengthened and without saṃprasāraṇa. prabhuḥ "master" is bhū with pra and ḍu, which loses its last vowel.

SK 3157 · भ्राजभासधुर्विद्युतोर्जिपॄजुग्रावस्तुवः क्विप् · bhrājabhāsadhurvidyutorjipṝjugrāvastuvaḥ kvip (A 3.2.177)

भ्राजभासधुर्विद्युतोर्जिपॄजुग्रावस्तुवः क्विप्

bhrājabhāsadhurvidyutorjipṝjugrāvastuvaḥ kvip · A 3.2.177

विभ्राट् । भाः । भासौ । धूः । धुरौ । विद्युत् । उर्क् । पूः । पुरौ । दृशिग्रहणस्यापकर्षाज्जवतेर्दीर्घः । जूः । जुवौ । जुवः । ग्रावशब्दस्य धातुना समासः सूत्रे निपात्यते । ततः क्विप् । ग्रावस्तुत् ॥

Rule. After eight bases, the affix kvip comes, in the agents of SK 3114. They are bhrāj "to shine", bhās "to shine", dhurv "to injure", dyut "to shine", ūrj "to be strong", pṝ "to fill", ju "to go", and stu "to praise" with grāvan. This kvip is the limit of the heading ā kveḥ (SK 3114), one commentator notes.

Example. The running example:

vi + bhrāj + kvip → nothing of the affix remains → vibhrāj

→ ṣ for j (vraśca…, SK 294) → then ḍ, then ṭ at the end → vibhrāṭ.

The other forms, with the commentators' steps:

bhās → bhāḥ, bhāsau.
dhurv → the v is lost (rāl lopaḥ, SK 2655) → dhur → lengthening (SK 433) → dhūḥ, dhurau.
vi + dyut → vidyut.
ūrj → k for j (SK 378) → ūrk. One commentator: the last consonant of the cluster is not lost, by the restriction SK 280.
pṝ → ur for ṝ (SK 2494) → pur → lengthening → pūḥ, purau.

Applies when / blocked by. Why does ju have a long vowel, jūḥ, juvau, juvaḥ? Bhaṭṭoji: the word dṛśi "is seen" is drawn back (apakarṣa) from SK 3158. One commentator: only the tradition's explanation supports this.

Watch out. A finished word like grāvan cannot compound with a bare root before the affix. Bhaṭṭoji: the compound is fixed by nipātana in the sūtra. Then kvip: grāvastut. One commentator adds tuk, the affix being p-marked.

SK 3158 · अन्येभ्योऽपि दृश्यते · anyebhyo 'pi dṛśyate (A 3.2.178)

अन्येभ्योऽपि दृश्यते

anyebhyo'pi dṛśyate · A 3.2.178

क्विप् । छित् । भिद् । दृशिग्रहणं विध्यन्तरोपसंग्रहार्थम् । क्वचिद्दीर्घः क्वचिदसंप्रसारणं क्वचिद्द्वे क्वचिद्ध्रस्वः । तथा च वार्तिकम् ।

क्विब्वचिप्रच्छयायतस्तुकटप्रुजुश्रीणां दीर्घोऽसंप्रसारणं च

(वार्तिकम्) ॥ किब्वचीत्यादिना उणादिसूत्रेण केषांचित्सिद्धे तच्छीलादौ तृना बाधा मा भूदिति वार्तिके ग्रहणम् । वक्तीति वाक् । पृच्छतीति प्राट् । आयतं स्तौतीति आयतस्तूः । कटं प्रवते कटप्रूः । जूरुक्तः । श्रयति हरिं सा श्रीः ।

द्युतिगमिजुहोतीनां द्वे च

(वार्तिकम्) ॥ दृशिग्रहणादभ्याससंज्ञा । दिद्युत् । जगत् ।

जुहोतेर्दीर्घश्च

(वार्तिकम्) ॥ जुहूः । दॄ भये । अस्य ह्रस्वश्च । दीर्यति ददृत् ।

ध्यायतेः संप्रसारणं च

(वार्तिकम्) ॥ धीः ॥

Rule. kvip "is seen" after other roots too. Bhaṭṭoji supplies kvip. The heading ends with SK 3157's kvip, but the senses of SK 3114 are still read here. One of his commentators says so, and Bhaṭṭoji's remark on tṛn below presupposes it.

Example.

chid + kvip → chit; bhid + kvip → bhit.

Applies when / blocked by. Bhaṭṭoji: the word dṛśi "is seen" takes in other operations. In some cases lengthening, in some no saṃprasāraṇa, in some reduplication, in some shortening. Four vārttikas follow.

The first, kvib vacipracchyāyatastukaṭaprujuśrīṇāṃ dīrgho 'saṃprasāraṇaṃ ca: kvip after vac, pracch, stu with āyata, pru with kaṭa, ju and śrī, with lengthening and no saṃprasāraṇa.

vac + kvip → lengthening, no saṃprasāraṇa → vāc → vāk "speech" (vaktīti vāk).
pracch + kvip → no saṃprasāraṇa → prāṭ "one who asks".
āyatastūḥ "one who praises the extended"; kaṭaprūḥ "one who goes to the mat".
jūḥ has been stated under SK 3157.
śrī: śrayati hariṃ sā śrīḥ "she who resorts to Hari is Śrī".

Bhaṭṭoji explains why these forms are in a vārttika. Some already come by the Uṇādi sūtra "kvib vaci…". They are restated so that tṛn does not block them in the senses of SK 3114.

The second, dyutigamijuhotīnāṃ dve ca: for dyut, gam and hu, reduplication with kvip. Bhaṭṭoji: the first part gets the name abhyāsa because of dṛśi in the sūtra.

dyut → reduplication; saṃprasāraṇa of the first part (SK 2344) → didyut.
gam → reduplication; the m is lost (gamaḥ kvau, SK 2986); tuk → jagat.

The third, juhoter dīrghaś ca: for hu, lengthening too → juhūḥ "ladle": the ladle, spoken of as the agent, as one commentator explains.

Bhaṭṭoji adds dṝ "to fear", with shortening (hrasvaś ca): dīryati "fears" → dadṛt.

The fourth, dhyāyateḥ saṃprasāraṇaṃ ca: for dhyai, saṃprasāraṇa with kvip → lengthening (SK 2559) → dhīḥ.

Watch out. Why must dṛśi give the name abhyāsa? Both commentators answer. pūrvo 'bhyāsaḥ (SK 2178) names only a reduplication taught in the sixth book.

One commentator rejects a reading of śrīḥ as an object noun, "she to whom they resort". It conflicts with kartari kṛt (SK 2832). The other meets a charge against the fourth vārttika, dhyāyateḥ saṃprasāraṇaṃ ca: that it repeats an Uṇādi sūtra. It does not, he says, since the Uṇādi sūtras are another sage's work, not Pāṇini's.

SK 3159 · भुवः संज्ञाऽन्तरयोः · bhuvaḥ saṃjñāntarayoḥ (A 3.2.179)

भुवः संज्ञाऽन्तरयोः

bhuvaḥ saṃjñā'ntarayoḥ · A 3.2.179

मित्रभूर्नाम कश्चित् । धनिकाधमर्णयोरन्तरे यस्तिष्ठति विश्वासार्थं स प्रतिभूः ॥

Rule. After bhū "to be", kvip comes only when the word is a name (saṃjñā), or means one who stands between (antara). One of Bhaṭṭoji's commentators supplies kvip and calls the rule a restriction (niyama).

Example.

mitra + bhū + kvip → mitrabhūḥ, the name of someone.

prati + bhū + kvip → pratibhūḥ "surety". Bhaṭṭoji: one who stands between a creditor and a debtor, for the sake of trust.

Applies when / blocked by. Is pratibhūḥ not also a name? No, say both commentators. A name lasts as long as its bearer. A man ceases to be a pratibhū once the debt is repaid. One commentator adds: one who stands between two villages is not a pratibhū. The word dṛśi, still in force, keeps the sense narrow.

SK 3160 · विप्रसम्भ्यो ड्वसंज्ञायाम् · viprasambhyo ḍv asaṃjñāyām (A 3.2.180)

विप्रसम्भ्यो ड्वसंज्ञायाम्

viprasambhyo ḍvasaṃjñāyām · A 3.2.180

एभ्यो भुवो डुः स्यान्न तु संज्ञायाम् । विभूर्व्यापकः । प्रभुः स्वामी । संभुर्जनिता । संज्ञायां तु विभूर्नाम कश्चित् ।

मितद्र्वादिभ्य उपसंख्यानम्

(वार्तिकम्) ॥ मितं द्रवतीति मितद्रुः । शतद्रुः । शंभुः । अन्तर्भावितण्यर्थोऽत्र भवतिः ॥

Rule. After bhū with vi, pra or sam, the affix ḍu comes, but not in a name. Bhaṭṭoji supplies bhū. Its marker is ḍ; u remains.

Example. The running example:

pra + bhū + ḍu → pra + bhū + u

→ the ḍ-marked affix: the ṭi of bhū (its last vowel, with anything after it) is lost → pra + bh + u → prabhuḥ "master".

vibhuḥ "pervading"; sambhuḥ "begetter".

In a name, kvip comes by SK 3159: vibhūḥ, the name of someone.

Applies when / blocked by. One of Bhaṭṭoji's commentators: the ṭi is lost by force of the ḍ marker alone, though the stem is not a bha stem.

A vārttika, mitadrvādibhya upasaṃkhyānam: ḍu should be added for mitadru and the rest.

mitaṃ dravatīti mitadruḥ "one who runs a measured pace"; śatadruḥ; śambhuḥ.

Watch out. In śambhuḥ, Bhaṭṭoji says, bhavati carries the causative sense within it. "One who causes happiness (śam) to be." One commentator reports other analyses; scholars discuss this.

IV.O. ṣṭran and itra: affixes of the object, the instrument and the agent, with the iṭ-bar — running example: `dā + ṣṭran (karaṇa) → dā-tra → dātram`; `śas + ṣṭran → (no iṭ, 3163) śastram`; `pū + itra → (guṇa, av) pavitram`

The heading of habit is over. The last affixes of the prakaraṇa are ṣṭran (SK 3161–3164) and itra (SK 3165–3167). They leave aside the agent sense of SK 2832 for the object and the instrument, and in SK 3167 the agent again. Bhaṭṭoji inserts one rule from the seventh book, SK 3163, which bars iṭ before tra. The running example has three forms. dātram "sickle" is dāp "to cut" with ṣṭran in the sense of the instrument. śastram "weapon" is śas with ṣṭran, with no iṭ. pavitram "strainer" is pū with itra.

SK 3161 · धः कर्मणि ष्ट्रन् · dhaḥ karmaṇi ṣṭran (A 3.2.181)

धः कर्मणि ष्ट्रन्

dhaḥ karmaṇi ṣṭran · A 3.2.181

धेटो घाञश्च कर्मण्यर्थे ष्ट्रन् स्यात् । धात्री जनन्यामलकीवसुमत्युपमातृषु ॥

Rule. After dhā, the affix ṣṭran comes in the sense of the object (karman). Bhaṭṭoji: dhā means dheṭ "to suck" and dhāñ "to place, to nourish". One commentator: here karmaṇi names the sense, not an upapada.

Example.

dhā + ṣṭran → the markers ṣ and n go

→ the ṭ was only the t made retroflex by the ṣ; with the ṣ it goes back to t → dhā + tra

→ feminine ṅīṣ, because the affix is ṣ-marked → dhātrī.

Bhaṭṭoji: dhātrī means "mother", "āmalakī", "earth" and "foster-mother".

One commentator: dhātrī is "one who is sucked", a wet-nurse.

SK 3162 · दाम्नीशसयुयुजस्तुतुदसिसिचमिहपतदशनहः करणे · dāmnīśasayuyujastutudasisicamihapatadaśanahaḥ karaṇe (A 3.2.182)

दाम्नीशसयुयुजस्तुतुदसिसिचमिहपतदशनहः करणे

dāmnīśasayuyujastutudasisicamihapatadaśanahaḥ karaṇe · A 3.2.182

दावादेः ष्ट्रन् स्यात्करणेऽर्थे । दान्त्यनेन दात्रम् । नेत्रम् ॥

Rule. After thirteen roots, ṣṭran comes in the sense of the instrument (karaṇa). They are dāp, nī, śas, yu, yuj, stu, tud, si, sic, mih, pat, daṃś and nah. ṣṭran carries over from SK 3161.

Example. The running example:

dāp + ṣṭran → dā + tra → dātram "sickle".

Bhaṭṭoji: dānty anena "they cut with it".

nī + tra → guṇa → netram "eye".

The other eleven forms are given under SK 3163.

