Most chapters so far have taught the language of ordinary usage (loka), with a few Vedic rules on the way. This chapter gathers the rules that hold only in the Veda (chandas). It has 263 sūtras, SK 3387 to SK 3649. Bhaṭṭoji takes them in the order of the Aṣṭādhyāyī, book by book. Most of them are exceptions: an earlier rule is relaxed, reversed or extended "in the Veda" (chandasi). So each section here names the earlier rule it changes, by its SK number. This first stretch, SK 3387–3438, covers books 1 to 3. After it you will be able to explain the singular punarvasu nakṣatram and the bha-stem nabhasvat. You will explain the case-endings of yavāgvā 'gnihotraṃ juhoti, and the root-forms hanati, dāti and praṇak. You will build the Vedic aorists ajūgupatam and akaram, the kṛtya niṣṭarkya, and the upapada nouns havyavāhanaḥ, havyavāḷ and śvetavāḥ. You will place the udātta of vṛṣṭí. You will form the Vedic mood leṭ (tāriṣat, patāti, mādayaite), see how vikaraṇas are exchanged (bhedati), and read infinitives such as jīváse and dṛśe.
á) and an independent svarita with a grave (à). An anudātta is left unmarked. Accent is marked only where a rule is about accent.punarvasū nakṣatram → punarvasu nakṣatram (singular for two); nabhas + vati → (bha: no ru) nabhasvatBook 1 of the Aṣṭādhyāyī defines names and general conditions. Its Vedic rules relax those definitions. Two stars may be named in the singular. pati may be a ghi outside a compound. A stem may be bha where it would be pada, or both at once. Upasargas may follow the verb or stand apart from it. The running examples are punarvasu nakṣatram, "the asterism Punarvasu", and nabhasvat, "like the sky".
The kaumudī heads the whole section vaidikī prakriyā, "the Vedic procedure". It also marks the books of the Aṣṭādhyāyī. Book 1 runs from SK 3387 to SK 3393, book 2 from SK 3394 to SK 3402, and book 3 begins at SK 3403.
Rule. In the Veda, punarvasu denoting two stars may take the singular. The word "optionally" (anyatarasyām) is carried forward from SK 817 (chapter 18), the Subodhinī says. Ordinary usage has only the dual.
Example.
punarvasu, naming two stars → the dual is expected → punarvasū nakṣatram
→ this rule allows the singular → punarvasu nakṣatram
Applies when / blocked by. The Subodhinī explains why the dual is expected. Punarvasu is a cluster of stars whose members stand out. So for two stars the dual would come, and this rule gives the singular.
Rule. The same holds for viśākhā: denoting two stars, it may take the singular in the Veda. Bhaṭṭoji says only "as before".
Example.
viśākhe nakṣatram (dual) → or, by this rule → viśākhā nakṣatram (singular)
Watch out. The Subodhinī notes that the sūtra is neutral on ordinary usage. So Amara's singular viśākhā outside the Veda is not barred.
Rule. In the Veda, pati construed with a sixth-case word is optionally a ghi. The word pati is carried forward from SK 257 (chapter 8), the Subodhinī says. SK 257 allows pati to be a ghi only in a compound. This rule relaxes that limit.
Example.
kṣetrasya pati + ṭā ("with the lord of the field")
→ pati is ghi by this rule
→ ṭā becomes nā (SK 244)
→ kṣetrasya patinā vayam
Applies when / blocked by. The Subodhinī gives two counter-examples. Without a sixth-case word: mayā patyā. Outside the Veda: grāmasya patye.
Watch out. Bhaṭṭoji splits this sūtra in two (yogavibhāga). The second part is just "vā", with chandasi carried into it. So every operation is optional in the Veda. The later bahulaṃ chandasi rules (such as SK 3401) are only an elaboration of this. (The 1915 print reads prapañcaḥ, "elaboration".)
Example (vārttika on SK 231). A vārttika on yaci bham (SK 231, chapter 8) makes nabhas, aṅgiras and manus bha before vat:
nabhasā tulyam "like the sky" → nabhas + vati (SK 1778, chapter 37)
→ nabhas is bha, not pada
→ no ru for the s (SK 162 needs a pada)
→ nabhasvat
Likewise aṅgirasvad aṅgiraḥ and manuṣvad agne. In manuṣvat, bha status lets s become ṣ (SK 212). manus itself takes the uṇādi usi by bāhulaka.
Example (second vārttika). vṛṣan is bha before vasu and aśva: vṛṣaṇvasuḥ, "whose wealth is raining", and vṛṣaṇaśvaḥ, "whose horse is virile".
Ordinary usage has vṛṣavasuḥ and vṛṣāśvaḥ (Subodhinī).
Rule. ayasmaya and the words of its list are correct in the Veda. They take both names, bha and pada, as each form requires.
Example.
ṛc + an affix beginning with v → pada status gives kutva, c → k (SK 378)
→ bha status blocks jaśtva of that k (SK 84)
→ stem ṛkvat- → instrumental ṛkvatā (sa ṛkvatā gaṇena)
vāc + an affix ina → pada status gives jaśtva, c → j
→ bha status blocks kutva
→ vājina (nainaṃ hinvanty api vājineṣu)
Bhaṭṭoji and his commentators do not name the affixes; they are inferred from the finished forms.
Applies when / blocked by. By proximity only the nearer name, bha, would apply (the maxim anantarasya). The vārttika "they have both names as well" is the authority for giving both, the Subodhinī says.
Watch out. The unit ends by quoting te prāg dhātoḥ (SK 2230, chapter 44). That rule is the frame for the next two sūtras.
Rule. In the Veda, the gati and upasarga words may also stand after the root. By api "also", they may still stand before it. This is the exception to SK 2230 (chapter 44), which puts them before the root.
Example. Bhaṭṭoji gives no comment and no example here. The examples for this rule and the next stand under SK 3392.
Rule. The gati and upasarga words may also stand separated from the root.
Example.
haribhyāṃ yāhy oka ā "come with the two bay horses to our home": ā comes after yāhi, with okaḥ between.
ā mandrair indra haribhir yāhi "come, Indra, with the gladdening bay horses": ā comes before yāhi, separated from it.
Without these rules the form would be āyāhi (Subodhinī).
Applies when / blocked by. The Subodhinī says the two mantras illustrate both sūtras together. Which example serves which rule is not stated by Bhaṭṭoji or his commentators; it is inferred. That "in the Veda" continues from SK 3391 is also inferred.
Watch out. The Subodhinī reports that the Bhāṣya can do without both rules. It reads SK 2230 as limiting the names themselves to words placed before the root.
Rule. A liṭ after indh "to kindle" and bhū "to be" is kit, even a liṭ that is pit. "liṭ" and "kit" are carried forward from SK 2242 (chapter 44), the Subodhinī says.
Example.
sam + indh + liṭ (e) → kit
→ the nasal is lost (SK 415)
→ the root is doubled, as in every liṭ (not stated in this section)
→ samīdhe (samīdhe dasyuhantamam); likewise īdhe
bhū + ṇal → ṇal is pit, yet this rule makes it kit, so no vṛddhi
→ the augment vuk (SK 2174)
→ the root is doubled, as in every liṭ (not stated in this section)
→ babhūva
Applies when / blocked by. The Subodhinī says why the rule is needed. SK 2242 does not reach indh, which ends in a conjunct. Nor does it reach babhūva, because ṇal is pit.
Watch out. A vārttika then rejects the sūtra. indh is used only in the Veda. And the vuk of bhū is nitya, so it comes first anyway. The wording of Bhaṭṭoji's gloss here is doubtful; it is taken as "a liṭ … also", which fits his own example babhūva.
yavāgvā 'gnihotraṃ juhoti (third case for the object of hu); vanaspatīnām for vanaspatibhyaḥ (sixth for fourth)Book 2 teaches the case-endings (chapter 16) and the gender of compounds (chapter 18). In the Veda some of these choices loosen. The object of hu may take the third case. The sixth case may stand for the fourth, and for the instrument of yaj. Two dvandvas take the gender of their first member. The running examples are yavāgvā 'gnihotraṃ juhoti, "he offers the oblation, the gruel", and te vanaspatīnām, "these are for the trees".
Rule. In the Veda the object of hu "to offer" takes the third case. By ca, it takes the second as well. "Object" (karmaṇi) is carried forward from SK 537 (chapter 16), the Subodhinī says. SK 537 alone would give only the second case.
Example.
object of juhoti: yavāgū "gruel" → third case → yavāgvā
object of juhoti: agnihotra → second case → agnihotram
→ yavāgvā 'gnihotraṃ juhoti
Here agnihotra means the oblation itself, "that which is offered to Agni". The two words stand in apposition. The sense is: "he casts the oblation, the gruel, into the fire, with the deity in view."
Watch out. The Subodhinī reports that the Bhāṣya rejects this sūtra.
Rule. In Brāhmaṇa usage, the object of div "to stake" takes the second case. This is an exception to the sixth case, which SK 619 (chapter 16) gives for that object.
Example.
object of dīvyeyuḥ "they should stake": go "cow" → SK 619 would give the sixth → this rule gives the second → gām
→ gām asya tadahaḥ sabhāyāṃ dīvyeyuḥ "on that day they should play for his cow in the hall"
Applies when / blocked by. The Subodhinī says div with an upasarga is already covered by SK 620. So this rule is for div without one.
Rule. In the Veda the sixth case stands, variously, in the sense of the fourth. The fourth is the case of the recipient (chapter 16).
Example.
"for the Moon" → fourth expected → sixth → puruṣamṛgaś candramasaḥ
"for the trees" → fourth vanaspatibhyaḥ expected → sixth → te vanaspatīnām
Applies when / blocked by. A vārttika gives the reverse: the fourth in the sense of the sixth. yā kharveṇa pibati tasyai kharvaḥ: tasyai (fourth) stands for tasyāḥ (sixth).
Watch out. Because the rule says bahulam "variously", the fourth may also stand. The Subodhinī mentions candramase as such a case.
Rule. The same holds for the instrument (karaṇa) of yaj "to sacrifice". In the Veda it may take the sixth case, variously. Normally the instrument takes the third (chapter 16).
Example.
instrument of yajate: ghṛta "ghee" → third → ghṛtena yajate
→ or, by this rule, sixth → ghṛtasya yajate
Rule. In the Veda the dvandvas hemantaśiśirau and ahorātre take the gender of their prior member. This overrides paravalliṅgaṃ dvandvatatpuruṣayoḥ (SK 812, chapter 18), which gives the gender of the last member.
Example.
hemantaś ca śiśiraṃ ca → prior member masculine → hemantaśiśirau (ordinary usage: hemantaśiśire)
ahaś ca rātriś ca → the compound-final ac (SK 787) → the i of rātri is lost before ac (Subodhinī) → prior member ahan neuter → ahorātre (ordinary usage: ahorātrau, by SK 814)
The change of ahan to aho- is not stated in this section.
Applies when / blocked by. The Subodhinī says the dual of the sūtra is not binding. So the plural ahorātrāṇi is also correct.
Watch out. Bhaṭṭoji places this sūtra (A 2.4.28) after SK 3398 (A 2.4.39), against Aṣṭādhyāyī order. It is taken here with the other Book 2 gender and case rules. The Subodhinī says Bhaṭṭoji's reversal hints that SK 3433, the exchange of gender, already covers it. The unit ends by quoting adiprabhṛtibhyaḥ śapaḥ (SK 2423), the frame for SK 3400.
ad + luṅ → (ghasḷ; cli lost by 3402; no aṭ) ghastām; han + śap (no luk) → hanati; pra + naś + luṅ (cli lost) → praṇakThe verb chapters taught where a vikaraṇa disappears by luk. In the Veda these losses are loose. ghasḷ may replace ad anywhere. The luk of śap in the ad class may fail, or apply outside it. So may ślu in the hu class. In a mantra, cli of the aorist is lost after certain roots. The running examples are ghastām, hanati and praṇak.
Rule. In the Veda ghasḷ replaces ad "to eat", variously. ghasḷ is carried forward from SK 2427 (chapter 45), which gives it only in luṅ and before san. ad is carried forward from SK 3080 (chapter 67).
Example.
ad + luṅ tām → ghasḷ for ad → ghas + cli + tām
→ cli lost by luk (SK 3402)
→ no augment aṭ (SK 3546, taught later in this chapter)
→ ghastām (ghastān nūnam)
Applies when / blocked by. The Subodhinī objects that ghastām is a luṅ form, already covered by SK 2427. So Bhaṭṭoji adds a second example, sagdhiḥ. The Subodhinī derives it:
ad + ktin → ghasḷ → ghas + ti
→ the penultimate a is lost (SK 3550)
→ the s is lost between consonants
→ t becomes dh; gh becomes g
→ gdhi "eating"
→ sa for samāna (SK 1012, chapter 25) → sagdhiḥ "eating together" (sagdhiś ca me)
Rule. In the Veda the luk of śap after ad and the rest (SK 2423, chapter 45) applies variously. It may fail inside the class. It may apply outside it.
Example.
han + śap + ti → luk does not apply → vṛtraṃ hanati (ordinary usage: hanti)
śī + śap + te → luk does not apply → ahiḥ śayate (ordinary usage: śete); the guṇa of ī to e, and ay for that e, are not stated in this section
trai + śap + dhvam (not an ad-class root) → luk applies → trādhvaṃ no devāḥ (ordinary usage: trāyadhvam)
Watch out. For "here there is no luk" the 1915 print reads luṅ na. The unit ends by quoting juhotyādibhyaḥ śluḥ (SK 2489), the frame for SK 3401.
Rule. In the Veda ślu after hu and the rest (SK 2489, chapter 46) applies variously. ślu is the loss of śap that brings doubling of the root.
Example.
dā + śap + ti (a hu-class root) → no ślu, so no doubling → dāti priyāṇi cid vasu (ordinary usage: dadāti)
vaś + ti (an ad-class root) → ślu applies "elsewhere too" → doubling
→ i for the vowel of the reduplicated syllable (SK 3598)
→ ś becomes ṣ, t becomes ṭ
→ pūrṇāṃ vivaṣṭi
The steps for vivaṣṭi are the Subodhinī's.
Rule. In a mantra, li (that is, cli) is lost by luk after these roots: ghas, hvṛ, ṇaś, vṛ, dah, roots ending in ā, vṛj, kṛ, gam and jan. Normally cli stands in luṅ (SK 2221, chapter 44) and is not lost.
Example.
ghas + cli + jhi → cli lost → the penultimate a is lost (SK 2363) → s becomes ṣ (SK 2410) → akṣan (akṣann amīmadanta hi)
The augment a, the ending an, and k before ṣ are not stated in this section.
pra + naś + cli + t → cli lost → the t is lost → ś becomes k (SK 431) → n becomes ṇ after pra → praṇak (dhūrtiḥ praṇaṅ martyasya)
The loss of t and the ṇ are the Subodhinī's steps.
Other forms: mā hvar; āvaḥ (vṛ); mā na ādhak (dah); ā prāḥ (prā, an ā-final root); parāvark (vṛj); akran (kṛ); anu gman (gam).
Applies when / blocked by. The word "mantra" also stands for the Brāhmaṇa: ajñata vā asya dantāḥ (jan). And "optionally" (vibhāṣā) is carried forward. So the luk may fail: ajaniṣṭa hi ṣaḥ keeps its cli.
Watch out. Outside a mantra the Subodhinī gives the ordinary forms: aghasat, ahvārṣīt, anaśat, avārīt, adhākṣīt, aprāsīt, avarjīt, akārṣīt, agamat, ajani, ajaniṣṭa.
gup + luṅ → (caṅ for cli; reduplication, long abhyāsa) ajūgupatam; kṛ + luṅ → (aṅ for cli; guṇa) akaramBook 3 begins with the vikaraṇas. Chapter 44 taught that cli stands in luṅ (SK 2221) and that caṅ or aṅ replace it after certain roots. The Vedic rules shift these choices. Six ready-made forms add ām with an auxiliary: akaḥ in four, kriyāt and akran in the other two. gup may take caṅ. Four ṇic-final roots refuse caṅ. Four roots may take aṅ. The running examples are ajūgupatam, "you two have guarded", and akaram, "I have made".
Rule. Six forms are taught ready-made (nipātana) in the Veda. Each is optional: "optionally" is carried forward from SK 2465 (chapter 45), the Subodhinī says. So the ordinary forms also stand in the Veda. The first four add ām in luṅ, with akaḥ as the auxiliary verb (anuprayoga).
Example.
abhyut + sad + ṇic → ām in luṅ + akaḥ → abhyutsādayām akaḥ (ordinary usage: abhyudasīṣadat)
pra + jan + ṇic → prajanayām akaḥ (= prājījanat)
ci → ām, reduplication, and k for c → cikayām akaḥ (= acaiṣīt)
ram + ṇic → ramayām akaḥ (= arīramat)
The other two: pāvayāṃ kriyāt, an āśīrliṅ form (ordinary usage: pāvyāt), and vidām akran (= avediṣuḥ).
Applies when / blocked by. The Subodhinī explains akaḥ. It is kṛ in luṅ with tip. Its cli is lost by SK 3402, and the t of tip is lost.
Watch out. For "reduplication" the 1915 print reads dvirvacanam.
Rule. In the Veda, after bare gup "to guard", caṅ optionally replaces cli. "Bare" means without the affix āya, the Subodhinī says.
Example.
gup + cli + thas → tam replaces thas (Subodhinī) → caṅ for cli
→ the root is reduplicated, and the vowel of the reduplicated syllable is long
→ the augment a before the root (not stated in this section)
→ ajūgupatam (gṛhān ajūgupataṃ yuvam "you two have guarded the houses")
Bhaṭṭoji glosses it as agauptam.
Applies when / blocked by. The Subodhinī counts the forms. Ordinary usage has three: agauptam, agopiṣṭam and, with āya, agopāyiṣṭam. The Veda adds the caṅ form, so it has four.
Rule. After ūnaya, dhvanaya, elaya and ardaya, caṅ does not replace cli. This blocks SK 2312 (chapter 44), which would give caṅ after these ṇic-final roots.
Example.
mā + ūnaya + cli + s → no caṅ → no augment aṭ after mā (SK 2228)
→ mā … kāmam ūnayīḥ "do not leave the desire unfulfilled", second person singular (ordinary usage: aunanat)
The steps that give the ending -īḥ are not stated in this section.
mā + dhvanaya + cli + t → no caṅ → mā tvāgnir dhvanayīt (ordinary usage: adidhvanat)
Applies when / blocked by. That the rule holds in the Veda is not stated by Bhaṭṭoji or his commentators; it is inferred. It rests on the Subodhinī's contrast with the ordinary forms aunanat, adidhvanat, ārdidat and aililat.
Rule. In the Veda, after kṛ, mṛ, dṛ and ruh, aṅ optionally replaces cli. The Subodhinī says "cli" comes from SK 2221, "aṅ" from SK 2438 and "optionally" from SK 2269.
Example.
kṛ + cli + mip → aṅ for cli → guṇa of ṛ before aṅ (SK 2406) → the augment a; mip becomes am (these two steps are not stated in this section) → akaram (idaṃ tebhyo 'karaṃ namaḥ "I have made this homage to them")
mṛ → amarat; dṛ → adarat; ruh → yat sānoḥ sānum āruhat
Applies when / blocked by. The ordinary forms are akārṣīt, amṛta, adārīt and arukṣat (Subodhinī). amarat is parasmaipada although mṛ is not. The Subodhinī puts this down to exchange (vyatyaya, SK 3433).
nis + kṛt + ṇyat (for kyap) → (initial and final swapped; ṣ for the s of nis) niṣṭarkyaChapter 66 taught the kṛtya affixes yat, kyap and ṇyat, "to be done". One long Vedic sūtra lists seventeen kṛtya forms taught ready-made. Some take a different affix from the one expected. Some add a lengthening, a substitution or a swap of sounds. The running example is niṣṭarkya, from kṛt "to cut" with nis.
Rule. In the Veda seventeen kṛtya forms are taught ready-made (nipātana). Bhaṭṭoji names the root and affix of each.
Example.
nis + kṛt → kyap would come (SK 2859, chapter 66) → ṇyat comes instead
→ the first and last consonants of the root are swapped
→ ṣ for the s of nis
→ niṣṭarkya (niṣṭarkyaṃ cinvīta paśukāmaḥ)
The guṇa ar and the ṭ after ṣ are not stated by Bhaṭṭoji or his commentators; they are inferred.
The other sixteen.
devahūya: hve or hu with deva as upapada, kyap, and a long vowel (spardhante vā u devahūye).praṇīya, unnīya: nī after pra and ut, with kyap. ucchiṣya: śiṣ after ut, with kyap.marya, staryā, dhvarya: mṛ, stṛ and dhvṛ with yat. staryā occurs only in the feminine.khanya and khānya: khan with yat and with ṇyat.devayajyā: yaj with ya (devayajyāyai).āpṛcchya: pṛcch after āṅ, with kyap. pratiṣīvya: siv with kyap and ṣ.brahmavādya: vad with brahman, and ṇyat. Ordinary usage has kyap and yat here, by SK 2854 (chapter 66).bhāvya, stāvya: bhū and stu with ṇyat.upacāyyapṛḍam: ci after upa, with ṇyat and the substitute āy, when pṛḍa follows.Applies when / blocked by. A vārttika limits upacāyyapṛḍam to gold (hiraṇya). Otherwise the form is upaceyapṛḍam. pṛḍa comes from the root pṛḍ with the affix ka. Bhaṭṭoji says the ka is the one taught for roots with an ik penultimate, that is, SK 2897 (chapter 67).
Watch out. A nipātana fixes the form. Bhaṭṭoji's notes say which affix and which special change are accepted; ordinary rules still do the rest. For example, the Subodhinī gives āpṛcchya its saṃprasāraṇa, and pratiṣīvya its long ī, by general rules.
havya + vah: (ñyuṭ, outside the pāda's middle) havyavāhanaḥ, against (ṇvi, mid-pāda) havyavāḷChapter 67 taught kṛt affixes added to a root when a word stands beside it as upapada. The Vedic rules add new upapada affixes. in comes after four roots, ṇvi after sah and vah, ñyuṭ after vah with certain words, and viṭ after five roots. A pāda is a quarter of a verse. The running example contrasts havyavāhanaḥ and havyavāḷ, both "bearer of the oblations".
Rule. In the Veda, in comes after van, san, rakṣ and math when an object word is the upapada. in is carried forward from SK 2938 (chapter 67), the Subodhinī says.
Example.
brahma vanati → brahman + van + in → brahmavanim (brahmavaniṃ tvā kṣatravanim)
gāṃ sanoti → go + san + in → ṣ for s (SK 1022, chapter 25) → goṣaṇim (uta no goṣaṇiṃ dhiyam)
pathirakṣayaḥ, pathirakṣī "guardians of the path"; havirmathīnām "of the disturbers of the oblation"
The Subodhinī says brahmavanim and goṣaṇim are accusative singulars of the in-final stems. The ṇ of goṣaṇim and the loss of the n of brahman are not explained here.
Applies when / blocked by. The Subodhinī says the roots are van and san of the bhū class. The like-sounding roots of the tan class carry a marker and are not meant.
Rule. In the Veda, ṇvi comes after sah "to overpower" when a word with a case-ending is the upapada. ṇvi is carried forward from SK 2976 (chapter 67), the Subodhinī says.
Example.
pṛtanā + sah + ṇvi → ṣ for the s of sah (SK 335, chapter 11) → pṛtanāṣāṭ "overpowering hostile hosts"
The other steps to -ṣāṭ are not stated in this section: the vṛddhi of the a (the affix is ṇit), the loss of the affix, which leaves nothing, and ṭ for the final h.
Watch out. The Subodhinī asks about classical turāsāham, which this Vedic rule cannot give. It explains that word with vic after the ṇic-final root.
Rule. ṇvi also comes after vah "to carry", with an upapada, in the Veda. Bhaṭṭoji says "as before".
Example.
ditya + vah + ṇvi → dityavāṭ
As with pṛtanāṣāṭ, the vṛddhi of the a, the loss of the affix and ṭ for the final h are not stated in this section.
Applies when / blocked by. Bhaṭṭoji says the rule is split off "for the sake of what follows". That this lets vah alone run on into SK 3411 and SK 3412 is not stated by Bhaṭṭoji or his commentators; it is inferred.
Rule. When kavya, purīṣa or purīṣya is the upapada, ñyuṭ comes after vah in the Veda. Bhaṭṭoji's gloss supplies "after vah" and "in the Veda".
Example.
kavya + vah + ñyuṭ → kavyavāhanaḥ
Likewise purīṣavāhanaḥ and purīṣyavāhanaḥ.
The vṛddhi of vah and the ana of yu are not stated by Bhaṭṭoji here; they are inferred.
Applies when / blocked by. The Subodhinī has no note on this sūtra. That vah comes from SK 3410 and "in the Veda" from SK 3409 is inferred.
Rule. When havya is the upapada, ñyuṭ comes after vah, if the word does not stand inside a pāda. Inside a pāda only ṇvi comes, by SK 3410.
Example.
havya + vah + ñyuṭ (not mid-pāda) → havyavāhanaḥ (agniś ca havyavāhanaḥ)
havya + vah + ṇvi (mid-pāda) → havyavāḷ (havyavāḷ agnir ajaraḥ pitā naḥ)
Watch out. The Subodhinī explains the ḷ of havyavāḷ. By the Prātiśākhya, a ḍ between two vowels is pronounced ḷ. The 1915 print reads the sūtra's last word as one compound, anantaḥpādam.
Rule. In the Veda, viṭ comes after jan, san, khan, kram and gam. Then ā replaces the root's final nasal, by viḍvanor anunāsikasyāt (SK 2982, chapter 67).
Example.
ap + jan + viṭ → ā for the n → abjāḥ "born in the waters"; likewise gojāḥ
go + san + viṭ → ā for the n → ṣ for s (SK 3645, later in this chapter) → goṣāḥ (goṣā indo nṛṣā asi); likewise nṛṣāḥ
bisakhāḥ "digger of lotus-stalks"; dadhikrāḥ; agregāḥ "going in front"
Applies when / blocked by. Bhaṭṭoji does not match each form to its root. Taking them in the sūtra's order is not stated by Bhaṭṭoji or his commentators; it is inferred. In agregāḥ the seventh case of agre is kept, by SK 966 (chapter 24), the Subodhinī says.
śveta + vah: (ḍas where pada status is coming; lengthening, SK 425) śvetavāḥ, against (ṇvin; vṛddhi) śvetavāhauThe upapada rules continue. In a mantra, ṇvin comes in three fixed compounds and after yaj with ava. A vārttika puts ḍas in its place wherever the stem will be a pada. A rule from the Tripādī keeps their long vowel in address too. Then come vic, the three affixes manin, kvanip and vanip, and kvip after han. At SK 3419 Bhaṭṭoji recalls the Vedic kṛt rules already taught in chapter 67. The running example is śveta + vah, "Indra, whom white horses carry".
Rule. In a mantra, ṇvin comes in three ready-made compounds. They are vah with śveta, śaṃs with uktha, and dāś with puras. A vārttika adds: ḍas comes after these instead, where the stem is to be a pada. Bhaṭṭoji explains: "where the status of pada is going to come about, ḍas is an exception to ṇvin."
Example.
śveta + vah + ḍas + su (the stem will be a pada) → śvetavas → the vowel before s is lengthened (SK 425), with ru for the final s → śvetavāḥ
The loss of su is not shown step by step here.
śveta + vah + ṇvin + au → vṛddhi → śvetavāhau; likewise śvetavāhaḥ
uktha + śaṃs + ḍas → the n is lost → ukthaśāḥ "the sacrificer, who recites the hymns"
uktha + śaṃs + ṇvin + au → the n is lost → vṛddhi (SK 2282, chapter 44) → ukthaśāsau, ukthaśāsaḥ
puras + dāś + ḍas → ḍ for the d → puroḍāḥ "what is given in front"
Applies when / blocked by. Bhaṭṭoji gives the division but does not assign each form. Which forms have ḍas and which ṇvin is not stated by Bhaṭṭoji or his commentators; it is inferred. The Subodhinī's two derivations fit it. SK 425 needs an as-final stem, which ḍas gives. SK 2282 needs a ṇ-marked affix, which ṇvin is.
Rule. When ava is the upapada, ṇvin comes after yaj, in a mantra.
Example.
ava + yaj → avayāḥ, avayājau, avayājaḥ
Bhaṭṭoji gives only the forms. That avayāḥ has the ḍas of SK 3414, and the others ṇvin with vṛddhi, is inferred.
Applies when / blocked by. The Subodhinī says this rule is split from SK 3414 for the sake of SK 3417. As one rule, only yaj with ava would have continued. SK 3417 wants plain yaj.
Rule. avayāḥ, śvetavāḥ and puroḍāḥ are taught ready-made with the long vowel even in the vocative singular (sambuddhi). By ca, ukthaśāḥ too.
Example.
śvetavas + vocative su → SK 425 does not apply in address → this rule fixes the long vowel → śvetavāḥ "O Indra, whom white horses carry"
Applies when / blocked by. The Subodhinī raises an objection. Once ḍas is made, SK 425 already gives śvetavāḥ, so the rule seems useless. The answer: SK 425 carries the word asambuddhau, "not in address". So in address it does not lengthen, and this rule is needed.
Watch out. This is a rule of the Tripādī, the last three quarters of book 8. Bhaṭṭoji takes it up here, beside the forms of SK 3414–3415.
Rule. When upa is the upapada, vic comes after yaj in the Veda. yaj continues from SK 3415.
Example.
upa + yaj + vic → upayaṭ
Applies when / blocked by. The Subodhinī asks why chandasi is stated, since mantre still continues. Its answer: chandas takes in the Brāhmaṇa too, which "mantra" alone would not. It quotes a line: "the name mantra for those that prompt the act; the word brāhmaṇa for the rest".
Rule. In the Veda, after roots ending in ā, three affixes come: manin, kvanip and vanip. A word with a case-ending, or an upasarga, must be the upapada. By ca, vic comes as well.
Example.
su + dā + manin → sudāmā
su + dhā + kvanip → ī for ā (SK 2462, chapter 45) → sudhīvā; likewise supīvā
bhūri + dā + vanip → bhūridāvā; likewise ghṛtapāvā
kīlāla + pā + vic → kīlālapāḥ "drinker of water"
Bhaṭṭoji does not match forms to affixes. The matching is inferred, from the ī that SK 2462 gives before a k-marked affix.
Watch out. The unit ends by recalling brahmabhrūṇavṛtreṣu kvip (SK 2998, chapter 67). That rule gives kvip after han only with brahman, bhrūṇa or vṛtra. SK 3419 relaxes it.
Rule. In the Veda, kvip comes after han variously, even with other upapadas. This extends SK 2998 (chapter 67), which allows only three upapadas.
Example.
mātṛ + han + kvip → mātṛhā "mother-slayer"; likewise pitṛhā
The recap of the Vedic kṛt rules. Bhaṭṭoji then runs through Vedic rules already taught in chapter 67, with their Vedic examples:
ahaṃ dyāvāpṛthivī ā tatāna "I have spread out heaven and earth". Its vṛddhi comes by SK 2283 (Subodhinī).cakrāṇā vṛṣṇi.rā + kvasu → doubling, as in liṭ (not stated in this section) → iṭ (SK 3096) → with nañ "not" → ararivān "ungenerous". Before aghāyuḥ, the n becomes ru (SK 3631, later in this chapter) and the ā before it is nasalized (SK 3632) → ararivāṁ aghāyuḥ (yo no agne ararivāṁ aghāyuḥ). The derivation is the Subodhinī's.vīrudhaḥ pārayiṣṇavaḥ. bhuvaś ca (SK 3118): the same after bare bhū, bhaviṣṇuḥ.agha + kyac "wishes harm" → final a becomes ā (SK 3590) → aghāya → the affix u by kyāc chandasi (SK 3150) → aghāyuḥ.jave yābhir yūnaḥ, ūrvor me javaḥ, devasya savituḥ save.Watch out. The Subodhinī says ac and ap differ only in accent (svare bhedaḥ). Bhaṭṭoji marks no accent here, so none is shown.
vṛṣ + ktin (udātta) → vṛṣṭí: vṛṣṭíṃ diváḥ; su + upa + sad + yuc → sūpasadanaḥChapter 69 taught ktin for feminine action nouns (SK 3272) and khal for "easy" and "hard" (SK 3305). In a mantra, ktin after eight roots carries the udātta. In the Veda, yuc replaces khal after roots of motion, and after other roots too. The running examples are vṛṣṭí "rain", with the udātta on its last syllable, and sūpasadanaḥ "easy to approach".
Rule. In a mantra, ktin comes after vṛṣ, iṣ, pac, man, vid, bhū, vī and rā, and it is udātta. ktin is carried forward from SK 3272 (chapter 69), the Subodhinī says. So the high tone falls on the syllable of the affix, ti or its inflected form.
Example.
vṛṣ + ktin → affix udātta → vṛṣṭí → vṛṣṭíṃ diváḥ "the rain of heaven": udātta on ṭí.
iṣ + ktin → iṣṭí → sumnám iṣṭáye: udātta on ṭá of the dative iṣṭáye.
pac + ktin → paktí → pácāt paktī́r utá: udātta on tī́.
man + ktin → the nasal is lost (SK 2428, chapter 45) → matí → iyáṃ te návyasī matíḥ: udātta on tí.
bhū + ktin → bhūtíḥ: udātta on tí.
vī + ktin → vītí → ágna ā́ yāhi vītáye "Agni, come for the meal": udātta on tá.
rā + ktin → rātí → rātaú syāmobháyāsaḥ "may we both be in your giving": udātta on taú.
vid + ktin → vittiḥ "being". vittiḥ is cited without accent marks, so it is left unmarked.
Applies when / blocked by. The Subodhinī glosses the roots. vṛṣ is "to sprinkle", iṣ "to wish", man "to know", vid "to be", rā "to give". ktin is k-marked, so vid takes no guṇa.
Watch out. Other accents in these phrases, such as on diváḥ, belong to the other words. The rule concerns only the ktin syllable. The word udāttaḥ of this sūtra still reaches ūtiyūti… (A 3.3.97) and the next sūtras in Aṣṭādhyāyī order. Bhaṭṭoji taught those earlier, as SK 3274–3276 (chapter 69). For syāmobháyāsaḥ (syāma + ubhayāsaḥ) the 1915 print reads the words run together.
Rule. In the Veda, after roots meaning motion, yuc comes when īṣat, dus or su is the upapada. It is an exception to khal (SK 3305, chapter 69).
Example.
su + upa + sad + yuc (ana) → sūpasadanaḥ → sūpasadano 'gniḥ "Agni, easy to approach"
This analysis is not given by Bhaṭṭoji or his commentators; it is inferred.
Applies when / blocked by. Bhaṭṭoji names the upapadas only as "īṣat and the rest". That they are those of SK 3305, and that yuc is the affix of SK 3309, is inferred. The Subodhinī has no note on this sūtra.
Rule. yuc is seen in the Veda after other roots too, not only those meaning motion.
Example.
su + vid + yuc (ana) → feminine → suvedanā → suvedanām akṛṇor brahmaṇe gām "you made the cow easy to find for the brahman"
This analysis is inferred; Bhaṭṭoji gives only the mantra.
Applies when / blocked by. The Subodhinī has no note on this sūtra.
tṛ + leṭ tip → (i lost, 3426; sip, ṇit; iṭ; vṛddhi; aṭ) tāriṣat; pat + leṭ tip → (āṭ; i kept) patāti; mādaya + leṭ ātām → (ṭi → e; ā → ai, 3429) mādayaiteChapters 44 and 65 tied each lakāra to its time and sense. In the Veda luṅ, laṅ and liṭ may stand for any time. A further lakāra, leṭ, is used in the senses of liṅ. The next rules build its forms. It takes the vikaraṇa sip, an augment aṭ or āṭ, and endings that drop i or s or change ā and e to ai. All these rules hold in the Veda (chandasi, SK 3423). Bhaṭṭoji repeats the word in some glosses (SK 3424) and leaves it out in others. The running examples are tāriṣat "may he lengthen", patāti "let it fall" and mādayaite "the two shall delight".
Rule. In the Veda, when the senses of verbs are related, luṅ, laṅ and liṭ optionally come in any time. In the other option, each time takes its own lakāra, as taught in chapters 44 and 65.
Example.
luṅ: devo devebhir āgamat (gam takes aṅ by SK 2343).
luṅ in the sense of loṭ: idaṃ tebhyo 'karaṃ namaḥ "let me make this homage to them" (akaram by SK 3406).
laṅ: agnim adya hotāram avṛṇītāyaṃ yajamānaḥ "today this sacrificer … Agni as hotṛ".
liṭ: adyā mamāra, which means adya mriyate "today he dies".
Applies when / blocked by. The condition "when the senses of verbs are related" echoes the heading dhātusambandhe pratyayāḥ (SK 2824, chapter 65). That link is not stated by Bhaṭṭoji or his commentators; it is inferred. The Subodhinī admits the relation in these examples "is to be looked for".
Watch out. The verbs āgamat and avṛṇīta are left untranslated on purpose. The point of the rule is that their form does not fix their time.
Rule. In the Veda leṭ comes after a root in the senses of liṅ. These are injunction (vidhi) and the rest, and the relation of cause and effect and the rest.
Example. Bhaṭṭoji gives no example here. The forms of leṭ come under SK 3426–3431.
Applies when / blocked by. "Injunction and the rest" points to SK 2208 (chapter 44). "Cause and effect" points to SK 2813 (chapter 65). These links are inferred. The Subodhinī has no note on this sūtra.
Rule. Before leṭ, the vikaraṇa sip comes variously.
Example. The sūtra stands alone, with no comment by Bhaṭṭoji and no note in the Subodhinī. Its forms, tāriṣat, joṣiṣat and sāviṣat, come under SK 3427.
Rule. The i of the parasmaipada endings of leṭ is optionally lost (lopa). The Subodhinī says "of leṭ" comes from SK 3427 and "optionally" from SK 3430.
