apo bhi — "apo bhi." From "aca upasargāt taḥ" (A 7.4.47) the substitute t is
carried forward (anuvṛtti). This [rule] stands within the aṅga-adhikāra.
Consequently the affix that follows the aṅga — "bhi," in the locative — qualifies
[the environment], and by the tad-ādi-vidhi [an affix beginning with bh is
meant]. He states this with "apas takāraḥ…" [there is a t-kāra for ap]. Why "when
an affix [follows]"? [To exclude] abbhakṣaḥ [where no affix follows, hence no t].
"adbhiḥ" — the p becomes t, then jaśtva [t → d] before bh — this is the sense.
— End of the p-final [stems].
Now the ś-final [stems]. "dik" — from "diśa atisarjane" (the root diś, to
assign), kvin by "ṛtvigādi…"; su-elision (sulopa); by "vraśca…" ś → ṣ; that ṣ by
jaśtva → ḍ; that [ḍ] by "kvin-pratyayasya kuḥ" gutturalization (kutva) → g; that
[g] optionally cartva [→ k] — this is the sense [dik / dig]. "dṛśyante arthā
anayā" (that by which meanings are seen): so analysed, since dṛś belongs to the
sampad-ādi class, kvip [after dṛś]. Then su-elision; anticipating ṣatva by
"vraśca…," he says "tyadādiṣu…": "anyatrāpi" (also elsewhere) means even in the
absence of a tyadādi upapada. Because [the rule] is framed relying on the
bahuvrīhi "that from which kvin is enjoined" — this is the sense. "dṛg" — the six
[operations,] g etc., as before. — End of the ś-final.
Now the ṣ-final [stems]. "tviḍ" — "tviṣa dīptau" (the root tviṣ, to shine), kvip,
su-elision, jaśtva-ṣṭva-cartva [ṣ → ḍ → ṭ] — this is the sense. "sajūr" — "juṣī
prītisevanayoḥ" [the root juṣ], kvip; the s-substitution of saha by "sahasya saḥ
saṃjñāyām," or by the nipātana "sasajuṣo…," the sa-substitution of saha;
su-elision; by "sasajuṣo ruḥ" the s → ru; by "r̄vor upadhāyāḥ" lengthening [→
sajūr]. "āṅaḥ śāsu icchāyām" [the root śās with āṅ], kvip; by "aśāsaḥ kvāv
upasaṃkhyānam" the i-substitution of the upadhā; by "śāsivasighasīnāṃ ca" the
s → ṣ; from the stem āśiṣ, elision of the su. Having treated this much as
established (siddhavat), he says "ṣatvasya…": "āśīr" — the ṣ having become ru, by
"r̄vor upadhāyāḥ" there is lengthening — this is the sense. — End of the ṣ-final.
Now the s-final [stems]. "asau" — the word adas, even in the feminine, has the
same form in su as in the masculine — this is the sense. When "adas au" stands,
he shows the derivation with "tyadādyatvam…": the tyadādi a-substitution [of the
final of adas], pararūpa, ṭāp, śī-substitution for auṅ, guṇa by "ād guṇaḥ,"
[yielding] "ade"; because the e is lengthened, ū-substitution, and the d → m —
this is the sense. The a-substitution, pararūpa, and ṭāp are to be understood
everywhere a vibhakti is present. "amūr" — in jas: a-substitution, pararūpa,
ṭāp; pūrvasavarṇa-dīrgha; ū-substitution and m-substitution (matva, d→m) [→ amūḥ]. Because
ṭāp is present, there being no a-final stem, there is no śī-substitution for jas;
and there being no e, no i-substitution (ittva). "amūm" — like the masculine;
the ū-substitution is the difference. "amū" [nom./acc. du.] — in auṭ, like au [of
the masculine]. "amūr" [acc. pl.] — in śas the form is like in jas; because it is
feminine, no na-substitution (natva). "amuyā" — when "adas ā" (ṭā) stands:
a-substitution, pararūpa, ṭāp; by "āṅi cāpaḥ" et-substitution [→ e]; ay-ādeśa
[e → ay]; ū-/m-substitution (matva, d→m) [→ amuyā] — this is the sense. "amūbhyām" —
a-substitution, pararūpa, ṭāp; ū-/m-substitution (matva, d→m) — this is the sense. Likewise
"amūbhiḥ." "amuṣyai" — when "adas e" (ṅe) stands: a-substitution, pararūpa, ṭāp;
the syāṬ-augment and shortening by "syāḍ ḍhrasvau"; u-/m-substitution (matva, d→m); ṣatva —
this is the sense. "amūbhyaḥ" too is like bhyām [in its operations]. Because ṭāp
is present, there being no a-final stem, there is no et-substitution. "amuṣyāḥ"
[abl. sg.] — in ṅasi and ṅas it is like [ṅe]; the ru-substitution and visarga are
the difference. "amuyoḥ" — in os [gen./loc. du.]: a-substitution, pararūpa, ṭāp;
by "āṅi cāpaḥ" et-substitution; ay-ādeśa; u-substitution (mutva) — this is the
sense. "amūṣām" — in ām: a-substitution, pararūpa, ṭāp; the ām-augment [suṬ];
shortening by "syāḍ ḍhrasvau"; u-substitution (mutva) — this is the sense.