Applies when / blocked by. Why is the root written dām? One of Bhaṭṭoji's commentators answers. It is dāp "to cut", with m for p before the nasal (SK 116). So it is not dā "to give".

Watch out. daś in the sūtra is daṃś "to bite". The commentators explain its n-less form differently.

SK 3163 · तितुत्रतथसिसुसरकसेषु च · titutratathasisusarakaseṣu ca (A 7.2.9)

तितुत्रतथसिसुसरकसेषु च

titutratathasisusarakaseṣu ca · A 7.2.9

एषां दशानां कृत्प्रत्ययानामिण्न स्यात् । शस्त्रम् । योत्रम् । योक्त्रम् । स्तोत्रम् । तोत्रम् । सेत्रम् । सेक्त्रम् । मेढ्रम् । पत्रम् । दंष्ट्रा । नद्ध्री ॥

Rule. Ten kṛt affixes do not take iṭ: ti, tu, tra, ta, tha, si, su, sara, ka and sa. This is a seventh-book rule, placed here out of order. One of Bhaṭṭoji's commentators gives the reason. iṭ might be expected for śas and pat before tra.

Example. The running example:

śas + ṣṭran → śas + tra → no iṭ (this rule) → śastram "weapon".

The forms of SK 3162, as Bhaṭṭoji lists them:

yotram, yoktram, stotram, totram, setram, sektram, meḍhram, patram, daṃṣṭrā, naddhrī.

nah + tra → dh for h, dh for t → d for the first dh (jaś) → naddhra → ṅīṣ → naddhrī "a leather strap".

Applies when / blocked by. One of Bhaṭṭoji's commentators illustrates all ten affixes. For example ti in tantiḥ, tu in saktuḥ, sa in vatsaḥ.

Watch out. ṣṭran is ṣ-marked, so why daṃṣṭrā and not ṅīṣ? One commentator says ṅīṣ after a ṣ-marked affix is not obligatory. Others give ṭāp through the aja list. Scholars discuss this.

SK 3164 · हलसूकरयोः पुवः · halasūkarayoḥ puvaḥ (A 3.2.183)

हलसूकरयोः पुवः

halasūkarayoḥ puvaḥ · A 3.2.183

पूङ्पूञोः करणे ष्ट्रन् स्यात् तच्चेत्करणं हलसूकरयोरवयवः । हलस्य सूकरस्य वा पोत्रम् । मुखमित्यर्थः ॥

Rule. After pū "to purify", ṣṭran comes in the sense of the instrument. That instrument must be part of a plough or of a hog. Bhaṭṭoji: both pūṅ and pūñ.

Example.

pū + ṣṭran → guṇa → po + tra → no iṭ (SK 3163) → potram.

Bhaṭṭoji: the potra of a plough or of a hog, "that is, its fore-part" (mukha): the front of the plough, the snout of the hog.

Applies when / blocked by. Bhaṭṭoji says "if that instrument is part of a plough or a hog". One commentator: this rejects two views. That the plough and hog are what the word means. And that they are upapadas.

SK 3165 · अर्तिलूधूसूखनसहचर इत्रः · artilūdhūsūkhanasahacara itraḥ (A 3.2.184)

अर्तिलूधूसूखनसहचर इत्रः

artilūdhūsūkhanasahacara itraḥ · A 3.2.184

अरित्रम् । लवित्रम् । धवित्रम् । सवित्रम् । खनित्रम् । सहित्रम् । चरित्रम् ॥

Rule. After seven roots, the affix itra comes, in the sense of the instrument. They are ṛ (arti), lū, dhū, sū, khan, sah and car.

Example.

ṛ + itra → guṇa → ar + itra → aritram.

lū + itra → guṇa → lo + itra → av → lavitram.

dhavitram, savitram, khanitram, sahitram, caritram.

Watch out. One commentator takes dhū as a root of the kuṭ group. After such a root an affix counts as ṅ-marked (SK 2461), so there is no guṇa. His reading then presupposes dhuvitram. Bhaṭṭoji's text has dhavitram; scholars discuss this.

SK 3166 · पुवः संज्ञायाम् · puvaḥ saṃjñāyām (A 3.2.185)

पुवः संज्ञायाम्

puvaḥ saṃjñāyām · A 3.2.185

पवित्रम् । येनाज्यमुत्पूयते यच्चानामिकावेष्टनम् ॥

Rule. After pū, itra comes in the sense of the instrument, when the word is a name. Both commentators: "instrument" still holds. One commentator: both pūṅ and pūñ are taken.

Example.

pū + itra → guṇa → po + itra → av → pavitram.

Bhaṭṭoji: that by which the ājya (clarified butter) is purified, and that which is wrapped round the ring finger.

SK 3167 · कर्तरि चर्षिदेवतयोः · kartari carṣidevatayoḥ (A 3.2.186)

कर्तरि चर्षिदेवतयोः

kartari carṣidevatayoḥ · A 3.2.186

पुव इत्रः स्यात् ऋषौ करणे देवतायां कर्तरि । ऋषिर्वेदमन्त्रः । तदुक्तमृषिणेति दर्शनात् । पूयतेऽनेनेति पवित्रम् ॥ देवतायां त्वग्निः पवित्रं स मा पुनातु ॥

Rule. After pū, itra comes in the instrument sense and in the agent sense (kartṛ). The ca brings in the instrument. Bhaṭṭoji pairs them in order (yathāsaṃkhya). The instrument goes with a ṛṣi, the agent with a deity (devatā).

Example.

For a ṛṣi, the instrument: pūyate 'nena "one is purified by it" → pavitram.

Bhaṭṭoji: ṛṣi here means a Vedic mantra, since one finds "tad uktam ṛṣiṇā". One commentator names the Pāvamānī hymns.

For a deity, the agent: agniḥ pavitraṃ sa mā punātu "Agni is the purifier; may he purify me."

Applies when / blocked by. Bhaṭṭoji's pairing in order follows the Kāśikā. One commentator reports the Śabdakaustubha, which denies the pairing on the Bhāṣya's authority; scholars discuss this.

Watch out. In agniḥ pavitram, pavitram is neuter and singular, beside a masculine. One of Bhaṭṭoji's commentators says the word is used generically.

End of the prakaraṇa. With SK 3167 the pūrvakṛdanta ends. The printed text closes it with the words iti pūrvakṛdantaṃ samāptam "here the pūrvakṛdanta is concluded". The next chapter, chapter 68, is the Uṇādi list.

