Chapter 67 taught the kṛt affixes that mostly name the agent, such as kārakaḥ and kartā. This chapter turns to affixes that name the action itself or one of the other participants in it. The chapter has 218 sūtras, SK 3169 to SK 3386. It opens by fixing the time and the sense of the uṇādi words of chapter 68 (tantuḥ, vartma). Next come the affixes of purpose: tumun (draṣṭum) and ṇvul (darśakaḥ). Then comes the large family of action nouns in ghañ, ac and ap (pākaḥ, cayaḥ, lavaḥ), with the rules that decide between them. The second half teaches the feminine action nouns in ktin (kṛtiḥ) and the neuter ones in lyuṭ (hasanam). It also teaches khal (duṣkaraḥ), ktvā and lyap (bhuktvā, prakṛtya), and ṇamul (smāraṃ smāram). After the first stretch, SK 3169–3219, you will be able to say in which time and which sense an uṇādi word stands. You will form purpose expressions such as kṛṣṇaṃ draṣṭuṃ yāti and kāṇḍalāvo vrajati. You will build ghañ nouns such as pādaḥ, pākaḥ, rāgaḥ, sphāraḥ, upādhyāyaḥ, saṃrāvaḥ, prastāraḥ and kāyaḥ. You will know why saṃrāvaḥ has ā but ravaḥ has a. And you will build the reciprocal noun vyāvakrośī and the pervasive noun sāṃrāviṇam.
cayaḥ.karaḥ, lavaḥ.tataḥ 'sau bhavati → tan + tun (agent) → tantuḥ; vṛttaṃ tat → vṛt + manin (karman) → vartmaChapter 68 listed the uṇādi affixes but did not say when they apply. Six sūtras now fix that. They come in the present and in names, sometimes in the past, and in a few fixed words in the future. Two short lists of fixed words stand in the recipient and source senses. Every other kāraka is open to them. The running examples are tantuḥ "thread", "that which becomes extended", and vartma "path", "that which is traversed".
Rule. The uṇādi affixes come variously (bahulam), in the present and in names. The section opens with the heading atha uttarakṛdantam, "now the later section on kṛt-final words". Bhaṭṭoji adds: "Some, even though not prescribed, are to be inferred."
Example. Bhaṭṭoji then quotes a verse. In names, one should infer the root and the affix that follows it. One should know the marker of such an affix from what it does. "This is the teaching for the uṇādis."
The Tattvabodhinī gives cases. The uṇādi list gives ulac after hṛṣ. Since śaṅkulā is in use, ulac is inferred after śaṅk as well. No sūtra teaches phiḍa or phiḍḍa, yet they are inferred after ṛ: ṛphiḍaḥ, ṛphiḍḍaḥ. They are taken to be k-marked, because the word shows no guṇa.
Applies when / blocked by. "Present" is carried forward from vartamāne laṭ (SK 2151). "In names" is carried forward from puvaḥ saṃjñāyām (SK 3166). The Tattvabodhinī adds that "after a dhātu" (SK 2829), "affix" (SK 180) and kṛt (SK 374) also continue. So every uṇādi affix is a kṛt after a root. By SK 2832 it is in the agent sense unless a rule says otherwise. The Bālamanoramā says "names" means conventional words, whose meaning is not built up from their parts.
Watch out. Here bahulam is what lets an affix come where no sūtra prescribes it. The Bālamanoramā says Bhaṭṭoji's sentence "some, even though not prescribed…" states the purpose of bahulam. The Tattvabodhinī says bahulam makes the incomplete uṇādi list complete.
Rule. The uṇādi affixes are seen also in the past sense. Bhaṭṭoji gives no comment on this sūtra.
Applies when / blocked by. The Tattvabodhinī answers an objection. Why carry "present" into SK 3169 at all, only to add the past here? The answer: these two sūtras show that the uṇādis come mostly in the present, and only now and then in the past or future.
Rule. gamī and the words of its list are fixed forms (nipātana) in the future sense. Bhaṭṭoji gives no comment.
Example. The Tattvabodhinī gives gamī grāmam, "one who will go to the village". It also gives āgāmī, "one who will come".
Applies when / blocked by. The commentators differ on the affix; scholars discuss this. The Tattvabodhinī derives gamī with the uṇādi affix ini. After āṅ that affix is ṇ-marked, so āgāmī has vṛddhi of the penultimate a. The Bālamanoramā takes the affix as ṇini. In gamī the nipātana cancels the vṛddhi its ṇ would bring. It adds prasthāyī, "one who will set out".
Rule. dāśa and goghna are fixed forms in the sense of the recipient (sampradāna).
Example.
"They give to him" → dāśaḥ.
"One kills a cow for him" → goghno 'tithiḥ, "a guest for whom a cow is killed".
Watch out. A kṛt normally names the agent (SK 2832). Here the word names the one for whom the action is done.
Rule. bhīma and the words of its list are fixed forms in the sense of the apādāna, the "from which".
Example. Bhaṭṭoji lists bhīmaḥ, bhīṣmaḥ, praskandanaḥ, prarakṣaḥ, mūrkhaḥ, khalatiḥ.
Applies when / blocked by. The Tattvabodhinī analyses each with "from which" (asmāt). For example, praskandana is "that from which one leaps down", and khalati "bald-headed" is "that from which [something] slips". It derives bhīṣma and mūrkha with uṇādi affixes. But it derives praskandana with lyuṭ, and prarakṣa with the ac of the paca-group (SK 2896). So these "fixed forms" do not all use uṇādi affixes.
Rule. In kārakas other than those two, the recipient and the source, the uṇādi affixes come.
Example.
"It becomes extended (tata)" → tan + tun (agent) → tantuḥ "thread".
"That is traversed (vṛtta)" → vṛt + manin (object) → guṇa → vartman → vartma "path".
"That is moved over (carita)" → carma "hide" (object sense; the affix is not named).
Applies when / blocked by. Bhaṭṭoji says the word "those two" (tābhyām) is there to take in the recipient as well. The Tattvabodhinī explains: without it, the sūtra would look back only to the source of SK 3173. The Tattvabodhinī gives the analyses of tantu and vartma used above.
Watch out. vartma and carma name the object, not the agent. This sūtra is what allows it. The guṇa of vṛt is not stated by Bhaṭṭoji or his commentators; it is inferred.
dṛś + tumun → (am augment: dra-ś; ś → ṣ, t → ṭ) draṣṭum: kṛṣṇaṃ draṣṭuṃ yāti; dṛś + ṇvul → (vu → aka, guṇa) darśakaḥ: kṛṣṇaṃ darśako yātiOne action is often done for the sake of another: "he goes, to see Kṛṣṇa". SK 3175 gives two affixes for the second action, tumun and ṇvul. Four more sūtras give tumun beside words of desire, ability, capability and time, where there is no purpose. Two last sūtras let the action nouns, and aṇ with an object, express the purpose too. The running example is dṛś "to see". With tumun it gives draṣṭum; with ṇvul it gives darśakaḥ.
dṛś + tumun → markers go → dṛś + tum
→ the augment am after the vowel → dṛ-a-ś + tum
→ ṛ becomes r before the vowel (SK 47) → draś + tum
→ ś becomes ṣ → draṣ + tum
→ t after ṣ becomes ṭ (SK 113) → draṣṭum.
The rules for the augment am and for ś → ṣ are not taught here. Bhaṭṭoji does not spell out these steps here.
Rule. Suppose a word for an action stands as upapada, and that action has another action as its purpose. Then tumun and ṇvul come after the root of the second action, in the future sense. "Future" is carried forward from SK 3171. The Bālamanoramā explains the markers. In tumun the n is a marker, and the u after m is only for pronunciation: tum is left. In ṇvul the ṇ and l are markers. The vu becomes aka (SK 1247).
Example.
kṛṣṇaṃ draṣṭuṃ yāti "he goes to see Kṛṣṇa": yāti is the upapada; dṛś takes tumun.
Bhaṭṭoji: because it ends in m, the form is an avyaya (SK 449).
dṛś + ṇvul → dṛś + aka → guṇa of the light penultimate (SK 2189), with r (SK 70) → darśakaḥ.
kṛṣṇaṃ darśako yāti "he goes as one who will see Kṛṣṇa".
Applies when / blocked by. Bhaṭṭoji: here tṛc and the rest do not come by the vāsarūpa rule (SK 2830), because ṇvul is stated again. The Bālamanoramā explains. ṇvul already comes in the agent sense in every time by SK 2895. Stating it again here shows that the other agent affixes do not come in this setting.
Both commentators say tumun is in the sense of the bare action (bhāva), because it makes an avyaya. ṇvul is in the agent sense. The Bālamanoramā adds that kṛṣṇam keeps the second case. Beside the avyaya draṣṭum the sixth case is barred by SK 627. Beside darśaka, an aka in the future, it is barred by SK 628.
Watch out. Both conditions are needed. The Tattvabodhinī gives two counter-examples. In "the matted locks are for begging", the locks are a thing, not an action. In "the stick will fall on you as you run", the running causes the fall but is not done for its sake.
Rule. When a word meaning desire stands as upapada, with the same agent as the root, tumun comes after the root. "Desire" is carried forward from icchārtheṣu liṅloṭau (SK 2814), as the Bālamanoramā says.
Example. icchati bhoktum "he wishes to eat". Likewise vaṣṭi bhoktum, vāñchati bhoktum.
Applies when / blocked by. Bhaṭṭoji: this sūtra is for an upapada that is not an action done for a purpose. In icchati bhoktum the wishing is not done for the sake of eating. Eating is only what the wish is about. So SK 3175 does not reach this case.
The agent must be one and the same. The Tattvabodhinī: "he wishes his son's reading" has two agents. Without this condition, one would wrongly say "he wishes his son to read" with tumun.
Rule. tumun comes after a root when one of these stands as upapada: śak, dhṛṣ, jñā, glā, ghaṭ, rabh, labh, kram, sah, arh, or a word meaning "is" (asti).
Example.
śaknoti bhoktum "he is able to eat".
Likewise dhṛṣṇoti bhoktum and so on.
asti bhoktum, bhavati bhoktum, vidyate bhoktum "there is [occasion] to eat".
Applies when / blocked by. Bhaṭṭoji: the word "meaning" (artha) goes only with asti, because it stands next to it. So any word meaning "is" counts, but for śak and the rest only the named roots count. Both commentators say this sūtra, too, is for an upapada that is not done for a purpose.
Rule. tumun (carried forward from SK 3177) comes when a word stands as upapada that expresses completeness (paryāpti) in the sense of capability. Bhaṭṭoji: paryāpti is fullness.
Example. paryāpto bhoktum "fully able to eat". pravīṇo bhoktum, kuśalo bhoktum, paṭur bhoktum "skilled, clever, adept at eating".
Applies when / blocked by. Bhaṭṭoji gives two counter-examples. Why "words of completeness"? alaṃ bhuktvā "enough of eating": here alam forbids, so no tumun. Why "in the sense of capability"? paryāptaṃ bhuṅkte "he eats amply": abundance is meant, not the eater's ability.
Rule. When a word meaning time stands as upapada, tumun comes. Bhaṭṭoji: naming the synonyms kāla, samaya and velā is to indicate the meaning, so any word for time counts.
Example. kālo bhoktum, samayo bhoktum, velā bhoktum, anehā bhoktum "it is time to eat".
Applies when / blocked by. Bhaṭṭoji: the senses of bidding and the rest (SK 2817) are carried forward here. So there is no tumun in a mere report: "'Time cooks beings': such is the report." Bhaṭṭoji names only "bidding and the rest". That the example has the sense "the time has come" is not stated by Bhaṭṭoji or his commentators; it is inferred.
Rule. The affixes that express the bare action also come in the future, when an action done for a purpose stands as upapada. Bhaṭṭoji: these are ghañ and the rest, taught under the heading bhāve (SK 3184).
Example. yāgāya yāti "he goes for sacrifice". The Tattvabodhinī glosses it as yaṣṭum "to sacrifice". Both commentators derive the fourth case of yāgāya from SK 582.
Applies when / blocked by. The Bālamanoramā explains why the rule is needed. After a tumun-meaning, the vāsarūpa option does not operate. So without this sūtra the action noun would be blocked by tumun.
Rule. When an object word stands as upapada, and an action done for a purpose stands as upapada too, aṇ comes. Bhaṭṭoji calls it an exception (apavāda) to ṇvul.
Example.
kāṇḍa + lū + aṇ → vṛddhi of the final vowel (SK 254) → lau → āv before a (SK 61) → kāṇḍalāvaḥ.
kāṇḍalāvo vrajati "he goes to cut stalks".
Applies when / blocked by. Bhaṭṭoji: because this rule is later (paratva), it blocks ka and the rest. So ka of SK 2915 does not come after an ā-final root here:
kambala + dā + aṇ → y after ā (SK 2761) → kambaladāyaḥ.
kambaladāyo vrajati "he goes to give blankets".
The Bālamanoramā: the ca means both upapadas are needed together. It reads the rule twice. Read once, it restores the aṇ of SK 2913 against the ṇvul of SK 3175. Read again as a fresh rule, it is later than ka and so blocks it. The Bālamanoramā adds that from here on "an action done for a purpose" stops, and only "future" goes on.
Watch out. The aṇ names the agent, by SK 2832. The Tattvabodhinī notes that Bhaṭṭoji names ṇvul alone: tumun is in the bhāva sense, so it could not compete here.
pad + ghañ → (vṛddhi of the penultimate a) pādaḥ; pac + ghañ → (vṛddhi; c → k, SK 2863) pākaḥ; rañj + ghañ (bhāva) → (n lost) raj-a → rāgaḥThe affix ghañ is a with two markers, gh and ñ. The ñ brings vṛddhi: of a final vowel (SK 254), and of a penultimate a (SK 2282). A light penultimate i, u or ṛ takes guṇa instead (SK 2189). The gh turns a root-final c or j into k or g (SK 2863). Two sūtras give ghañ in the agent sense after named roots. Then two headings open: bhāve (SK 3184), the bare action, and akartari ca kārake (SK 3186), a kāraka other than the agent. Both run as far as kṛtyalyuṭo bahulam (A 3.3.113), which Bhaṭṭoji taught earlier as SK 2841. In this chapter that means every rule up to the place of A 3.3.113. Every ghañ rule after the headings works inside them. Between and after the headings stand rules that change the root before ghañ. The 1915 edition prints these sūtras (SK 3186–3189) in a different order. The running examples are pādaḥ "foot", pākaḥ "cooking" and rāgaḥ "colouring".
pac + ghañ → markers go → pac + a
→ vṛddhi of the penultimate a (SK 2282) → pāc + a
→ c becomes k before the gh-marked affix (SK 2863) → pākaḥ.
Rule. ghañ comes after pad, ruj, viś and spṛś, in the agent sense. The Bālamanoramā: the agent sense holds "as before", that is, by the default of kartari kṛt (SK 2832). Bhaṭṭoji: "future" has now ceased to be carried forward.
Example.
pad + ghañ → vṛddhi of the penultimate a → pādaḥ "it goes": the foot.
ruj + ghañ → guṇa (SK 2189) → roj → j becomes g (SK 2863) → rogaḥ "it breaks": disease.
viś + ghañ → guṇa → veśaḥ "it enters".
spṛś + ghañ → guṇa, with r (SK 70) → sparśaḥ "it touches".
Applies when / blocked by. Both commentators explain pādaḥ. The foot is really the instrument "by which one goes". Here the speaker presents the instrument as the agent. The Bālamanoramā adds that ghañ is now a heading. It runs until just before erac (SK 3231).
Watch out. These four nouns name the agent. Most ghañ nouns from SK 3184 on name the action or another kāraka.
Rule. ghañ (from SK 3182) comes after sṛ when the agent is enduring (sthira). Bhaṭṭoji: "sṛ" stands without its case-ending. The Bālamanoramā: it stands for the fifth case, "after sṛ".
Example.
sṛ + ghañ → vṛddhi of the final vowel, with r → sāraḥ.
"It goes on into another time", that is, it lasts: sāraḥ, the firm core.
Applies when / blocked by. A vārttika (a supplementary rule) adds three senses: disease, fish and strength.
atīsāraḥ, a disease. Bhaṭṭoji: here sṛ carries the causative sense within it. It means "it makes blood and the like flow excessively". The Tattvabodhinī: ati is lengthened by SK 1044.visāraḥ, a fish. The Tattvabodhinī: "that which moves about variously".sāraḥ in the sense of strength, and of the firm part.Both commentators note that sthira describes the agent; it is not an upapada. Their proof is that sāra is listed among words that are both masculine and neuter. The Tattvabodhinī says the vārttika is for things that do not endure.
Watch out. Do not read sthire as a word standing beside the root. It is the sense of the affix, an enduring agent.
Rule. ghañ (from SK 3182) comes after a root in the sense of the bare action (bhāva). Bhaṭṭoji: this is the meaning of the root once it has reached the accomplished state (siddhāvasthā). The sūtra is a heading. Bhaṭṭoji says under SK 3187 that it runs as far as kṛtyalyuṭo bahulam (SK 2841).
Example.
pac + ghañ → vṛddhi → pāc → c becomes k (SK 2863) → pākaḥ "cooking".
Dual: pākau "two cookings".
Applies when / blocked by. Both commentators explain "accomplished state". In a verb such as pacati, the action is still in process (sādhyāvasthā). Such an action cannot take gender or number. The Bālamanoramā quotes the Bhāṣya: an action expressed by a kṛt shows itself like a thing. That is why pākaḥ can have a dual.
Watch out. bhāva here is the action itself, not "existence". pākaḥ means the cooking, not the cook.
Rule. When ghañ follows, ā replaces the e, o, ai or au (the ec) of sphur and sphul.
Example.
sphur + ghañ → guṇa of the light penultimate (SK 2189) → sphor
→ ā replaces the o → sphāraḥ.
Likewise sphul → sphālaḥ.
Bhaṭṭoji then gathers four more points on ghañ nouns.
parīhāraḥ.nīkāśaḥ, anūkāśaḥ. With pra, no lengthening: prakāśaḥ. The Bālamanoramā calls SK 1045 a restriction (niyama) on SK 1044.śamaḥ. But ācam, kam and vam do take it: ācāmaḥ, kāmaḥ, vāmaḥ.viśrāma is "not Pāṇinian": śram is barred from vṛddhi by SK 2763.Applies when / blocked by. The Tattvabodhinī offers a way to keep viśrāma. Take the causative stem viśrāmi and add ac; the ā then belongs to the causative. It says that if the causative meaning has faded, poets' use of viśrāmaḥ is correct. Bhaṭṭoji's own verdict stands; scholars discuss this.
The guṇa step before the ā is not stated by Bhaṭṭoji or his commentators; it is inferred from the sūtra's wording "the ec".
Rule. ghañ (from SK 3182) also comes to denote a kāraka other than the agent, when the word is a name (saṃjñā). The sūtra is a heading. Bhaṭṭoji says under SK 3187 that it runs, with bhāve, as far as kṛtyalyuṭo bahulam (SK 2841).
Example. The example follows under SK 3187: prāsaḥ "a dart". lābhaḥ "gain", given there too, illustrates that "name" holds only as a rule (see SK 3187).
Applies when / blocked by. The Bālamanoramā: "kārake" names the meaning of the affix; it is not an upapada. "Name" means a conventional word, so that rāgaḥ (SK 3187) fits.
Watch out. From here on, a ghañ noun may name the action, the object, the instrument, the locus, the source or the recipient. It never names the agent, unless a rule says so.
Rule. When ghañ follows in the sense of the bare action or the instrument, the n of rañj is lost. "Of rañj" is carried forward from SK 2397. "Is lost" comes from SK 415.
Example.
rañj + ghañ (action) → n lost → raj + a
→ vṛddhi of the penultimate a (SK 2282) → rāj
→ j becomes g (SK 2863) → rāgaḥ "colouring".
In the locus sense the n stays: raṅgaḥ, the stage, "that in which one delights".
Applies when / blocked by. Bhaṭṭoji adds three points that belong to SK 3186.
prāsaḥ "a dart": "it is hurled". The Bālamanoramā takes it as the instrument: "that by which enemies are hurled down".ko bhavatā lābho labdhaḥ "what gain have you gained?". The Bālamanoramā: lābha is ghañ in the bhāva sense.The commentators report the Bhāṣya differently on "name"; scholars discuss this. The Tattvabodhinī says the Bhāṣya drops the word and relies on usage. The Bālamanoramā says the Bhāṣya keeps it as holding only as a rule.
Watch out. The n is lost only in the action and instrument senses. In the locus, raṅga keeps it.
Rule. In the sense of speed (java), syada is a fixed form (nipātana). Bhaṭṭoji: from syand before ghañ, the loss of n and the absence of vṛddhi are given by nipātana.
Example.
syand + ghañ → n lost, no vṛddhi (nipātana) → syadaḥ "speed".
Elsewhere, in the sense of flowing: syandaḥ.
Applies when / blocked by. The Bālamanoramā explains why a nipātana is needed. ghañ is not k- or ṅ-marked, so SK 415 cannot remove the n. A plain rule removing the n would leave a penultimate a, and that would take vṛddhi. The nipātana gives both results at once.
Rule. avoda, edha, odma, praśratha and himaśratha are fixed forms (nipātana).
Example.
avodaḥ "wetting": from ava + und with ghañ; the n is lost.
edhaḥ "fuel": from indh with ghañ; the n is lost and i takes guṇa.
odma "moistening": from und.
praśrathaḥ, himaśrathaḥ: Bhaṭṭoji says the n of śranth is lost and vṛddhi is absent.
Applies when / blocked by. Both commentators fill in the steps. In avoda, after the n is lost, a + u merge as o. In edha, the guṇa is itself part of the nipātana, overriding na dhātulopa…, a prohibition that would otherwise bar the guṇa in edha. odma takes the uṇādi affix man, with n-loss and guṇa again by nipātana. praśratha and himaśratha take ghañ.
upa + adhi + iṅ + ghañ (apādāna) → (vṛddhi ai → āy) upādhyāyaḥ; sam + ru + ghañ → saṃrāvaḥ, against ap in ravaḥUnder the two headings, ghañ is the general affix. But SK 3231 gives ac after i-final roots, and SK 3232 gives ap after roots such as ru. These later rules are exceptions to ghañ, so they would win. The sūtras of this group win back ghañ in set conditions: a stated quantity, the root iṅ, a named upasarga, or (SK 3195) no upasarga. SK 3193, a sound-change rule, is placed here for its ghañ example. The difference shows in the vowel: ghañ brings vṛddhi, ac and ap bring guṇa. The running examples are upādhyāyaḥ "teacher" and saṃrāvaḥ "clamour", against plain ravaḥ "sound".
upa + adhi + i + ghañ → vṛddhi of the final i (SK 254) → upa + adhi + ai + a
→ ai becomes āy before a (SK 61) → upa + adhi + āya
→ i becomes y (SK 47) → upa + adhyāya
→ a + a merge (SK 85) → upādhyāyaḥ.
sam + ru + ghañ → vṛddhi → rau → āv (SK 61) → saṃrāvaḥ.
ru + ap → guṇa → ro → av → ravaḥ.
Rule. ghañ comes after any root, in a kāraka other than the agent, when a quantity (parimāṇa) is stated. Both commentators explain: the quantity measures what the affix denotes. Bhaṭṭoji: the rule is there to block ac and ap.
Example.
ni + ci + ghañ → vṛddhi, āy → nicāyaḥ: ekas taṇḍulanicāyaḥ "one heap of rice grains". (Not nicaya with ac.)
nis + pū + ghañ → vṛddhi, āv → niṣpāvaḥ: dvau śūrpaniṣpāvau "two measures of grain cleaned with a winnowing basket".
kṝ + ghañ → vṛddhi, with r → kāraḥ: dvau kārau "two scattered heaps". (Not kara with ap.)
Applies when / blocked by. A vārttika: dāra and jāra take ghañ in the agent sense, and the ṇi of the causative undergoes luk (disappears without trace).
"They rend [the heart]" → dāri + ghañ → ṇi lost by luk → vṛddhi → dārāḥ "a wife".
"They make [a family] waste away" → jari + ghañ → jārāḥ "paramours".
Both commentators explain the luk. An ordinary loss of ṇi would count as still present by SK 50 (sthānivat). It would then stand between root and affix, and the ghañ vṛddhi would fail. Nor can the vṛddhi of the causative serve for jāra. For jṝ, SK 2568 shortens that vowel. With luk, nothing is left to count, and ghañ gives its own vṛddhi.
The commentators give different reasons for "any root" (sarvebhyaḥ). The Bālamanoramā says it lets ghañ set aside only exceptions tied to the root, not those tied to a meaning. So the feminine ekā tilocchritiḥ keeps its ktin. The Tattvabodhinī says it lets ghañ block ap as well as ac.
Watch out. The quantity must measure the thing the word names: one heap, two measures. Without a measure, ac would come by SK 3231, giving nicaya; Bhaṭṭoji does not give that form here, and it is inferred.
Rule. ghañ also comes after iṅ "to study". Bhaṭṭoji calls it an exception to the ac of SK 3231. The Bālamanoramā: iṅ always has adhi before it, and the ghañ names a kāraka other than the agent.
Example.
"Having come near him, one studies from him" → upa + adhi + iṅ + ghañ → upādhyāyaḥ "teacher".
The Bālamanoramā: the teacher is the apādāna, by SK 592.
Applies when / blocked by. Two vārttikas follow.
First: in the apādāna and the feminine, ghañ after iṅ is added, and the result takes ṅīṣ optionally: upādhyāyā / upādhyāyī "a female teacher". The Tattvabodhinī: these are for a woman who herself teaches. The commentators explain why the vārttika is needed. SK 3191 sets aside only ac. In the feminine the domain belongs to ktin (SK 3272), so ghañ would not come without this vārttika.
Second: śṝ takes ghañ in the senses wind, colour and covering. Bhaṭṭoji: "śṝ" is cited without a case-ending. The Bālamanoramā calls it an exception to ap.
śāraḥ "wind": ghañ in the instrument sense.śāraḥ "colour": here the root means "to make variegated".nīśāraḥ, seen in a half-verse: "like a cow, one who has not made a nīśāra is mostly lean in the cold season". akṛtanīśāraḥ means one who has made no wrap. The ī of nī is by SK 1044.Watch out. upādhyāya names the source of learning, not the agent. The ghañ is in the apādāna sense.
Rule. When an upasarga stands as upapada, ghañ comes after ru "to sound". The Bālamanoramā calls it an exception to ap.
Example.
sam + ru + ghañ → vṛddhi, āv → saṃrāvaḥ "clamour".
Without an upasarga: ravaḥ "sound". The Tattvabodhinī: this has ap by SK 3232.
Rule. After abhinis, the s of stan "to sound" is replaced by a retroflex (SK 210), when the word names a sound. This rule belongs to the last three pādas of the Aṣṭādhyāyī (the tripādī). It stands here because its example is a ghañ noun.
Example.
abhi + nis + stan + ghañ → the s of stan becomes ṣ
→ t after ṣ becomes ṭ (SK 113)
→ vṛddhi of the penultimate a → abhiniṣṭānaḥ, a kind of sound (varṇa).
The result has a single ṣ where the s of nis meets the new ṣ; this step is not stated by Bhaṭṭoji or his commentators.
Not a name of a sound: abhiniḥstanati mṛdaṅgaḥ "the drum resounds".
Applies when / blocked by. The change of t to ṭ is not stated by Bhaṭṭoji or his commentators; it is inferred from SK 113.
Rule. When sam stands as upapada, ghañ comes after yu, dru and du. The Bālamanoramā calls it an exception to ap.
Example.
sam + yu + ghañ → vṛddhi, āv → saṃyāvaḥ, "that which is mixed with molasses and the like": a kind of cake made of flour.
Likewise saṃdrāvaḥ, saṃdāvaḥ.
Rule. After śri, nī and bhū, when no upasarga stands before them, ghañ comes. The Bālamanoramā calls it an exception to ac and ap.
Example.
śri + ghañ → vṛddhi, āy → śrāyaḥ. Likewise nāyaḥ, bhāvaḥ.
With an upasarga, no ghañ: praśrayaḥ, praṇayaḥ, prabhavaḥ.
Applies when / blocked by. Bhaṭṭoji answers two objections.
prabhāvo rājñaḥ "the king's might" correct? It is a prādi compound (SK 761), pra with bhāvaḥ: prakṛṣṭo bhāvaḥ, "eminent being".rājño nayaḥ "the king's policy" correct, with no upasarga? By bāhulaka. The Bālamanoramā: by the bahulam of SK 2841.Watch out. Which affix the counter-examples take is not stated by Bhaṭṭoji. That praśrayaḥ and praṇayaḥ have ac and prabhavaḥ has ap is inferred from the Bālamanoramā's "exception to ac and ap".
Rule. When vi stands as upapada, ghañ comes after kṣu "to sneeze" and śru "to hear". The Bālamanoramā calls it an exception to ap.
Example. vikṣāvaḥ, viśrāvaḥ. Without vi: kṣavaḥ, śravaḥ.
Rule. When ava or ud stands as upapada, ghañ comes after nī "to lead". The Bālamanoramā calls it an exception to ac.
Example. avanāyaḥ "leading down"; unnāyaḥ "leading up".
Applies when / blocked by. How is unnayaḥ "conjecture" correct, without vṛddhi? Bhaṭṭoji: by bāhulaka.
Rule. When pra stands as upapada, ghañ comes after dru "to run", stu "to praise" and sru "to flow". The Bālamanoramā calls it an exception to ap.
Example. pradrāvaḥ, prastāvaḥ, prasrāvaḥ. Without pra: dravaḥ, stavaḥ, sravaḥ.
Rule. When nir or abhi stands as upapada, ghañ comes after pū "to cleanse" and lū "to cut". The Bālamanoramā calls it an exception to ap.
Example.
"It is cleansed with winnowing baskets and the like" → niṣpāvaḥ, a kind of grain.
abhilāvaḥ.
Otherwise: pavaḥ, lavaḥ.
Applies when / blocked by. The Tattvabodhinī, on SK 3190, notes that niṣpāva already has ghañ by this rule. Yet with a stated quantity it may come by SK 3190 as well, on the strength of "any root".
Rule. When ud or ni stands as upapada, ghañ comes after gṝ "to swallow". The word graḥ is the fifth case of gṝ, "after gṝ". The Bālamanoramā calls it an exception to ap.
Example.
ud + gṝ + ghañ → vṛddhi, with r → udgāraḥ. Likewise nigāraḥ.
Otherwise: garaḥ.
Rule. After kṝ "to scatter", when grain is concerned, ghañ comes with ud or ni as upapada. The Bālamanoramā: "kṝ" stands with its fifth-case ending dropped.
Example. utkāro nikāro dhānyasya "the scattering of grain".
