With this chapter the book turns from verbs to the kṛdanta, a word that ends in a kṛt. A kṛt is an affix added to a root that is not a verb-ending (SK 374, chapter 11). The chapter opens a new heading, dhātoḥ "after a dhātu" (SK 2829). It is the second dhātu-adhikāra, a heading that runs to the end of the third book (adhyāya) of the Aṣṭādhyāyī. Every affix taught under it comes after a root. The first kṛt affixes taught are the *kṛtya*s, named by SK 2831: tavyat, tavya, anīyar, yat, kyap and ṇyat, with kelimar added by a vārttika. A kṛtya comes only in the sense of the bare action (bhāva) or of the object (karman), by SK 2833: cetavyaḥ dharmas tvayā "dharma is to be accumulated by you". After this chapter you will be able to build edhitavyam, cayanīyaḥ, prayāṇīyam, deyam, śapyam, ālambhyaḥ, aryaḥ beside āryaḥ, brahmodyam, stutyaḥ, kheyam and bhṛtyāḥ. Later in the chapter come kāryam beside kṛtyam, mārgyaḥ, pākyam beside avaśyapācyam, lāvyam, ānāyyaḥ and the agent bhavyaḥ. You will know which of the competing affixes yat, kyap and ṇyat a root takes, and when tavyat and anīyar still come beside them. Bhaṭṭoji also teaches, at the point where the kṛtyas first need them, rules from other parts of the Aṣṭādhyāyī. They give the ṇ of prayāṇīyam (SK 2835–2840), the wider range ("variously") of SK 2841, the num of ālambhyaḥ (SK 2845–2846), the ī of deyam (SK 2843), the t of stutyaḥ (SK 2858) and the velar g of mārgyaḥ (SK 2863). The chapter has sixty-six sūtras, SK 2829 to SK 2894.
edh + tavyat → edhitavyam (tvayā); ci + tavyat → cetavyaḥ dharmas tvayāFive rules set up the frame for every kṛt affix. SK 2829 restates the heading "after a dhātu". SK 2830 is a paribhāṣā: an exception that differs in form from the general affix blocks it only optionally. SK 2831 gives the name kṛtya to the affixes of this chapter. SK 2832 says that a kṛt denotes the agent. SK 2833 at once restricts the kṛtyas to the bare action and the object. The running example has two roots. edh "to grow" takes no object, so edhitavyam (tvayā) names the bare action: "growing is to be done by you". ci "to gather" takes an object, so cetavyaḥ dharmas tvayā agrees with the object: "dharma is to be accumulated by you".
Rule. This is a heading (adhikāra): "after a dhātu". It runs to the end of the third book (adhyāya) of the Aṣṭādhyāyī. Every affix taught under it is added after a root. Bhaṭṭoji then cites two rules, SK 781 (tatropapadaṃ saptamīstham) and SK 374 (kṛd atiṅ). They belong to this heading. One of his commentators says they "have been explained before".
Example.
edh + tavyat → the affix stands after the dhātu edh → edhitavyam.
ci + tavyat → the affix stands after the dhātu ci → cetavyaḥ.
The steps inside these forms are taught at SK 2834.
Applies when / blocked by. The word dhātoḥ already ran from an earlier heading, dhātor ekāco halādeḥ… (SK 2629). Why say it again? One of Bhaṭṭoji's commentators answers. The kṛt and upapada names must hold only inside this new heading. Otherwise the sya of kariṣyati would count as a kṛt and take a case-ending. And every word in the seventh case would count as an upapada. The same commentator adds that the name ārdhadhātuka goes only to affixes taught while "after a dhātu" is in force.
Watch out. This is the second "after a dhātu" heading, not the first. It reaches to the end of the third book. One of Bhaṭṭoji's commentators traces that limit to the Bhāṣya.
Rule. This is a paribhāṣā. Inside this dhātu heading an exception affix (apavāda) usually sets aside the general affix (utsarga). But if the exception is not of the same form (asarūpa), it blocks the general affix only optionally. The one exclusion is the strī-adhikāra, the heading that begins with striyāṃ ktin (SK 3272) in a later chapter. An exception taught there always blocks.
Example. One of Bhaṭṭoji's commentators gives these.
bhū ends in a vowel, so yat comes by SK 2842 → bhavyam.
yat and tavyat differ in form, so tavyat still comes → bhavitavyam.
kṛ takes ṇyat by SK 2872 → kāryam.
tavyat and anīyar still come → kartavyam, karaṇīyam.
Applies when / blocked by. Bhaṭṭoji glosses astriyām as "except one stated in the strī-adhikāra". Both commentators derive this from the svarita of strī, and reject other readings. Why the rule is a paribhāṣā and not a heading, they explain differently; scholars discuss this.
Watch out. "Same form" is judged without the markers. So yat and ṇyat count as the same form (sarūpa), and this rule gives no option between them (SK 2844).
Rule. This is a heading and a name. The affixes taught from here up to ṇvul are called kṛtya. Both of Bhaṭṭoji's commentators say ṇvul means ṇvultṛcau (SK 2895). That rule opens the next chapter. One says the Bhāṣya makes this clear. The other says it is the nearest rule naming ṇvul.
Example.
ci + tavyat → cetavyaḥ: tavyat is taught before SK 2895, so it is a kṛtya.
It also stands under "after a dhātu" and is no verb-ending, so it is a kṛt.
One of Bhaṭṭoji's commentators says tavyat, tavya and anīyar are both kṛt and kṛtya.
Applies when / blocked by. The name is used elsewhere. One of Bhaṭṭoji's commentators lists SK 2817, SK 2822 and SK 2823, all in chapter 65.
Rule. A kṛt affix comes in the sense of the agent (kartṛ). Bhaṭṭoji cites it as what would otherwise hold: "a kṛt affix should be in the agent". One of his commentators explains: tavyat and the rest are kṛts, so the agent sense would fall to them. The next rule prevents this. The same commentator calls this rule a statement of meaning (arthanirdeśa).
Example.
ci + tavyat → a kṛt → by this rule alone it would mean the agent
→ cetavya would mean "one who gathers".
SK 2833 prevents this.
Watch out. This rule stays in force for the kṛt affixes taught later. The kṛtyas are taken out of it by SK 2833.
Rule. The kṛtya affixes, the affix kta, and the affixes with the sense of khal come only in "those two" senses. Those two are bhāva (the bare action) and karman (the object), named in SK 2152. Bhaṭṭoji: "these come only in bhāva and karman". kta and the khal-sense affixes are taught in later chapters.
Example.
edh takes no object, so the kṛtya is in bhāva:
edh + tavyat → edhitavyam tvayā "growing is to be done by you".
One of Bhaṭṭoji's commentators glosses it as "the act of growing, of which you are the agent".
ci takes an object, so the kṛtya is in karman:
ci + tavyat → cetavyaḥ dharmas tvayā "dharma is to be accumulated by you".
Here cetavyaḥ agrees with dharmaḥ, the object. The agent tvayā "by you" is in the third case.
Applies when / blocked by. One of Bhaṭṭoji's commentators says the word eva "only" shuts out the agent. The other draws a further point from eva. The more specific name kṛtya does not cancel the name kṛt; the affixes are still kṛts. As one commentator puts it at SK 2854: after a root that takes an object, a kṛtya is in karman; after one that takes none, in bhāva.
Watch out. Where a kṛtya word does mean the agent, a special rule gives that sense. In this chapter that happens in vāstavyaḥ (a vārttika at SK 2834), in nipātanas such as ajaryam (SK 2853) and sūryaḥ (SK 2865), and in the words of SK 2894.
ci + tavyat → ce-tavya → cetavyaḥ; ci + anīyar → ce-anīya → cayanīyaḥ dharmas tvayāSK 2834 teaches the first three kṛtyas: tavyat, tavya and anīyar. Two vārttikas add the agent noun vāstavyaḥ and the affix kelimar. SK 2841 then lets the kṛtyas, and the affix lyuṭ, range "variously" beyond their taught senses. In the SK it stands after the ṇ rules of section C. It is taken here, next to the affixes it widens. The running example is ci "to gather" with two of the affixes: cetavyaḥ and cayanīyaḥ, both "to be accumulated".
Rule. The affixes tavyat, tavya and anīyar come after a dhātu. "After a dhātu" is carried forward from SK 2829. Bhaṭṭoji: "the t and the r are for the accent". Both of his commentators explain. The t of tavyat makes the accent svarita by titsvaritam (SK 3729). The r of anīyar puts the udātta on the next-to-last syllable. One commentator adds that plain tavya has the udātta on its first syllable.
Example.
ci + tavyat → the marker t goes → ci + tavya
→ tavya is ārdhadhātuka; the final i takes guṇa (SK 2168) → ce + tavya
→ it agrees with dharmaḥ → cetavyaḥ.
ci + anīyar → the marker r goes → ci + anīya
→ guṇa (SK 2168) → ce + anīya
→ e becomes ay before a vowel (SK 61) → cayanīyaḥ.
Bhaṭṭoji: cetavyaś cayanīyo vā dharmas tvayā.
edh + tavyat → edh + tavya
→ edh is seṭ, so the ārdhadhātuka tavya takes iṭ (SK 2184) → edh + i + tavya
→ edhitavyam. With anīyar: edhanīyam.
Bhaṭṭoji: edhitavyam, edhanīyaṃ tvayā.
Applies when / blocked by. In bhāva the word is singular "by the general rule" and neuter, says Bhaṭṭoji. One of his commentators traces both to dicta of the Bhāṣya. A bare action has no number of its own, so the singular is used.
A vārttika, vaser tavyat kartari ṇic ca: after vas "to dwell", tavyat comes in the agent sense, and it counts as marked with ṇ.
vas + tavyat → treated as ṇ-marked → the penultimate a takes vṛddhi (SK 2282) → vāstavyaḥ "one who dwells".
A second vārttika, kelimara upasaṃkhyānam: the affix kelimar is to be added to the list.
pac + kelimar → pacelimā māṣāḥ "beans fit to be cooked", that is, paktavyāḥ.
bhid + kelimar → bhidelimāḥ saralāḥ "sarala trees fit to be split", that is, bhettavyāḥ.
Bhaṭṭoji: the affix is in karman. One of his commentators says its k and r are markers; the k is why bhid takes no guṇa.
Watch out. The Vṛttikāra held that kelimar also comes in the karmakartṛ, an object that acts as its own agent. Bhaṭṭoji rejects this as contrary to the Bhāṣya. One of his commentators says why: the Bhāṣya explains bhidelimāḥ only as bhettavyāḥ "to be split".
Rule. The kṛtya affixes and the affix lyuṭ occur "variously" (bahulam). One of Bhaṭṭoji's commentators says what this means. They may come after other bases and in other senses than those they are taught for. lyuṭ is a kṛt affix taught in a later chapter.
Example. Bhaṭṭoji gives two.
snānty anena "they bathe with it" → snā + anīyar in the sense of the instrument (karaṇa) → snānīyaṃ cūrṇam "bathing powder".
dīyate 'smai "it is given to him" → dā + anīyar in the sense of the recipient (saṃpradāna) → dānīyo vipraḥ "a vipra to whom a gift is to be given".