Terms used in this chapter

kṛdanta — a word ending in a kṛt affix.
kṛt — an affix added to a root that is not a verb-ending (SK 374).
kṛtya — the affixes of chapter 66, taught up to ṇvul (SK 2895); they come only in bhāva and karman.
kartṛ — the agent. By kartari kṛt (SK 2832), a kṛt denotes the agent unless a rule says otherwise.
upapada — a word that a kṛt rule names in the seventh case; the affix comes only when it stands beside the root (SK 781).
niṣṭhā — the past-participle affixes kta and ktavatu, taught in a later part of this chapter.
śatṛ / śānac — the affixes of the present participle, taught in a later part of this chapter.
tācchīlika — an affix meaning "having that as one's habit", taught in a later part of this chapter.
liṭ — the lakāra of the perfect; its participles come in a later part of this chapter.
it — a marker sound, written in a root or affix as taught but not heard: the ṇ and l of ṇvul (chapter 1).
adhikāra — a heading whose words are read into the rules that follow it, such as SK 2829 "after a dhātu".
anuvṛtti — carrying a word of an earlier sūtra forward into a later one that does not repeat it (chapter 1).
upadeśa — the original statement of a form, as it stands in the grammar before any marker is dropped (chapter 1).
guṇa — the name of the vowels a, e, o; also the replacement of a vowel by one of them, as in kartā.
vṛddhi — the name of the vowels ā, ai, au; also the replacement of a vowel by one of them, as in kārakaḥ.
dhātu — a root (chapter 1).
upasarga — a prefix such as pra, ā, sam, construed with a verb (chapter 1).
vibhāṣā — "optionally": the grammar's word for an option (chapter 1).
saṃprasāraṇa — the replacement of a semivowel by its vowel, as in grah (SK 2412).
yathāsaṃkhya — pairing items in two lists in order, first with first, second with second (SK 128).
utsarga / apavāda — the general rule or affix / the exception that sets it aside.
ṇvul — the general agent affix aka (from vu), with vṛddhi: kārakaḥ (SK 2895).
tṛc — the general agent affix tṛ, with guṇa: kartā (SK 2895).
iṭ — the augment i before certain consonant-initial affixes: kuṭitā, eṣitā / eṣṭā (SK 2895).
num — the augment n inserted in certain roots: randhakaḥ, maṅktā (SK 2895).
āna — the ending of the participles taught with śānac; SK 2401 and SK 2348 do not apply before it (SK 2895).
ātmanepada / parasmaipada — the two sets of verb-endings, "middle" and "active".
sthānivat — "like the original": a substitute or a lost sound that still counts as what it replaced (pāpacakaḥ, SK 2895).
yaṅ / yaṅluk — the intensive affix / the intensive whose yaṅ is deleted: pāpacakaḥ, pāpācakaḥ.
yogavibhāga — splitting one sūtra into two rules (SK 2895, 2909, 2916).
bahulam / bāhulaka — "variously": a rule so marked applies in other senses or after other bases (SK 2895, 2932, 2958).
vārttika / Bhāṣya — a supplementary statement on a sūtra / Patañjali's great commentary.
gaṇa — a list of roots or words, named by its first member: the nandi-, grahi- and paca-groups (SK 2896).
lyu — an agent affix ana (from yu), after the nandi-group: nandanaḥ.
ṇini — an agent affix in, with vṛddhi, after the grahi-group: grāhī.
ac — an agent affix a, with guṇa and no vṛddhi: pacaḥ, aṃśaharaḥ. (Also the name for any vowel.)
ākṛtigaṇa — an open list, whose members are recognised from usage (pacādi, SK 2896; mūlavibhujādi, SK 2919).
nyaṅkvādi — the list headed by nyaṅku; śvapāka is read in it, so śvapākaḥ is also correct (SK 2896).
abhyāsa — the first copy of a doubled root.
āk / ūk — augments ā and ū added to an abhyāsa: carācaraḥ, pāṭūpaṭaḥ.
ka — an affix a marked with k: no guṇa, and a root-final ā is lost: kṣipaḥ, godaḥ.
ik / igupadha — the vowels i, u, ṛ, ḷ / having one of them as penultimate (SK 2897).
vāsarūpa — SK 2830: an exception of different form blocks the general affix only optionally (kṣepakaḥ beside kṣipaḥ).
śa — an affix a marked with ś, so sārvadhātuka: pibaḥ, paśyaḥ.
sārvadhātuka / ārdhadhātuka — an affix that is a verb-ending or ś-marked / any other affix after a root.
śap — the class-sign a that comes before a sārvadhātuka.
Haradatta — the author of the Padamañjarī, a commentary on the Kāśikā; his views are reported at SK 2899, 2932 and 2944.
kvip — an affix of which nothing remains: sāt (SK 2900).
sautra — a root found only in the sūtras, such as sāti (SK 2900).
ṇi / ṇic — the causative affix, leaving i on the root: eji, tāpi, dāri.
ślu — the deletion of the class-sign śap after the roots of the hu class, dā and dhā among them; before it the root is doubled: dadaḥ (SK 2901).
ṇa — an affix a marked with ṇ, with vṛddhi: jvālaḥ, dāyaḥ.
yuk — the augment y after a root-final ā: dāyaḥ, gāyanaḥ (SK 2761).
sāhacarya — "association": a word is taken in the sense of the word it is paired with (SK 2904).
vyavasthita-vibhāṣā — an option with fixed domains: each form keeps one meaning (grāhaḥ / grahaḥ, SK 2905).
Kāśikā — a commentary on the sūtras that Bhaṭṭoji quotes at SK 2905 and SK 2958.
tātsthya — "being located in it": a word transferred to what is in the thing (gṛhāḥ "wife", SK 2906).
ṣvun — an agent affix aka for a craftsman; its ṣ gives ṅīṣ: nartakī (SK 2907).
śilpa / śilpin — a craft, skill in an action / one who has it.
kvun — an Uṇādi affix aka, giving rajikā on the Bhāṣya's view (SK 2907).
Uṇādi — a list of affixes, cited by number, that form words such as rajikā (Uṇādi 190).
thakan — the affix thaka after gai: gāthakaḥ (SK 2908).
ṇyuṭ — an affix ana, ṭ-marked: gāyanaḥ, gāyanī (SK 2909).
vun — an affix aka for skill and for blessing: lavakaḥ, jīvakaḥ (SK 2911–2912).
samabhihāra / sādhukāritva — repetition / doing an action well (SK 2911).
āśis — a blessing, the property of the speaker (SK 2912).
ṭāp / ṅīp / ṅīṣ — the feminine affixes ā / ī / ī; ṅīp comes after a ṭ-marked or aṇ-stem (SK 470).
aṇ — the general affix a with an object upapada; its ṇ gives vṛddhi: kumbhakāraḥ (SK 2913).
karman — the object.
upapada-samāsa — the compound of an upapada with a kṛdanta (SK 782): kumbhakāraḥ.
anabhidhāna — "non-expression": a form is not used because it would not express the sense (SK 2913).
śeṣa — a mere relation, intended in place of the object (gaṅgādhara, SK 2913).
ḍa — an affix a marked with ḍ, which removes the ṭi of the root: brahmajyaḥ, giriśaḥ (SK 2915, 2929).
ṣṭutva — SK 113: a dental next to ṣ becomes retroflex, as the th of dviṣṭhaḥ (SK 2918).
ṭi — the last vowel of a stem with whatever follows it.
bha — a technical name of a stem before certain affixes; the loss of ṭi normally needs it (SK 2915).
sup — any case-ending; a "sup-word" is a word ending in one (SK 2916).
bhāva — the bare action: ākhūtthaḥ (SK 2916).
ārambha-sāmarthya — "the force of a rule's being made": a rule that would be pointless in one sense is read in another (SK 2916).
nipātana — a finished form stated in a sūtra, whose irregular features are thereby taught (praṣṭhaḥ, phalegrahiḥ, vācaṃyamaḥ).
iṇ — a group of sounds: every vowel except a and ā, plus h, y, v, r, l. After these, or a velar, s normally becomes ṣ (SK 2917).
akarmaka — intransitive, taking no object; with such roots only supi is construed (SK 2918).
ṭak — an affix a marked with ṭ and k: sāmagaḥ, sāmagī (SK 2922).
udyamana — lifting up (SK 2923).
liṅgaviśiṣṭa-paribhāṣā — the principle that a stem includes its feminine form; shown not universal by ghaṭa, ghaṭī (SK 2923, 2955).
vayas — age (SK 2924).
tācchīlya — habit, having that as one's nature (SK 2925, 2934).
aluk — non-deletion of a case-ending inside a compound: stamberamaḥ (SK 2927).
saṃjñā — here, a name of a particular thing (SK 2899, 2900, 2928).
adhikaraṇa — the locus, "that in which": khaśayaḥ, kurucaraḥ.
lomādi — the list headed by loman, after which the taddhita śa comes: giriśa "who has a mountain" (SK 2929).
taddhita — an affix added to a noun, not to a root.
ṭa — an affix a marked with ṭ; it gives ṅīp: kurucarī, yaśaskarī (SK 2930–2936).
lyap — the affix of an absolutive such as ādāya "having taken" (SK 2931).
hetu / ānulomya — a cause / compliance (SK 2934).
dyotya — conveyed, signalled, not named by the affix (SK 2934).
kaska list — a list of compounds in which a visarga becomes s before k: bhāskaraḥ, ahaskaraḥ (SK 2935).
paratva — being the later rule, which wins in a conflict (SK 2935).
in — an affix i marked with n: stambakariḥ, ātmambhariḥ (SK 2938–2940).
mum — the augment m added to a pūrvapada before a khit uttarapada (SK 2942).
Cāndra — a follower of the Cāndra grammar, which reads SK 2940 differently.
khaś — an affix a marked with kh and ś: sārvadhātuka and khit: janamejayaḥ (SK 2941–2952).
pūrvapada / uttarapada — the first / the second member of a compound.
khit — marked with kh; before an uttarapada with such an affix the pūrvapada is shortened and takes mum.
avyaya — an indeclinable; it takes neither the short vowel nor mum (SK 2942–2943).
pararūpa — the single substitute of two vowels that takes the form of the second.
luk — deletion of an affix or ending; sup-luk is the loss of a case-ending inside a compound (SK 2942).
hrasva — a short vowel (SK 2943, 2955).
ṅit — marked with ṅ, or counting as so marked; before such an affix guṇa is blocked (SK 2217). A sārvadhātuka without p, such as khaś, counts as ṅit (SK 2234): abhraṃlihaḥ (SK 2947).
nañ — the negative particle, appearing as a-: asūrya (SK 2951).
asamartha-samāsa — a compound whose members are not connected in sense: asūrya- (SK 2951).
śyan — the class-sign ya of the div class; barred in iraṃmadaḥ (SK 2952).
khac — an affix a marked with kh and c: ārdhadhātuka and khit: priyaṃvadaḥ, dviṣantapaḥ (SK 2953–2960).
ḍit — marked with ḍ; khac may be ḍit: vihaṅgaḥ (SK 2953).
Vṛtti — an explanation of the sūtras, whose view one commentator reports at SK 2953.
vrata — a vow: vācaṃyamaḥ (SK 2956).
tadantavidhi — applying a rule to a word ending in the item named: abhayaṅkaraḥ (SK 2960, SK 26).
kṛñ — the root kṛ "to make", cited with its marker ñ.
pacādi — the class headed by pac, which takes the affix ac (SK 2896): kṣemakarāḥ.
karaṇa — the instrument: āśitaṃbhavaḥ odanaḥ, āḍhyaṃkaraṇam.
lyuṭ — a kṛt affix ana, the general affix of the instrument and the bare action: āśitabhavanam.
ghañ — a kṛt affix forming action nouns; blocked at SK 2962 as being of the same form as khac.
sāman — a chant of the Sāmaveda, such as rathaṃtara.
vyutpattimātra — a mere etymology: an analysis whose parts' meaning is not carried into the word (SK 2963).
avagraha — the pause that splits a compound in recitation.
sakarmaka / akarmaka — transitive / intransitive.
antarbhāvita-ṇyartha — having the causative sense taken in without the causative affix: dam, yudh, kṣi (SK 2963, 3005, 3015).
ṇi — the causative affix; dhāri is a ṇi-final root (SK 2963).
ḍittva-sāmarthya — "the force of the ḍ-marking": the ṭi goes even without bha, else the marker would be pointless (SK 2965).
kriyāviśeṣaṇa — a word that qualifies the action: panna in pannagaḥ.
upasaṃkhyāna — an addition to a rule, made by a vārttika.
prapañca — elaboration: a rule that only spells out what another already gives (SK 2965, 2983).
khal — a kṛt affix after su, dur and īṣat in the object sense (SK 3305): sugamaḥ.
āśīḥ — a wish, the condition of SK 2966.
ṅīp — the feminine affix ī after a ṭ-marked stem: patighnī.
arśaādi — the class headed by arśas, which takes ac in the sense "having": lakṣaṇa "having a mark" (SK 1933).
gauṇa — figurative: the "wife-killer" of SK 2969 has only a mark presaging the death.
mūlavibhujādi — a list of words formed with ka by the vārttika under SK 2919: śatrughnaḥ.
bahulaka — "by the word bahula": the latitude of SK 2841, by which a kṛt may come beyond its taught range: coraghātaḥ.
nānubandhakṛtam asārūpyam — "a difference made by markers is no difference of form": so ṭak and aṇ count as the same form (SK 2970).
śakti — capability (SK 2971).
śilpin — an artisan (SK 2972).
cvi — an affix added to a noun in the sense of becoming what one was not, giving āḍhyī- in āḍhyīkaroti.
abhūtatadbhāva — becoming what one was not; the cvi sense (SK 2973).
khyun — a kṛt affix yu → ana, with mum, in the instrument sense: āḍhyaṃkaraṇam.
apavāda — an exception that sets aside a general rule or affix.
Kāśikā / Vṛtti — an early commentary on the Aṣṭādhyāyī; Bhaṭṭoji often reports and sometimes rejects its views.
Bhāṣya — Patañjali's great commentary on the Aṣṭādhyāyī; Bhaṭṭoji follows it against the Vṛtti (SK 2988).
khiṣṇuc / khukañ — kṛt affixes after bhū in the cvi sense: āḍhyaṃbhaviṣṇuḥ, āḍhyaṃbhāvukaḥ (SK 2974).
kvin / kañ / ksa — affixes after spṛś and dṛś, recalled at SK 2974: ghṛtaspṛk, sadṛśaḥ, tādṛkṣaḥ.
karmakartṛ — an object that acts as its own agent: sadṛk, sino grāsaḥ.
ṇvi — a kṛt affix that leaves nothing but its vṛddhi: aṃśabhāk.
viṭ — a kṛt affix that leaves nothing, after ad: āmāt, kravyāt.
apṛkta — an affix of a single sound; the v left by kvip, ṇvi, viṭ and vic is one, and is lost (SK 375).
kap — a kṛt affix after duh, with gh for the root's final: kāmadughā.
tuk — the augment t after a short vowel before a p-marked kṛt (SK 2858): sukṛt, aṅgagat.
ṇatva — the change of n to ṇ: agraṇīḥ, he prāṇ.
Amarakośa — a Sanskrit thesaurus; Bhaṭṭoji quotes it at SK 2978 and SK 3009.
niyama — a restriction: a rule that repeats what another gives is read as limiting it (SK 2978, 2998–3004).
pṛṣodarādi — a class of words with irregular forms admitted as found in use: kravya for pakvamāṃsa.
manin / kvanip / vanip / vic — Vedic affixes freed for all roots by SK 2980: suśarmā, prātaritvā, vijāvā, roṭ.
vaś — a pratyāhāra that includes m and v; no iṭ before a kṛt beginning with it (SK 2981).
tadādi-vidhi — reading a sound in a rule as "beginning with that sound" (SK 2981).
savarṇadīrgha — the merging of two like vowels into one long vowel: vijāvā.
kvi — kvip and similar affixes named together, before which SK 2985 and SK 2986 work.
ūṭh — the substitute ū for v or for a vowel and v: akṣadyūḥ, jūḥ (SK 2561, SK 2654).