Applies when / blocked by. Not of grain: bhikṣotkaraḥ "a heap of alms", puṣpanikaraḥ "a heap of flowers". That these take the ap of SK 3232 is not stated by Bhaṭṭoji or his commentators; it is inferred.
Watch out. ud and ni are carried forward from SK 3200. Bhaṭṭoji's paraphrase supplies them.
pra + stṛ + ghañ → (vṛddhi) prastāraḥ, against ap in barhiṣaḥ prastaraḥ; ni + ci + ghañ → (k for the initial c) nikāyaḥThe sūtras of this group also win back ghañ, but now the condition is a meaning as well as an upasarga. stṛ "to spread" takes ghañ with pra outside sacrifice, and with vi for spreading out. grah "to seize" takes it with ud, and with sam for the fist. nī, iṇ and śī take it in gambling, propriety and one's turn. ci "to pile" takes it for things picked by hand, and with k for its c in four further senses. Where a counter-example is given, it usually shows the guṇa form of ac or ap. The running examples are prastāraḥ (from stṛ "to spread") and nikāyaḥ "an abode".
pra + stṛ + ghañ → vṛddhi of the final ṛ, with r (SK 254, SK 70) → prastāraḥ.
ni + ci + ghañ → k for the initial c (SK 3213) → ni + ki + a
→ vṛddhi, ai → āy (SK 61) → nikāyaḥ.
Rule. In sacrifice, when sam stands as upapada, ghañ comes after stu "to praise".
Example. "The place where the Chandogas, the singers of the Sāman, come together and praise": that place is a saṃstāvaḥ.
Applies when / blocked by. Both commentators call this ghañ an exception to lyuṭ in the locus sense, taught later in this chapter. Outside sacrifice: saṃstavaḥ "acquaintance".
Rule. When pra stands as upapada, ghañ comes after stṛ "to spread", but not in sacrifice. Bhaṭṭoji: the word is divided as a-yajñe "not in sacrifice". "In sacrifice" is already in context from SK 3202, so reading it again would be pointless.
Example.
pra + stṛ + ghañ → vṛddhi, with r → prastāraḥ.
In sacrifice: barhiṣaḥ prastaraḥ "the prastara of the sacred grass", a particular handful.
Applies when / blocked by. The Bālamanoramā reports that, for ritualists, prastāra is a particular arrangement of bricks.
Watch out. The two forms differ by one vowel. prastāraḥ (ghañ, vṛddhi) is outside sacrifice; prastaraḥ (guṇa) is the sacrificial handful.
Rule. When vi stands as upapada, ghañ comes after stṛ, in the sense of spreading out (prathana), when speech is not concerned. The Bālamanoramā: the words are divided vau aśabde.
Example. paṭasya vistāraḥ "the breadth of a cloth".
Applies when / blocked by. Why "spreading out"? tṛṇavistaraḥ "a strewing of grass". Why "not speech"? granthavistaraḥ "the prolixity of a book".
Rule. ghañ also comes after vi + stṛ in the name of a metre (chandas). Bhaṭṭoji: "after stṛ" is carried forward. The Bālamanoramā supplies vi from SK 3204.
Example.
"The syllables are spread out in it" → vi + stṛ + ghañ (locus) → viṣṭāra.
viṣṭāra + paṅkti → karmadhāraya (a descriptive compound) → viṣṭārapaṅktiḥ, a metre.
The ṣ comes from the next rule, SK 3206.
Applies when / blocked by. Both commentators: the rule is needed because a metre is a matter of speech, which SK 3204 excludes. The Tattvabodhinī reports that some do not carry vi forward here, because of usages such as prastārapaṅktiḥ.
Rule. In the name of a metre, the s of stṛ after vi, in a ghañ form, becomes ṣ.
Example. vi + stāra → ṣ for s → viṣ + tāra → t after ṣ becomes ṭ (SK 113) → viṣṭāra. Bhaṭṭoji: "hence the ṣ" of SK 3205.
Applies when / blocked by. The Bālamanoramā: SK 2123 (sātpadādyoḥ) would bar this ṣ, and this rule overrides it. It mentions vṛkṣāsanayor viṣṭaraḥ (SK 3233), the sūtra just before this one in the Aṣṭādhyāyī; that is where "stṛ after vi" comes from. This edition gives SK 3233 later in this chapter.
Watch out. Two sūtras share the words chandonāmni ca. SK 3205 (A 3.3.34) gives the affix; SK 3206 (A 8.3.94) gives the ṣ. The ṭ step is not stated by Bhaṭṭoji or his commentators; it is inferred.
Rule. When ud stands as upapada, ghañ comes after grah "to seize".
Example. ud + grah + ghañ → vṛddhi of the penultimate a → udgrāhaḥ.
Applies when / blocked by. Both commentators call it an exception to the ap that SK 3234 gives after grah.
Rule. When sam stands as upapada, ghañ comes after grah, where the fist is concerned.
Example. mallasya saṃgrāhaḥ "the wrestler's grip". The Bālamanoramā: the firm grip of the fist.
Applies when / blocked by. Not of the fist: dravyasya saṃgrahaḥ "the collecting of goods".
Rule. ghañ comes after nī "to lead" with pari, in gambling (dyūta). It comes after iṇ "to go" with ni, for propriety (abhreṣa). The pairs go in order.
Example.
pari + nī + ghañ → vṛddhi, āy → pariṇāyaḥ: pariṇāyena śārān hanti "he strikes the pieces by moving them all round". The ṇ after pari is not explained by Bhaṭṭoji here.
ni + i + ghañ → vṛddhi, āy → ni + āya → i becomes y before a vowel (SK 47) → nyāyaḥ: eṣo 'tra nyāyaḥ "this is the proper course here".
Applies when / blocked by. Outside these senses: pariṇayaḥ "marriage"; nyayaḥ "ruin". The Bālamanoramā calls the rule an exception to ac. The Tattvabodhinī: dyūta is playing with dice and the like; abhreṣa is doing what is due.
Rule. When pari stands as upapada, ghañ comes after iṇ, in the sense of anupātyaya. Bhaṭṭoji: anupātyaya is not passing over what has come in due order.
Example.
pari + i + ghañ → vṛddhi, āy → pari + āya → i becomes y (SK 47) → paryāyaḥ.
tava paryāyaḥ "your turn".
Applies when / blocked by. Otherwise: kālasya paryayaḥ "the lapse of time", that is, a passing over.
Rule. When vi or upa stands as upapada, ghañ comes after śī "to lie", in the sense of one's turn (paryāya).
Example. tava viśāyaḥ "your turn [to lie down]"; tava rājopaśāyaḥ "your turn to lie beside the king".
Applies when / blocked by. Otherwise: viśayaḥ "doubt"; upaśayaḥ "lying nearby". The Bālamanoramā calls the rule an exception to ac.
The Tattvabodhinī asks why "turn" is named, when the sense could be carried from SK 3210. Its answer: the word makes this a fresh injunction. So when complete pervasion is meant, this ghañ still comes. It sets aside the inuṇ of SK 3218, though that rule is later.
Rule. ghañ comes after ci "to gather", for taking with the hand, but not for theft. Bhaṭṭoji: "taking with the hand" indicates that the thing taken is near at hand.
Example. puṣpapracāyaḥ "the gathering of flowers".
Applies when / blocked by. Why "with the hand"? "He gathers with a stick the fruits at the top of the tree": pracayam. Why "not theft"? puṣpapracayaś cauryeṇa "the gathering of flowers by theft". The Bālamanoramā: the flowers must be within reach of a man standing on the ground.
Rule. ghañ comes after ci for an abode, a piling, the body or a heaping up. Also, k replaces the initial sound of the root. The Bālamanoramā: bhāve and akartari ca kārake still continue. Bhaṭṭoji: heaping up (upasamādhāna) is the meaning of the root itself. The other three describe the kāraka that the affix denotes.
Example.
ci + ghañ → k for the initial c → ki + a → vṛddhi, āy → kāya.
Abode: kāśīnikāyaḥ "the abode Kāśī".
Piling: ākāyam agniṃ cinvīta "one should pile the fire-altar as an ākāya".
Body: kāyaḥ, "in it bones and the like are piled up".
Heap: gomayanikāyaḥ "a heap of cow-dung".
In other senses: cayaḥ "a pile".
Applies when / blocked by. Bhaṭṭoji explains why the sūtra says "of the initial" and not just "k for c". In the intensive with yaṅ deleted (yaṅluk), the root is doubled and has two c's. Only the first becomes k: gomayānāṃ nikecāyaḥ "heaping up cow-dung again and again".
The Bālamanoramā says k has to be taught here because SK 2863 changes only a c that stands right before the affix. In ci the i stands between.
The commentators read ākāya differently; scholars discuss this. The Bālamanoramā takes it as the object, "what is piled up". The Tattvabodhinī takes it as the locus, "that in which the bricks are piled".
Watch out. Without the k rule, ci + ghañ would keep its c. Compare puṣpapracāyaḥ (SK 3212) with nikāyaḥ.
Rule. ghañ also comes after ci, with k for the initial, for a multitude (saṃgha) not arranged one above another. Bhaṭṭoji supplies "ghañ after ci, and k for the initial" from SK 3213. He defines saṃgha as a multitude of living beings.
Example. bhikṣunikāyaḥ "a community of mendicants".
Applies when / blocked by. Why "not one above another"? sūkaranicayaḥ "a pile of pigs". Why "multitude of beings"? jñānakarmasamuccayaḥ "the combination of knowledge and ritual action".
The Tattvabodhinī explains the pigs. When suckling they lie one on another, so the counter-example holds. When they stand apart, like mendicants, ghañ comes for them too.
vi + ava + kruś + ṇac → vyavakrośa → (añ; ādivṛddhi, no aic) vyāvakrośa → (ṅīp) vyāvakrośī; sam + ru + inuṇ → (vṛddhi) saṃrāvin → (aṇ; ādivṛddhi; in kept) sāṃrāviṇamTwo special action nouns close this stretch. Each is made in two steps: a kṛt after the root, then a taddhita after the kṛt-stem. For a reciprocal action (karmavyatihāra), "doing to one another", the kṛt is ṇac and the taddhita añ. For complete pervasion (abhividhi), the kṛt is inuṇ and the taddhita aṇ. Both taddhitas bring vṛddhi of the first vowel of the whole stem (SK 1075). The running examples are vyāvakrośī "mutual reviling" and sāṃrāviṇam "noise all round".
vi + ava + kruś + ṇac → guṇa (SK 2189) → vi + ava + krośa
→ i becomes y (SK 47) → vyavakrośa
→ + añ (SK 3216) → vṛddhi of the first vowel → vyāvakrośa
→ + ṅīp (SK 470); the final a drops before the ī, a step not stated in this section → vyāvakrośī.
Rule. When the action is reciprocal, ṇac comes after the root, in the feminine. Bhaṭṭoji: in the feminine gender, ṇac in the bare-action sense. ṇac leaves a; its ṇ is a marker.
Example. vi + ava + kruś + ṇac → vyavakrośa, then the next two rules.
Applies when / blocked by. Both commentators take ṇac in the bhāva sense only. The Bālamanoramā: akartari ca kārake is not carried forward here. The Tattvabodhinī: in other kārakas it would not convey the sense. It rejects an earlier author's "bhāva and the rest".
Watch out. ṇac alone does not make the finished word. The añ of SK 3216 and the ṅīp of SK 470 follow.
Rule. After a word ending in ṇac, añ comes in the feminine. The Bālamanoramā: añ is a taddhita in the stem's own sense (svārthika). It brings vṛddhi of the first vowel, and then ṅīp (SK 470). Bhaṭṭoji gives no comment; the examples stand under SK 3217.
Example. vyavakrośa + añ → vṛddhi of the first vowel → vyāvakrośa → ṅīp (the final a drops before the ī) → vyāvakrośī.
Applies when / blocked by. The Tattvabodhinī: this rule shows why ṇac has its c marker. The c lets this rule name it apart from other affixes.
Rule. In a reciprocal action, the ai or au (the aic) does not come. The Tattvabodhinī: this is the aic that SK 1098 would place before a y or v. Bhaṭṭoji: "here aic does not occur".
Example. vyāvakrośī "mutual reviling", from kruś "to call out". vyāvahāsī "mutual laughing", from has "to laugh".
Applies when / blocked by. The Tattvabodhinī explains the result. With the aic barred, only the plain vṛddhi of the first vowel remains. The root with its upasargas counts as ending in ṇac. So when añ follows, the vṛddhi falls on the upasarga: vya becomes vyā. ṇac is a kṛt; añ is a taddhita.
The Bālamanoramā glosses the two words as agents: "one who reviles another", "one who laughs at another". Bhaṭṭoji's rule under SK 3215 gives ṇac in the bhāva sense, so the words name the action: "mutual reviling", "mutual laughing".
Watch out. Do not add ai before the y of vy-. In this reciprocal noun the vṛddhi is simply a → ā.
Rule. When complete pervasion (abhividhi) is meant, inuṇ comes in the bhāva sense. The Bālamanoramā: abhividhi is connection with the whole. inuṇ leaves in. Bhaṭṭoji gives no comment.
Example. sam + ru + inuṇ → vṛddhi of the final u (SK 254), āv → saṃrāvin.
Applies when / blocked by. The Tattvabodhinī gives the counter-example saṃrāvaḥ, without pervasion. It says "bhāva" is named again so that ghañ and kta do not come beside inuṇ by the vāsarūpa option. It adds that lyuṭ is wanted beside it, and comes by bāhulaka.
Rule. After a word ending in inuṇ, the taddhita aṇ comes, in the stem's own sense. Bhaṭṭoji cites SK 1245 (inaṇy anapatye): before this aṇ the stem keeps its form.
Example.
saṃrāvin + aṇ → vṛddhi of the first vowel (SK 1075) → sāṃrāvin + a
→ the ṭi in is kept (SK 1245), not lost by SK 679 → sāṃrāvina
→ sāṃrāviṇaṃ vartate "there is a noise all round".
Applies when / blocked by. Both commentators supply the steps above. The Tattvabodhinī: sam signals the pervasion. It says the word is neuter by nature.
Watch out. SK 679 would normally drop the in before a taddhita. SK 1245 keeps it, because this aṇ does not mean "descendant". The ṇ for n is not explained by Bhaṭṭoji here.
ava + grah + ghañ → avagrāhaḥ (a curse), against ap in avagrahaḥ padasya; pra + grah → pragrahaḥ / pragrāhaḥ (rein)This group goes on with ghañ, the affix a marked with gh and ñ. Like the ghañ rules before it, each sūtra names a root and an upasarga beside it, and most name a sense. The affix then comes in the bare action (bhāva, SK 3184) or in a kāraka other than the agent (SK 3186). The Bālamanoramā makes ghañ a heading from SK 3182 up to just before erac (SK 3231), so Bhaṭṭoji simply supplies "ghañ comes". Most of these rules set aside the affix ap, which SK 3234 gives after grah and vṛ, and SK 3232 after roots in ṛ and u. The two affixes are easy to tell apart. ghañ is ñ-marked, so it brings vṛddhi: grāhaḥ. ap brings none: grahaḥ. From SK 3225 to SK 3230 the ghañ is only optional, and ap is the other form. The running example is grah "to seize". With ava, in a curse, it takes ghañ: avagrāhas te bhūyāt "may you be overpowered!". Without a curse it takes ap: avagrahaḥ padasya "the separation of a word". With pra, for a rein, both come: pragrahaḥ / pragrāhaḥ.
Rule. When a curse (ākrośa) is meant, grah takes ghañ with ava or ni beside it. Bhaṭṭoji glosses ākrośa as śāpa, a curse.
Example.
ava + grah + ghañ → the markers gh and ñ go → ava + grah + a
→ the affix is ñ-marked, so the penultimate a takes vṛddhi (SK 2282) → ava + grāh + a → avagrāhaḥ.
avagrāhas te bhūyāt "may avagrāha befall you!", that is, may you be overpowered (abhibhava).
nigrāhas te bhūyāt "may nigrāha befall you!", that is, may you be obstructed (bādha).
Applies when / blocked by. Why "when a curse is meant"? Without a curse there is no ghañ. Bhaṭṭoji gives avagrahaḥ padasya and nigrahaś corasya. The Bālamanoramā explains both. avagraha of a word is reading a compound with its members apart, as puraḥ hitam for purohitam. nigraha is restraint: "the restraint of a thief". These forms have ap. The Bālamanoramā calls this rule an exception to the ap of SK 3234. The Tattvabodhinī agrees: the rule is made although SK 3234 would give ap.
Watch out. The word avagraha has two meanings and two affixes. In a curse it is avagrāhaḥ, with vṛddhi. For splitting a compound it is avagrahaḥ, without.
Rule. With pra beside it, when a desire to obtain (lipsā) is meant, grah takes ghañ. The Bālamanoramā supplies both grah and ghañ. Neither names the rule it comes from; it is taken here to come from SK 3220.
Example.
pra + grah + ghañ → vṛddhi of the penultimate a (SK 2282) → pragrāhaḥ.
pātrapragrāheṇa carati bhikṣuḥ "the mendicant goes about taking up his bowl". The Bālamanoramā says "to obtain alms" is understood. The Tattvabodhinī glosses: having taken up the alms-bowl.
Applies when / blocked by. This is again an exception to the ap of SK 3234 (the Bālamanoramā). Why "desire to obtain"? The Tattvabodhinī answers with devadattasya pragrahaḥ, "Devadatta's intense attachment". There no obtaining is meant, so ap comes.
Rule. With pari beside it, in connection with the sacrifice (yajña), grah takes ghañ. The Bālamanoramā supplies grah and ghañ, and calls this an exception to the ap of SK 3234.
Example.
pari + grah + ghañ → vṛddhi (SK 2282) → parigrāhaḥ.
uttaraḥ parigrāhaḥ "the later parigrāha". Bhaṭṭoji explains: taking possession of the altar (vedi) with the sphya.
Applies when / blocked by. The Bālamanoramā explains the ritual. The sphya is a sword-shaped wooden implement. parigrāha is drawing a line all round the altar ground with it. A first parigrāha is made with certain mantras, and this is the second, hence "later".
Rule. With ni beside it, when grain (dhānya) is meant, vṛ "to choose, cover" takes ghañ. Bhaṭṭoji notes that vṛ in the sūtra has lost its fifth-case ending: it means "after vṛ". The Tattvabodhinī adds: in bhāva and the other senses.
Example.
ni + vṛ + ghañ → vṛddhi of the final ṛ (SK 254), with r (SK 70) → ni + vār + a
→ the upasarga is lengthened before ghañ (SK 1044, as the Bālamanoramā says) → nīvārāḥ "wild rice".
Applies when / blocked by. Why "grain"? Bhaṭṭoji gives nivarā kanyā "a maiden eminently to be chosen". Here ap comes in the object sense, by SK 3234 (the Tattvabodhinī).
But a feminine object noun should take ktin, the feminine affix of SK 3272, since that rule comes later. So how does ap stand in a feminine? Bhaṭṭoji answers: even in the domain of ktin, ap comes by bāhulaka, the wide scope that SK 2841 allows. He compares pravarā gauḥ "an excellent cow". The Bālamanoramā cites the usage of elders as proof, and says the gender rests on usage. The Tattvabodhinī says the rule "words in ghañ and ap are masculine" holds, mostly, only in the bhāva sense. That leaves room for a feminine object noun like nivarā.
Watch out. nīvārāḥ has a long ī. The lengthening is the work of SK 1044, not of this rule.
Rule. With ud beside them, the roots śri "to resort", yu "to join", pū "to purify" and dru "to run" take ghañ. The Bālamanoramā calls it an exception to ac and ap.
Example.
ud + śri + ghañ → vṛddhi (SK 254) → ud + śrai + a → āy for ai (SK 61), and the usual sandhi d + ś → cch → ucchrāyaḥ.
udyāvaḥ, utpāvaḥ, uddrāvaḥ, with āv for the vṛddhi au (SK 61).
For the last, the 1915 print reads uddrāvaḥ, with two *d*s, as ud + drāva requires.
Applies when / blocked by. Bhaṭṭoji asks how the phrase patanāntāḥ samucchrayāḥ "elevations end in falling" is possible. samucchraya has ud and śri, yet no vṛddhi. His answer: by bāhulaka. The missing vṛddhi suggests the ac of SK 3231, though no one says so.
Rule. With āṅ (ā) beside them, ru "to sound" and plu "to float" take ghañ optionally (vibhāṣā). The Bālamanoramā reads rupluvoḥ as a sixth-case word with fifth-case sense: "after ru and plu".
Example.
ā + ru + ghañ → vṛddhi (SK 254) → ā + rau + a → āv (SK 61) → ārāvaḥ.
In the other option: ā + ru + ap → guṇa (SK 2168) → ā + ro + a → av (SK 61) → āravaḥ.
Likewise āplāvaḥ / āplavaḥ.
Applies when / blocked by. The Bālamanoramā calls this ghañ an exception to ap. So the second form of each pair is the ap of SK 3232, since ru and plu end in u. Bhaṭṭoji does not say so here.
The word vibhāṣā is carried forward into the next rules. SK 3226 says so. It reaches SK 3230.
Rule. With ava beside it, when obstruction of rain (drought) is meant, grah takes ghañ optionally. Bhaṭṭoji: "vibhāṣā is carried forward" from SK 3225.
Example.
ava + grah + ghañ → vṛddhi (SK 2282) → avagrāhaḥ "drought".
Other option: ava + grah + ap → avagrahaḥ "drought".
Applies when / blocked by. In the other option ap comes by SK 3234 (the Bālamanoramā). Why "obstruction of rain"? Bhaṭṭoji gives avagrahaḥ padasya "the separation of a word", with ap only. The Bālamanoramā: reading the two members of a compound apart.
Watch out. Compare SK 3220. In a curse, ava + grah takes ghañ always. For drought it takes ghañ or ap. For splitting a compound it takes only ap.
Rule. With pra beside it, grah takes ghañ optionally, when what the affix denotes belongs to merchants. Bhaṭṭoji narrows it: "that is to say, the string of a balance" (tulāsūtra).
Example.
tulā + pra + grah + ghañ → vṛddhi → tulāpragrāheṇa carati "he goes about with the balance-string".
Other option, with ap: tulāpragraheṇa.
Applies when / blocked by. The Bālamanoramā calls it an exception to the ap of SK 3234. It says the narrowing to the balance-string comes from interpretation (vyākhyāna).
Rule. And when a rein (raśmi) is meant, grah with pra takes ghañ optionally. pra comes from SK 3227, the option from SK 3225, grah from SK 3226. The Tattvabodhinī explains the rein: the rope for restraining horses yoked to a chariot.
Example.
pra + grah + ap → pragrahaḥ "rein".
pra + grah + ghañ → vṛddhi (SK 2282) → pragrāhaḥ "rein".
Applies when / blocked by. The Bālamanoramā: ca "and" gathers in what has been stated. The Tattvabodhinī remarks that grah is named twice over in this stretch (SK 3220 and SK 3226, after udi grahaḥ, SK 3207), though it could have been stated once. Pāṇini could have grouped these rules straight after udi grahaḥ and named grah only there. He did not arrange them so.
Watch out. This is the second half of the running example. With pra, grah takes ghañ only for a desire to obtain (SK 3221). It takes ghañ optionally for the balance-string (SK 3227) and the rein (here).
Rule. With pra beside it, vṛ (vṛṇoti) takes ghañ optionally, when a covering is meant. Bhaṭṭoji: "vibhāṣā and pra do continue".
Example.
pra + vṛ + ghañ → vṛddhi (SK 254), with r (SK 70) → pra + vār + a → the upasarga is lengthened (SK 1044) → prāvāraḥ "covering cloth".
Other option: pra + vṛ + ap → guṇa with r → pravaraḥ.
Applies when / blocked by. The Bālamanoramā: prāvāra is a cloth; the other form has ap by SK 3234. Why "covering"? The Tattvabodhinī gives pravaro gauḥ "an excellent cow", where only ap comes.
Rule. With pari beside it, bhū takes ghañ when disregard (avajñāna) is meant. Bhaṭṭoji gives two forms, so the ghañ is optional.
Example.
pari + bhū + ghañ → vṛddhi (SK 254) → pari + bhau + a → āv (SK 61) → paribhāvaḥ "disregard".
Other option: pari + bhū + ap → guṇa → av → paribhavaḥ "disregard".
Applies when / blocked by. Why "disregard"? Bhaṭṭoji: sarvato bhavanaṃ paribhavaḥ, "being all round is paribhava", with ap only. The Tattvabodhinī quotes the Amarakośa, which lists both forms for disrespect.
Watch out. No one says in so many words that the option of SK 3225 reaches this rule. Bhaṭṭoji's two forms show that it does. The Tattvabodhinī proposes a different order for these rules, and that proposal assumes it. The other form, paribhava, is the ap of SK 3232, bhū ending in ū.
ci + ac → (guṇa, ay) cayaḥ; kṛ + ap → (guṇa) karaḥ; lū + ap → lavaḥ; pra + ad + ap → (ghasḷ) praghasaḥHere the heading of ghañ ends. Two new affixes take its place in the same senses, bhāva and the non-agent kārakas. SK 3231 gives ac after roots in i or ī. SK 3232 gives ap after roots in ṛ or u. Both are a. Neither is marked with ñ, so neither brings vṛddhi. Instead the root vowel takes guṇa (SK 2168). ap then runs on to SK 3265, and the sūtras up to there add roots to it. In those sūtras Bhaṭṭoji or the Bālamanoramā simply supplies "ap comes" (the supplied remainder, śeṣa). They do not say where it comes from. It is taken to be the ap of SK 3232, since ap is taught there and nowhere between. The running examples: ci "to gather" gives cayaḥ "gathering"; kṛ gives karaḥ; lū "to cut" gives lavaḥ; ad "to eat" with pra gives praghasaḥ, with the substitute ghasḷ.
Rule. After a root ending in the i-vowel (i or ī), the affix ac comes. The Bālamanoramā adds the senses: bhāva and a kāraka other than the agent. It is an exception to ghañ.
Example.
ci + ac → the marker c goes → ci + a → guṇa (SK 2168) → ce + a → ay for e (SK 61) → cayaḥ "gathering".
ji + ac → jayaḥ "victory".
Applies when / blocked by. A vārttika: ac should also be added for bhaya and the rest, "to exclude kta and the rest in the neuter". Bhaṭṭoji's forms: bhayam "fear", varṣam "rain".
The Bālamanoramā explains. bhī ends in ī, so erac already gives bhaya. Why the vārttika, then? In the neuter, kta and lyuṭ are also taught for the bare action. They come later in the grammar, so they would win (paratva). Nor does the vāsarūpa option (SK 2830) save ac. That option works only for an exception, and ac is no exception to kta or lyuṭ. So the vārttika enjoins ac afresh in the neuter. The Tattvabodhinī makes the same point. It adds that lyuṭ also occurs, from the Bhāṣya's phrase vṛṣabho varṣaṇāt: varṣaṇam.
Watch out. ac brings guṇa, not vṛddhi. Write cayaḥ, not cāyaḥ. The form nikāyaḥ of SK 3213 has ghañ by its own rule.
Rule. After a root ending in the ṛ-vowel or the u-vowel, the affix ap comes. The Bālamanoramā adds: in bhāva and a kāraka other than the agent, as an exception to ghañ.
Example.
kṛ + ap → the marker p goes → guṇa with r (SK 2168, SK 70) → kar + a → karaḥ.
Also garaḥ, śaraḥ.
yu + ap → guṇa → yo + a → av (SK 61) → yavaḥ.
Also lavaḥ, stavaḥ, pavaḥ.
Applies when / blocked by. lū and pū end in long ū. Do they count? The Tattvabodhinī says yes. SK 3199 gives ghañ after pū and lū with nir or abhi, as an exception. That exception would be pointless unless ap came after them otherwise. So SK 3199 is an indicator (jñāpaka). That is why Bhaṭṭoji cites lavaḥ and pavaḥ.
Scholars discuss the d in ṛdoḥ. The Tattvabodhinī reads "ṛ and ut", and says the d is not a t after ṛ. The Bālamanoramā reads "ṛt and u". Bhaṭṭoji's own gloss, "ending in the ṛ-vowel or the u-vowel", settles the forms either way.
Watch out. Our text writes the sūtra with long ṝ. The 1915 print reads short ṛ, as every cross-reference does. Where no special rule intervenes, i/ī-final roots take ac (SK 3231) and ṛ/u-final roots take ap (here), in place of ghañ.
Rule. In the senses "tree" and "seat", the form viṣṭara is given by nipātana. Bhaṭṭoji: the ṣ for the s of stṛ after vi is given by nipātana in these two senses. (The 1915 print names the root straḥ, that is stṛ.)
Example.
vi + stṛ + ap → guṇa with r → vi + star + a → ṣ for s by this nipātana → viṣṭaraḥ "a tree; a seat".
The t becomes ṭ after ṣ (SK 113).
Applies when / blocked by. Why "tree"? Bhaṭṭoji gives vākyasya vistaraḥ "the expansion of a sentence", with plain s. He gives no counter-example for "seat".
Watch out. This is a rule of the eighth book. Bhaṭṭoji does not say why it stands here. Presumably viṣṭara is an ap-form of stṛ under SK 3232. Compare SK 3206, where viṣṭāra, with ghañ, gets its ṣ in the name of a metre.
Rule. And after grah "to seize", vṛ "to choose", dṛ "to tear", ci with nis, and gam "to go", the affix ap comes. Bhaṭṭoji supplies "ap comes" and calls it "an exception to ghañ and ac".
Example.
grah + ap → grahaḥ.
vṛ + ap → guṇa with r → varaḥ; dṛ + ap → daraḥ.
nis + ci + ap → guṇa, ay, and the usual sandhi ś for s before c → niścayaḥ.
gam + ap → gamaḥ.
Applies when / blocked by. Exception to which? The Bālamanoramā and the Tattvabodhinī agree. After niś-ci (an i-final root) ap sets aside ac. After the others it sets aside ghañ. The Tattvabodhinī notes that ci with pra, for taking by hand, has ghañ by SK 3212.
Three vārttikas follow.
vaśaḥ "will", raṇaḥ "battle". The Bālamanoramā: raṇa is "that in which they make sound". vaś is Vedic, but it is used in ordinary speech too.pra + sthā + ka → loss of ā (SK 2372) → prasthaḥ "a prastha measure", "that in which grain is set". vi + han + ka → penultimate lost (SK 2363) → gh for h (SK 358) → vighnaḥ "obstacle", "that in which minds are struck". The Bālamanoramā says the vārttika continues in the Bhāṣya with a list of roots, including sthā, snā, pā, vyadh, han and yudh.cakram, ciklidam, caknasaḥ. The Bālamanoramā derives cakram: ka after kṛñ; no guṇa, since ka is k-marked; doubling; then the usual changes in the reduplicated part. It glosses: "by which the slaying of enemies is done". The reading of the last word, caknasaḥ, is uncertain in the 1915 print.Watch out. Bhaṭṭoji's one phrase, "an exception to ghañ and ac", covers two cases. niś-ci would otherwise take ac (SK 3231), like plain cayaḥ. For the other roots, the Bālamanoramā and the Tattvabodhinī say, the affix set aside is ghañ. So write grahaḥ and gamaḥ, not grāhaḥ or gāmaḥ.