SK 2833 allows only bhāva and karman. This rule lets anīyar reach these two other senses.
Applies when / blocked by. One of Bhaṭṭoji's commentators adds that the Bhāṣya says the same of every kṛt: "kṛto bahulam". Its example is pādahārakaḥ.
Watch out. In SK order this rule stands after SK 2840. Later rules lean on it: SK 2858 forms ājyam with kyap in the instrument sense "by bahulaka", that is, by this "variously".
pra + yā + anīyar → prayānīya → prayāṇīyamThe n of anīyar often follows an upasarga that contains r, as in pra + yā + anīyar. SK 235 cannot turn it into ṇ, because the upasarga and the kṛt are not one pada. Six rules from the eighth book handle it. SK 2835 gives ṇ to the n of a kṛt after a vowel. SK 2836–2837 make it optional after certain roots. SK 2838 restricts it for roots with num. SK 2839 gives an optional ṇ to the root's own n in three roots. SK 2840 lists roots after which there is none. The running example: in prayānīya the r of pra is the cause, and the n of anīya follows the vowel ā. So the n becomes ṇ: prayāṇīyam.
Rule. An n that stands in a kṛt and comes after a vowel (ac) becomes ṇ, when the cause of the change (nimitta, an r or ṣ) stands in an upasarga. Bhaṭṭoji: "in place of the n which comes after a cause standing in an upasarga, comes after an ac, and stands in a kṛt, ṇ is substituted".
Three things are carried forward. "n becomes ṇ" comes from SK 235. "After an upasarga" comes from SK 859. Bhaṭṭoji's gloss reads it as "standing in an upasarga", and both commentators explain it so. One commentator also carries forward SK 197, so certain sounds may stand between the cause and the n.
Example.
pra + yā + anīyar → pra + yā + anīya
→ the cause: the r of pra, in the upasarga
→ between it and the n: a, y, ā, all allowed by SK 197
→ the n stands in the kṛt anīya and follows the vowel ā
→ this rule: ṇ → prayāṇīyam.
Applies when / blocked by. Why "after an ac"? Bhaṭṭoji's counter-example is pramagnaḥ. One of his commentators derives it from masj with kta. There the n of the kṛt follows g, a consonant, so it stays n.
A vārttika, nirviṇṇasyopasaṃkhyānam: the form nirviṇṇa is to be added.
nis + vid + kta → SK 3016, taught later, makes the t and the d before it both n → nirvinna
→ the second n follows n, not a vowel, so this rule cannot reach it
→ the vārttika: the second n becomes ṇ → nirvinṇa
→ ṣṭutva (SK 113): the first n becomes ṇ too → nirviṇṇaḥ.
Bhaṭṭoji: "the latter becomes ṇ; the former by ṣṭutva".
Watch out. "After an ac" describes the n, not the kṛt. Both of Bhaṭṭoji's commentators construe it so. One of them adds that otherwise forms like prayāpaṇam would get no ṇ.
Rule. The n of a kṛt optionally becomes ṇ when that kṛt is taught after a root ending in ṇi. The cause still stands in an upasarga. kṛtyacaḥ (SK 2835) and "n becomes ṇ" (SK 235) are carried forward. The word ṇeḥ says after what the kṛt is taught (vihitaviśeṣaṇa); it does not demand that the kṛt stand right after ṇi.
Example.
yā + ṇic → puk (SK 2570) → yāpi
→ pra + yāpi + anīyar → the ṇi is lost (SK 2313) → pra + yāp + anīya
→ this rule, optionally → prayāpaṇīyam / prayāpanīyam.
Applies when / blocked by. Why "taught after"? So that the option holds even when yak stands between. yak is the affix put between root and ending when the sense is bhāva or karman. Bhaṭṭoji: prayāpyamāṇaṃ paśya "look at what is being made to go". Both of his commentators trace it: yāpi in karman, śānac for laṭ (the present-tense lakāra), yak, the augment m (muk), loss of ṇi. The kṛt śānac does not follow ṇi directly, yet it was taught after a ṇi-final root. Whether the Bhāṣya's other route to this result works, scholars discuss this.
Watch out. For ṇatva, dur does not count as an upasarga. Bhaṭṭoji refers to a vārttika given at SK 2231: duryānam, duryāpanam, with no ṇ.
Rule. After a root that begins with a consonant (hal) and has an ic as its penultimate, the n of a kṛt after a vowel optionally becomes ṇ. An ic is any vowel other than a and ā. A root with such a penultimate is called ijupadha. "Optionally" continues from SK 2836.
Example.
pra + kup + anīyar → kup begins with k and has u as its penultimate
→ this rule, optionally → prakopaṇīyam / prakopanīyam.
Applies when / blocked by. Bhaṭṭoji gives two counter-examples. In them, both commentators say, SK 2835 makes ṇ obligatory. Not consonant-initial: prohaṇīyam, from ūh. Not ijupadha: pravapaṇīyam, from vap.
Watch out. halaḥ here means "beginning with a consonant", not "ending in one". Both of Bhaṭṭoji's commentators give the reason. A root with an ic as penultimate always ends in a consonant, so "ending in a consonant" would add nothing.
Rule. This is a niyama. Take a root that contains num. After it, the n of a kṛt becomes ṇ only if the root begins with an ic and ends in a consonant. Bhaṭṭoji: "if it has num, then only for the n standing in that kṛt which is taught after a hal-final root beginning with an ic". Both of his commentators say halaḥ is carried forward and here means "ending in a consonant". "Optionally" has stopped.
Example.
pra + ikh + anīyar → ikh is marked with i, so it takes num → pra + iṅkh + anīya
→ it begins with i and ends in a consonant → preṅkhaṇīyam.
Bhaṭṭoji's counter-example: mag "to creep", marked with i, which does not begin with an ic → pramaṅganīyam, no ṇ.
Applies when / blocked by. Why is this a restriction? SK 2835 alone would already give preṅkhaṇīyam. Both commentators say so. So the rule's work is to deny ṇ to the other num-roots, like mag.
Bhaṭṭoji adds that num here stands for (upalakṣaṇa) any anusvāra. The same holds in SK 197.
pra + inv + ana (ana is lyuṭ) → the num stays n before v, so it is no anusvāra
→ it is therefore not a permitted intervener under SK 197 either → no ṇ: prenvanam.
pra + umbh + ana (lyuṭ again) → the anusvāra is the root's own, not from num → ṇ: prombhaṇam.
Watch out. In SK 2837 halaḥ meant "consonant-initial". Here the same word means "consonant-final", because a root beginning with an ic cannot begin with a consonant.
Rule. The n of the roots niṃs, nikṣ and nind optionally becomes ṇ when a kṛt follows. Here it is the root's own n that changes. Both of Bhaṭṭoji's commentators say "when a kṛt follows" is carried forward from SK 2835. One adds that "after a vowel" is not.
Example.
pra + niṃs + tavya → iṭ → pra + niṃs + i + tavya
→ the r of pra, then the root's n; a kṛt follows
→ this rule, optionally → praṇiṃsitavyam / praniṃsitavyam.
Rule. After bhā, bhū, pū, kam, gam, pyāy and vep, the n of a kṛt does not become ṇ.
Example.
pra + bhā + anīyar → SK 2835 would give ṇ → this rule forbids it → prabhānīyam.
pra + bhū + anīyar → guṇa, av → prabhavanīyam.
Applies when / blocked by. A vārttika, pūña eveha grahaṇam iṣyate: only pūñ is meant here. So pūṅ takes ṇ: prapavaṇīyaḥ somaḥ "soma to be purified".
A second vārttika, ṇyantabhādīnām upasaṃkhyānam: the same holds for these roots with ṇi added: prabhāpanīyam.
Bhaṭṭoji adds a point on khyā. He recalls a statement made at SK 2437: "for the ś of khśāñ, y optionally" (a vārttika). It is to be understood as standing after the ṇatva section. So the change of ś to y is asiddha for ṇatva. The ś still counts as standing between the cause and the n, and ś is not among the sounds SK 197 allows. So there is no ṇ: prakhyānīyam. One of his commentators says the Bhāṣya on SK 2436 makes this clear.
Watch out. prakhyānīyam is not one of the listed roots. It escapes ṇ by the asiddha y, not by this rule.
dā + yat → dī-ya → deyam; śap + yat → śapyam (not ṇyat)The fourth kṛtya is yat. SK 2842 gives it after a root ending in a vowel: ceyam. SK 2843 turns a root-final ā into ī before it: deyam. The general affix after a root ending in a consonant is ṇyat, taught at SK 2872. SK 2844 and SK 2847–2848 replace it with yat after certain consonant-final roots: śapyam, śakyam, gadyam. SK 2845–2846 give labh a num, which sends it back to ṇyat: ālambhyaḥ. The running example is dā "to give": dā + yat → dī-ya → deyam "to be given". And śap "to curse": śapyam, not ṇyat.
Rule. After a root that ends in a vowel (ac), the affix yat comes. "After a dhātu" is carried forward from SK 2829. Bhaṭṭoji glosses it "after a dhātu ending in an ac". One commentator explains that ac qualifies the root.
Example.
ci + yat → the marker t goes → ci + ya
→ guṇa before the ārdhadhātuka (SK 2168) → ce + ya → ceyam.
ji + yat → jeyam.
Applies when / blocked by. Bhaṭṭoji says the word ac could be left out. Both of his commentators explain. ṇyat is taught specifically after roots ending in ṛ or a consonant (SK 2872). So yat is left with the vowel-final roots anyway. One commentator adds that SK 2830 gives no option here, since yat and ṇyat count as the same form.
Separating it as its own rule (yogavibhāga) was not needed either, Bhaṭṭoji says. yat could have been read in SK 2834 along with tavyat and the rest.
Watch out. A root ending in ṛ, such as kṛ, takes ṇyat by SK 2872: kāryam. Root-final ā takes yat, and before that yat the ā becomes ī (SK 2843).
Rule. Before yat, ī replaces a root-final ā. "ā" is carried forward from SK 2372 (āto lopa iṭi ca). Bhaṭṭoji: "when yat follows, ī replaces ā; then guṇa".
Example.
dā + yat (SK 2842) → this rule → dī + ya
→ guṇa (SK 2168) → de + ya → deyam "to be given".
glai + yat → ā for the final ai (SK 2370) → glā + ya
→ this rule → glī + ya → guṇa → gleyam.
Applies when / blocked by. A vārttika, takiśasicatiyatijanibhyo yad vācyaḥ: yat also comes after tak, śas, cat, yat and jan, though they end in consonants. takyam, śasyam, catyam, yatyam, janyam.
For jan, Bhaṭṭoji says, yat is only for the accent. ṇyat would give the same form janyam. Its vṛddhi is barred by SK 2512 (janivadhyoś ca). One of his commentators names the accent: yat puts the udātta on the first syllable, by SK 3701 (yato 'nāvaḥ).
A second vārttika, hano vā yad vadhaś ca vaktavyaḥ: after han, yat comes optionally, and vadha replaces han.
han + yat → vadha for han → vadhyaḥ.
In the other alternative, ṇyat (SK 2872) → ghātyaḥ.
One of Bhaṭṭoji's commentators traces ghātyaḥ. t replaces the n (SK 2574). gh replaces the h (SK 358). The penultimate a takes vṛddhi.