jhal — a consonant that is not a nasal or a semivowel; some rules apply only before an affix beginning with one.
jāti — a genus, such as brāhmaṇa, kṣīra, uṣṭra (SK 2988).
tṛn — a kṛt affix of the habitual agent: āmantrayitā.
navya — "the moderns", later grammarians reported by a commentator (SK 2988).
upamāna — the standard of comparison (SK 2989).
ābhīkṣṇya / paunaḥpunya — repetition, "again and again" (SK 2991).
sārvadhātuka — the name of verb-endings and ś-marked affixes; before one the class sign comes.
ātmamāna — thinking that has oneself as its object (SK 2993).
puṃvadbhāva — the use of the masculine form of a stem for the feminine (SK 831); blocked in kāliṃmanyā.
anavyaya / avyaya — not an indeclinable / an indeclinable: divā (SK 2993).
ic — the vowels other than a and ā.
ekāc — having a single vowel (SK 2994, SK 3015).
am — the augment of SK 2994, treated like the case-affix am: gāṃmanyaḥ.
bhūte — "in the past": the heading of SK 2995, up to SK 2151.
vyākhyāna — explanation, the authority for where a heading stops (SK 2995).
svarita — the mark by which a word runs on into later rules: ṇini into SK 2996.
kutsita — contemptible; added by a vārttika at SK 3003, and by the Kāśikā at SK 2997.
ini — a kṛt affix in without vṛddhi: somavikrayī (SK 3003).
Śabdakaustubha — a grammatical work cited by a commentator at SK 3005.
kāraka — a participant in the action: agent, object, instrument, locus and others (SK 3011).
sarvopādhivyabhicāra — the lapse of every condition, granted by api in SK 3011: parikhā.
kta / ktavatu — the niṣṭhā affixes ta and tavat: stutaḥ, kṛtavān.
anyonyāśraya — mutual dependence of two things, each needing the other first (SK 3012).
bhāvinī saṃjñā — a name that "will be": given to an affix once it has come (SK 3012).
niṣṭhā-t — the t of ta or tavat, which SK 3015–3024 replace with n.
Tripādī — the last three quarter-chapters (*pāda*s) of the eighth book, A 8.2–8.4; its rules are asiddha to earlier ones.
asiddha — treated as not having taken effect: the niṣṭhā-n, and a bahiraṅga vṛddhi (SK 3016).
kutva — the change of c or j to a velar: bhugnaḥ, samaknaḥ.
śryukaḥ kiti — SK 2381: no iṭ before a k-marked affix after śri and the uk-final roots: śritaḥ.
uk — the vowels u, ṛ, ḷ, short or long.
bahiraṅga — conditioned by something outside the operation in view; such a change is asiddha to it (SK 3016).
saṃyoga — a cluster of consonants (SK 3017).
yaṇ — the semivowels y, v, r, l.
pvādi / lvādi — the twenty-two kryādi roots from pūñ / the same twenty-one without pūñ, beginning with lūñ: lūnaḥ.
nitya — obligatory.
gaṇasūtra — a rule stated within the root list: svādaya oditaḥ.
odit — a root marked with o: bhujo, ṭuośvi, ohāk (SK 3019).
pāṭhasāmarthya — the force of a word's being read in a list: ḍīṅ among the odit roots (SK 3019).
dravamūrti / sparśa — the hardening of a liquid / touch, here the cold touch (SK 3020–3021).
vyavasthitavibhāṣā — a fixed option, holding only in fixed cases (SK 3023).
apādāna — the point of departure: kūpāt in udaktam udakaṃ kūpāt.
ktvā — the kṛt affix of the absolutive, taught in a later chapter (SK 3025).
ārdhadhātuka — the name of an affix after a root other than a verb-ending or a ś-marked affix.
ṣatva — the change of s (or ś) to ṣ: withheld in vṛkṇaḥ, pariskandaḥ, pratistabdhaḥ.
siddha — counting as having taken effect: the niṣṭhā substitute, for four operations (SK 3025).
prācyabharata — the eastern Bharatas, among whom pariskandaḥ is used (SK 3026).
karaṇa / hetu — the instrument / the cause of an action; neither is the agent.
pūrvarūpa — the merging of a following vowel into the one before it: sti-ā → sti.
ādit / īdit / udit / ūdit / odit / ñīt — a root whose marker is ā / ī / u / ū / o / ñi. The marker decides iṭ and n in the niṣṭhā.
seṭ / aniṭ / veṭ — taking iṭ / not taking it / taking it optionally. Said of roots and of affixes.
kit — marked with k. Before a kit affix the root takes no guṇa (SK 2217).
maṇḍūkapluti — "frog's leap": carrying a word forward past the rules between.
rūḍha — fixed by usage, without regard to the sense of the parts: bhitta "fragment".
lakṣaṇā — secondary signification: a word standing for something linked to its own sense.
yasya vibhāṣā — SK 3025: a root that is optionally seṭ anywhere takes no iṭ before a niṣṭhā. It holds only for one-vowelled roots.
punarvidhi — a renewed prescription: a rule that says again what another rule gave, so that a ban cannot remove it.
yaṅluk — the intensive stem whose affix yaṅ has been elided: śeśyitaḥ.
ādikarman — the beginning of an action: prakṛtaḥ kaṭaṃ saḥ "he has begun the mat".
bhāva / karman / kartṛ — the bare action / the object / the agent, the three senses a kta may denote.
takrakauṇḍinya-nyāya — the maxim of the buttermilk and Kauṇḍinya: naming one item specially shuts out the others.
upadhā — the penultimate sound of a stem.
vikaraṇa / śap — the class sign between root and ending / the class sign a of the first class.
ṇic / ṇi — the causative affix / its remainder i after the markers go: bhāvi.
jñāpaka — a feature of the grammar that makes known an unstated principle.
anitya — not binding: a rule or teaching that may fail to apply.
āgama-śāstra — the teaching of augments, such as iṭ; Bhaṭṭoji calls it not binding (SK 3058).
luk / lopa — two kinds of elision. After lopa the stem still counts the lost affix (pratyayalakṣaṇa); SK 3066 needs luk.
pratyayalakṣaṇa — the principle that a lost affix still has its effect on the stem.
mit — a root that shortens its vowel before ṇi (SK 2568): damitaḥ.
ātmanepada — the middle-voice endings.
svāṅga — a part of one's own body.
ti kiti — "before a t-initial kit affix", the condition carried from SK 3074 through group III.K.
śtip — the ending -ti used to cite a root, as dyati, dadhāti.
ghu — the name of the roots dā and dhā; dāp and daip are not ghu.
cartva — the plain voiceless stop put for a stop before a voiceless sound (khari ca, SK 121): dath-ta → datta.
sannipāta maxim — a change made because of a combination cannot destroy that combination.
uttarapada — the second member of a compound or of an upasarga-word.
āśrayāt siddham — "counted as done because it is the basis": a Tripādī change seen by a rule that depends on it (SK 3079).
mūrdhanya — retroflex; apadāntasya mūrdhanyaḥ (SK 210) is the heading of the ṣ rules.
gotra — a line of descent; a gotra name marks a descendant.
matvarthīya — an affix in the sense "having it": bhuktāḥ "having eaten".
vartamāne / bhūte — the headings "in the present" (SK 2151) and "in the past" (SK 2995).
kālasāmānya — time in general, with no tense fixed (SK 3090).
ṅvanip — a past agent affix -van after su and yaj: sutvā, yajvā.
atṛn — a past affix -at after jṝ: jaran.
ugit — an affix marked with u, ṛ or ḷ; it brings num (SK 361).
lakāra — a tense-and-mood affix whose name begins with l, such as laṭ (present), liṭ, luṅ and lṛṭ (future). A verb-ending, or here a participle affix such as kvasu or śatṛ, replaces it.
bhūte / bhūtasāmānya — "in the past", the heading of SK 2995 / the past in general, without the conditions of SK 2171.
anadyatana / parokṣa — not of today / out of the speaker's sight.
chandas / bhāṣā — the Veda / ordinary speech.
dhātusaṃbandha — a connection between verbs; SK 3423 needs it, SK 3093 does not.
kānac — an affix that replaces liṭ, leaving āna: cakrāṇa, anūcānaḥ (SK 3094, SK 3098).
lādeśa — a substitute of a lakāra, such as kānac, kvasu, śatṛ and śānac; the option of SK 2830 does not hold among them.
anantarasya — the maxim "of the nearest": a word refers to the nearest item it can (SK 3094, SK 3100).
kvasu / vasu — the affix that replaces liṭ, leaving vas (SK 3095) / its name without the k, used in SK 3096, SK 3099 and SK 3105.
trimunimatam — "the view of the three sages" of grammar; Bhaṭṭoji uses it for the view that kvasu and kānac are Vedic only (SK 3095).
luṅ — the lakāra of the past in general; it comes in the other option of SK 3097, and after māṅ (SK 2219).
halādiśeṣa — the rule (SK 2179) by which only the first consonant of the abhyāsa stays.
kṛādi-niyama — a restriction that brings iṭ back where SK 2981 had barred it; SK 3096 then restricts it again.
laṭ — the lakāra of the present, taught by vartamāne laṭ (SK 2151).
sāmānādhikaraṇya — co-reference: two words referring to the same thing.
prathamānta / aprathamānta — a word ending in the first case / one not ending in it.
muk — the augment m given to an a before āna (SK 3101): pacamānam, yajamānaḥ.
adhikavidhāna — prescribing more than the wording states; the restated laṭ of SK 3100 does this: san brāhmaṇaḥ (SK 3101).
māṅ / ākrośa — the prohibitive particle mā / reviling (vārttika at SK 3101).
sambodhana — address, the vocative (SK 3102).
prapañcārtha — "for elaboration": a rule that only spells out what is already obtained (SK 3102).
lakṣaṇa / hetu — a mark that makes an action known / a cause, covering both the result aimed at (phala) and the producing cause (kāraṇa) (SK 3103).
anyatarasyām — "optionally", carried over from SK 2197 into SK 3105.
sat — the name of śatṛ and śānac (SK 3106).
lṛṭ — the lakāra of the future, taught by lṛṭ śeṣe ca (SK 2193); sat may replace it (SK 3107).
śānan — an independent present affix after pū and yaj, leaving āna: yajamānaḥ (SK 3108).
pratyāhāra — a short name that takes in a whole stretch of a list, from its first item to a final marker. The "tṛn" of SK 627 runs from the t of śatṛ to the n of tṛn (SK 3108, SK 3115).
cānaś — an independent affix leaving āna, for disposition, age or capacity: bhogaṃ bhuñjānaḥ (SK 3109).
tācchīlya / vayas / śakti — habitual disposition / age / capacity (SK 3109).
akṛcchrin — one for whom the action is without difficulty (SK 3110).
sattra / yajamāna — a long Vedic rite with many masters / the master of a sacrifice (SK 3112).
tacchīla / taddharman / tatsādhukārin — one whose habit is the action / whose duty it is / who does it well (SK 3114).
abhividhi — inclusion of the limit: the ā of ā kveḥ includes kvip (SK 3114).
iṣṇuc — an affix of habit leaving iṣṇu: alaṃkariṣṇuḥ, bhaviṣṇuḥ (SK 3116–3118).
niraṅkuśāḥ kavayaḥ — "poets are unrestrained": Bhaṭṭoji's answer to a non-Vedic aprabhaviṣṇu (SK 3118).
gsnu — an affix leaving snu, marked with g: jiṣṇuḥ, sthāsnuḥ, bhūṣṇuḥ (SK 3119).
praśleṣa — reading an extra sound into a rule by coalescence: a g in SK 2217 and SK 2381 (SK 3119).
knu — an affix leaving nu, marked with k: gṛdhnuḥ (SK 3120).
ghinuṇ — an affix leaving in; its gh brings the velar, its ṇ vṛddhi: śamī, saṃparkī (SK 3121–3125).
anubandha — a marker; on the Bhāṣya's view the u of ghinuṇ is one (SK 3121).
apakarṣa — drawing a word back from a later sūtra: jhal from SK 314 (SK 3121), dṛśi from SK 3158 (SK 3157).
prāyika — generally valid, with exceptions (SK 3121, SK 3133).
vuñ — an affix of habit leaving vu → aka: nindakaḥ, ādevakaḥ (SK 3126–3127).
kaṇḍu list — a list of roots that take the affix yak, such as asūy (SK 3126).
yuc — an affix of habit leaving yu → ana: calanaḥ, vartanaḥ (SK 3128–3133).
anudātteta — a root whose marker is an anudātta vowel (SK 3129).
halādi — beginning with a consonant (SK 3129).
Dhātupāṭha — the traditional list of roots, with their meanings and markers (SK 3130, SK 3147).
yaṅ — the intensive affix; caṅkramya, yāyajya and yāyāya end in it (SK 3130, SK 3146, SK 3156).
parasparam — "mutually": the affixes of habit do not alternate with one another (SK 3130).
tadanta — reading a sound as "a root ending in it": yaḥ means "ending in y" (SK 3132).
nandi list / lyu — the list of SK 2896, whose roots take lyu (ana): madhusūdanaḥ (SK 3133).
ra — an affix of habit: namraḥ, ajasram, dīpraḥ (SK 3147).
ukañ — an affix of habit leaving uka, with vṛddhi: pātukaḥ (SK 3134).
ṣākan — an affix of habit leaving āka; its ṣ brings ṅīṣ: jalpākaḥ, varākī (SK 3135).
āluc — an affix of habit leaving ālu: spṛhayāluḥ, śayāluḥ (SK 3138).
adanta — ending in a; said of the cur-class roots spṛha, gṛha, pata (SK 3138).
ru / kmarac / ghurac / kurac — affixes of habit leaving ru / mara / ura / ura: dāruḥ, ghasmaraḥ, bhaṅguraḥ, viduraḥ (SK 3139–3142).
kvarap — an affix of habit leaving vara; k bars guṇa, p brings tuk: itvaraḥ, naśvaraḥ (SK 3143–3144).
ūka — an affix of habit: jāgarūkaḥ, yāyajūkaḥ (SK 3145–3146).
nuk — the augment n on the reduplicated syllable of certain intensives (SK 2638): jañjapūkaḥ.
u — an affix of habit: cikīrṣuḥ, bhikṣuḥ, sumnayuḥ (SK 3148–3150).
san — the desiderative affix: cikīrṣa (SK 3148).
garga list — a list of words; reading vijigīṣu in it shows that san in SK 3148 is the affix.
kya — the denominative affixes kyac, kyaṅ and kyaṣ: sumnaya (SK 3150).
ki / kin — affixes treated like liṭ, differing only in accent: papiḥ, cakriḥ (SK 3151).
liḍvat — "like liṭ": the root undergoes what it would before liṭ, such as reduplication (SK 3151).
nīk — an augment of the reduplicated syllable that pāpati lacks by nipātana (SK 3151).
najiṅ — an affix of habit leaving naj: svapnak, tṛṣṇak (SK 3152).
āru — an affix of habit: śarāruḥ, vandāruḥ (SK 3153).
kru / klukan / krukan — affixes after bhī: bhīruḥ, bhīlukaḥ, bhīrukaḥ (SK 3154).
varac — an affix of habit leaving vara: sthāvaraḥ, īśvaraḥ, yāyāvaraḥ (SK 3155–3156).
puṃyoga — the feminine for "the wife of": īśvarī (SK 3155).
vare luptaṃ na sthānivat — "a loss before vara is not treated as the original", a clause of SK 51 (SK 3156).
dṛśi — the word "is seen" (dṛśyate) in SK 3158, which takes in extra operations.
ḍu — an affix leaving u; its ḍ brings loss of the ṭi: prabhuḥ (SK 3160).
ṣṭran — an affix leaving tra, in the sense of the object or the instrument: dhātrī, dātram, śastram (SK 3161–3164).
aja list — the list of words headed by aja, whose feminine takes ṭāp (ā); some explain daṃṣṭrā through it (SK 3163).
itra — an affix in the sense of the instrument, or of the agent for a deity: khanitram, pavitram (SK 3165–3167).
ṛṣi / devatā — here a Vedic mantra / a deity (SK 3167).