Example.
tip (ti) of leṭ → i lost → t → as in tāriṣat (SK 3427)
Watch out. Bhaṭṭoji places this rule before the two rules it borrows from. So its words come up from below, in the order of the Aṣṭādhyāyī. Being optional, the i may also stay: patāti (SK 3427).
Rule. leṭ takes the augment aṭ or āṭ. They come by turns, not together, the Subodhinī says. Both are pit: "pit" is carried forward by a "frog's leap" (maṇḍūkapluti) from SK 2204 (chapter 44), skipping the rules between.
Example.
tṛ + sip + tip → i of tip lost (SK 3426)
→ sip is ṇit by the vārttika below → vṛddhi of ṛ (SK 254) → tār
→ the augment iṭ; the s of sip becomes ṣ after i (not stated in this section) → tār-i-ṣ
→ the augment aṭ → tāriṣat (pra ṇa āyūṃṣi tāriṣat "may he lengthen our life-spans")
Likewise joṣiṣad dhi and ā sāviṣat.
pat + āṭ + ti → no sip, and the i kept → patāti (patāti didyut "let the missile fall"); likewise bhavāti.
Applies when / blocked by. A vārttika says sip is variously ṇit. That is what brings vṛddhi in tāriṣat. The Subodhinī gives the steps for tṛ: loss of the i of tip, sip, iṭ. The rule for iṭ is not named.
Watch out. For the vārttika the 1915 print reads sibbahulaṃ ṇit, and for the mantra patāti, with a short first a.
Rule. The s of the uttama endings of leṭ is optionally lost. The uttama endings are those used for "I" and "we".
Example.
kṛ + u + āṭ + vas → kṛ takes guṇa with r, kar
→ āṭ is pit, so not ṅit, and u takes guṇa, o
→ that o becomes av before ā (not stated in this section) → karavā + vas
→ s lost → karavāva; or kept → karavāvaḥ
The steps are the Subodhinī's. u comes by SK 2466 (chapter 45).
Watch out. Bhaṭṭoji ends with "ṭer etvam", ṭi becomes e (SK 2233). This is the first step of the next example, mādayaite.
Rule. The ā of the endings of leṭ becomes ai.
Example.
mad + ṇic → mādaya + ātām
→ the ṭi, the last vowel with what follows, becomes e (SK 2233) → mādaya + āte
→ the first ā becomes ai → mādayaite (sutebhiḥ suprayasā mādayaite)
Applies when / blocked by. The Subodhinī says only the first ā of ātām is meant. The second has already become e by SK 2233.
Watch out. The augment āṭ does not become ai. If it did, Pāṇini would have taught aiṭ directly. The rule's very wording rules it out (vidhisāmarthya). So: yo yajāti yajāta it, with ā kept.
Rule. Elsewhere, the e of leṭ optionally becomes ai. "Elsewhere" means outside the domain of SK 3429, the Subodhinī says.
Example.
īś + iṭ (ātmanepada, "I") → no loss of i, since SK 3426 is for parasmaipada
→ ṭi becomes e (SK 2233) → īśe
→ e becomes ai → īśai (paśūnām īśai "I shall be lord of the animals")
Likewise grahā gṛhyāntai "the cups shall be taken".
These steps are the Subodhinī's.
Applies when / blocked by. Bhaṭṭoji asks: why "elsewhere"? His answer is mādayaite. There the ā has become ai by SK 3429. The final e stays e, because it falls in the domain of SK 3429. This explanation is not stated by Bhaṭṭoji or his commentators; it is inferred.
Watch out. The sūtra itself says only "optionally, of e, elsewhere". That "ai" and "leṭ" are carried in from SK 3429 and SK 3424 is inferred.
Rule. leṭ also comes in the sense of a bargain (paṇabandha) and of apprehension (āśaṅkā). The Subodhinī defines the bargain: "if you do this for me, I will give you this".
Example.
bargain: aham eva paśūnām īśai "I alone shall be lord of the creatures"
apprehension: nej jihmā yantau narakaṃ patāma "lest, going crookedly, we fall into hell"
In patāma the s of mas is lost by SK 3428, the Subodhinī says.
Watch out. The dual yantau stands beside the plural patāma; the sense is plural. The unit ends by quoting halaḥ śnaḥ śānaj jhau (SK 2557, chapter 52), the frame for SK 3432.
grah + śnā + hi → (śnā → śāyac; saṃprasāraṇa; h → bh) gṛbhāya; bhid + śap (for śnam) → bhedati; vac + aṅ + āśīrliṅ → (um, SK 2454) vocemaThree sūtras from the vikaraṇa section of book 3 (A 3.1.84–86) come in here, then A 3.4.117. śāyac may replace śnā. The vikaraṇas may be exchanged (vyatyaya); a verse widens this to eleven things. aṅ comes before the benedictive. And any affix after a root may count as both sārvadhātuka and ārdhadhātuka. The running examples are gṛbhāya "seize", bhedati "he splits" and vocema "may we speak".
Rule. In the Veda śāyac also replaces śnā before hi. This extends SK 2557 (chapter 52), which gives śānac there after a consonant. By api "also", śānac comes too.
Example.
grah + śnā + hi → śāyac (āya) for śnā
→ r becomes ṛ (saṃprasāraṇa) → bh for h
→ hi is lost, as in badhāna → gṛbhāya (gṛbhāya jihvayā madhu "seize the honey with the tongue")
bandh + śnā + hi → śānac (āna) → the nasal lost (SK 415) → hi lost → badhāna (badhāna deva savitaḥ "bind, O god Savitṛ")
Applies when / blocked by. A vārttika, hṛgrahor bhaś chandasi, gives bh for the h of hṛ and grah in the Veda. Its examples are gṛbhṇāmi te "I take" and madhvā jabhāra "he carried".
Watch out. Bhaṭṭoji does not label gṛbhāya and badhāna. That the first shows śāyac and the second śānac is inferred. The saṃprasāraṇa in gṛbhāya and the loss of hi in badhāna are the Subodhinī's steps.
Rule. In the Veda the vikaraṇas are exchanged (vyatyaya) variously. One vikaraṇa comes where another was due, or two or three come together. The vikaraṇas were taught in chapters 44–52, such as śap (SK 2167).
Example.
bhid (rudh class) → śap where śnam was due → bhedati (āṇḍā śuṣṇasya bhedati), for bhinatti
mṛ → jarasā marate patiḥ, for mriyate. The Subodhinī derives mriyate with the tud-class vikaraṇa śa; that marate has śap in its place is inferred.
nī + loṭ → two vikaraṇas, śap and sip → indro vastena neṣatu
tṝ + liṅ → three vikaraṇas, u, śap and sip → taruṣema, meaning tarema
The verse. Bhaṭṭoji quotes a verse. Pāṇini wants exchange of eleven things, and bahulam achieves it:
dhuri dakṣiṇāyāḥ, for dakṣiṇasyām.ye … takṣati, for takṣanti.icchate for icchati; yudhyati for yudhyate.madhos tṛptāḥ, for madhunaḥ.viyūyāḥ, for viyūyāt.śvo 'gnīn ādhāsyamānena, lṛṭ where luṭ was due.adukṣat, for adhukṣat.mitra vayaṃ ca sūrayaḥ, for mitrā.annādāya has ac where aṇ was due.Applies when / blocked by. In annādāya the difference shows in the word-division (avagraha) of the pada-text. With aṇ it would be anna-ādāya; with ac it is anna-adāya (Subodhinī). yaṅ is a pratyāhāra made with the ṅ of SK 3434, beginning from the syllable ya (Bhaṭṭoji does not say which).
Watch out. Bhaṭṭoji himself names only upagraha, nara, kāla, kartṛ and yaṅ. Assigning the other examples by the order of the verse is inferred; the Subodhinī confirms ac for mitra. For viyūyāt the 1915 print reads viyūyād.
Rule. In the Veda aṅ comes after a root before the benedictive liṅ (āśīrliṅ). Normally no vikaraṇa comes there, since SK 2215 (chapter 44) makes āśīrliṅ ārdhadhātuka.
Example.
vac + aṅ + mas (āśīrliṅ) → u by vaca um (SK 2454, chapter 45) → voc- → vocema (mantraṃ vocema agnaye "may we speak a hymn to Agni")
Applies when / blocked by. A vārttika gives ak after dṛś instead. With aṅ, dṛś would take guṇa by SK 2406 (chapter 44). That rule is conditioned by aṅ only, so ak leaves the ṛ: pitaraṃ ca dṛśeyaṃ mātaraṃ ca "may I see father and mother".
Watch out. The endings of vocema and dṛśeyam need yās to become iy. The Subodhinī explains this under SK 3435: liṅ counts as sārvadhātuka too. Connecting that remark with these forms is inferred.
Rule. In the Veda an affix taught under the heading "after a root" bears both names, sārvadhātuka and ārdhadhātuka. Normally tiṅ and ś-marked affixes are sārvadhātuka (SK 2166) and the rest ārdhadhātuka (SK 2187, chapter 44).
Example.
vardhaya + antu → the ending counts as ārdhadhātuka → ṇi is lost → vardhantu tvā suṣṭutayaḥ, meaning vardhayantu "may the praises make you grow"
vi + śru + ire → the ending counts as sārvadhātuka → śnu and śṛ for śru → yaṇ (SK 2387) → viśṛṇvire
Bhaṭṭoji names neither the root śru nor the rules for ṇi-loss and for śnu and śṛ; they are inferred.
The quotation of SK 3151. Bhaṭṭoji here quotes ādṛgamahanajanaḥ kikinau liṭ ca (SK 3151, chapter 67). After ā-final roots, roots ending in ṛ or ṝ, gam, han and jan, ki and kin come, treated like liṭ:
bhṛ + ki → doubling, as in liṭ (the reduplicated syllable is ba) → babhrir vajram "bearer of the thunderbolt"
Likewise papiḥ somam, dadir gāḥ, jagmir yuvā, jaghnir vṛtram, jajñiḥ.
tṝ + ki → doubling → u for the ṛ (SK 3578, later in this chapter) → taturiḥ; likewise jaguriḥ.
Watch out. Being treated like liṭ already makes ki and kin kit. The k-marker is still added, Bhaṭṭoji says, to block the guṇa of SK 2383. The Subodhinī adds that liṅ, too, is sārvadhātuka under this rule. So yās becomes iy, and y is lost before a consonant.
jīv + ase → jīváse (śarádo jīváse dhāḥ, udātta on va); pra + yā + kai → prayai = prayātumChapter 69 taught tumun for purpose (SK 3175): draṣṭum "to see". In the Veda fifteen other affixes carry the same sense (tumartha). Some of them fix the accent. Then five forms are taught ready-made. The running examples are jīváse "to live", with the udātta on va, and prayai, which means prayātum "to set forth".
Rule. In the sense of tumun, fifteen affixes come in the Veda. They are se, sen, ase, asen, kse, kasen, adhyai, adhyain, kadhyai, kadhyain, śadhyai, śadhyain, tavai, taveṅ and taven. The Subodhinī says the sense of tumun is the bare action (bhāva).
Example.
vac + se → k for c, ṣ for s → vakṣe rāyaḥ "to speak of riches".i + sen → guṇa → tā vām eṣe "to go to you two".pra + i + kse → kse is kit, so no guṇa → preṣe "to go forth".jīv + ase → śarádo jīváse dhāḥ "grant us to live many autumns". The udātta is on va, the first syllable of the affix.gávām iva śriyáse "to shine like the cows". The udātta is on ya.jaṭháram pṛṇádhyai "to fill the belly". With adhyai the udātta is on ṇa. With adhyain the word is ādyudātta instead.āhuvádhyai "to invoke". With kadhyai the udātta is on va. With kadhyain the word takes the accent of a nit affix. Taking the one example for both affixes of each pair is inferred.rādhasaḥ saha mādayadhyai "to delight together with bounty". śadhyain: vāyave pibadhyai "for Vāyu, to drink".dātavā u (dātavai + u) "to give". taveṅ: sūtave "to bring forth", with no guṇa because the affix is ṅit. taven: kartave, meaning kartum "to do".Applies when / blocked by. The Subodhinī says the n of sen is there to make the word ādyudātta. By the n of kasen, śriyáse too should be ādyudātta. The mantra recites it with the udātta on the middle syllable instead. The Subodhinī calls this the exchange of accent "by bāhulaka", which SK 3433 promised.
Watch out. In mādayadhyai and pibadhyai, śap comes by exchange where yak was due (Subodhinī). The first ā of āhuvádhyai is unmarked. From the quotation alone it may be udātta, or the unmarked level that follows a svarita earlier in the mantra; this is left open. Bhaṭṭoji's remark concerns only the affix.
Rule. prayai, rohiṣyai and avyathiṣyai are taught ready-made in the sense of tumun.
Example.
pra + yā + kai → prayai = prayātum "to set forth"
ruh + iṣyai → rohiṣyai = roḍhum "to climb"
nañ + vyath + iṣyai → avyathiṣyai = avyathitum "to be unshaken"
The affixes are the Subodhinī's; Bhaṭṭoji gives only the meanings.
Rule. dṛśe and vikhye are also taught ready-made in the sense of tumun. Bhaṭṭoji glosses them as draṣṭum "to see" and vikhyātum "to look out upon".
Example.
dṛś + ke → ke is kit, so no guṇa → dṛśe
vi + khyā + ke → the ā of khyā is lost → vikhye
These steps are the Subodhinī's.
Applies when / blocked by. That "taught ready-made in the sense of tumun" continues from SK 3437 and SK 3436 is inferred. The Subodhinī finds the purpose of making this a separate sūtra "open to question".
Watch out. Part I ends here. The next sūtra, śaki ṇamulkamulau (SK 3439), continues both tumarthe and chandasi.
vi + bhaj + ṇamul → (ṇit: vṛddhi of the penultimate a) vibhājam: vibhājaṃ nāśakat; vi + car + iṭ + tosun → vicaritoḥ: īśvaro vicaritoḥPart I ended with the Vedic affixes that stand in the sense of tumun, the infinitive affix (SK 3436). Six more sūtras finish the third book of the Aṣṭādhyāyī. Four of them give further infinitive affixes after fixed roots or beside fixed upapadas. A kṛt is an affix added directly to a root, other than a verb ending. An upapada is a word that a kṛt rule names, which must stand beside the root (SK 781, chapter 18). Two sūtras give affixes in the sense of the kṛtya affixes, the "to be done" affixes of chapter 66, such as tavyat (SK 2834).
None of these six sūtras contains the word chandasi, "in the Veda". The Subodhinī supplies it throughout. That it is carried from chandasi luṅlaṅliṭaḥ (SK 3423) is not stated by Bhaṭṭoji or his commentators; it is inferred. The phrase tumarthe, "in the sense of tumun", is carried from SK 3436.
The running examples are two infinitives. In the Veda, śak "to be able" as upapada may take ṇamul. Bhaṭṭoji glosses the result with the ordinary infinitive in tumun, vibhaktum. For comparison only: the ordinary rule for tumun beside śak is SK 3177 (chapter 69); Bhaṭṭoji does not cite it here. The derivation:
vi + bhaj + ṇamul → the markers ṇ, u, l go → vi + bhaj + am
→ ṇamul is marked with ṇ, so the penultimate a takes vṛddhi (ata upadhāyāḥ, SK 2282, chapter 44) → vibhājam
→ vibhājaṃ nāśakat "he could not divide".
In the same way īśvara "able" as upapada takes tosun:
vi + car + tosun → the augment iṭ → vicaritoḥ → īśvaro vicaritoḥ "able to range about".
Rule. When śak "to be able" is the upapada, ṇamul and kamul come after a root, in the sense of tumun, in the Veda.
Example.
vi + bhaj + ṇamul → vṛddhi of the penultimate a, because ṇamul is ṇ-marked (SK 2282) → vibhājam.
vibhājaṃ nāśakat "he could not divide". Bhaṭṭoji glosses it with the ordinary infinitive vibhaktum.
apa + lup + kamul → no guṇa, because kamul is k-marked → apalupam.
apalupaṃ nāśakat "he could not cut off" = apaloptum.
The vṛddhi and the absence of guṇa are explained by the Subodhinī. The Subodhinī glosses lup as "to cut".
Applies when / blocked by. The sense tumarthe comes from SK 3436, as Bhaṭṭoji's own gloss shows. As a cross-reference only, not cited by Bhaṭṭoji here: compare the ordinary tumun beside śak (SK 3177, chapter 69).
Watch out. Both affixes leave am. They differ only in their markers: ṇ of ṇamul brings vṛddhi, k of kamul blocks guṇa.
Rule. When īśvara "able" is the upapada, tosun and kasun come after a root, in the sense of tumun, in the Veda. The Subodhinī supplies "after a root, in the Veda".
Example.
vi + car + tosun → the augment iṭ (the Subodhinī: "with iṭ") → vicaritoḥ.
īśvaro vicaritoḥ "able to range about" = vicaritum.
vi + likh + kasun → no guṇa, because kasun is k-marked → vilikhaḥ.
īśvaro vilikhaḥ "able to scratch" = vilekhitum.
Applies when / blocked by. The Subodhinī adds the word class. Forms in tosun and kasun are avyayas, indeclinable words, by ktvātosunkasunaḥ (SK 450, chapter 14). So their case ending is lost by luk (avyayād āpsupaḥ, SK 452). luk is the loss of an affix without trace.
Watch out. The Subodhinī notes one more consequence of that luk. The a before the final s of vilikhas is not lengthened, as it would be in an ordinary as-stem. The prohibition na lumatā… stops it.
Rule. In the sense of the kṛtya affixes, the Veda uses tavai, ken, kenya and tvan. The Subodhinī explains the sense: the kṛtyas stand for the bare action (bhāva) and the object (karman), by tayor eva… (SK 2833, chapter 66). These four affixes come there.
Example. Bhaṭṭoji does not label the examples. They match the affixes in the sūtra's order, and the Subodhinī's analyses agree.
mlecch + iṭ + tavai → na mlecchitavai "one is not to speak barbarously" = na mlecchitavyam.
ava + gāh + ken → avagāhe "to be plunged into".
didṛkṣa + kenya → the final a of the desiderative stem is lost → didṛkṣeṇyaḥ "to be desired to be seen".
kṛ + tvan → kartvam "to be done" = kṛtyam, in bhūry aspaṣṭa kartvam.
Applies when / blocked by. tavai was already taught in the sense of tumun (SK 3436). The Subodhinī says it is taught again so that it may also stand for the object, which differs from the bare action.
Watch out. The 1915 print marks a svarita on the final syllable of didṛkṣeṇyaḥ. The rule itself says nothing about accent.
Rule. The form avacakṣe is a nipātana: a finished form stated in the sūtra, whose irregular steps are simply accepted. It stands in the kṛtya sense, in the Veda.
Example. ripuṇā nāvacakṣe "not to be looked down upon by the enemy". Bhaṭṭoji glosses it as avakhyātavyam, a kṛtya form.
Applies when / blocked by. The Subodhinī analyses the form as ava + cakṣ + eś. The affix eś is marked with ś, so it counts as sārvadhātuka, a "full-verb" affix. The substitute khyāñ replaces cakṣ only before an ārdhadhātuka affix. So khyāñ does not come, and the root keeps its form cakṣ.
Watch out. kṛtyārthe from SK 3441 still holds here. The Subodhinī says it ceases at the next sūtra.
Rule. tosun comes after eight roots: sthā, iṇ, kṛñ, vad, car, hu, tam and jan. The root must denote the point that marks off another action (bhāvalakṣaṇa). The sense is that of tumun, in the Veda.
Example.
ā + sam + sthā + tosun → āsaṃsthātoḥ.
āsaṃsthātoḥ sīdanti "they sit until the completion" = āsamāpteḥ sīdanti.
ud + i + tosun → udetoḥ "(of) the rising".
apa + kṛ + tosun → apakartoḥ.
The 1915 print adds four more: pravaditoḥ, pracaritoḥ, hotoḥ, ātamitoḥ.
kāmam ā vijanitoḥ saṃbhavāmaḥ "we come together at will until the giving birth".
Bhaṭṭoji closes: "so the Veda (śruti)".
Applies when / blocked by. The Subodhinī says kṛtyārthe has ceased here, and tumarthe continues. The completion, the rising and the rest mark the main action by being its limit (avadhi). The Subodhinī adds that sthā with sam is fixed in the sense "completion".
Watch out. udetoḥ stands bare, without the ā of āsaṃsthātoḥ. So it means only "(of) the rising", not "until the rising". Neither Bhaṭṭoji nor the Subodhinī supplies a governing word.
Rule. After sṛp "to go (creep)" and tṛd "to injure (pierce)", kasun comes in the sense of tumun, in the Veda. The root must again denote the point that marks off another action. Bhaṭṭoji says bhāvalakṣaṇe holds "all the same" (ity eva).
Example.
vi + sṛp + kasun → visṛpaḥ.
purā krūrasya visṛpo virapśin "before the creeping-forth of the cruel one, O mighty one".
ā + tṛd + kasun → ātṛdaḥ.
purā jatrubhya ātṛdaḥ "before the piercing of the collarbones".
The Subodhinī glosses visṛpaḥ as "from going": it is construed with purā, "before".
Applies when / blocked by. kasun is k-marked, as in SK 3440, so the root takes no guṇa. That this explains visṛpaḥ is not stated by Bhaṭṭoji or his commentators; it is inferred.
Watch out. The third book of the Aṣṭādhyāyī ends here. Bhaṭṭoji marks it: "Here ends the third adhyāya."
rātri + ṅīp → rā́trī (Veda; udātta on rā), against ordinary rātri + ṅīṣ → rātrī́ (udātta on the final ī); vi + bhu + ṅīṣ → (yaṇ) vibhvīBook 4 opens with the feminine affixes. Chapter 15 taught them for ordinary Sanskrit (SK 453 onward). Five sūtras and three vārttikas give the Vedic exceptions. A vārttika is a supplementary rule added to a sūtra, which Bhaṭṭoji quotes. The 1915 print begins the section with the heading "the fourth adhyāya".
Two feminine affixes leave the same ī: ṅīp and ṅīṣ. They differ in accent. The udātta is the high accent of a syllable, the anudātta the low one, and the svarita a combination of the two. In this chapter the udātta is written with an acute (á), an independent svarita with a grave (à), and an anudātta is left unmarked.
The running example is rātri "night". In the Veda it takes ṅīp, and the udātta stays on the first syllable, rā: rā́trī. In ordinary Sanskrit it takes ṅīṣ, and the udātta falls on the final ī: rātrī́.
Rule. After rātri, ṅīp comes in the Veda, except before the plural ending jas.
Example.
rātri + ṅīp → rā́trī, with the udātta on rā.
rā́trī vy àkhyad āyatī́ "Night, approaching, has looked forth".
In this line rā of rā́trī is udātta, vy à carries an independent svarita, and the final tī of āyatī́ is udātta.
Applies when / blocked by. Bhaṭṭoji: "in ordinary language, however," the gaṇasūtra kṛdikārāt gives ṅīṣ. Then the word is antodātta, with the udātta on its final syllable: rātrī́. A gaṇasūtra is a rule inside a word-list. The Subodhinī explains ajasau as a seventh-case form, "when not jas". The i in it is only for pronunciation.
Watch out. The Subodhinī explains Bhaṭṭoji's words loke tu, "in ordinary language, however". Without them the classical rātryaḥ would be impossible, and so would rātryaḥ in the Veda. loke gives ṅīṣ in ordinary language. tu, which gathers what the sūtra leaves out, adds ṅīṣ in the Veda before jas: rātryaḥ.
Rule. In the Veda, after bahu and the rest, ṅīṣ is obligatory. bahvādibhyaś ca (SK 503, chapter 15) gives the list, and anyato ṅīṣ (SK 497) gives the affix, as the Subodhinī says.
Example.
bahu + ṅīṣ → u becomes v before the vowel (SK 47) → bahvī.
bahvīṣu hi tvā "for … you among many".
Applies when / blocked by. Bhaṭṭoji: "the word nitya is stated for the sake of the following rule". The Subodhinī explains. The rule would be obligatory merely by being taught (ārambhasāmarthya). So nitya "obligatorily" adds nothing here, and it is meant to carry into SK 3447.
Rule. After bhu, ṅīṣ comes obligatorily in the Veda. The bhu meant is the one formed with the affix ḍu by viprasaṃbhyo ḍv asaṃjñāyām (SK 3160, chapter 67).
Example.
vi + bhu + ṅīṣ → u becomes v (SK 47) → vibhvī.
Likewise prabhvī.
Applies when / blocked by. Bhaṭṭoji gives the reason: utaḥ "after a short u" is carried forward. The Subodhinī names its source, voto guṇavacanāt (SK 502). It notes that svayaṃbhū is therefore excluded. nityam is carried from SK 3446. Without it, the ṅīṣ here could be read as optional, like that of SK 502.
Watch out. The sūtra writes bhuvaḥ, with uv. For a ḍu stem one would expect bhoḥ, by guṇa before the ending (gher ṅiti, SK 245). Bhaṭṭoji calls the uv a licence of the sūtra (sautra).
Vārttika. mudgalāc chandasi lic ca: after mudgala, in the Veda, ṅīṣ comes and it is l-marked (lit). Bhaṭṭoji: the ṅīṣ is lit, and the augment ānuk comes before it. The Subodhinī says the vārttika is read under indravaruṇa… (SK 505, chapter 15).
mudgala + ānuk + ṅīṣ → mudgalānī.
By liti (SK 3676, chapter 71), the syllable before a lit affix is udātta. That syllable is the ā of ānuk: mudgalā́nī, with the udātta on lā.
rathī́r abhūn mudgalā́nī "Mudgala's wife became the charioteer". Here thī of rathī́r and lā of mudgalā́nī are udātta.
The 1915 print reads littvam ānuk, "lit-hood, and ānuk".
Rule. In the Veda the form dīrghajihvī is a nipātana.
Example.
dīrgha-jihvā + ṅīṣ → dīrghajihvī.
āsurī vai dīrghajihvī devānāṃ yajñavāṭ "Dīrghajihvī, an asura woman, truly [was] the carrier-off of the gods' sacrifice".
Here dīrghajihvī is the name, and āsurī "an asura woman" is the common noun.
Applies when / blocked by. Bhaṭṭoji: ṅīṣ is not otherwise obtained, because the word has a conjunct consonant before its final (hv before ā). The Subodhinī names the blocking rule. svāṅgāc copasarjanād asaṃyogopadhāt (SK 510, chapter 15) gives ṅīṣ to body-part compounds only without such a conjunct. The 1915 print reads aprāpto ṅīṣ, "ṅīṣ, not obtained".
Rule. In the Veda, after kadru and kamaṇḍalu, the feminine affix ūṅ comes. ūṅ is carried from ūṅ utaḥ (SK 521, chapter 15).
Example. kadru + ūṅ → kadrū; kamaṇḍalu + ūṅ → kamaṇḍalū: kadrūś ca vai kamaṇḍalūḥ.
Vārttika. ūṅ is also to be stated for guggulu, madhu, jatu and patayālu: guggulūḥ, madhūḥ, jatūḥ, patayālūḥ. The Subodhinī says their feminine gender is a Vedic interchange (vyatyaya, see SK 3433).
Vārttika. Bhaṭṭoji then recalls avyayāt tyap (SK 1324, chapter 29), which gives tyap after an indeclinable. A second vārttika adds: in the Veda, āviṣṭya too.
āvis + tyap → s becomes ṣ (hrasvāt tādau taddhite, SK 1325) → t after ṣ becomes ṭ (ṣṭutva, SK 113) → āviṣṭya.
āviṣṭyo vardhate "the manifest one grows".
Watch out. The Subodhinī explains why the vārttika is needed. Under SK 1324 tyap is restricted by amehakvatasitrebhyaḥ to a short list of indeclinables. So āvis would not get it.
varṣā + ṭhañ → (ṭha → ika, SK 1170; ādivṛddhi) vā́rṣikam, ādyudātta with udātta on vā; śara + mayaṭ → śaramayamFive sūtras come from the taddhita rules of chapters 29 and 30. A taddhita is an affix added to a noun stem, not to a root. The first gives ṭhañ after varṣā "the rains", where ordinary Sanskrit has ṭhak (SK 1389). The next two extend ṭhañ to vasanta and hemanta. Two more give mayaṭ "made of" in the Veda and limit it.
The running example is varṣā "the rains". The affix ṭhañ leaves ṭha, and ṭha becomes ika (ṭhasyekaḥ, SK 1170). The first vowel takes vṛddhi before a ñ-marked taddhita (ādivṛddhi, SK 1075); here it is already ā. The result is vā́rṣikam, with the udātta on the first syllable, vā. Such a word is ādyudātta, "udātta at the start".
Rule. In the Veda, ṭhañ comes after varṣā. It is an exception (apavāda) to the ṭhak of varṣābhyaṣ ṭhak (SK 1389, chapter 29). An apavāda is a special rule that sets aside a general one.
Example.
varṣā + ṭhañ → ṭha becomes ika (SK 1170) → vā́rṣikam, with the udātta on vā.
Applies when / blocked by. Bhaṭṭoji says only: "the difference is in the accent". The Subodhinī explains. With ṭhañ the word is ādyudātta by ñnity ādir nityam (SK 3686, chapter 71): vā́rṣikam. With ṭhak it would be antodātta by kitaḥ (SK 3712, chapter 71), with the udātta on the final syllable: vārṣikám.
Watch out. The two affixes give the same letters, vārṣikam. Only the place of the udātta shows which one was used.
Rule. After vasanta "spring" also, ṭhañ comes in the Veda. ṭhañ is carried from SK 3450.
Example. vasanta + ṭhañ → ika, vṛddhi of the first vowel (SK 1075) → vāsantikam.
Rule. After hemanta "winter" also, ṭhañ comes in the Veda.
Example. hemanta + ṭhañ → ika, vṛddhi e → ai (SK 1075) → haimantikam.
Applies when / blocked by. Why not one sūtra for vasanta and hemanta together? Bhaṭṭoji: the split (yogavibhāga) is for the sake of the following rule. The Subodhinī names it: sarvatrāṇ ca talopaś ca (SK 1390, chapter 29). Only hemanta, not vasanta, is to run on into that rule.
Watch out. Bhaṭṭoji next recalls another Vedic rule from chapter 29, śaunakādibhyaś chandasi (SK 1486). After śaunaka and the rest, ṇini comes in the sense "promulgated by". It is an exception to cha and aṇ. The Subodhinī says ṇini is carried from SK 1483.
śaunakinaḥ "those who study what Śaunaka promulgated"; likewise vājasaneyinaḥ.
Why "in the Veda"? Outside it, śaunakīyā śikṣā "the Śikṣā promulgated by Śaunaka", with cha (vṛddhāc chaḥ, SK 1337).
The print gives that sūtra as śaunakādibhyachandasi, a misprint for śaunakādibhyaś chandasi.
Rule. In the Veda, after a stem of two vowels, mayaṭ comes in the sense of a product (vikāra), "made of".
Example.
śara + mayaṭ → śaramayam: śaramayaṃ barhiḥ "the barhis made of śara reed".
parṇa + mayaṭ → with the feminine ī → parṇamayī: yasya parṇamayī juhūḥ "he whose juhū ladle is made of parṇa wood".
śa-ra and par-ṇa have two vowels each. Bhaṭṭoji does not say so; it is inferred.
Applies when / blocked by. The Subodhinī explains why the rule is needed. mayaḍ vaitayor bhāṣāyām… (SK 1523, chapter 30) gave mayaṭ only for ordinary language (bhāṣā). So it was not available in the Veda, and this sūtra enjoins it there.
Rule. The mayaṭ of SK 3453 does not come after a stem containing u, nor after vardhra and bilva. Bhaṭṭoji: utvat means "having the vowel u".
Example. The Subodhinī says the general affix aṇ comes instead.
muñja + aṇ → vṛddhi of the first vowel (SK 1075) → mauñjaṃ śikyam "a carrying-sling of muñja grass".
vardhra + aṇ → vārdhra → ṅīp (ṭiḍḍhāṇañ…, SK 470) → vārdhrī rajjuḥ "a rope of hide". Bhaṭṭoji: vardhra is a hide, and its product is vārdhrī.
bilva + aṇ → bailvo yūpaḥ "a sacrificial post of bilva wood".
That muñja is the u-stem of the rule is not stated by Bhaṭṭoji; it is inferred.
Watch out. Bhaṭṭoji ends by citing sabhāyā yaḥ (SK 1657, chapter 32) with no comment. It is the rule that the next sūtra presupposes.
megha + yat (bhava) → meghya: meghyāya ca vidyutyāya ca; agra + gha → (gha → iya, SK 475) agriyaḥThe rest of book 4 is one Vedic stretch. At SK 3456 Bhaṭṭoji declares a heading (adhikāra): "in the Veda" holds to the end of the fourth adhyāya. An adhikāra is a heading whose words are read into the rules that follow it. So every sūtra from SK 3457 to SK 3490 is Vedic, even where it does not say so.
SK 3456 also sets up a sense: bhava, "being there", after a seventh-case word. In ordinary Sanskrit this sense takes aṇ and the rest (tatra bhavaḥ, SK 1428, chapter 29). In the Veda it takes yat. Then seven sūtras give rival affixes after particular stems, and one (SK 3462) gives yat again. The sense bhava holds through SK 3464.
The running example is megha "cloud". meghe bhavaḥ, "what is in the cloud", is megha + yat → meghya. The rivals often replace their first letters by fixed sounds (āyaneyīnīyiyaḥ…, SK 475, chapter 15). So gha becomes iya, cha becomes īya and ḍha becomes eya: agra + gha → agriyaḥ.
Rule. In the Veda, ḍha comes after sabhā "assembly", as a seventh-case word, in the sense sādhu "good at, fit for".
Example.
sabhā + ḍha → ḍha becomes eya (SK 475) → the final ā is lost (yasyeti ca, SK 311) → sabheya.
sabheyo yuvā "a young man fit for the assembly".
Applies when / blocked by. The general rule sabhāyā yaḥ (SK 1657, chapter 32) gives ya after sabhā in this sense. The Subodhinī supplies the stem, the seventh case and the sense. That sabhā comes from SK 1657, and the case and sense from tatra sādhuḥ (SK 1650), is not stated by Bhaṭṭoji or his commentators; it is inferred.
Rule. In the Veda, after a seventh-case word, yat comes in the sense bhava, "being there".
Example.
megha + yat → the final a is lost (SK 311) → meghya.
vidyut + yat → vidyutya.
meghyāya ca vidyutyāya ca "to the one in the cloud and to the one in the lightning".
Applies when / blocked by. Bhaṭṭoji: this yat is an exception, as the case may be, to two sets of affixes. One is aṇ and the rest, given by tatra bhavaḥ (SK 1428, chapter 29). The other is gha and the rest. The Subodhinī names SK 1428.
Bhaṭṭoji adds: "In the alternative those too occur, because in the Veda all rules are optional." His instance: Muñjavat is a mountain, and what is there is mauñjavataḥ, with aṇ. The mantra reads somasyeva maujavatasya bhakṣaḥ "like the draught of Soma from Muñjavat". The text keeps maujavata there, without the nasal.
Watch out. Bhaṭṭoji closes: "The heading 'in the Veda' runs to the end of the fourth adhyāya." The yat itself belongs to the heading prāg ghitād yat (SK 1626, chapter 32). Bhaṭṭoji says so only at SK 3465; here it is inferred.
Rule. After pāthas and nadī "river", as seventh-case words, ḍyaṇ comes in the sense bhava, in the Veda.
Example. ḍyaṇ leaves ya. Its ḍ removes the ṭi, the last vowel of the stem with what follows (ṭeḥ, SK 316, chapter 10). Its ṇ brings vṛddhi of the first vowel (SK 1075).
pāthas + ḍyaṇ → the ṭi as is lost → pāthya.
pāthasi bhavaḥ → pāthyaḥ: tam u tvā pāthyo vṛṣā.
nadī + ḍyaṇ → the ṭi ī is lost → vṛddhi → nādya.
nadyāṃ bhavaḥ → nādyaḥ: cano dadhīta nādyo giro me "may the one in the river take delight in my songs".
Applies when / blocked by. The seventh case and the sense bhava are stated by the Subodhinī. It gives pāthas as "water", citing the Amarakośa, and adds that the Kāśikā takes it as "sky".
Rule. After veśanta and himavat, as seventh-case words, aṇ comes in the sense bhava, in the Veda. Bhaṭṭoji: "bhave" is carried.
Example. The Subodhinī: veśanta is a pond, formed from the root viś with an affix (jhac). The waters that are there are vaiśantyaḥ.
veśanta + aṇ → vṛddhi of the first vowel (SK 1075) → vaiśanta → ṅīp (SK 470) → vaiśantī.
vaiśantībhyaḥ svāhā "svāhā to the waters that are in ponds".
himavat + aṇ → haimavata → ṅīp → haimavatī.
haimavatībhyaḥ svāhā "svāhā to the waters that are on the Himālaya".
The feminine ṅīp step is not stated by Bhaṭṭoji here; it is inferred (compare vārdhrī, SK 3454).
Rule. After srotas "stream", as a seventh-case word, ḍyat and ḍya come optionally, in the sense bhava, in the Veda. Bhaṭṭoji: in the alternative, yat.
Example.
srotas + ḍyat or ḍya → the ṭi as is lost (SK 316) → srotya.
srotasi bhavaḥ "what is in a stream" → srotyaḥ.
srotas + yat → srotasyaḥ.
Applies when / blocked by. Bhaṭṭoji: "between ḍyat and ḍya, however, the difference is in the accent". He does not say what it is. The following is not stated by Bhaṭṭoji or his commentators; it is inferred from the accent rules of chapter 71. With ḍya, the affix is udātta on its first syllable (ādyudāttaś ca, SK 3708): srotyá, udātta on ya. With ḍyat, a t-marked affix is svarita (tit svaritam, SK 3729): srotyà, independent svarita on ya.
Watch out. Some copies print the plural-looking srotyāḥ. The singular srotyaḥ is required beside srotasi bhavaḥ and srotasyaḥ.