Watch out. vadha comes only together with yat. Both of Bhaṭṭoji's commentators make this point: with ṇyat the root stays han.
Rule. After a root that ends in a labial (pu: p, ph, b, bh, m) and has short a as its penultimate (adupadha), yat comes. Bhaṭṭoji: "it is an exception (apavāda) to ṇyat". Without this rule these consonant-final roots would take ṇyat by SK 2872.
Example.
śap + yat → śap ends in p and has a before it → śapyam "to be cursed".
labh + yat → labhyam "to be obtained".
Applies when / blocked by. One of Bhaṭṭoji's commentators tests each condition. Not labial-final: pākyam. Not a as penultimate: kopyam. Long ā as penultimate: āpyam, since the a of the rule is tapara.
Bhaṭṭoji adds the principle nānubandhakṛtam asārūpyam: "a difference made by markers (anubandha) is no difference of form". yat and ṇyat differ only in the marker ṇ. So SK 2830 does not let ṇyat come as an option. One of his commentators says this principle is set out in the Bhāṣya on SK 2830. But tavyat and the rest do differ in form, and they still come: Bhaṭṭoji says so.
Watch out. So beside śapyam there is no ṇyat form. Only tavyat and the rest stand beside it.
Rule. After the upasarga āṅ (ā), the root labh takes the augment num. The condition is that a y-initial affix is "in view" (vivakṣita). Two words are carried forward. num comes from SK 2262 (idito num dhātoḥ). labh comes from SK 2582 (labheś ca). One of Bhaṭṭoji's commentators calls the seventh case of yi a "seventh of domain" (viṣayasaptamī), as the Bhāṣya makes clear.
Example.
ā + labh, a y-initial kṛtya in view
→ this rule, before any affix is added: num → ā + lambh
→ lambh no longer has a as its penultimate, so SK 2844 cannot give yat
→ ṇyat (SK 2872) → ālambhyo gauḥ "a cow to be seized for sacrifice".
Bhaṭṭoji: "once num has been added, only ṇyat comes".
Applies when / blocked by. Why "in view" and not "when it follows"? One of Bhaṭṭoji's commentators answers. If num waited for yat to follow, yat would already have come by SK 2844. So num is made first, while the affix is only in view.
Watch out. With yat the written form would be the same: ālambhya. Both of Bhaṭṭoji's commentators say the difference is in the accent. ṇyat gives the svarita of its t; yat would give the udātta on the first syllable.
Rule. After the upasarga upa, labh also takes num when a y-initial affix is in view, in the sense of praise. num, labh and "y-initial in view" are carried forward from SK 2845.
Example.
upa + labh, praise meant, y-initial affix in view
→ this rule: num → upa + lambh → ṇyat → upalambhyaḥ sādhuḥ "a good man, worth reaching".
One of Bhaṭṭoji's commentators says the praise is conveyed by the word sādhu.
Applies when / blocked by. Why "in praise"? Bhaṭṭoji's counter-example: upalabdhuṃ śakya upalabhyaḥ "upalabhya: what can be obtained". With no praise there is no num, and yat comes by SK 2844.
Rule. yat also comes after śak and sah. One of Bhaṭṭoji's commentators says the sixth case in the sūtra stands for the fifth, "after these two". The rule is an exception to ṇyat.
Example.
śak + yat → śakyam.
sah + yat → sahyam.
Neither root ends in a labial, so SK 2844 would not reach them. Without this rule they would take ṇyat.
Rule. yat also comes after gad, mad, car and yam, when they have no upasarga. It is an exception to ṇyat.
Example.
gad + yat → gadyam. mad + yat → madyam. car + yat → caryam.
With an upasarga, ṇyat comes: one of Bhaṭṭoji's commentators gives pragādyam.
Applies when / blocked by. A vārttika, carer āṅi cāgurau: after car with āṅ too, yat comes, unless a teacher (guru) is meant. ācaryo deśaḥ "a region to be gone to". But ācāryo guruḥ "a teacher", with ṇyat.
Why name yam? It ends in a labial and has a as its penultimate, so SK 2844 already gives it yat. Bhaṭṭoji: it is named for a restriction. With an upasarga, yam gets no yat: prayāmyam, with ṇyat.
Yet after ni, Bhaṭṭoji says, yat does come. He points to wording found in a vārttika: tena na tatra bhaved viniyamyam. This explains usages such as aniyamyasya nāyuktiḥ and tvayā niyamyā nanu divyacakṣuṣā "surely they are to be restrained by you, who have the divine eye". His second answer: niyamya means "proper for restraint" (niyame sādhuḥ). Both commentators explain this. niyama is formed first with the affix ap (SK 3240), and then the taddhita yat of SK 1650 (tatra sādhuḥ) is added. Whether the first answer holds, scholars discuss this.
ṛ + yat → ar-ya → aryaḥ "master, vaiśya"; against ṛ + ṇyat → ār-ya → āryaḥFive sūtras state finished words with yat, each in a fixed sense. Such a ready-made form is a nipātana: the sūtra quotes the word, and whatever is irregular in it is taken as taught. Most of these words would otherwise take ṇyat, or another affix. Outside the stated sense, the regular affix comes back. The running example is the root ṛ "to go". With yat it gives aryaḥ "a master" or "a vaiśya" (SK 2851). With the regular ṇyat it gives āryaḥ "one to be approached".
Rule. Three words are formed by nipātana with yat. avadya means "blameworthy". paṇya means "to be traded". varyā means "to be chosen", where there is no restriction (anirodha). The senses go with the words in order.
Example.
avadya: nañ + vad + yat → avadyaṃ pāpam "sin, the blameworthy".
nañ is the negative particle; at the head of a word it appears as a-.
paṇya: paṇ + yat → paṇyā gauḥ "a cow for sale", meaning "to be traded".
varyā: vṛ + yat → śatena varyā kanyā "a girl to be chosen by a hundred".
Bhaṭṭoji says the root is vṛṅ.
Applies when / blocked by. For avadya the nipātana does two jobs, Bhaṭṭoji says. vad with nañ as upapada would take both yat and kyap, by SK 2854. The nipātana allows only yat, and only in the sense of blame. Why "blameworthy"? anudyaṃ gurunāma "the guru's name, not to be uttered". That name is not blameworthy but unfit to be spoken. So kyap comes, with saṃprasāraṇa (SK 2409), as one of Bhaṭṭoji's commentators explains. Bhaṭṭoji cites a smṛti verse for the sense. One who seeks the good should not utter his own name, his guru's, a miser's, or those of his eldest child and his wife.
For paṇya, in the other sense the root takes ṇyat: pāṇyam "worthy of praise".
For varyā, outside the sense of no restriction there is no yat. Bhaṭṭoji's contrast is vṛtyā anyā "another girl is vṛtyā", with kyap (SK 2857) and tuk, from vṛñ "to choose". One of his commentators gives this derivation. vṛṅ itself takes the regular ṇyat (SK 2858).
Watch out. varyā is feminine only. Both of Bhaṭṭoji's commentators say the masculine takes ṇyat: vāryāḥ ṛtvijaḥ.
Rule. vahya is formed by nipātana with yat after vah, when it denotes the instrument (karaṇa). One of Bhaṭṭoji's commentators calls this an exception to ṇyat.
Example.
vahanty anena "they carry with it" → vah + yat in the instrument sense → vahyaṃ śakaṭam "a cart".
Applies when / blocked by. Why "instrument"? Bhaṭṭoji's counter-example is vāhyam, which he glosses voḍhavyam "to be carried". It has ṇyat, so the penultimate a takes vṛddhi.
Rule. arya is formed by nipātana in the senses "master" and "vaiśya". Bhaṭṭoji: after ṛ "to go", yat. It is an exception to ṇyat.
Example.
ṛ + yat → guṇa of ṛ, with r after it (SK 2168, SK 70) → ar + ya
→ aryaḥ "a master" or "a vaiśya".
One of Bhaṭṭoji's commentators gives this step.
Applies when / blocked by. Why "in these two senses"? Bhaṭṭoji's counter-example: āryo brāhmaṇaḥ, meaning "one to be approached". There ṛ ends in ṛ, so it takes ṇyat (SK 2872).
ṛ + ṇyat → vṛddhi (SK 254), with r (SK 70) → ār + ya → āryaḥ.
Watch out. aryaḥ and āryaḥ are two words from one root. Only the affix and the sense differ: yat gives guṇa, ṇyat gives vṛddhi.
Rule. upasaryā is formed by nipātana with yat after sṛ with upa. It is said of a female whose time for conception (prajana) has come. Bhaṭṭoji: "if she is one whose time for conceiving has come".
Example.
upa + sṛ + yat → guṇa, with r → upa + sar + ya → upasaryā gauḥ.
Bhaṭṭoji glosses it: "fit to be approached by a bull for impregnation".
Applies when / blocked by. Why "whose time has come for conception"? Bhaṭṭoji's counter-example: upasāryā kāśī, meaning "Kāśī, to be reached". One of his commentators says that form has ṇyat, in karman.
Rule. ajarya is formed by nipātana: yat after jṝ "to wear out", with nañ, in the agent sense. The condition is that saṃgata "association" is the noun it qualifies. Bhaṭṭoji: "after jīryati preceded by nañ, yat in the agent, if saṃgata is the qualificand".
Example.
na jīryati "it does not wear out" → nañ + jṝ + yat, agent sense → ajaryam.
Bhaṭṭoji quotes Bhaṭṭi: tena saṃgatam āryeṇa rāmājaryaṃ kuru drutam "Rāma, quickly make an unaging association with that noble one". Here ajaryam qualifies saṃgatam.
In a line he cites from another poem, mṛgair ajaryaṃ jarasopadiṣṭam…, the word saṃgatam is not written. Bhaṭṭoji says it is to be supplied. One of his commentators names the poet as Kālidāsa.
Applies when / blocked by. Why "saṃgata"? Bhaṭṭoji's counter-example: ajaritā kambalaḥ "a blanket that does not wear out". One of his commentators says it has the agent affix tṛc.
In bhāva, ṇyat comes even when saṃgata is the agent: ajāryaṃ saṃgatena "there is no wearing out on the part of the association".
Watch out. The agent sense here goes against SK 2833, which allows a kṛtya only in bhāva and karman. Both of Bhaṭṭoji's commentators say the agent sense comes from the nipātana itself.
stu + kyap → stu-t-ya → stutyaḥ; brahman + vad + kyap → brahma-udya → brahmodyamThe fifth kṛtya is kyap. Its two markers do two jobs. The k blocks guṇa and vṛddhi (SK 2217): mṛjyaḥ, duhyam. The p brings the augment tuk after a short vowel (SK 2858): stutyaḥ. SK 2854–2856 give kyap after vad, bhū and han with a noun as upapada. SK 2857 gives it after six roots, among them stu, and SK 2858 adds the t. SK 2859–2862 give it after roots with ṛ as penultimate, and after khan, bhṛ and mṛj. The running example has two words. stu + kyap → stu-t-ya → stutyaḥ "to be praised". And brahman + vad + kyap → brahma-udya → brahmodyam "the speaking of the Veda".