Rule. When an upasarga stands beside it, the root ad "to eat" takes ap. Bhaṭṭoji supplies only "ap comes". The Bālamanoramā calls it an exception to ghañ.
Example.
Bhaṭṭoji gives none here. The forms come under SK 3236: pra + ad + ap → ghasḷ for ad (SK 3236) → praghasaḥ.
Applies when / blocked by. Without an upasarga, ap does not come by this rule. Then ghañ comes: ghāsaḥ (see SK 3236).
Rule. And before ghañ or ap, the root ad is replaced by ghasḷ. The Bālamanoramā: "adaḥ" is carried forward from SK 3080, and "ghasḷ" from SK 2427. The ḷ of ghasḷ is a marker, leaving ghas.
Example.
pra + ad + ap → ghas for ad → pra + ghas + a → praghasaḥ.
vi + ad + ap → vighasaḥ "leavings of food".
ad + ghañ → ghas for ad → vṛddhi of the penultimate a (SK 2282) → ghāsaḥ "fodder".
Applies when / blocked by. Bhaṭṭoji asks: why "with an upasarga"? and answers ghāsaḥ. That condition belongs to SK 3235, not to this rule. Bhaṭṭoji leaves the steps implicit. Without an upasarga ap does not come, so ghañ comes, and ghasḷ still replaces ad before it. The two commentaries gloss vighasa differently. The Bālamanoramā: food left by guests. The Tattvabodhinī: what remains of the vaiśvadeva offering. Both gloss ghāsa by the Amarakośa, as young grass.
Watch out. This rule is from the second book. It is placed here because the ap-forms of SK 3235 need it.
Rule. When ni stands beside it, ad takes the affix ṇa, and ap too. Bhaṭṭoji: "with ni as upapada, ṇa comes after ad, and ap". The Bālamanoramā notes that ṇa in the sūtra has lost its first-case ending.
Example.
ni + ad + ṇa → the marker ṇ goes → vṛddhi of the penultimate a (SK 2282) → ni + ād + a → y for i before a vowel → nyādaḥ.
ni + ad + ap → ghas for ad (SK 3236) → nighasaḥ.
Applies when / blocked by. Bhaṭṭoji gives only the two forms. No one spells out the steps of nyādaḥ; they follow from the ṇ-marker. The Bālamanoramā gives the ghasḷ step for nighasaḥ.
Watch out. ghasḷ comes only before ghañ or ap. Before ṇa the root stays ad, so nyādaḥ has no gh.
Rule. After vyadh "to pierce" and jap "to mutter", when there is no upasarga, ap comes. Bhaṭṭoji supplies "ap comes".
Example.
vyadh + ap → vyadhaḥ.
jap + ap → japaḥ.
Applies when / blocked by. "But when there is an upasarga": āvyādhaḥ, upajāpaḥ "sowing dissension". These have the vṛddhi of ghañ. The vṛddhi shows the affix, though Bhaṭṭoji does not name it. The Tattvabodhinī glosses upajāpa as breaking of counsel. The Bālamanoramā glosses it as slander.
Watch out. "Without an upasarga" is carried into SK 3239.
svan + ap → svanaḥ / svan + ghañ → (vṛddhi) svānaḥ; with an upasarga only ghañ: prasvānaḥThese rules give ap after roots with a penultimate a. Such roots would otherwise take ghañ, with vṛddhi of that a (SK 2282). So most rules here yield pairs: an ap-form without vṛddhi and a ghañ-form with it. From SK 3239 to SK 3242 the ap is optional (vā). SK 3243 makes it obligatory (nityam) and so ends the option. The running example is svan "to sound". Without an upasarga it gives svanaḥ with ap or svānaḥ with ghañ. With pra only ghañ comes: prasvānaḥ.
Rule. After svan "to sound" and has "to laugh", ap comes optionally, when there is no upasarga. In the other option, ghañ. "No upasarga" is carried from SK 3238.
Example.
svan + ap → svanaḥ.
svan + ghañ → vṛddhi of the penultimate a (SK 2282) → svānaḥ.
Likewise hasaḥ / hāsaḥ.
Applies when / blocked by. Bhaṭṭoji: "no upasarga" does apply. With an upasarga, only ghañ: prasvānaḥ, prahāsaḥ.
Watch out. The option is open only for the bare roots. Once pra is added, vṛddhi is required.
Rule. After yam "to restrain", ap comes optionally when sam, upa, ni or vi stands beside it. By ca "and", also when there is no upasarga. Bhaṭṭoji: "with these, and without an upasarga, ap optionally after yam". The Bālamanoramā explains that ca adds "without an upasarga".
Example.
sam + yam + ap → saṃyamaḥ; sam + yam + ghañ → vṛddhi → saṃyāmaḥ.
Also upayamaḥ / upayāmaḥ, niyamaḥ / niyāmaḥ, viyamaḥ / viyāmaḥ.
Without an upasarga: yamaḥ / yāmaḥ.
Applies when / blocked by. In each pair the first form has ap and the second ghañ, as at SK 3239. Bhaṭṭoji does not name the affixes.
Watch out. With other upasargas, such as ā, the option is not given here.
Rule. With ni beside them, gad "to speak", nad "to roar", paṭh "to recite" and svan "to sound" take ap optionally. Bhaṭṭoji supplies "ap optionally".
Example.
ni + gad + ap → nigadaḥ; with ghañ → nigādaḥ.
Also ninadaḥ / ninādaḥ, nipaṭhaḥ / nipāṭhaḥ, nisvanaḥ / nisvānaḥ.
Applies when / blocked by. svan was already in SK 3239, but only without an upasarga. This rule extends its option to ni. The Tattvabodhinī quotes the Amarakośa's list of words for sound, which has both ninada and nināda.
Rule. After kvaṇ "to sound", ap comes optionally: with ni, without an upasarga, and also when the sound is of the vīṇā. Bhaṭṭoji: "the mention of the vīṇā is for the sake of pra and the rest".
Example.
ni + kvaṇ + ap → nikvaṇaḥ; with ghañ → nikvāṇaḥ.
Without an upasarga: kvaṇaḥ / kvāṇaḥ.
Of the vīṇā: prakvaṇaḥ / prakvāṇaḥ.
Applies when / blocked by. The Bālamanoramā sets out the point. "ni" comes from SK 3241 and "no upasarga" from SK 3238. Those two would cover the vīṇā's sound too, as long as no other upasarga is present. So "vīṇā" must add something new. It lets kvaṇ take the option with any upasarga, such as pra, when the vīṇā is meant.
Watch out. The ap-option with pra and the like holds only for the sound of the vīṇā. Elsewhere such forms would have ghañ alone; Bhaṭṭoji gives no example of this, so it is inferred.
Rule. After paṇ "to trade", when a quantity (parimāṇa) is meant, ap comes always (nityam). The Bālamanoramā: the word nityam stops the "optionally" of SK 3239.
Example.
mūlaka + paṇ + ap → mūlakapaṇaḥ "a bunch of radishes".
śākapaṇaḥ "a bunch of greens".
Bhaṭṭoji explains: for trade, a measured handful of radishes and the like is tied up. That is what the word denotes.
Applies when / blocked by. The Bālamanoramā: "for trade" means "for sale". The handful is what is traded, so ap is in the object sense: mūlakānāṃ paṇaḥ. Why "a quantity"? Bhaṭṭoji gives pāṇaḥ. Its vṛddhi marks it as a ghañ-form, though he does not name the affix.
Watch out. The word nityam runs on into SK 3244, so the ap there is also obligatory.
Rule. After mad "to be intoxicated", when a word that is not an upasarga stands beside it, ap comes, always. The Tattvabodhinī: "with a non-upasarga word as upapada, ap comes obligatorily, in bhāva and the other senses".
Example.
dhana + mad + ap → dhanamadaḥ "intoxication with wealth".
Applies when / blocked by. "But when there is an upasarga": unmādaḥ, with ud. Its vṛddhi shows ghañ; Bhaṭṭoji leaves the affix unnamed.
Rule. In the sense of joy (harṣa), the forms pramada and sammada are given by nipātana. The Bālamanoramā: they are ap-forms.
Example.
pra + mad + ap → pramadaḥ "joy".
sam + mad + ap → saṃmadaḥ "joy".
Bhaṭṭoji gives no positive example. The Bālamanoramā gives kanyānāṃ pramadaḥ "the joy of maidens" and kokilānāṃ sammadaḥ "the joy of cuckoos".
Applies when / blocked by. Why "joy"? Bhaṭṭoji gives pramādaḥ "negligence" and saṃmādaḥ, with the vṛddhi of ghañ. Scholars discuss why Pāṇini stated finished forms here. The Tattvabodhinī says it keeps out double-prefixed forms like prasaṃmada and saṃpramada. The Bālamanoramā says it captures the conventional sense.
Watch out. With an upasarga SK 3244 does not apply, so mad takes ghañ (pramādaḥ). Only in the sense of joy are pramada and saṃmada stated with ap.
Rule. After aj "to drive" with sam or ud beside it, when animals are concerned, ap comes. Bhaṭṭoji: aj with sam means a collection; with ud, driving. The 1915 print reads utpūrvaś ca "and preceded by ud".
Example.
sam + aj + ap → samajaḥ "a herd of animals".
ud + aj + ap → udajaḥ "the driving of animals".
Applies when / blocked by. SK 2292 replaces aj by vī, but it says "not when ghañ or ap follows". So vī does not come here. Why "animals"? Bhaṭṭoji gives samājo brāhmaṇānām "an assembly of brāhmaṇas" and udājaḥ kṣatriyāṇām "the driving forth of kṣatriyas" (so the 1915 print). These have ghañ, with vṛddhi. Ghañ is equally exempt from vī.
Watch out. Both forms keep the root aj. Do not substitute vī before ap or ghañ.
ni + hvā + ap → (saṃprasāraṇa hu; guṇa, av) nihavaḥ, against ghañ in pra + hvā + ghañ → (vṛddhi; āy) prahvāyaḥThis group gives ap in special senses. SK 3247 states the finished form glaha. SK 3248 gives ap after sṛ for conception. SK 3249–3252 give ap after hvā "to call", with saṃprasāraṇa. Saṃprasāraṇa replaces a semivowel by its vowel: the v of hvā becomes u. The running example is hvā with ni: nihavaḥ. With pra, a prefix not listed, ghañ comes instead, with vṛddhi and āy: prahvāyaḥ.
Rule. With reference to gaming, the form glaha is given by nipātana for the stake. Bhaṭṭoji: the word akṣa "dice" indicates gaming (devana). For what is to be taken as a stake (paṇa), glaha is the form.
Example.
grah + ap (SK 3234) → l for r by this nipātana → glahaḥ "the stake".
akṣasya glahaḥ "the stake in dicing".
vyātyukṣīm abhisaraṇaglahām adīvyan "they played at mutual water-splashing, with a rendezvous as the stake". The Bālamanoramā assigns the line to Māgha's poem.
Applies when / blocked by. The Bālamanoramā explains. If akṣa were the upapada, the plural akṣeṣu would not fit. So akṣa stands for gaming in general, and glaha is used even where no dice are thrown, as in the verse. The stake is what the loser must give the winner, so it is the object. The form is grah with ap by SK 3234, and l for r.
Why "gaming"? Bhaṭṭoji gives pādasya grahaḥ "the clasping of the feet", as in salutation. That has ap without the l.
Watch out. glaha is the stake, not the dice.
Rule. After sṛ "to go", in the sense of conception (prajana), ap comes. Bhaṭṭoji glosses prajana as the first taking of the embryo.
Example.
upa + sṛ + ap → guṇa with r → upa + sar + a → upasaraḥ.
gavām upasaraḥ "the covering of the cows".
Applies when / blocked by. Bhaṭṭoji asks: how then are avasaraḥ and prasaraḥ formed, where no conception is meant? With ghañ there would be vṛddhi. His answer: they take the affix gha in the locus sense, by SK 3296. The Bālamanoramā explains the locus: a time in which one moves.
Scholars discuss the word "first". The Tattvabodhinī says it follows the Kāśikā. Others, it reports, insist only on the taking of the embryo, first or not.
Watch out. avasaraḥ and prasaraḥ have no vṛddhi, but they are not ap-forms of this rule. They take gha (SK 3296).
Rule. After hvā "to call", with ni, abhi, upa or vi beside it, ap comes, and saṃprasāraṇa takes place. The Bālamanoramā calls it an exception to ghañ.
Example.
ni + hvā + ap → saṃprasāraṇa: u for the v → ni + h-u-ā + a
→ the u and ā become u (pūrvarūpa) → ni + hu + a
→ guṇa of the u → ni + ho + a → av (SK 61) → nihavaḥ.
Also abhihavaḥ, upahavaḥ, vihavaḥ.
These steps are the Bālamanoramā's.
Applies when / blocked by. Why "with these"? Bhaṭṭoji gives prahvāyaḥ. The Bālamanoramā derives it with ghañ, vṛddhi and the substitute āy.
Watch out. Saṃprasāraṇa here is taught by this rule itself. It does not come before every affix after hvā.
Rule. With āṅ beside it, when battle is meant, hvā takes ap with saṃprasāraṇa. hvā and saṃprasāraṇa come from SK 3249.
Example.
ā + hvā + ap → saṃprasāraṇa, pūrvarūpa, guṇa, av (as in SK 3249) → āhavaḥ "battle".
Bhaṭṭoji: "they are challenged in it" (āhūyante 'smin).
Applies when / blocked by. Why "battle"? Bhaṭṭoji gives āhvāyaḥ, with ghañ. The analysis "in it" makes āhava a locus; Bhaṭṭoji does not name the kāraka, so this is inferred.
Rule. The form āhāva is given by nipātana when a drinking-trough (nipāna) is meant. Bhaṭṭoji: for hvā with āṅ, saṃprasāraṇa, ap and vṛddhi, if a water-receptacle is meant.
Example.
ā + hvā + ap → saṃprasāraṇa → ā + hu + a → vṛddhi (by this nipātana) → ā + hau + a → āv (SK 61) → āhāvaḥ.
Bhaṭṭoji quotes a half-verse: "āhāva means a drinking-trough, a reservoir beside a well". He does not name its source; it is the Amarakośa's definition of āhāva.
Applies when / blocked by. The Bālamanoramā: Bhaṭṭoji's paraphrase states the net result of the nipātana.
Watch out. ap does not bring vṛddhi. The ā in āhāva comes only from the nipātana.
Rule. For hvā without an upasarga, in bhāva, saṃprasāraṇa and ap take place. Bhaṭṭoji's paraphrase supplies hvā, saṃprasāraṇa and ap.
Example.
hvā + ap → saṃprasāraṇa, pūrvarūpa, guṇa, av → havaḥ, the bare action of hvā.
Applies when / blocked by. The words "in bhāva" and "without an upasarga" run on into SK 3253, the first rule of the next group.
han + ap (bhāva) → vadhaḥ; abhra + han + ap (mūrti) → ghana → abhraghanaḥ; pari + han + ap (karaṇa) → gha → parighaḥ → (l) palighaḥThis group gives ap after han "to strike". Mostly a substitute replaces the root, or the whole root-and-affix is stated as a finished form. The substitutes are vadha (SK 3253), ghana (SK 3254, 3259, 3260) and gha (SK 3261). Several words are stated by nipātana: antarghana, praghaṇa, udghana, apaghana, upaghna, saṃgha, udgha, nigha. Where these rules do not apply, han takes ghañ. Then n becomes t (SK 2574), h becomes gh (SK 358), and the a takes vṛddhi: ghātaḥ. The running examples: vadhaḥ "slaying"; abhraghanaḥ "the solidity of a cloud"; parighaḥ "a bar", which may also be palighaḥ.
Rule. After han without an upasarga, in bhāva, ap comes, and vadha replaces han. Bhaṭṭoji: the substitute vadha is end-acute (antodātta). "In bhāva" and "without an upasarga" come from SK 3252.
Example.
han + ap → vadha for han → vadha + a → the final a of vadha is lost before the affix's a (a step no commentator states) → vadhaḥ "slaying".
vadhena dasyum "with the slaying, the dasyu". The Bālamanoramā: a Ṛgvedic mantra, cited to show the accent.
"By ca, ghañ too": han + ghañ → t for n (SK 2574), gh for h → vṛddhi → ghātaḥ.
Applies when / blocked by. Scholars discuss the word ca. Bhaṭṭoji takes it to bring in ghañ as well. The Bālamanoramā takes it as joining vadha to ap, out of its natural order, and its text lacks "by ca, ghañ". Both statements stand in our text and in the 1915 print. The Bālamanoramā adds why the rule is needed. A separate root vadh could give the form by the vārttika "ka in the sense of ghañ" (SK 3234). But this rule fixes the condition "without an upasarga".
Watch out. vadha is a substitute for han, not a separate root here.
Rule. When solidity (mūrti) is meant, han takes ap, and ghana replaces it. Bhaṭṭoji glosses mūrti as hardness (kāṭhiṇya). The Bālamanoramā: "in bhāva" is carried forward.
Example.
abhra + han + ap → ghana for han → abhraghanaḥ "the solidity of a cloud".
Applies when / blocked by. Bhaṭṭoji asks: how then is saindhavaghanam ānaya "bring a block of rock-salt" said? One cannot bring a hardness. His answer: the word for the property indicates the thing that bears it. The Bālamanoramā puts the objection the same way.
Watch out. Because the affix is in bhāva, abhraghana means the hardness of a cloud, not the cloud. The thing-sense comes only by indication.
Rule. The form antarghana is given by nipātana when a place is meant. Bhaṭṭoji: it is the name of a particular village of the Vāhīkas.
Example.
antar + han + ap → ghana for the base → antarghanaḥ.
The Bālamanoramā gives these steps.
Applies when / blocked by. In a name, SK 857 would make the n into ṇ after antar. The Bālamanoramā: the nipātana blocks that. Bhaṭṭoji records antarghaṇa as a variant reading.
Rule. The forms praghaṇa and praghāṇa are given by nipātana when a part of a house is meant. Bhaṭṭoji: at the doorway are two prakoṣṭhas, the alindas, an inner and an outer. This nipātana is for the outer one.
Example.
pra + han + ap (object sense) → praghaṇaḥ.
In the other option, vṛddhi → praghāṇaḥ.
Bhaṭṭoji: "it is struck hard by the feet of people entering".
Applies when / blocked by. The Bālamanoramā: the prakoṣṭhas are oblong raised platforms fit for entering. The Tattvabodhinī quotes the Amarakośa: praghāṇa, praghaṇa and alinda mean the chamber outside the door.
Watch out. The ṇ of praghāṇa, after a long ā, can come only from the nipātana; that of praghaṇa could also come from SK 359. No commentator says either; this is an inference.
Rule. The form udghana is given by nipātana for atyādhāna. Bhaṭṭoji glosses atyādhāna as placing on top. The Bālamanoramā: ap after han with ud, with gh for h.
Example.
ud + han + ap (locus) → gh for h → udghanaḥ.
Bhaṭṭoji: the block of wood on which other pieces are placed and then pared.
Applies when / blocked by. Bhaṭṭoji: ap is in the locus sense (adhikaraṇa), "that on which".
Rule. The form apaghana is given by nipātana for a limb (aṅga). Bhaṭṭoji: a limb is a part of the body. Here, "they say", not every part but the hand and the foot.
Example.
apa + han + ap (instrument) → gh for h → apaghanaḥ.
The Bālamanoramā: "that with which one strikes away" (apahanyate 'nena).
Applies when / blocked by. Bhaṭṭoji: ap is in the instrument sense (karaṇa). "Something else is apaghātaḥ": where no limb is meant, the ordinary form comes. Its vṛddhi shows ghañ, though Bhaṭṭoji does not name the affix. The Bālamanoramā says the narrowing to hand and foot rests only on the explanation (vyākhyāna).
Rule. With ayas "iron", vi or dru "tree" beside it, han takes ap in the instrument sense, and ghana replaces it.
Example.
ayas + han + ap → ghana → ayoghanaḥ "a hammer": "iron is struck with it".
vighanaḥ, drughanaḥ.
Applies when / blocked by. Bhaṭṭoji: "according to some, drughaṇaḥ", with ṇ by SK 857. This is the name of the axe; dru is "tree". The Bālamanoramā explains the two views. On one, drughana is not a name, so SK 857 does not apply. On the other, it is a name, and the ṇ comes.
Watch out. "According to some" reports a view about the word, not a different reading of the text.
Rule. With stamba "clump" beside it, han takes ka in the instrument sense. In the other option it takes ap, with ghana. Bhaṭṭoji: "only in the instrument sense" (karaṇa ity eva).
Example.
stamba + han + ka → the penultimate a is lost (SK 2363) → stamba + hn + a → gh for h → stambaghnaḥ.
stamba + han + ap → ghana → stambaghanaḥ.
Applies when / blocked by. The Bālamanoramā: the substitute ghana goes only with ap. So it belongs to the second option, not to ka. Outside the instrument sense: stambaghātaḥ, with ghañ and the velar gh (the Bālamanoramā).
Watch out. ka is k-marked, so the penultimate a of han drops: -ghna, not -ghana.
Rule. With pari beside it, han takes ap in the instrument sense, and gha replaces it. The Bālamanoramā: "karaṇe", "hanaḥ" and "ap" are carried forward.
Example.
pari + han + ap → gha for han → parighaḥ "a bar": "one strikes with it" (parihanyate 'nena).
Applies when / blocked by. The r of pari may become l before gha by the next rule: palighaḥ.
Rule. Before the word gha or the word aṅka, l optionally replaces the r of pari. The Bālamanoramā: "ro laḥ" comes from SK 2350, and "optionally" from SK 2541.
Example.
parighaḥ → l for r → palighaḥ; or it stays parighaḥ.
pari + aṅka → y for i → paryaṅkaḥ; with l, palyaṅkaḥ.
Applies when / blocked by. Bhaṭṭoji: the word gha here is not the technical name gha of SK 2003 (the comparative and superlative affixes tarap and tamap). This rests on the explanation (vyākhyāna). The Bālamanoramā states the objection. By a maxim, a technical name is preferred to a plain word. So gha might mean tarap and tamap. The explanation settles that it means the word gha.
Watch out. This is a rule of the eighth book, placed here for palighaḥ.
Rule. The form upaghna is given by nipātana when āśraya is meant. Bhaṭṭoji: ap after han with upa, and the penultimate (upadhā) is lost. The word āśraya indicates proximity.
Example.
upa + han + ap → penultimate a lost → upa + hn + a → gh → upaghna.
parvatopaghnaḥ "the place near the mountain": "it is reached by the mountain", that is, approached through nearness.
Applies when / blocked by. The Bālamanoramā: Bhaṭṭoji's paraphrase states the result of the nipātana. That āśraya means proximity rests on the explanation.
Rule. The forms saṃgha and udgha are given by nipātana in the senses of a group and of praise, respectively. The Bālamanoramā: with sam or ud, ap after han; the loss of the ṭi (the final vowel with what follows it) and the gh are by nipātana.
Example.
sam + han + ap (bhāva) → saṃghaḥ "a group". Bhaṭṭoji: "coming together".
ud + han + ap (object) → udghaḥ "excellent". Bhaṭṭoji: "it is known as excellent".
Applies when / blocked by. Bhaṭṭoji explains the sense "known". Roots meaning "go" also mean "know", and han is one of them. The Bālamanoramā cites the maxim: all roots of going mean knowing. The Tattvabodhinī quotes the Amarakośa, where udgha is among the words for "excellent". Our text writes the form uddhaḥ here; the sūtra's own form is udgha.
Rule. The form nigha is given by nipātana for nimita, "measured all round". The Bālamanoramā: here too ap, the loss of the ṭi and gh are by nipātana.
Example.
ni + han + ap → nighaḥ.
nighā vṛkṣāḥ "trees of equal height and girth". Bhaṭṭoji: they are known as without distinction. The 1915 print has the plural verb hanyante.
Applies when / blocked by. The Bālamanoramā: ni here means "all round", and han again means "know".
ḍupacaṣ → pac + ktri → paktri → (map, SK 1570) paktrimam; yaj + naṅ → (ñ by SK 111) yajñaḥ; pra + dhā + ki → (ā lost) pradhiḥThe last group of Part I gives five small affixes. ktri and athuc come after roots with a special marker (it) in their dictionary form: ḍu or ṭu. Both come only in bhāva, by their nature. naṅ and nan come after a few named roots. ki comes after the ghu roots, the dā- and dhā-type roots (SK 2373). The running examples: ḍupacaṣ "to cook" gives paktrimam "produced by cooking"; yaj "to sacrifice" gives yajñaḥ "sacrifice"; dhā with pra gives pradhiḥ "felloe".
Rule. After a root that has ḍu as its marker, the affix ktri comes. Bhaṭṭoji: it comes only in bhāva, by its nature. The Bālamanoramā: "bhāve" is carried forward, but not "akartari ca kārake". Its k is a marker.
Example.
ḍupacaṣ → the markers ḍu and ṣ go → pac + ktri → pac + tri → k for c (not stated by Bhaṭṭoji or his commentators; it is inferred) → paktri
→ map always follows (SK 1570) → paktrimam "produced by cooking" (pākena nirvṛttam).
ḍuvap → vap + tri → saṃprasāraṇa before a k-marked affix (SK 2409, the Bālamanoramā) → uptri → uptrimam "produced by sowing".
Applies when / blocked by. Bhaṭṭoji: because of the word nitya "always" in SK 1570 (ktrer mam nityam), ktri is always followed by map. The Bālamanoramā: ktri is never used alone.
For this reason the word is never analysed with the ktri-word itself (paktryā nirvṛttam), but with another word, pāka: pākena nirvṛttam. So the Bālamanoramā; our text and the 1915 print lack the "not", and scholars discuss this. The Tattvabodhinī supports bhāva-only from the same analysis. It is pākena, with the action noun, not pakvena "by the cooked thing".
Watch out. ktri never stands alone. Write paktrimam, never paktriḥ.
Rule. After a root that has ṭu as its marker, the affix athuc comes. Bhaṭṭoji: this too comes only in bhāva, by its nature.
Example.
ṭuvepṛ → the markers ṭu and ṛ go → vep + athuc → the marker c goes → vep + athu → vepathuḥ "trembling".
śvayathuḥ "swelling".
Applies when / blocked by. Bhaṭṭoji gives śvayathuḥ without its root or steps.
Rule. After yaj "to sacrifice", yāc "to ask", yat "to strive", vicch, pracch "to ask" and rakṣ "to protect", the affix naṅ comes. The Tattvabodhinī: "bhāva and the non-agent kāraka" continue. That is why, in yajñena yajñam ayajanta devāḥ, yajña is explained as "what is offered".
Example.
yaj + naṅ → the marker ṅ goes → yaj + na → ñ for n after j (SK 111, the Bālamanoramā) → yajñaḥ "sacrifice".
yāc + na → ñ for n after c, the same step as in yajñaḥ (SK 111) → with the feminine ending ā, yācñā "request". yat + na → yatnaḥ "effort".
vicch + na → ś for cch (SK 2561, the Tattvabodhinī) → viśnaḥ.
pracch + na → ś for cch (SK 2561), as in viśnaḥ (the Bālamanoramā) → praśnaḥ "question".
rakṣ + na → ṇ for n after ṣ (a step Bhaṭṭoji leaves unstated) → rakṣṇaḥ "protection".
Applies when / blocked by. Why no saṃprasāraṇa in praśnaḥ? Before a ṅ-marked affix, SK 2412 would turn the r of pracch into ṛ. Bhaṭṭoji: the wording praśne cāsannakāle (SK 2777) is an indicator (jñāpaka). Pāṇini himself writes praśna there, without saṃprasāraṇa. The Bālamanoramā spells out the argument this way.
Then what is the ṅ-marker for? Bhaṭṭoji: it blocks guṇa in viśnaḥ.
Watch out. yācñā is feminine. Bhaṭṭoji gives it without comment.
Rule. After svap "to sleep", the affix nan comes. The Bālamanoramā: in bhāva and the non-agent kāraka. The marker n is for the accent.
Example.
svap + nan → the marker n goes → svap + na → svapnaḥ "sleep; dream".
Watch out. nan and naṅ both leave na. They differ only in the marker, and so in its effect.
Rule. When an upasarga stands beside it, a ghu root takes the affix ki. The Bālamanoramā: in bhāva and the non-agent kāraka. The k-marker is for the loss of the root's ā.
Example.
pra + dhā + ki → the marker k goes → pra + dhā + i → ā lost (SK 2372) → pradhiḥ "felloe".
antardhiḥ "concealment".
upādhiḥ: "that by which something is placed near" (upādhīyate 'nena).
Applies when / blocked by. antar is not in the list of upasargas. The Tattvabodhinī: a vārttika gives it upasarga status for this rule. Scholars discuss whether ki can denote the agent. Mādhava allows it by bāhulaka. The Tattvabodhinī rejects this: Bhaṭṭoji's instrument analysis of upādhi hints at that.
Rule. And when an object word (karman) stands beside it, a ghu root takes ki in the locus sense (adhikaraṇa). The Bālamanoramā: the object is the upapada and the locus is the meaning, by the explanation.
Example.
jala + dhā + ki → ā lost → jaladhiḥ "the ocean": "waters are placed in it" (jalāni dhīyante 'smin).
Applies when / blocked by. "ghu" and "ki" come from SK 3270. Here no upasarga is needed. The object word takes the place of the upasarga as upapada.
Watch out. Do not read karmaṇi as the sense of the affix. The affix denotes the locus; the object is the word beside the root. Part II opens with the feminine ktin (SK 3272).
kṛ + ktin → kṛ-ti → kṛtiḥ; pra + sthā + ktin → (i, SK 3074) prasthitiḥThe second half of the chapter opens with a new heading, striyām "in the feminine". Every affix taught by the sūtras from SK 3272 to SK 3289 makes a feminine noun. (Some vārttikas under SK 3285 are not tied to the heading.) The two older headings still hold: bhāve (SK 3184), the bare action, and akartari ca kārake (SK 3186), a kāraka other than the agent. So these affixes make feminine action nouns, or feminine nouns for an instrument, a place and the like. The general feminine affix is ktin, which leaves ti. Beside it come kyap (ya), śa (a) and a few finished forms stated in the sūtras (nipātana). The running example is kṛ "to do": kṛ + ktin gives kṛtiḥ "making". With sthā "to stand" and pra, the root vowel becomes i by SK 3074: prasthitiḥ "setting out".