Rule. After sagarbha, sayūtha and sanuta, as seventh-case words, yan comes in the sense bhava, in the Veda. It is an exception to the yat of SK 3456, as the Subodhinī says.
Example. Bhaṭṭoji derives the stems. All three are karmadhāraya compounds with samāna "same". A karmadhāraya is a compound whose members refer to the same thing. samāna becomes sa by samānasya chandasi… (SK 1012, chapter 25).
samāna + garbha → sagarbha → + yan → sagarbhya: anubhrātā sagarbhyaḥ "the brother of the same womb".
samāna + yūtha → sayūtha → + yan → sayūthya: anusakhā sayūthyaḥ "the companion of the same herd".
samāna + nuta → sanuta → + yan → sanutya: yo naḥ sanutya uta vā jighatnuḥ.
nuta is nuti, "praise": kta in the neuter, in the sense of the bare action (napuṃsake bhāve ktaḥ, SK 3090, chapter 67).
Watch out. The 1915 print reads jighatnuḥ in the mantra.
Rule. After tugra, as a seventh-case word, ghan comes in the sense bhava, in the Veda. Bhaṭṭoji: "in the alternative, yat as well."
Example.
tugra + ghan → gha becomes iya (SK 475) → tugriya.
tugra + yat → tugrya.
The Bahvṛcas, the Ṛgvedins, read ā vaḥ śamaṃ vṛṣabhaṃ tugryāsu "(you) helped Śama, the bull, among the tugryās". Another śākhā, a Vedic recension, reads tugriyāsu.
Bhaṭṭoji cites the Kāśikā: tugra means wealth, sky, sacrifice or "best".
That tugryāsu shows yat and tugriyāsu shows ghan is not stated by Bhaṭṭoji; it is inferred.
Watch out. The text spaces the mantra ā vaḥ. In the Ṛgveda it is the single verb āvaḥ, "you helped", and it is read so here.
Rule. After agra "front", as a seventh-case word, yat comes in the sense bhava, in the Veda.
Example. agra + yat → the final a is lost (SK 311) → agrya: agryaḥ "what is at the front".
Applies when / blocked by. yat already comes by SK 3456. The Subodhinī says it is taught again so that gha and the rest should not block it.
Watch out. The 1915 print gives this sūtra bare, with no explanation. The forms agryaḥ, agriyaḥ, agrīyaḥ belong to SK 3463.
Rule. After agra, gha and cha also come, in the sense bhava, in the Veda. Bhaṭṭoji: by ca "too", yat as well.
Example. agre bhavaḥ "what is at the front":
agra + yat → agryaḥ.
agra + gha → gha becomes iya (SK 475) → agriyaḥ.
agra + cha → cha becomes īya (SK 475) → agrīyaḥ.
Bhaṭṭoji gives the three forms without labels. Matching them to the affixes is inferred.
Rule. After samudra "sea" and abhra "cloud", as seventh-case words, gha comes in the sense bhava, in the Veda.
Example.
samudra + gha → gha becomes iya (SK 475) → samudriya: samudriyā apsaraso manīṣiṇam "the apsarases of the sea … the wise one".
abhra + gha → abhriya: nānadato abhriyasyeva ghoṣāḥ "like the roars of the thundering one in the cloud".
The step gha → iya is not stated by Bhaṭṭoji here; it is inferred.
Applies when / blocked by. This is the last rule with the sense bhava. The Subodhinī says at SK 3465: "tatra bhava has ceased".
Watch out. The Subodhinī notes the order of the compound samudrābhra. Pāṇini did not write abhrasamudra. This shows that the rule fixing which member of a compound stands first is not invariable. The 1915 print reads apsaraso in the mantra.
asura + yat (svam) → (a lost, SK 311) asuryam, against asura + aṇ (māyā) → (ādivṛddhi; ṅīp) āsurī māyāFrom SK 3465 the sense bhava has ceased. The affix is still yat, from the heading prāg ghitād yat (SK 1626, chapter 32). Bhaṭṭoji says so at SK 3465: "prāg ghitād yat all the same". Each sūtra now names its own relation: "given on", "share or task of", "slayer of", "praise of", "own of". One sūtra gives aṇ instead.
The running example is asura. "The asura's own" is asura + yat → the final a is lost (SK 311) → asurya. "The asura's māyā" is asura + aṇ → vṛddhi of the first vowel (SK 1075) → āsura → the feminine ṅīp → āsurī.
Rule. After barhis "sacrificial grass", as a seventh-case word, yat comes in the sense "given on it", in the Veda.
Example.
barhis + yat → barhiṣya.
barhiṣyeṣu nidhiṣu priyeṣu "among the dear treasures given on the barhis".
Applies when / blocked by. Bhaṭṭoji: "prāg ghitād yat (SK 1626) all the same (ity eva)". So the affix is yat. The Subodhinī: "tatra bhava has ceased". The sense is now "given", not "being there".
Watch out. The s of barhis becomes ṣ in barhiṣya. Bhaṭṭoji does not explain this step.
Rule. After dūta "envoy", as a sixth-case word, yat comes when his share (bhāga) or his task (karman) is meant, in the Veda. Bhaṭṭoji: bhāga means a share (aṃśa).
Example. dūta + yat → the final a is lost (SK 311) → dūtyam.
Applies when / blocked by. The Subodhinī explains what the rule sets aside. For "share", it is a special statement against the aṇ of tasyedam (SK 1500, chapter 29). For "task", it is a statement against the affix ya (not yat) of the vārttika dūtavaṇigbhyāṃ ca. That vārttika is read under sakhyur yaḥ (SK 1791, chapter 37). The Subodhinī distinguishes dūtyaḥ "the envoy's share" from dūtyam "the envoy's task".
Watch out. The 1915 print reads bhāgo 'ṃśaḥ, "bhāga is a share".
Rule. After rakṣas and yātu, as sixth-case words in the plural, yat comes in the sense "that which slays them" (hananī), in the Veda.
Example.
rakṣas + yat → rakṣasya → feminine → rakṣasyā.
yā te agne rakṣasyā tanūḥ "that body of yours, O Agni, which slays the rakṣases".
The Subodhinī adds: likewise yātavyā.
Applies when / blocked by. The Subodhinī explains the plural. It shows the kind of praise meant: the praise lies in slaying many rakṣases. It also says the two synonyms are not reduced to one (ekaśeṣa), because the sūtra names the words themselves.
Rule. After revatī, jagatī and haviṣyā, yat comes in the sense of praising them, in the Veda. Bhaṭṭoji glosses praśasya as praśaṃsana, "praising".
Example.
revatī + yat → revatyam "the praising of the revatīs".
jagatī + yat → jagatyam.
haviṣyā + yat → the final ā is lost (yasyeti ca, SK 311) → haviṣy + ya → one y is dropped (halo yamāṃ yami lopaḥ, SK 60) → haviṣyam.
The steps for haviṣyam are the Subodhinī's.
Applies when / blocked by. The Subodhinī: praśasya is kyap in the sense of the bare action, from pra + śaṃs. So it means "praising", not "praiseworthy". That is why Bhaṭṭoji writes praśaṃsane.
Rule. After asura, as a sixth-case word, yat comes in the sense "its own" (sva), in the Veda.
Example.
asura + yat → the final a is lost (SK 311) → asurya.
asuryaṃ devebhir dhāyi viśvam "all that is the asura's own has been set in place by the gods".
The 1915 print reads devebhir.
Applies when / blocked by. The Subodhinī gives two derivations of asura. It is "not a sura", the opponent of the suras. Or it is from the root as with an affix (urac): "one who throws or is thrown".
Rule. When māyā is meant, aṇ comes after asura, in the Veda. The Subodhinī calls it an exception to the previous sūtra.
Example.
asura + aṇ → vṛddhi of the first vowel (SK 1075) → āsura → ṅīp (ṭiḍḍhāṇañ…, SK 470) → āsurī.
āsurī māyā "the māyā of the asura".
Applies when / blocked by. The Subodhinī says only that aṇ comes after asura, against SK 3469. That the sixth case and the sense "its own" carry over from SK 3469 is not stated by Bhaṭṭoji or his commentators; it is inferred. An exception presupposes the same case and sense.
Watch out. The Subodhinī glosses māyā as "that by which one measures out": the power to make the unreal appear.
varcasvat (a mantra containing varcas) + yat, luk of matup → varcas-ya → varcasyāḥ [iṣṭakāḥ]In the Vedic altar-building (cayana), bricks (iṣṭakā) are laid with mantras. Three sūtras name a brick after a word in its mantra. A mantra that contains the word varcas is called varcasvat, with the affix matup "having". matup is taught in chapter 39 (tad asyāsty asminn iti matup, SK 1894). The brick then takes yat, and the matup is removed by luk. luk is the loss of an affix without trace.
The running example: "these bricks have a varcasvat mantra for their laying". varcasvat + yat → luk of matup → varcas + ya → varcasyāḥ "the varcasya bricks".
Rule. Take a first-case word ending in matup, "containing that word". Suppose it names the mantra by which "these" are laid, and "these" are bricks. Then yat comes after it, and the matup takes luk. This holds in the Veda.
Example.
varcasvān upadhāno mantra āsām iṣṭakānām "a mantra containing varcas is the laying mantra of these bricks".
→ varcasvat + yat → luk of matup → varcasyāḥ.
Likewise a mantra containing ṛtu is ṛtumat: → ṛtavyāḥ.
Bhaṭṭoji gives only the phrase and the forms. The steps are implicit.
Applies when / blocked by. The Subodhinī explains iti. It is restrictive: a mantra with many words is named after one of them only. The Subodhinī also says matoḥ "of matup" serves SK 3472. There only matup is elided, not the ini of aśvin.
Watch out. All three conditions are needed. The Subodhinī gives counter-examples. A varcasvat mantra used to consecrate, not to lay, gets no yat. Nor does a hand that has fingers, though it lays bricks. Nor does gravel, which is laid but is not brick.
Rule. Under the conditions of SK 3471, aṇ comes after aśvimat, "containing the word aśvin", in the Veda. The matup again takes luk.
Example. The Subodhinī gives the steps.
aśva + ini (ata iniṭhanau, SK 1922, chapter 39) → aśvin → + matup → aśvimat.
aśvimat + aṇ → luk of matup → aśvin + a.
The in stays before aṇ (inaṇy anapatye, SK 1245, chapter 28) → vṛddhi of the first vowel → āśvina → feminine → āśvinī.
āśvinīr upadadhāti "he lays the āśvinī bricks".
Rule. Bhaṭṭoji: the whole of SK 3471 is carried forward. The word matoḥ is repeated and read the second time as a fifth-case word, "after matup". So: for vayasyā bricks, matup comes after mūrdhan that already ends in matup. The first matup takes luk.
Example. The vayasyā bricks are laid with a mantra containing vayas. The mantra meant has both words: mūrdhā vayaḥ… (the Subodhinī). So it is both mūrdhavat and vayasvat.
mūrdhavat + matup → luk of the first matup → mūrdhan + vat → the augment nuṭ (ano nuṭ, SK 3601) → mūrdhanvat → feminine → mūrdhanvatī.
mūrdhanvatīr upadadhāti "he lays the mūrdhanvatī bricks".
The 1915 print supplies the clause about the mantra with both words, and the Subodhinī's comment supports it.
Applies when / blocked by. The Subodhinī explains the rule's purpose. By SK 3471, vayasvat gives yat. So mūrdhavat would give yat too. This sūtra teaches matup against that yat. ano nuṭ is taught later in this chapter (Part V).
Watch out. The sūtra writes mūrdhnaḥ "after mūrdhan", where mūrdhavataḥ was due. The Subodhinī says Pāṇini wrote it with the coming luk of matup in view. The text reads mātoḥ for matoḥ; matoḥ is meant.
nabhas + yat (matvarthe) → nabhasyaḥ [māsaḥ]; ojas + kha → (kha → īna) ojasīnam [ahaḥ]Five sūtras give affixes "in the sense of matup" (matvarthe): "that is in it", "having". Ordinary Sanskrit uses matup itself (SK 1894, chapter 39). The Veda uses yat, ña, kha and yal after particular stems, for particular things. matvarthe holds from SK 3474 through SK 3478.
The running example is nabhas "cloud". "That is in it" gives nabhas + yat → nabhasyaḥ, the rainy month. With kha → īna (SK 475), ojas "vigour" gives ojasīnam, "a day having vigour".
Rule. After a first-case word, yat comes in the sense of matup, when a month (māsa) or a body (tanū) is meant. This holds in the Veda.
Example. Bhaṭṭoji: nabhas means cloud.
nabhas + yat → nabhasyaḥ "the month that has clouds in it" (tad asminn asti).
ojas + yat → ojasyā tanūḥ "a body possessed of vigour".
Applies when / blocked by. The Subodhinī says the first case follows from the mention of the matup sense.
Watch out. The sūtra's tanū has a long ū. The Subodhinī gives the reason. A short-u tanu would be a ghi stem. A ghi stem stands first in a dvandva, a compound of co-ordinate members (dvandve ghi, SK 903, chapter 20). Since māsa stands first, tanu cannot be meant.
Rule. After madhu, ña comes in the sense of matup, and by ca yat as well, in the Veda. The Subodhinī supplies the sense and the yat.
Example.
madhu + ña → vṛddhi of the first vowel (SK 1075) → guṇa of the final u (or guṇaḥ, SK 847, chapter 19) → o becomes av before the vowel a (eco 'yavāyāvaḥ, SK 61, chapter 3) → mādhavaḥ.
madhu + yat → guṇa of the final u (SK 847) → o becomes av before y (vānto yi, SK 63, chapter 3) → madhavyaḥ.
The guṇa and av steps for madhavyaḥ are the Subodhinī's. For mādhavaḥ the same two steps are not stated by Bhaṭṭoji or his commentators; they are inferred.
Watch out. The words "by ca, yat" are not in the 1915 print. The sense rests on the Subodhinī.
Rule. After ojas "vigour", yat and kha come in the sense of matup, when a day (ahan) is meant, in the Veda.
Example.
ojas + yat → ojasyam ahaḥ "a day possessing vigour".
ojas + kha → kha becomes īna (SK 475) → ojasīnam.
Applies when / blocked by. The Subodhinī meets an objection. Why name yat? Why not just say "kha too" (khaś ca)? Because ca would then add kha only to the ña of SK 3475, the rule just before. So yat is named.
Rule. After bhaga preceded by veśas or yaśas, yal comes in the sense of matup, in the Veda.
Example. Bhaṭṭoji: veśas is strength, and that itself is bhaga.
veśobhaga + yal → veśobhagyaḥ; with kha (SK 3478) → veśobhagīnaḥ.
yaśobhaga + yal → yaśobhagyaḥ; with kha → yaśobhagīnaḥ.
Applies when / blocked by. Bhaṭṭoji: "respective correspondence (yathāsaṃkhya) is not intended". So the first base is not tied to yal and the second to kha. Both affixes come after both bases. The Subodhinī states yal and kha together here.
Watch out. The Subodhinī says the l of yal is "for the accent". By liti (SK 3676), the syllable before an l-marked affix is udātta. That would give veśobhágyaḥ, with the udātta on bha. The form is not stated by Bhaṭṭoji or his commentators; it is inferred. A footnote in the 1915 print, citing the Manoramā, says the Vṛttikāra split off "kha ca" as a separate sūtra. The Bhāṣya is said to read yal and kha in one sūtra. Scholars discuss this.
Rule. kha also comes, under the same conditions as yal in SK 3477.
Example. veśobhaga + kha → kha becomes īna (SK 475) → veśobhagīnaḥ; likewise yaśobhagīnaḥ.
Applies when / blocked by. Bhaṭṭoji: the split of the rule (yogavibhāga) is "for the sake of what follows, and for excluding respective order". Which following rule is meant is not said. The ca of SK 3479 is the nearest candidate; this is inferred.
sahasra + gha → (gha → iya; a lost, SK 311) sahasriya: sahasriyāso apāṃ normayaḥThree sūtras give Vedic affixes for three new relations: "made by those of old", "prepared with water" and "equal to a thousand". A fourth gives gha after sahasra in the sense of matup as well.
The running example is sahasra "a thousand". sahasra + gha → gha becomes iya (SK 475) → the final a is lost (SK 311) → sahasriya, "equal to a thousand".
Rule. After pūrva, in the sense "made by those of old" (pūrvaiḥ kṛtam), ina and ya come, in the Veda.
Example.
pūrva + ina → pūrviṇa: gambhīrebhiḥ pathibhiḥ pūrviṇebhiḥ "by deep paths made by those of old".
pūrva + ya → pūrvya: ye te panthāḥ savitaḥ pūrvyāsaḥ "your paths, O Savitṛ, that were made by those of old".
The Subodhinī glosses both: pūrviṇāḥ = "made by those of old", and pūrvyāsaḥ = pūrviṇāḥ.
Applies when / blocked by. The sūtra's dual inayau names two affixes, ina and ya. What its ca adds is not stated by Bhaṭṭoji or the Subodhinī.
Watch out. The n of ina becomes ṇ after the r of pūrva. Bhaṭṭoji does not state this step.
Rule. After ap "water", as a third-case word, yat comes in the sense "prepared with it", in the Veda. The Subodhinī supplies the third case and yat.
Example.
ap + yat → apya.
yasyedam apyaṃ haviḥ "whose is this oblation prepared with water".
Rule. After sahasra, as a third-case word, gha comes in the sense saṃmita, "commensurate with a thousand", in the Veda. Bhaṭṭoji: the meaning is "equal to a thousand".
Example.
sahasra + gha → gha becomes iya (SK 475) → the final a is lost (yasyeti ca, SK 311) → sahasriya.
sahasriyāso apāṃ normayaḥ "equal to a thousand, like the waves of the waters".
The loss of a is the Subodhinī's step.
Watch out. The Subodhinī glosses saṃmita as "like, equal". It records the variant reading samitau for the sūtra. The 1915 print reads sahasriyāso, with sra.
Rule. After sahasra, gha also comes in the sense of matup, in the Veda.
Example. sahasram asyāsti "he has a thousand" → sahasra + gha → sahasriyaḥ.
Applies when / blocked by. The Subodhinī: this is an exception to tapaḥsahasrābhyāṃ vinīnī (SK 1909, chapter 39). That rule gives vini and ini after tapas and sahasra in the matup sense.
soma + ya → somyaḥ (arhati: somyo brāhmaṇaḥ; mayaṭ sense: somyaṃ madhu); madhu + yat → (guṇa, av) madhavyaḥFour sūtras give ya and yat for "deserves", for the senses of mayaṭ, "consisting of", and for "collection". Then two affixes come "in their own sense" (svārthe): yat after chandas (a vārttika) and gha after nakṣatra. Such an affix is svārthika: it adds no meaning of its own.
The running example is soma. soma + ya → the final a is lost (SK 311) → somya. With "deserves" it gives somyo brāhmaṇaḥ "a brāhmaṇa who deserves soma". With the mayaṭ sense it gives somyaṃ madhu "honey consisting of soma".
Rule. After soma, as a second-case word, ya comes in the sense "deserves" (arhati), in the Veda. The Subodhinī supplies the second case.
Example. soma + ya → somyo brāhmaṇaḥ "a brāhmaṇa who deserves soma". Bhaṭṭoji: the meaning is "worthy of the sacrifice" (yajñārha).
Watch out. The Subodhinī names the affix yat. The sūtra says yaḥ, and Bhaṭṭoji repeats yaḥ at SK 3484. Follow the sūtra: the affix is ya.
Rule. After soma, ya also comes in the sense of the affix mayaṭ, in the Veda.
Example. somyaṃ madhu = somamayam "consisting of soma".
Applies when / blocked by. The Subodhinī lists the senses of mayaṭ: "come from" (āgata), "product" (vikāra), "part" (avayava) and "abundantly present" (prakṛta). For "come from" the base stands in the fifth case. For product and part it stands in the sixth. For abundance it stands in the first.
Rule. After madhu, yat comes in the sense of mayaṭ, in the Veda. Bhaṭṭoji says yat, not the ya of SK 3483–3484.
Example.
madhu + yat → guṇa (SK 847) → av before y (SK 63) → madhavyaḥ.
The meaning is madhumayaḥ "consisting of honey".
Applies when / blocked by. The Subodhinī: the Vṛttikāra carries forward only yat, not ya. Neither Bhaṭṭoji nor the Subodhinī names where this yat comes from.
Rule. After vasu, yat comes in the sense of a collection (samūha), in the Veda. Bhaṭṭoji: by ca, also in the sense of mayaṭ.
Example. vasu + yat → guṇa (SK 847) → av before y (SK 63) → vasavyaḥ.
These steps are not stated by Bhaṭṭoji here; they are inferred, as for madhavyaḥ (SK 3485).
Vārttika. akṣarasamūhe chandasa upasaṃkhyānam: chandas must be added, in the sense of a collection of syllables. Bhaṭṭoji: yat comes in its own sense after chandas meaning a collection of syllables. He counts the syllables:
āśrāvaya has four; astu śrauṣaṭ has four; yaja has two; ye yajāmahe has five; the vaṣaṭ call has two.
This, he says, is the seventeen-syllable chāndasyaḥ.
Watch out. The text reads chāndasyaḥ, with vṛddhi. yat is not ñ- or ṇ-marked, so one would expect chandasya. This is inferred; no commentary discusses the form.
Rule. After nakṣatra "lunar mansion", gha comes in its own sense (svārthe), in the Veda.
Example. nakṣatra + gha → gha becomes iya (SK 475) → nakṣatriya.
nakṣatriyebhyaḥ svāhā "svāhā to the nakṣatras".
Applies when / blocked by. The Subodhinī: samūha of SK 3486 is not carried forward. Since no sense is named, the affix comes in its own sense.
śiva + tātil → śivatātiḥ ("he makes welfare", 3489; "the state of welfare", 3490)The fourth book ends with three sūtras on tātil. Its l is a marker, so it leaves tāti. It comes in its own sense after sarva and deva. After śiva, śam and ariṣṭa it means "one who makes" and "the state of".
The running example is śiva "welfare". śivaṃ karoti "he makes welfare" → śiva + tātil → śivatātiḥ. śivasya bhāvaḥ "the state of welfare" → śivatātiḥ as well.
Rule. After sarva "all" and after deva "god", tātil comes in its own sense, in the Veda. The Subodhinī takes the two as separate bases, not as one word sarvadeva.
Example.
sarva + tātil → sarvatāti: savitā naḥ suvatu sarvatātim "may Savitṛ impel for us the whole".
deva + tātil → devatāti: pradakṣiṇid devatātim urāṇaḥ "going round to the right, making the gods".
Since the affix is svārthika, sarvatāti and devatāti mean the same as sarva and deva.
Rule. After śiva, śam and ariṣṭa, standing as the object, tātil comes in the sense kara, "one who makes". This holds in the Veda.
Example. Bhaṭṭoji: kara is "one who makes", with the affix ac of the paca-group (SK 2896, chapter 67).
śivaṃ karotīti → śiva + tātil → śivatātiḥ "maker of welfare".
yābhiḥ śantātī bhavatho dadāśuṣe "the aids by which you two become bringers of welfare to the worshipper".
atho ariṣṭatātaye "and also for the maker of safety".
Applies when / blocked by. The Subodhinī explains the sixth case in śivaśamariṣṭasya. The derived word itself means kara, "maker": the two refer to the same agent. So the sixth case is that of the object, by ubhayaprāptau karmaṇi (SK 624, chapter 16). That tātil is carried from SK 3488 is not stated by Bhaṭṭoji or his commentators; it is inferred.
Watch out. The reading dadāśuṣe "to the worshipper" is chosen. The 1915 print is unclear at this letter, and dadāśuṣe is the only form that makes sense.
Rule. After śiva, śam and ariṣṭa, tāti also comes in the sense of the state (bhāva), in the Veda.
Example.
śivasya bhāvaḥ "the state of welfare" → śivatātiḥ.
Likewise śantātiḥ, ariṣṭatātiḥ.
Applies when / blocked by. Bhaṭṭoji's gloss names śiva and the rest. That they are carried from SK 3489 is inferred.
Watch out. The same word śivatātiḥ thus has two senses: "maker of welfare" (SK 3489) and "the state of welfare" (SK 3490). The 1915 print closes with "Here ends the fourth adhyāya." The heading "in the Veda" set up at SK 3456 ends here too.
saptan + añ → (ṭi lost, nas taddhite) sāpta → sāptāni; saṃvatsara + kha (īna) → saṃvatsarīṇaḥWith this part the Vedic rules reach the fifth book (adhyāya) of the Aṣṭādhyāyī. The 1915 print heads the stretch pañcamo 'dhyāyaḥ, "the fifth book". No Vedic heading runs through it. Bhaṭṭoji's heading ā caturthasamāpteḥ (SK 3456) stopped at the end of book 4. So every rule here names chandasi, "in the Veda", itself.
These six rules add Vedic taddhita affixes. A taddhita is an affix added to a noun stem to make a new stem. Five of the six rules work inside a sense from chapters 34–36: "that is its measure", "deserves that", "accomplished by" and "that has come to it". SK 3496 gives vati in its own sense. The running examples are sāptāni "heptads", from saptan "seven", and saṃvatsarīṇaḥ "lasting a year", from saṃvatsara.
Rule. In the Veda, the affix añ comes after saptan "seven". It comes in the sense "that is its measure" (tad asya parimāṇam, SK 1723, chapter 34), and in the sense "a group" (varge). Bhaṭṭoji names both senses as carried forward. That varge is the word of SK 1726 (chapter 34) is not stated by Bhaṭṭoji or his commentators; it is inferred.
Example. Bhaṭṭoji quotes sapta sāptāni asṛjat "he created seven heptads". The Subodhinī, the commentary on these Vedic rules, derives sāptāni:
saptan + añ
→ the ṭi, the last vowel with what follows it (an), is lost before a taddhita (SK 679, chapter 17) → sapt + a → sāpta
→ neuter, first case plural: jas becomes śi (SK 312, chapter 10) → sāpta + i
→ the augment num (SK 314, chapter 10), and the vowel before it lengthened → sāptāni.
The long ā of the first syllable is the vṛddhi of a ñ-marked taddhita (SK 1075, chapter 26). That step is not stated by Bhaṭṭoji or his commentators; it is inferred.
Applies when / blocked by. Three vārttikas follow. A vārttika is a supplementary statement on a sūtra, which Bhaṭṭoji quotes.
pañcadaśan + ḍini → pañcadaś + in → pañcadaśino 'rdhamāsāḥ "half-months of fifteen days". Likewise triṃśino māsāḥ "months of thirty days".viṃśino 'ṅgirasaḥ "the Aṅgirases, whose measure is twenty". The Subodhinī: twenty sub-lineages, told apart by their pravara lists (pravara, the list of ancestral seers a lineage invokes). Here viṃśati + ḍini → ti is lost before a ḍ-marked affix (SK 844, chapter 19) → viṃśa + in → the a is lost (SK 311, chapter 10) → viṃśin.tvāvataḥ purūvaso "of one like you, O you of many goods"; na tvāvān anyaḥ "there is no other like you"; yajñaṃ viprasya māvataḥ "the sacrifice of a sage like me". The Subodhinī: tva and ma come by SK 1373 (chapter 29), and the ā by SK 430 (chapter 11).Rule. In the Veda, the affix yat comes after any noun stem whatever (prātipadika) in the sense "deserves that" (tad arhati). Bhaṭṭoji names this sense. That it is the sense of SK 1728 (chapter 34) is not stated by Bhaṭṭoji or his commentators; it is inferred.
Example. Bhaṭṭoji gives sādanyaṃ vidathyam "deserving of the house, deserving of the sacrifice". The Subodhinī explains both.
sadana + yat "it deserves a house" → the first vowel lengthened by anyeṣām api dṛśyate (SK 3539, below) → sādanyam.
vidatha + yat "it deserves the sacrifice" (vidatha is the sacrifice) → vidathyam.
Rule. In the Veda, the affix cha comes after a stem ending in vatsara "year". It comes in the senses that begin with "accomplished" (nirvṛtta). In the word, cha is replaced by īya.
Example.
idvatsara + cha (īya) → idvatsarīyaḥ.
The Subodhinī: it means "accomplished by an idvatsara", or "engaged, hired, having lasted, or going to last for an idvatsara". So it takes the senses of tena nirvṛttam (SK 1743) and tam adhīṣṭo bhṛto bhūto bhāvī (SK 1744), both in chapter 35. The idvatsara is one of the years of the five-year cycle (yuga). The Subodhinī names idāvatsara, saṃvatsara and parivatsara as years of the same cycle.
Rule. After a stem ending in vatsara that has sam or pari before it, the affix kha comes. By the word ca "and", cha (from SK 3493) comes too. The senses are the same "accomplished" group. In the word, kha is replaced by īna.
Example. Bhaṭṭoji gives the forms in pairs:
saṃvatsara + kha (īna) → saṃvatsarīṇaḥ; saṃvatsara + cha (īya) → saṃvatsarīyaḥ.
parivatsara + kha → parivatsarīṇaḥ; parivatsara + cha → parivatsarīyaḥ.
Applies when / blocked by. "In the Veda", "ending in vatsara" and the senses all come forward from SK 3493. The Subodhinī states this.
Rule. In the Veda, the affix ghas comes after the word ṛtu "season", in the sense "that has come to it". The Subodhinī: the sense is carried forward from samayas tad asya prāptam (SK 1768), and ṛtu from ṛtor aṇ (SK 1769), both in chapter 36. In the word, ghas appears as iya.
Example. bhāga ṛtviyaḥ "a share whose season has come". The Subodhinī derives it:
ṛtu + ghas
→ the s is a marker; by siti ca (SK 1252, chapter 28) the stem counts as a pada, not as a bha
→ so guṇa by or guṇaḥ (SK 847, chapter 19) does not apply
→ u becomes v before the vowel (yaṇ) → ṛtv + iya → ṛtviyaḥ.
Watch out. Guṇa by SK 847 does not apply here. A stem before a vowel-initial taddhita is normally a bha, and a bha in u takes guṇa (SK 847). The s of ghas stops that: it makes the stem a pada.
Rule. In the Veda, the affix vati comes in its own sense after an upasarga, a verbal prefix such as ud or ni. The prefix must denote a sādhana, a factor of an action, qualified by a verbal meaning. That this vati is the affix of tena tulyaṃ kriyā ced vatiḥ (SK 1778, chapter 37) is not stated by Bhaṭṭoji or his commentators; it is inferred.
Example. yad udvato nivataḥ "when (you go) from the heights and from the depths". Bhaṭṭoji: the meaning is "from what has gone up (udgata), from what has gone off (nirgata)".
ud + vati → udvat → fifth case udvataḥ; ni + vati → nivataḥ.
Applies when / blocked by. The Subodhinī explains "in the sense of a verb". Where a verb is expressed, the prefix qualifies that verb, as in āgacchati "he comes". Where none is expressed, the prefix itself stands for a factor qualified by the action, as in niṣkauśāmbiḥ "gone out of Kauśāmbī". The affix comes only in the second case, where no verb is expressed.
pañcan + ḍaṭ (+ thaṭ augment) → pañcatham; kim + thā → (ka for kim) kathāThese seven rules are Vedic additions to chapters 39–42. One gives a Vedic ordinal: pañcatham beside pañcamam "the fifth". One gives Vedic possessive affixes such as vin, ī and vanip. The rest give Vedic adverbs from pronouns and other words: idā "now", tarhi "then", kathā "how", pratnathā "as of old". The running examples are pañcatham and kathā.
Rule. In the Veda, the affix ḍaṭ takes the augment thaṭ, and also maṭ. An augment is a sound added to an affix or stem, not a separate affix. The ḍaṭ meant is the one of tasya pūraṇe ḍaṭ (SK 1849, chapter 39), "the filler of". It must follow a word ending in n that does not begin with a numeral (nāntād asaṃkhyādeḥ, SK 1850). The Subodhinī names both rules. That the word in n is a numeral, such as pañcan, is not stated here; it is inferred from the ḍaṭ rule.
Example. The Subodhinī: for "the filler of five" (pañcānāṃ pūraṇam), ḍaṭ is added with the augment thaṭ.
pañcan + ḍaṭ with thaṭ → pañcatham "the fifth".
pañcan + ḍaṭ with maṭ → pañcamam.
The 1915 print divides the sūtra thaṭ ca chandasi.
Applies when / blocked by. SK 1850 gives the augment maṭ: pañcamam. The Veda adds thaṭ.
Bhaṭṭoji also cites here chandasi paripanthiparipariṇau paryavasthātari (SK 1889, chapter 39). In the Veda paripanthin and paripariṇ are fixed forms (nipātana) for "enemy". A nipātana is a finished form stated in a sūtra, whose irregular steps are simply accepted. Bhaṭṭoji glosses paryavasthātṛ as "enemy" (śatru). Examples: apa tyaṃ paripanthinam "(drive) away that waylayer"; mā tvā paripariṇau vidan "may the two adversaries not find you". The Subodhinī: the affix ini comes after paryavasthātṛ in its own sense, and panthi and pari replace avasthātṛ, both by nipātana.
Rule. In the Veda, the affix vini comes variously (bahulam) in the sense of matup, "possessing". bahulam means the rule holds in some Vedic words and not in others. The Subodhinī: vini is carried forward from asmāyāmedhāsrajo viniḥ (SK 1928, chapter 39).
Example. Two vārttikas follow.
ubhaya + vin → final a lengthened → ubhayāvin → second case ubhayāvinam: maṃhiṣṭham ubhayāvinam "the most bountiful one, possessing both".
śunam aṣṭrāvy acarat "the one possessing the aṣṭrā went prosperously". The Subodhinī: aṣṭrā is a synonym of daṃṣṭrā.
ī: rathīr abhūt "he became a chariot-owner"; sumaṅgalīr iyaṃ vadhūḥ "this bride is auspicious".
vanip: maghavānam īmahe "we entreat the bountiful one".
Applies when / blocked by. The Subodhinī: only dvaya, ubhaya and hṛdaya really need the lengthening; the other three are long already. It forms mekhalāvī, dvayāvī, rujāvī and hṛdayāvī the same way. In sumaṅgalī, su and maṅgala are first compounded by suḥ pūjāyām, and then ī comes.
Watch out. With vanip, magha "wealth" gives maghavan. With matup the same sense gives maghavat (Subodhinī). Keep the two stems apart.
Rule. In the Veda, after idam "this" and tad "that", the affixes dā and rhil come, in that order. The Subodhinī: "those two" are idam and tad, which stand next to each other in idamo rhil (SK 1965) and tado dā ca (SK 1968), chapter 40. In the sūtra the sixth case "of those two" stands for the fifth case "after those two". This exchange of forms is vyatyaya.
Example. The Subodhinī pairs them: dā after idam, rhil after tad.
idam + dā → idam is replaced by iś (SK 1949, chapter 40) → idā "now": idā hi va upastutim.
tad + rhil → tarhi "then".
Watch out. The pairing is the reverse of chapter 40. There SK 1965 gives rhil after idam, and SK 1968 gives dā after tad. The Veda adds the crossed pair.
Rule. In the Veda, the affix thā comes after kim "what" in the sense of cause (hetu), and in the sense of manner. The Subodhinī: "manner" is carried forward from prakāravacane thāl (SK 1971), and kim from kimaś ca (SK 1973), both in chapter 40.
Example.
kim + thā
→ thā falls under the heading prāg diśo vibhaktiḥ (SK 1947, chapter 40), so it counts as a vibhakti, the name of a case ending, which SK 1947 also gives to these adverbial affixes
→ before a vibhakti kim becomes ka (SK 342, chapter 11) → kathā.
Cause: kathā grāmaṃ na pṛcchasi "for what reason do you not ask about the village?"
Manner: kathā dāśema "in what manner should we worship?"
The Subodhinī gives both glosses.
Rule. In the Veda, paśca and paścā are fixed forms (nipātana) for avara "later, behind", in the sense of the affix astāti. The Subodhinī: paśca replaces avara, and the affixes a and ā are given by nipātana. It adds that astāti comes after words of direction in the senses of direction, place and time (SK 1974, chapter 41).
Example. paśca hi saḥ "for he (is) behind"; no te paścā "not after you".
Bhaṭṭoji also cites here tuś chandasi (SK 2007, chapter 41). In the Veda, iṣṭhan and īyasun come after a word ending in tṛc or tṛn. These are the affixes of "most" and "more".
kartṛ + iṣṭhan → tṛ is lost by tur iṣṭhemeyaḥsu (SK 2008, chapter 41) → kar + iṣṭha → kariṣṭhaḥ: āsutiṃ kariṣṭhaḥ "most active (in making) the pressing".
dohīyasī dhenuḥ "a cow that is the better milker". The Subodhinī: the affix comes after the feminine dogdhrī. The feminine ī is removed because the stem is treated as masculine (puṃvadbhāva). Then tṛ is lost by SK 2008.
Rule. In the Veda, the affix thāl comes after pratna "old", pūrva "former", viśva "all" and ima "this". Bhaṭṭoji supplies the sense: "like" (iva). That this sense comes down from ive pratikṛtau (SK 2051, chapter 42) is not stated by Bhaṭṭoji or his commentators; it is inferred.
Example. taṃ pratnathā pūrvathā viśvathemathā "him, as of old, as before, as always, as now".
pratna + thāl → pratnathā; likewise pūrvathā, viśvathā, imathā.
viśvathemathā is viśvathā + imathā joined in recitation.
Rule. In the Veda, the affix amu comes as well. It comes after kim, after a word ending in e, after a verb form, after an indeclinable (avyaya) and after a word ending in a gha affix. Bhaṭṭoji says this list carries forward from SK 2004 (chapter 41). The sense is excess (prakarṣa), but not excess of a substance, as SK 2004 states.
Example. prataṃ naya prataram "lead him forth, farther".
pra + tarap "further forth" → pratara
→ pratara ends in tarap, one of the gha affixes → pratara + amu → prataram.