Rule. After vad "to speak", kyap comes, and by the word ca "and", yat too. Two conditions hold. A word ending in a case-ending (sup) is the upapada, and there is no upasarga. Bhaṭṭoji takes the sense from the next sūtra: "in bhāva" is drawn back into this rule. Drawing a word back from a later sūtra is called apakarṣa.
Example.
brahman "the Veda" as upapada → brahman + vad + kyap
→ kyap is k-marked (kit), so the v of vad takes saṃprasāraṇa: u for the v (SK 2409)
→ the following a merges into it (SK 330) → udya
→ brahman appears as brahma before it → brahma + udya
→ a and u merge into o (SK 69) → brahmodyam.
With yat: brahmavadyam.
Bhaṭṭoji: "brahman is the Veda; the meaning is the speaking of it".
Applies when / blocked by. With an upasarga only ṇyat comes: anuvādyam, apavādyam. One of Bhaṭṭoji's commentators traces the condition "no upasarga" to the Bhāṣya.
Bhaṭṭoji records a second view. "Some" (eke) say the two affixes are in karman, not bhāva. Both of his commentators give the grounds of the two views. For the first: vad takes an object, so its kṛtya could not otherwise be in bhāva. And bhū takes none, so "bhāva" in SK 2855 would be idle unless it served here. For the second: the Bhāṣya calls the word "bhāva" in SK 2855 "for the following rules". Which view is better, scholars discuss this.
Watch out. vad takes an object, so by SK 2833 its kṛtya would normally be in karman. The bhāva sense of brahmodyam therefore needs the drawn-back word.
Rule. After bhū "to be", in bhāva, kyap comes. The conditions of SK 2854 continue: a word with a case-ending as upapada, and no upasarga.
Example.
brahmaṇo bhāvaḥ "the state of brahman" → brahman + bhū + kyap
→ kyap is k-marked, so no guṇa → brahmabhūyam.
One of Bhaṭṭoji's commentators says the neuter is from usage.
Applies when / blocked by. Bhaṭṭoji gives two counter-examples: bhavyam with no upapada, prabhavyam with an upasarga. One of his commentators says they have yat (SK 2842). He derives bhavyam with guṇa and av (SK 63). He glosses prabhavyam as "majesty".
Watch out. "Bhāva" here is part of this rule. Whether it also runs back into SK 2854 is the question discussed there.
Rule. After han "to strike", in bhāva, kyap comes, and t replaces the final of the root. The conditions continue: a word with a case-ending as upapada, and no upasarga. Bhaṭṭoji: "t is the substitute for the final".
Example.
brahmaṇo hananam "the killing of a brahman" → brahman + han + kyap
→ t for the final n → brahma + hat + ya
→ the feminine, which Bhaṭṭoji says is from usage → brahmahatyā.
Applies when / blocked by. Each condition has a counter-example from Bhaṭṭoji's commentators. Not bhāva: ghātyo vṛṣalaḥ, from one commentator. No upapada: ghātaḥ, from the other. With an upasarga: praghātaḥ, from both.
Watch out. The t replaces only the final sound, the n. The ha of han stays.
Rule. kyap comes after i "to go" (iṇ), stu, śās, vṛ, dṛ and juṣ. Bhaṭṭoji: "after these, kyap". Both of his commentators say the three conditions of SK 2854–2856 have stopped. No upapada is needed, an upasarga is allowed, and the sense is not limited to bhāva. One commentator says eti means the root iṇ only.
Example. Bhaṭṭoji gives the examples under the next sūtra, SK 2858.
stu + kyap → no guṇa (the k) → stu + ya → then SK 2858.
Applies when / blocked by. i and stu end in vowels, so SK 2842 would give them yat. vṛ and dṛ end in ṛ, and śās and juṣ in consonants, so SK 2872 would give these four ṇyat. This rule replaces both with kyap.
Watch out. kyap is also carried forward from SK 2854. Why name it again here? Bhaṭṭoji answers under SK 2858.
Rule. Before a kṛt marked with p (pit), a short vowel takes the augment tuk, that is, t. kyap is pit and a kṛt, so every short root-final vowel before it gets t.
Example. These are the examples of SK 2857.
i + kyap → the short i takes tuk → i + t + ya → ityaḥ "to be gone to".
stu + kyap → the short u takes tuk → stu + t + ya → stutyaḥ "to be praised".
śās + kyap → i for ā by SK 2486 → śis + ya
→ its s becomes ṣ (a rule taught elsewhere) → śiṣyaḥ.
vṛ + kyap → vṛtyaḥ. Also ādṛtyaḥ (dṛ with ā) and juṣyaḥ.
Applies when / blocked by. Bhaṭṭoji: the vṛ of SK 2857 is vṛñ, not vṛṅ. One of his commentators calls this a vārttika. From vṛṅ the regular ṇyat comes: vāryāḥ ṛtvijaḥ "the officiating priests are to be engaged".
Why does SK 2857 name kyap again? Bhaṭṭoji: to block even the later ṇyat of SK 2886. That rule gives ṇyat after a u-final root when necessity is meant. In avaśyastutyaḥ "who must be praised" both rules apply. By SK 175 the later rule, SK 2886, would win. The restated kyap prevents that. Both of his commentators explain it so.
Bhaṭṭoji quotes the Kāśikā: kyap comes optionally after śaṃs, duh and guh. śasyam / śaṃsyam, duhyam / dohyam, guhyam / gohyam. For śaṃs, he says, both forms are proved by two sūtras, SK 2009 and SK 3702. For duh and guh "the authority is still to be sought". One of his commentators adds that the option is not seen in the Bhāṣya.
A vārttika, āṅpūrvād añjeḥ saṃjñāyām upasaṃkhyānam: kyap after añj with āṅ, when a name is meant.
ā + añj + kyap, in the instrument sense by bahulaka ("variously", as in SK 2841)
→ the n is lost before the k-marked affix (SK 415) → ā + aj + ya → ājyam "clarified butter".
Watch out. tuk needs a short vowel. In śiṣyaḥ and juṣyaḥ the root ends in a consonant, so no t is added.
Rule. kyap also comes after a root whose penultimate is short ṛ (ṛdupadha), except kḷp and cṛt. One of Bhaṭṭoji's commentators says kyap is carried forward from SK 2857.
Example.
vṛt + kyap → no guṇa (the k) → vṛtyam.
vṛdh + kyap → vṛdhyam.
Applies when / blocked by. kḷp and cṛt are excepted: kalpyam, cartyam. They take ṇyat (SK 2872), with guṇa. One of Bhaṭṭoji's commentators explains why kḷp must be named. Its l is asiddha, and ṛ and ḷ count as savarṇa, so kḷp counts as ṛdupadha.
Why the t after ṛ in the sūtra (taparakaraṇa)? It limits the rule to short ṛ. kṝt, with long ṝ, is not taken: kīrtyam. kṝt belongs to the curādi class, whose roots "are not always ṇic-final". Without ṇic, ṇyat comes; with ṇic, only yat.
Watch out. By this rule mṛj would always take kyap. SK 2862 makes that optional.
Rule. After khan "to dig", kyap comes, by the word ca, and ī replaces the final of the root. One of Bhaṭṭoji's commentators says the ī replaces the final n.
Example.
khan + kyap → ī for the final n → kha + ī + ya
→ a and ī merge into e (SK 69) → kheyam "to be dug".
Applies when / blocked by. Bhaṭṭoji says a short i would have been a sound reading (supaṭha). One of his commentators gives the reason: with short i too, SK 69 gives kheyam. The other reports that the Bhāṣya rejects the long vowel.
Watch out. The ī here is a substitute for the root's final, not an augment.
Rule. After bhṛñ, kyap comes, when the word is not a name (saṃjñā). One of Bhaṭṭoji's commentators supplies kyap.
Example.
bhṛ + kyap → no guṇa → the short ṛ takes tuk (SK 2858) → bhṛ + t + ya → bhṛtyāḥ "workers".
Bhaṭṭoji: the meaning is "those who are to be maintained" (bhartavyāḥ). It is a word for an action, not a name.
Applies when / blocked by. A vārttika, samaś ca bahulam: after bhṛñ with sam, variously. saṃbhṛtyāḥ / saṃbhāryāḥ. Bhaṭṭoji: this option holds only where no name is meant.
Why "not a name"? bhāryā nāma kṣatriyāḥ "the Kṣatriyas called Bhārya", with ṇyat.
Then how is bhāryā "wife" right? Bhaṭṭoji raises the objection. It is a name, so kyap should come by SK 3276 (saṃjñāyāṃ samaja…), a later rule. He answers: bhāryā comes from a different root. It is from ḍubhṛñ of the juhotyādi class, or from bhṝ of the kryādi class. These take ṇyat. kyap comes only from bhṛñ of the bhvādi class. bhṝ ends in long ṝ, so it is not bhṛñ at all. ḍubhṛñ carries the extra marker ḍu. The paribhāṣā tadanubandhakagrahaṇe nātadanubandhakasya keeps it out: a root named with certain markers does not take in a root with other markers.
Watch out. bhṛtya is fixed by usage for servants, so is it not a name? One of Bhaṭṭoji's commentators raises this doubt. Bhaṭṭoji's "a word for an action, not a name" answers it.
Rule. After mṛj, kyap comes optionally. In the other alternative, ṇyat. One of Bhaṭṭoji's commentators says why the rule is needed. mṛj has ṛ as penultimate, so SK 2859 would give kyap always.
Example.
mṛj + kyap → no guṇa (the k) → mṛjyaḥ.
In the other alternative: mṛj + ṇyat → mārgyaḥ.
The g of mārgyaḥ comes from SK 2863, taught in section I of this chapter with the other velar rules.
Watch out. mṛjyaḥ keeps the j. Only before ṇyat does it become g.
sṛ + kyap → sur-ya → sūryaḥ; prati + grah + kyap → pratigṛhyamGroup F taught kyap by general rules. This group gives words that Pāṇini simply lists as formed with kyap. Such a listing is a nipātana: the sūtra hands down a ready-made form, and whatever is irregular in it is accepted because the sūtra says so. A nipātana can fix an unusual sound, an unusual sense, or both. The running example is sūryaḥ "the sun", from sṛ "to move". It shows both kinds of irregularity: u for the root's ṛ, and kyap in the sense of the agent. The group ends with kyap after grah "to take", as in the Vedic pratigṛhyam.
Rule. Seven words are given by nipātana as ending in kyap: rājasūya, sūrya, mṛṣodya, rucya, kupya, kṛṣṭapacya and avyathya. Bhaṭṭoji explains each one.
Example. The running example, sūrya:
sṛ "to move" + kyap, in the sense of the agent (by the nipātana)
→ u replaces the ṛ (by the nipātana); the u takes r after it (SK 70, raparatva) → sur + ya
→ the u is lengthened before a consonant by hali ca (SK 354) → sūrya
→ sūryaḥ "it moves (sarati) in the sky", the sun.
Bhaṭṭoji's commentators both give the step with r and hali ca. Bhaṭṭoji offers a second route as well. The root may be ṣū "to impel", of the tudādi class: the sun impels the world to activity. Then kyap takes the augment ruṭ, an inserted r: sū + r + ya → sūryaḥ.