One fact shapes the whole feminine section. SK 2830 (vā 'sarūpo 'striyām) lets a dissimilar exception block the general affix only optionally, but "not in the feminine" (astriyām). So inside this section an exception blocks ktin completely. Where both forms are wanted, Pāṇini or a vārttika has to say so separately. The Bālamanoramā and the Tattvabodhinī point this out at SK 3272, 3277 and 3287.
Rule. In the feminine, the affix ktin comes after a root, in the bare action and in a kāraka other than the agent. striyām is a heading. The Bālamanoramā says it runs up to just before kṛtyalyuṭo bahulam (SK 2841), which is re-cited at SK 3289. Bhaṭṭoji calls ktin an exception (apavāda) to ghañ. The Bālamanoramā names that ghañ: the one of halaś ca (SK 3300), taught later in this part. So in feminine bhāva ktin comes; in masculine bhāva, ghañ.
Example.
kṛ + ktin → the markers k and n go → kṛ + ti → kṛti → kṛtiḥ "making".
ci + ktin → citiḥ "piling". stu + ktin → stutiḥ "praise".
sphāy + ktin → sphātiḥ "swelling". The Bālamanoramā: the y is lost by lopo vyor vali, and SK 3163 (titutra…) bars iṭ.
Applies when / blocked by. Bhaṭṭoji: "ac and ap, however, it sets aside, because it is later (paratva)". ac comes after i-final roots (SK 3231) and ap after ṛ- and u-final roots (SK 3232). The Tattvabodhinī explains. Their own field is the masculine: cayaḥ, lavaḥ. After ci and stu in the feminine, both ktin and ac or ap could apply; ktin wins as the later rule. Hence citiḥ, stutiḥ. (The 1915 print reads the dual bādhete, "they set aside", which would make ac and ap the winners; citiḥ and stutiḥ show that ktin wins.)
Bhaṭṭoji calls the word sphītikāma an error. Both commentators give the reason: sphī replaces sphāy only before a niṣṭhā (SK 3044), not before ktin. He then offers a rescue. From the kta-form sphīta, ṇic is added in the sense "to make or call it sphīta". Then the uṇādi affix i comes by aca iḥ (uṇādi 578). The Bālamanoramā spells it out: this gives sphīti, and "one who desires sphīti" is sphītikāma.
Four vārttikas follow.
śrūyate 'nayā "one hears by it" → śrutiḥ "the ear". From yaj: yaj + ktin → saṃprasāraṇa (SK 2409) → ij + ti → ṣ for j (vraścabhrasja…, says the Tattvabodhinī) → iṣ + ti → ṭ for t (SK 113) → iṣṭiḥ "that by which one sacrifices". From iṣ, also iṣṭiḥ; from stu, stutiḥ.kṝ + ti → i with r for ṝ → kir + ti → lengthening → kīr + ti → n for t → kīr + ni → ṇ → kīrṇiḥ. Likewise gīrṇiḥ, lūniḥ, dhūniḥ, pūniḥ. Bhaṭṭoji adds two more. SK 3073 (hlādo niṣṭhāyām) is split, so that hlādaḥ alone gives a short vowel before ktin too: prahlanniḥ. And by ti ca (SK 3037): cūrtiḥ, phultiḥ. In both, u replaces the a of car and phal. Which rule lengthens the u of cūrti, the Tattvabodhinī and the Bālamanoramā say differently; scholars discuss this.apacitiḥ "honouring".sampat "prosperity", vipat "misfortune". A further vārttika: "ktin also is desired": sampattiḥ, vipattiḥ. Both commentators explain why this must be said. In the feminine the vāsarūpa option (SK 2830) does not work, so kvip would block ktin entirely.Watch out. ktin is the feminine counterpart of masculine ghañ, ac and ap. The same root can make both: masculine cayaḥ, feminine citiḥ.
Rule. After sthā "to stand", gā "to sing", pā "to drink" and pac "to cook", ktin comes in feminine bhāva. Bhaṭṭoji: it is an exception to aṅ. Both commentators name the two *aṅ*s. One is that of ātaś copasarge (SK 3283), which would come after sthā, gā, pā with an upasarga. The other is that of ṣidbhidādibhyo 'ṅ (SK 3281), which would come after pac. The Bālamanoramā notes that kvip from the vārttika above does not run on here. It is not marked svarita, the mark that lets a word run on.
Example.
pra + sthā + ktin → i for ā before a k-marked affix (SK 3074) → pra + sthi + ti → prasthitiḥ "setting out".
Also upasthitiḥ "presence".
sam + gai + ktin → ī (SK 2462) → saṅgītiḥ "singing together".
sampītiḥ "drinking together". pac + ktin → paktiḥ "cooking".
Applies when / blocked by. How then are avasthā and saṃsthā formed, with aṅ? This rule should block it. Bhaṭṭoji answers: from the indicator (jñāpaka) vyavasthāyām. A jñāpaka is a wording in Pāṇini that shows a point he did not state. Bhaṭṭoji does not say where the word stands; it is SK 218, where Pāṇini himself uses vyavasthā.
The Tattvabodhinī limits the exception. An exception stated before other rules sets aside only those immediately following it. So ktin blocks the aṅ rules but not the later ṇvul and iñ (SK 3285, 3287). Those still come after sthā: the Bhāṣya says kāṃ tvaṃ sthāyikām….
Rule. Six words are stated as finished ktin-forms (nipātana): ūti, yūti, jūti, sāti, heti, kīrti. Bhaṭṭoji: "the statement is for the sake of the accent". The word udātta "high-pitched" runs on from A 3.3.96, a Vedic sūtra that the SK gives at SK 3420. So the ktin here is udātta, and these words are end-acute (antodātta). Both commentators add: otherwise, by ñnity ādir nityam (SK 3686), a ktin-word would be acute on its first syllable.
Example.
av + ktin → ūṭh by SK 2654 → ū + ti → ūtiḥ. The Bālamanoramā: ūṭh replaces the v together with the a before it.
yūtiḥ, jūtiḥ: the Bālamanoramā takes them from yu and ju. Their long vowel is also by nipātana.
sātiḥ: from so (syati). SK 3074 would give i; its absence is by nipātana. Or from san (sanoti): ā comes by SK 2504, and the nipātana is only for the accent.
hetiḥ: from han or from hi. The Tattvabodhinī: from han, the n becomes i; from hi, there is guṇa.
kīrtiḥ: the Tattvabodhinī takes it from the ṇic-stem kīrtay-. There yuc would come by SK 3284; ktin is set down instead.
Watch out. These are not new meanings. They are ordinary ktin-words whose accent, and in some cases whose shape, the nipātana fixes.
Rule. After vraj "to go" and yaj "to sacrifice", the affix kyap comes in feminine bhāva. kyap leaves ya. The Tattvabodhinī: udātta still runs on. The p of kyap is there for the augment tuk (t) in the following rules (SK 2858).
Example.
vraj + kyap → vraj + ya → vrajyā "wandering".
yaj + kyap → saṃprasāraṇa (SK 2409, the Bālamanoramā) → ij + ya → ijyā "sacrifice".
Rule. When a name (saṃjñā) is meant, kyap comes in the feminine after nine roots: aj with sam, sad with ni, pat with ni, man, vid, ṣuñ (su "to press"), śīṅ "to lie", bhṛñ "to bear" and iṇ "to go". It comes in bhāva and the other non-agent senses, and it is udātta.
Example.
samajanty asyām "they come together in it" → samajyā "an assembly".
niṣīdanty asyām "they sit down in it" → niṣadyā "a market". The Bālamanoramā: the ṣ is by SK 2271.
nipatanty asyām "they fall down on it" → nipatyā "slippery ground".
manyate 'nayā "one knows by it" → manyā "the vein at the side of the neck". The Tattvabodhinī: by it one is known to be angry.
vidanty anayā "they know by it" → vidyā "knowledge".
sutyā "the soma-pressing". śayyā "a bed". bhṛtyā "maintenance".
īyate 'nayā "one goes by it": i + kyap → the augment tuk → i + t + ya → ityā "a palanquin".
The Bālamanoramā: sutyā likewise has tuk.
Applies when / blocked by. A vārttika: vī does not replace aj before kyap. The Bālamanoramā: otherwise SK 2292 would give vī. Hence samajyā, not a form with vī.
Why "when a name is meant"? The Tattvabodhinī: otherwise ktin comes: matiḥ, smṛtiḥ.
Rule. Bhaṭṭoji splits this sūtra into two rules (yogavibhāga). First, kṛñaḥ: after kṛ, kyap comes in the feminine. Second, śa ca: after kṛ, the affix śa comes, and by ca "and" also ktin.
Example.
kṛ + kyap → tuk → kṛ + t + ya → kṛtyā "a deed; a rite".
kṛ + śa → riṅ for ṛ (SK 2367) → kri + a → iyaṅ → kriy + a → kriyā "action".
kṛ + ktin → kṛtiḥ "making".
The Bālamanoramā explains kriyā. śa is ś-marked, so it is sārvadhātuka and counts as ṅ-marked. So there is no guṇa, and iyaṅ follows.
Applies when / blocked by. Why split? The Bālamanoramā: read as one sūtra, ca would bring in kyap, giving "after kṛ, śa and kyap". Then kṛtiḥ with ktin could not come, since the vāsarūpa option does not work in the feminine. Split, the first part gives kyap, and the ca of the second part brings ktin.
Watch out. All three feminines of kṛ are correct: kṛtyā, kriyā, kṛtiḥ.
Rule. The form icchā is a nipātana. From iṣ "to wish", in bhāva only, the affix śa comes, and the expected yak does not. Both commentators note "in bhāva only", not in other kārakas. That, says the Tattvabodhinī, is why iṣṭiḥ was given for the instrument at SK 3272. (Bhaṭṭoji writes the root as īṣ; the Tattvabodhinī reads iṣ, the root of icchati.)
Example.
iṣ + śa (bhāva) → no yak → icchā "wish".
The Bālamanoramā: yak would otherwise come, since śa is sārvadhātuka.
Applies when / blocked by. A vārttika adds four more (upasaṃkhyāna, "they should be counted in"): paricaryā, parisaryā, mṛgayā, aṭāṭyā. In these, śa and yak are set down together.
paricaryā "attendance, worship", from car with pari.parisaryā "moving about", from sṛ with pari. Here guṇa too is by nipātana. The Bālamanoramā: śa counts as ṅ-marked and would block guṇa.mṛgayā "hunting". In the 1915 print Bhaṭṭoji adds that mṛga "to search" is an a-final root of the curādi class, and that keeping its a is also set down. He does not say which rule would remove it; it is ato lopaḥ (SK 2308), before ṇic. Then with yak after śa, ṇi is lost: mṛga + ṇic + śa + yak → ṇi lost → mṛga + ya + ā → mṛgayā.aṭāṭyā "roaming about", from aṭ. With yak after śa, the piece ṭya is doubled. The y of the first copy is removed, and its vowel is lengthened: a + ṭya-ṭya → a + ṭa-ṭya → a + ṭā-ṭya → aṭāṭyā. The Tattvabodhinī: the y goes by nipātana, since this doubling is not one that makes an abhyāsa. Both commentators give another form too: with a by SK 3279 after the intensive stem aṭāṭya (the root doubled, with the intensive affix yaṅ), aṭāṭā.A second vārttika: after jāgṛ "to wake", a comes optionally; in the other option, śa with yak. jāgarā, jāgaryā "wakefulness". Both commentators: in jāgaryā the guṇa is by SK 2480.
Watch out. icchā has no yak; paricaryā and its companions have it. Both kinds are nipātanas, so neither can be extended to other roots.
cikīrṣa + a → cikīrṣā; jṝ + aṅ → (guṇa) jarā; kāri + yuc → kār-ana → kāraṇāThree more affixes give feminines ending in ā. The affix a comes after a root that itself ends in an affix, such as a desiderative, and after a consonant-final root with a heavy vowel. The affix aṅ (a with the marker ṅ) comes after roots marked with ṣ, after bhid and its group, after five curādi roots, and after ā-final roots with an upasarga. The affix yuc (yu, which becomes ana) comes after ṇic-stems and two other roots. Each is an exception to ktin in its own field. Among themselves, yuc (SK 3284) sets aside the a of SK 3279, and the aṅ of SK 3282 sets aside yuc. The running examples are cikīrṣā "the wish to do", jarā "old age" and kāraṇā, from the causative of kṛ.
Rule. After a root that ends in an affix (pratyaya), the affix a comes in the feminine. The Bālamanoramā: bhāva and the non-agent kārakas still hold. Such roots are derived stems: a desiderative in san, a denominative in kāmyac, and the like.
Example.
cikīrṣa + a → cikīrṣā "the wish to do" (the desiderative of kṛ).
putrakāmya + a → putrakāmyā "the wish for a son" (kāmyac after putra).
Watch out. A ṇic-stem also ends in an affix. But the Bālamanoramā says SK 3284 gives yuc after it, setting this a aside. So kāraṇā, from the causative of kṛ, has yuc, not this a.
Rule. After a root that ends in a consonant (hal) and contains a heavy (guru) vowel, a comes in the feminine. A heavy vowel is a long vowel, or a short one before a consonant cluster. Both commentators explain the wording. "Heavy" belongs only to vowels, and the root is consonant-final, so it cannot itself be heavy. "Containing a heavy vowel" (gurumat) is meant. Bhaṭṭoji glosses it so.
Example.
īh + a → īhā "effort".
ūh + a → ūhā "reasoning".
Applies when / blocked by. Why "heavy"? bhaktiḥ: bhaj has a light vowel, so ktin comes. Why "consonant-final"? nītiḥ: nī ends in a vowel, so ktin comes.
A vārttika: only after a root that is seṭ in the niṣṭhā. (A seṭ root takes the augment iṭ.) So the a does not come in āptiḥ or in dīptiḥ. The Bālamanoramā: āp is not seṭ before the niṣṭhā, so only ktin comes. For dīp Bhaṭṭoji does not give the reason; it is SK 3039: dīp is ī-marked, so it takes no iṭ in the niṣṭhā (dīptaḥ). Bhaṭṭoji also cites titutra… (SK 3163) for dīptiḥ. That rule only explains why the ktin itself takes no iṭ.
A second vārttika amends SK 3163: "except grah and its group". So these roots do take iṭ before ktin: nigṛhītiḥ, nipaṭhitiḥ. The Bālamanoramā lists the group from the Bhāṣya: nigṛhīti, upasnihiti, nikuciti, nipaṭhiti.
Watch out. Three tests must all pass for a: a consonant at the end, a heavy vowel, and iṭ in the niṣṭhā. āptiḥ passes the first two and fails the third.
Rule. After roots marked with ṣ (ṣit) and after bhid "to split" and its group, aṅ comes in the feminine. The Bālamanoramā: in bhāva and the non-agent kārakas.
Example.
jṝṣ + aṅ → guṇa by SK 2406 (ṛdṛśo 'ṅi guṇaḥ) → jar + a → jarā "old age".
trapūṣ + aṅ → trapā "shame".
bhidā, chidā, mṛjā.
The Bālamanoramā on jarā: the ṅ of aṅ would block guṇa (SK 2217), but SK 2406 overrides that.
Applies when / blocked by. Bhaṭṭoji: bhidā is used only for splitting. bhittiḥ "a wall" is a different word. The Tattvabodhinī: bhitti is "what is split", with ktin. It adds that chidā means only dividing in two, and that mṛj is listed because it is not ṣit.
A gaṇasūtra (a rule read inside the list): "and saṃprasāraṇa of krap". krap + aṅ → kṛpā "compassion".
The Tattvabodhinī raises labdhiḥ. labh is ṣit, so only aṅ should come. It answers that ktin is also allowed, from the Bhāṣya's own use of anupalabdheḥ, or by bāhulaka (the latitude of "variously").
Rule. After cinti "to think", pūji "to worship", kathi "to tell", kumbi and carc "to discuss", aṅ comes in the feminine. Bhaṭṭoji: it is an exception to yuc. The Tattvabodhinī: the yuc of SK 3284. The Bālamanoramā: these are curādi roots, and ṇi is lost before aṅ.
Example.
cinti + aṅ → ṇi lost → cint + a → cintā "thought".
pūjā "worship", kathā "story", kumbā, carcā "discussion".
Applies when / blocked by. Both commentators say ca takes in an unlisted root: tul (tolayati), giving tulā. Bhaṭṭoji does not give it.
Rule. After a root ending in ā, with an upasarga beside it, aṅ comes in the feminine. Bhaṭṭoji: it is an exception to ktin.
Example.
pra + dā + aṅ → the root's ā is lost (SK 2372, the Bālamanoramā) → pra + d + a → pradā "giving".
upadā "a present".
Applies when / blocked by. Bhaṭṭoji adds that śrat and antar are treated as upasargas. So śraddhā "faith" and antardhā "concealment" have aṅ. And by upasarge ghoḥ kiḥ (SK 3270), ki also comes: antardhiḥ. The Bālamanoramā cites two vārttikas behind this. It adds that antar also brings ṇ as an upasarga does: antarṇayati.
Watch out. SK 2898 has the same wording (ātaś copasarge, A 3.1.136). That rule gives ka in the agent (prajñaḥ). This one gives feminine aṅ in bhāva (pradā).
Rule. After a root ending in ṇi, after ās "to sit" and after śranth "to loosen", yuc comes in the feminine. yuc leaves yu, which becomes ana (SK 1247). Bhaṭṭoji: it is an exception to a. The Bālamanoramā names the a of SK 3279, for the ṇi-final roots. For ās and śranth no commentator names the rule; it would be the a of SK 3280.
Example.
kāri + yuc → kāri + yu → yu → ana (SK 1247) → ṇi lost → kār + ana → kāraṇā.
Likewise hāraṇā, āsanā "sitting", śranthanā "loosening".
Applies when / blocked by. The Tattvabodhinī: āsanā is "that in which one sits". ṛhalor ṇyat also gives āsyā. This is no breach of the feminine bar on the vāsarūpa option. The exception yuc is in the feminine section, but the general rule ṇyat is not.
Three vārttikas:
ghaṭṭanā, vandanā "salutation", vedanā. The Tattvabodhinī: vid "to gain" is meant; vid "to know" gives saṃvittiḥ.anveṣaṇā "search". The Tattvabodhinī: in the sense of wish, only icchā (SK 3278).paryeṣaṇā / parīṣṭiḥ.Watch out. SK 3282 already took five ṇi-final curādi roots out of this rule. So cintā, not a yuc-form.
pra + chard + ṇvul → (vu → aka) pra-chard-aka → pracchardikāThe last affixes of the feminine section come in special settings. ṇvul, the agent affix of chapter 67 (SK 2895), returns here in the feminine for names of diseases and other names. iñ (i) comes in a question and its reply, ṇvuc (aka again) for a turn, a desert, a debt or an arising, and ani in a curse. Then Bhaṭṭoji re-cites two sūtras taught earlier, SK 2841 and SK 3090, in their Aṣṭādhyāyī place. With them the feminine heading closes, and so do the headings bhāve and akartari ca kārake. The running example is chard "to vomit" with pra: pra + chard + ṇvul → vu becomes aka → pracchardikā "vomiting", the name of a disease.
Rule. When the name of a disease is formed, ṇvul comes after a root in the feminine, variously (bahulam). The Tattvabodhinī: it is an exception to ktin and the rest. The Bālamanoramā: "in the feminine" still holds.
Example.
pra + chard + ṇvul → markers go → pra + chard + vu → vu → aka (SK 1247) → pra-chard-aka → (feminine -ikā) pracchardikā "vomiting".
pravāhikā "dysentery". vicarcikā "scabies".
Applies when / blocked by. "Variously" lets the affix fail in some cases: śirortiḥ "headache". Both commentators: this is ktin after ard "to hurt", with no iṭ by SK 3163.
Several vārttikas follow. They are not all tied to disease.
āsikā "sitting", śāyikā "lying down".paciḥ, pacatiḥ "the root pac". Both commentators: pacatiḥ has śap, although it does not denote the agent. The Bālamanoramā: "in the feminine" is not connected here.akāraḥ "the sound a", kakāraḥ "the sound k". Bhaṭṭoji: "when designated" still holds. Both commentators: the a in ka is only for pronouncing it. kāra counts as a kṛt, so the word is a nominal stem. Its k is not a marker.rephaḥ "the sound r". The Bālamanoramā: "in the feminine" does not continue here. So the vāsarūpa option works, and rakāraḥ with kāra is also correct.matvarthīyaḥ. Bhaṭṭoji: the a of matvartha is lost by the word "variously".ājiḥ, ātiḥ. The Bālamanoramā: in ājiḥ, vī does not replace aj, by "variously".vāpiḥ, vāsiḥ. Bhaṭṭoji: the difference from iṇ is in the accent. The Bālamanoramā: iñ is ñ-marked, so SK 3686 gives the first syllable the acute.kṛṣiḥ, giriḥ. The Bālamanoramā: ik is k-marked, so there is no guṇa.The Tattvabodhinī: iṇ and the affixes after it are exceptions to the ṇvul of the first vārttika.
Watch out. bahulam in this sūtra does real work. It lets ṇvul fail (śirortiḥ), and it licenses odd forms in the vārttikas (ājiḥ, matvarthīyaḥ).
Rule. When a name is meant, ṇvul comes after a root, in the feminine. Bhaṭṭoji: "here ṇvul comes after a root".
Example.
uddālakapuṣpabhañjikā, a name: "uddālaka-flower-breaking". The affix is ṇvul, which becomes aka, in the feminine.
Applies when / blocked by. The Tattvabodhinī says uddālaka is the śleṣmātaka tree. It takes this ṇvul in the sense of the place: the activity in which the flowers are broken. Others take it in bhāva: "the breaking of uddālaka flowers". Scholars discuss this.
Rule. When a question (paripraśna) or a reply (ākhyāna) is understood, iñ comes, and by ca also ṇvul. Bhaṭṭoji: since "optionally" is said, the other affixes also come, as they otherwise would. Both commentators: in the sūtra ākhyāna stands first, as the shorter word. In sense the question comes first, so Bhaṭṭoji explains in that order.
Example.
Question: kāṃ tvaṃ kāriṃ kārikāṃ kriyāṃ kṛtyāṃ kṛtiṃ vā akārṣīḥ? "Which deed did you do?"
Reply: sarvāṃ kāriṃ kārikāṃ kriyāṃ kṛtyāṃ kṛtiṃ vā akārṣam "I did every deed."
The Bālamanoramā names the five forms: kāri with iñ, kārikā with ṇvul, kriyā with śa (SK 3277), kṛtyā with kyap, kṛti with ktin.
Likewise gaṇim, gaṇikām, gaṇanām "counting"; pācim, pācikām, pacām, paktim "cooking".
Applies when / blocked by. Why say "optionally" at all? The Bālamanoramā: in the feminine the vāsarūpa option does not work. Without vibhāṣā, only iñ and ṇvul would come, and no other affix. It also explains three forms. gaṇim has no vṛddhi, since the lost a of the curādi stem still counts. gaṇanām has yuc (SK 3284). pacām has aṅ, pac being ṣit (SK 3281).
Rule. When a turn (paryāya), deserving (arha), a debt (ṛṇa) or an arising (utpatti) is to be indicated, ṇvuc comes optionally, in the feminine. Bhaṭṭoji glosses paryāya as "succession by turns", that is, order. arha is "deserving", that is, fitness. The Bālamanoramā: the affix denotes bhāva alone.
Example.
Turn: bhavata āsikā, śāyikā, agragāmikā "it is your turn to sit, to lie down, to go in front".
Deserving: bhavān ikṣubhakṣikām arhati "you deserve an eating of sugarcane".
Debt: ikṣubhakṣikāṃ me dhārayati "he owes me an eating of sugarcane".
Arising: ikṣubhakṣikā udapādi "an eating of sugarcane came about".
Watch out. Bhaṭṭoji's text, in the 1915 print too, reads upapādi in the last example. That is no regular aorist of upa-pad. The Bālamanoramā reads udapādi, from ut + pad, and the sense "arising" (utpatti) needs it.
Rule. "Optionally" (vibhāṣā) now stops. When a curse (ākrośa) is meant and nañ "not" stands as upapada, ani comes in the feminine.
Example.
nañ + jīv + ani → a + jīv + ani → ajīvaniḥ: ajīvanis te śaṭha bhūyāt "may you have no livelihood, you rogue!"
aprayāṇiḥ "may you have no setting out!"
The Bālamanoramā: ajīvani means ajīvana, and aprayāṇi means aprayāṇa. (Bhaṭṭoji's text, in the 1915 print too, reads āprayāṇiḥ; nañ gives a-, and the Bālamanoramā reads aprayāṇiḥ.)
Applies when / blocked by. The Bālamanoramā: ktin does not come beside ani, since the vāsarūpa option does not hold in the feminine.
Bhaṭṭoji then re-cites, in the Aṣṭādhyāyī's order, two sūtras taught earlier.
rājñā bhujyante "they are eaten by the king" → rājabhojanāḥ śālayaḥ "rice grains that are king-food". The Bālamanoramā: this is lyuṭ in the object sense.The Bālamanoramā closes the feminine section here: "iti stryadhikāraḥ".
Watch out. From SK 3290 on, the affixes are no longer feminine by heading, and bhāva is no longer carried automatically. Each rule states or carries its own sense.
has + lyuṭ → (yu → ana) hasanam beside hasitam; go + duh + lyuṭ → godohanīWith the feminine closed, Bhaṭṭoji turns to lyuṭ. It is yu with the markers l and ṭ, and yu becomes ana (SK 1247). SK 3290 gives it in the neuter for the bare action, beside the kta of SK 3090: hasanam and hasitam, both "laughing". SK 3291 uses it with an object word, to force a compound. SK 3293 gives it for the instrument and the place. Two sound rules from the eighth book are placed here, for forms in -ana after antar. One rule from the second book is placed here too, for aj before yu. The running examples are hasanam and godohanī "a milking-pot", the vessel in which the cow is milked.
Rule. In the neuter, for the bare action, lyuṭ comes too, and by ca also kta. "In the neuter, in bhāva" runs on from napuṃsake bhāve ktaḥ (SK 3090), just re-cited at SK 3289.
Example.
has + lyuṭ → markers go → has + yu → yu → ana (SK 1247) → hasanam "laughing".
With kta: hasitam "laughing".
Applies when / blocked by. Bhaṭṭoji: "the splitting is for the sake of the following rule". The Bālamanoramā explains. One rule, "in the neuter bhāva, kta and lyuṭ", would have been enough. The two are kept apart so that lyuṭ alone runs on into SK 3291.
Rule. lyuṭ comes, in the neuter bhāva, when an object word stands as upapada and its touch gives the agent bodily pleasure. Bhaṭṭoji: the preceding rule (SK 3290) would already give lyuṭ. So this statement is for an obligatory compound (nityasamāsa).
Example.
payas + pā + lyuṭ → payaḥ + pāna → obligatory compound → payaḥpānaṃ sukham "drinking milk is pleasant".
The Bālamanoramā: the drinking brings pleasure to one whose body is touched by the milk.
Applies when / blocked by. Why "of the agent"? guroḥ snāpanaṃ sukham "bathing the teacher is pleasant". Here the teacher is not the agent but the object, so there is no compound. The Tattvabodhinī: the teacher is the agent of the bathing, but in the causing-to-bathe the pupil is the agent.
Both commentators read the grammar of the sūtra the same way. In kartuḥ, the sixth case marks the object of the touching. yena is the agent, in the third case by SK 624. So: "that by which the agent, being touched, gets bodily pleasure". Bhaṭṭoji's gloss says exactly this.
Why "bodily"? The Tattvabodhinī: putrasya pariṣvajane sukham "there is pleasure in embracing a son". That pleasure is of the mind, so the rule does not apply.
Watch out. The form payaḥpānam would exist anyway. What this rule adds is that the compound cannot be undone into two words.
Rule. In place of aj "to drive", vī comes optionally when yu follows. Bhaṭṭoji supplies aj and vī. He does not say where they come from; it is ajer vy aghañapoḥ (SK 2292), the sūtra just before this one in Pāṇini's order.
Example.
pra + aj + lyuṭ → pravayaṇam (with vī) or prājanam (with aj).
Watch out. The yu is presumably that of lyuṭ (SK 3290–3291), which is why the rule is placed here. Bhaṭṭoji does not say so, and he gives no steps for the two forms.
Rule. lyuṭ also comes to express the instrument (karaṇa) and the place (adhikaraṇa). The Bālamanoramā: this names the sense of the affix. It is not an upapada. Bhaṭṭoji: karaṇādhikaraṇayoḥ is a heading up to before khal (SK 3305).
Example.
idhmapravraścanaḥ kuṭhāraḥ "an axe, the cutter of firewood": the thing by which firewood is cut.
go + duh + lyuṭ → guṇa (SK 2189) → doh + yu → yu → ana (SK 1247) → dohana → feminine ī, since lyuṭ is ṭ-marked (SK 470) → goḥ dohanī → godohanī sthālī "a pot, the milking-vessel of the cow": the thing in which the cow is milked.
Applies when / blocked by. Being later, this lyuṭ blocks ac and ap (the Bālamanoramā). The Tattvabodhinī: it is an exception to ghañ, and being later it also blocks the feminine affixes. Both quote a saying on such conflicts: "over ac and ap, the feminine affixes, khal and ana prevail; over the feminine affixes, khal and ana prevail".
Watch out. This heading carries into SK 3296–3304. The gha and ghañ forms there are instruments and places too.
Rule. After the word antar "within", ṇ replaces the n of han "to strike", when no place (deśa) is meant.
Example.
antar + hanana → ṇ for the n of han → antarhaṇanam "striking within".
But where a place is meant: antarhanano deśaḥ "a region where one strikes within".
Applies when / blocked by. Bhaṭṭoji: "preceded by a" (atpūrvasya) still holds. So there is no ṇ in antarghnanti "they strike within", where the n does not follow a. Bhaṭṭoji does not name the source of these words; it is hanter atpūrvasya (SK 359).
Why is the a marked with t (tapara)? So that only short a counts. In antaraghāni the n follows long ā, so no ṇ.
Rule. After antar, ṇ also replaces the n of ayana "going", when no place is meant.
Example.
antar + ayana → ṇ → antarayaṇam "going within".
Bhaṭṭoji: "when no place is meant" still holds. antarayano deśaḥ "a region through which one goes within".
danta + chādi + gha → (short a, ṇi lost) dantacchadaḥ; ā + kṛ + gha → ākaraḥ; ram + ghañ → rāmaḥStill under the heading "instrument and place" (SK 3293), Pāṇini gives masculine names. The affix is gha, a with the marker gh. Then come exceptions with ghañ, the a with gh and ñ that brings vṛddhi. The most general is halaś ca (SK 3300): after a consonant-final root, ghañ. Several nipātanas sit around it. One sixth-book rule, SK 3297, shortens the vowel of chādi before gha. The running examples are dantacchadaḥ "the lip", that by which the teeth are covered; ākaraḥ "a mine"; and rāmaḥ, he in whom the yogins delight.