The Subodhinī: tarap comes after pra, which itself means excess, for a further degree. Words ending in am and ām are listed among the indeclinables, so prataram is an avyaya. Its case ending is dropped (luk).
Watch out. The Subodhinī explains the u of amu. It is a marker, so that amu is not caught by ica ekāco 'mpratyayavac ca (SK 2994). Without the marker that rule's treatment would spread to it, with wrong results.
brahman + sāman + ṭac → (ṭi lost) brahmasāmam; hatā mātā yasya → (kap blocked) hatamātāThe last five rules of book 5 change compounds and add affixes in a word's own sense. One gives affixes "in their own sense" (svārthe), which leave the meaning as it was. One gives ṭac and one blocks kap at the end of a compound; such an affix is a samāsānta. One gives a fixed form, bahuprajāḥ. One gives a substitute inside a compound. The running examples are brahmasāmam, where a Vedic ṭac is added, and hatamātā, where the ordinary kap is blocked.
Rule. In the Veda, the affixes til and tātil come after vṛka "wolf" and jyeṣṭha "eldest, best", in their own sense. til goes with vṛka, and tātil with jyeṣṭha. Two bases and two affixes of equal number pair in order by yathāsaṃkhyam (SK 128, chapter 5). That this paribhāṣā is at work is not stated by Bhaṭṭoji or his commentators; it is inferred, and the examples bear it out.
Example.
vṛka + til → vṛkatiḥ: yo no durevo vṛkatiḥ "the ill-disposed wolf who (is set) against us".
jyeṣṭha + tātil → jyeṣṭhatātim: jyeṣṭhatātiṃ barhiṣadam "the most excellent one, seated on the barhis".
Applies when / blocked by. You met tātil in Part II, at SK 3488–3490 (śivatātiḥ). Bhaṭṭoji does not explain the word ca "also", and there is no Subodhinī on this sūtra.
Rule. In the Veda, the samāsānta ṭac comes after a neuter tatpuruṣa compound that ends in an or as. Bhaṭṭoji supplies "tatpuruṣa" and "ṭac". That they come from SK 786 and rājāhaḥsakhibhyaṣ ṭac (SK 788), chapter 18, is not stated by Bhaṭṭoji or his commentators; it is inferred.
Example. The Subodhinī labels the two examples.
Ending in an: brahman + sāman → brahmasāman + ṭac → the ṭi an is lost → brahmasāma → brahmasāmaṃ bhavati "it becomes the brahmasāman".
Ending in as: deva + chandas + ṭac → devacchandasa → neuter plural devacchandasāni "the metres of the gods".
Rule. In the Veda, bahuprajāḥ "having many offspring" is a fixed form (nipātana).
Example. bahuprajā nirṛtim āviveśa "having many offspring, he has entered into Nirṛti (dissolution)". The Subodhinī derives it:
bahavaḥ prajā yasya "he whose offspring are many": a bahuvrīhi → bahu-prajā
→ the affix asic → bahuprajā + as
→ the final ā is lost (SK 311, chapter 10) → bahupraj-as
→ the vowel before s lengthened (SK 425, chapter 11) → bahuprajās
→ s becomes ru, then visarga → bahuprajāḥ.
A bahuvrīhi is a compound that names something outside its own members.
Rule. In the Veda too, datṛ replaces danta "tooth" in a bahuvrīhi. The Subodhinī: "datṛ for danta" is carried forward from vayasi dantasya datṛ (SK 880, chapter 19).
Example. ubhayato dantā yasya "that which has teeth on both sides":
ubhayatas + danta → datṛ replaces danta → ubhayatodat- → second case plural ubhayatodataḥ.
ubhayatodataḥ pratigṛhṇāti "he receives those that have teeth on both sides".
Watch out. SK 880 gives datṛ only when the teeth show age (vayasi). The Subodhinī names only "datṛ for danta" as carried forward. So in the Veda the substitute comes without that condition. This conclusion is not stated as such by the Subodhinī; it is inferred.
Rule. In the Veda, after a bahuvrīhi ending in ṛ, the samāsānta kap does not come. By nadyṛtaś ca (SK 833, chapter 19) kap would come obligatorily; this rule blocks it (Subodhinī).
Example.
hatā mātā yasya "he whose mother has been slain"
→ bahuvrīhi hata-mātṛ; no kap
→ first case hatamātā.
The 1915 print closes the book here: iti pañcamo 'dhyāyaḥ, "here ends the fifth book".
Watch out. This is a prohibition, not an option. In the Veda the ṛ-final bahuvrīhi stands bare.
tuj + liṭ (kānac) → tu-tuj-āna → (abhyāsa lengthened) tūtujānaḥ; ā + hve + laṭ → (saṃprasāraṇa, uvaṅ) āhuveThe 1915 print now heads a new book: ṣaṣṭho 'dhyāyaḥ, "the sixth book". Its Vedic rules begin with the root. They touch reduplication, taught in chapter 44 (SK 2175). They touch saṃprasāraṇa, the change of y, v, r to i, u, ṛ. And they give substitutes for roots and stems, with one sūtra of nine fixed forms. The running examples are tūtujānaḥ, with a long reduplicated syllable, and āhuve "I call", with saṃprasāraṇa of hve.
Rule. For tuj and the roots like it, the abhyāsa is lengthened. The abhyāsa is the first, repeated part of a doubled root. Bhaṭṭoji: tujādi is an ākṛtigaṇa, a list known from its instances rather than closed by a full count.
Example. The Subodhinī: tuj, with liṭ replaced by kānac (āna).
tuj + kānac
→ the root is doubled (SK 2175, chapter 44) → tu-tuj-āna
→ the abhyāsa tu is lengthened → tūtujāna → tūtujānaḥ: prabharā tūtujānaḥ.
Others: sūrye māmahānam (mah "to honour", with kānac); dādhāra yaḥ pṛthivīm "who has upheld the earth"; sa tūtāva (tu, a root found only in sūtras, in liṭ).
Applies when / blocked by. Before this sūtra the 1915 print puts a vārttika on ekāco dve prathamasya (SK 2175). It reads chandasi veti vaktavyam, "it should be stated: optionally, in the Veda". So in the Veda the doubling itself is optional. The undoubled forms are yo jāgāra "who has kept awake" and dāti priyāṇi "he gives dear things". The Subodhinī: jāgāra is jāgṛ "to wake" in liṭ. dāti is dā in laṭ, with ślu for śap. There ślau (SK 2490, chapter 46) would make the doubling obligatory. Here it is optional.
Watch out. Two Vedic options meet here. The doubling of SK 2175 may fail (jāgāra). Where it does happen, the abhyāsa of tuj and the like is long (tūtujānaḥ).
Rule. In the Veda, hvā (hveñ) "to call" undergoes saṃprasāraṇa, variously. The Subodhinī: hvaḥ saṃprasāraṇam is carried forward from SK 2586 (chapter 54).
Example. indram āhuva ūtaye "I call Indra for help". The Subodhinī derives āhuve: hve with ā, laṭ, ātmanepada, first person singular.
ā + hve + laṭ (ending e)
→ śap is lost by "bahulaṃ chandasi"
→ saṃprasāraṇa → ā + hu + e
→ uvaṅ: u becomes uv before the vowel → āhuve.
The "bahulaṃ chandasi" that drops the śap is the earlier rule SK 3400 (Part I), not this one. The Subodhinī does not say which; it is inferred.
Applies when / blocked by. Two vārttikas follow.
tri + ṛc → saṃprasāraṇa of tri; the initial ṛ of ṛc is lost → tṛc → the samāsānta a (SK 940, chapter 23) → tṛcaṃ sūktam "a hymn of three verses". The Subodhinī: "that in which there are three verses".
Why "in the Veda"? Outside it: tryṛcāni.
revān: the Subodhinī says the v for m is by chandasīraḥ (SK 3600, Part V).
rayimān puṣṭivardhanaḥ "possessed of wealth, increasing prosperity": by bahulam, neither saṃprasāraṇa nor v.
Rule. In the Veda, kī replaces cāy (cāyṛ "to honour, to perceive"), variously. Bhaṭṭoji says bahulam is carried forward from SK 3510.
Example. ny anyaṃ cikyur na ni cikyur anyam "they discerned the one, they did not discern the other". Bhaṭṭoji: cikyuḥ is liṭ with the ending us.
cāy + liṭ (us)
→ kī replaces cāy → kī + us
→ the root is doubled (SK 2175, chapter 44) → kī-kī + us
→ the abhyāsa kī is shortened by hrasvaḥ (SK 2180, chapter 44) → ki-kī + us
→ the abhyāsa k becomes c by kuhoś cuḥ (SK 2245, chapter 44; Subodhinī) → ci-kī + us
→ ī becomes y before the vowel of us by eranekāco 'saṃyogapūrvasya (SK 272, chapter 8) → cikyus → cikyuḥ.
The shortening and the y are not stated by Bhaṭṭoji or his commentators; they are inferred.
Applies when / blocked by. Because bahulam runs on, the substitute fails in agniṃ jyotir nicāyya. The Subodhinī: nicāyya is cāy with ktvā, replaced by lyap in a compound (SK 3332, chapter 69).
Rule. Nine forms are fixed forms (nipātana) in the Veda. Bhaṭṭoji names the root, the tense and the ending of each. The Subodhinī gives each irregular step and the ordinary-language form.
Example.
indraś ca viṣṇo yad apaspṛdhethām "when you two, Indra and Viṣṇu, contended": spardh in laṅ, with āthām. The Subodhinī: doubling, saṃprasāraṇa of the r and loss of the a are by nipātana; the ordinary form is aspardhethām.arkam ānṛcuḥ "they sang a song of praise"; vasūny ānṛhuḥ "they were worthy of riches": arc and arh in liṭ, with us. Ordinary forms: ānarcuḥ, ānarhuḥ.cicyuṣe: cyuṅ in liṭ, with thās. The Subodhinī: saṃprasāraṇa of the abhyāsa, and no iṭ, by nipātana. Ordinary form: cucyuviṣe.yas tityāja "who abandoned": tyaj with ṇal. Saṃprasāraṇa of the abhyāsa by nipātana. Ordinary form: tatyāja.śrātās ta indra somāḥ "cooked, O Indra, are your soma drinks"; śritā no grahāḥ: śrīñ "to cook" in the niṣṭhā, the past participle. The Subodhinī: śrā replaces śrīñ in śrātāḥ.āśīraṃ duhe and madhyata āśīrtaḥ: śrīñ with ā, with kvip and in the niṣṭhā respectively. The Subodhinī: the root becomes śir. In āśīrtaḥ the niṣṭhā t does not become n, by nipātana, and the vowel is lengthened by hali ca (SK 354).Watch out. For apaspṛdhethām scholars discuss this. The Subodhinī reports others who read apa + spardh, with no augment a by a Vedic rule. On their view the ordinary form is apāspardhethām.
Rule. In the Veda, ā optionally replaces the ec vowel of khid (khida "to be dejected"). ec is the name for e, o, ai, au. The Subodhinī: "ā" and "ec" come from ādeca upadeśe 'śiti (SK 2370, chapter 44), and "optionally" from vibhāṣā līyateḥ (SK 2509, chapter 47).
Example. Bhaṭṭoji: cikhāda means cikheda "he was dejected".
cikheda → ā for its e → cikhāda.
The Subodhinī: the parasmaipada ending here is by exchange (vyatyaya).
Watch out. Both forms stand in the Veda: cikhāda and cikheda.
Rule. In the Veda, śīrṣan replaces the word śiras "head".
Example. śīrṣṇaḥ śīrṣṇo jagataḥ. The Subodhinī derives śīrṣṇaḥ:
śiras + ṅas → śīrṣan replaces śiras → śīrṣan + as
→ the a of an is lost by allopo 'naḥ (SK 234, chapter 8) → śīrṣn + as
→ n becomes ṇ after ṣ by raṣābhyāṃ no ṇaḥ samānapade (SK 235, chapter 8) → śīrṣṇaḥ.
The Subodhinī mentions another route to the ṇ, by SK 197 (chapter 8). That route holds on the view that a substitute counts as the original even for an operation conditioned by what precedes.
vārāhī + au → vārāhī (pūrvasavarṇadīrgha) or vārāhyau (yaṇ); yāni → (śi lost) yāThree rules loosen vowel changes in the Veda. SK 3515 makes optional a long vowel that chapter 8 bars, and its option reaches two more rules. SK 3516 drops the neuter ending śi. SK 3517 gives ay before yat in two fixed words. The running examples are the dual vārāhī beside vārāhyau, and yā for yāni.
Rule. In the Veda, after a long vowel, before jas or an ic vowel (ic, any vowel except a and ā), the pūrvasavarṇadīrgha is optional. The pūrvasavarṇadīrgha is a single long vowel of the same kind as the preceding one, standing for both (SK 164, chapter 7). The Subodhinī: the words come from nād ici (SK 165, chapter 7) and dīrghāj jasi ca (SK 239, chapter 8). In ordinary usage SK 239 bars this long vowel after a long vowel before jas or ic.
Example.
vārāhī + au (dual) → with the long vowel → vārāhī; without it, ī becomes y (yaṇ) → vārāhyau.
The Subodhinī: vārāhī is "made of a boar", with añ and ṅīp, the feminine affix ī.
Before jas: mānuṣīr īḷate viśaḥ "the human clans praise".
Applies when / blocked by. Bhaṭṭoji: the option also joins the next two sūtras, by vākyabheda. vākyabheda is splitting one sentence, so that a word also reads into a neighbouring rule.
śamīṃ ca śamyaṃ ca; sūrmyaṃ suṣirām iva "like a hollow tube".ijyamānaḥ, yajyamānaḥ.The Subodhinī: in śamyam, when the single vowel fails, the long vowel does not come either; only yaṇ. For yajyamānaḥ: yaj with śānac (the participle affix āna), yak (the affix ya) and muk (the augment m), and saṃprasāraṇa (SK 2412, chapter 44). Without the pūrvarūpa, yaṇ comes.
Watch out. This sūtra is 6.1.106. Bhaṭṭoji moves it ahead of 6.1.70 (SK 3516). Separately, its chandasi reaches 6.1.115, prakṛtyāntaḥpādam (SK 3518), by skipping 6.1.107–114, which are not Vedic. The Subodhinī calls this leap maṇḍūkapluti, "the frog's jump".
Rule. In the Veda, śi is lost, variously. śi is the substitute for jas and śas after a neuter stem (SK 312, chapter 10). The word "lost" (lopaḥ) is Bhaṭṭoji's gloss.
Example. yā te gātrāṇām (yā for yāni); tā tā piṇḍānām (tā tā for tāni tāni). The Subodhinī derives yā:
yad + jas → śi (SK 312) → yad + i
→ śi is lost by this rule
→ the lost ending still has its effects (pratyayalakṣaṇa): a for the d by tyadādīnām aḥ → ya
→ num (SK 314, chapter 10); lengthening by sarvanāmasthāne ca; loss of the n → yā.
Applies when / blocked by. A vārttika: in the Veda, in eman and the like, the single substitute is pararūpa, the form of the following vowel. apāṃ tveman; apāṃ tvodman.
Rule. In the Veda, bhayya and pravayyā are fixed forms (nipātana). The Subodhinī: for bhī and for vī with pra, ay replaces the vowel before the affix yat.
Example.
bhī + yat → ay for ī → bhay + ya → bhayyaḥ "that from which one is afraid" (bibhety asmāt). The Subodhinī: yat here is in the sense of apādāna, the source, by kṛtyalyuṭo bahulam (SK 2841, chapter 66).
pra + vī + yat → pravayyā. Bhaṭṭoji: the nipātana holds only in the feminine. Elsewhere: praveyam.
Why "in the Veda"? Outside it: bheyam, praveyam.
Applies when / blocked by. A vārttika adds hradayyā: hradayyā āpaḥ "waters that are in a lake (hrada)". The affix is yat by bhave chandasi (SK 3456, Part II). The Subodhinī: ay replaces the a. The reading is uncertain (hṛda / hrada); hrada "lake" gives the sense.
upaprayantaḥ + adhvaram → upaprayanto adhvaram (o kept; no pūrvarūpa by eṅaḥ padāntād ati, SK 86)In ordinary sandhi a word-final e or o swallows a following short a. This is the pūrvarūpa of eṅaḥ padāntād ati (SK 86, chapter 3): te + avadan → te 'vadan. Chapter 4 taught prakṛtibhāva, where a vowel "stays as it is" (prakṛtyā) and no sandhi takes place. These eight rules give the Vedic cases. First comes the Ṛgveda, inside a quarter-verse (pāda). Then come the Yajus formulas, which have no pādas. Last comes the particle ā before a vowel. The running example is upaprayanto adhvaram "approaching the sacrifice", where the o stays before a.
Rule. Inside a pāda of a ṛc, an eṅ (e or o) stays as it is before an a. This holds only when that a is not followed by v or y. The Subodhinī: eṅ is carried forward from SK 86 and recast as the subject. chandasi comes from vā chandasi (SK 3515) by the frog's jump. The pāda meant is that of a ṛc only, not of a classical verse.
Example.
upaprayanto + adhvaram → by SK 86 the a would merge into the o → here the o stays → upaprayanto adhvaram.
Also sujāte aśvasūnṛte.
Applies when / blocked by. Why "inside a pāda"? etāsa ete 'rcanti. The Subodhinī: the e of ete ends one pāda, and the a begins the next. So both sounds are not inside one pāda, and SK 86 applies.
Why "not followed by v or y"? te 'vadan: the a is followed by v, so SK 86 applies.
Watch out. Both conditions must hold. The two sounds must be in the same pāda, and the a must not stand before v or y.
Rule. Before avyāt, avadyāt, avakramuḥ, avrata, ayam, avantu and avasyu-, an eṅ stays as it is. It stays even though the a is followed by v or y.
Example.
vasubhir no avyāt "may he protect us with the Vasus".
mitramaho avadyāt "O you of friendly greatness, from blame".
mā śivāso avakramuḥ; te no avrata.
śatadhāro ayaṃ maṇiḥ "this jewel of a hundred streams".
te no avantu "may they help us"; kuśikāso avasyavaḥ "the Kuśikas, seeking help".
The Subodhinī: the seven words are quoted forms (anukaraṇa) joined into one compound. avyāt is the benedictive of av "to protect". avakramuḥ is kram with ava in liṭ, left undoubled by the Vedic option (see SK 3509).
Applies when / blocked by. Bhaṭṭoji adds a note. The Bahvṛcas, the reciters of the Ṛgveda, do not keep the vowel in te no 'vantu, rathatūḥ, so 'yam agāt, te 'ruṇebhiḥ and the like. He accounts for this by bāhulaka, the latitude of "variously" in Vedic rules. In the Prātiśākhya, the phonetic manual of the Veda, it is stated outright.
Rule. In the Yajus, the word uras, ending in eṅ (uro), stays as it is before a. The Yajus is the Veda of prose sacrificial formulas.
Example. uro antarikṣam "the wide mid-air".
Applies when / blocked by. The Subodhinī raises an objection: SK 3518 already does this, so the rule is useless. Bhaṭṭoji's answer: the Yajus has no pādas. So this rule is for an eṅ that is not inside a pāda. That "stays as it is" and "before a" come from SK 3518 is not stated by Bhaṭṭoji or his commentators; it is inferred from his gloss.
Rule. In the Yajus, these words stay as they are before a: āpo, juṣāṇo, vṛṣṇo, varṣiṣṭhe, and ambe and ambāle when they stand before ambike.
Example.
āpo asmān mātaraḥ.
juṣāṇo agnir ājyasya "Agni, enjoying the clarified butter".
vṛṣṇo aṃśubhyām "with the two stalks of the bull".
varṣiṣṭhe adhi nāke "upon the highest vault".
ambe ambāle ambike.
The Subodhinī: the words are quoted forms. āpo quotes a jas form, juṣāṇo a su form, vṛṣṇo a śas form, varṣiṣṭhe a ṅi form. "Before ambike" qualifies ambe and ambāle.
Watch out. Bhaṭṭoji: from this very wording, the shortening of ambārthanadyor hrasvaḥ (SK 267, chapter 8) does not take place. SK 267 shortens the address form of words meaning "mother". The sūtra itself shows ambe and ambāle unshortened. How the wording proves this is not stated by Bhaṭṭoji or his commentators; it is inferred.
Rule. In the Yajus, two kinds of eṅ stay as they are before a. One is an eṅ in the word aṅga "limb". The other is any eṅ before the a that begins aṅga.
Example. prāṇo aṅge aṅge adīdhyat.
prāṇo + aṅge: an eṅ before the initial a of aṅga → prāṇo aṅge.
aṅge + aṅge: the e of aṅga, and also before aṅga → aṅge aṅge.
aṅge + adīdhyat: the e of aṅga before an ordinary a → aṅge adīdhyat.
Watch out. The Subodhinī explains ca. Without it the sense would be narrow: the eṅ of aṅga only before the initial a of aṅga. That covers aṅge aṅge but not aṅge adīdhyat or prāṇo aṅge. With ca the two conditions stand apart, and all three junctions are covered.
Rule. In the Yajus, an eṅ stays as it is before an anudātta a that is followed by a ku consonant (k, kh, g, gh, ṅ) or by dh.
This rule turns on accent. A Vedic syllable is udātta (raised), anudātta (not raised) or svarita (falling from a raised tone). Here the udātta is marked with an acute (á) and an independent svarita with a grave (à). An anudātta, and a svarita that only follows an udātta, are left unmarked.
Example.
ayáṃ só agníḥ: the udātta syllables are yaṃ, so and gni. The initial a of agníḥ is anudātta, and g follows it. So the o of só stays.
ayáṃ só adhvaráḥ: the udātta of adhvaráḥ is on ra. Its initial a is anudātta; after the udātta só it is recited as a dependent svarita, which is left unmarked. And dh follows it.
The Subodhinī: agni has the udātta on its last syllable, being formed with the uṇādi affix ni. The uṇādi affixes are a separate list of affixes that form noun stems from roots. adhvara has it on its last syllable by the accent of the stem.
Applies when / blocked by. Why "anudātta"? atho 'gre rudre. Bhaṭṭoji: the word agra has the udātta on its first syllable (ágra, udātta a). The a is not anudātta, so SK 86 merges it.
Why "followed by ku or dh"? so 'yam agnimataḥ: the a of ayam is followed by y, so SK 86 applies.
Watch out. Both conditions are needed. The a must be anudātta, and the next consonant must be a ku sound or dh.
Rule. In the Yajus, an eṅ stays as it is before avapathāḥ, when the initial a of avapathāḥ is anudātta.
Example. trī́ rudrébhyo avapathāḥ: the udātta syllables are trī and dre. The verb is wholly anudātta.
Bhaṭṭoji: avapathāḥ is vap in laṅ, with thās. It loses its accent by tiṅṅatiṅaḥ (SK 3935, chapter 71). So its first a is anudātta, and the o of rudrébhyo stays. That trī́ bears the udātta is read from the marking; it is not stated by Bhaṭṭoji or his commentators; it is inferred.
Applies when / blocked by. Why "anudātta"? yád rudrébhyo 'vapathāḥ: the udātta syllables are yad and dre. Bhaṭṭoji: after yad, nipātair yadyadi- (SK 3937, chapter 71) blocks the loss of accent. The verb keeps its own accent, so its first a is not anudātta, and SK 86 merges it. The verb's own udātta falls on the augment a (ávapathāḥ), by luṅlaṅlṛṅkṣv aḍ udāttaḥ (SK 2206, chapter 44). Bhaṭṭoji says only that the loss of accent does not happen; the place of the udātta is not stated by him or his commentators; it is inferred.
Rule. In the Veda, before a vowel, the particle āṅ (ā) becomes nasal (anunāsika) and stays as it is.
Example. abhra ā̃ apaḥ; gabhīra ā̃ ugraputre. The ā̃ does not merge with the following vowel.
Applies when / blocked by. The Subodhinī asks whether the ā of abhra ā̃ apaḥ is really āṅ. A verse gives the ā marked with ṅ four senses. There the ā has none of them; it expresses the sense of the seventh case. Following the Bhāṣya (the Mahābhāṣya, Patañjali's great commentary on the sūtras), the Subodhinī answers that ā lacks the ṅ only in a sentence and in recollection. Everywhere else it is āṅ. That the rule draws on SK 90 (chapter 4) and SK 3518 is not stated by Bhaṭṭoji or his commentators; it is inferred from his gloss.
A vārttika: in the Veda, īṣā akṣa and the like stay as they are. īṣā akṣo hiraṇyayaḥ; jyā iyam; pūṣā aviṣṭu.
Watch out. The nasal ā̃ is not a marker. The Subodhinī: a nasal vowel is a marker only in the teaching form of an element (upadeśe 'janunāsika it), and this is not such a form.
syaḥ + bhānuḥ → (su lost before a hal) sya bhānuḥ; hari + candra → (suṭ) hariścandraḥTwo short rules close the Vedic part of 6.1. One extends to sya the loss of su that chapter 7 taught for eṣaḥ and saḥ. The other gives the augment s before candra in a compound. The running examples are eṣa sya bhānuḥ and hariścandraḥ.
Rule. In the Veda, the case ending su of sya is lost before a consonant (hal), variously. sya is the first-case form of the pronoun tyad "that".
Example. eṣa sya bhānuḥ. The Subodhinī:
tyad + su → a for the d (the tyadādi rule) → t becomes s before su → syaḥ
→ before the bh of bhānuḥ the su is lost by this rule → sya bhānuḥ.
In eṣa the su is lost by etattadoḥ sulopo 'kor anañsamāse hali (SK 176, chapter 7).
Applies when / blocked by. SK 176 covers etad and tad only. This rule brings sya in, in the Veda. That "before a consonant" is carried from SK 176 is not stated by Bhaṭṭoji or his commentators; it is inferred from his gloss.
Rule. In a mantra, a verse or formula of the Veda, the augment suṭ (s) comes to candra "bright, moon". candra must be the second member of a compound (uttarapada), after a short vowel. The Subodhinī: suṭ is carried forward from suṭ kāt pūrvaḥ (SK 2553, chapter 51).
Example.
hari + candra → the augment s before candra → hari-s-candra → hariścandraḥ: hariścandro marudgaṇaḥ.
Also suścandra dasma.
The s appears as ś before c; that step is not stated here; it is inferred.
Applies when / blocked by. The Subodhinī: the augment goes to candra, not to the vowel before it. When a rule names one item in the fifth case and another in the seventh, the fifth-case naming is stronger. So Bhaṭṭoji writes "to the word candra as uttarapada".
saha + māda → (sadha for saha) sadhamāde; aṣṭan + pāda → (ā for the final) aṣṭāpadīThese rules belong to 6.3, the section on the prior member of a compound (pūrvapada) and its changes. Chapter 25 taught the ordinary rules of this section. The Veda adds a fixed dvandva, a substitute for saha, two substitutes for ku, three fixed forms of sah, and ā for aṣṭan. The running examples are sadhamāde "at the common feast" and aṣṭāpadī "eight-footed".
Rule. In the Veda, pitarāmātarā "father and mother" is a fixed form (nipātana) of a dvandva, a compound of members joined by "and". By ca, the reverse order stands too.
Example. ā mā gantāṃ pitarā mātarā ca "may father and mother come to me". The Subodhinī: in the first member, ā is substituted by nipātana. In the second, the case ending is replaced by ā (SK 3561, Part IV), with guṇa by ṛto ṅisarvanāmasthānayoḥ (SK 275, chapter 8).
Reverse order: na mātarā pitarā nū cid iṣṭau.
Applies when / blocked by. Bhaṭṭoji cites here samānasya chandasy amūrdhaprabhṛtyudarkeṣu (SK 1012, chapter 25). In the Veda, sa replaces samāna "same", except before mūrdhan, prabhṛti and udarka. The Subodhinī: sa is carried from sahasya saḥ saṃjñāyām (SK 1009, chapter 25).
samāno garbhaḥ → sagarbhaḥ → "being there" (tatra bhavaḥ), with yat → sagarbhyaḥ.
Counter-examples (Subodhinī): samānamūrdhā, samānaprabhṛtayaḥ, samānodarkāḥ.
A vārttika on viṣvagdevayoś ca ṭer adryañcatāv apratyaye (SK 418, chapter 11): in the Veda, in the feminine, it applies variously. Bhaṭṭoji: the substitute adri for the ṭi of viṣvak and deva comes variously.
Without adri: viśvācī ca ghṛtācī ca.
With adri: devadrīcī nayata devayantaḥ; also kadrīcī, from kim.
The Subodhinī: viśvācī is "going to all": añc with kvin (an affix that, like kvip, leaves nothing of itself), ṅīp, loss of the a of añc, and lengthening by cau. Here adri was due but fails by bāhulaka. In kadrīcī "going badly" it does come, for the ṭi of kim. The Subodhinī takes viṣvak and deva to stand for pronouns as well.
Rule. In the Veda, sadha replaces saha "with" before māda or stha as the second member. The Subodhinī: saha is carried from sahasya sadhriḥ (SK 422, chapter 11).
Example.
saha + māda → sadhamāda → seventh case sadhamāde: indra cāsmin sadhamāde.
saha + stha → sadhastham: somaḥ sadhastham.
The Subodhinī: sadhamāda is "where the gods rejoice together", the sacrifice, made with ghañ. sadhastha is "what stands together", with ka by āto 'nupasarge kaḥ.
Watch out. The Subodhinī: the a of māda in the sūtra is only for pronunciation. So the rule also covers mād, the kvip form: arvāg indra sadhamādo vahantu.
Rule. In the Veda, when pathin "path" is the second member, kava and kā replace ku "bad".
Example. kavapathaḥ, kāpathaḥ, and also kupathaḥ, all "a bad path".
Applies when / blocked by. The source rules would be koḥ kat tatpuruṣe 'ci (SK 1027) and kavaṃ coṣṇe (SK 1033) for ku and kava. kā would come from kā pathyakṣayoḥ (SK 1030). All three are in chapter 25. This link is not stated by Bhaṭṭoji or his commentators; it is inferred. The third form kupathaḥ shows the substitutes are not obligatory. Bhaṭṭoji does not say where the option comes from. There is no Subodhinī on this sūtra.
Rule. In the nigama, the Vedic text, sāḍhyai, sāḍhvā and sāḍhā are fixed forms. Bhaṭṭoji: all three are from sah "to overcome". The first two are made with ktvā, the third with tṛn.
Example. The Subodhinī derives sāḍhyai. The replacement of ktvā by dhyai is itself part of the nipātana.
sah + ktvā → dhyai for ktvā → sah + dhyai
→ h becomes ḍh by ho ḍhaḥ (SK 324, chapter 11) → saḍh + dhyai
→ dh becomes ḍh after the retroflex (ṣṭutva) → saḍh + ḍhyai
→ the first ḍh is lost by ḍho ḍhe lopaḥ (SK 2335, chapter 44) → sa + ḍhyai
→ the a before the lost ḍh is lengthened by ḍhralope pūrvasya dīrgho 'ṇaḥ (SK 174, chapter 7) → sāḍhyai.
sāḍhvā is formed likewise (Subodhinī).
With tṛn: marudbhir ugraḥ pṛtanāsu sā́ḷhā "fierce with the Maruts, the conqueror in battles". The udātta is on the first syllable, sā.
Applies when / blocked by. The Subodhinī explains the choice of tṛn over tṛc by accent. With tṛc the word would have the udātta on its last syllable (antodātta). The mantra has it on the first (ādyudātta), and tṛn gives that. The Subodhinī also takes iti in the sūtra as "of this kind". So the nipātana reaches the past participle too: aṣāḷho agne vṛṣabhaḥ.
Watch out. The mantra writes ḷh, not ḍh. Bhaṭṭoji quotes the Prātiśākhya: a ḍ between two vowels becomes ḷ, and a ḍh there becomes ḷh. So sāḍhā is recited sāḷhā.
Rule. In the Veda too, ā replaces the final of aṣṭan "eight" before a second member. That this comes from aṣṭanaḥ saṃjñāyām (SK 1046, chapter 25), which gives it in names, is not stated by Bhaṭṭoji or his commentators; it is inferred.
Example. aṣṭau pādā asyāḥ "she whose feet are eight":
aṣṭan + pāda (bahuvrīhi) → ā for the final → aṣṭā-pāda
→ the final of pāda is lost by saṃkhyāsupūrvasya (SK 879, chapter 19) → aṣṭā-pād
→ ṅīp by pādo 'nyatarasyām (SK 457, chapter 15) → aṣṭā-pād + ī
→ pād becomes pad before the ī by pādaḥ pat (SK 414, chapter 11) → aṣṭāpadī "eight-footed".
This last step is not stated by Bhaṭṭoji or his commentators; it is inferred.
aśva + matup → (a lengthened) aśvāvatī; abhi + su + naḥ → (i lengthened) abhī ṣu ṇaḥNine of the last ten rules of this part lengthen vowels. The word "long vowel" (dīrgha) runs through SK 3533–3540 from ḍhralope pūrvasya dīrgho 'ṇaḥ (SK 174, chapter 7). The first rules lengthen the final of a stem before matup, before case endings, and of particles and short verb forms in a mantra. SK 3539 opens the lengthening to other words as well. Two rules of 6.4 then make the lengthening of chapter 8 optional in the Veda. Last comes the fixed form janitā. The running examples are aśvāvatī "rich in horses" and abhī ṣu ṇaḥ.
Rule. In a mantra, before the affix matup, the final vowel of soma, aśva, indriya and viśvadevya is lengthened.
Example.
aśva + matup → the final a lengthened → aśvā + mat
→ m becomes v after the ā by mādupadhāyāś ca mator vo 'yavādibhyaḥ (SK 1897, chapter 39) → aśvāvat-
→ feminine ī → aśvāvatī → second case aśvāvatīm.
The v for m is not stated by Bhaṭṭoji or his commentators; it is inferred.
aśvāvatīṃ somāvatīm "rich in horses, rich in soma".
indriyāvān madintamaḥ "possessed of vigour, most exhilarating".
viśvakarmaṇā viśvadevyāvatā.
Watch out. The Subodhinī lists the bases as five: soma, aśva, indriya, viśva, devya. The example viśvadevyāvatā shows the single stem viśvadevya.
Rule. In a mantra, the final of oṣadhi "plant" is lengthened before a case ending (vibhakti) other than one of the first case.
Example.
oṣadhi + bhyas → i lengthened → oṣadhībhyaḥ: yad oṣadhībhyaḥ.
oṣadhi + su → oṣadhīṣu: adadhāty oṣadhīṣu.
Applies when / blocked by. The Subodhinī meets an objection by accent. Could the ī come from the feminine affix ṅīṣ by kṛdikārād aktinaḥ, making this rule needless? No. ṅīṣ would put the udātta on the last syllable. The word wants it on the first: óṣadhībhyaḥ, óṣadhīṣu, with the udātta on o. That accent comes from the phiṭ-sūtra laghāvante… (phiṭ 42). So the ī must be this rule's lengthening.
Watch out. Before a first-case ending the rule does not apply.
Rule. In a ṛc, the final of tu, nu, gha, makṣu, taṅ, ku, tra and uruṣya is lengthened.
Example.
ā tū na indra; nū martaḥ.
uta vā ghā syālāt "or else from the wife's brother". The Subodhinī: syāla is the wife's brother, in the fifth case.
makṣū gomantam īmahe.
bharatā jātavedasam (taṅ).
kūmanāḥ (ku); atrā te bhadrā, yatrā naś cakrā (tra).
uruṣyā ṇaḥ. The Subodhinī: uruṣya is a kaṇḍvādi stem in yak, "to protect", in loṭ. Its hi is lost by ato heḥ (SK 2202, chapter 44). The ṇ of ṇaḥ is by SK 3649 (Part V).
Applies when / blocked by. Bhaṭṭoji: taṅ means the substitute for the ending tha, in the alternative where it is marked with ṅ. The Subodhinī: in loṭ the second person plural tha becomes ta by loṭo laṅvat (SK 2198) and tasthasthamipāṃ… (SK 2199), chapter 44. That ta is ṅ-marked: bharatā.
Not here: śṛṇota grāvāṇaḥ "hear, O pressing-stones". The Subodhinī: there the substitute is tap by taptanaptanathanāś ca (SK 3567, Part IV). Being p-marked, it is not ṅ-marked.
Watch out. gha here is the word gha itself. It is not the name gha for the affixes tarap and tamap. The Subodhinī: in the Veda, lengthening is seen only for the word gha.
Rule. In a ṛc, a word ending in an ik vowel (i, u, ṛ, ḷ) is lengthened before the particle su (suñ). Bhaṭṭoji: "in a ṛc" and "long vowel" still apply. That they come from SK 3535 just above is not stated by Bhaṭṭoji or his commentators; it is inferred.
Example.
abhi + su + naḥ
→ the final i of abhi lengthened → abhī su naḥ
→ s becomes ṣ by suñaḥ (SK 3644, Part V) → abhī ṣu naḥ
→ n becomes ṇ by naś ca dhātusthoruṣubhyaḥ (SK 3649, Part V) → abhī ṣu ṇaḥ.
abhī ṣu ṇaḥ sakhīnām.
Rule. In a mantra, the final a of a verb form (tiṅanta) of two vowels is lengthened.
Example. vidmā hi cakrā jarasam.
vidma "we know" → vidmā. The Subodhinī: vid in laṭ, with ma for mas by vido laṭo vā (SK 2464, chapter 45).
cakra "you (plural) made" → cakrā. The Subodhinī: liṭ, second person plural.
Watch out. Bhaṭṭoji says "in a mantra". The Subodhinī words it "in a ṛc".
Rule. In a mantra, the final of a nipāta, a particle, is lengthened.
Example. eva → evā: evā hi te. The Subodhinī: eva is a nipāta because it is listed in the cādi group, the list of particles that begins with ca "and".
Rule. A long vowel is seen for other words too, in a prior member. Bhaṭṭoji glosses: "for others too, standing in a prior member". The Bālamanoramā, on SK 846 where this sūtra is cited, supplies "long vowel" from SK 174. It adds that "others" means others than those named from nahivṛtivṛṣi… (SK 1037, chapter 25) on.