The other six words:
rājasūyaḥ and rājasūyam, because it is in the ardharcādi list.mṛṣā + vad + kyap → saṃprasāraṇa: u for the v (SK 2409), the a merging into it (SK 330) → mṛṣā + udya
→ ā + u → o (SK 69) → mṛṣodyam "falsely spoken".
It takes the gender of the noun it describes: ucchrāyasaundaryaguṇā mṛṣodyāḥ "merits of height and beauty, falsely spoken of".
rocate "it pleases" → rucyaḥ "pleasing".kupyam is wealth other than gold and silver. Otherwise the form is gopyam "to be guarded".kṛṣṭapacyāḥ. Here kyap is in the sense of the karmakartṛ, an object that acts as its own agent.na vyathate "it does not waver" → avyathyaḥ.Applies when / blocked by. How can rājasūya mean a locus, when a kṛt means the agent (SK 2832)? One commentator answers: by the nipātana. The same commentator says the agent sense of avyathya is also by the nipātana. Both commentators give the same answer for the long ū of rājasūya. Its root ṣuñ "to press" has a short u. The lengthening rule SK 2298 does not apply before a kṛt.
Why "invariably" for mṛṣodya? SK 2854 allowed yat beside kyap after vad with a sup-final upapada (brahmavadyam). One commentator notes that the nipātana shuts yat out here.
Watch out. Where the object is a pure object, kṛṣṭapacya is not used. One says kṛṣṭapākyāḥ, "to be cooked on ploughed land". One of Bhaṭṭoji's commentators gives its steps: ṇyat, vṛddhi of the penultimate a, and k for c by SK 2863 (taught in group I below).
Rule. bhidya and uddhya are given by nipātana when a river (nada) is meant. Bhaṭṭoji: kyap comes after bhid "to break" and ujh "to let go", and dh replaces the jh of ujh. Bhaṭṭoji's commentators both supply the sense: the agent.
Example.
bhid + kyap, agent → bhidyaḥ: "it breaks its bank" (bhinatti kūlam), a river.
ujh + kyap, agent → dh for jh (by the nipātana) → uddhyaḥ: "it lets the water out" (ujjhaty udakam), a river.
Applies when / blocked by. Why "a river"? Otherwise the agent is expressed by another kṛt: bhettā "breaker", ujjhitā "one who abandons". Here again the kṛtya means the agent by the nipātana, as with sūrya and avyathya (SK 2865).
Rule. puṣya and siddhya are given by nipātana when a nakṣatra, a lunar mansion, is meant. Bhaṭṭoji: kyap comes in the sense of the locus (adhikaraṇa).
Example.
puṣ + kyap, locus → puṣyaḥ: "in it aims prosper" (puṣyanty asminn arthāḥ).
sidh + kyap, locus → siddhyaḥ: "in it they succeed" (siddhyanti).
Applies when / blocked by. Why "a lunar mansion"? One of Bhaṭṭoji's commentators answers. Otherwise the locus sense takes lyuṭ: poṣaṇam, sedhanam. The two words name the same mansion. The sūtra still lists both, because each word stands for its own form.
Rule. vipūya, vinīya and jitya are given by nipātana, in the senses muñja grass, kalka and hali, respectively. Bhaṭṭoji: kyap comes after pūṅ "to cleanse", nīñ "to lead" and ji "to conquer".
Example.
vi + pūṅ + kyap → vipūyo muñjaḥ: muñja grass "to be cleaned" for making rope and the like.
vi + nīñ + kyap → vinīyaḥ kalkaḥ: a medicinal herb "ground to a paste". Or else kalka means "sin".
ji + kyap → tuk after the short i (SK 2858) → ji-t-ya → jityo haliḥ: a hali "to be drawn by force". A hali is a thick log for levelling ploughed ground.
Applies when / blocked by. In other senses there is no kyap. Bhaṭṭoji gives vipavyam, vineyam, jeyam. One commentator adds that pū here is only pūṅ of the bhvādi class, because the sūtra pairs it with nī and ji.
Rule. After grah "to take" with prati or api before it, kyap comes. Bhaṭṭoji adds no gloss of his own; he passes straight to a vārttika.
vārttika chandasīti vaktavyam: "it should be stated: in the Veda". So this kyap is Vedic only.
Example.
prati + grah + kyap → the r of grah becomes ṛ (saṃprasāraṇa) → pratigṛhyam.
Likewise apigṛhyam.
In ordinary usage (loka): pratigrāhyam, apigrāhyam.
Applies when / blocked by. One of Bhaṭṭoji's commentators names the ordinary forms' affix: ṇyat. Why does Bhaṭṭoji include a Vedic rule at all? The other commentator answers: to show, with "in ordinary usage, however", what the ordinary language says.
Watch out. Some read "in the Veda" as part of the sūtra itself. Bhaṭṭoji's commentators say it is a vārttika. The author of the Kāśikā, a commentary on the Aṣṭādhyāyī, put it into the sūtra. The Kāśikā's readings are sometimes rejected in this chapter: by Bhaṭṭoji's commentators here and at SK 2885, and by Bhaṭṭoji himself at SK 2882.
Rule. kyap also comes after grah in four senses: pada "a word", asvairin "not independent", bāhyā "outside" and pakṣya "on a side". Bhaṭṭoji's commentators both supply "after grah, kyap".
Example.
avagṛhyam, "a word to be split by avagraha", the pause that separates the members of a compound in recitation. pragṛhyaṃ padam, "a word to be kept apart", one that does not coalesce with the next vowel.gṛhyakāḥ śukāḥ "kept parrots", made dependent by a cage. One commentator says the ka is the affix kan in the sense of pity.grāmagṛhyā senā, "an army outside the village".āryagṛhyaḥ, "taken as one of theirs by the Āryas", one who has taken their side.Applies when / blocked by. bāhyā stands in the feminine in the sūtra. So, Bhaṭṭoji says, kyap does not come in the masculine or neuter for this sense. pakṣya is "that which is on a side" (pakṣe bhavaḥ). It takes the taddhita yat (an affix after a noun, not the kṛtya yat), because pakṣa is in the digādi list.
kṛ + ṇyat → kār-ya → kāryam; against kṛ + kyap → kṛ-t-ya → kṛtyamThis group reaches the general rule for ṇyat. Its ṇ brings vṛddhi; its t is for the accent. SK 2871 gives kyap as an option after kṛ and vṛṣ. To state the other alternative, Bhaṭṭoji pulls ṛhalor ṇyat (SK 2872) ahead of its place. Then come two more nipātanas, yugya and amāvasyā, and two vārttikas on sṛj. The running example is kṛ "to do": kāryam with ṇyat, kṛtyam with kyap. Both mean "to be done".
Rule. After kṛ "to do" and vṛṣ "to rain", kyap comes optionally. Bhaṭṭoji then says "in the other alternative —", and the next rule follows.
Example.
kṛ + kyap → tuk after the short ṛ, since kyap is pit (SK 2858) → kṛ-t-ya → kṛtyam.
vṛṣ + kyap → kyap is kit, so no guṇa → vṛṣyam.
Applies when / blocked by. Why is the option needed? Both commentators explain. The two roots stand differently. kṛ ends in ṛ, so without this rule it would take only ṇyat (SK 2872). vṛṣ has ṛ as its penultimate, so it would take only kyap (SK 2859). This rule makes kyap optional for both.
Rule. After a dhātu ending in ṛ or ṝ (ṛvarṇa), or ending in a consonant (hal), the affix ṇyat comes. "After a dhātu" is carried forward from SK 2829. One commentator says so, and adds that the rule then means "ending in" these sounds.
Example.
kṛ + ṇyat → ṇyat is marked with ṇ, so the final ṛ takes vṛddhi (SK 254), ā with r after it (SK 70) → kār-ya → kāryam.
vṛṣ + ṇyat → the light penultimate ṛ takes guṇa, a with r (SK 2189) → varṣ-ya → varṣyam.
These are the ṇyat alternatives that SK 2871 leaves open. So kṛ has kṛtyam / kāryam, and vṛṣ has vṛṣyam / varṣyam.
Applies when / blocked by. In the Aṣṭādhyāyī this rule stands after SK 2873 and SK 2874. Bhaṭṭoji places it here to state the alternative of SK 2871. One of his commentators says every copy of the text reads it in this place. Another notes that Bhaṭṭoji takes it up again in its own order, under SK 2874.
Watch out. The ṛ of the sūtra is the sound ṛ, not the root ṛ "to go". One commentator gives reasons, among them: it is paired with hal, a class of sounds, and the rule acts on what follows it.
Rule. yugya is given by nipātana in the sense of patra, a draught animal. Bhaṭṭoji glosses patra as vāhana, "a carrier". The nipātana gives kyap, and ku for the j of yuj "to yoke".
Example.
yuj + kyap → g for j (by the nipātana) → yug-ya → yugyo gauḥ "an ox to be yoked".
Applies when / blocked by. yuj ends in a consonant, so SK 2872 would give ṇyat. Both commentators give that form: yogya. This rule is an exception (apavāda) to ṇyat.
Rule. After vas "to dwell" with amā "together" as upapada, ṇyat comes in the sense of the locus. After vṛddhi has applied, a short a is optionally given by nipātana.
Example.
amā + vas + ṇyat, locus
→ vṛddhi of the penultimate a (SK 2282) → amā-vās-ya
→ by the nipātana, optionally, short a again → amā-vas-ya
→ amāvāsyā / amāvasyā: "[the time] in which the moon and the sun dwell together" (amā saha vasataḥ), the new moon.
Applies when / blocked by. One commentator explains why Bhaṭṭoji says "after vṛddhi has applied". The short form is then the same word with one sound changed. So the taddhita of amāvāsyāyā vā (SK 1403) still reaches amāvasyā.
ṛhalor ṇyat again. Bhaṭṭoji now takes up SK 2872 in its Aṣṭādhyāyī order. One commentator says so. He gives pac "to cook":
pac + ṇyat (SK 2872)
→ vṛddhi of the penultimate a (SK 2282) → pāc-ya
→ k for c (SK 2863, group I below) → pāk-ya → pākyam "to be cooked".
vārttika pāṇau sṛjer ṇyad vācyaḥ: "ṇyat should be stated after sṛj 'to let out' with pāṇi 'hand' as upapada". It is an exception to the kyap of SK 2859, which sṛj would take for its penultimate ṛ.
pāṇi + sṛj + ṇyat → guṇa of the light ṛ (SK 2189) → sarj → g for j (SK 2863) → pāṇisargyā rajjuḥ "a rope made with the hands".
vārttika samavapūrvāc ca: "and after sṛj with sam and ava". One commentator supplies "sṛjer ṇyat" from the previous vārttika. So samavasargyā.
Watch out. Why is the first vārttika needed, if SK 2872 already gives ṇyat after a consonant? One commentator answers. kyap of SK 2859 is itself an exception to ṇyat, and would block it. The vārttika restores ṇyat against that exception.
mṛj + ṇyat → mārj-ya → mārgyaḥ; pac + ṇyat → pāk-ya → pākyam, but avaśya + pac + ṇyat → avaśyapācyamṇyat has a second effect besides vṛddhi. Before it, a c or j of the root is replaced by a velar: this is kutva, "the substitution of a ku". The ku is the velar class, k, kh, g, gh, ṅ. The same happens before an affix marked with gh, such as ghañ. SK 2863 states this, and SK 2864 adds a list of words. In the Kaumudī both stand right after SK 2862, to give mārgyaḥ. This chapter teaches them here, with the rules that withhold the change (SK 2875–2885). Bhaṭṭoji places those after SK 2874; one commentator calls them an incidental topic. The running example: mārgyaḥ and pākyam show kutva; avaśyapācyam "which must be cooked" keeps its c.