"amūṣu" — in sup [loc. pl.]: a-substitution, pararūpa, ṭāp; ū-/m-substitution (matva, d→m);
ṣatva — this is the sense. — End of the s-final.
* End of the feminine prakaraṇa. * Now the neuter prakaraṇa. *
Now the h-final [stems]. In the bahuvrīhi "su-anaḍvāhaḥ" (su + anaḍvah), "the
family whose oxen are fine," from the word su-anaḍuh, being neuter, sup is
derived. Objection: there, before su, one would expect ām by "caturanaḍuhoḥ," am
by "am saṃbuddhau," and num by "sāv anaḍuhaḥ" — since even when su is elided by
"halṅyādi…," pratyaya-lakṣaṇa remains. To this he says "svamor luk": because of
paratva, blocking the halṅyādi-elision, there is luk by "svamor napuṃsakāt." And
consequently, [su] being elided by luk, pratyaya-lakṣaṇa is absent, so ām etc. do
not occur — this is the sense. "datva" — namely by "vasusraṃsu…"; "iti śeṣaḥ"
[supply this]. Because the datva-rule stands within the pada-adhikāra, it
operates also on [a stem] ending in [anaḍuh] — this is the sense. Nor should one
raise the prohibition "na lumatā…" there either, for they say the datva, being
conditioned by pada-hood [that consists in] being sup-final, is not an
aṅga-operation [and so escapes that prohibition]. "svanaḍut" — after datva,
optional cartva. "uraḥ-prabhṛtibhyaḥ kaḥ" should not be suspected [to apply],
because in that gaṇa only the singular "anaḍvān" is read; this will be stated in
the bahuvrīhi-adhikāra in the mūla itself. "svanaḍuhī" — by "napuṃsakāc ca" the
śī-substitution for auṅ. "svanaḍvāṃhi" — by "jaśśasoḥ śiḥ" the śi-substitution;
that being a sarvanāmasthāna, ām by "caturanaḍuhoḥ," num by "napuṃsakasya
jhalacaḥ," and anusvāra by "naś ca…" [→ svanaḍvāṃhi] — this is the sense.
Whatever is to be said here has been stated in the treatment of the masculine.
— End of the h-final.
Now the v-final [stems]. In the bahuvrīhi "vimalā dyauḥ = ākāśaṃ yasya ahnaḥ"
("the day whose sky is clear"), with sub-luk [elision of the internal sup], from
the word vimaladiv, luk of su. Having treated this much as established, he says
"diva ut": "aharvimaladyu" — the v becomes u, the i becomes yaṇ — this is the
sense. The mention of "ahar" (day) is to indicate the neuter gender. Objection:
when "vimaladiv au" stands, by "napuṃsakāc ca" the śī-substitution [would give]
the form "vimaladivī" that he will state; that is wrong, because "diva ut" (the
u-substitution) would be occasioned. Nor can one say that the whole aggregate
"vimaladivī," being sup-final, is a pada, so the v is not pada-final in the form
of the word div; for, relying by pratyaya-lakṣaṇa on the vibhakti elided after
the word div per "supo dhātu…," the entity "div" is [itself] a pada. Nor can one
say that, [the vibhakti] being elided by luk, there is no pratyaya-lakṣaṇa;
because the pada-saṃjñā, being a property of the sup-final [word] and not an
aṅga-property, the prohibition "na lumatā…" does not operate there. Otherwise, in
"rājñaḥ puruṣaḥ = rājapuruṣaḥ" and the like, the n-elision etc. in the prior
member would not occur. Anticipating this, he says "antarvartinīm…":
"uttarapadatve ca" — in a compound, when the pada-saṃjñā is to be effected for
the latter member, on the elision of the internal (antarvartin) vibhakti, the
prohibition of pratyaya-lakṣaṇa is to be stated, excepting the pada-ādi
operations [operations conditioned by "beginning of a pada"]. This is a vārttika.