Summary

SK	what it does	example
2895	ṇvul and tṛc after any root, in the agent sense; verb rules shown on them; bahulam split off; vārttika: no iṭ after kram in ātmanepada	`kārakaḥ`, `kartā`; `pādahārakaḥ`; `prakrantā`
2896	lyu, ṇini, ac after the nandi-, grahi- and paca-groups; reduplicated forms by vārttika	`nandanaḥ`, `grāhī`, `pacaḥ`; `carācaraḥ`
2897	ka after an igupadha root and after jñā, prī, kṝ; ṇvul/tṛc beside it	`kṣipaḥ`, `jñaḥ`, `priyaḥ`, `kiraḥ`; `kṣepakaḥ`
2898	ka after an ā-final root with an upasarga, against ṇa	`suglaḥ`, `prajñaḥ`
2899	śa after pā, ghrā, dhmā, dheṭ, dṛś; not after ghrā in a name	`pibaḥ`, `jighraḥ`, `dhamaḥ`, `paśyaḥ`; `dhayā kanyā`
2900	śa after limp, vind, dhāri … sāhi without an upasarga; vārttikas ni + limp, go + vind	`limpaḥ`, `dhārayaḥ`; `sāt`; `nilimpāḥ`, `govindaḥ`
2901	śa optionally after dā, dhā without an upasarga	`dadaḥ`, `dadhaḥ` / `dāyaḥ`; `pradaḥ`
2902	ṇa optionally after jval … kas without an upasarga; tan always	`jvālaḥ` / `jvalaḥ`; `ujjvalaḥ`; `avatānaḥ`
2903	ṇa invariably after śyā, ā-finals, vyadh … śvas	`dāyaḥ`, `vyādhaḥ`, `avaśyāyaḥ`, `śvāsaḥ`
2904	ṇa after du (dunoti) and nī without an upasarga	`dāvaḥ`, `nāyaḥ`; `davaḥ`; `praṇayaḥ`
2905	ṇa optionally after grah, with fixed meanings	`grāhaḥ` (aquatic) / `grahaḥ` (planet)
2906	ka after grah when the agent is a house	`gṛham`; `gṛhāḥ` "wife"
2907	ṣvun for a craftsman (only nṛt, khan, rañj)	`nartakaḥ`, `nartakī`, `rajakaḥ`
2908	thakan after gai, for a craftsman	`gāthakaḥ`
2909	ṇyuṭ after gai, for a craftsman	`gāyanaḥ`, `gāyanī`
2910	ṇyuṭ after hā when rice or time is the agent	`hāyano vrīhiḥ`, `hāyano varṣam`
2911	vun after pru, sṛ, lū for repetition, i.e. skill	`pravakaḥ`, `sarakaḥ`, `lavakaḥ`
2912	vun in a blessing	`jīvakaḥ`, `nandakaḥ`
2913	aṇ with an object upapada; the upapada compound; vārttika ṇa after śīli etc.	`kumbhakāraḥ`; `māṃsaśīlā`
2914	aṇ after hvā, vā, mā, against ka	`svargahvāyaḥ`, `tantuvāyaḥ`
2915	ka after an ā-final root without an upasarga, with an object upapada; vārttika ḍa	`godaḥ`; `gosaṃdāyaḥ`; `brahmajyaḥ`, `āhvaḥ`
2916	split: ka after ā-finals with any sup-word; ka after sthā in bhāva	`dvipaḥ`, `samasthaḥ`; `ākhūtthaḥ`
2917	praṣṭha with ṣ for "going in front"	`praṣṭhaḥ` / `prasthaḥ`
2918	ṣ for the s of stha after eighteen words; from here karmaṇi and supi run on	`dviṣṭhaḥ`, `triṣṭhaḥ`
2919	ka after parimṛj, apanud with tunda, śoka (laziness, happiness); vārttika mūlavibhujādi	`tundaparimṛjaḥ`, `śokāpanudaḥ`; `mūlavibhujaḥ`
2920	ka after dā-roots and jñā with pra	`sarvapradaḥ`, `pathiprajñaḥ`; `gosaṃpradāyaḥ`
2921	(ka) after khyā with sam	`gosaṃkhyaḥ`
2922	ṭak after gai, pā without an upasarga; pā only with surā, śīdhu	`sāmagaḥ`, `sāmagī`; `surāpī`; `kṣīrapā`
2923	ac after hṛ not meaning lifting; vārttikas grah	`aṃśaharaḥ` / `bhārahāraḥ`; `śaktigrahaḥ`, `sūtragrahaḥ`
2924	ac after hṛ for lifting, when age is conveyed	`kavacaharaḥ kumāraḥ`
2925	ac after ā + hṛ for habit	`puṣpāharaḥ`
2926	ac after arh, against aṇ, for the feminine in ā	`pūjārhā brāhmaṇī`
2927	(ac) after ram, jap with stamba, karṇa, for an elephant, an informer	`stamberamaḥ`, `karṇejapaḥ`
2928	ac after any root with śam, for a name, even against ṭa	`śambhavaḥ`, `śaṃvadaḥ`, `śaṃkarā`
2929	ac after śī with a locus; vārttikas pārśva etc., uttāna etc.; ḍa with giri (Veda)	`khaśayaḥ`, `pārśvaśayaḥ`, `uttānaśayaḥ`, `giriśaḥ`
2930	ṭa after car with a locus	`kurucaraḥ`, `kurucarī`
2931	ṭa after car with bhikṣā, senā, ādāya	`bhikṣācaraḥ`, `ādāyacaraḥ`; `sahacarī` (pacādi caraṭ)
2932	ṭa after sṛ with puras, agratas, agre	`purassaraḥ`, `agresaraḥ`
2933	ṭa after sṛ with pūrva as agent	`pūrvasaraḥ` / `pūrvasāraḥ`
2934	ṭa after kṛ when cause, habit or compliance is conveyed	`yaśaskarī vidyā`, `śrāddhakaraḥ`, `vacanakaraḥ`
2935	ṭa after kṛ with twenty-seven words; vārttika ac with kim, yad, tad, bahu	`divākaraḥ`, `bhāskaraḥ`, `dhanuṣkaraḥ`; `kiṃkarā`
2936	ṭa after kṛ with the word karman, for wages	`karmakaraḥ` / `karmakāraḥ`
2937	no ṭa with śabda, śloka … pada	`śabdakāraḥ`
2938	in after kṛ with stamba, śakṛt, when rice or a calf is the agent	`stambakarir vrīhiḥ`, `śakṛtkarir vatsaḥ`
2939	in after hṛ with dṛti, nātha, when an animal is the agent	`dṛtihariḥ`, `nāthahariḥ`
2940	nipātana phalegrahi, ātmambhari; ca adds kukṣimbhari	`phalegrahiḥ`, `ātmambhariḥ`, `kukṣimbhariḥ`
2941	khaś after eji (causative ej)	`janamejayaḥ` (SK 2942)
2942	mum for arus, dviṣat, a vowel-final non-avyaya before a khit uttarapada	`janamejayaḥ`; `vātamajāḥ`
2943	short vowel for the final of a non-avyaya pūrvapada before a khit uttarapada; then mum	`śunindhayaḥ`, `tilantudaḥ`
2944	khaś after dheṭ with stana, dhmā/dheṭ with nāsikā	`stanandhayaḥ`, `stanandhayī`, `nāsikandhamaḥ`
2945	khaś after dhmā, dheṭ with nāḍī, muṣṭi (and ghaṭī, khārī, kharī)	`nāḍindhamaḥ`, `muṣṭindhayaḥ`, `ghaṭindhamaḥ`
2946	khaś after ut + ruj, vah with kūla	`kūlamudrujaḥ`, `kūlamudvahaḥ`
2947	khaś after lih with vaha, abhra (no guṇa)	`vahaṃlihaḥ`, `abhraṃlihaḥ`
2948	khaś after pac with a measure	`prasthampacā`, `khārimpacaḥ`
2949	khaś after pac with mita, nakha	`mitampacā`, `nakhampacā`
2950	khaś after tud with vidhu, arus	`vidhuntudaḥ`, `aruntudaḥ`
2951	khaś after dṛś, tap with asūrya, lalāṭa	`asūryampaśyāḥ`, `lalāṭantapaḥ`
2952	nipātana ugrampaśya, irammada, pāṇindhama	`ugraṃpaśyaḥ`, `iraṃmadaḥ`, `pāṇiṃdhamaḥ`
2953	khac after vad with priya, vaśa; vārttikas gam, viha, optional ḍit	`priyaṃvadaḥ`, `mitaṅgamaḥ`, `vihaṅgaḥ`
2954	khac after tāpi with dviṣat, para	`dviṣantapaḥ` (SK 2955)
2955	short vowel for the penultimate before ṇi + khac	`dviṣantapaḥ`, `paraṃtapaḥ`; `dviṣatītāpaḥ`
2956	khac after yam with vāc, for a vow	`vācaṃyamaḥ` (SK 2957)
2957	nipātana: the am of vāc, pur	`vācaṃyamaḥ` / `vāgyāmaḥ`; `puraṃdaraḥ`
2958	khac after dāri with pur, sah with sarva	`puraṃdaraḥ`, `sarvaṃsahaḥ`; `bhagaṃdaraḥ`
2959	(khac) after kaṣ with sarva, kūla, abhra, karīṣa	`sarvaṅkaṣaḥ`, `kūlaṅkaṣā`
2960	(khac) after kṛ with megha, ṛti, bhaya (and words ending in bhaya)	`meghaṅkaraḥ`, `bhayaṃkaraḥ`, `abhayaṅkaraḥ`
2961	aṇ and khac after kṛñ with kṣema, priya, madra	`kṣemaṅkaraḥ` / `kṣemakāraḥ`
2962	khac after bhū with āśita, in instrument or bare action	`āśitaṃbhavaḥ odanaḥ`
2963	khac in names after bhṛ, tṝ, vṛ, ji, dhāri, sah, tap, dam	`viśvambharaḥ`, `ariṃdamaḥ`
2964	khac after gam in a name	`sutaṃgamaḥ`
2965	ḍa after gam with anta etc.; ṭi lost by the ḍ-marker; vārttikas	`antagaḥ`, `uragaḥ`, `vihagaḥ`
2966	ḍa after han when a wish is meant; vārttikas on -ghāṭa	`śatruhaḥ`; `dārvāghāṭaḥ`
2967	ḍa after apa-han with kleśa, tamas, without a wish	`tamopahaḥ sūryaḥ`
2968	ṇini after han with kumāra, śīrṣa; śīrṣa by nipātana	`kumāraghātī`, `śīrṣaghātī`
2969	ṭak after han with jāyā, pati, the agent bearing a mark	`jāyāghno nā`, `patighnī`
2970	ṭak after han with a non-human agent	`pittaghnaṃ ghṛtam`
2971	ṭak after han with hastin, kapāṭa, for capability	`hastighno nā`
2972	nipātana pāṇigha, tāḍagha for an artisan; vārttika rājagha	`pāṇighaḥ`, `rājaghaḥ`
2973	khyun after kṛñ in the cvi sense, instrument	`āḍhyaṃkaraṇam`
2974	khiṣṇuc, khukañ after bhū in the cvi sense; recall of SK 432, 373, 429	`āḍhyaṃbhaviṣṇuḥ`; `sadṛk`, `tādṛkṣaḥ`
2975	kvip after sad, dviṣ and ten more, with or without upasarga	`mitradviṭ`, `pradviṭ`; `agraṇīḥ`
2976	ṇvi after bhaj	`aṃśabhāk`
2977	viṭ after ad, with an upapada other than anna	`āmāt`; but `annādaḥ`
2978	viṭ after ad with kravya, blocking aṇ	`kravyāt`
2979	kap after duh, with gh for its final	`kāmadughā`
2980	manin, kvanip, vanip, vic after any root	(see 2981–2982)
2981	no iṭ before a kṛt beginning with vaś	`suśarmā`, `prātaritvā`
2982	ā for a nasal before viṭ and van	`vijāvā`, `avāvā`; `roṭ`
2983	kvip after any root	`ukhāsrat`, `vāhabhraṭ`
2984	ṇ for the pada-final n of an after an upasarga; more kvip forms	`he prāṇ`; `mitraśīḥ`, `gīḥ`
2985	short vowel for the ā of chādi before is, man, tran, kvi; more kvip forms	`tanucchat`; `akṣadyūḥ`, `janauḥ`
2986	nasal of gam lost before kvi; vārttikas	`aṅgagat`; `agregūḥ`
2987	ka and kvip after sthā, blocking ac	`śaṃsthaḥ`, `śaṃsthāḥ`
2988	ṇini for habit, upapada not a genus; vārttikas	`uṣṇabhojī`; `sādhukārī`, `brahmavādī`
2989	ṇini when the upapada's agent is the standard of comparison	`uṣṭrakrośī`
2990	ṇini for a vow	`sthaṇḍilaśāyī`
2991	ṇini, variously, for repetition	`kṣīrapāyiṇa uśīnarāḥ`
2992	ṇini after divādi man	`darśanīyamānī`
2993	khaś (and ṇini) after man when one thinks of oneself	`paṇḍitaṃmanyaḥ`, `kāliṃmanyā`
2994	am after a one-vowel ic-final first member before a khit	`gāṃmanyaḥ`, `naraṃmanyaḥ`; `śrimanyam`
2995	heading: "in the past", up to SK 2151	`somayājī`
2996	ṇini after yaj with an instrument upapada	`somayājī`
2997	ṇini after han with an object upapada; heading karmaṇi to SK 3006	`pitṛvyaghātī`
2998	kvip after han with brahman, bhrūṇa, vṛtra; a restriction	`brahmahā`, `vṛtrahā`
2999	kvip after kṛñ with su, karman etc.; threefold restriction	`sukṛt`, `śāstrakṛt`
3000	kvip after suñ with soma; fourfold restriction	`somasut`
3001	kvip after ci with agni	`agnicit`
3002	kvip after ci in the object sense, naming a fire-site	`śyenacit`
3003	ini after vi-krī with an object; vārttika: contemptible object	`somavikrayī`
3004	kvanip after dṛś, in the past; a restriction	`pāradṛśvā`
3005	kvanip after yudh, kṛñ with rājan	`rājayudhvā`, `rājakṛtvā`
3006	the same with saha; karmaṇi stops	`sahayudhvā`, `sahakṛtvā`
3007	ḍa after jan with a seventh-case upapada	`sarasijam`, `mandurajaḥ`
3008	ḍa after jan with a fifth-case upapada, not a genus	`saṃskārajaḥ`
3009	ḍa after jan with an upasarga, in a name	`prajā`
3010	ḍa after anu-jan with an object	`pumanujā`
3011	ḍa after jan seen elsewhere; every condition may lapse	`dvijaḥ`, `ajaḥ`; `parikhā`
3012	name: kta and ktavatu are niṣṭhā	`snātam`, `kṛtavān`
3013	a niṣṭhā in the past: kta in bhāva/karman, ktavatu in the agent	`snātaṃ mayā`, `stutaḥ`, `kṛtavān`
3014	long vowel for kṣi before a niṣṭhā in the agent	`kṣī-tavat`
3015	n for the niṣṭhā-t after long kṣī; recall of SK 2381; vārttika ūrṇu	`kṣīṇavān`; `kṣitaḥ`; `ūrṇutaḥ`
3016	n for the niṣṭhā-t after r or d, and for the d	`bhinnaḥ`, `śīrṇaḥ`; not `kārtiḥ`
3017	n after the ā of a cluster-initial root with a semivowel	`glānaḥ`, `drāṇaḥ`
3018	n after the lvādi; vārttikas du, gu, pūñ, si	`lūnaḥ`, `jīnaḥ`; `dūnaḥ`, `pūnāḥ`
3019	n after an o-marked root	`bhugnaḥ`, `ucchūnaḥ`, `uḍḍīnaḥ`
3020	saṃprasāraṇa of śyaiṅ before a niṣṭhā for hardening or touch	`śi-ta`
3021	n after śyaiṅ except for (cold) touch	`śīnaṃ ghṛtam` / `śītaṃ jalam`
3022	saṃprasāraṇa of śyaiṅ after prati	`pratiśīnaḥ`
3023	the same, optionally, after abhi, ava; a fixed option	`abhiśyānam` / `abhiśīnam`
3024	n after añc, except with an apādāna	`samaknaḥ`; `udaktam … kūpāt`