Rule. A ku, a velar, replaces c and j before a suffix marked with gh (a ghit) or before ṇyat. Bhaṭṭoji's gloss: "of c and of j there is kutva".
vārttika niṣṭhāyām aniṭa iti vaktavyam: "it should be stated: of a root that is aniṭ in the niṣṭhā". The niṣṭhā is the past-participle affix kta (and ktavatu). A root is aniṭ there if it takes no augment iṭ (i) before it; a root that takes iṭ is seṭ.
Example. The running example, from mṛj "to wipe", in the ṇyat alternative left by SK 2862:
mṛj + ṇyat
→ vṛddhi for mṛj by mṛjer vṛddhiḥ (SK 2473) → mārj-ya
→ g for j (this rule) → mārg-ya → mārgyaḥ "to be wiped".
pākyam from pac goes the same way (SK 2874): pāc-ya → pāk-ya.
Applies when / blocked by. The vārttika limits the rule. garj "to roar" takes iṭ in the niṣṭhā (garjita). So it is not aniṭ, and keeps its j: garjyam, "not here", says Bhaṭṭoji.
Watch out. Do not pair the sounds "respectively", c with ghit and j with ṇyat. Bhaṭṭoji's commentators both point to Pāṇini's own word rāga in tena raktaṃ rāgāt (SK 1202). There j becomes g before ghañ, a ghit. So each sound changes before both. One commentator adds that this is why Bhaṭṭoji writes "of c and of j".
Rule. Kutva also occurs in nyaṅku and the other words of its list.
Example.
ni + añc + the affix u, by the Uṇādi rule nāv añceḥ ("after añc with ni")
→ k for c (this rule)
→ nyaṅkuḥ, a kind of deer.
Applies when / blocked by. The affix u is neither ghit nor ṇyat. So SK 2863 cannot give the velar here; the list does.
Rule. In a dhātu that begins with a ku, a velar, c and j do not take kutva. "cajoḥ ku" is carried forward from SK 2863, one commentator says. So "na" here negates that kutva.
Example.
garj + ṇyat (SK 2872) → kutva would be due by SK 2863
→ garj begins with g, a velar: no kutva (this rule)
→ garjyam.
Applies when / blocked by. Bhaṭṭoji now reports a different view, that of the vārttikakāra, the author of the vārttikas. He completed SK 2863 with "niṣṭhāyām aniṭaḥ". He then rejected this rule and the next ones. One of Bhaṭṭoji's commentators lists them: this rule, SK 2876, and the mention of yāc and ruc in SK 2882. The two views give different results.
grokyam.One commentator explains why gruc and gluñc are aniṭ. Their iṭ is only optional before the affix ktvā. This is the kṛt affix of the absolutive ("having done"), taught in a later chapter. A root of that kind takes no iṭ in the niṣṭhā.
Which view wins? Bhaṭṭoji: yathottaraṃ munīnāṃ prāmāṇyam, "among the sages, the later is authority". The vārttikakāra comes after Pāṇini, so his result is accepted. For garjyam both views agree.
Watch out. Do not treat the conflict as an option between two forms. One commentator says Bhaṭṭoji's "nevertheless" answers exactly that doubt. The later sage decides.
Rule. For aj "to drive" and vraj "to go" too, there is no kutva. Bhaṭṭoji glosses only "no kutva". One commentator draws "ku" from SK 2863 and "na" from SK 2875, in a note on SK 2879.
Example.
sam + aj + ghañ → kutva would be due before the ghit (SK 2863)
→ no kutva (this rule); vṛddhi of the penultimate a (SK 2282) → samājaḥ.
pari + vraj + ghañ → likewise → parivrājaḥ.
One of Bhaṭṭoji's commentators identifies the affix as ghañ. One of them calls this rule and those after it "clear".
Rule. bhuja and nyubja are given by nipātana, in the senses "hand" (pāṇi) and "disease" (upatāpa), respectively. Bhaṭṭoji: "in these two senses these two forms are *nipāta*s." Bhaṭṭoji glosses upatāpa as roga, disease.
Example.
bhuj "to eat" + ghañ by halaś ca (SK 3300), in the instrument sense → bhujaḥ "the hand", "that by which one eats" (bhujyate 'nena).
ni + ubj → nyubjaḥ "a disease in which people are bent down" (nyubjanty asmin).
One commentator takes nyubja as ubj "to be straight" with ni.
Applies when / blocked by. In other senses kutva comes: bhogaḥ, samudgaḥ. Both show the velar g.
Rule. prayāja and anuyāja are given by nipātana when a component of the sacrifice (yajñāṅga) is meant. They keep the j of yaj "to sacrifice".
Example.
pañca prayājāḥ "the five prayājas" (preliminary offerings).
trayo 'nuyājāḥ "the three anuyājas" (closing offerings).
Applies when / blocked by. Why "a component of the sacrifice"? Otherwise: prayāgaḥ, anuyāgaḥ. These show the velar g that the nipātana withholds.
Rule. For vañc in the sense of going, there is no kutva. One of Bhaṭṭoji's commentators says "no kutva" fills the gap. "ku" comes from SK 2863, and "na" from SK 2875.
Example.
vañc + ṇyat, "going" → no kutva (this rule) → vañcyam "a place to be gone to".
Applies when / blocked by. Why "in the sense of going"? In its other sense kutva comes: vaṅkyaṃ kāṣṭham, "a piece of wood to be bent", made crooked.
Rule. oka is given by nipātana from the root uc, before the affix ka. Bhaṭṭoji: guṇa and kutva are given for uc when ka follows.
Example.
uc + ka → guṇa: o for u (by the nipātana) → kutva: k for c (by the nipātana) → okaḥ.
The word means a bird (śakunta) and a vṛṣala, a śūdra.
Applies when / blocked by. The ka is that of igupadhajñāprīkiraḥ kaḥ (SK 2897), taught after a root with an ik vowel as its penultimate. The ik vowels are i, u, ṛ, ḷ, short or long. Bhaṭṭoji adds a point. The form oka would arise anyway with ghañ. So this rule is for the accent: the final vowel is to be udātta, high-pitched. One commentator explains: with ghañ the first vowel would be udātta, which is not wanted.
Rule. Before ṇya (the ya of ṇyat) conveying necessity (āvaśyaka), there is no kutva. Bhaṭṭoji glosses only "no kutva".
Example.
avaśya + pac + ṇyat, "necessarily"
→ vṛddhi of the penultimate a (SK 2282) → avaśya-pāc-ya
→ necessity is conveyed: no kutva (this rule)
→ avaśyapācyam "that which must necessarily be cooked".
Compare pākyam "to be cooked", with kutva (SK 2874).
Applies when / blocked by. One commentator names the source of ṇyat here: āvaśyakādhamarṇyayor ṇiniḥ (SK 3311) with kṛtyāś ca (SK 3312). The same commentator adds a point on the word "necessity". It names the sense, not a word that must be present. Where context alone shows necessity, the c still stays.
Rule. Before ṇya, there is no kutva for yaj "to sacrifice", yāc "to ask", ruc "to shine", pravac "to proclaim" and ṛc "to praise". Bhaṭṭoji's gloss: "before ṇya there is no kutva".
Example.
yaj + ṇyat → vṛddhi (SK 2282) → yāj-ya → no kutva (this rule) → yājyam.
Likewise yācyam; ruc + ṇyat → guṇa (SK 2189) → rocyam; and pravācyam, a particular kind of text.
ṛc + ṇyat → guṇa of the light ṛ (SK 2189) → arc-ya → no kutva → arcyam.
Applies when / blocked by. ṛc has ṛ as its penultimate. So it should take kyap (SK 2859), not ṇyat. Bhaṭṭoji: the mention of ṛc in this very sūtra is a jñāpaka, an indication, that it takes ṇyat. A rule about ṇya after ṛc would be pointless otherwise. One commentator adds why the option between unlike affixes (SK 2830) cannot let ṇyat in. kyap and ṇyat count as of the same form (sarūpa). So the indication is the only source.
vārttika tyajeś ca: "and for tyaj 'to abandon'": tyājyam.
The Kāśikā reads "tyajipūjyoś ca", adding pūj "to honour". Bhaṭṭoji calls the mention of pūj questionable (cintya). The Bhāṣya does not state it. It says only: "in the ṇyat section, tyaj is to be added".
Watch out. pravācyam is the name of a text. One commentator notes that the next rule would therefore not protect it. Naming pravac here supplies the protection.
Rule. For vac "to speak", before ṇya, there is no kutva, unless the word is a name (saṃjñā) of a speech-form (śabda).
Example.
vac + ṇyat → vṛddhi (SK 2282) → vāc-ya → no kutva (this rule) → vācyam "a thing to be said".
As the name of a speech-form, kutva comes: vāk-ya → vākyam, the technical term "sentence".
Applies when / blocked by. Bhaṭṭoji's commentators define vākya in slightly different words; scholars discuss this. Both take it as a technical name, so kutva applies.
Watch out. pravācya is also a name, of a text. So the exception in this rule would give it kutva. It keeps its c only because SK 2882 names pravac specifically. One commentator explains this.
Rule. prayojya and niyojya are given by nipātana in the sense "able to be …" (śakya). The nipātana withholds kutva from the j of yuj before ṇya.
Example.
pra + yuj + ṇyat → guṇa of the light u (SK 2189) → pra-yoj-ya → no kutva (by the nipātana) → prayojyaḥ "able to be employed" (prayoktuṃ śakyaḥ).
Likewise niyojyaḥ "able to be commanded" (niyoktuṃ śakyaḥ): a servant.
Applies when / blocked by. How can a ṇyat-word mean "able to"? One commentator answers. śaki liṅ ca (SK 2823) gives the kṛtya affixes in the sense of capability too.
Rule. bhojya is given by nipātana when something to be eaten (bhakṣya) is meant. Kutva is withheld from the j of bhuj before ṇya.
Example.
bhuj + ṇyat → guṇa (SK 2189) → bhoj-ya → no kutva (by the nipātana) → bhojyam "food".
Otherwise, Bhaṭṭoji says, bhogyam. One commentator glosses it "to be protected, or to be enjoyed". He adds that bhakṣya covers anything eaten, so bhojyā yavāgūḥ "gruel to be eaten" is also correct.
Applies when / blocked by. Here the incidental kutva rules end. One commentator marks the point: "this was incidental; now the main topic", ṇyat.
vārttika ṇyatprakaraṇe lapidabhibhyāṃ ceti vaktavyam: "in the ṇyat section, it should be stated: and after lap 'to speak' and dabh 'to harm'".
lap + ṇyat → vṛddhi (SK 2282) → lāpyam.
dabh + ṇyat → dābhyaḥ.