Example. pūruṣaḥ "man". daṇḍādaṇḍi "a fight in which they struck each other with sticks" (SK 846, chapter 19). Earlier, sādanya (SK 3492) got its long first vowel from this rule.
Applies when / blocked by. The Bālamanoramā on SK 846 limits the rule. In a bahuvrīhi, the final of the prior member is lengthened only where an action is done mutually (karmavyatihāra). It draws this from the word "is seen" (dṛśyate). There is no Subodhinī on this sūtra.
Rule. In the Veda, before nām, a long vowel comes optionally: "both ways". The Subodhinī: nāmi (SK 209, chapter 8) and "long vowel" from SK 174 are carried forward. SK 209 gives the lengthening before nām; here it is optional.
Example. dhātā dhātṝṇām "the Ordainer of ordainers": so the Bahvṛcas recite it. The Taittirīyas read only the short vowel: dhātṛṇām.
Rule. In the nigama, a penultimate vowel (upadhā) that has ṣ before it is optionally lengthened. This happens before a sarvanāmasthāna ending, other than the ending of address (sambuddhi). The sarvanāmasthāna endings are the "strong" endings of chapter 8. That the condition comes from sarvanāmasthāne cāsambuddhau (SK 250, chapter 8) is not stated by Bhaṭṭoji or his commentators; it is inferred from his gloss.
Example. ṛbhukṣāṇam and ṛbhukṣaṇam, both second case singular of ṛbhukṣin.
ṛbhukṣin + am
→ the i becomes a by ito 't sarvanāmasthāne (SK 366, chapter 11; Subodhinī) → ṛbhukṣan + am
→ the penultimate a, after ṣ, optionally lengthened → ṛbhukṣān + am / ṛbhukṣan + am
→ n becomes ṇ after ṣ (SK 235, chapter 8, as at SK 3514) → ṛbhukṣāṇam / ṛbhukṣaṇam.
That the lengthened vowel is this a, and the ṇ step, are not stated by Bhaṭṭoji or his commentators; they are inferred. The Subodhinī: ṛbhukṣin is a fixed form of the uṇādi list.
Applies when / blocked by. Why "in the nigama"? Outside it only the long vowel stands: takṣā, takṣāṇau.
Rule. In a mantra, janitā is a fixed form. Bhaṭṭoji: the loss of ṇi before tṛc with iṭ is established by nipātana.
Example. yo naḥ pitā janitā "who is our father, our begetter". The Subodhinī: janitā means janayitā, the agent noun of the causative of jan.
jan + ṇic → jani
→ jani + iṭ + tṛc
→ ṇi lost by nipātana → jan-i-tṛ → janitā.
The regular form keeps the ṇi: janayitā.
Watch out. ṇer aniṭi (SK 2313, chapter 44) drops ṇi only before an affix without iṭ. Here iṭ is present, so only the nipātana can drop it. Part IV opens with śamitā yajñe (SK 3543), a fixed form of the same kind.
naś + luṅ → (āṭ for an aṅga not beginning with a vowel) ānaṭ; dhā + liṭ, jha → ire → re → dadhrePart IV takes the Vedic rules of the aṅga. The aṅga is the base to which an affix is added, and book 6, pāda 4 changes it. Chapter 44 taught the ordinary rules that these Vedic rules bend. The past tenses take the augment aṭ (SK 2206). An augment (āgama) is a sound added to a word without replacing anything. A root beginning with a vowel takes āṭ instead (SK 2254). Neither comes when the verb is construed with the prohibitive mā (māṅ, SK 2228). And in the perfect (liṭ) the ending jha becomes ire (SK 2241). In the Veda, āṭ reaches roots that begin with a consonant: naś + luṅ gives ānaṭ. The augments may also fail, or come with mā. And ire is shortened to re: dadhre, "they received".
Rule. In the sacrifice (yajñe), the word śamitā is a nipātana. A nipātana is a finished form stated in a sūtra; its irregular steps are accepted as given. Bhaṭṭoji glosses the word as śamayitā.
Example.
śam + ṇic + iṭ + tṛc → the ṇi is dropped by nipātana → śam-i-tṛ → śamitā.
With the ṇi kept, the form is śamayitā, the word Bhaṭṭoji uses to give the sense.
Applies when / blocked by. The Subodhinī says only "the nipātana is as before". It points back to janitā in SK 3542, the rule just before (Part III of this chapter). There too the ṇi is dropped before tṛc with iṭ.
Watch out. The nipātana is restricted to the sacrifice (yajñe).
Rule. In the Veda, the vowel of yu and plu is made long before the affix lyap. Bhaṭṭoji says "when lyap follows" is carried forward. It comes from lyapi laghupūrvāt (SK 3336, chapter 69). lyap is the substitute of ktvā "having done" in a compound (SK 3332, chapter 69).
Example.
vi + yu + lyap → u lengthened → viyūya.
vi + plu + lyap → viplūya.
Applies when / blocked by. Only yu and plu are named. The long vowel is the Vedic form; Bhaṭṭoji gives no further conditions.
Watch out. Bhaṭṭoji ends this unit by recalling āḍ ajādīnām (SK 2254, chapter 44): "āṭ for aṅgas beginning with a vowel". That recall does not belong to this rule. It is the frame for the next sūtra, which extends āṭ. Bhaṭṭoji often closes a unit this way, quoting the earlier rule the next sūtra continues.
Rule. In the Veda, āṭ is seen also on aṅgas that do not begin with a vowel. Bhaṭṭoji's gloss is one word: anajādīnām, "also for those not beginning with a vowel". Ordinarily SK 2254 (chapter 44) gives āṭ only to vowel-initial roots.
Example.
naś + luṅ t (the ending tip, reduced to t) → cli is lost by mantre ghasa… (SK 3402, Part I of this chapter)
→ āṭ added → ā-naś-t
→ the single-sound t lost (halṅyābbhyo…, SK 252, chapter 8) → ā-naś
→ ś → ṣ by vraśca… (SK 294, chapter 9); naśer vā (SK 431, chapter 11) does not apply
→ ṣ → ḍ (jaśtva) → ḍ → ṭ (cartva) → ānaṭ.
vṛñ + luṅ → cli lost → guṇa → var → āṭ → āvar → r → visarga → āvaḥ.
Applies when / blocked by. Bhaṭṭoji gives only the two forms. The steps above are the Subodhinī's, except the loss of the ending. That step is presumably the same as for ās at SK 3586. cli is the vikaraṇa of luṅ, the aorist. jaśtva softens a final consonant, and cartva makes it voiceless again.
Watch out. Bhaṭṭoji closes by recalling na māṅyoge (SK 2228, chapter 44): aṭ and āṭ do not come when the verb is construed with māṅ. That is the frame for SK 3546.
Rule. In the Veda, aṭ and āṭ are bahulam, "variable". bahulam lets a rule fail where it should apply and apply where it should fail. So the augments may be absent without māṅ. And they may be present with māṅ. Ordinarily aṭ (SK 2206) and āṭ (SK 2254) always come, and SK 2228 removes them only with māṅ.
Example.
janiṣṭhā ugraḥ sahase turāya, "you were born mighty, for strength, for overpowering".
jan + luṅ thās → no aṭ, though there is no māṅ → janiṣṭhāḥ.
mā vaḥ kṣetre parabījāny avāpsuḥ, "let not the seeds of others be sown in your field".
vap + luṅ → aṭ, though construed with mā → avāpsuḥ.
Applies when / blocked by. The Subodhinī analyses avāpsuḥ. It is luṅ of vap in the object (karman) sense. It is parasmaipada by vyatyaya, "interchange", which in the Veda lets one form stand for another. cli becomes sic (SK 2222, chapter 44), and the root takes vṛddhi.
Watch out. The Subodhinī doubts the second example. Bhaṭṭoji cites it following the Kāśikā. But in the recitation of the Veda the word is only vāpsuḥ, with no aṭ. "Another example of māṅ with the augment must be looked for."
Rule. In the Veda, re replaces ire. Ordinarily ire replaces the third-person plural ending jha of liṭ (SK 2241, chapter 44).
Example.
dhā + liṭ jha → reduplication → dadhā + jha
→ ire for jha (SK 2241) → dadhā + ire
→ re for ire → dadhā + re
→ the ā is lost (āto lopa iṭi ca, SK 2372, chapter 44) → dadhre.
prathamaṃ dadhra āpaḥ, "the waters received first (the embryo)".
Applies when / blocked by. Why is the ā still lost? SK 2372 needs an ending that begins with a vowel, and re does not. Bhaṭṭoji answers: the change to re falls in the ābhīya section. That is a stretch of book 6 in which one rule is asiddha, "treated as not yet done", for another. So the loss of ā still sees ire, and applies.
Watch out. Why does the sūtra say irayoḥ, in the dual? re can itself take the augment iṭ. Then i + re gives ire once more, and re must replace it again. The dual names both *ire*s: the original and the one made by iṭ. The Subodhinī adds that, because of the dual, the paribhāṣā (working maxim) on derived (lākṣaṇika) versus directly taught (pratipadokta) forms does not stop the second change.
vibhū + am → vibhvam / vibhuvam; pat + liṭ ma (+ iṭ) → (upadhā lost) paptima; śru + hi → (dhi; lengthened) śrudhīThese six rules undo three ordinary rules of earlier chapters. na bhūsudhiyoḥ (SK 273, chapter 8) keeps bhū and sudhī from yaṇ, the change of i, u to y, v. In the Veda both treatments occur: vibhvam and vibhuvam. Next, tan and pat lose the a before their last consonant, as in paptima "we have flown". Last, hujhalbhyo her dhiḥ (SK 2425, chapter 45) gives dhi for the imperative ending hi after some roots. The Veda extends dhi to śru, kṛ and others: śrudhī "hear!".
Rule. In the Veda, bhū and sudhī go "both ways" (ubhayathā). They take yaṇ, and they also take iyaṅ and uvaṅ. iyaṅ turns i, ī into iy, and uvaṅ turns u, ū into uv. Ordinarily SK 273 (chapter 8) bars yaṇ for these two, so only iyaṅ and uvaṅ come.
Example.
vibhū + am → yaṇ (ū → v) → vibhvam: vaneṣu citraṃ vibhvam, "wondrous, all-pervading, in the woods".
vibhū + am → uvaṅ → vibhuvam.
sudhī → yaṇ → sudhyaḥ, written sudhyo3 navyam agne in the mantra; or iyaṅ → sudhiyaḥ.
Applies when / blocked by. A vārttika follows. A vārttika is a supplementary rule that Bhaṭṭoji quotes after a sūtra. It says: for tanū and the rest, in the Veda, iyaṅ and uvaṅ are variable.
tanvaṃ puṣema, "may we nourish the body" → or tanuvam.
tryambakam → or triyambakam.
The Subodhinī explains. tanuvam has uvaṅ, which tanū would not otherwise get, since it is not a root. tanvam has yaṇ. That happens where vā chandasi (SK 3515, Part III of this chapter) withholds the single substitute that am would make. tryambaka is Rudra, "he whose three eyes".
Watch out. The "3" in sudhyo3 marks a pluta, a vowel of three measures; this rule does not make it.
Rule. In the Veda, tan and pat lose their upadhā before a kit or ṅit affix. The upadhā is the sound just before the last sound of a word; here, the a. A kit or ṅit affix is one marked with k or ṅ.
Example.
vi + tan + liṭ ire → reduplication → vi-ta-tan-ire → a lost → vitatnire: vitatnire kavayaḥ, "the poets have spread out".
pat + liṭ ma → iṭ → pa-pat-i-ma → a lost → paptima: śakunā iva paptima, "like birds, we have flown".
Applies when / blocked by. The Subodhinī gives the analysis. The a-loss is asiddha. Yet ata ekahalmadhye… (SK 2260, chapter 44) does not apply. That rule would give e for the a and drop the reduplicated syllable. It is kept out because the loss is expressly taught.
Watch out. Ordinary usage has vitenire and petima (Subodhinī). There SK 2260 does apply.
Rule. ghas and bhas also lose their upadhā a before a kit or ṅit affix. This holds before an affix beginning with a consonant (hali), and, by ca "and", before one beginning with a vowel.
Example.
ad + ktin → ghas for ad by bahulaṃ chandasi (SK 3398, Part I)
→ ghas + ti → a lost → ghs + ti
→ s lost by jhalo jhali (SK 2281, chapter 44) → gh + ti
→ t → dh (SK 2280, chapter 44) → gh + dhi → gh → g (jaśtva) → gdhi
→ samāna "common" → sa (SK 1012, chapter 25) → sagdhiḥ: sagdhiś ca me, "and my common meal".
bhas + loṭ tām, with ślu → doubling first, by ślau (SK 2490, chapter 46) → babhas-tām
→ a lost → s lost → t → dh → jaśtva → babdhām: babdhāṃ te harī dhānāḥ, "let your two bay horses chew the grains".
Applies when / blocked by. Bhaṭṭoji gives the forms; the steps are the Subodhinī's. In babdhām the doubling comes first by bāhulaka, the latitude of bahulam. It overrides the a-loss, though that rule is later and always applicable.
Watch out. Bhaṭṭoji closes by recalling hujhalbhyo her dhiḥ (SK 2425, chapter 45). After hu and after an aṅga ending in a consonant other than a nasal or semivowel (jhal), dhi replaces the imperative hi. That is the frame for SK 3551.
Rule. In the Veda, dhi replaces hi after śru, śṛṇu, pṝ, kṛ and vṛ. "dhi for hi" is carried forward from SK 2425 (chapter 45).
Example.
śru + śap + hi → śap lost (bahulaṃ chandasi, SK 3400, Part I) → śru + dhi
→ lengthened by anyeṣām api dṛśyate (SK 3539, Part III) → śrudhī: śrudhī havam, "hear the call".
śṛṇu + hi → dhi → śṛṇudhī: śṛṇudhī giraḥ, "hear the songs".
pṝ + hi → śap lost → ṝ → ur (SK 2494, chapter 46) → long ū (hali ca, SK 354, chapter 11) → pūr + dhi → rāyaspūrdhi, "fill (us) with wealth".
kṛ + hi → kṛdhi: uruṇaskṛdhi, "make wide (room) for us".
Applies when / blocked by. The steps are the Subodhinī's. śṛṇu is śru with śnu, where śru becomes śṛ. After śṛṇu the hi would normally be lost by utaś ca pratyayāt… (SK 2334, chapter 44). It is not lost, because this rule names śṛṇu. In uruṇaskṛdhi, the ṇ comes by SK 3649 and the s for visarga by SK 3635 (both in Part V).
Watch out. Bhaṭṭoji gives no example for vṛ; the Subodhinī comments on one, apāvṛdhi. The Subodhinī glosses pṝ as "to protect"; the root also means "to fill".
Rule. In the Veda, hi, the substitute of the ending sip, is optionally apit. Ordinarily ser hy apic ca (SK 2201, chapter 44) makes hi always apit. apit means "not marked with p". So in the Veda hi may also count as pit.
Example. Bhaṭṭoji gives no worked form. His gloss is three words: "hi is optionally apit."
Applies when / blocked by. This is a rule of book 3, placed among the rules of book 6. An apit ending counts as ṅit, by sārvadhātukam apit (SK 2234, chapter 44). When hi is not apit, it is not ṅit. Then the next rule, aṅitaś ca, can work on it. That is presumably why the rule stands here. The Subodhinī on SK 3553 does say that in rārandhi the hi is not ṅit.
Watch out. There is no Subodhinī on this sūtra.
Rule. dhi replaces hi also when the hi is not ṅit. "dhi for hi" is carried forward from SK 2425 (chapter 45). The Vedic setting is understood from the section, not restated.
Example.
ram + hi → parasmaipada by vyatyaya → śap → ślu → doubling → ra-ram
→ the reduplicated syllable lengthened (the tujādi rule, SK 3509, Part III) → rā-ram + hi
→ dhi → rāram + dhi → the m heard as n before dh → rārandhi, "rejoice".
asme prayandhi, "hold out to us": yam, with luk of śap.
yuyodhi jātavedaḥ, "keep away, O Jātavedas": yu, with ślu of śap.
Applies when / blocked by. luk and ślu both remove an affix; ślu also causes doubling. The Subodhinī says the luk for yam comes by bahulaṃ chandasi (SK 3400). The ślu for yu comes likewise (SK 3401, Part I).
Watch out. In rārandhi the m is not dropped; it is only heard as n before dh. anudāttopadeśa… (SK 2428, chapter 45) drops such a nasal only before a ṅit affix. Here hi is not ṅit (Subodhinī).
ātman + āṅ (= ṭā) → (initial ā lost) tmanā; ṛju + iṣṭhan → (ra for ṛ) rajiṣṭhamThree rules close book 6. Each changes a stem before an ending or affix. ātman loses its first sound before the instrumental ending: tmanā. ṛju "straight" may take ra for its ṛ, by the rule that chapter 37 taught for other stems (SK 1785): rajiṣṭham. And five Vedic words are given ready-made, among them mādhvī and hiraṇyaya.
Rule. In mantras, the first sound of ātman is lost before āṅ. āṅ is the older grammarians' name for ṭā, the third-case singular ending. SK 244 (āṅo nā 'striyām, chapter 8) uses the same name. The word "loss" (lopa) is supplied by Bhaṭṭoji's paraphrase.
Example.
ātman + ṭā → ātman + ā → the initial ā lost → tman + ā → tmanā: tmanā deveṣu, "by himself, among the gods".
Applies when / blocked by. The rule is restricted to mantras (mantreṣu), the verses and formulas of the Veda. Bhaṭṭoji does not name the rule from which lopa is carried.
Watch out. The ending is named by its old name āṅ, not ṭā.
Rule. In the Veda, ra optionally replaces the ṛ of ṛju. This happens before the affixes iṣṭhan, iman and īyasun. "ra for ṛ" is carried forward from ra ṛto halāder laghoḥ (SK 1785, chapter 37). The three affixes are carried from turiṣṭhemeyassu (SK 2008, chapter 41). The Subodhinī names both sources.
Example.
ṛju + iṣṭhan → ra for ṛ → rajiṣṭham: tvaṃ rajiṣṭham anu neṣi, "you lead along the straightest (path)".
Without the change → ṛjiṣṭham.
Applies when / blocked by. vibhāṣā means "optionally". Both forms are correct in the Veda.
Rule. In the Veda, five words are nipātanas: ṛtvya, vāstvya, vāstva, mādhvī and hiraṇyaya. Four of them have yaṇ (u → v) for the final of the stem. The fifth loses the ma of the affix mayaṭ.
Example.
"What is in the season (ṛtu)": ṛtu + yat → yaṇ → ṛtvyam.
"What is on the site (vāstu)": vāstu + yat → vāstvyam; vāstu + aṇ → vāstvam.
madhu + aṇ, feminine → yaṇ → mādhvī: mādhvīr naḥ santv oṣadhīḥ, "may the plants be sweet for us".
hiraṇya + mayaṭ → ma lost → hiraṇyaya: hiraṇyayena savitā rathena, "Savitṛ, with his golden chariot".
Applies when / blocked by. The Subodhinī names the affixes: yat after ṛtu, and aṇ and yat after vāstu. Ordinary derivation would give guṇa (oḥ guṇaḥ, SK 847, chapter 19). The results would be ṛtavya, vāstavya, vāstava and mādhavī. Bhaṭṭoji names only the yaṇ.
Watch out. How does hiraṇya keep its final a before ya? Scholars discuss this. The Subodhinī first says the loss of ma is asiddha. So yasyeti ca (SK 311, chapter 10) does not drop the stem's a. Lengthening by SK 2298 (chapter 44) is kept off by a paribhāṣā, a working maxim, called aṅgavṛtta. Alternatively, only the m is lost. Then the stem's a goes by yasyeti ca, and the a we hear is the affix's. This rule ends book 6.
duh + laṭ jha → ate → (ruṭ) duh-r-ate → (t lost, SK 3563) duhre / duhrate; deva + bhis → devebhiḥBook 7 opens with rules on endings. Four ordinary rules are loosened here. śīṅo ruṭ (SK 2442, chapter 45) adds the augment r to an ending after śī. ato bhisa ais (SK 203, chapter 8) replaces bhis by ais after a. adḍ ḍatarādibhyaḥ… (SK 315, chapter 10) replaces su and am by adḍ after itara and its group: itarat. And SK 3332 (chapter 69) puts lyap for ktvā in a compound. The running examples are duhre "they yield", with r added, and devebhiḥ "with the gods", with bhis kept.
Rule. In the Veda, the augment ruṭ comes variously (bahulam). The frame is śīṅo ruṭ (SK 2442, chapter 45), which Bhaṭṭoji quotes at the head of the unit. ruṭ adds r at the front of the ending.
Example.
duh + laṭ jha → at for jha; ṭi → e → duh + ate
→ ruṭ → duh + r + ate
→ t lost, optionally, by lopas ta ātmanepadeṣu (SK 3563, below) → duh + r + a + e
→ a + e → e by pararūpa (ato guṇe, SK 191, chapter 8) → duhre: dhenavo duhre, "the cows yield milk".
Without the t-loss → duhrate: ghṛtaṃ duhrate, "they yield ghee".
adṛśram asya, "his (rays) have appeared": dṛś, luṅ, with ruṭ.
Applies when / blocked by. The steps of duhre are the Subodhinī's, up to the loss of t. The last step, pararūpa, is the one the Subodhinī gives for the parallel aduhra (SK 3563). pararūpa is the single substitute that takes the form of the second sound. The ṭi is the last vowel with what follows it. In adṛśram, says the Subodhinī, the first-person singular mip stands for the third-person plural, by vyatyaya. Then ruṭ is added.
Watch out. Bhaṭṭoji closes by recalling ato bhisa ais (SK 203, chapter 8), the frame for SK 3558.
Rule. In the Veda, ais replaces bhis after a only variably. So bhis may stay. The frame is ato bhisa ais (SK 203, chapter 8), recalled at the end of SK 3557.
Example.
deva + bhis → no ais → e for the stem's a (bahuvacane jhaly et, SK 205, chapter 8) → devebhiḥ: agnir devebhiḥ, "Agni, with the gods".
Ordinary usage → deva + bhis → ais → devaiḥ.
Applies when / blocked by. Bhaṭṭoji gives the example without comment. There is no Subodhinī on this sūtra.
Watch out. Two sūtras in a row read bahulaṃ chandasi. Each continues a different frame: SK 3557 the augment ruṭ, SK 3558 the substitute ais.
Rule. In the Veda, after itara "other", adḍ does not replace the endings su and am. Ordinarily adḍ ḍatarādibhyaḥ pañcabhyaḥ (SK 315, chapter 10) gives adḍ to itara.
Example.
itara + am → no adḍ → ato 'm (SK 309, chapter 10) → itaram: vārtraghnam itaram, "the other, belonging to the slayer of Vṛtra".
Bhaṭṭoji: "Why 'in the Veda'?" Ordinary usage → itarat kāṣṭham, "the other piece of wood".
Applies when / blocked by. The Subodhinī confirms the example as the am-form. With adḍ absent, am comes by SK 309.
Watch out. Bhaṭṭoji closes by recalling samāse 'nañpūrve ktvo lyap (SK 3332, chapter 69). That is the frame for SK 3560.
Rule. In the Veda, ktvā also stands where lyap would replace it. Ordinarily SK 3332 (chapter 69) requires lyap in a compound whose first member is not nañ "not".
Example.
pari + dhā + ṇic + ktvā → compounded with pari → lyap expected → ktvā kept → paridhāpayitvā: yajamānaṃ paridhāpayitvā, "having made the sacrificer put on (the garment)".
Applies when / blocked by. The Subodhinī reads more into api "also". It brings lyap too, even where no compound would call for it. Otherwise the sūtra would have said just vā chandasi. So the authors of the Kalpasūtras, the ritual manuals, following Vedic usage, write ājyenākṣiṇī ajya, with lyap and no compound.
Watch out. The Vedic form keeps ktvā inside a compound. Ordinary usage would put lyap there, by SK 3332.
panthānaḥ → (su) panthāḥ; vyoman + ṅi → (luk) vyoman; dhīti + ṭā → (pūrvasavarṇa) dhītīOne sūtra with two vārttikas lets a noun ending (sup) be replaced by one of seventeen Vedic substitutes. Chapters 8 to 13 taught the ordinary endings these stand for. Each example below is set against the ordinary form it replaces. The running examples are panthāḥ for panthānaḥ "paths", vyoman for vyomani "in heaven", and dhītī for dhītyā "with thought". This group also has the only accent statements of Part IV. The udātta, the high tone, is marked with an acute (á). Only the udātta of the two forms discussed is marked; elsewhere the mantras are printed without accents.
Rule. In the Veda, any of eleven substitutes may replace a sup ending. They are su, luk, pūrvasavarṇa, ā, āt, śe, yā, ḍā, ḍyā, yāc and āl. luk is loss of the ending. pūrvasavarṇa is a long vowel like the one before. "In the Veda" is carried forward from SK 3560.
Example.
ṛjavaḥ santu panthāḥ, "may the paths be straight", for panthānaḥ.parame vyoman, "in the highest heaven", for vyomani: the ṅi is lost.dhītī matī suṣṭutī, "with thought, with mind, with fair praise", for dhītyā, matyā, suṣṭutyā.yā surathā rathītamobhā devā divispṛśā aśvinā, "the two gods with fine chariots, best of charioteers, both touching heaven, the Aśvins", for yau surathau … divispṛśau.yā deva vidma tā tvā, "you whom, O god, we know as such", for yam and tam.natād brāhmaṇam, "the bowed brāhmaṇa", for natam.na yuṣme vājabandhavaḥ, "not in you, O you whose kin is strength", for yuṣmāsu; asme indrābṛhaspatī, "to us, O Indra and Bṛhaspati", for asmabhyam.uruyā, dhṛṣṇuyā, for uruṇā, dhṛṣṇunā.nābhā pṛthivyāḥ, "at the navel of the earth", for nābhau.tā anuṣṭhyocyāvayatāt, "have those (ribs) taken apart in order". anuṣṭhā is anu + sthā with aṅ, as in vyavasthā; ḍyā replaces the third-case singular.sādhuyā, for sādhu.vasantā yajeta, "he should sacrifice in spring", for vasante.Applies when / blocked by. Two vārttikas add six more substitutes. An upasaṃkhyāna is a vārttika that "adds to the count".
Vārttika 1 adds iyā, ḍiyāc and ī:
urviyā, dārviyā, for uruṇā, dāruṇā.sukṣetriyā, for sukṣetriṇā; the ṭi is lost.dṛtiṃ na śuṣkaṃ sarasī́ śayānam, "like a dried-up skin lying in the pond". Some say ī replaces ṅi in sarasī́.Vārttika 2 adds āṅ, ayāc and ayār:
pra bāhavā sisṛtam, "stretch forth with (your) arms", for bāhunā. gher ṅiti (SK 245, chapter 8) gives guṇa: bāhu + ā → bāho + ā → bāhavā.svapnayā, for svapnena.sa naḥ sindhum iva nāváyā, "… us, as across a river by boat", for nāvā.Watch out. Two forms carry accent statements.
sarasī́ is antodātta: its final syllable sī is udātta. On the view that ī replaces ṅi, the word would have been ādyudātta, udātta on its first syllable. So its final udātta has to be put down to vyatyaya. Bhaṭṭoji prefers another account. sarasī is a feminine in ṅīṣ, and the ṅi is simply lost (luk). "Another example of the ī substitute must be sought."nāváyā has udātta on its penultimate syllable va. That is the accent of a r-marked affix (upottamaṃ riti, SK 3733, chapter 71), and ayār is r-marked.The Subodhinī adds several points. The t of āt is not a marker, by na vibhaktau tusmāḥ (SK 190, chapter 8). In na yuṣme, the rest of the stem goes by śeṣe lopaḥ (SK 385, chapter 11). asme takes no ay before a vowel, since śe is pragṛhya, exempt from sandhi. ḍā, ḍyā and ḍiyāc are ḍ-marked, and a ḍ-marked affix drops the stem's ṭi. In svapnayā the a of ayāc prevents lengthening by supi ca (SK 202, chapter 8); pararūpa follows. The Subodhinī also says panthāḥ is already covered by vyatyayo bahulam (SK 3433, Part I). This sūtra spells that out.
yā and tā in yā deva vidma tā tvā presumably show the substitute ā, since āt would give yāt. In the step bāho + ā → bāhavā, av presumably replaces o before the vowel. The Subodhinī defends aṅ in anuṣṭhā: by sthāgāpāpaco bhāve (SK 3273, chapter 69) ki should come. It answers that the wording vyavasthāyām in SK 218 (chapter 8) shows aṅ is allowed too. It also keeps pra in pra bāhavā a separate word, against the Kāśikā's compound prabāhuṇā.
kṛṇuta → (tāt for ta) kṛṇutāt; as + mas → (final i) smasi; vadh-i-s + am → (maś) m → (īṭ; sic lost) vadhīmChapter 44 taught the verb endings (tiṅ) and their ordinary substitutes. Here the Veda replaces some of them again. am, the first-person singular of the past, becomes m: vadhīm "I slew". The t of a middle ending may be lost: aduhra, śaye. dhvam becomes dhvāt. The second-person plural ta has five Vedic substitutes, among them tāt: kṛṇutāt "make!". And mas "we" gains a final i: smasi "we are", and emasi "we come". Three of these sūtras do not repeat "in the Veda": SK 3565, SK 3567 and SK 3568. For them the Vedic setting is understood from the neighbouring rules and the mantra examples.
Rule. maś replaces the am that stands for mip, the first-person singular ending. Ordinarily tasthasthamipām… (SK 2199, chapter 44) gives am for mip. The a of maś is only for pronunciation. Its ś is a marker, and an affix marked with ś replaces the whole of what it replaces. So maś is just m, and it takes the place of all of am.
Example.
han + luṅ mip → vadha for han (SK 2433–2434, chapter 45) → sic (SK 2222) and iṭ → am for mip (SK 2199) → vadh-i-s-am
→ maś for am → vadh-i-s-m
→ īṭ (astisicaḥ…, SK 2225, chapter 44) → vadh-i-s-ī-m
→ sic lost (iṭa īṭi, SK 2266) → vadh-i-ī-m → i + ī → ī → vadhīm
→ no aṭ in the Veda → vadhīṃ vṛtram, "I slew Vṛtra".
Ordinary usage → avadhiṣam.
Applies when / blocked by. Bhaṭṭoji gives only the maś and īṭ steps; the rest are the Subodhinī's. The Subodhinī explains the ś. Without it, maś would replace only the final m, by alo 'ntyasya (SK 42, chapter 2). Is m for m not pointless anyway? No: such a substitute could serve to block anusvāra, the nasal sound written ṃ, as in SK 126 (chapter 5).
Watch out. The am here is the substitute of mip. It is not the second-case singular am of nouns (Subodhinī). Bhaṭṭoji does not repeat "in the Veda" here; he restates it at the next sūtra.
Rule. In the Veda, a t that is in an ātmanepada (middle) ending is lost. Bhaṭṭoji's gloss opens with "in the Veda" (chandasi), carried forward.
Example.
duh + laṅ jha → at for jha (ātmanepadeṣv anataḥ, SK 2258, chapter 44) → ruṭ (SK 3557) → a-duh-r-ata
→ t lost → a-duh-r-a-a → a + a by pararūpa (SK 191) → aduhra: devā aduhra, "the gods milked".
Ordinary usage → aduhata.
śete → t lost → śe + e → ay for the first e → śaye: dakṣiṇataḥ śaye, "he lies on the south side".
Applies when / blocked by. Bhaṭṭoji: "Why 'in the ātmanepada'?" utsaṃ duhanti, "they milk the spring". Its ending anti is parasmaipada, the active set of endings. So its t stays.
Watch out. This is the rule that SK 3557 used, optionally, for duhre. The steps of aduhra and śaye are the Subodhinī's.
Rule. In the Veda, dhvāt replaces dhvam, the second-person plural middle ending. The Subodhinī glosses: "dhvāt is substituted for dhvam in the Veda."
Example.
vṛ + ṇic + loṭ dhvam → vārayadhvam → dhvāt → vārayadhvāt: antar evoṣmāṇaṃ vārayadhvāt, "keep the heat within".
Applies when / blocked by. The root is vṛ with ṇic, in loṭ, the imperative (Subodhinī).
Watch out. The substitute ends in t, not m. Compare tāt for ta in SK 3566.
Rule. The word yajadhvainam is a nipātana. Before enam "him", the final m of dhvam is lost.
Example.
yajadhvam + enam → m lost → yajadhva + enam → a + e → ai → yajadhvainam: yajadhvainaṃ priyamedhāḥ, "worship him, O Priyamedhas".
Applies when / blocked by. Bhaṭṭoji states only the loss of m. The joining of a + e into ai is presumably the ordinary sandhi.
Watch out. The vṛttikāra, the author of the Kāśikā gloss, said that y also replaces v by nipātana. Bhaṭṭoji calls that an oversight. The Subodhinī agrees. The reading with v is not in doubt, and the Vedabhāṣya, a commentary on the Veda, too says the sūtra teaches only the loss of m.
Rule. In the Veda, tāt replaces ta, the second-person plural ending. It is not the third-person singular ta. The Subodhinī gives two reasons. In the Veda one follows what is attested. And the neighbours dhvam (SK 3564) and mas (SK 3568) are plural endings too.
Example.
kṛṇuta → tāt → kṛṇutāt: gātram asyānūnaṃ kṛṇutāt, "make his limb whole".
gamayata → gamayatāt: sūryaṃ cakṣur gamayatāt, "send his eye to the sun".
Applies when / blocked by. The Subodhinī derives both. kṛṇu- is kṛvi with the affix u (dhinvikṛṇvyor a ca, SK 2332, chapter 44). a replaces the v, and that a is lost (ato lopaḥ, SK 2308). gamaya- is gam with ṇic: gam + ṇic → vṛddhi → gāmi → shortened by mitāṃ hrasvaḥ (SK 2568, chapter 53) → gami → gamaya-. The shortening applies because gam, a root ending in am, counts as mit by a gaṇasūtra of the ghaṭādi group. The ending is the loṭ second-person plural.
Watch out. The ordinary forms are kṛṇuta and gamayata.
Rule. tap, tanap, tana and thana also replace ta. Bhaṭṭoji: "'in place of ta' continues here too".
Example.
śṛṇuta → śṛṇota: śṛṇota grāvāṇaḥ, "hear, O pressing-stones".sunotana pacata brahmavāhase dadhātana draviṇaṃ citram asme, "press (the soma), cook for the one who carries the prayer; grant bright wealth to us".marutas taj jujuṣṭana, "O Maruts, enjoy that", for juṣadhvam.viśve devāso maruto yatiṣṭhana, "you who are of whatever number".Applies when / blocked by. The Subodhinī explains tap. śṛṇu- is śru with śnu, where śru becomes śṛ (SK 2386, chapter 44). tap is p-marked, so it is not ṅit, and the u takes guṇa. In dadhātana the ā is kept, because śnābhyastayor ātaḥ (SK 2483, chapter 45) needs a ṅit ending. sunotana is ṣu with śnu (SK 2523, chapter 48). In jujuṣṭana, juṣ takes parasmaipada by vyatyaya, and ślu. In yatiṣṭhana, yad takes the Vedic affix ḍati. thana replaces the ta of as "to be".
Watch out. Bhaṭṭoji labels tap, tanap and thana. jujuṣṭana is presumably his example of tana, but he does not name it.
Rule. mas, the first-person plural ending, takes a final i. In effect mas gets the augment ik: i marked with k, which goes at the end (ādyantau ṭakitau, SK 36, chapter 2).
Example.
i + mas → final i → imasi (ordinary imaḥ). The mantra has it as emasi: namo bharanta emasi, "bringing homage, we come". The Subodhinī analyses emasi only as iṇ "to go" with mas ending in i; it does not account for the e.
as + mas → a lost (śnasor allopaḥ, SK 2469, chapter 45) → smas → smasi: tvam asmākaṃ tava smasi, "you are ours, we are yours". Ordinary → smaḥ.
Applies when / blocked by. Bhaṭṭoji explains the wording. masi is cited without a case-ending. idanta means "having i as its end part". The Subodhinī adds that the s stays: i is added, it does not replace. Had i been meant to replace the s, the sūtra would have said so. The sūtra avoids "mas ik" so as not to be confused with the sound-group ik.
Watch out. The i is an augment, not a substitute. The s of mas stays, so the ending is heard as masi.
gam + ktvā → (nasal lost, SK 2428) gatvā → (yak) gatvāya; pā + ktvā → (ī, SK 2462) pītvā → (ī for ā) pītvīktvā "having done" was taught in chapter 69. The Veda gives it three new shapes. It may take ya at the end: gatvāya "having gone". After yaj it may end in īnam: iṣṭvīnam. Or its ā may become ī: pītvī "having drunk". None of these three sūtras says "in the Veda". The Subodhinī states it on SK 3569, and for the other two it is understood from that.
Rule. In the Veda, ktvā takes the augment yak. Being k-marked, yak goes at the end: tvā becomes tvāya.
Example.
gam + ktvā → the nasal lost (anudāttopadeśa…, SK 2428, chapter 45) → gatvā → yak → gatvāya: divaṃ suparṇo gatvāya, "the eagle, having gone to heaven".
Applies when / blocked by. The Subodhinī: "ktvā takes the augment yak in the Veda." The steps are its own.
Watch out. The ordinary form is gatvā.
Rule. The word iṣṭvīnam is a nipātana. After yaj, īnam replaces the final of ktvā. The Subodhinī says "after yaj" must be supplied.
Example.
yaj + ktvā → saṃprasāraṇa (vacisvapiyajādīnāṃ kiti, SK 2409, chapter 44) → ij + tvā
→ j → ṣ (SK 294, chapter 9) → iṣ + tvā → ṣṭutva → iṣṭvā
→ īnam for the final ā → iṣṭvīnam: iṣṭvīnaṃ devān, "having worshipped the gods".