Rule. In the masculine, when a name is meant, gha generally (prāyeṇa) comes after a root. The Bālamanoramā: "instrument and place" still holds, and this gha is an exception to the lyuṭ of SK 3293. Bhaṭṭoji gives no gloss or example under this sūtra.
Example.
Its example stands under SK 3297: ākurvanty asmin → ā + kṛ + gha → ākaraḥ "a mine". The Tattvabodhinī: the place where people come and carry on their business, a place of origin.
Watch out. "Generally" leaves room. Not every masculine name of an instrument or place has gha; SK 3299–3303 give ghañ in this field.
Rule. Before gha, chādi (the ṇic-stem of chad "to cover") takes a short vowel, if it does not have two or more upasargas. The Bālamanoramā fixes the details. "Short" runs on from khaci hrasvaḥ (SK 2955). "Penultimate" runs on from SK 2364. The vowel shortened is the penultimate of chad, the base of ṇic. It rejects the view that ṇi itself is shortened and so escapes loss.
Example.
dantāś chādyante 'nena "the teeth are covered by it":
danta + chādi + gha → short a → ṇi lost → dantacchadaḥ (with cch) "the lip".
Applies when / blocked by. Bhaṭṭoji gives ākaraḥ right after it. It is gha by SK 3296 and has nothing to do with chādi.
Watch out. Shortening is barred only with two or more upasargas. With none or one, as in dantacchadaḥ, it applies.
Rule. gocara, saṃcara, vaha, vraja, vyaja, āpaṇa and nigama are set down as ending in gha. Bhaṭṭoji: an exception to the ghañ of halaś ca (SK 3300), taught next. The Bālamanoramā: that is the point of the rule, since SK 3296 would already give gha.
Example.
gāvaś caranty asmin "cows graze in it" → gocaro deśaḥ "pasture".
saṃcaranty anena "they go about by it" → saṃcaro mārgaḥ "a path".
vahanty anena "they carry with it" → vahaḥ skandhaḥ "the shoulder".
vrajaḥ.
vyajas tālavṛntam "a palm-leaf fan". By the nipātana, vī does not replace aj.
āpaṇaḥ paṇyasthānam "a place for wares".
nigacchanty anena "they arrive by it" → nigamaś chandaḥ "the Veda".
By ca: kaṣaḥ, nikaṣaḥ.
Watch out. Without this rule, roots like car and vah would take ghañ with vṛddhi by SK 3300. The nipātana keeps the short a: gocaraḥ.
Rule. With ava beside them, tṝ "to cross" and stṝ "to spread" take ghañ. Both commentators: the frame continues, so this is in the instrument or place, in the masculine, for a name. It is an exception to gha.
Example.
ava + tṝ + ghañ → vṛddhi → avatāraḥ "the way down" (of a well and the like).
ava + stṝ + ghañ → avastāraḥ "a curtain".
Rule. After a root ending in a consonant (hal), ghañ comes. Bhaṭṭoji: an exception to gha.
Example.
ramante yogino 'smin "the yogins delight in him":
ram + ghañ → vṛddhi of the penultimate a (SK 2282) → rāmaḥ.
apamṛjyate 'nena vyādhyādiḥ "disease and the like are wiped away by it" → apāmārgaḥ, a plant.
vimārgaḥ samūhanī "a broom".
Applies when / blocked by. No commentator states the frame here. The examples are a place (rāmaḥ) and instruments (apāmārgaḥ, vimārgaḥ), and Bhaṭṭoji's "exception to gha" implies it.
Watch out. This is the ghañ that ktin (SK 3272) and the nipātanas of SK 3298 set aside. It is the general masculine affix after consonant-final roots in this field.
Rule. adhyāya, nyāya, udyāva and saṃhāra are also set down, as ending in ghañ. The Bālamanoramā: an exception to the gha of SK 3296. It names the roots: iṅ with adhi, iṇ with ni, yu with ut, hṛ with sam.
Example.
adhīyate 'smin "they study in it" → adhyāyaḥ "a lesson".
niyanti anena "they go in by it" → nyāyaḥ.
udyavanti anena "they mix together with it" → udyāvaḥ.
saṃharanti anena "they draw together with it" → saṃhāraḥ.
Applies when / blocked by. A vārttika adds three more (upasaṃkhyāna): avahāraḥ, ādhāraḥ, āvāyaḥ.
Watch out. Some versions write ādhāra and āvāya into the sūtra itself. The Tattvabodhinī calls that insertion incoherent. The sūtra has four words; the other three come from the vārttika.
Rule. After añc with ut, ghañ comes, but not in reference to water. The Tattvabodhinī: halaś ca (SK 3300) would give ghañ anyway, so the rule exists for the water ban.
Example.
ghṛtam udacyate 'smin "ghee is drawn up in it":
ghṛta + ut + añc + ghañ → k for c (SK 2863, the Bālamanoramā) → ghṛtodaṅkaḥ "a leather vessel".
Applies when / blocked by. Why "not water"? udakodañcanaḥ "a water-drawer". The Tattvabodhinī: when ghañ is barred, gha is barred too. The two would give the same form and the same accent. So lyuṭ comes.
Rule. The ghañ-form ānāya is used when a net (jāla) is meant.
Example.
ānīyante matsyādayo 'nena "fish and the like are brought in by it" → ā + nī + ghañ → ānāyaḥ "a net".
Applies when / blocked by. Why "a net"? Otherwise ānayanaḥ, "a means of bringing".
Watch out. The Tattvabodhinī: jāla itself is not set down here. It is made from jal with ṇa, by SK 2902.
Rule. After khan "to dig", gha comes, and by ca also ghañ.
Example.
ā + khan + gha → ākhanaḥ. ā + khan + ghañ → vṛddhi → ākhānaḥ.
Applies when / blocked by. Bhaṭṭoji: giving the affix the marker gh is a jñāpaka that gha comes after other roots too. Hence bhagaḥ, padam and the like. Both commentators explain. The gh serves to bring a velar for c or j (SK 2863), and khan has neither. So here the gh must be pointing beyond this rule. The Tattvabodhinī adds that padam is neuter. Under SK 3296 it would have to be masculine, so its gha comes by this indication.
A vārttika: after khan, ḍa, ḍara, ika and ikavaka. ākhaḥ, ākharaḥ, ākhanikaḥ, ākhanikavakaḥ. The Bālamanoramā: in the first two the ḍ-marker removes the final vowel and what follows it (ṭi). Bhaṭṭoji: "these are words for a digging tool". The Tattvabodhinī: all six, counting ākhanaḥ and ākhānaḥ.
dus + kṛ + khal → (guṇa; ṣ) duṣkaraḥ kaṭo bhavatā; su + pā + yuc → supānaḥThe heading "instrument and place" stops at SK 3305. Now come words meaning "hard to do" or "easy to do". The upapada is īṣat "slightly", dus "badly" or su "well". The affix is khal, a with the markers kh and l. Like a kṛtya, it comes only in the bare action or the object (SK 2833), so the doer stands in the third case: duṣkaraḥ kaṭo bhavatā "the mat is hard for you to make". After ā-final roots yuc replaces khal: supānaḥ "easy to drink". Two rules from the seventh book are placed here. They give labh an n (supralambhaḥ) and withhold it after bare su and dur (sulabham). One rule from the eighth book bars ṇ after a pada-final ṣ.
Rule. Bhaṭṭoji: karaṇādhikaraṇayoḥ (SK 3293) has ceased. When īṣat, dus or su stands as upapada in the sense of difficulty (kṛcchra) or ease (akṛcchra), khal comes. By tayor eva… (SK 2833) it comes in bhāva and in the object sense. The Bālamanoramā: without SK 2833, kartari kṛt (SK 2832) would put it in the agent.
Example.
Difficulty: dus + kṛ + khal → markers go → dus + kṛ + a → guṇa → dus + kar + a → ṣ → duṣkaraḥ: duṣkaraḥ kaṭo bhavatā "the mat is hard for you to make".
Ease: īṣatkaraḥ, sukaraḥ "easy to make".
Applies when / blocked by. Which particle goes with which sense? Both commentators: by fitness (yogyatā), dus goes with difficulty, and īṣat and su with ease. Bhaṭṭoji's grouping of the examples shows this.
In bhavatā the third case marks the agent. The Tattvabodhinī: the sixth case is barred by SK 627.
Why "in the sense of difficulty or ease"? The Tattvabodhinī: īṣatkāryam "a small task" means neither, so khal does not come. It adds that the l of khal is for the accent, and the kh for mum in a later rule (SK 3308).
A vārttika: before khal and ac, ā does not replace the vowel of mi with ni, of mī, or of lī. Both commentators: SK 2508 and SK 2509 would otherwise give ā. With khal: īṣannimayaḥ, duṣpramayaḥ, suvilayaḥ. With ac (SK 3231): nimayaḥ, mayaḥ, layaḥ.
Watch out. A khal-word never denotes the doer. duṣkaraḥ describes the mat, "hard to make"; the maker stands in the third case.
Rule. Before khal and ghañ, labh "to obtain" takes the augment num (n) only after an upasarga. The Bālamanoramā: "of labh" and "num" run on from labheś ca (SK 2582) and idito num dhātoḥ (SK 2262). The Tattvabodhinī: SK 2582 alone would give num, so this rule restricts (niyama). Hence Bhaṭṭoji's "only after an upasarga".
Example.
With khal: su + pra + labh + khal → num → su + pra + lambh + a → supralambhaḥ. Likewise īṣatpralambhaḥ, duṣpralambhaḥ.
With ghañ: upa + ā + labh + ghañ → num → upālambhaḥ.
Applies when / blocked by. Why "after an upasarga"? lābhaḥ: no upasarga, so no num, and the ghañ brings vṛddhi.
Rule. After su and dur on their own, with no other upasarga, labh does not take num before khal and ghañ.
Example.
su + labh + khal → no num → sulabham "easy to get".
dur + labh + khal → durlabham "hard to get".
Applies when / blocked by. Why "on their own"? supralambhaḥ, atidurlambhaḥ. There su and dur come with another upasarga, so the bar does not apply.
How then are atisulabham and atidurlabham correct? Bhaṭṭoji: when su and ati are *karmapravacanīya*s. The Tattvabodhinī names the rules: SK 555 (suḥ pūjāyām) and SK 556 (atir atikramaṇe ca). The Bālamanoramā explains. A karmapravacanīya is not an upasarga. So su and dur stand as if alone, and the bar holds.
Watch out. "On their own" means without another upasarga, not without any word at all.
Rule. khal comes after bhū "to become" and kṛ "to make" when two upapadas stand together. One is an agent or an object word. The other is īṣat, dus or su. Bhaṭṭoji: pairing in order (yathāsaṃkhya) is not wanted. So bhū is not tied to the agent, nor kṛ to the object. The agent or object word stands right before the root, and īṣat and the rest before that.
A vārttika: only when the agent or object has the sense of cvi. The Bālamanoramā: that sense is "becoming what one was not" (abhūtatadbhāva).
Example.
Bhaṭṭoji: since khal is kh-marked, mum comes.
anāḍhyenāḍhyena duḥkhena bhūyate "it is become rich with difficulty by one not rich":
dus + āḍhya + bhū + khal → guṇa (SK 2168) → bho + a → av for o (SK 61) → bhav + a → mum after āḍhya → dur-āḍhyam-bhava → durāḍhyambhavam.
Likewise īṣadāḍhyambhavam, svāḍhyambhavam.
With kṛ: īṣadāḍhyaṅkaraḥ, durāḍhyaṅkaraḥ, svāḍhyaṅkaraḥ.
Applies when / blocked by. Why "in the sense of cvi"? āḍhyena subhūyate "it is easily been by a rich man". He was rich already, so there is no khal-compound.
Why is the rule needed? The Bālamanoramā: āḍhya stands between īṣat and the root, so SK 3305 alone would not apply.
The Tattvabodhinī says the Kāśikā and others accepted the pairing in order. It rejects them, citing the Bhāṣya on SK 128. For kṛ it gives two analyses, one with the agent and one with the object; the Bālamanoramā gives only the object one. They do not disagree in doctrine.
Watch out. The order is fixed: particle first, then the agent or object word, then the root. And mum follows the middle word: durāḍhyambhavam, durāḍhyaṅkaraḥ.
Rule. After a root ending in ā, with īṣat, dus or su in the sense of difficulty or ease, yuc comes. Bhaṭṭoji: an exception to khal. The Bālamanoramā: kartṛkarmaṇoḥ (SK 3308) does not run on, as it is not marked svarita.
Example.
su + pā + yuc → su + pā + yu → yu → ana (SK 1247) → supānaḥ "easy to drink".
īṣatpānaḥ somo bhavatā "the soma is easy for you to drink".
duṣpānaḥ: the Bālamanoramā takes the ṣ from SK 155.
Applies when / blocked by. The Bālamanoramā: there is no optional khal beside yuc. The vāsarūpa option does not work for affixes in the sense of kta, lyuṭ, tumun and khal.
A vārttika: in the common language (bhāṣā), yuc comes after śās, yudh, dṛś, dhṛṣ and mṛṣ. duḥśāsanaḥ, duryodhanaḥ and so on. The Tattvabodhinī fills in "and so on": durdarśanaḥ, durdharṣaṇaḥ, durmarṣaṇaḥ. The Bālamanoramā places this vārttika after two Vedic rules (SK 3421–3422). It ends the affixes in the sense of khal here.
Watch out. yuc here is not feminine. The feminine section ended at SK 3289, so īṣatpānaḥ is masculine, agreeing with somaḥ.
Rule. After a ṣ that ends a pada, ṇ does not replace n. A pada is a word, or a word-part, that counts as a finished word.
Example.
niṣpānam, sarpiṣpānam: the ṣ ends the pada niṣ or sarpiṣ, so the n of pāna stays.
Applies when / blocked by. Why "after ṣ"? nirṇayaḥ: after r, the ṇ comes. Why "ending a pada"? puṣṇāti: the ṣ is inside the word, so the ṇ comes.
Bhaṭṭoji: padānta is a seventh-case compound, "final in a pada" (pade antaḥ). Hence the bar does not apply in susarpiṣkeṇa. How exactly this analysis admits the ṇ there, he does not spell out.
Watch out. The bar needs both conditions: the sound is ṣ, and it ends a pada.
avaśyam + kṛ + ṇini → (vṛddhi; the m of avaśyam → ṅ) avaśyaṅkārī; bhū + ktic → bhūtiḥ; alam + dā + ktvā → (dad) alaṃ dattvāThis group joins two short runs of the Aṣṭādhyāyī. The first three rules (SK 3311–3313) close the third quarter of book 3. They give ṇini and the kṛtyas for necessity and debt, and ktic and kta for names that carry a blessing. Two rules of book 6 (SK 3314–3315) then say what happens to the root before ktic. Three rules of the fourth quarter (SK 3316, 3317, 3319), with one book-6 rule (SK 3318) between them, bring in ktvā, the kṛt affix that leaves -tvā on the root. Here it is used in three special senses, where one action does not come before another. The ordinary ktvā, for the earlier of two actions, is the next group. The running examples are kṛ with avaśyam "necessarily", which gives avaśyaṅkārī "one who necessarily does"; bhū with ktic, which gives the name bhūtiḥ; and dā with alam "enough", which gives alaṃ dattvā "enough of giving".
Rule. When necessity (āvaśyaka) or debt (ādhamarṇya) is meant, the affix ṇini comes after a dhātu. The Bālamanoramā says kartari kṛt (SK 2832) holds, so ṇini denotes the agent. In ṇini the ṇ and the final i are markers, leaving in.
Example.
avaśyam + kṛ + ṇini → markers go → avaśyam + kṛ + in
→ the ṇ-marked affix brings vṛddhi (SK 254), with r (SK 70) → avaśyam + kār + in
→ the m of avaśyam becomes ṅ before the k → avaśyaṅkārin
→ (first case) avaśyaṅkārī "one who necessarily does".
śatam + dā + ṇini → śataṃdāyī "one who owes a hundred".
Applies when / blocked by. The Bālamanoramā notes that avaśyam is an indeclinable (avyaya). In śataṃdāyī the word śatam is the object of giving. Beside a kṛt the object would normally take the sixth case (SK 623). Here the sixth case is barred by akenor bhaviṣyadādhamarṇyayoḥ (SK 628), so śatam keeps the second case.
Watch out. The ṅ of avaśyaṅkārī is the final m of avaśyam, standing before the k of kārī. The 1915 print reads avaśyaṅkārī, with the long final ī of the first case.
Rule. The kṛtya affixes also come when necessity or debt is meant. The kṛtyas are the affixes of chapter 66, such as ṇyat and yat. Bhaṭṭoji: "'when necessity or debt is meant' is carried forward". The words of SK 3311 thus continue by anuvṛtti.
Example.
sev + kṛtya → sevyaḥ: avaśyaṃ hariḥ sevyaḥ "Hari must necessarily be served".
dā + kṛtya → deyam: śataṃ deyam "a hundred is to be given", as a debt.
Applies when / blocked by. A kṛtya comes only in the bare action or the object (SK 2833, chapter 66). So in sevyaḥ the affix denotes Hari, the one served; in deyam it denotes the hundred, the thing owed. Contrast SK 3311, where ṇini names the doer or the debtor. The Bālamanoramā calls this rule self-evident and adds nothing.
Rule. After a dhātu, ktic and kta come when a blessing (āśis) is meant and the word is a name (saṃjñā). The Bālamanoramā draws "when a blessing is meant" from āśiṣi liṅloṭau (SK 2195). In ktic the k and c are markers, leaving ti. By titutra… (SK 3163) iṭ does not come.
Example.
bhavatāt "may he be"
→ bhū + ktic → markers go → bhū + ti
→ k-marked, so no guṇa (SK 2217); no iṭ (SK 3163)
→ bhūtiḥ, a name.
The kta example, devadattaḥ, is given under SK 3315.
Applies when / blocked by. The Bālamanoramā explains bhavatāt as the benedictive loṭ. It is the paraphrase of the name bhūtiḥ. It also says ktic here is in the agent sense. The Tattvabodhinī explains the marker c. It lets na ktici (SK 3314) name this affix alone, without taking in ktin (SK 3272). Scholars discuss whether the marker is needed for that.
Watch out. Both commentators note a point about the iṭ in bhūtiḥ. śryukaḥ kiti (SK 2381) would also bar it, and, being later, would be the proper bar here. Bhaṭṭoji cites titutra… because it is needed anyway for vantiḥ and tantiḥ (SK 3314). The result is the same.
Rule. Before ktic, two root changes do not happen. The penultimate vowel is not lengthened. The nasal is not lost. This holds for the roots taught without iṭ (aniṭ) and for van, tan and the rest. Bhaṭṭoji: "for aniṭ roots and for vanati, tanoti and the rest, lengthening and loss of the nasal do not take place before ktic".
Example.
yam + ktic → yam + ti
→ no loss of m, no lengthening → yantiḥ.
Likewise ram → rantiḥ, van → vantiḥ, tan → tantiḥ.
Applies when / blocked by. The Tattvabodhinī names the two rules that are stopped. Lengthening would come by anunāsikasya kvijhaloḥ kṅiti (SK 2666). Loss of the nasal would come by anudāttopadeśa… (SK 2428). The Bālamanoramā says SK 2428 is carried forward here. The Tattvabodhinī gives the roots' meanings: yam "to cease", ram "to sport", vanu "to beg", tanu "to extend".
Watch out. The Bālamanoramā adds that in rantiḥ the iṭ is barred only by titutra… (SK 3163). śryukaḥ kiti (SK 2381) does not reach ram. Compare bhūtiḥ under SK 3313.
Rule. Before ktic, the nasal of san "to give" (sanoti) is optionally replaced by ā. Its final n is also optionally lost. The Bālamanoramā draws the ā from viḍvanor anunāsikasyāt (SK 2982). It supplies "of the final n" for the loss.
Example.
sanutāt "may he give" → san + ktic → san + ti
→ ā for the nasal → sātiḥ
→ or the n lost → satiḥ
→ or neither → santiḥ.
Three names from one root.
Applies when / blocked by. Bhaṭṭoji then gives the kta of SK 3313. devā enaṃ deyāsuḥ "may the gods give him" → devadattaḥ, a name. The Bālamanoramā calls deyāsuḥ the benedictive liṅ. It says devadattaḥ has kta in the object sense, by tayor eva… (SK 2833).
Watch out. devadattaḥ does not illustrate this rule. It belongs to SK 3313. The Bālamanoramā explains why SK 3313 names kta at all. Otherwise ktic, the exception, would block kta. The option of SK 2830 (vāsarūpa) holds only for rules before striyāṃ ktin (SK 3272).
Rule. When alam or khalu, meaning prohibition, stands as upapada, ktvā comes after a dhātu. In ktvā the k is a marker, leaving tvā. The Bālamanoramā says this ktvā is in bhāva, the action itself, by the saying avyayakṛto bhāve: "indeclinable kṛts are in bhāva". Bhaṭṭoji: "the mention of the Eastern grammarians (prācām) is for honour".
Example.
alam + dā + ktvā → dā + tvā
→ do dad ghoḥ (SK 3077): dad for dā → dad + tvā
→ alaṃ dattvā "enough of giving".
pā + ktvā → ī for the ā by ghumāsthā… (SK 2462) → pītvā khalu "enough of drinking".
Applies when / blocked by. There is no upapada compound. Bhaṭṭoji: this follows from the restriction (niyama) of amaivāvyayena (SK 783). The Bālamanoramā says the point is to avoid lyap, which comes in a compound (SK 3332). Why "alam and khalu"? Otherwise prohibition is said another way: mā kārṣīt "let him not do it". Why "meaning prohibition"? With another sense of alam: alaṅkāraḥ "ornament".
Watch out. Why is prācām only for honour? The Bālamanoramā: the next rule says udīcām "of the Northerners", so this rule must be fixed, not optional. Both commentators accept alaṃ rodanena "enough of weeping", with no ktvā. The Bālamanoramā gets it by bahulam; the Tattvabodhinī by the vāsarūpa option. Scholars discuss this.
Rule. In the view of the Northern grammarians, ktvā comes after māṅ in the sense of exchange (vyatīhāra). The Tattvabodhinī says māṅ is the root meṅ, cited with its ā already made. Bhaṭṭoji: "this rule is for ktvā where there is no prior time (apūrvakāla)".
Example.
The form comes under the next rule:
apa + meṅ + ktvā → (SK 3332, 3318) → apamitya yācate "he asks, in order to exchange".
Applies when / blocked by. The Bālamanoramā explains Bhaṭṭoji's remark. One might think ktvā already comes by SK 3320, for the earlier of two actions. But the exchange is not earlier than the asking. So a separate rule is needed. The Tattvabodhinī adds that SK 3316, 3317 and 3319 all give ktvā in bhāva. Each relates the two actions in a different way, so none is redundant.
Watch out. udīcām makes the rule optional. SK 3318 shows the other construction.
Rule. Before lyap, i optionally replaces the final sound of meṅ (mayati). lyap is the substitute for ktvā in a compound (SK 3332, group II.I). The Bālamanoramā draws "before lyap" from na lyapi (SK 3335). It also draws "optionally" from vā 'nyasya saṃyogādeḥ (SK 2378).
Example.
apa + meṅ + ktvā → compound, so lyap (SK 3332) → apa + me + ya
→ this rule: i for the final → apa + mi + ya
→ tuk after a short vowel (SK 2858) → apamitya yācate.
The other form, without i, is apamāya yācate.
Applies when / blocked by. Bhaṭṭoji: "because SK 3317 says udīcām, the ordinary ktvā also comes". Then the asking is the earlier action, and ktvā goes on yāc by SK 3320: yācitvā apamayate "having asked, he exchanges". The Bālamanoramā explains the scene. A man asks for a cow, offering a buffalo in return. The exchange is not before the asking.
Watch out. In apamāya the root does not take ī. The Bālamanoramā says īd yati (SK 2843) is barred here by na lyapi (SK 3335). The sūtra already says "optionally"; the Bālamanoramā does not explain what the second "optionally" adds.
Rule. ktvā also comes after a dhātu when connection with the farther (para) or the nearer (avara) thing is meant. Bhaṭṭoji: "when the connection of the prior with the later, or of the later with the prior, is to be understood". The Bālamanoramā says this ktvā too is in bhāva.
Example.
aprāpya nadīṃ parvataḥ
"the mountain lies without reaching the river": the mountain's connection with the farther river.
atikramya parvataṃ sthitā nadī
"the river lies beyond the mountain": the river's connection with the nearer mountain.
Applies when / blocked by. Both commentators read the sentences as spoken by someone south of the Vindhya, with the Gaṅgā to its north. For him the Vindhya is reached without crossing the Gaṅgā. The Gaṅgā is reached only by crossing the Vindhya. The Bālamanoramā adds that not-reaching and crossing do not come before the standing. So SK 3320 cannot give the ktvā.
Watch out. These are positions in space, not steps in time. That is why the rule is separate from SK 3320.
bhuj + ktvā → bhuktvā vrajati; śī + iṭ + ktvā → (not kit: guṇa, ay) śayitvā against kṛtvāThis group teaches the ordinary ktvā. It comes after the root of the earlier of two actions done by the same agent: bhuktvā vrajati "having eaten, he goes". The rest of the group shapes the root before ktvā. ktvā is marked with k, so it is kit, and a kit affix stops guṇa (SK 2217). But a ktvā that has iṭ is not kit (SK 3322). Four rules (SK 3323–3326) make it kit again, always or optionally, for named roots. Other rules decide whether a nasal is lost, and whether iṭ comes at all. The running example is śīṅ "to lie down". Its ktvā has iṭ, so it is not kit, and the root takes guṇa: śayitvā. Compare kṛ, with no iṭ: kṛtvā.
Rule. Of two actions that have the same agent, ktvā comes after the root whose action is earlier (pūrvakāla). Bhaṭṭoji: "ktvā is in bhāva", by the saying avyayakṛto bhāve. The word ktvā is carried from SK 3316; this is not stated by Bhaṭṭoji or his commentators; it is inferred, though both gloss "ktvā comes".
Example.
bhuj + ktvā → bhuj + tvā → bhuktvā vrajati "having eaten, he goes".
snātvā bhuktvā pītvā vrajati "having bathed, eaten and drunk, he goes".
nam + ktvā → nasal lost by SK 2428 → viṣṇuṃ natvā stauti "having bowed to Viṣṇu, he praises him".
svṛ + ktvā → no iṭ → svṛtvā; likewise sūtvā, dhūtvā.
Applies when / blocked by. Bhaṭṭoji: "the duality is not essential". The sūtra says "of two", but three or more earlier actions may stand together. The Tattvabodhinī explains why. The finite verb names the main action. Each ktvā-form qualifies that main action, not the others. So snātvā pītvā bhuktvā vrajati can stand in any order.
For svṛ and the rest, svaratisūti… (SK 2279) would make iṭ optional. Bhaṭṭoji: iṭ is absent without option, by śryukaḥ kiti (SK 2381), through pūrvavipratiṣedha. That means the earlier rule wins here. The Bālamanoramā gives the reason. The section that bars iṭ begins before the section that gives iṭ. That order shows the bar is meant to prevail.
Watch out. The bhāva of ktvā is not the bhāva of ghañ. The Tattvabodhinī says the action of ghañ is accomplished (SK 3184). The action of ktvā is in process.
Rule. Before ktvā, the roots skand and syand do not lose their n. The Bālamanoramā draws the loss of n from śnān nalopaḥ (SK 2544) and the "not" from nāñceḥ pūjāyām (SK 424). ktvi is the seventh case of ktvā.
Example.
skand + ktvā → no iṭ; the n stays → skantvā.
syand + ktvā → syand is marked with ū (ūdit), so iṭ is optional
→ without iṭ → syantvā
→ with iṭ → syanditvā.
Applies when / blocked by. skantvā has no iṭ. The Bālamanoramā says the root is taught with an anudātta vowel. The Tattvabodhinī cites ekāca… (SK 2246). It also says why skand needs naming. Its marker is the whole ir (skandir), not i alone. So it does not count as i-marked, and the n would otherwise be lost.
Watch out. Bhaṭṭoji glosses syand as ūdit for the option of iṭ. Only syand has that option here.
Rule. A ktvā that has iṭ (seṭ) is not kit. The Bālamanoramā draws "kit" from asaṃyogāl liṭ kit (SK 2242).
Example.
śī + ktvā → iṭ → śī + i + tvā
→ this rule: not kit, so guṇa is not stopped → śe + i + tvā
→ e becomes ay before a vowel (SK 61) → śayitvā.
Without iṭ: kṛ + ktvā → the affix stays kit, no guṇa → kṛtvā.
Applies when / blocked by. Bhaṭṭoji: "why say 'with iṭ'? kṛtvā". The next four rules are exceptions to this one. Chapter 67 drew on this rule for the niṣṭhā (group III.F).
Watch out. The rule takes away only kit-hood. The iṭ itself comes by the general iṭ rules.
Rule. After mṛḍ, mṛd, gudh, kuṣ, kliś, vad and vas, a ktvā with iṭ is kit. Bhaṭṭoji: "after these, a seṭ ktvā is kit". The Bālamanoramā calls the rule an exception to SK 3322. "kit" is carried from SK 2242.
Example.
mṛḍ + ktvā → iṭ → mṛḍ + i + tvā
→ kit, so no guṇa of the light penultimate → mṛḍitvā.
kliś → iṭ optional by kliśaḥ ktvāniṣṭhayoḥ (SK 3049) → kliśitvā (with iṭ, kit) / kliṣṭvā (without iṭ).
vad → kit, so saṃprasāraṇa → uditvā.
vas → iṭ by SK 3046; saṃprasāraṇa; ṣ by SK 2410 → uṣitvā.
Applies when / blocked by. Bhaṭṭoji adds forms made kit by rudavida… (SK 2609), another exception to SK 3322: ruditvā, viditvā, muṣitvā, gṛhītvā. The Tattvabodhinī explains gṛhītvā: saṃprasāraṇa by SK 2412, and the iṭ lengthened by SK 2562. For vad and vas the Bālamanoramā names vacisvapi… (SK 2409) for the saṃprasāraṇa.
Watch out. Bhaṭṭoji illustrates only mṛḍ of the first four roots. The Bālamanoramā says mṛditvā and the rest should be given too. The Tattvabodhinī gives the meanings: mṛḍ "to gladden", mṛd "to crush", gudh "to wrap", kuṣ "to draw out", kliś "to torment", vad "to speak", vas "to dwell".