Both roots end in a labial and have a as penultimate. So SK 2844 would give yat. Both commentators say the vārttika sets that yat aside. Bhaṭṭoji adds that dabh is not read among the dhātus. It is accepted on the strength of the vārttika.
Watch out. Jayāditya, an author of the Kāśikā, reads lap into SK 2887 instead. One of Bhaṭṭoji's commentators rejects that reading as baseless and against the Bhāṣya. In the Kaumudī, lap comes by this vārttika.
lū + ṇyat → lau-ya → lāvyam; ā + nī + ṇyat → ānāyyaḥ, against āneyaḥThe kutva rules were a side topic. Bhaṭṭoji now returns to ṇyat. SK 2886 and SK 2887 give it after roots that would otherwise take yat. SK 2888–2893 are nipātanas, mostly words of the Vedic ritual. Several show āy where the root has ī or i: nī gives ānāyya, ci gives saṃcāyya. The running example has two parts. lāvyam "which must be cut" shows ṇyat after a u-final root. ānāyyaḥ shows the ritual āy.
Rule. After a dhātu ending in u or ū (uvarṇa), ṇyat comes when necessity is to be conveyed. Bhaṭṭoji glosses necessity as avaśyaṃbhāva, "necessary occurrence".
Example.
lū "to cut" + ṇyat, necessity
→ ṇyat is marked with ṇ, so the final ū takes vṛddhi (SK 254) → lau-ya
→ āv for au before the y of an affix (SK 63) → lāv-ya
→ lāvyam "which must necessarily be cut".
Likewise pū "to purify" → pāvyam "which must necessarily be purified".
Applies when / blocked by. Without necessity, a u-final root takes yat by SK 2842. This rule overrides that only in the sense of necessity.
"Necessity" names the sense, not a word that must be present. One commentator says so: if it meant "with the word avaśya", lāvyam alone would not be formed. The word may still be added: avaśyalāvyam.
Watch out. stu "to praise" also ends in u. But it takes kyap, even in necessity: avaśyastutyaḥ. The kyap restated in SK 2857 beats even this later ṇyat (see SK 2858).
Rule. ṇyat also comes after su with ā (āsu), and after yu, vap, rap, trap and cam. One commentator supplies "ṇyat" as the missing word.
Example. Bhaṭṭoji names the roots meant:
ā + su + ṇyat → vṛddhi (SK 254) → āv (SK 63) → āsāvyam.yāvyam.vāpyam, rāpyam, trāpyam, cāmyam.Applies when / blocked by. One of Bhaṭṭoji's commentators says what this rule overrides. For āsu and yu, aco yat (SK 2842) would give yat. For rap, trap and cam, por adupadhāt (SK 2844) would. The same commentator explains why the yu is "to mix". yuñ "to bind" carries a marker, ñ, and the sūtra's yu has none.
Rule. ānāyya is given by nipātana in the sense "impermanent" (anitya). Bhaṭṭoji: ṇyat after nī "to lead" with ā, and the substitute āy, are given by nipātana.
Example.
ā + nī + ṇyat
→ vṛddhi of the final ī (SK 254) → ā-nai-ya
→ āy for ai (by the nipātana) → ā-nāy-ya
→ ānāyyaḥ.
Applies when / blocked by. Bhaṭṭoji limits the word to a particular southern fire (dakṣiṇāgni). It is brought (ānīyate) from the household fire (gārhapatya). It is impermanent, because it is not kept constantly burning. Anything else brought is āneyaḥ: a pot and the like. So too is the southern fire brought from the house of a vaiśya or the like.
One commentator explains "impermanent". The fire is brought anew each day for the rite, and becomes ordinary fire once the rite is over.
Watch out. Divide the sūtra as ānāyyaḥ anitye, "in the impermanent". One commentator says so, against reading ānāyyo nitye. Why must āy come by nipātana? See SK 2890.
Rule. praṇāyya is given by nipātana when there is no saṃmati. One commentator gives its parts: ṇyat, vṛddhi and āy.
Example.
pra + nī + ṇyat → vṛddhi (SK 254) → āy (by the nipātana) → praṇāyyaḥ.
Applies when / blocked by. Bhaṭṭoji gives saṃmati two senses, and so two uses of the word:
praṇāyyaś coraḥ, a thief, "unworthy of affection".praṇāyyo 'ntevāsī, a pupil who is "detached".Otherwise: praṇeyaḥ.
Bhaṭṭoji calls the first saṃmati "an activity of the object" (karmavyāpāra). One commentator explains: it is the thief who is not loved. The other says both senses are meant at once.
Rule. Four words are given by nipātana, in four senses, respectively. pāyya means a measure; sānnāyya, an oblation; nikāyya, a dwelling; dhāyyā, a sāmidhenī verse.
Example.
mā "to measure" + ṇyat, instrument → p for the initial m (by the nipātana) → pā-ya
→ the augment yuk (y) by āto yuk (SK 2761) → pā-y-ya → pāyyam.
sam + nī + ṇyat → vṛddhi → āy for ai, and long ā in sam (by the nipātana) → sānnāyyam.
ni + ci + ṇyat, locus → vṛddhi → ni-cai-ya
→ āy for ai, and k for the root's initial c (by the nipātana) → nikāyyaḥ.
Applies when / blocked by. Why must āy be given by nipātana? One commentator answers for nikāyya. ai becomes āy only before a vowel (SK 61). Before the consonant y nothing would replace it. So the nipātana supplies āy.
One commentator names the dwelling: a granary. Both commentators say the dhāyyā is the verse pṛthupājā amartyaḥ, inserted among the verses for laying the fuel.
Rule. kuṇḍapāyya and saṃcāyya are given by nipātana when a sacrifice (kratu) is meant.
Example.
kuṇḍa + pā + ya, locus → the augment yuk → kuṇḍapāyyaḥ: a sacrifice "in which soma is drunk with a kuṇḍa" (kuṇḍena pīyate 'smin).
sam + ci + ṇyat → vṛddhi → āy (by the nipātana) → saṃcāyyaḥ: a sacrifice "which is accumulated" (saṃcīyate).
Applies when / blocked by. Bhaṭṭoji names no affix for kuṇḍapāyya. One commentator takes ṇyat, and yuk then comes by SK 2761. The other takes yat, with yuk given by the nipātana; yat fixes a different accent. Scholars discuss this. For saṃcāyya both give ṇyat and āy. Outside a sacrifice, one commentator gives kuṇḍapānam and saṃceyam.
Rule. paricāyya, upacāyya and samūhya are given by nipātana when the agni is meant. Bhaṭṭoji: they are correct forms for a particular site built for holding the fire.
Example. Bhaṭṭoji's commentators give the parts:
pari + ci + ṇyat → vṛddhi → āy (by the nipātana) → paricāyyaḥ; likewise upacāyyaḥ.
sam + vah + ṇyat → saṃprasāraṇa: u for the v, the a merging into it (by the nipātana) → sam-uh-ya
→ lengthening (by the nipātana) → samūhyaḥ.
Applies when / blocked by. Elsewhere, Bhaṭṭoji says: pariceyam, upaceyam, saṃvāhyam.
Watch out. agni here is not the flame. Both commentators say it is the site built of bricks to hold the fire. One cites the śruti, the Vedic text: samūhyaṃ cinvīta paśukāmaḥ, "one who desires cattle should build the samūhya".
Rule. citya and agnicityā are also given by nipātana. "When the agni is meant" is carried forward from SK 2892.
Example.
ci "to build" → cityaḥ: the agni altar "that is built up" (cīyate).
agni + ci → agnicityā: "the building of the agni".
Applies when / blocked by. Bhaṭṭoji names no affix. His commentators differ on the affix, and on whether "agni" qualifies both words or only citya; scholars discuss this. One commentator notes that elsewhere the form is ceyam.
Bhaṭṭoji then recalls two kṛtya rules from chapter 65. They are taken up with SK 2894 below.
bhū + yat → bhav-ya → bhavyaḥ "one who becomes" / bhavyam anena; gai → gā + yat → geyaḥ sāmnām / geyaṃ sāma anenaSK 2833 confined the kṛtya affixes to bhāva and karman. The last sūtra of the chapter lets seven kṛtya-words mean the agent as well, by option. Bhaṭṭoji then recalls three kṛtya rules from chapter 65. They show that the option of SK 2830 does not always hold after the strī-adhikāra (SK 3272). The running example is bhavya from bhū: bhavyaḥ "one who becomes" (agent) or bhavyam anena "it is to be become by him" (bhāva).
Rule. bhavya, geya, pravacanīya, upasthānīya, janya, āplāvya and āpātya are given by nipātana, optionally in the sense of the agent. "In the sense of the agent" is carried forward from kartari kṛt (SK 2832). One commentator says so.
In the other alternative SK 2833 holds. After a transitive (sakarmaka) root the word is in karman, the object. After an intransitive (akarmaka) root it is in bhāva, the action itself.
Example.
bhū + yat (SK 2842) → guṇa (SK 2168) → bho-ya → av for o before y (SK 63) → bhav-ya
→ agent: bhavyaḥ "one who becomes" (bhavati);
→ or bhāva, since bhū is intransitive: bhavyam anena "it is to be become by him".
gai "to sing" → ā for ai (SK 2370) → gā + yat (SK 2842)
→ ī for ā before yat (SK 2843) → guṇa (SK 2168) → geya
→ agent: geyaḥ sāmnām ayam "this one is a singer of sāmans", the Vedic chants;
→ or karman, since gai is transitive: geyaṃ sāma anena "the sāman is to be sung by him".
Applies when / blocked by. Bhaṭṭoji ends with "and so on". Bhaṭṭoji's commentators fill it in, each word in both senses:
pravacanīyo guruḥ "the teacher who expounds" / pravacanīyo vedaḥ "the Veda to be expounded";upasthānīyaḥ śiṣyo guroḥ "the pupil who attends the teacher" / upasthānīyo guruḥ śiṣyeṇa "the teacher to be attended by the pupil";janyaḥ / janyam anena;āplāvyaḥ / āplāvyam anena, with ṇyat by SK 2886;āpātyaḥ / āpātyam anena, with ṇyat by SK 2872.(The last pair, and the SK numbers for the ṇyat, are one commentator's. For vedaḥ the other commentator has svādhyāyaḥ, Vedic recitation.)
Watch out. In geyaḥ sāmnām ayam the object sāmnām is in the sixth case, not the second. One commentator explains: the object goes with a kṛt-word, and is not expressed by it. ayam "this one" is in the first case, because the agent is what geyaḥ expresses.
Recalled: SK 2817, praiṣātisargaprāptakāleṣu kṛtyāś ca. In the senses of a command (praiṣa), a permission (atisarga) and "the time has come" (prāptakāla), the kṛtya affixes come too:
tvayā gantavyam / gamanīyam / gamyam "you are to go".
The affix loṭ is taught specially for these senses. A special rule would block the general kṛtyas. So Pāṇini restates the kṛtyas here, so that loṭ does not block them.
But SK 2830 already lets a general affix stand beside a special affix of unlike form. Why restate? Bhaṭṭoji draws an indication (jñāpaka) from it. After the strī-adhikāra, the option of SK 2830 sometimes does not hold. The strī-adhikāra is the heading that begins with striyāṃ ktin (SK 3272). SK 2817 stands after it.