Applies when / blocked by. saṃprasāraṇa turns the semivowel of a root into its vowel: ya → i. ṣṭutva turns a dental next to ṣ into a retroflex: t → ṭ. The steps are the Subodhinī's.
Watch out. The ordinary form is iṣṭvā.
Rule. snātvī and forms "of that kind" are nipātanas. In them ī replaces the ā of ktvā. Bhaṭṭoji: the word ādi here means "of that kind" (prakāra), not "and the following".
Example.
snātvā → ī for ā → snātvī: svinnaḥ snātvī malād iva, "like one sweating who, having bathed, (is freed) from dirt".
pā + ktvā → ī for the root's ā (ghumāsthā…, SK 2462, chapter 45) → pītvā → pītvī: pītvī somasya vāvṛdhe, "having drunk of the soma, he grew strong".
Applies when / blocked by. The step for pītvā is the Subodhinī's.
Watch out. Two different *ī*s appear in pītvī. The first replaces the root's ā (SK 2462). The second replaces the ā of ktvā (this rule).
deva + jas → (asuk) deva-as-as → devāsaḥ; go + ām → (nuṭ, at the end of a pāda) gonāmThese six rules add or change sounds in noun stems before endings. Chapters 8 to 13 taught the ordinary rules. jas, the nominative plural, may take as in front: devāsaḥ "gods". ām, the genitive plural, takes n after śrī, grāmaṇī and go. asthi "bone" and its group take an or ī where ordinary usage does not allow it. And dṛś, svavas and svatavas take a nasal before su. SK 3572 and SK 3576 do not say "in the Veda"; for them it is understood from the neighbouring rules.
Rule. After an aṅga ending in a, the ending jas takes the augment asuk. asuk is the sound-sequence as; its u and k are markers.
Example.
deva + jas → asuk → deva-as-as → devāsaḥ, "gods".
brāhmaṇa + jas → brāhmaṇāsaḥ, "brāhmaṇas".
Applies when / blocked by. The Subodhinī: once asuk is added, the s of jas is heard. The final s becomes visarga. The sūtra says jaseḥ, not jasaḥ, out of deference to earlier teachers.
Watch out. asuk does not replace jas. The ending stays, and asuk adds a second as beside it.
Rule. In the Veda, after śrī and grāmaṇī, the ending ām takes the augment nuṭ (n). Bhaṭṭoji writes āmo nuṭ, "nuṭ to ām", as what continues.
Example.
śrī + ām → nuṭ → śrīṇām: śrīṇām udāro dharuṇo rayīṇām, "the bountiful one of fortunes, the upholder of riches".
sūtagrāmaṇī + ām → sūtagrāmaṇīnām, "of charioteers and village-headmen".
Applies when / blocked by. The Subodhinī explains why śrī needs this rule. By vāmi (SK 304, chapter 9), śrī is a nadī only optionally before ām. A nadī is a feminine stem in ī or ū that takes special endings. Where śrī is a nadī, SK 208 gives nuṭ anyway. So śrīṇām illustrates the other case, where it is not a nadī. sūtagrāmaṇīnām is a dvandva, "charioteers and village-headmen".
Watch out. śrīṇām is cited for the case where śrī is not a nadī. Where it is a nadī, the same form comes by SK 208 without this rule.
Rule. After go "cow", ām takes nuṭ at the end of a pāda. A pāda is a quarter of a verse. The Subodhinī says it is the pāda of a ṛk, a verse of the Ṛgveda, since the heading is "in the Veda".
Example.
go + ām → nuṭ → gonām: vidmā hi tvā gopatiṃ śūra gonām, "for we know you, hero, as the lord of cows".
Applies when / blocked by. Bhaṭṭoji: "Why 'at the end of a pāda'?" gavāṃ śatā pṛkṣayāmeṣu: gavām is not pāda-final, so no nuṭ. And even at a pāda's end it sometimes fails: virājaṃ gopatiṃ gavām. The reason: "in the Veda all rules are optional". That is the principle that SK 3389 set up (Part I of this chapter).
Watch out. The ordinary genitive plural of go is gavām.
Rule. In the Veda, anaṅ (an) for asthi and the rest is seen also where its conditions fail. The base rule is asthidadhisakthyakṣṇām anaṅ udāttaḥ (SK 322, chapter 10). It gives anaṅ only before a vowel-initial ending from ṭā on.
Example.
asthi + bhis → anaṅ, though bhis begins with a consonant → asthan + bhis → asthabhiḥ: indro dadhīco asthabhiḥ, "Indra, with the bones of Dadhyañc".
Applies when / blocked by. The Subodhinī explains api dṛśyate "is seen also". It lets every condition of SK 322 be set aside. So anaṅ comes even with no ending at all. In asthanvantam, asthi takes matup, anaṅ is put in, matup takes nuṭ (ano nuṭ, SK 3601, Part V), and the n of anaṅ is lost.
Watch out. This sūtra has the same wording as SK 3545, chandasy api dṛśyate. Each extends a different base rule: SK 3545 extends āṭ, this one anaṅ.
Rule. ī replaces the final of asthi and the rest before a dual ending. Bhaṭṭoji: "'of asthi and the rest' continues". That ī replaces the final vowel is shown by the example.
Example.
akṣi + bhyām → ī for i → akṣībhyām: akṣībhyāṃ te nāsikābhyām, "from your two eyes, from your two nostrils".
Applies when / blocked by. There is no Subodhinī on this sūtra.
Watch out. Ordinary usage has akṣibhyām, with short i.
Rule. In the Veda, dṛś, svavas and svatavas take the augment num (n) before su. num is carried from āc chīnadyor num (SK 445, chapter 13). "Before su" is carried from sāv anaḍuhaḥ (SK 332, chapter 11). The Subodhinī names both sources.
Example.
kīdṛś + su → num → kīdṛ-n-ś → the final ś lost → n → ṅ (kvinpratyayasya kuḥ, SK 377, chapter 11) → kīdṛṅ
→ ṅ doubled before a vowel (ṅamuṭ, SK 134, chapter 5) → kīdṛṅṅ indraḥ, "what sort of Indra".
svavas + su → num → lengthening (sāntamahataḥ…, SK 317, chapter 10) → s lost → svavān.
Likewise svatavān.
Applies when / blocked by. The Subodhinī gives the steps. kīdṛś is dṛś with kvin, with kim as upapada (SK 429, chapter 11); kim becomes kī (SK 1018, chapter 25). svavas is av + asun: "whose protection (avas) is good". svatavas is from the root tu "grow", known only from a sūtra: "whose strength (tavas) is his own". In svavān the loss of s is asiddha, so the n is not lost.
Watch out. Bhaṭṭoji closes by recalling udoṣṭhyapūrvasya (SK 2494, chapter 46). After a labial, ut replaces a root-final ṝ. That is the frame for SK 3578.
hvṛ + kta → (hru) hruta → ahrutam; gras + kta → (iṭ by nipātana) grasitaThese rules change roots, and the augment iṭ (i), before affixes. Chapters 44 and 67 taught the ordinary rules. A final ṝ may become ur after any consonant: taturiḥ. hvṛ "to be crooked" becomes hru before the past participle: ahrutam. Then come lists of fixed forms. Some take iṭ where ordinary usage refuses it: grasita "swallowed". Others lack iṭ where ordinary usage requires it: viśastā "carver". Several of these sūtras do not say "in the Veda". For them it is understood from the setting and from Bhaṭṭoji's contrast with "ordinary usage" (bhāṣā).
Rule. In the Veda, ut replaces a root-final ṝ, variously. ut is carried from udoṣṭhyapūrvasya (SK 2494, chapter 46), recalled at the end of SK 3577. "A root-final ṝ" comes from ṝta id dhātoḥ (SK 2390, chapter 44). Ordinarily such a ṝ is replaced by i (SK 2390), and by u only after a labial (SK 2494).
Example.
tṝ + kin (SK 3151, chapter 67) → ut for ṝ → tur-
→ doubled, tṝ itself being doubled by sthānivadbhāva (SK 2243, chapter 44) → urat in the reduplicated syllable (SK 2244) → taturiḥ.
jaguriḥ parācaiḥ.
Applies when / blocked by. The steps are the Subodhinī's. sthānivadbhāva is "being like the original": a substitute counts as what it replaced. So the doubling works on tṝ, not on tur.
Watch out. In both examples the ṝ follows a non-labial (t, g). So "preceded by a labial" presumably does not carry forward here.
Rule. In the Veda, hru replaces hvṛ before a niṣṭhā. The niṣṭhā affixes are kta and ktavatu, the past participles. "Before a niṣṭhā" is carried from śvīdito niṣṭhāyām (SK 3039, chapter 67).
Example.
a + hvṛ + kta → hru for hvṛ → a + hru + ta → ahrutam: ahrutam asi havirdhānam, "you are the havirdhāna, not bent crooked".
Applies when / blocked by. The Subodhinī glosses ahrutam as na hrutam, "not bent".
Watch out. The next two sūtras limit this one.
Rule. The word aparihvṛtāḥ is a nipātana. It shows the absence of hru, which SK 3579 would bring.
Example.
a + pari + hvṛ + kta → no hru → aparihvṛtāḥ: aparihvṛtāḥ sanuyāma vājam, "unbent, may we win the prize".
Applies when / blocked by. The Subodhinī says the sūtra names the form in the plural because only that form is found in use.
Rule. Of Soma, the word hvaritaḥ is a nipātana. It has iṭ and guṇa. The Subodhinī adds that the absence of hru is also understood.
Example.
hvṛ + iṭ + kta → guṇa (ṛ → ar) → no hru → hvaritaḥ: mā naḥ somo hvaritaḥ, "let Soma not be made crooked towards us".
Applies when / blocked by. The nipātana holds only of Soma (some).
Watch out. This nipātana fixes three things: iṭ, guṇa, and no hru. Bhaṭṭoji names only the first two; the third is the Subodhinī's.
Rule. Eighteen forms are nipātanas. Bhaṭṭoji sorts them into five sets.
1. grasu, skambhu and stambhu take iṭ before kta.
2. cat and kas take no iṭ before kta.
3. śas, śaṃs and śās take no iṭ before tṛc.
4. tṛ, vṛṅ and vṛñ take the augments uṭ and ūṭ before tṛc.
5. ujjvala and three other roots take i for śap.
Example.
yuvaṃ śacībhir grasitām amuñcatam, "you two by your powers freed the swallowed one"; viṣkabhite ajare, "the two, propped apart, unageing"; yena svaḥ stabhitam, "by whom the heaven was propped up"; satyenottabhitā bhūmiḥ, "by truth the earth is held up".catto itaś cattām utaḥ; tridhā ha śyāvam aśvinā vikastam, "O Aśvins, the dark one, cut apart in three"; uttānāyā hṛdayaṃ yad vikastam, "the heart of the one lying supine, which is cut open".ekas tvaṣṭur aśvasyā viśastā, "one is the carver of Tvaṣṭṛ's horse"; grāvagrābha uta śaṃstā, "the grāvagrābha and the reciter"; praśāstā potā, "the praśāstṛ, the potṛ".tarutāraṃ rathānām, "the overcomer of chariots"; tarūtāram; varūtāram; varūtrībhiḥ suśaraṇo no astu, "with the protectresses may he be a good shelter to us".ujjvaliti, kṣariti, vamiti, amiti.Applies when / blocked by. Set 1 overrides two ordinary rules. These roots are u-marked (udit), so by udito vā (SK 3328, chapter 69) their iṭ is optional before ktvā. Then yasya vibhāṣā (SK 3025, chapter 67) bars iṭ before a niṣṭhā. The nipātana gives iṭ all the same. uttabhita is listed apart from stabhita for a reason. iṭ must not come when an upasarga other than ut precedes. In ordinary usage (bhāṣā) the forms are: grasta, skabdha, stabdha, uttabdha, catita, vikasita; viśasitā, śaṃsitā, śāsitā; tarītā, taritā, varītā, varitā; ujjvalati, kṣarati, vamati, amati.
Bhaṭṭoji then brings in a related rule, babhūthātatanthajagṛbhmavavartheti nigame (SK 2527, chapter 48), with its own examples. These four forms lack iṭ in the Veda (nigama):
vidmā tam utsaṃ yata ābabhūtha, "we know that spring from which you have come into being";
yenāntarikṣam urv ātatantha, "by which you have stretched out the wide atmosphere";
jagṛbhmā te dakṣiṇam indra hastam, "we have grasped your right hand, Indra";
tvaṃ jyotiṣā vi tamo vavartha, "you with light have uncovered the darkness".
Ordinary usage: babhūvitha, ātenitha, jagṛhima, vavaritha.
Watch out. The sūtra's plural vikastāḥ refers to the many nipātanas. So a singular form such as vikastam is also correct. varūtrībhiḥ is listed only for completeness: varūtṛ is already given, and its feminine follows by ṅīp (ṛnnebhyo ṅīp, SK 306, chapter 9). Some read kṣamiti "endures" after kṣariti, from kṣamūṣ; then the count is nineteen. For vavartha, the Subodhinī notes that vṛ already lacks iṭ by kṛsṛbhṛvṛ… (SK 2293, chapter 44). So naming it restricts it: "only in the Veda". Hence ordinary vavaritha.
Rule. saniṃ sasanivāṃsam is a nipātana. It is the root san, with sanim before it. Two things are fixed. The affix kvasu takes iṭ. And there is no e-substitution and no loss of the reduplicated syllable.
Example.
san + kvasu, with sanim before it → reduplication → iṭ → sasanivāṃsam: añjitvāgne saniṃ sasanivāṃsam.
Applies when / blocked by. The Subodhinī: the iṭ is a nipātana because neḍ vaśi kṛti (SK 2981, chapter 67) would bar it. The e-substitution and abhyāsa-loss withheld are presumably those of ata ekahalmadhye… (SK 2260, chapter 44). abhyāsa is the reduplicated syllable.
Watch out. A vārttika is added here. For pāvaka and the rest, in the Veda, the a before an affix's k does not become i.
pū + ṇvul + ṭāp → no i → pāvakāḥ: hiraṇyavarṇāḥ śucayaḥ pāvakāḥ, "gold-coloured, bright, purifying (waters)".
The Subodhinī: pāvaka is "they purify" or "they cause to purify", from pū or its causative with ṇvul, then ṭāp. The rule that would give i is presumably pratyayasthāt kāt… (SK 463, chapter 15). It would make pāvikā.
dhā + leṭ tip → (ślu doubling) dadhā-ti → (ā lost) dadh-ti → (aṭ; i lost) dadhat; as + laṅ tip → (āṭ; no īṭ; t lost) ās → ā idamThese four rules touch the root before verb endings. They bend rules of chapters 44, 46, 52 and 53. leṭ is the Vedic mood taught in Part I (SK 3424 onward). Before it, dā and dhā may lose their ā: dadhat "he shall bestow". mī "to violate" keeps a short vowel: pramiṇanti. The augment īṭ (SK 2225) may fail: ā idam for āsīt idam "this was". And a rule that is optional in ordinary usage becomes fixed before caṅ.
Rule. Before leṭ, the final ā of a ghu root is optionally lost. The ghu roots are dā and dhā (dādhā ghv adāp, SK 2373, chapter 44). The Vedic setting is not repeated; it is understood from the context.
Example.
dhā + leṭ tip → doubling by ślau (SK 2490, chapter 46) → dadhā + ti
→ ā lost → dadh + ti
→ aṭ by leṭo 'ḍāṭau (SK 3427, Part I) → dadh + a + ti
→ i lost by itaś ca lopaḥ parasmaipadeṣu (SK 3426, Part I) → dadhat: dadhad ratnāni dāśuṣe, "he shall bestow treasures on the worshipper".
dā → dadat: somo dadad gandharvāya, "Soma shall give to the Gandharva".
Applies when / blocked by. The loss is optional. Without it: yad agnir agnaye dadāt, "what Agni shall give to Agni". The steps are the Subodhinī's. It says dadhat has the sense of dadhāt, and dadat is the form of dadāti.
Rule. In the Veda (nigama), a short vowel replaces the ī of mī (mīnāti) before a śit affix. A śit affix is one marked with ś, such as śnā. "A short vowel" is carried from pvādīnāṃ hrasvaḥ (SK 2558, chapter 52), the Subodhinī says. Bhaṭṭoji's gloss is "before a śit, a short vowel".
Example.
pra + mī + śnā + anti → short i for ī
→ the ā of nā lost before the vowel (śnābhyastayor ātaḥ, SK 2483, chapter 45)
→ ṇ for n (hinumīnā, SK 2530, chapter 48) → pramiṇanti: pramiṇanti vratāni, "they violate the ordinances".
Ordinary usage (loke) → pramīṇanti.
Applies when / blocked by. The condition "before a śit" presumably belongs to ṣṭhivuklamucamāṃ śiti (SK 2320, chapter 44), which runs on through SK 2558. The loss of the ā of nā is presumably that of SK 2483, which applies before a ṅit ending such as anti.
Watch out. Bhaṭṭoji closes by recalling astisico 'pṛkte (SK 2225, chapter 44). After as and after sic, an apṛkta ending takes the augment īṭ. An apṛkta is an affix of a single sound. That is the frame for SK 3586.
Rule. In the Veda, the augment īṭ comes variously. īṭ is carried from SK 2225 (chapter 44), recalled at the end of SK 3585.
Example.
as + laṅ tip → āṭ (SK 2254) → luk of śap → ās + t
→ no īṭ → the single-sound t lost (halṅyābbhyo…, SK 252, chapter 8) → ās
→ ru and visarga → āḥ
→ in saṃhitā, y for the ru (bhobhago…, SK 167, chapter 7) → āy idam
→ the y lost (lopaḥ śākalyasya, SK 67, chapter 3) → sarvam ā idam, "all this was".
Ordinary usage → āsīt.
gobhir akṣāḥ: the absence of iṭ to sic is Vedic; aṭ; the rest as before.
Applies when / blocked by. saṃhitā is the joined recitation, where sandhi applies between words. ru is the r that replaces a final s. Bhaṭṭoji then recalls hrasvasya guṇaḥ (SK 242) and jasi ca (SK 241), both of chapter 8. Two vārttikas follow.
Vārttika 1: in the Veda, jasi ca and rules of its kind are optional, up to, but not including, ṇau caṅy upadhāyā hrasvaḥ (SK 2314, chapter 44). The Subodhinī says "of its kind" takes in rules named in the sūtra before, such as SK 242.
adhā śatakratvo yūyam, "then you, of a hundred powers": śatakratvaḥ for śatakratavaḥ. Guṇa by jasi ca fails. The long vowel of prathamayoḥ pūrvasavarṇaḥ (SK 164, chapter 7) fails too, by vā chandasi (SK 3515, Part III). So yaṇ applies.paśve nṛbhyo yathā gave, "for the beast, for men, as for the cow": paśve for paśave. Guṇa by gher ṅiti (SK 245) fails, and yaṇ applies.Vārttika 2: where nābhyastasyāci (SK 2503, chapter 46) prohibits, read "variably in the Veda".
anuṣag jujoṣat, "may he take pleasure, in due order". The Subodhinī: juṣ, leṭ, parasmaipada by vyatyaya, tip; the i lost; aṭ; ślu of śap by vyatyaya; doubling.Watch out. What SK 2503 prohibits is presumably guṇa: the o of jujoṣat.
Rule. In the Veda, before caṅ, ṛ replaces an ṛ-vowel in the upadhā always (nityam). Bhaṭṭoji: "in the domain of the Veda, before caṅ, ṛ always replaces an ṛ-vowel in the upadhā". caṅ is the substitute of cli in the reduplicated aorist.
Example.
vṛdh + ṇic + caṅ → ṛ substituted for the ṛ-vowel in the upadhā, always → avīvṛdhat.
Applies when / blocked by. In ordinary usage the ṛ is only optional, by ur ṛt (SK 2567, chapter 53). That rule works under ṇau caṅy upadhāyā hrasvaḥ (SK 2314, chapter 44). nityam makes it fixed in the Veda.
Watch out. There is no Subodhinī on this sūtra. Bhaṭṭoji gives the form without steps.
mitra + kyac → (no ī, no dīrgha) mitraya- → (u, SK 3150) mitrayuḥ; aśva + kyac → (ā) aśvāya- → aśvāyantaḥkyac makes a denominative verb, "to wish for X", from a noun (chapter 58). Ordinarily a final a becomes ī before it (kyaci ca, SK 2658): putrīyantaḥ "desiring sons". In the Veda this ī often fails, and other stem changes come instead. The Vedic adjective in u (kyāc chandasi, SK 3150, chapter 67) is common here: mitrayuḥ "seeking friends". The running examples are mitrayuḥ, with no change to the stem, and aśvāyantaḥ "desiring horses", with ā.
Rule. In the Veda, an a-final stem other than putra takes no ī before kyac (kyaci ca, SK 2658, chapter 58). Nor is its a lengthened (akṛtsārvadhātukayor dīrghaḥ, SK 2298, chapter 44).
Example.
mitra + kyac → no ī, no lengthening → mitraya- → u by kyāc chandasi (SK 3150, chapter 67) → mitrayuḥ.
Applies when / blocked by. Bhaṭṭoji: "Why 'other than putra'?" putrīyantaḥ sudānavaḥ, "the bountiful ones, desiring sons": putra keeps the ī. A vārttika widens the exception: "it should be stated as 'other than putra and the like'". So janīyanto nv agravaḥ, "desiring jana", also keeps the ī. That jana is one of "the like" is shown by Bhaṭṭoji's example, not stated.
Watch out. Why name both the ī and the lengthening? The Subodhinī explains. If only the ī were barred, lengthening would come back. When an exception is barred, the general rule stands again. The proof (jñāpaka) that both are barred is SK 3590. A jñāpaka is a wording that shows an unstated principle.
Rule. durasyuḥ, draviṇasyuḥ, vṛṣaṇyati and riṣaṇyati are nipātanas before kyac. The Subodhinī names the substitutions: duras for duṣṭa, draviṇas for draviṇa, vṛṣaṇ for vṛṣa, and riṣaṇ for riṣṭa.
Example.
duṣṭa + kyac → duras for duṣṭa → durasya- → u → durasyuḥ.
Likewise draviṇasyuḥ; and the verbs vṛṣaṇyati, riṣaṇyati.
Applies when / blocked by. Bhaṭṭoji: "In ordinary usage, since the affix u does not occur there": duṣṭīyati, draviṇīyati, vṛṣīyati, riṣṭīyati. The Subodhinī adds that the point is only the base. With tṛc, which has the same meaning as u, one gets duṣṭīyitā and so on.
Watch out. "In the Veda" is not in the sūtra. It is understood from SK 3588 and from Bhaṭṭoji's contrast with ordinary usage. The u in durasyuḥ is presumably the u of SK 3150, given SK 3588 and Bhaṭṭoji's remark on the affix u.
Rule. In the Veda, ā replaces the final of aśva "horse" and agha "harm" before kyac.
Example.
aśva + kyac → ā → aśvāya- → aśvāyanto maghavan, "desiring horses, O Maghavan".
agha + kyac → aghāya- → u (SK 3150) → aghāyavaḥ: mā tvā vṛkā aghāyavaḥ, "let not the wolves, intent on harm, (reach) you".
Applies when / blocked by. The Subodhinī: aśvāyantaḥ is laṭ replaced by śatṛ, the present participle. aghāyavaḥ has kyac in the sense "wishing (harm) for another", by the vārttika under SK 3419 (Part I).
Watch out. Bhaṭṭoji: this sūtra is the jñāpaka for SK 3588. SK 3588 bars lengthening as well as ī. Otherwise lengthening alone would give the ā, and stating ā here would be pointless.
Rule. In the Yajus of the Kaṭha branch (śākhā), ā replaces the final of deva and sumna before kyac.
Example.
deva + kyac → ā → devāyanto yajamānāḥ, "the sacrificers, longing for the gods".
sumna + kyac → sumnāyanto havāmahe, "longing for favour, we invoke".
Applies when / blocked by. Bhaṭṭoji: here yajus does not mean the mantras alone. It stands for the Yajurveda. So the rule applies also to a mantra in verse (ṛk) form within the Yajurveda. It even applies in the Ṛgveda, if that mantra is found in the Kaṭha Yajus. Bhaṭṭoji: "Why 'in the Yajus'?" devāñ jigāti sumnayuḥ, "longing for favour, he goes to the gods": no ā.
Watch out. The counter-example seems to have both "Yajus" and "Kaṭha". Haradatta answers that the Bahvṛcas, the followers of the Ṛgveda, have a Kaṭha branch too. The counter-example comes from there. The Subodhinī adds why the sūtra says kāṭhake: so that ā does not come in another branch of the Yajurveda.
Rule. In the domain of a ṛk, the final of kavi, adhvara and pṛtanā is lost before kyac.
Example.
kavi + kyac → final i lost → kavya-: sa pūrvayā nividā kavyatāyoḥ.
adhvara + kyac → final a lost → adhvarya- → adhvaryum: adhvaryuṃ vā madhupāṇim.
pṛtanā + kyac → final ā lost → pṛtanya- → pṛtanyum: damayantaṃ pṛtanyum, "subduing the one who seeks battle".
Applies when / blocked by. The Subodhinī notes that adhvaryu also stands in the mṛgayvādi list, the list of words headed by mṛgayu. That is a different derivation.
Watch out. Bhaṭṭoji closes by recalling dadhāter hiḥ (SK 3076, chapter 67), "hi for dhā", and jahāteś ca ktvi (SK 3331, chapter 69), "and for hā before ktvā". They are the frame for SK 3593.
hā + ktvā → (hi) hitvā / hītvā; su + dhā + kta → (i for ā) sudhitamTwo sūtras close Part IV. Chapter 69 made hā "to leave" into hi before ktvā without option (SK 3331). In the Veda it is optional: hitvā or hītvā. Chapter 67 made dhā into hi (SK 3076). In five fixed forms the Veda has dhi instead: sudhitam "well placed". Two vārttikas then extend apo bhi (SK 442, chapter 12), t for the final of ap before bh, to Vedic stems in s: mādbhiḥ.
Rule. In the Veda, hi optionally replaces hā (jahāti) before ktvā. "hi before ktvā" is carried from jahāteś ca ktvi (SK 3331, chapter 69), recalled at the end of SK 3592.
Example.
hā + ktvā → hi → hitvā śarīram, "leaving the body".
Without hi → ī for ā (ghumāsthā…, SK 2462, chapter 45) → hītvā.
Applies when / blocked by. The Subodhinī gives the ī of hītvā. It also reports a reading hātvā in some copies. There the ī of SK 2462 fails because the form is Vedic.
Watch out. Ordinarily SK 3331 gives only hitvā. The Vedic option adds hītvā.
Rule. sudhita, vasudhita, nemadhita, dhiṣva and dhiṣīya are nipātanas. In each, i replaces the ā of dhā. For the first three this holds before kta, after su, vasu and nema.
Example.
su + dhā + kta → i for ā → sudhitam: garbhaṃ mātā sudhitaṃ vakṣaṇāsu, "the mother (bears) the well-placed embryo in her womb".
vasudhitam agnau; nemadhitā na pauṃsyā.
With ktin too: uta śvetaṃ vasudhitiṃ nireke.
dhiṣva vajraṃ dakṣiṇa indra haste, "place the vajra in your right hand, O Indra", for dhatsva.
suretā reto dhiṣīya, for dhāsīya: āśīrliṅ, the benedictive, with the ending iṭ, which a replaces (iṭo 't, SK 2257, chapter 44).
Applies when / blocked by. Bhaṭṭoji then recalls apo bhi (SK 442, chapter 12): t for the final of ap before bh. Two vārttikas extend it.
Vārttika 1: "It should be stated for mās in the Veda." mās + bhis → t, voiced to d → mādbhiḥ. śaradbhiḥ is cited beside it.
Vārttika 2: "It is wanted for svavas and svatavas, and for uṣas." svavadbhiḥ, svatavadbhiḥ, and samuṣadbhir ajāyathāḥ, "you were born together with the dawns".
Bhaṭṭoji then derives the stems. svavas is av + asun, "those whose protection is good". svatavas is from the root tu, "those whose strength is their own". uṣas has the uṇādi affix asi (mithune 'siḥ). Haradatta takes it by vaseḥ kic ca; by the method of the Pañcapādī, the uṇādi sūtras in five chapters, it was explained earlier by uṣaḥ kit.
Watch out. Bhaṭṭoji does not repeat "in the Veda" in his gloss here. The Subodhinī adds detail. vasudhita and nemadhita are karmadhāraya compounds, by Haradatta; nema means "half". The Vedabhāṣya takes vasudhita as "giver of wealth". mās replaces māsa by SK 228 (chapter 8). Only mādbhiḥ shows the mās rule; śaradbhiḥ seems to be cited as the word beside it in the mantra. Bhaṭṭoji closes by recalling na kavater yaṅi (SK 2641, chapter 56), the frame for SK 3595 in Part V.
kṛṣ + yaṅ → (abhyāsa ka keeps its k; rīk) karīkṛṣyateThe verb chapters taught how a doubled root behaves. The first copy of the root is the abhyāsa, the reduplicated syllable. Its velars turn palatal (kuhoś cuḥ, SK 2245, chapter 44), its consonants are replaced by abhyāse car ca (SK 2182), and so on. Four sūtras now close the seventh book of the Aṣṭādhyāyī with Vedic exceptions to those rules. Two of them are nipātanas: finished forms stated in the sūtra, whose irregular steps are simply accepted. The running example is the intensive of kṛṣ "to plough". In the Veda its abhyāsa keeps its k: karīkṛṣyate, against ordinary carīkṛṣyate.
Rule. In the Veda (chandas), the root kṛṣ does not take the palatal for its abhyāsa before yaṅ. yaṅ is the affix of the intensive, "again and again" (chapter 56). The change of a velar to a palatal is called cutva.
Example.
kṛṣ + yaṅ → the root is doubled; the abhyāsa is ka, with the augment rīk → ka-rī-kṛṣ-ya
→ kuhoś cuḥ (SK 2245) would make the k of the abhyāsa c; this rule stops it
→ karīkṛṣyate.
Applies when / blocked by. The words "before yaṅ, no cutva of the abhyāsa" come from na kavater yaṅi (SK 2641, chapter 56). The Subodhinī says Bhaṭṭoji quoted that rule at the end of SK 3594 so that its words run on into this one. Outside the Veda the cutva does apply: carīkṛṣyate kṛṣīvalaḥ "the ploughman ploughs again and again" (Subodhinī).
Watch out. Bhaṭṭoji gives only the form. The Subodhinī names rugrikau ca luki (SK 2652, chapter 57) for the augment rīk. A reader may compare rīg ṛdupadhasya ca (SK 2644), which gives rīk before yaṅ itself; that comparison is not stated by Bhaṭṭoji or his commentators; it is inferred.
Rule. Eighteen Vedic verb forms are taught as nipātanas. Most are intensives whose yaṅ has been deleted (yaṅluk, chapter 57), or forms with ślu, the deletion of the class-sign that doubles the root (chapter 46). In most, something about the abhyāsa is irregular. Bhaṭṭoji glosses each form in the order of the sūtra.
Example. One form worked through, alarṣi "you rush on":
ṛ (iyarti) + laṭ, sip → the root is doubled (ślu)
→ the abhyāsa keeps an l in place of r, against halādiḥ śeṣaḥ (SK 2179, chapter 44)
→ the abhyāsa does not take the i of artipipartyoś ca (SK 2493, chapter 46; so the Subodhinī)
→ alarṣi: alarṣi yudhma khajakṛt purandara "you rush on, O fighter, maker of the battle-din, breaker of forts".
The ending sip (second person) is not binding: alarti dakṣa uta "and the able one rushes on".
The other seventeen, with what each nipātana grants:
dādharti, dardharti, dardharṣi: from dhṛṅ, or from dhāri, the same root with ṇic.bobhūtu: bhū in yaṅluk, imperative (loṭ), with no guṇa.tetikte: tij in yaṅluk, with a middle (ātmanepada) ending.āpanīphaṇat: phaṇ with ā, yaṅluk, present participle (śatṛ); the abhyāsa takes the augment nīk: anv āpanīphaṇat "bounding along after".saṃsaniṣyadat: syand with sam, yaṅluk, śatṛ; the abhyāsa takes nik, and the root's s becomes ṣ.karikrat: kṛ in yaṅluk; the abhyāsa takes no cutva.kanikradat: krand in the aorist (luṅ), with aṅ for the aorist sign, doubling, no cutva, and the augment nik: kanikradaj januṣam "he roared out his birth", that is, akrandīt.bharibhrat: bhṛ (bibharti); the abhyāsa takes no jaśtva, the replacement that abhyāse car ca (SK 2182) would make: vi yo bharibhrad oṣadhīṣu "who carries it about among the plants".davidhvataḥ: dhvṛ in yaṅluk, śatṛ; the abhyāsa takes vik, and the root's ṛ is lost: davidhvato raśmayaḥ sūryasya "the flickering rays of the sun".davidyutat: dyut; the abhyāsa takes no saṃprasāraṇa, has a, and takes vik: davidyutad dīdyac chośucānaḥ "flashing, shining, blazing".taritrataḥ: tṝ in śatṛ with ślu; the abhyāsa takes rik: suhorjā taritrataḥ.sarīsṛpatam: sṛp in śatṛ with ślu, second case singular; the abhyāsa takes rīk.varīvṛjat: vṛj in śatṛ with ślu; the abhyāsa takes rīk.marmṛjya: mṛj in the perfect (liṭ), ending ṇal; the abhyāsa takes ruk, the root takes yuk.āganīganti: gam with ā, present (laṭ), with ślu; no cutva, and the augment nīk: vakṣyantīved āganīganti karṇam "as if about to speak, it comes close to the ear".Applies when / blocked by. The Subodhinī names the rules these forms set aside. karikrat escapes kuhoś cuḥ (SK 2245). bharibhrat escapes abhyāse car ca (SK 2182). davidyutat escapes dyutisvāpyoḥ saṃprasāraṇam (SK 2344). For davidhvataḥ it adds that the nasal of the participle is barred by nābhyastāc chatuḥ (SK 427).
Watch out. bobhūtu teaches something beyond itself. The Subodhinī objects that bhūsuvos tiṅi (SK 2224) already forbids guṇa here. It answers that the nipātana is a jñāpaka, an indication: elsewhere a yaṅluk form of bhū does take guṇa. So ordinary Sanskrit has bobhavīti. The 1915 print gives taritrataḥ and sarīsṛpatam separate glosses.
Rule. In the Vedic text (nigama) the form sasūva is a nipātana. It is the perfect of sū "to give birth".
Example.
sū + liṭ → the active (parasmaipada) ending, by nipātana
→ the root is doubled; the abhyāsa takes a for its u: sa-sū
→ the augment vuk → sasūva.
gṛṣṭiḥ sasūva sthaviram "the young cow has borne a sturdy calf".
Applies when / blocked by. In ordinary Sanskrit (bhāṣā) the form is suṣuve, a middle form with the abhyāsa su.
Watch out. The Subodhinī says this sūtra ought to have stood among the eighteen forms of SK 3596.
Rule. In the Veda, i variously takes the place of the abhyāsa's vowel. bahulam means "variously": sometimes it applies, sometimes not.
Example.
vaś + laṭ, tip → the class-sign is deleted (ślu), and the root is doubled → va-vaś-ti
→ i for the abhyāsa's vowel → vi-vaś-ti
→ ś becomes ṣ (SK 294), and t after ṣ becomes ṭ (ṣṭutva) → vivaṣṭi.
pūrṇāṃ vivaṣṭi "he longs for the full ladle". Bhaṭṭoji: this is the form of vaś "to desire".
Applies when / blocked by. The i is the it of bhṛñām it (SK 2496, chapter 46), carried into this rule. The Subodhinī, on the same form at SK 3401, derives vivaṣṭi with this rule. Here the seventh book ends.
akṣi + matup → (anaṅ) akṣan-mat → (n lost; nuṭ, asiddha) akṣaṇ-vat → akṣaṇvantaḥThe eighth book begins. One sūtra doubles four prefixes just to fill a line of verse. Three more change the possessive affix matup and the comparative affixes in the Veda. matup (-mat, -vat) means "having": chapter 39 taught it, and SK 1897 there makes its m a v after certain stems. The comparative and superlative affixes tarap and tamap are called gha (SK 2003, chapter 41). The running example is akṣaṇvantaḥ "having eyes", from akṣi "eye".
Rule. The prefixes pra, sam, upa and ud are doubled to fill out the pāda, the metrical line of a verse.
Example.
pra ayam agniḥ → pra doubled → pra prāyam agniḥ "forth, forth goes this Agni".
saṃsam id yuvase "you join all together".
upopa me parā mṛśa "touch me closely".
kiṃ nod ud u harṣase "why do you not exult?"
Applies when / blocked by. The doubling adds no meaning (Subodhinī). Its only work is to fill the line. Such usage is not found in ordinary language, so the Subodhinī says the rule holds only in the Veda. That the doubling is the one taught by sarvasya dve (SK 2139, chapter 43) is not stated by Bhaṭṭoji or his commentators; it is inferred.
Watch out. The sūtra word prasamupodaḥ is a collective compound (samāhāra-dvandva) naming the four prefixes. It ought to take the closing affix ṭac by SK 930. The Subodhinī says it does not, because rules for closing affixes are not obligatory. The 1915 print reads pra prāyam with both copies of pra. Here the eighth book begins.
Rule. In the Veda, after a stem ending in an i-vowel or in r, v replaces the m of matup.
Example.
hari + matup → m becomes v → harivat "having bay horses"
→ dative harivate: harivate haryaśvāya "to him who has bay horses, whose horses are bay".
gir + matup → m becomes v → gir-vat → the i is lengthened before r (SK 433) → gīrvān.
Applies when / blocked by. Bhaṭṭoji's gloss supplies "the m of matup becomes v". That this comes from mād upadhāyāś ca… (SK 1897, chapter 39) is not stated by Bhaṭṭoji or his commentators; it is inferred. The ī of the sūtra stands for the whole i-class, short and long. The Subodhinī supplies the lengthening rule for gīrvān.