Rule. After a root that ends in th or ph and has n as its penultimate (upadhā), a ktvā with iṭ is optionally kit. Bhaṭṭoji: "a seṭ ktvā is optionally kit". The Bālamanoramā says this is an option where SK 3322 would deny kit-hood outright.
Example.
śranth + ktvā → iṭ → śranth + i + tvā
→ kit: the n is lost → śrathitvā
→ not kit: the n stays → śranthitvā.
gumph → guphitvā / gumphitvā.
Applies when / blocked by. Bhaṭṭoji: "why 'having n as penultimate'? kothitvā, rephitvā". These roots, kuth "to stink" and riph, have no penultimate n. So their seṭ ktvā stays not kit by SK 3322. The light penultimate takes guṇa. The Tattvabodhinī adds that the option of ralo vyupadhāt… (SK 2617) does not reach them either, because of the word nopadha.
Watch out. Kit-hood here decides the n. When the affix is kit the n is lost; when it is not, the n stays.
Rule. After vañc, luñc and ṛt, a ktvā with iṭ is also optionally kit. Bhaṭṭoji: "a seṭ ktvā is optionally kit". The Bālamanoramā calls it an exception to SK 3322.
Example.
vañc + ktvā → iṭ → kit: n lost → vacitvā; not kit → vañcitvā.
luñc → lucitvā / luñcitvā.
ṛt → kit: no guṇa → ṛtitvā; not kit: guṇa → artitvā.
Applies when / blocked by. The Tattvabodhinī gives the roots. vañcu is "to deceive", luñcu "to remove". ṛti is a root found only in the sūtras, meaning "loathing". ṛter īyaṅ (SK 2422) applies only optionally before ārdhadhātuka affixes. Where it does not apply, this rule makes kit-hood optional.
Watch out. For vañc and luñc the option shows in the nasal. For ṛt it shows in the vowel.
Rule. After tṛṣ, mṛṣ and kṛś, in the view of Kāśyapa, a ktvā with iṭ is optionally kit. Bhaṭṭoji: "after these, a seṭ ktvā is optionally kit". The Tattvabodhinī draws "optionally" from SK 3324. So the name Kāśyapa is for honour.
Example.
tṛṣ + ktvā → iṭ → kit: no guṇa → tṛṣitvā; not kit → tarṣitvā.
mṛṣitvā / marṣitvā; kṛśitvā / karśitvā.
Applies when / blocked by. Bhaṭṭoji gathers other roots whose ktvā takes or lacks kit-hood or iṭ:
dyutitvā / dyotitvā, likhitvā / lekhitvā. Not sevitvā (sev does not end in ral), vartitvā (no u or i as penultimate), eṣitvā (no initial consonant), bhuktvā (no iṭ).uṣitvā; kṣudhitvā / kṣodhitvā.añcitvā. In the sense of going, aktvā as well.lubhitvā / lobhitvā. In the other sense, lubdhvā.Watch out. The printed text says "but lubdhvā in the sense of bewildering". That inverts SK 3048. The Bālamanoramā reads "not in the sense of bewildering" and calls the other a wrong reading. In the sense of greed, iṭ is optional by tīṣasaha… (SK 2340), so lubhitvā occurs there too, meaning "having longed for". The Bālamanoramā also explains eṣitvā: with no option, SK 3322 keeps it not kit, so it takes guṇa.
Rule. After jṝ and vraśc, ktvā takes iṭ. Bhaṭṭoji: "to ktvā following these two, iṭ is added".
Example.
jṝ + ktvā → iṭ → the iṭ optionally long by vṝto vā (SK 2391) → jarītvā / jaritvā.
vraśc + ktvā → iṭ → not kit (SK 3322), so no saṃprasāraṇa → vraścitvā.
Applies when / blocked by. Both commentators say why the rule is needed. For jṝ, śryukaḥ kiti (SK 2381) would bar iṭ. For vraśc, which is ūdit, iṭ would be optional. This rule makes it fixed for both.
Watch out. In vraścitvā there is no saṃprasāraṇa by grahijyā… (SK 2412). Both commentators give the reason: with iṭ the affix is not kit (SK 3322).
Rule. After a root marked with u (udit), ktvā takes iṭ optionally. The Bālamanoramā draws "to ktvā" from SK 3327. It calls the rule an option for what was not otherwise obtained (aprāptavibhāṣā).
Example.
śam + ktvā → with iṭ → śamitvā
→ without iṭ → the penultimate lengthened by SK 2666 → śāntvā.
div + ktvā → without iṭ → dyūtvā; with iṭ → devitvā.
Applies when / blocked by. The Bālamanoramā explains div. It is udit (divu), so iṭ is optional. Kit-hood is optional by SK 2617. Without iṭ, the affix is kit: v becomes ūṭh (ū) by cchvoḥ śūḍ… (SK 2561), and i becomes y: dyūtvā. With iṭ and not kit, the penultimate takes guṇa: devitvā.
Watch out. With iṭ there is no ūṭh even if the affix is kit. The Bālamanoramā: the affix then does not begin with a jhal consonant.
Rule. Before a ktvā beginning with a jhal consonant, the penultimate of kram is optionally lengthened. Bhaṭṭoji: "before a jhal-initial ktvā, the penultimate of kram is optionally lengthened". The Bālamanoramā gives four words carried forward: "penultimate" (SK 370), "optionally" (SK 2622), "long" (SK 174) and "before a jhal" (SK 2666).
Example.
kram + ktvā → no iṭ → kram + tvā
→ lengthened → krāntvā; not lengthened → krantvā.
With iṭ → kramitvā.
Applies when / blocked by. Bhaṭṭoji: "why 'before a jhal'? kramitvā". With iṭ the affix begins with i. The Bālamanoramā: kram is udit, so iṭ is optional (SK 3328); SK 2666 would make the lengthening fixed, and this rule makes it optional. Bhaṭṭoji adds pūṅaś ca (SK 3050): iṭ optional after pū: pavitvā / pūtvā.
Watch out. Lengthening and iṭ exclude each other here. With iṭ there is no lengthening at all.
Rule. Before ktvā, roots ending in j, and naś, optionally lose their n. The Bālamanoramā draws the loss of n from SK 2544 and "before ktvā" from SK 3321.
Example.
bhañj + ktvā → bhaktvā / bhaṅktvā.
rañj + ktvā → raktvā / raṅktvā.
naś + ktvā → num by masjinaśor jhali (SK 2517) → this rule: optional loss → naṣṭvā / naṃṣṭvā.
With iṭ by radhādibhyaś ca (SK 2515) → naśitvā.
Applies when / blocked by. A vārttika: "it should be stated: before a jhal-initial ktvā". Bhaṭṭoji: "so not here: añjitvā". añj is ūdit and takes iṭ optionally. With iṭ the affix begins with i, so the n stays. Without iṭ: aktvā / aṅktvā.
Bhaṭṭoji adds more root changes before ktvā:
khātvā; with iṭ, khanitvā.ditvā, sitvā, mitvā, sthitvā.hitvā.Watch out. In SK 3074 the i is a substitute, not the augment iṭ. The Bālamanoramā identifies the roots: do "to cut", ṣo "to end", mā and sthā.
Rule. Before ktvā, hā "to abandon" (jahāti) is replaced by hi. The hi comes from dadhāter hiḥ (SK 3076); this is not stated by Bhaṭṭoji or his commentators; it is inferred, though the Bālamanoramā calls this rule the injunction of hi.
Example.
hā + ktvā → hi for hā → hitvā "having abandoned".
But hāṅ "to go" → hātvā.
ad + ktvā → jagdh for ad by ado jagdhiḥ… (SK 3080) → jagdhvā.
Applies when / blocked by. The Bālamanoramā explains hāṅas tu. The rule names jahāti, the form of ohāk "to abandon". ohāṅ "to go" would have been named jihīte. So hāṅ keeps its ā. For jagdhvā it gives dh for the t of tvā (SK 2280), and optional loss of the root's dh (SK 71).
Watch out. The same hitvā also comes from dhā by SK 3076 (under SK 3330). The two forms look alike but come from different roots.
pra + kṛ + ktvā → (lyap) pra-kṛ-ya → (tuk, SK 2858) prakṛtyaWhen a ktvā-form is compounded with an indeclinable such as an upasarga, ktvā is replaced by lyap. In lyap the l and p are markers, leaving ya. Because lyap is p-marked, a root ending in a short vowel takes the augment t (tuk) before it, by hrasvasya piti kṛti tuk (SK 2858). The other rules of this group are rules of the aṅga, the base before the affix. Some come from book 6 and are placed here because lyap triggers them. They decide which root changes still happen before ya, and which are stopped. The running example is kṛ with pra: prakṛtya.
Rule. In a compound whose prior member is an indeclinable other than nañ (the negative a-), lyap replaces ktvā. Bhaṭṭoji: "in a compound other than one with nañ, whose prior member is an indeclinable, lyap is the substitute for ktvā". The Tattvabodhinī explains "not nañ": an indeclinable other than nañ, yet like it.
Example.
pra + kṛ + ktvā → compound → lyap for ktvā → pra + kṛ + ya
→ tuk after the short ṛ (SK 2858) → pra + kṛ + t + ya
→ prakṛtya.
Applies when / blocked by. Bhaṭṭoji: "why 'not with nañ'? akṛtvā". The negative compound keeps ktvā. Next, paramakṛtvā also keeps ktvā. Bhaṭṭoji: "because an exclusion (paryudāsa) is taken, not here". anañ means "a word other than nañ, of the same kind", that is, another indeclinable. parama is not an indeclinable. The Tattvabodhinī: with a plain negation, lyap would come in paramakṛtvā too.
Watch out. Both commentators say pra in prakṛtya means excellence. It is compounded with the ktvā-form without option, by kugatiprādayaḥ (SK 761).
Rule. When ṣ for s (ṣatva) or the augment tuk is to be done, a single substitute (ekādeśa) counts as not made (asiddha). Bhaṭṭoji: "the rule that teaches the single substitute is asiddha". The Bālamanoramā supplies "single substitute", since the sūtra follows the section of ekaḥ pūrvaparayoḥ (SK 68).
Example.
adhi + i + ktvā → lyap → adhi + i + ya
→ the two i merge into ī, but that single ī is asiddha for tuk
→ tuk sees a short i (SK 2858) → adhītya.
pra + i + ktvā → lyap → e from a + i is asiddha → tuk → pretya.
Applies when / blocked by. Bhaṭṭoji first gives ṣatva: ko 'sicat "who sprinkled?". Both commentators explain. o and a merge into o by SK 86. Without this rule, the s would stand after a sound of the iṇ group (the vowels other than a/ā, with the semivowels and h) and become ṣ (SK 212). With the merged o asiddha, the s still follows a, so there is no ṣ.
The Bālamanoramā says adhītya is from iṅ "to study", always with adhi; pretya is pra + iṇ. Without this rule there would be no short vowel, and no tuk.
Watch out. The Bālamanoramā says what counts as asiddha: the rule, not the operation. The Mahābhāṣya settles on that, and Bhaṭṭoji writes "the rule that teaches".
Rule. Before lyap, the loss of the nasal taught in anudāttopadeśa… (SK 2428) is optional. It covers roots taught with an anudātta vowel, and van, tan and the rest. Bhaṭṭoji: "this is a restricted option (vyavasthita-vibhāṣā)". For m-final aniṭ roots it is optional. For n-final aniṭ roots and for van and the rest it is fixed.
Example.
ā + gam + ktvā → lyap → ā + gam + ya
→ m lost → ā + ga + ya → tuk → āgatya
→ m kept → āgamya.
praṇatya / praṇamya.
Fixed loss: prahatya (han), pramatya (man), pravatya (van), vitatya (tan).
Applies when / blocked by. The Bālamanoramā lists the roots. The m-final aniṭ roots are gam, nam, ram and yam. The n-final aniṭ roots are man and han. It also notes that the Mahābhāṣya does not show this sūtra. It reads two vārttikas under SK 2428 to the same effect.
Bhaṭṭoji then cites ado jagdhiḥ… (SK 3080) and a paribhāṣā, a working maxim of the grammar: "lyap, though outer (bahiraṅga), blocks even inner (antaraṅga) operations". It is known from the mention of lyap in SK 3080. The Bālamanoramā explains: jagdh would come before lyap anyway, being inner. So naming lyap there shows the maxim. Hence, before lyap:
vidhāya: no hi (SK 3076).pradāya: no dad (SK 3077).prakhanya: no ā (SK 2504), and no iṭ.prasthāya: no i (SK 3074).prakramya: no lengthening (SK 3329).āpṛcchya: no ś (śūṭh, SK 2561).pradīvya: no ūṭh (SK 2561), and no iṭ.Bhaṭṭoji ends the list with prapaṭhya.
Watch out. Bhaṭṭoji names seven blocked operations; the last is iṭ. Both commentators give the missing iṭ to pradīvya; the Tattvabodhinī also to prakhanya. Neither assigns an operation to prapaṭhya, so this chapter assigns none. For the second form both commentators read pradāya, which shows the missing dad; the printed text has pradhāya.
Rule. Before lyap, ghu, mā, sthā and the rest do not take ī. The Bālamanoramā draws the roots from ghumāsthāgāpājahātisāṃ hali (SK 2462) and the ī from īd yati (SK 2843).
Example.
pra + dheṭ + ktvā → lyap → pra + dhā + ya → no ī → pradhāya.
Likewise pramāya, pragāya, prapāya, prahāya, prasāya.
Applies when / blocked by. Bhaṭṭoji adds three related forms:
pramāya, nimāya, upadāya.vilāya / vilīya.uttārya, vicārya.The Bālamanoramā says dheṭ names the root of pradhāya; its ā comes by SK 2370. prasāya is from ṣo "to end".
Watch out. Both commentators meet an objection from Śrīharṣa's nipīya "having drunk". It has ī before lyap. They answer that it comes from pīṅ "to drink" of the div class, so there is no fault.
Rule. Before lyap, ay replaces ṇi when the sound before ṇi is preceded by a light (laghu) vowel. The Bālamanoramā draws ṇi from ṇer aniṭi (SK 2313) and ay from ayām antālvāyya… (SK 2311). It calls the rule an exception to the loss of ṇi.
Example.
vi + gaṇi + ktvā → lyap → vi + gaṇ-i + ya
→ the a of ga is light → ay for ṇi → vi + gaṇ-ay + ya
→ vigaṇayya.
praṇamayya; prabebhidayya.
Applies when / blocked by. Bhaṭṭoji: "why 'after a light vowel'? sampradhārya". There the sound before ṇi, the r of dhāri, is preceded by a long ā, so ṇi is lost. The Bālamanoramā derives two forms. gaṇa is a-final in the cur class; ṇic, loss of a, lyap, ay: vigaṇayya. prabebhidayya is from the intensive of bhid, with ṇic, and ay for ṇi. The Tattvabodhinī says nam is mit, so its vowel is shortened before ṇi: praṇamayya.
Watch out. The Tattvabodhinī meets an objection on vigaṇayya. The lost a might count as still present for what precedes. Then ṇi would not follow a light vowel. It answers that this counting is not fixed; scholars discuss this.
Rule. For āp "to obtain" (āpnoti), ay for ṇi before lyap is optional. Bhaṭṭoji: "for āpnoti, ay is optionally substituted for ṇi before lyap".
Example.
pra + āpi + ktvā → lyap → pra + āp-i + ya
→ ay → prāpayya
→ ṇi lost → prāpya.
Applies when / blocked by. The Bālamanoramā explains why the rule is needed. The ā of āp is long, so SK 3336 does not apply. This is an option for what was not otherwise obtained (aprāptavibhāṣā). The Tattvabodhinī adds that āp here is the fifth-class root. The āpi made from iṅ with ṇic is not meant.
Watch out. The Tattvabodhinī uses the rule to defend Śrīharṣa's avāpya. It can stand for avāpayya, a causative, by this option. That ṇi, ay and lyap continue from SK 3336 is not stated by Bhaṭṭoji or his commentators; it is inferred from Bhaṭṭoji's gloss.
Rule. Before lyap, kṣi is lengthened. The Bālamanoramā draws "before lyap" from SK 3336 and "long" from yupluvor dīrghaś chandasi (SK 3544).
Example.
pra + kṣi + ktvā → lyap → pra + kṣi + ya
→ this rule: ī → prakṣīya.
Applies when / blocked by. Bhaṭṭoji gives only the form prakṣīya, with no t.
Watch out. The word "long" is drawn from SK 3544, a rule placed in the Vedic section of the SK.
Rule. Before lyap, veñ "to weave" takes no saṃprasāraṇa. The Bālamanoramā draws veñ from veñaḥ (SK 2415) and "no saṃprasāraṇa" from na saṃprasāraṇe saṃprasāraṇam (SK 363).
Example.
pra + veñ + ktvā → lyap → pra + ve + ya
→ ā by ādeca upadeśe 'śiti (SK 2370) → no saṃprasāraṇa → pravāya.
Applies when / blocked by. lyap replaces ktvā, which is kit. So saṃprasāraṇa would come. The Tattvabodhinī names vacisvapi… (SK 2409) as its source; the Bālamanoramā names grahijyā… (SK 2412). Scholars discuss this.
Watch out. The next two rules extend the same bar to jyā and vyeñ.
Rule. Before lyap, jyā also takes no saṃprasāraṇa. The Tattvabodhinī: "saṃprasāraṇa does not occur".
Example.
pra + jyā + ktvā → lyap → no saṃprasāraṇa → prajyāya.
Applies when / blocked by. Bhaṭṭoji gives only the example. The Bālamanoramā says the saṃprasāraṇa of grahijyā… (SK 2412) is stopped. Without this rule the y of jyā would become i.
Watch out. SK 3339 covers only veñ, so jyā needs its own rule; the saṃprasāraṇa it stops is that of grahijyā… (SK 2412).
Rule. Before lyap, vyā, that is vyeñ "to cover", also takes no saṃprasāraṇa. The Tattvabodhinī says it is explained like SK 3340.
Example.
upa + vyeñ + ktvā → lyap → upa + vye + ya
→ ā by SK 2370 → no saṃprasāraṇa → upavyāya.
Applies when / blocked by. Bhaṭṭoji gives only the example. The Bālamanoramā names the root vyeñ saṃvaraṇe and the stopped rule, grahijyāvayi… (SK 2412).
Watch out. The next rule makes an exception for vyeñ after pari.
Rule. After pari, vyeñ optionally takes saṃprasāraṇa before lyap. Bhaṭṭoji: "for vyeñ after pari, saṃprasāraṇa optionally occurs before lyap". The Bālamanoramā lists the words carried forward: the root, "before lyap" and "saṃprasāraṇa". That the "not" of SK 3339–3341 does not continue is not stated by Bhaṭṭoji or his commentators; it is inferred from Bhaṭṭoji's gloss.
Example.
pari + vyeñ + ktvā → lyap → saṃprasāraṇa of y, then the a merges → pari + vi + ya
→ tuk (SK 2858) would come after the short i
→ but halaḥ (SK 2559) lengthens the i, and wins as the later rule (paratva) → parivīya.
Without saṃprasāraṇa → parivyāya.
Applies when / blocked by. Bhaṭṭoji then meets an objection. How do mukhaṃ vyādāya svapiti "having opened his mouth, he sleeps" and netre nimīlya hasati "having closed his eyes, he laughs" stand? The opening and closing are not earlier than the sleeping and laughing. His answer: the sleeping and laughing go on after the opening and closing. Taking that later portion in view, the ktvā (here lyap) of the earlier time comes.
Watch out. Both commentators extend the answer. In a sentence like mām upetya tu kaunteya punarjanma na vidyate, a word such as sthitasya must be supplied. That gives the two actions one agent, as SK 3320 requires.
smṛ + ṇamul → (vṛddhi) smār-am → (dvitva, SK 2140) smāraṃ smāraṃ namati śivamThis group brings in a second affix for the earlier of two actions. SK 3320 (samānakartṛkayoḥ pūrvakāle) gave ktvā after the root of the earlier action when both actions have one agent: bhuktvā vrajati "having eaten, he goes". SK 3343 adds ṇamul in that same domain, for an action done again and again. ṇamul leaves am, and its ṇ brings vṛddhi, so smṛ gives smāram. Like ktvā, it is an indeclinable kṛt affix and denotes the bare action (bhāva), as the Bālamanoramā says at SK 3345. SK 3344 and SK 3345 then limit both affixes beside yad, agre, prathama and pūrva. From SK 3346 on, the root is kṛñ "to do" (kṛ named with its marker ñ). It takes khamuñ in abuse and ṇamul beside set words such as svādu, anyathā and yathā. The group ends with ṇamul after dṛś, vid, jīv, the causative pūri and knopi. The running example is smāraṃ smāraṃ namati śivam "remembering him again and again, he bows to Śiva".
Bhaṭṭoji gives the finished forms. Most steps shown in this part are not stated by Bhaṭṭoji or his commentators; they are inferred. These are the vṛddhi and guṇa of the root (SK 254, SK 2189, SK 2282), the loss of a causative ṇi, and plain sandhi. Where a commentator does state a step, his name is given.
Rule. When repetition (ābhīkṣṇya, "again and again") is to be expressed, ṇamul comes, and ktvā too. It works in the domain of SK 3320: after the root of the earlier of two actions that share one agent. The Tattvabodhinī explains Bhaṭṭoji's "in the earlier domain" (pūrvaviṣaye) in just this way. The ṇamul-form or ktvā-form is then doubled (dvitva), by nityavīpsayoḥ (SK 2140).
Example.
smṛ + ṇamul → the markers ṇ, u and l go → smṛ + am
→ the ṇ-marked affix brings vṛddhi of the final vowel (SK 254), with r (SK 70) → smār + am → smāram
→ doubled (SK 2140) → smāraṃ smāraṃ namati śivam.
With ktvā: smṛtvā smṛtvā (namati śivam is construed again, says the Bālamanoramā).
Further forms:
pā + ṇamul → y after root-final ā (yuk, SK 2761) → pāyaṃ pāyam;bhuj + ṇamul → guṇa of the light penultimate (SK 2189) → bhojaṃ bhojam;śru + ṇamul → vṛddhi, au → āv (SK 61) → śrāvaṃ śrāvam.Applies when / blocked by. Three earlier rules shape the root before ṇamul.
gam + ṇic → vṛddhi (SK 2282) → gāmi → shortened by mitāṃ hrasvaḥ (SK 2568) → gami. Before ṇi followed by ṇamul, SK 2762 lengthens the vowel optionally. The ṇi is then lost → gamaṃ gamam / gāmaṃ gāmam.labh: SK 2765 (vibhāṣā ciṇṇamuloḥ) gives the augment num (n) optionally → lambhaṃ lambham / lābhaṃ lābham (vṛddhi of the penultimate a, SK 2282). The option is restricted (vyavasthita-vibhāṣā): with an upasarga, num is invariable → pralambhaṃ pralambham.jāgṛ: SK 2480 gives guṇa, not the expected vṛddhi → jāgaraṃ jāgaram. The causative of jāgṛ gives the same form. Its ṇi is lost, says the Bālamanoramā.Watch out. Why does the sūtra need ca for ktvā? The general option of SK 2830 (vāsarūpa) might seem to give it already. Scholars discuss this, and the two commentaries give different reasons. The Tattvabodhinī's point is useful: if vāsarūpa worked here, the finite verb (laṭ and the rest) would also come. It does not; under this sūtra only ṇamul and ktvā are used. The Tattvabodhinī also calls the earlier grammarians' optional vṛddhi for causative jāgṛ non-Pāṇinian.
Rule. When yad "when" is the upapada, the affixes for the earlier time, ktvā and ṇamul, do not come. The condition: the sentence holding the earlier and the later action needs no further sentence (anākāṅkṣa, "not expecting").
Example.
yad ayaṃ bhuṅkte tataḥ paṭhati "when he eats, then he recites".
This is one sentence with two verbs, bhuj and paṭh, and it needs no other. So no ktvā or ṇamul: both verbs are finite.
Counter-example: yad ayaṃ bhuktvā vrajati tato 'dhīte "when he goes, having eaten, then he studies".
Here the clause with bhuj and vraj expects the further clause tato 'dhīte. So the bar does not apply, and bhuktvā stands.
Applies when / blocked by. The Tattvabodhinī meets an objection. A maxim might suggest that only the nearest affix, ṇamul, is barred. But the sūtra means "whatever is obtained for the earlier time", so ktvā is barred as well. Only the specially taught ktvā and ṇamul are barred; the general finite forms (laṭ and the rest) still come, as in bhuṅkte.
Watch out. anākāṅkṣe is a negative compound with ākāṅkṣa, an ac-noun of the paca group (SK 2896). Bhaṭṭoji shows this by his gloss "if it does not expect" (nākāṅkṣate cet), says the Tattvabodhinī.
Rule. When agre "first", prathama or pūrva is the upapada, ktvā and ṇamul come optionally, for the earlier of two actions with one agent. Bhaṭṭoji says "in repetition" (ābhīkṣṇye) is not carried forward from SK 3343.
Example.
agre bhojaṃ vrajati "having eaten first, he goes"; agre bhuktvā.
prathamaṃ bhojam, prathamaṃ bhuktvā; pūrvaṃ bhojam, pūrvaṃ bhuktvā.
The other option, the finite verb (laṭ and the rest): agre bhuṅkte tato vrajati "he eats first, then he goes".
Applies when / blocked by. In repetition the affixes are obligatory, with no finite alternative. SK 3343 prevails here by pūrvavipratiṣedha, the rule that in this conflict the earlier rule wins: agre bhojaṃ bhojaṃ vrajati "having eaten first again and again, he goes"; bhuktvā bhuktvā.
The Bālamanoramā explains the finite option. ktvā and ṇamul denote the bare action; when the agent is to be expressed, laṭ and the rest come.
Watch out. agre bhojam is two words, not a compound. The Tattvabodhinī says SK 783 (amaivāvyayena) restricts the compound here. And if agre itself means "first", what do the affixes add? The Tattvabodhinī answers: agre puts his eating before that of the other eaters; ktvā or ṇamul puts it before the going. Scholars discuss what kind of word agre is.
Rule. When an object (karman) is the upapada and abuse (ākrośa) is understood, kṛñ takes the affix khamuñ. Here kṛ means "to utter". The Bālamanoramā notes that karmaṇi names the upapada, while ākrośe names the sense conveyed.
Example.
caura + kṛ + khamuñ → the markers go → caura + kṛ + am
→ the ñ-marker brings vṛddhi (SK 254) → caura + kār + am
→ the kh-marker brings the augment m to the upapada (mum, SK 2942, says the Bālamanoramā) → cauraṅkāram.
cauraṅkāram ākrośati "he reviles him, calling him 'thief'". Bhaṭṭoji: the meaning is "having uttered the word cora".
Watch out. Do not read kṛ here as "do". He does not make anyone a thief; he says the word.
Rule. When a word meaning svādu "sweet" is the upapada, kṛñ takes ṇamul, for the earlier of two actions with one agent. The prior member is given its m-ending (svādum) by nipātana, a form fixed by the sūtra itself.
Example.
asvāduṃ svāduṃ kṛtvā bhuṅkte "having made sweet what was not sweet, he eats" (the 1915 print reads asvāduṃ svāduṃ).
svādum + kṛ + ṇamul → vṛddhi → svādum + kāram → svāduṅkāraṃ bhuṅkte.
sampanna and lavaṇa are synonyms of svādu: sampannaṅkāram, lavaṇaṅkāram.
By vāsarūpa (SK 2830), ktvā comes too: svāduṃ kṛtvā bhuṅkte.
Applies when / blocked by. The Bālamanoramā says kṛñ is carried forward from SK 3346, but khamuñ is not, since it is not marked for carrying (svarita). So ṇamul is stated, and the condition of SK 3320 carries forward. The word svādumi takes in its meaning, so synonyms count.
Watch out. Why fix the m by nipātana instead of using khamuñ with its mum? Both commentators answer. When svādu is a cvi-form ("made sweet"), it is an indeclinable. mum is not given to indeclinables, and other unwanted changes would follow. The fixed form svādum avoids them. The Tattvabodhinī adds that the fixed m holds before any bhāva affix, so svāduṃ kṛtvā also keeps it.
Rule. When anyathā "otherwise", evam "so", katham "how" or ittham "thus" is the upapada, kṛñ takes ṇamul. The condition: kṛñ is siddhāprayoga, "one whose non-use is established". Bhaṭṭoji explains: it has no purpose, so it does not deserve to be used.
Example.
anyathākāram, evaṅkāram, kathaṅkāram.
ittham + kṛ + ṇamul → itthaṅkāraṃ bhuṅkte. Bhaṭṭoji: it means itthaṃ bhuṅkte "he eats thus".
Counter-example: śiro 'nyathā kṛtvā bhuṅkte "having turned his head the other way, he eats". Here kṛ is needed for the sense, so no ṇamul.
Applies when / blocked by. Bhaṭṭoji supplies "kṛñ takes ṇamul" in his gloss. That these come from SK 3346 and SK 3347 is not stated by Bhaṭṭoji or his commentators; it is inferred. The Bālamanoramā says the earlier-time condition of SK 3320 cannot apply here. It also notes, from the Bhāṣya's own usage anyathā kṛtvā coditām, that ktvā comes as well.
Watch out. If kṛ adds nothing, why use it? The Bālamanoramā answers: its meaning is not intended; it is used so that the ṇamul-form is correct. The Tattvabodhinī explains the counter-example. Without kṛ, "otherwise" would describe the eating, not the turning of the head.
Rule. When yathā "as" or tathā "so" is the upapada, kṛñ takes ṇamul in a reply made with resentment (asūyā). Bhaṭṭoji says kṛñaḥ and siddhāprayogaḥ continue from SK 3348: again kṛ adds no meaning.
Example.
yathākāram ahaṃ bhokṣye tathākāraṃ bhokṣye, kiṃ tavānena "however I shall eat, so I shall eat — what is that to you?"
Applies when / blocked by. The Tattvabodhinī sets the scene: someone with no standing to ask has asked how he will eat. It adds that ktvā is also available here, optionally, by vāsarūpa (SK 2830).
Rule. When an object is the upapada and its totality (sākalya) is understood, dṛś "see" and vid "know, find" take ṇamul. The Bālamanoramā says the condition of SK 3320 still carries forward.
Example.
kanyā + dṛś + ṇamul → guṇa (SK 2189) → kanyādarśaṃ varayati "he courts every girl he sees". Bhaṭṭoji: the meaning is "all the girls".
brāhmaṇa + vid + ṇamul → brāhmaṇavedaṃ bhojayati "he feeds every brāhmaṇa he knows of". Bhaṭṭoji: whatever brāhmaṇa he knows, finds or considers, he feeds them all.
Applies when / blocked by. Both commentators say Bhaṭṭoji's gloss covers vid in three senses: "know", "find" and "consider". vid in the sense "to be" is not taken. The Bālamanoramā takes varayati from vara īpsāyām "to desire" (cur class).