3025	no iṭ before a niṣṭhā after a root with optional iṭ; vārttika: the substitute siddha for ṣatva etc.	`samaknaḥ`; `vṛkṇaḥ`
3026	nipātana pariskanda without ṣ among the eastern Bharatas	`pariskandaḥ` / `pariṣkandaḥ`
3027	no ṣ in pratistabdha, nistabdha	`pratistabdhaḥ`, `nistabdhaḥ`
3028	after div, n for the niṣṭhā-t, except in desire to conquer	`dyūnaḥ`; but `dyūtam`
3029	after nir-vā, n, unless wind is the agent	`nirvāṇo 'gniḥ`; but `nirvāto vātaḥ`
3030	after śuṣ, k	`śuṣkaḥ`
3031	after pac, v; the v is asiddha for kutva	`pakvaḥ`
3032	after kṣai, m	`kṣāmaḥ`
3033	saṃprasāraṇa for styā after pra before a niṣṭhā	`pra-styā-ta → prastī-ta`
3034	after pra-styā, optionally m	`prastīmaḥ` / `prastītaḥ`; `styānaḥ`
3035	without an upasarga, phulla, kṣīva, kṛśa, ullāgha by nipātana	`phullaḥ`, `phullavān`, `kṣīvaḥ`
3036	no iṭ to a niṣṭhā after an ādit root	`phal-ta` (no iṭ)
3037	u for the a of car, phal before a t-initial kit affix; vārttika utphulla, saṃphulla	`praphultaḥ`, `prakṣīvitaḥ`
3038	optional n after nud, vid, und, trā, ghrā, hrī	`nunnaḥ` / `nuttaḥ`; `vinnaḥ` / `vittaḥ`
3039	no iṭ to a niṣṭhā after śvi and an īdit root	`unnaḥ` / `uttaḥ`; `trāṇaḥ` / `trātaḥ`
3040	no n after dhyā, khyā, pṝ, mūrch, mad	`dhyātaḥ`, `pūrtaḥ`, `mūrtaḥ`, `mattaḥ`
3041	vitta (from vid lābhe) for "enjoyed" and "known"	`vittaṃ dhanam`; but `vinnaḥ`
3042	bhitta "a fragment"	`bhittam`; but `bhinnam`
3043	ṛṇa, with n, in a debtor's transaction	`ṛṇam`; but `ṛtam`
3044	sphī for sphāy before a niṣṭhā	`sphītaḥ`
3045	iṭ to a niṣṭhā after nir-kuṣ, against SK 3025	`niṣkuṣitaḥ`
3046	obligatory iṭ after vas, kṣudh before ktvā and a niṣṭhā	`uṣitaḥ`, `kṣudhitaḥ`
3047	iṭ after añc "to honour"	`añcitaḥ`; but `aktaḥ`
3048	obligatory iṭ after lubh "to bewilder"	`lubhitaḥ`; but `lubdhaḥ`
3049	optional iṭ after kliś	`kliśitaḥ` / `kliṣṭaḥ`
3050	optional iṭ after pūṅ, against SK 2381	`pū-i-ta` / `pū-ta`
3051	after pūṅ, a seṭ ktvā and niṣṭhā are not kit	`pavitaḥ`; `pūtaḥ`
3052	after śīṅ, svid, mid, kṣvid, dhṛṣ, a seṭ niṣṭhā is not kit; vārttika: niṣṭhā in ādikarman	`śayitaḥ`; yaṅluk `śeśyitaḥ`
3053	the kta of ādikarman denotes the agent, and also bhāva and karman	`prakṛtaḥ kaṭaṃ saḥ`
3054	in bhāva and ādikarman, optional iṭ after an ādit root	`prasveditaś caitraḥ`; `prasvinnaḥ`
3055	after mṛṣ "to forbear", a seṭ niṣṭhā is not kit	`marṣitaḥ`; but `apamṛṣitam`
3056	after a root with penultimate u, in bhāva and ādikarman, optionally not kit; only śap roots	`dyutitam` / `dyotitam`; `gudhitam`
3057	ṇi lost before a seṭ niṣṭhā; further forms with and without iṭ	`bhāvitaḥ`; `śūnaḥ`, `gūḍhaḥ`, `patitaḥ`, `daridritaḥ`
3058	eight aniṭ nipātanas for eight senses	`kṣubdhaḥ`, `svāntam`, `lagnam`, `bāḍham`; but `kṣubhitam`
3059	dhṛṣ, śas aniṭ only in impudence	`dhṛṣṭaḥ`; but `dharṣitaḥ`
3060	dṛḍha "stout, strong" by nipātana	`dṛḍhaḥ`; but `dṛṃhitaḥ`
3061	parivṛḍha "master" by nipātana	`parivṛḍhaḥ`; but `parivṛṃhitaḥ`
3062	kaṣ aniṭ for "painful", "hard to penetrate"	`kaṣṭaṃ śāstram`; but `kaṣitam`
3063	ghuṣ aniṭ except for "making known by words"	`ghuṣṭā rajjuḥ`; `ghuṣitaṃ vākyam`
3064	ard aniṭ after sam, ni, vi	`samarṇaḥ`; but `arditaḥ`
3065	ard aniṭ after abhi for "not far"	`abhyarṇam`; but `abhyarditam`
3066	vṛtta from causative vṛt for "studied": no iṭ, luk of ṇic	`vṛttaṃ chandaḥ`; but `vartitā rajjuḥ`
3067	śṛ for śrā, śrapi before kta, of milk or oblation	`śṛtaṃ kṣīram`; but `śrāṇam`, `śrapitam`
3068	seven optional nipātanas from causative stems	`dāntaḥ` / `damitaḥ`; `pūrṇaḥ` / `pūritaḥ`
3069	optional iṭ after ruṣ, am, tvar, saṃ-ghuṣ, ā-svan	`ruṣṭaḥ` / `ruṣitaḥ`; `tūrṇaḥ` / `tvaritaḥ`
3070	optional iṭ after hṛṣ of hair; vārttika: astonishment, being thwarted	`hṛṣṭaṃ` / `hṛṣitaṃ loma`
3071	apacita from cāy, optionally	`apacitaḥ` / `apacāyitaḥ`
3072	optional pī for pyāy before a niṣṭhā, with fixed domains	`pīnaḥ` / `pyānaḥ`; `prapyānaḥ`; `āpīnam ūdhaḥ`
3073	short vowel for hlād before a niṣṭhā	`prahlannaḥ`
3074	i for the final of do, so, mā, sthā before a t-initial kit affix	`ditaḥ`, `sitaḥ`, `mitaḥ`, `sthitaḥ`
3075	the same i, optionally, for śo, cho; obligatory for a vow	`śitaḥ` / `śātaḥ`; `saṃśitaṃ vratam`
3076	hi for dhā	`abhihitam`, `nihitam`
3077	dath for the ghu dā	`dattaḥ`; but `dātaḥ`
3078	t for the vowel of dā after a vowel-final upasarga	`prattaḥ`, `avattaḥ`; also `pradattam`
3079	long vowel for an ik-final upasarga before that t	`nīttam`; also `dhītam`, `jātam`
3080	jagdhi for ad before lyap and a t-initial kit affix	`jagdhaḥ`; also `prakṣīṇaḥ saḥ`
3081	optional long ī for kṣi in imprecation and wretchedness	`kṣīṇāyuḥ` / `kṣitāyuḥ`
3082	ṣ for the s of snā after ni, nadī, of skill	`niṣṇātaḥ`, `nadīṣṇaḥ`
3083	ṣ after prati, qualifying sūtra	`pratiṣṇātaṃ sūtram`; but `pratisnātam`
3084	kapiṣṭhala as a gotra name	`kapiṣṭhalaḥ`; but `kapisthalam`
3085	ṣ for the s of sthala after vi, ku, śami, pari	`viṣṭhalam`, `kuṣṭhalam`
3086	kta in the agent after roots of going, akarmaka roots, and śliṣ … jṝ	`gaṅgāṃ gataḥ`, `garuḍam ārūḍhaḥ`
3087	kta in the locus after roots of remaining still, going, eating	`mukundasyāsitam idam`
3088	kta in the present after a ñīt root	`kṣviṇṇaḥ`, `iddhaḥ`
3089	kta in the present after roots of desire, understanding, honouring	`rājñāṃ mataḥ`, `pūjitaḥ`
3090	kta in neuter bhāva, in time in general	`jalpitam`, `hasitam`
3091	ṅvanip in the past after su, yaj	`sutvā`, `yajvā`
3092	atṛn in the past after jṝ; the niṣṭhā too, by vāsarūpa	`jaran`; `jīrṇaḥ`
3093	in the Veda, liṭ in the past in general	— (no Vedic example; see 3095)
3094	kānac optionally replaces that liṭ	`cakrāṇa`
3095	kvasu also replaces it; Vedic, but poets use them	`tasthivāṃsam`, `adhijagmuṣaḥ`
3097	in ordinary speech, after sad, vas, śru: optional liṭ, always replaced by kvasu	`niṣeduṣīm`, `śuśruvān`
3098	nipātanas upeyivān, anāśvān, anūcāna	`upeyivān`, `anāśuṣaḥ`, `anūcānaḥ`
3096	vasu takes iṭ only after roots still ekāc after reduplication, ā-final roots and ghas	`ādivān`, `dadivān`, `jakṣivān`; not `babhūvān`
3099	optional iṭ for vasu after gam, han, vid, viś; vārttika dṛś	`jagmivān` / `jaganvān`; `dadṛśivān` / `dadṛśvān`
3100	śatṛ, śānac replace laṭ, co-referent with a non-first-case word	`pacantaṃ caitraṃ paśya`
3101	muk for an a before āna; śatṛ sometimes with a first-case word; vārttika māṅ	`pacamānam`; `san brāhmaṇaḥ`; `mā jīvan`
3102	also in address	`he pacan`, `he pacamāna`
3103	also for a mark or a cause of another action	`śayānā bhuñjate`, `arjayan vasati`
3104	ī for the first sound of āna after ās	`āsīnaḥ`
3105	vasu optionally for the śatṛ of vid "to know"	`vidan` / `vidvān`, `viduṣī`
3106	name: śatṛ and śānac are sat	`pacat`, `pacamāna`
3107	sat optionally for lṛṭ; a fixed option	`kariṣyantaṃ kariṣyamāṇaṃ paśya`
3108	śānan after pū, yaj, in the present	`pavamānaḥ`, `yajamānaḥ`
3109	cānaś for disposition, age, capacity	`bhogaṃ bhuñjānaḥ`, `śatrūn nighnānaḥ`
3110	śatṛ after iṅ, dhāri, when the action is easy	`adhīyan`, `dhārayan`
3111	śatṛ after dviṣ for an enemy	`dviṣan`
3112	śatṛ after su for pressing in a sacrifice	`sarve sunvantaḥ`
3113	śatṛ after arh in praise	`arhan`
3114	heading: affixes of habit, duty, skill, up to kvip	`kartā kaṭān`
3115	tṛn	`kartā kaṭān`
3116	iṣṇuc after alaṃkṛñ and eleven more	`alaṃkariṣṇuḥ`, `sahiṣṇuḥ`
3117	in the Veda, iṣṇuc after a ṇi-final root	`pārayiṣṇavaḥ`
3118	in the Veda, iṣṇuc after bhū	`bhaviṣṇuḥ`
3119	gsnu after glā, ji, sthā, bhū; vārttika daṃś	`jiṣṇuḥ`, `sthāsnuḥ`, `bhūṣṇuḥ`; `daṅkṣṇavaḥ`
3120	knu after tras, gṛdh, dhṛṣ, kṣip	`trasnuḥ`, `gṛdhnuḥ`
3121	ghinuṇ after śam and seven more; no vṛddhi	`śamī`, `pramādī`, `unmādī`
3122	ghinuṇ after twenty-seven bases	`saṃparkī`, `tyāgī`, `abhyāghātī`
3123	ghinuṇ with vi after kaṣ, las, katth, srambh	`vilāsī`, `vikatthī`
3124	ghinuṇ after laṣ with apa or vi	`apalāṣī`, `vilāṣī`
3125	ghinuṇ with pra after six roots	`pralāpī`, `pravāsī`
3126	vuñ after ten bases; indication: no tṛc etc. in these senses	`nindakaḥ`, `hiṃsakaḥ`
3127	vuñ after devi, kruś with an upasarga	`ādevakaḥ`, `ākrośakaḥ`; but `kroṣṭā`
3128	yuc after intransitive roots of motion and sound	`calanaḥ`, `śabdanaḥ`; but `paṭhitā vidyām`
3129	yuc after intransitive, consonant-initial anudātteta roots	`vartanaḥ`, `vardhanaḥ`
3130	yuc after ju and nine more; indication: no mutual alternation	`javanaḥ`, `caṅkramaṇaḥ`
3131	yuc after roots meaning "be angry", "adorn"	`krodhanaḥ`, `bhūṣaṇaḥ`
3132	no yuc after a y-final root	`knūyitā`
3133	no yuc after sūd, dīp, dīkṣ; the bar on alternation is only general	`dīpitā`; `kamrā` / `kamanā`
3134	ukañ after laṣ and nine more	`pātukaḥ`, `ghātukaḥ`
3135	ṣākan after jalp and four more	`jalpākaḥ`, `varākī`
3136	ini after pra + ju	`prajavī`
3137	ini after ji and nine more	`jayī`, `paribhavī`
3138	āluc after spṛhi and six more; vārttika śīṅ	`spṛhayāluḥ`, `tandrāluḥ`; `śayāluḥ`
3139	ru after dā, dheṭ, si, śad, sad	`dāruḥ`, `seruḥ`
3140	kmarac after sṛ, ghas, ad	`ghasmaraḥ`
3141	ghurac after bhañj, bhās, mid	`bhaṅguraḥ`, `bhāsuraḥ`
3142	kurac after vid, bhid, chid	`viduraḥ`, `bhiduram`
3143	kvarap after iṇ, naś, ji, sṛ	`itvaraḥ`, `naśvaraḥ`
3144	nipātana gatvara: kvarap and loss of the nasal of gam	`gatvaraḥ`, `gatvarī`
3145	ūka after jāgṛ	`jāgarūkaḥ`
3146	ūka after yaj, jap, daś ending in yaṅ	`yāyajūkaḥ`, `dandaśūkaḥ`
3147	ra after nam and six more	`namraḥ`, `smeraḥ`, `ajasram`
3148	u after a san-stem, āśaṃs, bhikṣ	`cikīrṣuḥ`, `bhikṣuḥ`
3149	nipātanas vindu, icchu	`vinduḥ`, `icchuḥ`
3150	in the Veda, u after a kya-stem	`sumnayuḥ`
3151	in the Veda, ki, kin, treated like liṭ; vārttikas for ordinary speech	`papiḥ`, `jaghniḥ`; `cakriḥ`, `nemiḥ`
3152	najiṅ after svap, tṛṣ (and dhṛṣ)	`svapnak`, `tṛṣṇak`
3153	āru after śṝ, vand	`vandāruḥ`
3154	kru, klukan after bhī; vārttika krukan	`bhīruḥ`, `bhīlukaḥ`, `bhīrukaḥ`
3155	varac after sthā, īś, bhās, pis, kas	`sthāvaraḥ`, `īśvaraḥ`
3156	varac after yā ending in yaṅ; the ā survives	`yāyāvaraḥ`
3157	kvip after eight bases; the limit of the heading	`vibhrāṭ`, `dhūḥ`, `grāvastut`
3158	kvip after other roots too; four vārttikas	`chit`; `vāk`, `didyut`, `juhūḥ`, `dhīḥ`
3159	kvip after bhū only in a name or for "one between"	`mitrabhūḥ`, `pratibhūḥ`
3160	ḍu after bhū with vi, pra, sam, not in a name; vārttika mitadru	`prabhuḥ`, `vibhuḥ`; `śambhuḥ`
3161	ṣṭran after dhā in the sense of the object	`dhātrī`
3162	ṣṭran after thirteen roots in the sense of the instrument	`dātram`, `netram`
3163	no iṭ before ti, tu, tra and seven more kṛt affixes	`śastram`, `patram`, `daṃṣṭrā`
3164	ṣṭran after pū for part of a plough or hog	`potram`
3165	itra after ṛ and six more, instrument	`aritram`, `khanitram`
3166	itra after pū in a name	`pavitram`
3167	itra after pū: instrument for a ṛṣi, agent for a deity; end of the pūrvakṛdanta	`pavitram`; `agniḥ pavitram`