Rule. After a stem ending in an, matup takes the augment nuṭ, that is, an n placed before it.
Example.
akṣi + matup → akṣi becomes akṣan (anaṅ, SK 322, through chandasy api dṛśyate, SK 3575) → akṣan-mat
→ the stem's n is lost → akṣa-mat
→ nuṭ comes, on the strength of what the stem was before → akṣa-n-mat
→ v for m; n becomes ṇ → akṣaṇvantaḥ "having eyes".
Also asthanvantaṃ yad anasthā "that the boneless one bears the one with bones".
Applies when / blocked by. The Subodhinī explains why nuṭ does not spoil the other rules. nuṭ is taught in the Tripādī, the last three quarters of book 8, whose rules are asiddha, treated as not yet done, for the earlier rules (SK 12). So the n is lost first, and nuṭ comes after. And nuṭ does not stand between the stem and matup for the rule that makes m into v.
Watch out. The quoted line also has karṇavantaḥ "having ears". That word does not end in an and is not an instance of this rule. The Subodhinī adds that the Vedic word-by-word text writes akṣaṇ-vantaḥ. It cites the Mahābhāṣya: grammar does not follow those reciters; they must follow grammar.
Rule. After a stem ending in n, gha (tarap or tamap) takes the augment nuṭ.
Example.
supathin + tarap → the stem's n is lost → supathi-tara
→ nuṭ → supathi-n-tara → supathintaraḥ.
The Subodhinī: tarap comes by SK 2005; the n passes through anusvāra to the nasal of t's class.
Applies when / blocked by. Two vārttikas extend the rule. A vārttika is a supplementary statement to a sūtra.
bhūridāvan + tarap → n lost → bhūridāva-tara → tuṭ → bhūridāvattaro janaḥ "a man who gives more abundantly".
The Subodhinī derives the stem with vanip after dā (SK 3418).
rathin + tarap → n lost → rathi-tara → ī for the i → rathītaraḥ; rathītamaṃ rathīnām "the best charioteer among charioteers".
Watch out. In rathītaraḥ the ī replaces the i left after the n is lost, not the n itself. The Subodhinī says so: an ī for the n would be asiddha, and the vowels could not merge. The vārttika's stem is printed bhūridānnaḥ; the form bhūridāvattaraḥ and the Subodhinī's derivation show it is bhūridāvan.
amnas eva → (ru → y → lost) amna eva / (r) amnar evaThree short sūtras follow. The first fixes six Vedic past participles taught as finished forms; most keep a t where ordinary Sanskrit has n. The other two let four s-final words end either in ru or in a plain r. ru is the r that the earlier chapters put in place of a word-final s (sasajuṣo ruḥ, SK 162, chapter 7). It then changes further by the rules of chapter 7, for example to y before a vowel (SK 167). The running example is amnas "a little". With ru it gives amna eva; with r it gives amnar eva.
Rule. In the Veda, six past participles are nipātanas: nasatta, niṣatta, anutta, pratūrta, sūrta and gūrta. The past participle in -ta is called the niṣṭhā.
Example.
ni + sad + kta → by nipātana the niṣṭhā does not take n
→ s becomes ṣ (SK 2271, so the Subodhinī) → niṣattam: niṣattam asya carataḥ.
na + sad + kta → nasattam: nasattam añjasā.
In ordinary Sanskrit these would be niṣaṇṇam and asannam.
The others: anuttam from und with the negative na; pratūrtam from tvar or from turvī; sūrtam from sṛ; gūrtam from gurī.
Applies when / blocked by. The change these forms escape is radābhyāṃ niṣṭhāto naḥ pūrvasya ca daḥ (SK 3016, chapter 67). It makes the t of the niṣṭhā n after r and d, and the d before it n too. The Subodhinī gives the ordinary forms anunnam and gūrṇam.
Watch out. The Subodhinī draws a further point from anuttam. Because this form had to be taught for the Veda, the option of SK 3038 does not hold for und in ordinary Sanskrit.
Rule. In the Veda, the final s of amnas, ūdhas and avas is replaced either by ru or by r.
Example.
amnas eva → s becomes ru → ru becomes y before a vowel (SK 167) → y is lost (SK 67) → amna eva.
amnas eva → s becomes r → amnar eva.
Likewise ūdha eva / ūdhar eva, and ava eva / avar eva.
Applies when / blocked by. ru would come anyway by sasajuṣo ruḥ (SK 162). So, the Subodhinī says, the rule is really for the plain r. The plain r does not go on to y, so it stays. The Subodhinī glosses the words: amnas means "a little", avas means "protection".
Rule. The same choice, ru or r, holds for bhuvas when it is the great utterance (mahāvyāhṛti). The three great utterances name earth, mid-region and heaven; bhuvas is the middle one.
Example.
bhuvas iti → ru → y → lost → bhuva iti.
bhuvas iti → r → bhuvar iti.
Applies when / blocked by. The Subodhinī: the word must be the indeclinable bhuvas, "the mid-region". In bhuvo viśveṣu bhuvaneṣu the word is a verb form of bhū, so the rule does not touch it. There bhuvaḥ is an imperfect (laṅ) used for the present (SK 3423), with no augment a (SK 3546).
Watch out. Bhaṭṭoji gives only the two forms. That bhuva iti shows ru and bhuvar iti shows r, as at SK 3604, is not stated by Bhaṭṭoji or his commentators; it is inferred. The 1915 print reads bhuvar iti.
ye yajāmahe → (ṭi e pluta, udātta by SK 93) ye3 yajāmahe; akārṣaṃ hi → akārṣaṃ hi3A pluta vowel is a vowel held for three mātrās (morae), against one for a short and two for a long vowel. It is written with the figure 3 after it: ye3. The head of the pluta rules was taught in chapter 4. There vākyasya ṭeḥ pluta udāttaḥ (SK 93) says that the ṭi of a sentence becomes pluta and udātta (high-pitched). The ṭi is the last vowel of a word together with whatever follows it. SK 94–97 gave the pluta in a return greeting, in a call from afar, and elsewhere.
Thirteen sūtras now give the pluta in the ritual and in dialogue. The Subodhinī says at SK 3608 that the words "of the ṭi" are still carried from SK 93. At SK 3619 it says such a pluta would be udātta. So every pluta in this group except the om of SK 3608 is udātta, and it is marked with an acute: yé3, hí3. The rest of each sentence keeps its own accent, which these units do not mark.
Rule. The word om becomes pluta at a commencement (abhyādāna, that is, ārambha).
Example.
om agnim īḷe purohitam → o of om pluta → ó3m agnim īḷe purohitam "Om! I praise Agni, the household priest".
The udātta syllable is the pluta ó3.
Applies when / blocked by. It is not pluta when om does not open anything: om ity ekākṣaram "om, the single syllable".
Watch out. Only the vowel is prolated, not the m. The Subodhinī says the pluta always falls on a vowel, by acaś ca (SK 35, chapter 2). With half a mātrā for the m, ó3m comes to three and a half mātrās.
Rule. In a sacrificial act (yajñakarman), the word ye becomes pluta.
Example.
ye yajāmahe → e of ye pluta → yé3 yajāmahe "we who worship".
The udātta syllable is the pluta yé3.
Applies when / blocked by. Outside the sacrificial act the same words are said plainly: ye yajāmahe. The word yajñakarmaṇi runs on into SK 3608–3610.
Rule. In a sacrificial act, the praṇava, the syllable om, replaces the whole ṭi of the sentence.
Example. The Subodhinī's reading:
jinvati (jinv, present, third person singular) → its ṭi i is replaced by om → jinvato3m.
In the mantra: apāṃ retāṃsi jinvato3m "it quickens the seeds of the waters".
Applies when / blocked by. yajñakarmaṇi comes from SK 3607 (Subodhinī). Why does the sūtra say "of the ṭi" again, when SK 93 already gives it? The Subodhinī answers. A substitute replaces only the last sound of what it stands for, by alo 'ntyasya (SK 42, chapter 2). Where the ṭi ends in a consonant, om would replace only that consonant. Restating "of the ṭi" makes om replace the whole ṭi.
Watch out. Bhaṭṭoji's text, like the 1915 print, has jivanto3m. The Subodhinī reads jinvato3m and derives it from jinvati; the example above follows it. Bhaṭṭoji writes the substituted om as pluta but does not say here why. So its accent is not marked here.
Rule. In a sacrificial act, in a yājyā (a verse recited at the offering), only the ṭi of the last sentence becomes pluta.
Example.
jihvām agne cakṛṣe havyavāham → the ṭi am of the last word pluta → jihvām agne cakṛṣe havyavāhā́3m.
"O Agni, you have made your tongue the bearer of oblations." The udātta syllable is the pluta hā́3m.
Applies when / blocked by. Why "last"? A yājyā is a verse made of a group of sentences. Without "last", every sentence would take the pluta at its ṭi. It is wanted only for the very last.
Rule. In a sacrificial act, the first vowel of brūhi, preṣya, śrauṣaṭ, vauṣaṭ and āvaha becomes pluta.
Example.
agnaye 'nubrū́3hi.
agnaye gomayāni pré3ṣya.
astu śráu3ṣaṭ.
somasyāgne vīhī váu3ṣaṭ.
agnim ā́3vaha.
In each, the udātta syllable is the pluta: brū́3, pré3, śráu3, váu3, ā́3.
Watch out. This rule moves the pluta to the first vowel of the word, not the ṭi. The sūtra says only "the first". That it means the first vowel is not stated by Bhaṭṭoji or his commentators; it is inferred, since only a vowel can be pluta (SK 35). There is no Subodhinī on this sūtra.
Rule. In the agnīdh's directive (agnīt-preṣaṇa), the first vowel becomes pluta, and also the vowel after it. The agnīdh is a priest of the rite.
Example.
o śrāvaya → first vowel o pluta, and the next vowel ā pluta → ó3śrā́3vaya.
Both plutas are udātta: ó3 and śrā́3.
Applies when / blocked by. "The first" comes from SK 3610, as Bhaṭṭoji's gloss shows. Bhaṭṭoji and the Subodhinī read the compound as "the directive of the agnīdh". They do not say whether the agnīdh gives it or receives it.
Rule. In the answer to a question, the particle hi optionally (vibhāṣā) becomes pluta.
Example.
Question: akārṣīḥ kaṭam? "Did you make the mat?"
Answer: akārṣaṃ hi → hi pluta → akārṣaṃ hí3 "Of course I made it"; or plain akārṣaṃ hi.
The udātta syllable is the pluta hí3.
Applies when / blocked by. Only in an answer: kaṭaṃ kariṣyati hi "he will surely make the mat" answers nothing, so no pluta. Only for hi: kaṭaṃ karoti nanu "why, he is making the mat" uses nanu, so no pluta.
Rule. When one has refuted someone (nigṛhya) and then questions him about his claim (anuyoga), the ṭi of that sentence optionally becomes pluta.
Example.
Someone has said, "Today is the new-moon day". He has been refuted by reasoning. Now he is asked:
adyāmāvāsyety āttha → ṭi a of āttha pluta → adyāmāvāsyety ātthá3 "You say, 'Today is the new-moon day'?"
The udātta syllable is the pluta thá3.
Applies when / blocked by. The option comes from SK 3612, as Bhaṭṭoji's "optionally" (vā) shows. The Subodhinī: nigṛhya is an absolutive, "having dislodged him from his position"; anuyoga is putting that position back to him in words, "so you say this without proof". The question is asked with indignation.
Rule. In a threat (bhartsana), the doubled word becomes pluta at its ṭi. The sūtra names the āmreḍita, the second member of a doubled word (SK 83, chapter 3). Bhaṭṭoji says it stands for the whole doubled expression, both members.
Example.
dasyo dasyo → each ṭi o pluta → dasyó3 dasyó3 ghātayiṣyāmi tvām "Robber! robber! I will have you killed!"
caurá3 caurá3 "Thief! thief!"
The udātta syllables are the plutas syó3 and rá3.
Applies when / blocked by. The word is doubled by vākyāder āmantritasya… (SK 2143, chapter 43), so the Subodhinī. It explains Bhaṭṭoji's remark. The name āmreḍita belongs to the second member alone. Yet the pluta is wanted on both, and the second member alone is never used in a threat. So "āmreḍita" here indicates the whole doubled group (upalakṣaṇa).
Rule. In a threat, a verb form (tiṅanta, a word ending in a verb ending) joined with the particle aṅga becomes pluta at its ṭi. The verb must be ākāṅkṣa, "expectant": the sentence must go on.
Example.
aṅga kūja → ṭi a of kūja pluta → aṅga kūjá3, idānīṃ jñāsyasi jālma "Go on, coo! Now you will find out, you wretch!"
The udātta syllable is the pluta já3. The Subodhinī: aṅga shows impatience, and kūja is the imperative of kūj "to make an inarticulate sound".
Applies when / blocked by. Three counter-examples, none pluta:
aṅga devadatta mithyā vadasi "Come now, Devadatta, you are lying". Here aṅga devadatta expects the rest, but it is not a verb (Subodhinī).aṅga paca "Come, cook!" expects nothing further.aṅgādhīṣva, bhaktaṃ tava dāsyāmi "Come, study; I will give you rice". "In a threat" runs on from SK 3614.Rule. The ṭi of sentences under deliberation (vicāryamāṇa) becomes pluta. Bhaṭṭoji: deliberation (vicāra) is examining the true nature of a thing by the means of knowledge (pramāṇas).
Example.
hotavyaṃ dīkṣitasya gṛhe, na hotavyam iti
→ the ṭi of each sentence pluta → hotavyaṃ dīkṣitasya gṛhā́3i, na hotavyá3m iti
"Should one offer in the house of the initiated sacrificer, or should one not offer?"
The udātta syllables are the plutas hā́3 and vyá3.
Applies when / blocked by. Both alternatives, "one should offer" and "one should not offer", are being weighed, so both take the pluta. SK 3617 shows that this is the Vedic usage.
Watch out. gṛhā3i is gṛhe, a locative, whose e has been split into a pluta ā and an i. The Subodhinī names SK 3626 for that split.
Rule. But in ordinary Sanskrit (bhāṣā), of the sentences under deliberation, only the first becomes pluta.
Example.
ahir nu rajjur nu → only the first sentence pluta → ahir nú3 rajjur nu "Is it a snake? Or is it a rope?"
The udātta syllable is the pluta nú3.
Applies when / blocked by. "First" means first in the order spoken. Since the sūtra names bhāṣā, Bhaṭṭoji infers that SK 3616, which gives the pluta to every sentence, is for the Veda.
Watch out. The first sentence already has the pluta by SK 3616. So, the Subodhinī says, this rule is a niyama, a restriction: "only the first". It does not mean "the first, only in bhāṣā".
Rule. In pratiśravaṇa the ṭi of the sentence becomes pluta. Bhaṭṭoji gives the word three senses: acceptance, affirmation, and readiness to listen.
Example.
gāṃ me dehi bhoḥ "Give me a cow, sir." — hanta te dadāmí3 "Very well, I give it to you".nityaḥ śabdo bhavitum arhatí3 "Sound must be eternal".datta kim ātthá3 "Datta, what are you saying?"The udātta syllables are the plutas mí3, tí3 and thá3.
Watch out. Bhaṭṭoji lists the examples without labels. The Subodhinī assigns the last to readiness to listen. Pairing the first two with acceptance and affirmation is not stated by Bhaṭṭoji or his commentators; it is inferred from the examples.
agnibhūte (end of a question) → (e = a+i; a → pluta ā, + i) agnibhūta3i (anudātta pluta); agna3i āśā → (i → y) agna3yāśāThe pluta of SK 93 is udātta. Six sūtras now change that accent in set contexts. Some make the pluta anudātta (low-pitched); others make it svarita (a falling pitch). Marking: udātta takes an acute, an independent svarita takes a grave, and anudātta is left unmarked. Three more sūtras say what happens when a diphthong (ai, au, e, o) is to be pluta. Accents are not marked in SK 3625–3627. The figure 3 marks a three-mātrā vowel whether it is written a3 or ā3. The running example is the vocative agnibhūte "O Agnibhūti" at the end of a question. Its e splits into a pluta ā and an i: agnibhūta3i. Before a vowel that i becomes y: agna3yāśā.
Rule. At the end of a question (praśnānta) and in praise (abhipūjita), the pluta is anudātta. Bhaṭṭoji: this rule gives only the anudātta accent. The pluta itself is already given by dūrād dhūte ca (SK 95, chapter 4) and the other rules.
Example.
Question: agamaḥ pūrvān grāmān agnibhūte paṭo "Did you go to the eastern villages, Agnibhūti, clever one?"
→ the closing words take the pluta at their ṭi → agnibhūta3i, paṭa3u
→ this rule: those plutas are anudātta.
The pluta syllables ta3 and ṭa3 are anudātta, so they carry no mark.
Praise: śobhanaḥ khalv asi māṇavaka3 "You are fine indeed, boy". The pluta ka3 is anudātta.
Applies when / blocked by. The Subodhinī names the sources of these plutas. In praise the pluta comes from SK 95. In the question it comes from SK 3624, which reaches even a non-final word. In the question, the non-final words keep the svarita of SK 3624; only the closing words take the anudātta of this rule.
Watch out. agnibhūta3i and paṭa3u are agnibhūte and paṭo with the e and o split by SK 3626. Bhaṭṭoji does not derive them here.
Rule. When the particle cit is used in the sense of comparison (upamā), the ṭi of the sentence becomes an anudātta pluta.
Example.
agnicid bhāyāt "may he shine like fire" (= agnir iva bhāyāt)
→ ṭi āt of bhāyāt becomes an anudātta pluta → agnicid bhāyā3t.
The pluta syllable yā3 is anudātta, so it carries no mark.
Applies when / blocked by. Not where cit means something else: kathaṃcid āhuḥ "they say it with difficulty". Not where cit is not used: agnir māṇavako bhāyāt "may the boy shine like fire". There the comparison is understood, but cit is absent.
Watch out. Here the pluta itself is prescribed, not only its accent (Subodhinī, contrasting SK 3619). The pluta falls on the sentence's ṭi, on bhāyāt, not on cit. The Subodhinī: the iti in the sūtra makes "used" go with the word cit, so the pluta is not placed on cit.
Rule. In the phrase upari svid āsīt "was it above?", the ṭi becomes an anudātta pluta. The pluta itself is the one that vicāryamāṇānām (SK 3616) gives to a sentence under deliberation. This rule changes only its accent (Subodhinī).
Example.
upári svid āsī3t "was it above?"
The udātta syllable is pá of upári. The pluta sī3 is anudātta and unmarked. The ri after pá carries a svarita that only follows an udātta; as elsewhere in this chapter, such a svarita is left unmarked.
Contrast adháḥ svid āsī́3t "was it below?"
Here the udātta syllables are dhá of adháḥ and the pluta sī́3. That pluta is udātta by SK 3616.
Applies when / blocked by. Only the phrase with upari is named. The paired phrase with adhaḥ keeps the udātta pluta of SK 3616.
Watch out. Bhaṭṭoji glosses "the ṭi becomes an anudātta pluta". That the word "anudātta" is carried from SK 3619 is not stated by Bhaṭṭoji or his commentators; it is inferred.
Rule. When an āmreḍita, the second member of a doubled word, follows, the pluta is svarita. This holds where envy (asūyā), approval (sammati), anger (kopa) or contempt (kutsana) is meant.
Example.
abhirūpaka abhirūpaka → the first member's ṭi becomes a svarita pluta → abhirūpakà3 abhirūpaka, riktaṃ te ābhirūpyam "Handsome fellow, handsome fellow! Your handsomeness is empty".abhirūpakà3 abhirūpaka, śobhano 'si "… You are splendid".avinītakà3 avinītaka, idānīṃ jñāsyasi jālma "Ill-bred fellow, ill-bred fellow! Now you will find out, wretch".śāktīkà3 śāktīka, riktā te śaktiḥ "Spearman, spearman! Your spear is empty".In each, the pluta kà3 is svarita and marked with a grave.
Applies when / blocked by. The word is doubled by vākyāder āmantritasya asūyāsammatikopakutsanabhartsaneṣu (SK 2143, chapter 43), so the Subodhinī. Bhaṭṭoji's gloss supplies "pluta". That it comes from SK 93 is not stated by Bhaṭṭoji or his commentators; it is inferred.
Rule. In a breach of proper conduct (kṣiyā), a wish (āśīḥ) or a direction (praiṣa), the ṭi of an expectant verb form becomes a svarita pluta. "Expectant" means the sentence goes on after it.
Example.
svayaṃ ha rathena yāti → ṭi pluta, svarita → svayaṃ ha rathena yātì3, upādhyāyaṃ padātiṃ gamayati "He himself goes by chariot, and makes his teacher go on foot!"putrāṃś ca lapsīṣṭà3 dhanaṃ ca tāta "May he obtain sons, and wealth too, dear one".kaṭaṃ kurù3 grāmaṃ gaccha "Make the mat, then go to the village".The svarita plutas, marked with a grave, are tì3, ṣṭà3 and rù3.
Applies when / blocked by. Not where the verb expects nothing further. For a wish: dīrghāyur asi "you are long-lived". For a direction: agnīd agnīn vihara "Agnīdh, spread out the fires". The Subodhinī: a breach of conduct has no counter-example, since it is always expectant. "He goes by chariot" shows no breach until the second half is said.
Watch out. That "svarita" is carried from SK 3622 is not stated by Bhaṭṭoji or his commentators; it is inferred. Bhaṭṭoji's gloss itself says "svarita pluta". The 1915 print reads putrāṃś.
Rule. In a question (praśna) and in a narration (ākhyāna), the ṭi of every word (pada), not only the last, becomes a svarita pluta.
Example.
agamaḥ pūrvān grāmān → the ṭi of each word a svarita pluta → agamàḥ3 pūrvà3n grāmà3n "Did you go to the eastern villages?"agamam pūrvān grāmān → agamà3m pūrvà3n grāmà3n "I went to the eastern villages".Every pluta here is svarita and marked with a grave.
Applies when / blocked by. Bhaṭṭoji: this pluta comes on all the words. At the very end of a question, SK 3619 makes the pluta anudātta instead (Subodhinī on SK 3619). In the full question of SK 3619, agamaḥ pūrvān grāmān goes on with agnibhūte paṭo; those closing words take the anudātta pluta.
Watch out. That "svarita" is carried from SK 3622 is not stated by Bhaṭṭoji or his commentators; it is inferred. There is no Subodhinī on this sūtra itself.
Rule. Where ai or au is to be pluta, its i or u part is made pluta.
Example.
aitikāyana → the i part of ai pluta → ai3tikāyana.
aupagava → the u part of au pluta → au3pagava.
Bhaṭṭoji: here ai and au come to four mātrās.
Applies when / blocked by. It works wherever a pluta is prescribed, as in dūrād dhūte (SK 95) and the rest. The Subodhinī says these two plutas come from guror anṛto… (SK 97, chapter 4).
Watch out. The count of four mātrās: the Subodhinī takes ai as one mātrā of a plus one of i. When the i becomes pluta it has three, and the whole has four. Some divide ai as half a mātrā and one and a half, which would give three and a half. The Subodhinī keeps the equal division, because the Mahābhāṣya wants a pluta of four mātrās here.
Rule. Take an ec (e, o, ai, au) that is not pragṛhya, where a pluta is due, but not in a call from afar. Its first half becomes a pluta ā, and its second half becomes i or u. A pragṛhya vowel is one that stays unchanged before a vowel (chapter 4).
Example.
agnibhūte → e = a + i → first half a pluta ā, second half i → agnibhūta3i.
paṭo → o = a + u → paṭa3u.
Applies when / blocked by. A vārttika limits the rule to five contexts:
agamaḥ3 pūrvā3n grāmā3n agnibhūta3i;karoṣi paṭa3u;hotavyaṃ dīkṣitasya gṛha3i, na hotavya3m iti;āyuṣmān edhi agnibhūta3i "Be long-lived, Agnibhūti";stomair vidhemāgnayā3i "Let us worship Agni with praises".Why the list? Not in viṣṇubhūte ghātayiṣyāmi tvām "Viṣṇubhūti, I shall have you killed". So the sūtra's "not in a call from afar" need not be stated; the list already excludes it.
But "at the end of a word" (padānta) must be added. Otherwise the rule would reach bhadraṃ karoṣi gauḥ "you do well, ox". The Subodhinī: by SK 230, go is not a pada before the ending s, so its au is not word-final.
Not for a pragṛhya: śobhane māle "O two fine garlands", pragṛhya by SK 100.
A second vārttika: in the Veda, this split also applies to a vocative (āmantrita) where pluta is not otherwise due: agna3i patnīvaḥ "O Agni, accompanied by the wives". The Subodhinī: agne is a vocative by SK 411, and the vārttika covers only cases where no pluta is otherwise due.
Watch out. The Subodhinī matches the list to its rules. The question takes SK 3619 and SK 3624, and praise SK 3619. Deliberation takes SK 3616; the last two take SK 94 and SK 3609. That patnīvaḥ is one word, with ru by SK 3628, is not stated by Bhaṭṭoji or his commentators; it is inferred.
Rule. The i and u left by SK 3626 become y and v before a vowel (ac), in continuous utterance (saṃhitā).
Example.
agna3i + āśā → i becomes y → agna3yāśā.
paṭa3u + āśā → u becomes v → paṭa3vāśā.
Likewise agna3yindram, paṭa3vudakam.
Applies when / blocked by. Not before a consonant: agnā3i varuṇau. Not outside saṃhitā: agna3i indraḥ. Bhaṭṭoji: the word saṃhitāyām is a heading (adhikāra) to the end of the eighth book.
Watch out. The Subodhinī asks why this is taught, when i and u become y and v before a vowel anyway (SK 47). Bhaṭṭoji gives two answers.
Scholars discuss this. The Subodhinī weighs both at length and quotes a verse that favours the second.
harivan (sambuddhi) → (n → ru) harivaḥ → (u by SK 166) harivo medinam; devān acchā → (n → ru; ā nasalized) devā̃ acchāThe eighth book's third quarter opens with ru, now for a final n. Chapter 5 taught naś chavy apraśān (SK 140): a word-final n becomes ru before a sound of the group chav. With it came SK 136, which nasalizes the vowel before that ru, and SK 138, which makes the visarga s. Six sūtras now add Vedic cases. The running examples are harivo medinam "O lord of the bay horses, … the ally", and devā̃ acchā "towards the gods".
Rule. In the Veda, in the vocative singular (sambuddhi), a stem in matup or in vasu changes its final n to ru. By alo 'ntyasya (SK 42) the ru replaces the last sound, the n.
Example. The Subodhinī's steps:
harivat + su (vocative) → the augment num (SK 361) → harivant + s
→ s is lost, then the final t of the cluster → harivan
→ n becomes ru → ru becomes u before m (haśi ca, SK 166) → a + u become o (ādguṇaḥ, SK 69, chapter 3) → harivo medinaṃ tvā "… you, the ally, O lord of the bay horses".
marutvan → ru → indra marutva iha pāhi somam "Indra, accompanied by the Maruts, drink the soma here".
Applies when / blocked by. Bhaṭṭoji: the v of harivaḥ is by chandasīraḥ (SK 3600). The Subodhinī takes the v of marutvan from jhayaḥ (SK 1898, chapter 39). The word ru in the sūtra has no case-ending (Bhaṭṭoji).
Watch out. In marutva iha the ru has gone on to y and then been lost before the vowel. These steps are not stated by Bhaṭṭoji or his commentators here; they are inferred, as at SK 3604.
Rule. dāśvān, sāhvān and mīḍhvān are nipātanas. All three end in the affix kvasu.
Example.
mīḍhvas tokāya tanayāya "O bountiful one, for offspring and descendants".
Applies when / blocked by. This sūtra comes from the sixth book. The Subodhinī says Bhaṭṭoji brings it in "incidentally" (prasaṅgāt). The link is probably that mīḍhvaḥ is the vocative of a kvasu stem, with the ru of SK 3628. That link is not stated by Bhaṭṭoji or his commentators; it is inferred.
A vārttika adds van, stems in -van. That it adds them to the matup and vasu stems of SK 3628 is not stated by Bhaṭṭoji or his commentators; it is inferred from its example, the vocative prātaritvaḥ with ru. Both kvanip and vanip are covered, since no marker is named. So the rule about affixes with a given marker (anubandha) does not come in.
prātar + i + kvanip → the augment tuk (SK 2858) → prātaritvan "early-comer"
→ vocative, n becomes ru → yas tvāyantaṃ vasunā prātaritvaḥ "… who welcomes you as you come with wealth, O early-comer".
Bhaṭṭoji: kvanip comes after iṇ "to go". The Subodhinī gives anyebhyo 'pi dṛśyante (SK 2980) for it.
Rule. In the ṛcs (Vedic verses), the change of SK 140 is optional. A final n optionally becomes ru before a sound of the group chav that is followed by a sound of the group am.
Example.
paśūn tān cakre → the n of tān becomes ru → ru becomes visarga → visarga becomes s (SK 138), with a nasal before it → s becomes ś before c (stoḥ ścunā ścuḥ, SK 111, chapter 5; not named here) → paśūn tāṃś cakre "he made those animals".
Applies when / blocked by. naś chavy apraśān (SK 140) runs on into this rule; by itself it would make ru obligatory. So this rule is for the option (Subodhinī). The nasal before ru is optional here by SK 136, the Subodhinī says.
Watch out. The n of paśūn stays as it is. That it shows the option not taken is not stated by Bhaṭṭoji or his commentators; it is inferred.
Rule. An n after a long vowel optionally becomes ru before a sound of the group aṭ. The n and the aṭ must stand in the same pāda.
Example.
devān acchā → n becomes ru; the ā before it is nasalized (SK 3632) → the ru then goes before the vowel, as at SK 3604 (not stated here) → devā̃ acchā sumatī "towards the gods, with good thought".
mahān indraḥ → mahā̃ indro ya ojasā "Indra, who is great in might".
Applies when / blocked by. "Optionally" runs on from SK 3630. So the option may not be taken: ādityān yāciṣāmahe "we entreat the Ādityas". That the n is still the n of SK 140 is not stated by Bhaṭṭoji or his commentators; it is inferred. The Subodhinī: samāna here means "one", so the n and the aṭ must be in one pāda.
Watch out. The 1915 print reads devā̃ and mahā̃, with the nasal sign over the vowel, not an anusvāra. The Subodhinī adds that the option also accounts for the reading mahān hi saḥ of the Ṛgveda reciters.
Rule. Before a sound of the group aṭ, an ā standing before ru is always nasalized (anunāsika).
Example.
mahān indraḥ → n becomes ru (SK 3631) → the ā before it nasalized → the ru then goes before the vowel, as at SK 3604 (not stated here) → mahā̃ indraḥ.
Applies when / blocked by. SK 136 (chapter 5) makes this nasal optional. Before an aṭ, after ā, this rule makes it obligatory. Bhaṭṭoji places the rule after SK 3631 because SK 3631's examples need it. Bhaṭṭoji's gloss supplies "before ru" and "nasal". That they come from SK 3628 and SK 136 is not stated by Bhaṭṭoji or his commentators; it is inferred.
Watch out. Bhaṭṭoji doubts the rule's purpose. The Taittirīya school recites an anusvāra here. The earlier grammarians (prāñcaḥ) call that a Vedic interchange (vyatyaya). So the use of the sūtra is open to question (cintya). The Subodhinī: since the option of SK 136 is fixed in scope anyway, dropping this sūtra would be much simpler.
Rule. The n of svatavān optionally becomes ru when the word pāyu follows.
Example.
svatavān pāyuḥ → n becomes ru; the ā nasalized → bhuvas tasya svatavā̃ḥ pāyur agne "be, O Agni, his self-strong protector".
Applies when / blocked by. The Subodhinī: svatavas means "whose strength is his own", from the root tu "to grow" with asun. Its n is the augment num of dṛksvavassvatavasāṃ chandasi (SK 3577). That the ru comes from SK 3628 and the option from SK 3630 is not stated by Bhaṭṭoji or his commentators; it is inferred.
Watch out. The 1915 print reads bhuvas and svatavā̃ḥ, with the visarga that the ru leaves.
vācaḥ patim → (visarga → s before p, ṣaṣṭhī) vācas patimIn ordinary Sanskrit a visarga before k or p stays a visarga, or becomes the special breath sounds of kupvoḥ ≍ka≍pau ca (SK 142, chapter 5). The Veda often has s there instead: vācas patim, not vācaḥ patim. Six sūtras say where. The first is general and optional; the others name the following word or the case of the word before. The running example is vācas patim "Vācaspati, the lord of speech", where the sixth-case (ṣaṣṭhī) vācaḥ keeps an s before pati.
Rule. In the Veda, a visarga optionally becomes s before a sound of the k-class or the p-class (ku, pu). This does not happen before the word pra or before an āmreḍita.
Example.
ṛtaḥ kaviḥ → visarga becomes s before k → agne trātar ṛtas kaviḥ "O Agni, protector, seer of the ṛta".
viśvataḥ pṛthuḥ → girir na viśvatas pṛthuḥ "broad on all sides like a mountain".
Applies when / blocked by. Not before pra: agniḥ pra vidvān. Not before an āmreḍita: paruṣaḥ paruṣaḥ "joint by joint", a doubling for pervasion (vīpsā), so the Subodhinī. Bhaṭṭoji also gives vasunaḥ pūrvyas patiḥ "the ancient lord of wealth" as "not here". The visarga of vasunaḥ stays before p, while pūrvyas in the same line shows the s.
Watch out. That vasunaḥ shows the option not taken is not stated by Bhaṭṭoji or his commentators; it is inferred. Bhaṭṭoji's gloss supplies "visarga becomes s before ku and pu". No commentator says from which rule those words come.
Rule. A visarga becomes s before kaḥ, karat, karati, kṛdhi and kṛta, all forms of kṛ "to make". The visarga of aditiḥ is excepted.
Example.
apaḥ kaḥ → visarga becomes s → pradivo apas kaḥ "from of old you have done the work".
yathā no vasyasas karat "so that he may make us better off".
supeśasas karati "he makes them well-adorned".
uruṇas kṛdhi "make wide room for us".
somaṃ na cārũ maghavatsu nas kṛtam "make us dear among the patrons, like soma".
Applies when / blocked by. Not for aditi: yathā no aditiḥ karat "so that Aditi may make it for us".
Watch out. The Subodhinī derives all five from kṛ:
Whether "in the Veda" runs on from SK 3634 is not said by Bhaṭṭoji or his commentators.
Rule. The visarga of a fifth-case (ablative) word becomes s before pari when pari means adhi, "above".
Example.
divaḥ pari → visarga becomes s → divas pari prathamaṃ jajñe "he was first born above heaven".
Applies when / blocked by. Not when pari means something else: divas pṛthivyāḥ pary ojaḥ. There, the Subodhinī says, pari means "all round".
Rule. The visarga of a fifth-case word becomes s variously (bahulam) before pātu, a form of pā "to protect".
Example.
divaḥ pātu → visarga becomes s → sūryo no divas pātu "may Sūrya protect us from heaven".
Applies when / blocked by. Bhaṭṭoji: "of a fifth-case word" applies here too, carried from SK 3636.
Watch out. The Subodhinī reports two views of the sūtra word pātau. Some take it as naming the root pā; others, looking at the example, take it as imitating the imperative pātu.
Rule. The visarga of a sixth-case (genitive) word becomes s before pati, putra, pṛṣṭha, pāra, pada, payas and poṣa.
Example.
vācaḥ patim → visarga becomes s → vācas patiṃ viśvakarmāṇam "Vācaspati, the all-maker".
divas putrāya sūryāya "to Sūrya, the son of heaven".
divas pṛṣṭhaṃ bhandamānaḥ "… the back of heaven, rejoicing".
tamasas pāram asya "the far shore of this darkness".
parivīta iḻas pade "enveloped, in the place of Iḻā".
divas payo didhiṣāṇāḥ "longing for the milk of heaven".
rāyas poṣaṃ yajamāneṣu "increase of wealth among the sacrificers".
Applies when / blocked by. Bhaṭṭoji gives only examples. That the operation is "visarga becomes s", as in SK 3635–3636, is not stated by Bhaṭṭoji or his commentators here; it is inferred. The Subodhinī: in vācaspatim the sixth-case ending is not dropped, by tatpuruṣe kṛti bahulam (SK 972, chapter 24).
Watch out. ḻ writes the Vedic ḷa.
Rule. The visarga of iḍāyāḥ "of Iḍā" optionally becomes s before pati, putra and the rest of SK 3638.
Example.
iḻāyāḥ putraḥ → visarga becomes s → iḻāyās putraḥ; or iḻāyāḥ putraḥ "son of Iḻā".
iḻāyās pade or iḻāyāḥ pade "in the place of Iḻā".
Applies when / blocked by. The sūtra writes ḍ; the Vedic forms have ḻ. The 1915 print reads iḻāyās with ḻ.
Watch out. Bhaṭṭoji ends the unit by recalling nisas tapatāv anāsevane (SK 2403, chapter 44). It makes the s of nis retroflex before tapati, when the sense is not repetition: niṣṭaptaṃ rakṣo niṣṭaptā arātayaḥ "scorched is the demon, scorched are the hostile powers". With repetition: nistapati "he heats again and again". That rule is recalled because the next sūtras go on with the retroflex for s.
tribhis tvam → (s → ṣ before t-initial yuṣmad; t → ṭ) tribhiṣ ṭvam; abhi su naḥ → (abhi lengthened, SK 3536) abhī su naḥ → (ṣ, 3644) abhī ṣu ṇaḥ → (ṇ, 3649) abhī ṣu ṇaḥChapter 8 taught apadāntasya mūrdhanyaḥ (SK 210), the heading for a retroflex (mūrdhanya) substitute; with SK 335 it makes ṣ for s. iṇkoḥ (SK 211) gave the sounds after which it does so. Chapter 8 also taught raṣābhyāṃ no ṇaḥ samānapade (SK 235): n becomes ṇ after r or ṣ within one word. Ten sūtras close the Vedic rules with more cases of both. The s becomes ṣ across a word boundary, and nas "us" takes ṇ. The running examples are tribhiṣ ṭvam "you, with the three" and abhī ṣu ṇaḥ.