Watch out. The totality is of what the upapada names: all the girls who came within sight. It is not "all of one girl".
Rule. When yāvat "as much as, as long as" is the upapada, vind and jīv take ṇamul. The Tattvabodhinī says vinda names vid "to find", which has the nasal in vindati by SK 2542.
Example.
yāvat + vid + ṇamul → guṇa → yāvadvedaṃ bhuṅkte "he eats as much as he gets". Bhaṭṭoji: as much as he gets, that much (the 1915 print reads yāval labhate).
yāvat + jīv + ṇamul → yāvajjīvam adhīte "he studies as long as he lives".
Applies when / blocked by. The Bālamanoramā says "in the earlier time" (pūrvakāle) is not connected here. It does not fit and is not understood.
Watch out. Do not confuse this with SK 3350. "As much as he gets" is not totality. For totality, brāhmaṇavedam is already given by SK 3350, says the Tattvabodhinī.
Rule. When carman "hide" or udara "belly" is the object upapada, pūri takes ṇamul. Bhaṭṭoji says "as object" (karmaṇi) continues. pūri is the causative of pūr "to fill": the Tattvabodhinī says only the causative takes an object.
Example.
carma + pūri + ṇamul → the ṇi is lost → carmapūraṃ stṛṇāti "he spreads a covering, filling the hide".
udara + pūri + ṇamul → udarapūraṃ bhuṅkte "he eats, filling the belly".
Applies when / blocked by. Bhaṭṭoji does not name the source of karmaṇi. The Tattvabodhinī traces the karmaṇi of SK 3356 to SK 3350. That the same karmaṇi runs here is not stated by Bhaṭṭoji or his commentators; it is inferred. The Tattvabodhinī says "in the earlier time" is not connected here.
Watch out. udarapūraṃ bhuṅkte does not mean "he eats in order to fill the belly". The eating is qualified by the filling: udaraṃ pūrayitvā.
Rule. When the measure of rain is understood, pūri takes ṇamul with an object upapada. The ū of this pūri is optionally lost (lopa). Bhaṭṭoji says the measure is understood from the whole compound.
Example.
goṣpada + pūri + ṇamul → goṣpadapūram
or, with the ū lost → goṣpadapram.
goṣpadapūraṃ vṛṣṭo devaḥ / goṣpadapraṃ vṛṣṭo devaḥ "the god has rained just enough to fill a cow's hoofprint".
Applies when / blocked by. Why "of this" (asya)? So that the loss does not touch the upapada. The Bālamanoramā: asya means "of pūri". mūṣikābilapram "enough to fill a mouse-hole": the ū of mūṣikā stays.
Watch out. Bhaṭṭoji supplies pūri, the object and ṇamul without naming SK 3352. That they come from it is not stated by him or his commentators; it is inferred. The Bālamanoramā says smallness is understood: only a hoofprint was filled.
Rule. When a word meaning cloth (cela) is the object upapada, knopi takes ṇamul, the measure of rain being understood. Both commentators say cele takes in its meaning, so synonyms count.
Example.
cela + knopi + ṇamul → the ṇi is lost → celaknopaṃ vṛṣṭo devaḥ "the god has rained so much as to make the clothes sound (or be soaked)".
vastraknopam, vasanaknopam.
Applies when / blocked by. knopi is the causative of knūy, with the augment p (SK 2570). This is not stated by Bhaṭṭoji or his commentators; it is inferred. The Bālamanoramā's text as transmitted places the root in the cur class. Bhaṭṭoji does not name SK 3353 as the source of "measure of rain"; that too is inferred.
Watch out. The Tattvabodhinī gives two senses. Either the clothes make a noise, as clothes do in a downpour, or they become wet. The Bālamanoramā takes the "sound" sense only.
nimūla + kaṣ + ṇamul → (vṛddhi) nimūla-kāṣ-am → nimūlakāṣaṃ kaṣatiThese twelve rules (SK 3355–3366) form the kaṣādi group, "the rules beginning with kaṣ". In each, a root takes ṇamul beside a fixed upapada. Then the same root follows as the finite verb: nimūlakāṣaṃ kaṣati. The verb used after the ṇamul-form is the anuprayoga, "the word used after". SK 3367 closes the group by requiring that it be the very root that took ṇamul. In this group the earlier-time condition of SK 3320 is not connected, Bhaṭṭoji says at SK 3355. The upapada is first an object (SK 3355–3357), then an instrument (SK 3358–3361), then a place or a name (SK 3362–3363), then an agent (SK 3364–3365), and at SK 3366 a standard of comparison as object or, by ca, as agent. The running example is nimūlakāṣaṃ kaṣati "he destroys it down to the root".
Rule. When nimūla "down to the root" or samūla "root and all" is the object upapada, kaṣ "to destroy" takes ṇamul. Bhaṭṭoji says "as object" continues. He adds that he will state the anuprayoga for these rules (SK 3367). In this section "in the earlier time" is not connected.
Example.
nimūla + kaṣ + ṇamul → vṛddhi of the penultimate a (SK 2282) → nimūla + kāṣ + am → nimūlakāṣaṃ kaṣati.
samūlakāṣaṃ kaṣati.
Bhaṭṭoji: the meaning is nimūlaṃ kaṣati, samūlaṃ kaṣati.
Applies when / blocked by. Both verbs are kaṣ. Is this not mere repetition? Bhaṭṭoji answers: the root's meaning is one, but the qualifier nimūla makes a difference. So the two stand as qualifier and qualified, as particular and general. The Tattvabodhinī compares āmro vṛkṣaḥ "a mango tree": the particular qualifies the general. The Bālamanoramā compares mahendra beside indra.
Watch out. Bhaṭṭoji's "as object" names no source rule here. The Tattvabodhinī traces the karmaṇi of SK 3356 to SK 3350. That it runs through this rule too is not stated by Bhaṭṭoji or his commentators; it is inferred.
Rule. When śuṣka "dry", cūrṇa "powder" or rūkṣa "rough" is the object upapada, piṣ "to grind" takes ṇamul. The Tattvabodhinī says "as object" comes from SK 3350.
Example.
śuṣka + piṣ + ṇamul → guṇa of the light penultimate (SK 2189) → śuṣkapeṣaṃ pinaṣṭi.
Bhaṭṭoji: the meaning is śuṣkaṃ pinaṣṭi "he grinds it dry".
cūrṇapeṣam, rūkṣapeṣam.
Applies when / blocked by. The Bālamanoramā notes that the verb used after is piṣ itself (SK 3367).
Watch out. Why is snehana "moistener" not added to this list? The Tattvabodhinī answers: here the upapada is an object, but at SK 3359 it is an instrument.
Rule. When samūla "root and all", akṛta "not done" or jīva "living" is the object upapada, han, kṛñ and grah take ṇamul. Bhaṭṭoji says "as object" still applies. The Bālamanoramā says the pairing is in order (yathāsaṃkhya, SK 128): samūla with han, akṛta with kṛñ, jīva with grah.
Example.
samūlaghātaṃ hanti "he kills it root and all".
akṛta + kṛ + ṇamul → vṛddhi → akṛtakāraṃ karoti "he does what was not done".
jīva + grah + ṇamul → vṛddhi of the penultimate a → jīvagrāhaṃ gṛhṇāti "he seizes it alive".
Applies when / blocked by. Bhaṭṭoji explains jīva: "that which lives" (jīvati). It is made with the agent affix ka taught for roots with an ik penultimate (SK 2897). So the meaning is "he seizes it while it is living".
Watch out. jīva here is not the action noun "life" made with ghañ. That would not give "he seizes it living". Nor can ghañ have an agent sense, since SK 3186 excludes the agent. The Bālamanoramā raises both points.
Rule. When a word for the instrument (karaṇa) is the upapada, han "to strike" takes ṇamul.
Example.
pādaghātaṃ hanti. Bhaṭṭoji: the meaning is pādena hanti "he strikes with the foot".
Applies when / blocked by. Why the rule, when SK 3369 also gives ṇamul after injury-roots? Bhaṭṭoji: it secures two things. The verb used after is the same root (SK 3367). And the compound is obligatory (SK 782). When the ṇamul-form goes with a different verb, SK 3369 applies instead. The Bālamanoramā explains why. SK 3369 lies outside the kaṣādi group, so it would not give the same root after. And it follows SK 3368, so its compound is only optional (SK 784).
Watch out. The Bālamanoramā reports that the Bhāṣya states the purpose differently: han in a sense other than injury, and the obligatory compound. Scholars discuss this; Bhaṭṭoji's own wording is given above.
Rule. When a moistener (snehana) is the instrument upapada, piṣ takes ṇamul. Bhaṭṭoji glosses snehana as "that by which something is moistened". The Tattvabodhinī says "instrument" comes from SK 3358.
Example.
udaka + piṣ + ṇamul → guṇa → uda for udaka (SK 996, says the Bālamanoramā) → udapeṣaṃ pinaṣṭi.
Bhaṭṭoji: the meaning is udakena pinaṣṭi "he grinds with water".
Watch out. Compare SK 3356. There the upapada is the object (śuṣkapeṣam); here it is the instrument (udapeṣam).
Rule. When a word meaning "hand" is the instrument upapada, varti and grah take ṇamul. varti is the causative of vṛt "to turn". Both commentators say haste takes in its meaning, so synonyms count.
Example.
hasta + varti + ṇamul → the ṇi is lost → hastavartaṃ vartayati; karavartam. Bhaṭṭoji: the meaning is "he makes a pellet with the hand".
hasta + grah + ṇamul → hastagrāhaṃ gṛhṇāti "he takes it by the hand"; karagrāham, pāṇigrāham.
Applies when / blocked by. The Bālamanoramā says "instrument" continues. That it comes from SK 3358 is not stated by Bhaṭṭoji or his commentators; it is inferred.
Rule. When sva "property" is the instrument upapada, puṣ "to nourish" takes ṇamul. Bhaṭṭoji says "instrument" still applies. sva takes in its meaning. So ṇamul comes with sva itself, with a synonym, or with a particular kind of property.
Example.
sva + puṣ + ṇamul → guṇa (SK 2189) → svapoṣaṃ puṣṇāti "he nourishes with property".
A synonym: dhanapoṣam. A particular kind: gopoṣam.
Applies when / blocked by. The Bālamanoramā points to the Bhāṣya on SK 25 (svaṃ rūpaṃ śabdasya…) for this reading.
Watch out. Normally a word in a sūtra stands for its own form only (SK 25). Here sva is read for its meaning, which is why dhanapoṣam and gopoṣam are correct.
Rule. When a word for the place (adhikaraṇa) is the upapada, bandh "to bind" takes ṇamul.
Example.
cakra + bandh + ṇamul → cakrabandhaṃ badhnāti. Bhaṭṭoji: the meaning is cakre badhnāti "he binds him on the wheel".
Watch out. The verb after is again bandh (SK 3367): badhnāti.
Rule. bandh takes ṇamul when a name (saṃjñā) is meant. Both commentators say the rule is begun for an upapada that is not a place. So "place" of SK 3362 does not continue.
Example.
krauñcabandhaṃ baddhaḥ "bound in the krauñca-binding"; mayūrikābandham; aṭṭālikābandham.
Bhaṭṭoji: these are names of particular kinds of binding.
Rule. When jīva "living being" or puruṣa "man" is the upapada as agent (kartṛ), naś and vah take ṇamul, in order. The Bālamanoramā: naś goes with jīva, vah with puruṣa.
Example.
jīva + naś + ṇamul → vṛddhi of the penultimate a → jīvanāśaṃ naśyati. Bhaṭṭoji: the meaning is jīvo naśyati "the living being perishes".
puruṣa + vah + ṇamul → puruṣavāhaṃ vahati. Bhaṭṭoji: the meaning is puruṣo vahati "the man carries".
Watch out. Up to now the upapada was an object, instrument or place. Here it is the agent of the action.
Rule. When ūrdhva "upright" is the upapada as agent, śuṣ "to dry" and pūr "to fill" take ṇamul. Bhaṭṭoji glosses "ūrdhva as agent". That the agent sense is the kartroḥ of SK 3364 is not stated by Bhaṭṭoji or his commentators; it is inferred.
Example.
ūrdhva + śuṣ + ṇamul → guṇa → ūrdhvaśoṣaṃ śuṣyati. Bhaṭṭoji: a tree or the like dries up while still standing upright.
ūrdhva + pūr + ṇamul → ūrdhvapūraṃ pūryate. Bhaṭṭoji: a pot or the like, mouth upward just as it stands, fills with rain-water.
Watch out. The eva in Bhaṭṭoji's glosses matters: "while still standing upright".
Rule. When a word for the standard of comparison (upamāna) is the upapada as object, a root takes ṇamul. Bhaṭṭoji: by ca, also when it is the agent.
Example.
As object: ghṛta + ni-dhā + ṇamul → yuk (SK 2761) → ghṛtanidhāyaṃ nihitaṃ jalam "the water is put away as one puts away ghee". Bhaṭṭoji: guarded carefully, like ghee.
As agent: ajakanāśaṃ naṣṭaḥ. Bhaṭṭoji: he has vanished as an ajaka vanishes.
Watch out. In ghṛtanidhāyam the ghee is the object of ni-dhā; in ajakanāśam the ajaka is the agent of naś. The kaṣādi group ends here (SK 3367).
Rule. In the kaṣādi rules, the root used after the ṇamul-form (anuprayoga) is the very root after which ṇamul was prescribed (yathāvidhi). The Bālamanoramā gives the range: SK 3355 to SK 3366.
Example.
Bhaṭṭoji: "it has been exemplified in just that way". So nimūlakāṣaṃ kaṣati, śuṣkapeṣaṃ pinaṣṭi, pādaghātaṃ hanti: each verb repeats the root of the ṇamul-form. The Bālamanoramā says no separate example is needed.
Watch out. Outside this group the verb after may be another root. At SK 3368 the ṇamul-form of upa-daṃś is followed by bhuṅkte.
mūlaka + upa-daṃś + ṇamul → mūlakopadaṃśaṃ bhuṅkte (or mūlakenopadaṃśam)With SK 3368 the earlier-time condition of SK 3320 comes back. Bhaṭṭoji says it is connected "from here onward". The ṇamul-form again names the earlier of two actions with one agent, and the verb after it may be any root. Most rules name the upapada by its case. It is third case at SK 3368–3369 and seventh or third at SK 3370–3372. SK 3373 names it as the point of departure (apādāna). From SK 3374 to SK 3380 it is second case, SK 3375 apart. The compound of upapada and ṇamul-form is optional here, by SK 784 (tṛtīyāprabhṛtīny anyatarasyām). So the compound is optional; several examples show both forms. SK 3375 is a book-6 rule placed here for the ṇamul-form of apa-gur. The running example is mūlakopadaṃśaṃ bhuṅkte "he eats, biting on a radish".
Rule. When a third-case word is the upapada, daṃś "to bite" with upa takes ṇamul, for the earlier of two actions with one agent. Bhaṭṭoji: "from here onward 'in the earlier time' is connected". The Bālamanoramā says it is connected by a frog's leap (maṇḍūkapluti): it skips the rules where it does not fit (SK 3351–3367) and applies again here.
Example.
mūlaka + upa-daṃś + ṇamul → mūlaka + upadaṃśam → compounded (SK 784) → mūlakopadaṃśaṃ bhuṅkte.
Uncompounded: mūlakenopadaṃśaṃ bhuṅkte.
Applies when / blocked by. Why the third case on mūlaka? Bhaṭṭoji: the radish that is bitten is the instrument of eating. With "bite" it has no connection in the words. It does have a connection by sense, since it is in fact what is bitten. That much connection is enough for the affix and the compound. This is known from the force of the word tṛtīyāyām in the sūtra.
Watch out. Both commentators explain tṛtīyāyām. Had a direct link with the third-case word been meant, the sūtra would say "as instrument" (karaṇe), as SK 3358 does. Scholars discuss other readings of this sūtra; the Tattvabodhinī rejects them.
Rule. When a third-case word is the upapada, a root meaning injury (hiṃsā) takes ṇamul. The condition: it has the same object as the verb used after it.
Example.
daṇḍa + upa-han + ṇamul → daṇḍopaghātaṃ gāḥ kālayati "he drives the cows, striking them with a stick".
Uncompounded: daṇḍenopaghātam.
With taḍ "to strike": daṇḍatāḍam.
The cows are the object of both upahanti and kālayati. Both commentators take kālayati from kal "to drive" (cur class).
Applies when / blocked by. Why "having the same object"? daṇḍena coram āhatya gāḥ kālayati "having struck the thief with a stick, he drives the cows". The thief and the cows differ. So there is no ṇamul; ktvā comes, replaced by lyap after the prefix (SK 3332).
Watch out. Compare SK 3358 (pādaghātaṃ hanti), where the verb after is han itself and the compound is obligatory. Here the verb after is another root and the compound is optional.
Rule. When a seventh-case word is the upapada, pīḍ, rudh and karṣ with upa take ṇamul. By ca, a third-case word also counts, says the Bālamanoramā.
Example.
pārśvopapīḍaṃ śete "he lies pressing his sides": pārśvayor upapīḍam, pārśvābhyām upapīḍam.
vrajoparodhaṃ gāḥ sthāpayati "he keeps the cows, confining them in the pen": vrajena or vraje uparodham.
pāṇyupakarṣaṃ dhānāḥ saṃgṛhṇāti "he gathers the parched grain, drawing it in with the hand": pāṇāv upakarṣam, pāṇinopakarṣam.
Applies when / blocked by. Both commentators say karṣa names kṛṣ "to scratch" of the bhū class, with its guṇa. The tud class kṛṣ, used of ploughing, is excluded. It takes only ktvā: kṣetre upakṛṣya, halenopakṛṣya.
Watch out. Each example has three forms: the compound, the seventh case, and the third case.
Rule. When a third- or seventh-case word is the upapada, a root takes ṇamul if closeness (samāsatti) is understood. Bhaṭṭoji glosses it as sannikarṣa. The Bālamanoramā: contact with no interval.
Example.
keśa + grah + ṇamul → vṛddhi → keśagrāhaṃ yudhyante "they fight, seizing one another by the hair". Bhaṭṭoji: keśeṣu gṛhītvā.
hastagrāham. Bhaṭṭoji: hastena gṛhītvā.
Watch out. The Tattvabodhinī says the expression conveys closeness whether or not the hair is actually seized. The fighters are at the closest quarters.
Rule. Also when a measure (pramāṇa) is understood, a root takes ṇamul with a third- or seventh-case upapada. Bhaṭṭoji says "third or seventh" continues.
Example.
dvyaṅgula + ut-kṛṣ + ṇamul → guṇa → dvyaṅgulotkarṣaṃ khaṇḍikāṃ chinatti "he cuts off a small piece, marking it off at two finger-breadths".
Uncompounded: dvyaṅgulena or dvyaṅgule utkarṣam.
Applies when / blocked by. The Tattvabodhinī: dvyaṅgula "two fingers' breadth" takes the compound-final affix ac by SK 786. It glosses utkṛṣya as "marking off".
Rule. When a word for the point of departure (apādāna) is the upapada, a root takes ṇamul if haste is understood. Bhaṭṭoji: parīpsā means haste (tvarā).
Example.
śayyā + ut-sthā + ṇamul → th after ud (SK 118) → yuk (SK 2761) → śayyotthāyaṃ dhāvati "springing up from bed, he runs".
Watch out. Without haste there is no ṇamul. The bed is the point he leaves; the hurry is what licenses the affix.
Rule. Also when a second-case word is the upapada, a root takes ṇamul, haste being understood. Bhaṭṭoji says "in haste" continues.
Example.
yaṣṭi + grah + ṇamul → yaṣṭigrāhaṃ yudhyante "they fight, snatching up sticks".
loṣṭagrāham "snatching up clods".
Applies when / blocked by. The Tattvabodhinī explains the haste. They are so eager to fight that they snatch up the stick at hand. They do not wait for a weapon.
Rule. Before ṇamul, the diphthong (ec) of gur "to raise" with apa is optionally replaced by ā. Bhaṭṭoji names the root as gurī udyamane. The words "diphthong", "ā" and "optionally" are supplied in his gloss.
Example.
asi + apa-gur + ṇamul → guṇa of the light penultimate (SK 2189) → apa-gor-am
→ optionally ā for o → apa-gār-am
→ asyapagoraṃ yudhyante / asyapagāraṃ yudhyante "they fight with swords raised".
The Tattvabodhinī glosses: asim udyamya "having raised the sword".
Applies when / blocked by. Bhaṭṭoji gives no derivation. That the diphthong is the o made by guṇa is not stated by Bhaṭṭoji or his commentators; it is inferred. So is the source of "diphthong", "ā" and "optionally" in SK 2370 and SK 2509.
Watch out. This rule does not give ṇamul; it changes the root before it. It stands among the ṇamul rules because its forms need them.
Rule. When the upapada is a second-case word for a part of one's own body (svāṅga) that is not dhruva, a root takes ṇamul. Bhaṭṭoji says "second case" continues.
Example.
bhrū + vi-kṣip + ṇamul → guṇa → bhrūvikṣepaṃ kathayati "he speaks with a toss of the brows".
Uncompounded: bhruvaṃ vikṣepam.
Applies when / blocked by. Why "not dhruva"? śira utkṣipya "raising the head": no ṇamul. Bhaṭṭoji defines: that without which there is no life is dhruva.
Watch out. dhruva here means "vital", not "fixed". The head moves, yet it is dhruva, since one cannot live without it. The Bālamanoramā raises this very point.
Rule. Also when the second-case body-part word names something afflicted on every side, a root takes ṇamul. Bhaṭṭoji: this rule is for a vital (dhruva) part.
Example.
uras + prati-piṣ + ṇamul → guṇa → uraḥpratipeṣaṃ yudhyante "they fight crushing the chest". Bhaṭṭoji: they press hard on the whole chest.
urovidāraṃ praticaskare nakhaiḥ "he was mauled with claws, so that the chest was torn open".
Applies when / blocked by. The Bālamanoramā explains the need. The chest is vital, so SK 3376 does not reach it. This rule does, when the whole chest is afflicted. It takes praticaskare as the passive perfect of prati-kṝ, with the augment s by SK 2539–2540.
Watch out. SK 3376 is for a part that is not vital; SK 3377 is for a vital part, afflicted all over.
Rule. When a second-case word is the upapada, viś, pat, pad and skand take ṇamul. The sense is either "being pervaded" (vyāpyamāna) or "being repeated" (āsevyamāna). Bhaṭṭoji: pervasion is the complete connection of things such as houses with the action of entering and the rest. Repetition is the doing of the action again and again.
Example.
geha + anu-pra-viś + ṇamul → guṇa → gehānupraveśam āste "he stays, entering all the houses" or "entering the house again and again".
Likewise gehānuprapātam, gehānuprapādam, gehānuskandam.
Uncompounded, one word is doubled:
gehaṃ-geham anupraveśam;geham anupraveśam-anupraveśam.Applies when / blocked by. In the compound there is no doubling by SK 2140. The compound by its nature already expresses both senses.
Watch out. ṇamul for repetition was already taught at SK 3343. Why teach it again? Bhaṭṭoji: so that the second-case word is an upapada. The Bālamanoramā adds the gain: only then is the optional compound of SK 784 available.
Rule. When second-case time-words are the upapada, as "to throw" and tṛṣ "to thirst" take ṇamul. The roots must mean kriyāntara: "putting an interval into an action". Bhaṭṭoji: kriyāntara is what separates (vyavadhatte) an action.
Example.
dvyaha + ati-as + ṇamul → vṛddhi of the penultimate a → dvyahātyāsaṃ gāḥ pāyayati "he waters the cows, letting two days pass". Uncompounded: dvyaham atyāsam.
dvyaha + tṛṣ + ṇamul → guṇa → dvyahatarṣam; uncompounded dvyahaṃ tarṣam.
Bhaṭṭoji: having watered them today, he lets two days go by and waters them again.
Applies when / blocked by. The passing over and the making thirsty separate the cows' drinking by an interval. The Bālamanoramā glosses dvyahatarṣam as "having made them thirsty for two days".
Rule. When nāman "name" in the second case is the upapada, diś with ā and grah take ṇamul. Bhaṭṭoji says "second case" still holds.
Example.
nāman + ā-diś + ṇamul → guṇa → nāmādeśam ācaṣṭe "he tells, announcing the name".
nāman + grah + ṇamul → nāmagrāham āhvayati "he calls, taking the name".
Applies when / blocked by. The Bālamanoramā splits the sūtra nāmni ādiśigrahoḥ. Its sixth-case dual ādiśigrahoḥ stands for the fifth: "after ā-diś and grah".
uccaiḥ + kṛ → uccaiḥkṛtya (compound, lyap) / uccaiḥ kṛtvā / uccaiḥkāramIn the last six rules both affixes, ktvā and ṇamul, come together (named at SK 3381 and carried on). The root is kṛ "to do", or kṛ and bhū "to become", or bhū alone. The upapada is an indeclinable word: uccaiḥ, tiryak, a body-part word in -tas, a word in nā or dhā, tūṣṇīm, anvak. The compound is optional (SK 784), so each sense has three forms. Compounded, ktvā is replaced by lyap (SK 3332): uccaiḥkṛtya. Uncompounded, ktvā stays: uccaiḥ kṛtvā. With ṇamul: uccaiḥkāram. The running example is uccaiḥkāram apriyam ācaṣṭe "he tells the unpleasant thing loudly". After SK 3386 Bhaṭṭoji closes the chapter, and the second half (uttarārdha) of the Siddhānta-kaumudī, with two verses.
Rule. When an indeclinable (avyaya) is the upapada, kṛñ takes ktvā and ṇamul in "a telling contrary to what is intended" (ayathābhipretākhyāna). Bhaṭṭoji defines it: saying something unpleasant loudly, and something pleasant softly.
Example.
uccaiḥ + kṛ + ktvā → compounded (SK 784) → lyap for ktvā (SK 3332) → uccaiḥ + kṛ + ya → t after the short vowel (tuk, SK 2858) → uccaiḥkṛtya.
Uncompounded: uccaiḥ kṛtvā.
uccaiḥ + kṛ + ṇamul → vṛddhi → uccaiḥkāram apriyam ācaṣṭe "he tells the unpleasant thing loudly".
Likewise nīcaiḥkṛtya, nīcaiḥ kṛtvā, nīcaiḥkāraṃ priyaṃ brūte "he speaks the pleasant thing softly".
Applies when / blocked by. The Bālamanoramā: what is intended is to tell the unpleasant softly and the pleasant loudly; the reverse is not intended. It adds that uccaiḥkāram compounded and uncompounded differ only in accent.
Watch out. Why not say simply "in telling the unpleasant" (apriyākhyāne)? The Tattvabodhinī answers. Saying the pleasant softly would then be excluded, though nīcaiḥkāraṃ priyaṃ brūte is wanted. And saying the unpleasant softly would wrongly be included.
Rule. When tiryak "crosswise" is the upapada, kṛñ takes ktvā and ṇamul if completion (apavarga) is meant. Bhaṭṭoji glosses apavarga as samāpti "finishing".
Example.
tiryakkṛtya gataḥ, tiryak kṛtvā gataḥ, tiryakkāram. Bhaṭṭoji: the meaning is "he went, having finished the matter".
Applies when / blocked by. Why "completion"? tiryak kṛtvā kāṣṭhaṃ gataḥ "having put the log crosswise, he went". The Tattvabodhinī explains: the crooked log stood in front, and he set it aside. Here tiryak is literal, so only ktvā, uncompounded.
Watch out. The Bālamanoramā notes that the regular locative would be tiraści. The sūtra's tiryaci is a licence of the sūtra's own wording.
Rule. When a body-part word (svāṅga) ending in the affix tas is the upapada, kṛ and bhū take ktvā and ṇamul. Both commentators say the pairing is not one-to-one: each root takes both affixes.
Example.
mukhataḥkṛtya gataḥ "he went, having put it in front"; mukhataḥ kṛtvā; mukhataḥkāram.
mukha + tas + bhū + ktvā → compounded → lyap → mukhatobhūya; uncompounded mukhato bhūtvā.
bhū + ṇamul → vṛddhi, au → āv (SK 61) → mukhato bhāvam.
Applies when / blocked by. The Tattvabodhinī says mukhataḥ has tas in the sense of the seventh case. Why the word "affix" (pratyaya)? A mukhataḥ made otherwise, from a root with kvip, is not an affix-form, so only ktvā comes: mukhataḥ kṛtvā gataḥ. The Tattvabodhinī adds that a general maxim already secures this.
Watch out. Scholars discuss how the compound taspratyaya is to be analysed. The Tattvabodhinī records three readings and notes that all give the same upapada.
Rule. When the upapada ends in the affix nā, or in an affix with the sense of dhā, kṛ and bhū take ktvā and ṇamul. The upapada must carry the sense of cvi: making or becoming what a thing was not.
Example.
anānā nānā kṛtvā "making separate what was not separate":
nānākṛtya; nānā kṛtvā; nānākāram;vinākṛtya; vinā kṛtvā; vinākāram;nānābhūya; nānā bhūtvā; nānābhāvam.anekaṃ dravyam ekaṃ bhūtvā "many things having become one":
ekadhābhūya; ekadhā bhūtvā; ekadhābhāvam;ekadhākṛtya; ekadhā kṛtvā; ekadhākāram.Applies when / blocked by. The Bālamanoramā names the affixes. nā is taught after vi and nañ by SK 1828. dhā is the affix of SK 1988. "With the sense of dhā" also takes in dhamuñ and the like (SK 1991, says the Tattvabodhinī). Why the word "affix"? hiruk kṛtvā, pṛthag bhūtvā. hiruk and pṛthak "apart" have the same sense but are not affix-forms. So only ktvā, uncompounded.
Watch out. Scholars discuss whether "sense" (artha) attaches to nā as well as to dhā. The set of affixes is the same either way.
Rule. When tūṣṇīm "silently" is the upapada, bhū takes ktvā and ṇamul. kṛ does not.
Example.
tūṣṇīṃbhūya; tūṣṇīṃ bhūtvā; tūṣṇīṃbhāvam.
Applies when / blocked by. bhū would carry forward from SK 3383 anyway. Why name it? The Bālamanoramā: to stop kṛ. The Tattvabodhinī says instead that the mention of tūṣṇīm stops kṛ. Scholars discuss which word does the work; the result is the same.
Rule. When anvak "after, behind" is the upapada, bhū takes ktvā and ṇamul if conformity (ānulomya) is understood. Bhaṭṭoji glosses ānulomya as ānukūlya "compliance".
Example.
anvagbhūya āste; anvag bhūtvā; anvagbhāvam.
Bhaṭṭoji: "he stays, having become compliant", whether in front, at the side or behind.