Where does the option fail? Bhaṭṭoji: in the domains of kta, lyuṭ, tumun and the affixes in the sense of khal. One commentator illustrates:
hasitam, hasanaṃ chātrasya "the student's laughing": ghañ does not come where kta and lyuṭ serve;icchati bhoktum "he wishes to eat": loṭ does not come where tumun serves;īṣatpānaḥ "easily drunk": khal does not come where yuc serves. yuc is a kṛt affix ana, which comes in the sense of khal after roots in ā.Why "sometimes". One commentator gives the reason. If the option always failed after SK 3272, ktin would always be blocked by aṅ. Both are kṛt affixes forming feminine action nouns, taught in the strī-adhikāra (SK 3272, SK 3281). Then forms such as kṣānti and bhitti would not arise.
Recalled: SK 2822, arhe kṛtyatṛcaś ca. In the sense "deserving", the kṛtyas and tṛc come too:
stotum arhaḥ stutyaḥ "deserving to be praised", the object of praise;
stotā "deserving to praise", the agent of praise.
The kṛtyas and tṛc are named so that liṅ, taught for this sense, does not block them.
Recalled: SK 2823, śaki liṅ ca. In the sense "able", liṅ comes, and by "ca" the kṛtyas too:
voḍhavyaḥ "able to be carried" (voḍhuṃ śakyaḥ); likewise vahanīyaḥ, vāhyaḥ.
The "ca" brings in the kṛtyas so that liṅ does not block them.
Bhaṭṭoji closes with a point of economy (lāghava). This "ca" alone is enough to give the indication about SK 2830. So "kṛtyāś ca" in SK 2817, and "kṛtyatṛcaḥ" in SK 2822, need not be taken as its source. One commentator explains: resting it on those two would be heavier.
Bhaṭṭoji ends: "Here ends the section on the kṛtya affixes." The next chapter turns to the other kṛt affixes. It opens with ṇvultṛcau (SK 2895), where the kṛtya heading of SK 2831 stops.
edhitavyam.cetavyaḥ dharmaḥ.cetavyaḥ, cayanīyaḥ. The t and r are markers for the accent.edhitavyam.pacelimāḥ.prenvanam (SK 2838).snānīyam, vahyam.dānīyaḥ.yāpi.prayāpyamāṇam.prayāpyamāṇam (SK 2836).duryānam.ceyam, deyam.avadyam, ajaryam.ajaritā.stutyaḥ.stutyaḥ.mṛṣodyam, samūhyaḥ, or r to ṛ in pratigṛhyam.sūryaḥ (SK 2865).rājasūya, puṣyaḥ, amāvāsyā, nikāyyaḥ.nyaṅkuḥ (SK 2864).samājaḥ, bhujaḥ; it is a ghit (SK 2876, 2877).okaḥ (SK 2880).ānāyyaḥ, sānnāyyam, nikāyyaḥ, saṃcāyyaḥ, paricāyyaḥ.pāyyam; in kuṇḍapāyyaḥ by SK 2761 or by the nipātana, depending on the commentator (SK 2891).bhoktum (SK 2894).īṣatpānaḥ (SK 2894).| SK | what it does | example |
|---|---|---|
| 2829 | heading: "after a dhātu", to the end of the third adhyāya | edhitavyam, cetavyaḥ |
| 2830 | paribhāṣā: an asarūpa exception blocks the general affix only optionally, outside the strī-adhikāra | bhavyam / bhavitavyam |
| 2831 | heading and name: the affixes up to ṇvul are kṛtya | tavyat, yat, kyap … |
| 2832 | a kṛt denotes the agent | cetavya would be "one who gathers" |
| 2833 | the kṛtyas, kta and khal-sense affixes only in bhāva and karman | edhitavyam tvayā; cetavyaḥ dharmas tvayā |
| 2834 | tavyat, tavya, anīyar after a dhātu; vārttikas: vāstavya, kelimar | cetavyaḥ, cayanīyaḥ; vāstavyaḥ; pacelimāḥ |
| 2841 | the kṛtyas and lyuṭ occur variously | snānīyaṃ cūrṇam, dānīyo vipraḥ |
| 2835 | ṇ for the n of a kṛt after a vowel, with the cause in an upasarga | prayāṇīyam; vārttika nirviṇṇaḥ |
| 2836 | optional, for a kṛt taught after a ṇi-final root | prayāpaṇīyam / prayāpanīyam |
| 2837 | optional after a consonant-initial root with an ic as penultimate | prakopaṇīyam / prakopanīyam |
| 2838 | restriction: for a root with num, only after an ic-initial, consonant-final one | preṅkhaṇīyam; not pramaṅganīyam |
| 2839 | optional ṇ for the root's n in niṃs, nikṣ, nind before a kṛt | praṇiṃsitavyam / praniṃsitavyam |
| 2840 | no ṇ after bhā, bhū, pū and others | prabhānīyam, prabhavanīyam; prakhyānīyam |
| 2842 | yat after a vowel-final root | ceyam, jeyam |
| 2843 | ī for root-final ā before yat; vārttikas: yat after tak etc.; vadha | deyam, gleyam; vadhyaḥ / ghātyaḥ |
| 2844 | yat after a labial-final root with a as penultimate, against ṇyat | śapyam, labhyam |
| 2845 | num for labh after āṅ, a y-initial affix in view; so ṇyat | ālambhyo gauḥ |
| 2846 | the same after upa, in praise | upalambhyaḥ sādhuḥ / upalabhyaḥ |
| 2847 | yat after śak and sah | śakyam, sahyam |
| 2848 | yat after gad, mad, car, yam without an upasarga | gadyam, madyam, caryam; ācaryaḥ; prayāmyam |
| 2849 | nipātana: avadya, paṇya, varyā | avadyaṃ pāpam, paṇyā gauḥ, varyā kanyā |
| 2850 | nipātana: vahya, the instrument | vahyaṃ śakaṭam / vāhyam |
| 2851 | nipātana: arya, master or vaiśya | aryaḥ / āryaḥ |
| 2852 | nipātana: upasaryā, a female whose time for conception has come | upasaryā gauḥ / upasāryā |
| 2853 | nipātana: ajarya, agent, of an association | ajaryam / ajāryaṃ saṃgatena |
| 2854 | kyap and yat after vad with a sup-final upapada, no upasarga | brahmodyam, brahmavadyam |
| 2855 | kyap after bhū in bhāva, same conditions | brahmabhūyam / bhavyam |
| 2856 | kyap after han in bhāva, and t for its final | brahmahatyā |
| 2857 | kyap after i, stu, śās, vṛ, dṛ, juṣ | examples under 2858 |
| 2858 | augment tuk to a short vowel before a pit kṛt | ityaḥ, stutyaḥ, śiṣyaḥ, avaśyastutyaḥ; ājyam |
| 2859 | kyap after a root with short ṛ as penultimate, not kḷp, cṛt | vṛtyam, vṛdhyam; kalpyam |
| 2860 | kyap after khan, with ī for its final | kheyam |
| 2861 | kyap after bhṛñ, when not a name | bhṛtyāḥ; saṃbhṛtyāḥ / saṃbhāryāḥ |
| 2862 | kyap after mṛj optionally; else ṇyat | mṛjyaḥ / mārgyaḥ |
| 2865 | Seven kyap-words by nipātana: rājasūya, sūrya, mṛṣodya, rucya, kupya, kṛṣṭapacya, avyathya | sūryaḥ, mṛṣodyam; but gopyam, kṛṣṭapākyāḥ |
| 2866 | bhidya, uddhya (kyap, agent) of a river | bhidyaḥ, uddhyaḥ; but bhettā |
| 2867 | puṣya, siddhya (kyap, locus) of a lunar mansion | puṣyaḥ, siddhyaḥ |
| 2868 | vipūya, vinīya, jitya for muñja, kalka, hali | vipūyo muñjaḥ; else vipavyam |
| 2869 | kyap after grah with prati, api (Vedic, by vārttika) | pratigṛhyam; ordinary pratigrāhyam |
| 2870 | kyap after grah for pada, asvairin, bāhyā, pakṣya | pragṛhyam, grāmagṛhyā senā |
| 2871 | kyap optionally after kṛ, vṛṣ | kṛtyam, vṛṣyam |
| 2872 | ṇyat after a ṛ-final or consonant-final root | kāryam, varṣyam, pākyam |
| 2873 | yugya (kyap, kutva) for a draught animal | yugyo gauḥ |
| 2874 | amāvasyā (ṇyat, locus; short a optional); vārttikas ṇyat after sṛj | amāvāsyā / amāvasyā; pāṇisargyā |
| 2863 | ku for c, j before a ghit or ṇyat (vārttika: if aniṭ in the niṣṭhā) | mārgyaḥ; but garjyam |
| 2864 | Kutva also in nyaṅku etc. | nyaṅkuḥ |
| 2875 | No kutva in a velar-initial root (set aside by the vārttikakāra) | garjyam |
| 2876 | No kutva for aj, vraj | samājaḥ, parivrājaḥ |
| 2877 | bhuja (hand), nyubja (disease) without kutva | bhujaḥ; else bhogaḥ |
| 2878 | prayāja, anuyāja (parts of the sacrifice) without kutva | pañca prayājāḥ; else prayāgaḥ |
| 2879 | No kutva for vañc "go" | vañcyam; else vaṅkyam |
| 2880 | oka from uc + ka, with guṇa and kutva | okaḥ |
| 2881 | No kutva before ṇya of necessity | avaśyapācyam |
| 2882 | No kutva before ṇya for yaj, yāc, ruc, pravac, ṛc; vārttika tyaj | yājyam, arcyam, tyājyam |
| 2883 | No kutva for vac before ṇya, unless a name of a speech-form | vācyam; but vākyam |
| 2884 | prayojya, niyojya "able to be …" without kutva | prayojyaḥ, niyojyaḥ |
| 2885 | bhojya "food" without kutva; vārttika ṇyat after lap, dabh | bhojyam; else bhogyam; lāpyam |
| 2886 | ṇyat after a u-final root in the sense of necessity | lāvyam, pāvyam |
| 2887 | ṇyat after āsu, yu, vap, rap, trap, cam | āsāvyam, yāvyam, cāmyam |
| 2888 | ānāyya (ṇyat, āy): the impermanent southern fire | ānāyyaḥ; else āneyaḥ |
| 2889 | praṇāyya where there is no saṃmati | praṇāyyaś coraḥ; else praṇeyaḥ |
| 2890 | pāyya, sānnāyya, nikāyya, dhāyyā: measure, oblation, dwelling, sāmidhenī | pāyyam, nikāyyaḥ |
| 2891 | kuṇḍapāyya, saṃcāyya for a sacrifice | kuṇḍapāyyaḥ, saṃcāyyaḥ |
| 2892 | paricāyya, upacāyya, samūhya for the fire-site | samūhyaḥ; else saṃvāhyam |
| 2893 | citya, agnicityā | cityaḥ, agnicityā |
| 2894 | Seven kṛtya-words optionally in the agent sense; recall of SK 2817, 2822, 2823 | bhavyaḥ / bhavyam anena; voḍhavyaḥ |