Rule. Within a pāda (antaḥpādam), a retroflex replaces s before three kinds of word. They are forms of yuṣmad "you" beginning with t, the word tat "that", and tatakṣuḥ "they fashioned".
Example.
tribhis tvam → s becomes ṣ; t beside ṣ becomes ṭ → tribhiṣ ṭvaṃ deva savitaḥ "with the three …, you, O god Savitṛ".
The t-forms of yuṣmad are tvam, tvā, te and tava:
tebhiṣ ṭvā; ābhiṣ ṭe; apsv agne sadhiṣ ṭava "in the waters, O Agni, is your seat".
With tat: agniṣ ṭad viśvam. With tatakṣuḥ: dyāvā pṛthivī niṣ ṭatakṣuḥ "they fashioned heaven and earth".
Applies when / blocked by. Only within a pāda. tad agnis tad aryamā "that is Agni, that is Aryaman". yan ma ātmano mindābhūd agnis tat punar āhār jātavedā vicarṣaṇiḥ. Here, Bhaṭṭoji says, agniḥ ends the preceding pāda and does not stand inside it.
Watch out. "Beginning with t" comes from hrasvāt tādau taddhite (SK 1325, chapter 29), so the Subodhinī. It qualifies yuṣmad only; tat and tatakṣuḥ always begin with t. That "a retroflex replaces s" comes from SK 210 and SK 335 is not stated by Bhaṭṭoji or his commentators; it is inferred. So is the step t → ṭ (ṣṭutva, SK 113).
Rule. In the Yajurveda, in the view of some teachers, a retroflex replaces s before yuṣmad, tat and tatakṣuḥ. Bhaṭṭoji takes "of some" as an option.
Example.
arcibhis tvam → s becomes ṣ → arcibhiṣ ṭvam; or arcibhis tvam.
agniṣ ṭe agram. arcibhiṣ ṭatakṣuḥ.
Watch out. Neither Bhaṭṭoji nor the Subodhinī says that "within a pāda" runs on here, so it is not supplied. There is no Subodhinī on this sūtra.
Rule. In the Veda, the s of stuta "praised" and stoma "praise" optionally becomes ṣ after a conditioning sound in the word before.
Example.
nṛbhiḥ stutasya → the s of stuta becomes ṣ; t becomes ṭ → nṛbhiṣṭutasya; or nṛbhiḥ stutasya "of the one praised by men". (The visarga of nṛbhiḥ does not appear in the transmitted form; the text does not explain this.)
go + stoma → goṣṭomam; or gostomam.
Applies when / blocked by. Bhaṭṭoji: SK 3643 already gives this, so the sūtra is for elaboration (prapañca). The Subodhinī: "in the Veda" is still needed, for the rules that follow.
Watch out. In nṛbhiṣṭutasya the ṣ is the first sound of stuta, not the visarga of nṛbhiḥ. goṣṭomam shows this: it has no s but that of stoma. Bhaṭṭoji gives every example with and without ṣ, but does not say "optionally". That the rule is optional is not stated by Bhaṭṭoji or his commentators; it is inferred from the examples.
Rule. In the Veda, an s after a conditioning sound (nimitta) in the prior word (pūrvapada) optionally becomes ṣ.
Example.
divi sthaḥ → s becomes ṣ; th becomes ṭh → yad indrāgnī diviṣṭhaḥ "when, O Indra and Agni, you two are in heaven".
The option not taken: yuvaṃ hi sthaḥ svarpatī "for you two are lords of the light".
Applies when / blocked by. The Subodhinī: pūrvapada here means any preceding word, not only the first member of a compound. So the rule works within a sentence, as both examples show.
Watch out. The conditioning sound is probably that of iṇkoḥ (SK 211); the text does not say. That "in the Veda" runs on from SK 3642 through SK 3646 is not stated by Bhaṭṭoji or his commentators; it is inferred. The Subodhinī says only "for the following".
Rule. In the Veda, the s of the particle (nipāta) su, named suñ, becomes ṣ after a conditioning sound in the prior word.
Example.
abhi su naḥ → abhi lengthened (ikaḥ suñi, SK 3536) → abhī su naḥ
→ s of su becomes ṣ → abhī ṣu naḥ → n becomes ṇ (SK 3649) → abhī ṣu ṇaḥ.
Likewise ūrdhva ū ṣu ṇaḥ "upright, indeed, for us".
Applies when / blocked by. Bhaṭṭoji's gloss carries "after a conditioning sound in the prior word" from SK 3643. He does not repeat "optionally", so the rule is taken as obligatory. The Subodhinī gives the lengthening and the ṇ.
Rule. In the Veda, the s of the root san "to win" becomes ṣ when san no longer ends in n.
Example. The Subodhinī's steps:
go + san + viṭ (SK 3413) → ā for the n of san (SK 2982) → go-sā
→ s becomes ṣ → goṣāḥ: goṣā indo nṛṣā asi "you, O Indu, are a winner of cows, a winner of men".
Applies when / blocked by. Where san keeps its n, there is no ṣ: gosaniḥ. There the affix is in (SK 3408), and san keeps its n.
Rule. In the Veda, the s of the root sah "to overpower" becomes ṣ after pṛtanā "army" and ṛta.
Example.
pṛtanā + sāham → s becomes ṣ → pṛtanāṣāham "the overpowerer of armies".
ṛtāṣāham. By the word ca "and", also ṛtīṣāham.
Watch out. Bhaṭṭoji gives only the forms, and there is no Subodhinī. That the operation is "s becomes ṣ" is not stated by Bhaṭṭoji or his commentators here; it is inferred from the run of rules and the forms. Nothing is said about the long vowels at the end of pṛtanā-, ṛtā-, ṛtī-.
Rule. In the Veda, after ni, vi and abhi, when the augment a (aṭ) stands between, a retroflex optionally replaces s.
Example.
ni + a + sīdat → s becomes ṣ → nyaṣīdat; or nyasīdat.
vyaṣīdat or vyasīdat.
abhyaṣṭaut or abhyastaut, with vṛddhi by uto vṛddhir luki hali (SK 2443, chapter 45).
Applies when / blocked by. The Subodhinī: the word "not" (na) runs on from na raparasṛpi… (SK 3168, chapter 68). An option on a prohibition comes to an option on the change itself. Hence Bhaṭṭoji's gloss, "a retroflex replaces s", read with the sūtra's "optionally".
Rule. In the Veda, ṇ replaces an n that follows an avagraha ending in ṛ. An avagraha is a member that is recited separately, as in the word-by-word text.
Example.
nṛ-manāḥ → n becomes ṇ after nṛ → nṛmaṇāḥ "manly-minded".
pitṛ-yānam → pitṛyāṇam "the path of the fathers".
Applies when / blocked by. That this is the ṇ of raṣābhyāṃ no ṇaḥ samānapade (SK 235, chapter 8) is not stated by Bhaṭṭoji or his commentators; it is inferred. SK 235 itself needs both sounds in one word.
Watch out. Scholars discuss this. Bhaṭṭoji reads the compound as "an avagraha that ends in ṛ". The Subodhinī reads it as "the ṛ that is an avagraha". For it, ṛ is called avagraha because it is fit to be separated. The ṇ is made in the continuous text, while the separation belongs to the word-by-word text. The result is the same.
Rule. The n of nas "us" becomes ṇ after a conditioning sound in a root (dhātu), and after uru and ṣu.
Example.
rakṣa naḥ → rakṣa lengthened (dvyaco 'tas tiṅaḥ, SK 3537) → n becomes ṇ after the r and ṣ of the root rakṣ → agne rakṣā ṇaḥ "O Agni, protect us". Also śikṣā ṇo asmin "help us in this".uru ṇas kṛdhi "make wide room for us".abhī ṣu ṇaḥ; mo ṣu ṇaḥ.Applies when / blocked by. The Subodhinī: ṣu is the particle su already made ṣu by SK 3644. It is not the locative plural ending su. So indro dhartā gṛheṣu naḥ keeps its n. In every example nas is the substitute of asmad "we".
Watch out. Whether "in the Veda" runs on from SK 3648 is not said by Bhaṭṭoji or his commentators. With this sūtra the eighth book ends, and with it the Vedic section (vaidikī prakriyā).
babhūva).akaram).vibhaktum. Its sense, tumartha, is carried through SK 3439–3444 from SK 3436.vibhājam, apalupam (SK 3439).vicaritoḥ, vilikhaḥ (SK 3440, SK 3443, SK 3444).vibhājam, haimantikam.madhavyaḥ.vicaritoḥ.mlecchitavai, avagāhe, didṛkṣeṇyaḥ, kartvam (SK 3441).rā́trī against rātrī́.vā́rṣikam against vārṣikám.mudgalā́nī.mudgalānī.kadrūḥ, kamaṇḍalūḥ (SK 3449).āviṣṭyaḥ (SK 1324, SK 3449).śaunakinaḥ (SK 1486).śaramayam (SK 3453).mauñjam, āsurī.meghya, asurya.varcasvat.mūrdhanvatīḥ (SK 3601).mādhavaḥ, veśobhagyaḥ.pūrviṇa (SK 3479).sarvatātiḥ, śivatātiḥ (SK 3488–3490).idā, tarhi, kathā, pratnathā.prataram.vṛkatiḥ, jyeṣṭhatātim.hatamātā.bahuprajāḥ.śrātāḥ, āśīrtaḥ.vārāhī, aṣṭāpadī.yajyamānaḥ.abhra ā̃ apaḥ.hariścandraḥ.viśvācī.sadhamāda, sadhastha.avāpsuḥ, ānaṭ.śamitā, viśastā.gatvā, viyūya.vadh-i-s-am.sudhiyaḥ, vibhuvam.sudhyo3.śrudhī, rārandhi.rārandhi.rajiṣṭham.ṛtvyam, vāstvam.duhre.dhītī.á); elsewhere the mantras are printed without accents.sarasī́.nāváyā.asme.nābhā, sukṣetriyā.vadhīm.gatvāya.devāsaḥ.asthabhiḥ.kīdṛṅṅ indraḥ.taturiḥ.tṝ is doubled, not tur (SK 2243).ahrutam, sudhitam.sasanivāṃsam.varūtrī, pāvakā.pāvaka.ās-t.avīvṛdhat.aśvāyantaḥ.dhiṣīya.bobhūtu, karikrat.yé3).yātì3).prātaritvan (SK 3629).paruṣaḥ paruṣaḥ).| SK | what it does | example |
|---|---|---|
| 3387 | punarvasu for two stars may be singular | punarvasu nakṣatram |
| 3388 | the same for viśākhā | viśākhā nakṣatram |
| 3389 | pati with a sixth-case word is optionally ghi; all Vedic operations optional; nabhas etc. bha before vat | patinā, nabhasvat, vṛṣaṇvasuḥ |
| 3390 | ayasmaya and the rest have both names, bha and pada | ṛkvatā, vājina |
| 3391 | gati and upasarga may follow the root (against SK 2230) | (see 3392) |
| 3392 | they may stand separated from the root | haribhyāṃ yāhy oka ā |
| 3393 | liṭ after indh and bhū is kit; rejected by a vārttika | samīdhe, babhūva |
| 3394 | the object of hu takes the third case, or the second | yavāgvā 'gnihotraṃ juhoti |
| 3395 | in the Brāhmaṇa the object of div takes the second, not the sixth | gām … dīvyeyuḥ |
| 3396 | the sixth stands variously for the fourth (and the fourth for the sixth) | vanaspatīnām, tasyai |
| 3397 | the sixth stands variously for the instrument of yaj | ghṛtasya yajate |
| 3399 | hemantaśiśirau, ahorātre take the gender of the prior member | ahorātre |
| 3398 | ghasḷ replaces ad variously | ghastām, sagdhiḥ |
| 3400 | the luk of śap in the ad class applies variously | hanati, trādhvam |
| 3401 | ślu in the hu class applies variously | dāti, vivaṣṭi |
| 3402 | in a mantra, cli is lost after ghas, hvṛ, naś and the rest | akṣan, praṇak, akran |
| 3403 | six ready-made forms with ām and an auxiliary | abhyutsādayām akaḥ, vidām akran |
| 3404 | after bare gup, caṅ optionally for cli | ajūgupatam |
| 3405 | no caṅ after ūnaya, dhvanaya, elaya, ardaya | mā … ūnayīḥ |
| 3406 | aṅ optionally for cli after kṛ, mṛ, dṛ, ruh | akaram, āruhat |
| 3407 | seventeen ready-made kṛtya forms | niṣṭarkya, devahūya, upacāyyapṛḍam |
| 3408 | in after van, san, rakṣ, math with an object upapada | goṣaṇim, pathirakṣī |
| 3409 | ṇvi after sah with an upapada | pṛtanāṣāṭ |
| 3410 | ṇvi after vah too | dityavāṭ |
| 3411 | ñyuṭ after vah with kavya, purīṣa, purīṣya | kavyavāhanaḥ |
| 3412 | ñyuṭ with havya when not inside a pāda; inside, ṇvi | havyavāhanaḥ / havyavāḷ |
| 3413 | viṭ after jan, san, khan, kram, gam; ā for the nasal | abjāḥ, goṣāḥ, agregāḥ |
| 3414 | in a mantra, ṇvin in three compounds; ḍas where pada status comes | śvetavāḥ / śvetavāhau |
| 3415 | ṇvin after yaj with ava | avayāḥ, avayājau |
| 3416 | avayāḥ, śvetavāḥ, puroḍāḥ, ukthaśāḥ keep the long vowel in address | śvetavāḥ (vocative) |
| 3417 | vic after yaj with upa | upayaṭ |
| 3418 | manin, kvanip, vanip, and vic, after ā-final roots | sudāmā, sudhīvā, bhūridāvā, kīlālapāḥ |
| 3419 | kvip after han variously, even with other upapadas; recap of the Vedic kṛt rules | mātṛhā, aghāyuḥ, javaḥ |
| 3420 | in a mantra, ktin after eight roots, and the affix is udātta | vṛṣṭíṃ diváḥ (udātta on ṭí) |
| 3421 | yuc for khal after roots of motion | sūpasadanaḥ |
| 3422 | yuc after other roots too | suvedanām |
| 3423 | luṅ, laṅ, liṭ optionally in any time | adyā mamāra |
| 3424 | leṭ in the senses of liṅ | (see 3426–3431) |
| 3425 | sip variously before leṭ | tāriṣat |
| 3426 | the i of leṭ's parasmaipada endings optionally lost | tāriṣat |
| 3427 | leṭ takes aṭ or āṭ, both pit; sip variously ṇit | tāriṣat, patāti |
| 3428 | the s of leṭ's uttama endings optionally lost | karavāva / karavāvaḥ |
| 3429 | the ā of leṭ's endings becomes ai (not the augment āṭ) | mādayaite, yajāti |
| 3430 | elsewhere, the e of leṭ optionally becomes ai | īśai, gṛhyāntai |
| 3431 | leṭ also for a bargain and for apprehension | nej jihmā … patāma |
| 3432 | śāyac (and śānac) for śnā before hi; bh for the h of hṛ, grah | gṛbhāya, badhāna, jabhāra |
| 3433 | vikaraṇas, and eleven other things, are exchanged variously | bhedati, taruṣema, annādāya |
| 3434 | aṅ before āśīrliṅ; ak after dṛś | vocema, dṛśeyam |
| 3435 | an affix after a root is both sārvadhātuka and ārdhadhātuka | vardhantu, viśṛṇvire, babhriḥ |
| 3436 | fifteen affixes in the sense of tumun, some with fixed accent | jīváse (udātta on va), kartave |
| 3437 | prayai, rohiṣyai, avyathiṣyai in the sense of tumun | prayai = prayātum |
| 3438 | dṛśe, vikhye in the sense of tumun | dṛśe = draṣṭum |
| 3439 | with śak as upapada, ṇamul and kamul in the sense of tumun | vibhājaṃ nāśakat, apalupaṃ nāśakat |
| 3440 | with īśvara as upapada, tosun and kasun in the sense of tumun | īśvaro vicaritoḥ, īśvaro vilikhaḥ |
| 3441 | tavai, ken, kenya, tvan in the kṛtya sense | na mlecchitavai, avagāhe, didṛkṣeṇyaḥ, kartvam |
| 3442 | avacakṣe is a nipātana in the kṛtya sense | ripuṇā nāvacakṣe |
| 3443 | tosun after sthā, iṇ … jan marking the limit of an action | āsaṃsthātoḥ sīdanti, udetoḥ |
| 3444 | kasun after sṛp and tṛd in the same sense; end of book 3 | purā krūrasya visṛpaḥ, ātṛdaḥ |
| 3445 | ṅīp after rātri in the Veda, except before jas | rā́trī vy àkhyad āyatī́ |
| 3446 | ṅīṣ obligatory after bahu and the rest | bahvīṣu |
| 3447 | ṅīṣ obligatory after bhu in ḍu; vārttika: lit ṅīṣ with ānuk after mudgala | vibhvī, prabhvī; mudgalā́nī |
| 3448 | dīrghajihvī is a nipātana | āsurī vai dīrghajihvī |
| 3449 | ūṅ after kadru, kamaṇḍalu; vārttikas: guggulu etc., and tyap in āviṣṭya | kadrūḥ, guggulūḥ, āviṣṭyaḥ |
| 3450 | ṭhañ after varṣā, against ṭhak: a difference of accent | vā́rṣikam |
| 3451 | ṭhañ after vasanta | vāsantikam |
| 3452 | ṭhañ after hemanta, split off for SK 1390; ṇini by SK 1486 | haimantikam; śaunakinaḥ |
| 3453 | mayaṭ "made of" after a two-vowel stem | śaramayaṃ barhiḥ, parṇamayī |
| 3454 | no mayaṭ after a stem with u, vardhra, bilva | mauñjam, vārdhrī, bailvaḥ |
| 3455 | ḍha after sabhā in the sense "fit for" | sabheyo yuvā |
| 3456 | yat "being there" after a seventh-case word; the Vedic heading to SK 3490 | meghyāya, vidyutyāya; mauñjavataḥ |
| 3457 | ḍyaṇ after pāthas, nadī | pāthyaḥ, nādyaḥ |
| 3458 | aṇ after veśanta, himavat | vaiśantībhyaḥ, haimavatībhyaḥ |
| 3459 | ḍyat or ḍya after srotas, optionally; else yat | srotyaḥ, srotasyaḥ |
| 3460 | yan after sagarbha, sayūtha, sanuta | sagarbhyaḥ, sayūthyaḥ, sanutyaḥ |
| 3461 | ghan after tugra; else yat | tugriyāsu, tugryāsu |
| 3462 | yat after agra, taught again | agryaḥ |
| 3463 | gha and cha after agra, and yat | agryaḥ, agriyaḥ, agrīyaḥ |
| 3464 | gha after samudra, abhra; last rule of bhava | samudriyāḥ, abhriyasya |
| 3465 | yat after barhis "given on it" | barhiṣyeṣu |
| 3466 | yat after dūta for his share or task | dūtyam |
| 3467 | yat after rakṣas, yātu "that which slays them" | rakṣasyā tanūḥ, yātavyā |
| 3468 | yat after revatī, jagatī, haviṣyā "praising them" | revatyam, jagatyam, haviṣyam |
| 3469 | yat after asura "its own" | asuryam |
| 3470 | aṇ after asura when māyā is meant | āsurī māyā |
| 3471 | yat naming bricks after their laying mantra, with luk of matup | varcasyāḥ, ṛtavyāḥ |
| 3472 | aṇ after aśvimat, same conditions | āśvinīr upadadhāti |
| 3473 | matup after mūrdhavat for vayasyā bricks | mūrdhanvatīr upadadhāti |
| 3474 | yat in the matup sense for a month or a body | nabhasyaḥ, ojasyā tanūḥ |
| 3475 | ña, and yat, after madhu in the matup sense | mādhavaḥ, madhavyaḥ |
| 3476 | yat and kha after ojas for a day | ojasyam, ojasīnam |
| 3477 | yal after bhaga with veśas or yaśas before it | veśobhagyaḥ, yaśobhagyaḥ |
| 3478 | kha also, under the same conditions | veśobhagīnaḥ, yaśobhagīnaḥ |
| 3479 | ina and ya after pūrva "made by those of old" | pūrviṇebhiḥ, pūrvyāsaḥ |
| 3480 | yat after ap "prepared with water" | apyaṃ haviḥ |
| 3481 | gha after sahasra "equal to a thousand" | sahasriyāsaḥ |
| 3482 | gha after sahasra in the matup sense too | sahasriyaḥ |
| 3483 | ya after soma "deserves" | somyo brāhmaṇaḥ |
| 3484 | ya after soma in the mayaṭ sense | somyaṃ madhu |
| 3485 | yat after madhu in the mayaṭ sense | madhavyaḥ |
| 3486 | yat after vasu "collection" and in the mayaṭ sense; vārttika: yat after chandas | vasavyaḥ; chāndasyaḥ |
| 3487 | gha after nakṣatra in its own sense | nakṣatriyebhyaḥ |
| 3488 | tātil after sarva, deva in their own sense | sarvatātim, devatātim |
| 3489 | tātil after śiva, śam, ariṣṭa "one who makes" | śivatātiḥ, śantātī, ariṣṭatātaye |
| 3490 | tātil after the same in the sense of the state; end of book 4 | śivatātiḥ, śantātiḥ, ariṣṭatātiḥ |
| 3491 | añ after saptan for measure and group; vārttikas: ḍini, vatup | sāptāni, pañcadaśinaḥ, tvāvān |
| 3492 | yat after any stem for "deserves that" | sādanyam, vidathyam |
| 3493 | cha after a stem ending in vatsara, in the "accomplished" senses | idvatsarīyaḥ |
| 3494 | kha, and cha, after saṃvatsara and parivatsara | saṃvatsarīṇaḥ, saṃvatsarīyaḥ |
| 3495 | ghas after ṛtu for "that has come to it" | ṛtviyaḥ |
| 3496 | vati after an upasarga denoting a factor of the action | udvataḥ, nivataḥ |
| 3497 | the augment thaṭ, and maṭ, to ordinal ḍaṭ; SK 1889 cited | pañcatham, pañcamam |
| 3498 | vini variously; vārttikas: lengthening before vin; ī and vanip | ubhayāvinam, rathīḥ, maghavānam |
| 3499 | dā after idam, rhil after tad | idā, tarhi |
| 3500 | thā after kim for cause and manner | kathā |
| 3501 | paśca, paścā fixed for avara; SK 2007 cited | paśca hi saḥ, kariṣṭhaḥ |
| 3502 | thāl in the sense "like" after pratna, pūrva, viśva, ima | pratnathā, imathā |
| 3503 | amu in the setting of SK 2004 | prataram |
| 3504 | til after vṛka, tātil after jyeṣṭha | vṛkatiḥ, jyeṣṭhatātim |
| 3505 | ṭac after a neuter tatpuruṣa in an or as | brahmasāmam, devacchandasāni |
| 3506 | bahuprajāḥ fixed | bahuprajāḥ |
| 3507 | datṛ for danta in a bahuvrīhi | ubhayatodataḥ |
| 3508 | no kap after a ṛ-final bahuvrīhi | hatamātā |
| 3509 | the abhyāsa of tuj and the like lengthened; doubling optional (vārttika) | tūtujānaḥ; jāgāra |
| 3510 | saṃprasāraṇa of hvā variously; vārttikas on tri + ṛc and rayi | āhuve, tṛcam, revān |
| 3511 | kī for cāy variously | cikyuḥ; not nicāyya |
| 3512 | nine fixed forms | apaspṛdhethām, tityāja, āśīrtaḥ |
| 3513 | ā optionally for the e of khid | cikhāda / cikheda |
| 3514 | śīrṣan for śiras | śīrṣṇaḥ |
| 3515 | pūrvasavarṇadīrgha optional; the option reaches SK 194 and SK 330 | vārāhī / vārāhyau; śamīm / śamyam |
| 3516 | śi lost variously; vārttika: pararūpa in eman | yā, tā tā; apāṃ tveman |
| 3517 | bhayya, pravayyā fixed, with ay before yat; vārttika hradayyā | bhayyaḥ, pravayyā, hradayyāḥ |
| 3518 | eṅ stays before a inside a pāda, unless v or y follows | upaprayanto adhvaram |
| 3519 | eṅ stays before seven words, even with v or y | vasubhir no avyāt |
| 3520 | in the Yajus, uro stays before a | uro antarikṣam |
| 3521 | in the Yajus, āpo, juṣāṇo, vṛṣṇo, varṣiṣṭhe, ambe, ambāle stay | ambe ambāle ambike |
| 3522 | in the Yajus, eṅ of aṅga, and eṅ before aṅga, stays | prāṇo aṅge aṅge adīdhyat |
| 3523 | in the Yajus, eṅ stays before an anudātta a followed by ku or dh | ayáṃ só agníḥ |
| 3524 | in the Yajus, eṅ stays before an anudātta avapathāḥ | trī́ rudrébhyo avapathāḥ |
| 3525 | āṅ becomes nasal before a vowel and stays; vārttika īṣā akṣa | abhra ā̃ apaḥ, īṣā akṣaḥ |
| 3526 | the su of sya lost before a consonant, variously | eṣa sya bhānuḥ |
| 3527 | suṭ to candra after a short vowel, in a mantra | hariścandraḥ |
| 3528 | pitarāmātarā fixed; SK 1012 cited; vārttika on adri | pitarā mātarā, sagarbhyaḥ, devadrīcī |
| 3529 | sadha for saha before māda, stha | sadhamāde, sadhastham |
| 3530 | kava, kā for ku before pathin | kavapathaḥ, kāpathaḥ |
| 3531 | sāḍhyai, sāḍhvā, sāḍhā fixed from sah | sā́ḷhā |
| 3532 | ā for the final of aṣṭan before a second member | aṣṭāpadī |
| 3533 | final of soma, aśva, indriya, viśvadevya long before matup | aśvāvatīm |
| 3534 | final of oṣadhi long before a non-first-case ending | óṣadhībhyaḥ |
| 3535 | final of tu, nu, gha … uruṣya long in a ṛc | ā tū na indra, bharatā |
| 3536 | an ik-final word long before su | abhī ṣu ṇaḥ |
| 3537 | final a of a two-vowel verb form long | vidmā, cakrā |
| 3538 | final of a nipāta long | evā hi te |
| 3539 | lengthening seen in other prior members too | pūruṣaḥ, daṇḍādaṇḍi |
| 3540 | lengthening before nām optional | dhātṝṇām / dhātṛṇām |
| 3541 | penultimate after ṣ optionally long before a strong ending | ṛbhukṣāṇam / ṛbhukṣaṇam |
| 3542 | janitā fixed, with ṇi lost before iṭ | janitā |
| 3543 | śamitā in the sacrifice: ṇi dropped, as in janitā | śamitā (= śamayitā) |
| 3544 | long ū for yu, plu before lyap | viyūya, viplūya |
| 3545 | āṭ also for aṅgas not beginning with a vowel | ānaṭ, āvaḥ |
| 3546 | aṭ, āṭ variable: absent without māṅ, present with it | janiṣṭhāḥ, mā … avāpsuḥ |
| 3547 | re for ire; asiddha for the loss of ā | dadhre |
| 3548 | bhū, sudhī take yaṇ and iyaṅ/uvaṅ; vārttika on tanū | vibhvam / vibhuvam; tanvam / tanuvam |
| 3549 | tan, pat lose the upadhā before kit/ṅit | vitatnire, paptima |
| 3550 | ghas, bhas lose the upadhā before kit/ṅit | sagdhiḥ, babdhām |
| 3551 | dhi for hi after śru, śṛṇu, pṝ, kṛ, vṛ | śrudhī, śṛṇudhī, rāyaspūrdhi, uruṇaskṛdhi |
| 3552 | hi (for sip) is optionally apit | — |
| 3553 | dhi for hi also when hi is not ṅit | rārandhi, prayandhi, yuyodhi |
| 3554 | in mantras, ātman loses its first sound before āṅ (ṭā) | tmanā |
| 3555 | ra for the ṛ of ṛju, optionally, before iṣṭhan etc. | rajiṣṭham / ṛjiṣṭham |
| 3556 | nipātanas ṛtvya, vāstvya, vāstva, mādhvī, hiraṇyaya | mādhvīḥ, hiraṇyayena |
| 3557 | the augment ruṭ, variously | duhre, duhrate, adṛśram |
| 3558 | ais for bhis only variably | devebhiḥ |
| 3559 | no adḍ after itara | itaram (ordinary itarat) |
| 3560 | ktvā also where lyap is due | paridhāpayitvā |
| 3561 | seventeen Vedic substitutes for sup endings | panthāḥ, vyoman, dhītī, sarasī́, nāváyā |
| 3562 | maś (m) for the am of mip | vadhīm |
| 3563 | t of an ātmanepada ending lost | aduhra, śaye |
| 3564 | dhvāt for dhvam | vārayadhvāt |
| 3565 | yajadhvainam: m of dhvam lost before enam | yajadhvainam |
| 3566 | tāt for the second-person plural ta | kṛṇutāt, gamayatāt |
| 3567 | tap, tanap, tana, thana for ta | śṛṇota, sunotana, jujuṣṭana, yatiṣṭhana |
| 3568 | mas takes a final i | emasi, smasi |
| 3569 | ktvā takes yak | gatvāya |
| 3570 | īnam for the final of ktvā after yaj | iṣṭvīnam |
| 3571 | ī for the ā of ktvā in snātvī and the like | snātvī, pītvī |
| 3572 | jas takes asuk after an a-stem | devāsaḥ, brāhmaṇāsaḥ |
| 3573 | nuṭ to ām after śrī, grāmaṇī | śrīṇām, sūtagrāmaṇīnām |
| 3574 | nuṭ to ām after go at a pāda's end | gonām (not pāda-final gavām) |
| 3575 | anaṅ for asthi etc. even outside its conditions | asthabhiḥ, asthanvantam |
| 3576 | ī for the final of asthi etc. before a dual ending | akṣībhyām |
| 3577 | num for dṛś, svavas, svatavas before su | kīdṛṅ, svavān, svatavān |
| 3578 | ut for a root-final ṝ, variously, even without a labial | taturiḥ, jaguriḥ |
| 3579 | hru for hvṛ before a niṣṭhā | ahrutam |
| 3580 | aparihvṛtāḥ: no hru | aparihvṛtāḥ |
| 3581 | hvaritaḥ of Soma: iṭ, guṇa, no hru | hvaritaḥ |
| 3582 | eighteen nipātanas with or without iṭ, uṭ/ūṭ, i for śap | grasitām, viśastā, tarutāram, ujjvaliti |
| 3583 | saniṃ sasanivāṃsam: iṭ to kvasu; vārttika on pāvaka | sasanivāṃsam, pāvakāḥ |
| 3584 | final ā of dā, dhā optionally lost before leṭ | dadhat, dadat / dadāt |
| 3585 | short vowel for the ī of mī before a śit affix | pramiṇanti (ordinary pramīṇanti) |
| 3586 | īṭ variable; vārttikas: jasi ca etc. optional, nābhyastasyāci variable | ā idam, śatakratvaḥ, paśve, jujoṣat |
| 3587 | ṛ in the upadhā before caṅ always | avīvṛdhat |
| 3588 | no ī and no lengthening before kyac, except putra and the like | mitrayuḥ; putrīyantaḥ, janīyantaḥ |
| 3589 | nipātanas durasyu, draviṇasyu, vṛṣaṇyati, riṣaṇyati | durasyuḥ (ordinary duṣṭīyati) |
| 3590 | ā for the final of aśva, agha before kyac | aśvāyantaḥ, aghāyavaḥ |
| 3591 | ā for deva, sumna before kyac in the Kaṭha Yajus | devāyantaḥ, sumnāyantaḥ |
| 3592 | final of kavi, adhvara, pṛtanā lost before kyac in a ṛk | kavyatāyoḥ, adhvaryum, pṛtanyum |
| 3593 | hi for hā before ktvā, optionally | hitvā / hītvā |
| 3594 | i for the ā of dhā in five forms; vārttikas on mās, svavas, uṣas | sudhitam, dhiṣva, dhiṣīya; mādbhiḥ, samuṣadbhiḥ |
| 3595 | in the Veda, kṛṣ keeps the velar of its abhyāsa before yaṅ | karīkṛṣyate |
| 3596 | eighteen Vedic intensive and ślu forms with irregular abhyāsas | alarṣi, kanikradat, āganīganti |
| 3597 | sasūva, the Vedic perfect of sū | gṛṣṭiḥ sasūva sthaviram |
| 3598 | in the Veda, i variously for the abhyāsa's vowel | vivaṣṭi |
| 3599 | pra, sam, upa, ud doubled to fill the pāda | pra prāyam agniḥ |
| 3600 | in the Veda, v for the m of matup after i-vowels and r | harivate, gīrvān |
| 3601 | matup takes nuṭ after an an-stem | akṣaṇvantaḥ |
| 3602 | gha takes nuṭ after an n-stem; tuṭ after bhūridāvan; ī for rathin | supathintaraḥ, bhūridāvattaraḥ, rathītaraḥ |
| 3603 | six Vedic niṣṭhā forms taught as finished forms; most keep t | nasattam, niṣattam, gūrtam |
| 3604 | amnas, ūdhas, avas take ru or r | amna eva / amnar eva |
| 3605 | the same for bhuvas, the great utterance | bhuva iti / bhuvar iti |
| 3606 | om is pluta at a commencement | ó3m agnim īḷe purohitam |
| 3607 | ye is pluta in a sacrificial act | yé3 yajāmahe |
| 3608 | om replaces the whole ṭi in a sacrificial act | jinvato3m |
| 3609 | only the last ṭi of a yājyā is pluta | havyavāhā́3m |
| 3610 | the first vowel of brūhi, preṣya, śrauṣaṭ, vauṣaṭ, āvaha is pluta | agnaye 'nubrū́3hi |
| 3611 | in the agnīdh's directive, the first vowel and the next are pluta | ó3śrā́3vaya |
| 3612 | hi is optionally pluta in an answer | akārṣaṃ hí3 |
| 3613 | optional pluta in questioning a refuted speaker | adyāmāvāsyety ātthá3 |
| 3614 | a doubled word is pluta in a threat | dasyó3 dasyó3 |
| 3615 | in a threat, an expectant verb with aṅga is pluta | aṅga kūjá3 |
| 3616 | the ṭi of sentences under deliberation is pluta | gṛhā́3i, na hotavyá3m |
| 3617 | in bhāṣā, only the first alternative is pluta | ahir nú3 rajjur nu |
| 3618 | pluta in acceptance, affirmation, readiness to listen | hanta te dadāmí3 |
| 3619 | the pluta is anudātta at the end of a question and in praise | agnibhūta3i, māṇavaka3 |
| 3620 | an anudātta pluta when cit is used for comparison | agnicid bhāyā3t |
| 3621 | an anudātta pluta in upari svid āsīt | upári svid āsī3t |
| 3622 | a svarita pluta before an āmreḍita, in envy, approval, anger, contempt | abhirūpakà3 abhirūpaka |
| 3623 | a svarita pluta on an expectant verb in breach of conduct, wish, direction | kaṭaṃ kurù3 grāmaṃ gaccha |
| 3624 | in question and narration, every word's ṭi is a svarita pluta | agamàḥ3 pūrvà3n grāmà3n |
| 3625 | in a pluta ai, au, the i, u part is pluta | ai3tikāyana |
| 3626 | a pluta e, o splits into pluta ā + i, u | agnibhūta3i, paṭa3u |
| 3627 | that i, u becomes y, v before a vowel, in saṃhitā | agna3yāśā |
| 3628 | in the Veda, ru for the n of a matup or vasu stem in the vocative | harivo medinam |
| 3629 | dāśvān, sāhvān, mīḍhvān are nipātanas; van stems also take ru | mīḍhvas, prātaritvaḥ |
| 3630 | in the ṛcs, the ru of SK 140 is optional | paśūn tāṃś cakre |
| 3631 | n after a long vowel optionally ru before aṭ in one pāda | devā̃ acchā |
| 3632 | before aṭ, the ā before ru is always nasal | mahā̃ indraḥ |
| 3633 | the n of svatavān optionally ru before pāyu | svatavā̃ḥ pāyuḥ |
| 3634 | in the Veda, visarga optionally s before ku, pu, except pra and an āmreḍita | ṛtas kaviḥ |
| 3635 | visarga s before kaḥ, karat, karati, kṛdhi, kṛta, not for aditi | uruṇas kṛdhi |
| 3636 | a fifth-case visarga s before pari "above" | divas pari |
| 3637 | a fifth-case visarga variously s before pātu | divas pātu |
| 3638 | a sixth-case visarga s before pati, putra and the rest | vācas patim |
| 3639 | the same optionally for iḍāyāḥ; SK 2403 recalled | iḻāyās putraḥ / iḻāyāḥ putraḥ |
| 3640 | within a pāda, ṣ for s before t-forms of yuṣmad, tat, tatakṣuḥ | tribhiṣ ṭvam |
| 3641 | in the Yajus, the same optionally | arcibhiṣ ṭvam / arcibhis tvam |
| 3642 | in the Veda, ṣ optionally for the s of stuta, stoma | nṛbhiṣṭutasya, goṣṭomam |
| 3643 | ṣ optionally for s after a conditioning sound in the prior word | diviṣṭhaḥ |
| 3644 | ṣ for the s of the particle su | abhī ṣu ṇaḥ |
| 3645 | ṣ for the s of san when it has lost its n | goṣāḥ |
| 3646 | ṣ for the s of sah after pṛtanā, ṛta | pṛtanāṣāham |
| 3647 | after ni, vi, abhi with aṭ between, ṣ optional | nyaṣīdat / nyasīdat |
| 3648 | in the Veda, ṇ for n after an avagraha ending in ṛ | nṛmaṇāḥ |
| 3649 | ṇ for the n of nas after a root sound, uru, ṣu | agne rakṣā ṇaḥ |