← textbook contents reading page (Technical) →

67. Pūrvakṛdanta-prakaraṇa — The first kṛt affixes: agent nouns, upapada compounds, the past and present participles, and the affixes of habit

Before you start

I.A. The general agent affixes ṇvul and tṛc — running example: kṛ + ṇvul → kṛ + vu → kṛ + aka → (vṛddhi) kārakaḥ; kṛ + tṛc → (guṇa) kartṛ → kartā

SK 2895 · ण्वुल्तृचौ · ṇvultṛcau (A 3.1.133)

I.B. lyu, ṇini and ac for the three gaṇas — running example: nandi + lyu → (yu → ana) nandanaḥ; grah + ṇini → (vṛddhi) grāhī; pac + ac → pacaḥ

SK 2896 · नन्दिग्रहिपचादिभ्यो ल्युणिन्यचः · nandigrahipacādibhyo lyuṇinyacaḥ (A 3.1.134)

I.C. ka after an ik-penultimate root, jñā, prī and kṝ, and after ā-finals with an upasarga — running example: kṣip + ka → (k: no guṇa) kṣipaḥ; pra + jñā + ka → (ā-loss) prajñaḥ

SK 2897 · इगुपधज्ञाप्रीकिरः कः · igupadhajñāprīkiraḥ kaḥ (A 3.1.135)

SK 2898 · आतश्चोपसर्गे · ātaś copasarge (A 3.1.136)

I.D. śa: agents built on the sārvadhātuka stem — running example: pā + śa → (ś: sārvadhātuka; piba for pā, SK 2360; śap, pararūpa) pibaḥ

SK 2899 · पाघ्राध्माधेट्दृशः शः · pāghrādhmādheṭdṛśaḥ śaḥ (A 3.1.137)

SK 2900 · अनुपसर्गाल्लिम्पविन्दधारिपारिवेद्युदेजिचेतिसातिसाहिभ्यश्च · anupasargāl limpavindadhāripārivedyudejicetisātisāhibhyaś ca (A 3.1.138)

SK 2901 · ददातिदधात्योर्विभाषा · dadātidadhātyor vibhāṣā (A 3.1.139)

I.E. ṇa, optional, invariable and vyavasthita, and ka for the house — running example: jval + ṇa → (vṛddhi) jvālaḥ / jval + ac → jvalaḥ; dā + ṇa → (yuk) dāyaḥ

SK 2902 · ज्वलितिकसन्तेभ्यो णः · jvalitikasantebhyo ṇaḥ (A 3.1.140)

SK 2903 · श्याद्व्यधास्रुसंस्व्रतीणवसावहृलिहश्लिषश्वसश्च · śyādvyadhāsrusaṃsvratīṇavasāvahṛlihaśliṣaśvasaś ca (A 3.1.141)

SK 2904 · दुन्योरनुपसर्गे · dunyor anupasarge (A 3.1.142)

SK 2905 · विभाषा ग्रहः · vibhāṣā grahaḥ (A 3.1.143)

SK 2906 · गेहे कः · gehe kaḥ (A 3.1.144)

I.F. Agents with a stated quality: craft, rice and time, repetition, blessing — running example: nṛt + ṣvun → (vu → aka, guṇa) nartakaḥ → (ṣ: ṅīṣ) nartakī

SK 2907 · शिल्पिनि ष्वुन् · śilpini ṣvun (A 3.1.145)

SK 2908 · गस्थकन् · gas thakan (A 3.1.146)

SK 2909 · ण्युट् च · ṇyuṭ ca (A 3.1.147)

SK 2910 · हश्च व्रीहिकालयोः · haś ca vrīhikālayoḥ (A 3.1.148)

SK 2911 · प्रुसृल्वः समभिहारे वुन् · prusṛlvaḥ samabhihāre vun (A 3.1.149)

SK 2912 · आशिषि च · āśiṣi ca (A 3.1.150)

I.G. aṇ with an object upapada, and the upapada compound — running example: kumbha (karman) + kṛ + aṇ → (vṛddhi) kumbha + kāra → (upapada compound) kumbhakāraḥ

SK 2913 · कर्मण्यण् · karmaṇy aṇ (A 3.2.1)

SK 2914 · ह्वावामश्च · hvāvāmaś ca (A 3.2.2)

I.H. ka (and ḍa) with an upapada, and the ṣ of -ṣṭha — running example: go + dā + ka → (ā-loss, SK 2372) godaḥ; sama + sthā + ka → samasthaḥ; dvi + stha → (ṣ) dviṣṭhaḥ

SK 2915 · आतोऽनुपसर्गे कः · āto 'nupasarge kaḥ (A 3.2.3)