Applies when / blocked by. Why "conformity"? anvag bhūtvā tiṣṭhati. Bhaṭṭoji: it means "he stands, having got behind". Here anvak gives only a place, says the Tattvabodhinī. So there is no ṇamul, only ktvā.
Watch out. The regular locative of anvak would be anūci, with lengthening by SK 417. The Bālamanoramā calls anvaci a licence of the sūtra, like tiryaci (SK 3382). The Tattvabodhinī holds it no fault, since a quoted word need not undergo every operation.
The close of the uttarārdha. Two verses follow the last example.
1. Only a bare indication (diṅmātra) of the words of ordinary usage has been shown here. The full treatment, by the śāstra, is in the Śabdakaustubha.
2. May this Siddhānta-kaumudī, the work of Bhaṭṭoji Dīkṣita, be for the delight of the blessed Bhavānī and Viśvanātha.
Then come two colophons, supplied in brackets from the 1915 print: [iti uttarakṛdantam] "thus ends the Uttarakṛdanta", and [iti śrībhaṭṭojidīkṣitaviracitāyāṃ siddhāntakaumudyām uttarārdhaṃ samāptam] "thus ends the second half of the Siddhānta-kaumudī composed by the illustrious Bhaṭṭoji Dīkṣita".
tantuḥ, vartma (SK 3169).draṣṭum (SK 3175).darśakaḥ (SK 2895, SK 3175).pākaḥ (SK 3182, SK 3184).cayaḥ, ravaḥ. They are exceptions to ghañ.rogaḥ, ravaḥ.pādaḥ, saṃrāvaḥ.draṣṭum.vyāvakrośī, upādhyāyī.vartma "that which is traversed".śāraḥ "wind", "that by which leaves are shattered".dāśaḥ, goghnaḥ (SK 3172).bhīmaḥ (SK 3173), upādhyāyaḥ (SK 3191).raṅgaḥ, saṃstāvaḥ.draṣṭuṃ yāti, sam in saṃrāvaḥ.prabhāvaḥ.vyāvakrośa, sāṃrāvin.gamī, syadaḥ, edhaḥ.rājño nayaḥ, unnayaḥ.pākaḥ "cooking" (SK 3184).pākau.dārāḥ (SK 3190).ekas taṇḍulanicāyaḥ (SK 3190).nivṛtam (SK 3191).viṣṭārapaṅktiḥ.nikecāyaḥ (SK 3213).saṃrāvin (SK 3218).karaḥ, grahaḥ (SK 3232, 3234).cayaḥ (SK 3231).samucchrayāḥ, nivarā.ucchrāyaḥ and s → ś before c in niścayaḥ.praśnaḥ (SK 3268).prasthaḥ), and SK 2363 drops the penultimate a of han and a few other roots (vighnaḥ, stambaghnaḥ).praghasaḥ, ghāsaḥ (SK 3236).nyādaḥ (SK 3237).nihavaḥ.drughaṇaḥ.upaghna.paktrimam (SK 1570).pākena nirvṛttam for paktrimam.vepathuḥ (SK 3267).praśnaḥ, svapnaḥ (SK 3268–3269).pradhiḥ, jaladhiḥ (SK 3270–3271).rephaḥ, akāraḥ).sampat (SK 3272).kīrṇiḥ, lūniḥ (SK 3272).vrajyā, kṛtyā (SK 3275–3277).kṛtyā, ityā.kriyā (SK 3277).icchā, present in paricaryā (SK 3278).aṭāṭyā and aṭāṭā (SK 3278).jarā, cintā, pradā (SK 3281–3283).kāraṇā, supānaḥ (SK 3284, 3309).pracchardikā, āsikā (SK 3285, 3288).paciḥ, pacatiḥ (SK 3285).pacatiḥ (SK 3285).akāraḥ, kakāraḥ (SK 3285).matvarthīyaḥ (SK 3285).ājiḥ, vāpiḥ, kāriḥ (SK 3285, 3287).ajīvaniḥ (SK 3289).hasanam, godohanī (SK 3290, 3293).payaḥpānam (SK 3291).ākaraḥ, dantacchadaḥ (SK 3296).duṣkaraḥ (SK 3305).supralambhaḥ (SK 3306), durāḍhyambhavam (SK 3308).śayitvā.naṃṣṭvā (SK 2517).dyūtvā.pretya) or the affix of the same name.smāram, kāṣam. Taught first at SK 3343.smṛ → smār-, bhuj → bhoj-.smāraṃ smāram.pāyam.bhuṅkte.cauraṅkāram (SK 3346).goṣpadapram (SK 3353).nimūlakāṣaṃ kaṣati.mukhataḥ, here in the sense of the seventh case (SK 3383).| SK | what it does | example |
|---|---|---|
| 3169 | the uṇādi affixes come variously, in the present and in names; unlisted ones are inferred | śaṅkulā, ṛphiḍaḥ |
| 3170 | the uṇādis are seen also in the past | — |
| 3171 | gamī and the rest are fixed forms in the future | gamī grāmam, āgāmī |
| 3172 | dāśa and goghna are fixed forms in the recipient sense | dāśaḥ, goghno 'tithiḥ |
| 3173 | bhīma and the rest are fixed forms in the source sense | bhīmaḥ, bhīṣmaḥ, mūrkhaḥ |
| 3174 | in the other kārakas the uṇādis come | tantuḥ, vartma, carma |
| 3175 | tumun and ṇvul in the future, for an action done for the sake of another | kṛṣṇaṃ draṣṭuṃ yāti, kṛṣṇaṃ darśako yāti |
| 3176 | tumun with a word of desire having the same agent | icchati bhoktum |
| 3177 | tumun with śak, dhṛṣ … arh, or a word meaning "is" | śaknoti bhoktum, asti bhoktum |
| 3178 | tumun with a word of completeness meaning capability | paryāpto bhoktum; not alaṃ bhuktvā |
| 3179 | tumun with a word for time, in the senses of bidding and the rest | kālo bhoktum |
| 3180 | the bhāva affixes also express the purpose | yāgāya yāti |
| 3181 | aṇ with an object word and a purposive action; blocks ṇvul and ka | kāṇḍalāvo vrajati, kambaladāyo vrajati |
| 3182 | ghañ in the agent sense after pad, ruj, viś, spṛś | pādaḥ, rogaḥ, veśaḥ, sparśaḥ |
| 3183 | ghañ after sṛ for an enduring agent; vārttika: disease, fish, strength | sāraḥ, atīsāraḥ, visāraḥ |
| 3184 | heading: ghañ in the bhāva sense, the action as accomplished | pākaḥ, pākau |
| 3185 | ā for the ec of sphur, sphul before ghañ; notes on upasarga lengthening and SK 2763 | sphāraḥ, parīhāraḥ, śamaḥ, kāmaḥ |
| 3186 | heading: ghañ in a non-agent kāraka, in a name (as a rule) | prāsaḥ, lābhaḥ |
| 3187 | rañj loses its n before ghañ in the bhāva and instrument senses | rāgaḥ; but raṅgaḥ |
| 3188 | syada "speed" by nipātana: n lost, no vṛddhi | syadaḥ; syandaḥ |
| 3189 | avoda, edha, odma, praśratha, himaśratha by nipātana | avodaḥ, edhaḥ |
| 3190 | ghañ after any root when a quantity is stated, against ac and ap; vārttika: dāra, jāra | ekas taṇḍulanicāyaḥ, dvau kārau, dārāḥ |
| 3191 | ghañ after iṅ, against ac; vārttikas: feminine apādāna, śṝ | upādhyāyaḥ, upādhyāyī, śāraḥ |
| 3192 | ghañ after ru with an upasarga | saṃrāvaḥ; ravaḥ |
| 3193 | ṣ for the s of stan after abhinis, in a name of a sound | abhiniṣṭānaḥ |
| 3194 | ghañ after yu, dru, du with sam | saṃyāvaḥ, saṃdrāvaḥ |
| 3195 | ghañ after śri, nī, bhū without an upasarga | śrāyaḥ, nāyaḥ, bhāvaḥ; praśrayaḥ |
| 3196 | ghañ after kṣu, śru with vi | vikṣāvaḥ, viśrāvaḥ |
| 3197 | ghañ after nī with ava or ud | avanāyaḥ, unnāyaḥ |
| 3198 | ghañ after dru, stu, sru with pra | pradrāvaḥ, prastāvaḥ |
| 3199 | ghañ after pū, lū with nir or abhi | niṣpāvaḥ, abhilāvaḥ |
| 3200 | ghañ after gṝ with ud or ni | udgāraḥ, nigāraḥ; garaḥ |
| 3201 | ghañ after kṝ with ud or ni, of grain | utkāro dhānyasya; puṣpanikaraḥ |
| 3202 | ghañ after stu with sam, in sacrifice | saṃstāvaḥ; saṃstavaḥ |
| 3203 | ghañ after stṛ with pra, not in sacrifice | prastāraḥ; barhiṣaḥ prastaraḥ |
| 3204 | ghañ after stṛ with vi, for spreading out, not of speech | paṭasya vistāraḥ; granthavistaraḥ |
| 3205 | ghañ after vi + stṛ in the name of a metre | viṣṭārapaṅktiḥ |
| 3206 | ṣ for the s of vi + stṛ + ghañ in the name of a metre | viṣṭāra- |
| 3207 | ghañ after grah with ud | udgrāhaḥ |
| 3208 | ghañ after grah with sam, of the fist | mallasya saṃgrāhaḥ; dravyasya saṃgrahaḥ |
| 3209 | ghañ after pari + nī in gambling, ni + iṇ for propriety | pariṇāyaḥ, nyāyaḥ |
| 3210 | ghañ after pari + iṇ for a turn in due order | tava paryāyaḥ; kālasya paryayaḥ |
| 3211 | ghañ after śī with vi or upa, for one's turn | tava viśāyaḥ, rājopaśāyaḥ |
| 3212 | ghañ after ci for taking by hand, not theft | puṣpapracāyaḥ |
| 3213 | ghañ after ci, with k for the initial, for abode, piling, body, heap | nikāyaḥ, kāyaḥ, nikecāyaḥ |
| 3214 | the same for a multitude of beings not one above another | bhikṣunikāyaḥ; sūkaranicayaḥ |
| 3215 | ṇac in the feminine, bhāva sense, for a reciprocal action | vyavakrośa- |
| 3216 | taddhita añ after a ṇac-stem in the feminine | vyāvakrośa- |
| 3217 | no aic in a reciprocal action; plain ādivṛddhi on the upasarga | vyāvakrośī, vyāvahāsī |
| 3218 | inuṇ in the bhāva sense for complete pervasion | saṃrāvin |
| 3219 | taddhita aṇ after an inuṇ-stem; in kept by SK 1245 | sāṃrāviṇaṃ vartate |
| 3220 | ghañ after grah with ava or ni in a curse; otherwise ap | avagrāhas te bhūyāt; avagrahaḥ padasya |
| 3221 | ghañ after grah with pra for a desire to obtain | pātrapragrāheṇa carati |
| 3222 | ghañ after grah with pari in the sacrifice | uttaraḥ parigrāhaḥ |
| 3223 | ghañ after vṛ with ni for grain; otherwise ap, even in the feminine | nīvārāḥ; nivarā kanyā |
| 3224 | ghañ after śri, yu, pū, dru with ud | ucchrāyaḥ, uddrāvaḥ; samucchrayāḥ by bāhulaka |
| 3225 | optional ghañ after ru, plu with ā; the option runs to SK 3230 | ārāvaḥ / āravaḥ |
| 3226 | optional ghañ after grah with ava for drought | avagrāhaḥ / avagrahaḥ |
| 3227 | optional ghañ after grah with pra: the balance-string | tulāpragrāheṇa / tulāpragraheṇa |
| 3228 | the same for a rein | pragrāhaḥ / pragrahaḥ |
| 3229 | optional ghañ after vṛ with pra for a covering | prāvāraḥ / pravaraḥ |
| 3230 | optional ghañ after bhū with pari for disregard | paribhāvaḥ / paribhavaḥ |
| 3231 | ac after i/ī-final roots; vārttika: ac for bhaya etc. in the neuter | cayaḥ, jayaḥ; bhayam, varṣam |
| 3232 | ap after ṛ- and u-final roots | karaḥ, yavaḥ, lavaḥ |
| 3233 | viṣṭara (with ṣ) by nipātana: tree, seat | viṣṭaraḥ; vākyasya vistaraḥ |
| 3234 | ap after grah, vṛ, dṛ, niś-ci, gam; vārttikas: vaś, raṇ; ka for ghañ; doubling before ka | grahaḥ, niścayaḥ; raṇaḥ; vighnaḥ; cakram |
| 3235 | ap after ad with an upasarga | (praghasaḥ, SK 3236) |
| 3236 | ghasḷ for ad before ghañ or ap | praghasaḥ, vighasaḥ; ghāsaḥ |
| 3237 | ṇa and ap after ad with ni | nyādaḥ, nighasaḥ |
| 3238 | ap after vyadh, jap without an upasarga | vyadhaḥ, japaḥ; upajāpaḥ |
| 3239 | optional ap after svan, has without an upasarga | svanaḥ / svānaḥ; prasvānaḥ |
| 3240 | optional ap after yam with sam, upa, ni, vi, or alone | saṃyamaḥ / saṃyāmaḥ |
| 3241 | optional ap after gad, nad, paṭh, svan with ni | nigadaḥ / nigādaḥ |
| 3242 | optional ap after kvaṇ with ni, alone, or of the vīṇā | nikvaṇaḥ / nikvāṇaḥ; prakvaṇaḥ |
| 3243 | obligatory ap after paṇ for a quantity | mūlakapaṇaḥ; pāṇaḥ |
| 3244 | obligatory ap after mad with a non-upasarga upapada | dhanamadaḥ; unmādaḥ |
| 3245 | pramada, sammada by nipātana for joy | pramadaḥ; pramādaḥ |
| 3246 | ap after aj with sam, ud, of animals; no vī | samajaḥ, udajaḥ; samājaḥ |
| 3247 | glaha by nipātana: the stake in gaming | akṣasya glahaḥ |
| 3248 | ap after sṛ for conception | gavām upasaraḥ; avasaraḥ by SK 3296 |
| 3249 | ap and saṃprasāraṇa after hvā with ni, abhi, upa, vi | nihavaḥ; prahvāyaḥ |
| 3250 | the same with ā, for battle | āhavaḥ; āhvāyaḥ |
| 3251 | āhāva by nipātana: a drinking-trough | āhāvaḥ |
| 3252 | ap and saṃprasāraṇa after bare hvā in bhāva | havaḥ |
| 3253 | ap after bare han in bhāva, with vadha; ghañ too | vadhaḥ; ghātaḥ |
| 3254 | ap after han for solidity, with ghana | abhraghanaḥ |
| 3255 | antarghana by nipātana: a place | antarghanaḥ |
| 3256 | praghaṇa, praghāṇa by nipātana: the outer door-platform | praghaṇaḥ, praghāṇaḥ |
| 3257 | udghana by nipātana: the paring-block | udghanaḥ |
| 3258 | apaghana by nipātana: a limb | apaghanaḥ; apaghātaḥ |
| 3259 | ap after han with ayas, vi, dru, instrument, with ghana | ayoghanaḥ, drughanaḥ |
| 3260 | ka, or ap with ghana, after han with stamba, instrument | stambaghnaḥ, stambaghanaḥ |
| 3261 | ap after han with pari, instrument, with gha | parighaḥ |
| 3262 | optional l for the r of pari before gha, aṅka | palighaḥ, palyaṅkaḥ |
| 3263 | upaghna by nipātana: proximity | parvatopaghnaḥ |
| 3264 | saṃgha, udgha by nipātana: group, praise | saṃghaḥ, udghaḥ |
| 3265 | nigha by nipātana: evenly measured | nighā vṛkṣāḥ |
| 3266 | ktri after a ḍu-marked root, bhāva only, always with map | paktrimam, uptrimam |
| 3267 | athuc after a ṭu-marked root, bhāva only | vepathuḥ, śvayathuḥ |
| 3268 | naṅ after yaj, yāc, yat, vicch, pracch, rakṣ | yajñaḥ, praśnaḥ, yatnaḥ |
| 3269 | nan after svap | svapnaḥ |
| 3270 | ki after a ghu root with an upasarga | pradhiḥ, antardhiḥ, upādhiḥ |
| 3271 | ki after a ghu root with an object upapada, in the locus | jaladhiḥ |
| 3272 | heading striyām; ktin in feminine bhāva and non-agent senses, against ghañ, ac, ap; vārttikas: ktin in the instrument, niṣṭhāvat, ci for cāy, kvip and ktin after sampad | kṛtiḥ, citiḥ, śrutiḥ, kīrṇiḥ, apacitiḥ, sampat / sampattiḥ |
| 3273 | ktin in bhāva after sthā, gā, pā, pac, against aṅ | prasthitiḥ, saṅgītiḥ, paktiḥ |
| 3274 | six ktin-forms by nipātana, end-acute | ūtiḥ, sātiḥ, hetiḥ, kīrtiḥ |
| 3275 | kyap in feminine bhāva after vraj, yaj | vrajyā, ijyā |
| 3276 | kyap in a name after nine roots; no vī for aj | samajyā, vidyā, śayyā, ityā |
| 3277 | split rule: kyap, and śa with ktin, after kṛ | kṛtyā, kriyā, kṛtiḥ |
| 3278 | icchā by nipātana; vārttikas: śa + yak forms; a after jāgṛ | icchā, paricaryā, mṛgayā, aṭāṭyā, jāgarā |
| 3279 | a after a root ending in an affix | cikīrṣā, putrakāmyā |
| 3280 | a after a consonant-final root with a heavy vowel, if seṭ in the niṣṭhā | īhā, ūhā; but āptiḥ, nigṛhītiḥ |
| 3281 | aṅ after ṣit roots and bhid etc.; kṛpā | jarā, trapā, bhidā, kṛpā |
| 3282 | aṅ after cinti, pūji, kathi, kumbi, carc, against yuc | cintā, pūjā, kathā |
| 3283 | aṅ after an ā-final root with an upasarga; śrat, antar as upasargas | pradā, śraddhā, antardhiḥ |
| 3284 | yuc after ṇi-finals, ās, śranth; vārttikas for ghaṭṭ, vand, vid, iṣ | kāraṇā, āsanā, vedanā, anveṣaṇā |
| 3285 | ṇvul for disease names, variously; vārttikas: ṇvul, ik/śtip, kāra, ipha, cha, iṇ, iñ, ik | pracchardikā, śirortiḥ, pacatiḥ, kakāraḥ, rephaḥ, vāpiḥ |
| 3286 | ṇvul in a name | uddālakapuṣpabhañjikā |
| 3287 | iñ and ṇvul (and the others) in question and reply | kāriṃ kārikāṃ kriyāṃ kṛtyāṃ kṛtiṃ vā |
| 3288 | ṇvuc for turn, desert, debt, arising | bhavata āsikā, ikṣubhakṣikā |
| 3289 | ani in a curse with nañ; SK 2841 and SK 3090 re-cited; feminine heading ends | ajīvaniḥ, rājabhojanāḥ śālayaḥ |
| 3290 | lyuṭ (and kta) in neuter bhāva | hasanam, hasitam |
| 3291 | lyuṭ with an object giving bodily pleasure: obligatory compound | payaḥpānaṃ sukham |
| 3292 | vī optionally for aj before yu | pravayaṇam / prājanam |
| 3293 | lyuṭ for instrument and place; heading to SK 3305 | idhmapravraścanaḥ, godohanī |
| 3294 | ṇ for the n of han after antar, no place meant | antarhaṇanam; antarhanano deśaḥ |
| 3295 | ṇ in ayana after antar, no place meant | antarayaṇam; antarayano deśaḥ |
| 3296 | gha for masculine names of instrument or place, generally | ākaraḥ |
| 3297 | short vowel in chādi before gha, unless two upasargas | dantacchadaḥ |
| 3298 | seven gha-forms by nipātana, against ghañ | gocaraḥ, vyajaḥ, nigamaḥ, kaṣaḥ |
| 3299 | ghañ after tṝ, stṝ with ava | avatāraḥ, avastāraḥ |
| 3300 | ghañ after a consonant-final root, against gha | rāmaḥ, apāmārgaḥ |
| 3301 | four ghañ-forms by nipātana; three more by vārttika | adhyāyaḥ, nyāyaḥ, ādhāraḥ |
| 3302 | udaṅka with ghañ, not of water | ghṛtodaṅkaḥ; udakodañcanaḥ |
| 3303 | ānāya with ghañ for a net | ānāyaḥ; ānayanaḥ |
| 3304 | gha and ghañ after khan; gha after other roots; ḍa etc. by vārttika | ākhanaḥ / ākhānaḥ, padam, ākhaḥ |
| 3305 | khal with īṣat, dus, su in bhāva and object; no ā for mi, mī, lī | duṣkaraḥ, sukaraḥ, īṣannimayaḥ |
| 3306 | num in labh before khal, ghañ only after an upasarga | supralambhaḥ, upālambhaḥ; lābhaḥ |
| 3307 | no num after bare su, dur | sulabham, durlabham; atidurlambhaḥ |
| 3308 | khal after bhū, kṛ with agent/object (cvi sense) and īṣat etc.; mum | durāḍhyambhavam, svāḍhyaṅkaraḥ |
| 3309 | yuc after ā-final roots, against khal; vārttika on śās etc. | supānaḥ, duḥśāsanaḥ, duryodhanaḥ |
| 3310 | no ṇ after a pada-final ṣ | niṣpānam, sarpiṣpānam |
| 3311 | ṇini in the agent for necessity or debt; no sixth case by SK 628 | avaśyaṅkārī, śataṃdāyī |
| 3312 | the kṛtyas for necessity or debt | avaśyaṃ hariḥ sevyaḥ, śataṃ deyam |
| 3313 | ktic and kta for a name with a blessing; no iṭ by SK 3163 | bhūtiḥ, devadattaḥ |
| 3314 | before ktic: no lengthening, no loss of the nasal | yantiḥ, rantiḥ, vantiḥ, tantiḥ |
| 3315 | before ktic: san takes ā or loses n, optionally | sātiḥ, satiḥ, santiḥ |
| 3316 | ktvā with alam, khalu meaning prohibition; no upapada compound | alaṃ dattvā, pītvā khalu |
| 3317 | ktvā after māṅ for exchange, with no prior time; optional | apamitya yācate |
| 3318 | optional i for the final of meṅ before lyap; ordinary ktvā beside it | apamitya / apamāya; yācitvā apamayate |
| 3319 | ktvā for connection with the farther or nearer | aprāpya nadīṃ parvataḥ, atikramya parvataṃ sthitā nadī |
| 3320 | ktvā for the earlier of two actions with one agent, in bhāva | bhuktvā vrajati, natvā, svṛtvā |
| 3321 | skand, syand keep their n before ktvā | skantvā, syantvā / syanditvā |
| 3322 | a seṭ ktvā is not kit | śayitvā; kṛtvā |
| 3323 | a seṭ ktvā is kit after mṛḍ, mṛd, gudh, kuṣ, kliś, vad, vas | mṛḍitvā, uditvā, uṣitvā; ruditvā |
| 3324 | optionally kit after n-penultimate th/ph-final roots | śrathitvā / śranthitvā; kothitvā |
| 3325 | optionally kit after vañc, luñc, ṛt | vacitvā / vañcitvā, ṛtitvā / artitvā |
| 3326 | optionally kit after tṛṣ, mṛṣ, kṛś; further kit and iṭ cases | tṛṣitvā / tarṣitvā; dyutitvā; añcitvā; lubdhvā |
| 3327 | iṭ for ktvā after jṝ and vraśc | jarītvā / jaritvā, vraścitvā |
| 3328 | optional iṭ for ktvā after an udit root | śamitvā / śāntvā, devitvā / dyūtvā |
| 3329 | optional long penultimate of kram before a jhal-initial ktvā | krāntvā / krantvā; kramitvā |
| 3330 | optional loss of n for j-final roots and naś, before a jhal-initial ktvā | bhaktvā / bhaṅktvā, naṣṭvā / naṃṣṭvā; khātvā, hitvā |
| 3331 | hi for hā (jahāti) before ktvā | hitvā; hātvā; jagdhvā |
| 3332 | lyap for ktvā in a compound with an indeclinable other than nañ | prakṛtya; akṛtvā, paramakṛtvā |
| 3333 | the single substitute is asiddha for ṣatva and tuk | ko 'sicat; adhītya, pretya |
| 3334 | optional loss of the nasal before lyap (restricted); lyap blocks inner operations | āgatya / āgamya, prahatya; vidhāya, pradīvya |
| 3335 | no ī for ghu, mā, sthā etc. before lyap | pradhāya, prapāya; vilāya / vilīya; vicārya |
| 3336 | ay for ṇi before lyap after a light vowel | vigaṇayya, praṇamayya; sampradhārya |
| 3337 | optional ay for the ṇi of āp | prāpayya / prāpya |
| 3338 | long vowel for kṣi before lyap | prakṣīya |
| 3339 | no saṃprasāraṇa for veñ before lyap | pravāya |
| 3340 | no saṃprasāraṇa for jyā before lyap | prajyāya |
| 3341 | no saṃprasāraṇa for vyeñ before lyap | upavyāya |
| 3342 | optional saṃprasāraṇa for vyeñ after pari; lengthening by SK 2559 over tuk | parivīya / parivyāya; mukhaṃ vyādāya svapiti |
| 3343 | ṇamul (and ktvā) for repetition in the domain of SK 3320; the form is doubled | smāraṃ smāram, smṛtvā smṛtvā; gāmaṃ gāmam; pralambhaṃ pralambham |
| 3344 | with yad, no ktvā or ṇamul if the sentence expects no other | yad ayaṃ bhuṅkte tataḥ paṭhati |
| 3345 | with agre, prathama, pūrva, ktvā and ṇamul optional; obligatory in repetition | agre bhojaṃ vrajati / agre bhuṅkte tato vrajati |
| 3346 | khamuñ after kṛñ with an object upapada, in abuse | cauraṅkāram ākrośati |
| 3347 | ṇamul after kṛñ with svādu or a synonym; m by nipātana | svāduṅkāraṃ bhuṅkte, lavaṇaṅkāram |
| 3348 | ṇamul after a redundant kṛñ with anyathā, evam, katham, ittham | itthaṅkāraṃ bhuṅkte |
| 3349 | the same with yathā, tathā, in a resentful reply | yathākāram ahaṃ bhokṣye |
| 3350 | ṇamul after dṛś, vid with an object upapada, for totality | kanyādarśaṃ varayati |
| 3351 | ṇamul after vind, jīv with yāvat | yāvajjīvam adhīte |
| 3352 | ṇamul after causative pūri with carman, udara | udarapūraṃ bhuṅkte |
| 3353 | the same for a measure of rain; ū optionally lost | goṣpadapraṃ vṛṣṭo devaḥ |
| 3354 | ṇamul after knopi with a cloth-word, for a measure of rain | celaknopaṃ vṛṣṭo devaḥ |
| 3355 | ṇamul after kaṣ with nimūla, samūla; earlier time not connected | nimūlakāṣaṃ kaṣati |
| 3356 | ṇamul after piṣ with śuṣka, cūrṇa, rūkṣa | śuṣkapeṣaṃ pinaṣṭi |
| 3357 | ṇamul after han, kṛñ, grah with samūla, akṛta, jīva, in order | jīvagrāhaṃ gṛhṇāti |
| 3358 | ṇamul after han with an instrument upapada | pādaghātaṃ hanti |
| 3359 | ṇamul after piṣ with a moistener as instrument | udapeṣaṃ pinaṣṭi |
| 3360 | ṇamul after varti, grah with a hand-word as instrument | hastagrāhaṃ gṛhṇāti |
| 3361 | ṇamul after puṣ with sva (its meaning) as instrument | dhanapoṣam |
| 3362 | ṇamul after bandh with a place upapada | cakrabandhaṃ badhnāti |
| 3363 | ṇamul after bandh in a name | krauñcabandhaṃ baddhaḥ |
| 3364 | ṇamul after naś, vah with jīva, puruṣa as agent | jīvanāśaṃ naśyati |
| 3365 | ṇamul after śuṣ, pūr with ūrdhva as agent | ūrdhvaśoṣaṃ śuṣyati |
| 3366 | ṇamul with a standard of comparison as object or agent | ghṛtanidhāyaṃ nihitaṃ jalam |
| 3367 | in SK 3355–3366 the verb after is the same root | nimūlakāṣaṃ kaṣati |
| 3368 | earlier time connected again; ṇamul after upa-daṃś with a third-case upapada | mūlakopadaṃśaṃ bhuṅkte |
| 3369 | ṇamul after an injury-root sharing the object, third-case upapada | daṇḍopaghātaṃ gāḥ kālayati |
| 3370 | ṇamul after upa-pīḍ, -rudh, -karṣ with a seventh- or third-case upapada | pārśvopapīḍaṃ śete |
| 3371 | ṇamul with a third- or seventh-case upapada, for closeness | keśagrāhaṃ yudhyante |
| 3372 | the same for a measure | dvyaṅgulotkarṣaṃ khaṇḍikāṃ chinatti |
| 3373 | ṇamul with an apādāna upapada, in haste | śayyotthāyaṃ dhāvati |
| 3374 | ṇamul with a second-case upapada, in haste | yaṣṭigrāhaṃ yudhyante |
| 3375 | before ṇamul, ā optionally for the diphthong of apa-gur | asyapagoram / asyapagāram |
| 3376 | ṇamul with a second-case body-part word, not vital | bhrūvikṣepaṃ kathayati |
| 3377 | the same for a body part afflicted all over (vital too) | uraḥpratipeṣaṃ yudhyante |
| 3378 | ṇamul after viś, pat, pad, skand for pervasion or repetition | gehānupraveśam āste |
| 3379 | ṇamul after as, tṛṣ with second-case time-words, for an interval | dvyahātyāsaṃ gāḥ pāyayati |
| 3380 | ṇamul after ā-diś, grah with nāman | nāmagrāham āhvayati |
| 3381 | ktvā and ṇamul after kṛñ with an indeclinable, telling contrary to intent | uccaiḥkṛtya / uccaiḥ kṛtvā / uccaiḥkāram |
| 3382 | the same with tiryak, for completion | tiryakkāraṃ gataḥ |
| 3383 | ktvā and ṇamul after kṛ, bhū with a body-part word in -tas | mukhataḥkṛtya, mukhatobhūya |
| 3384 | the same with a word in nā or a dhā-sense affix, in the cvi sense | nānākṛtya, ekadhābhūya |
| 3385 | ktvā and ṇamul after bhū only, with tūṣṇīm | tūṣṇīṃbhūya, tūṣṇīṃbhāvam |
| 3386 | the same with anvak, for conformity; closing verses and colophons | anvagbhūya āste |