SK 2916 · सुपि स्थः · supi sthaḥ (A 3.2.4)

SK 2917 · प्रष्ठोऽग्रगामिनि · praṣṭho 'gragāmini (A 8.3.92)

SK 2919 · तुन्दशोकयोः परिमृजापनुदोः · tundaśokayoḥ parimṛjāpanudoḥ (A 3.2.5)

SK 2920 · प्रे दाज्ञः · pre dājñaḥ (A 3.2.6)

SK 2921 · समि ख्यः · sami khyaḥ (A 3.2.7)

I.I. ṭak after gā and pā; ac after hṛ, grah, arh, ram and jap, śam, śī — running example: sāman + gai + ṭak → (ā-loss) sāmagaḥ → (ṭ: ṅīp) sāmagī; aṃśa + hṛ + ac → (guṇa) aṃśaharaḥ (against aṇ in bhārahāraḥ)

SK 2922 · गापोष्टक् · gāpoṣ ṭak (A 3.2.8)

SK 2923 · हरतेरनुद्यमनेऽच् · harater anudyamane 'c (A 3.2.9)

SK 2924 · वयसि च · vayasi ca (A 3.2.10)

SK 2925 · आङि ताच्छील्ये · āṅi tācchīlye (A 3.2.11)

SK 2926 · अर्हः · arhaḥ (A 3.2.12)

SK 2927 · स्तम्बकर्णयो रमिजपोः · stambakarṇayo ramijapoḥ (A 3.2.13)

SK 2928 · शमि धातोः संज्ञायाम् · śami dhātoḥ saṃjñāyām (A 3.2.14)

SK 2929 · अधिकरणे शेतेः · adhikaraṇe śeteḥ (A 3.2.15)

I.J. ṭa after car and sṛ — running example: kuru (locus) + car + ṭa → kurucaraḥ → (ṭ: ṅīp) kurucarī

SK 2930 · चरेष्टः · careṣ ṭaḥ (A 3.2.16)

SK 2931 · भिक्षासेनादायेषु च · bhikṣāsenādāyeṣu ca (A 3.2.17)

SK 2932 · पुरोऽग्रतोऽग्रेषु सर्तेः · puro 'grato 'greṣu sarteḥ (A 3.2.18)

SK 2933 · पूर्वे कर्तरि · pūrve kartari (A 3.2.19)

I.K. ṭa after kṛ — running example: yaśas + kṛ + ṭa (hetu) → (s for visarga, SK 160) yaśaskara → (ṭ: ṅīp) yaśaskarī vidyā

SK 2934 · कृञो हेतुताच्छील्यानुलोम्येषु · kṛño hetutācchīlyānulomyeṣu (A 3.2.20)

SK 2936 · कर्मणि भृतौ · karmaṇi bhṛtau (A 3.2.22)

SK 2937 · न शब्दश्लोककलहगाथावैरचाटुसूत्रमन्त्रपदेषु · na śabdaślokakalahagāthāvairacāṭusūtramantrapadeṣu (A 3.2.23)

I.L. in after kṛ, hṛ, grah and bhṛ — running example: stamba + kṛ + in → (guṇa) stambakariḥ vrīhiḥ; ātman + bhṛ + in → (mum by nipātana) ātmambhariḥ

SK 2938 · स्तम्बशकृतोरिन् · stambaśakṛtor in (A 3.2.24)

SK 2939 · हरतेर्दृतिनाथयोः पशौ · harater dṛtināthayoḥ paśau (A 3.2.25)

SK 2940 · फलेग्रहिरात्मम्भरिश्च · phalegrahir ātmambhariś ca (A 3.2.26)

I.M. khaś, with the short vowel and mum before a khit uttarapada — running example: jana + ej-i + khaś → (śap, guṇa, ay, pararūpa) jana-ejaya → (sup-luk, mum) janamejayaḥ; śunī + dhaya → (hrasva) śuni-dhaya → (mum) śunindhayaḥ

SK 2941 · एजेः खश् · ejeḥ khaś (A 3.2.28)

SK 2942 · अरुर्द्विषदजन्तस्य मुम् · arurdviṣadajantasya mum (A 6.3.67)

SK 2943 · खित्यनव्ययस्य · khity anavyayasya (A 6.3.66)

SK 2944 · नासिकास्तनयोर्ध्माधेटोः · nāsikāstanayor dhmādheṭoḥ (A 3.2.29)

SK 2945 · नाडीमुष्ट्योश्च · nāḍīmuṣṭyoś ca (A 3.2.30)

SK 2946 · उदि कूले रुजिवहोः · udi kūle rujivahoḥ (A 3.2.31)

SK 2947 · वहाभ्रे लिहः · vahābhre lihaḥ (A 3.2.32)

SK 2948 · परिमाणे पचः · parimāṇe pacaḥ (A 3.2.33)

SK 2949 · मितनखे च · mitanakhe ca (A 3.2.34)

SK 2950 · विध्वरुषोस्तुदः · vidhvaruṣos tudaḥ (A 3.2.35)

SK 2951 · असूर्यललाटयोर्दृशितपोः · asūryalalāṭayor dṛśitapoḥ (A 3.2.36)

SK 2952 · उग्रम्पश्येरम्मदपाणिन्धमाश्च · ugrampaśyerammadapāṇindhamāś ca (A 3.2.37)

I.N. khac, and its short vowel — running example: priya + vad + khac → (mum) priyaṃvadaḥ; dviṣat + tāpi + khac → (hrasva, 2955) dviṣat-tapi → (ṇi-loss; mum; t lost) dviṣantapaḥ

SK 2953 · प्रियवशे वदः खच् · priyavaśe vadaḥ khac (A 3.2.38)

SK 2954 · द्विषत्परयोस्तापेः · dviṣatparayos tāpeḥ (A 3.2.39)

SK 2955 · खचि ह्रस्वः · khaci hrasvaḥ (A 6.4.94)

SK 2956 · वाचि यमो व्रते · vāci yamo vrate (A 3.2.40)

SK 2957 · वाचंयमपुरंदरौ च · vācaṃyamapuraṃdarau ca (A 6.3.69)

SK 2958 · पूःसर्वयोर्दारिसहोः · pūḥsarvayor dārisahoḥ (A 3.2.41)

SK 2959 · सर्वकूलाभ्रकरीषेषु कषः · sarvakūlābhrakarīṣeṣu kaṣaḥ (A 3.2.42)

SK 2960 · मेघर्तिभयेषु कृञः · meghartibhayeṣu kṛñaḥ (A 3.2.43)

I.A. The general agent affixes ṇvul and tṛc — running example: `kṛ + ṇvul → kṛ + vu → kṛ + aka → (vṛddhi) kārakaḥ`; `kṛ + tṛc → (guṇa) kartṛ → kartā`

I.B. lyu, ṇini and ac for the three gaṇas — running example: `nandi + lyu → (yu → ana) nandanaḥ`; `grah + ṇini → (vṛddhi) grāhī`; `pac + ac → pacaḥ`

I.C. ka after an ik-penultimate root, jñā, prī and kṝ, and after ā-finals with an upasarga — running example: `kṣip + ka → (k: no guṇa) kṣipaḥ`; `pra + jñā + ka → (ā-loss) prajñaḥ`

I.D. śa: agents built on the sārvadhātuka stem — running example: `pā + śa → (ś: sārvadhātuka; piba for pā, SK 2360; śap, pararūpa) pibaḥ`

I.E. ṇa, optional, invariable and vyavasthita, and ka for the house — running example: `jval + ṇa → (vṛddhi) jvālaḥ` / `jval + ac → jvalaḥ`; `dā + ṇa → (yuk) dāyaḥ`

I.F. Agents with a stated quality: craft, rice and time, repetition, blessing — running example: `nṛt + ṣvun → (vu → aka, guṇa) nartakaḥ → (ṣ: ṅīṣ) nartakī`

I.G. aṇ with an object upapada, and the upapada compound — running example: `kumbha (karman) + kṛ + aṇ → (vṛddhi) kumbha + kāra → (upapada compound) kumbhakāraḥ`

I.H. ka (and ḍa) with an upapada, and the ṣ of -ṣṭha — running example: `go + dā + ka → (ā-loss, SK 2372) godaḥ`; `sama + sthā + ka → samasthaḥ`; `dvi + stha → (ṣ) dviṣṭhaḥ`

I.I. ṭak after gā and pā; ac after hṛ, grah, arh, ram and jap, śam, śī — running example: `sāman + gai + ṭak → (ā-loss) sāmagaḥ → (ṭ: ṅīp) sāmagī`; `aṃśa + hṛ + ac → (guṇa) aṃśaharaḥ` (against aṇ in `bhārahāraḥ`)

I.J. ṭa after car and sṛ — running example: `kuru (locus) + car + ṭa → kurucaraḥ → (ṭ: ṅīp) kurucarī`

I.K. ṭa after kṛ — running example: `yaśas + kṛ + ṭa (hetu) → (s for visarga, SK 160) yaśaskara → (ṭ: ṅīp) yaśaskarī vidyā`

I.L. in after kṛ, hṛ, grah and bhṛ — running example: `stamba + kṛ + in → (guṇa) stambakariḥ vrīhiḥ`; `ātman + bhṛ + in → (mum by nipātana) ātmambhariḥ`

I.M. khaś, with the short vowel and mum before a khit uttarapada — running example: `jana + ej-i + khaś → (śap, guṇa, ay, pararūpa) jana-ejaya → (sup-luk, mum) janamejayaḥ`; `śunī + dhaya → (hrasva) śuni-dhaya → (mum) śunindhayaḥ`

I.N. khac, and its short vowel — running example: `priya + vad + khac → (mum) priyaṃvadaḥ`; `dviṣat + tāpi + khac → (hrasva, 2955) dviṣat-tapi → (ṇi-loss; mum; t lost) dviṣantapaḥ`

II.A. khac (and aṇ) after kṛñ, bhū, the "name" roots and gam — running example: `kṣema + kṛ + khac → kṣema-m-kara → kṣemaṅkaraḥ`; `kṣema + kṛ + aṇ → kṣemakāraḥ`

II.B. ḍa after gam and han: the ṭi lost by the ḍ-marker — running example: `anta + gam + ḍa → anta-g-a → antagaḥ`

II.C. han with ṇini and ṭak; the nipātanas pāṇigha, tāḍagha — running example: `jāyā + han + ṭak → jāyā-ghn-a → jāyāghnaḥ`

II.D. Making or becoming what one was not: khyun, khiṣṇuc, khukañ — running example: `anāḍhyam āḍhyaṃ kurvanty anena → āḍhya + kṛ + khyun → āḍhya-m-kar-ana → āḍhyaṃkaraṇam`

II.E. Zero and near-zero affixes with a sup upapada: kvip, ṇvi, viṭ, kap — running example: `mitra + dviṣ + kvip → mitra-dviṣ → mitradviṭ`

II.F. manin, kvanip, vanip, vic and kvip after any root — running example: `su + śṝ + manin → su-śar-man → suśarmā`; `vi + jan + vanip → vi-jā-van → vijāvā`

II.G. What kvip does to the root, and ka/kvip after sthā — running example: `aṅga + gam + kvip → aṅga-ga → aṅga-ga-t → aṅgagat`; `he prāṇ`; `śam + sthā + ka → śaṃsthaḥ`

II.H. ṇini: habit, comparison, vow, repetition, and man — running example: `uṣṇa + bhuj + ṇini → uṣṇa-bhoj-in → uṣṇabhojī`

II.I. Thinking of oneself: khaś, and am for an ic-final monosyllable — running example: `paṇḍita + man + śyan + khaś → paṇḍita-m-manya → paṇḍitaṃmanyaḥ`; `go + am + manyaḥ → gāṃmanyaḥ`

II.J. The bhūte heading; ṇini in the past; kvip of han under restriction — running example: `brahman + han + kvip → brahma-han → brahmahā`

II.K. kvip of kṛñ, su, ci, and ini of vi-krī — running example: `su + kṛ + kvip → su-kṛ-t → sukṛt`

II.L. kvanip after dṛś, yudh and kṛñ; karmaṇi ends — running example: `pāraṃ dṛṣṭavān → pāra + dṛś + kvanip → pāradṛśvā`

II.M. ḍa after jan — running example: `sarasi + jan + ḍa → sarasi-j-a → sarasijam`

II.N. The niṣṭhā: name and affixes — running example: `snā + kta → snātam (bhāva)`; `stu + kta → stutaḥ (karman)`; `kṛ + ktavatu → kṛtavān (agent)`

II.O. n for the t of the niṣṭhā — running example: `bhid + kta → bhid-ta → bhin-na → bhinnaḥ`; `kṣi + ktavatu → kṣī-tavat → kṣīṇavān`