"vimaladivī" — the form in auṅ. "vimaladivi" — in śas, by "jaśśaso śiḥ" the
śi-substitution, the form is established. Because it is not jhal-final, no num.
"dadhisecau" — kvip after siñcati (the root sic). "dadhnaḥ secau = dadhisecau"
("the two sprinklers of curds"). In the compound, on the luk of the sup of the
prior and latter members, sup is derived afresh from the compound — such is the
position. There, relying by pratyaya-lakṣaṇa on the vibhakti elided after the
latter member, the entity "sec" being a pada, its s being pada-initial, the
prohibition of ṣatva by "sāt padādyoḥ" is desired. If only "on being the latter
member, the prohibition of pratyaya-lakṣaṇa" were stated, then, the pada-saṃjñā
of the latter member "sec" being what is to be effected, there would be the
prohibition of pratyaya-lakṣaṇa; and so, "sec" not being sup-final and hence
non-pada, its s not being pada-initial, the prohibition of ṣatva would not obtain
— hence "excepting the pada-ādi operations" was said. Since the prohibition of
ṣatva is a pada-ādi operation, in effecting it there being no prohibition of
pratyaya-lakṣaṇa, relying on pratyaya-lakṣaṇa the entity "sec" is a pada, so the
prohibition of ṣatva of its s by "sāt padādyoḥ" is unobstructed. Objection: when
the latter member "sec" is a pada, kutva of its c would occur. To this he says
"cakārasya kutvaṃ tu na": but when kutva is to be effected, "sec" has no
pada-hood at all, because kutva, being a pada-anta (pada-final) operation and
hence distinct from a pada-ādi operation, the prohibition of pratyaya-lakṣaṇa
obtains in effecting it, so there is no pada-hood — this is the sense. Objection:
"dadhi siñcati" [analysed], from [the root] with an upapada, by vic, in the
upapada-compound the ṣatva is hard to avert, since — as will be stated — the
compounding of gati/kāraka/upapada with kṛt-forms precedes the derivation of sup
("prāk subutpatteḥ"), so the entity "sec," having no internal sup and hence being
non-pada-final, "sāt padādyoḥ" does not operate there. If so — [reply:] "there is
no denotation (anabhidhāna) of vic after sic with an upapada here," is how
Kaiyaṭa resolved it. On the view "padād ādiḥ padādiḥ" [that "pada-ādi" means the
beginning of the first member], even in an upapada-compound the ṣatva is easily
avoided. Nor should one object that even so, in "paramaś cāsau daṇḍī ca =
paramadaṇḍī, sa priyo yasya saḥ = paramadaṇḍipriyaḥ," the n-elision in the latter
member of the internal tatpuruṣa would fail — because by "uttarapadatve ca" the
prohibition of pratyaya-lakṣaṇa denies its pada-hood — because only the latter
member not encroached by middle-member-hood (madhyapadatva) is intended. Let this
rest. — End of the v-final.
Now the r-final [stems]. "vār" — the word vār, ending in repha, neuter. "āpaḥ
strī bhūmni vār vārī," says Amara. From the word vār, luk of su/am, the repha →
visarga — this is the sense. "catvāri" — the word catur is always plural-final;
its śi by "jaśśasoḥ śiḥ"; that being sarvanāmasthāna, ām for the base by
"caturanaḍuhoḥ." And that [ām], like am, [is placed] after the final vowel, the
u-kāra; then yaṇ for the u-kāra [u → v]; because it is not jhal-final, no num —
this is the sense [→ catvāri]. — End of the repha-final.
Now the m-final [stems]. From the word kim, on the luk of su/am, relying on
pratyaya-lakṣaṇa and since it is followed by a vibhakti, anticipating the
ka-substitution [by "kimaḥ kaḥ"], he says "na lumatā…" [which prohibits it]. And
[stating that] in the ka-substitution [that occurs] elsewhere than for su and am,
the forms are like [the pronoun] sarva, he says "ke kāni" [nom./acc. du. ke, pl.
kāni]. "idam" — the form on luk of su/am. The rules "idamo maḥ," "daś ca," etc.
do not apply, because of the prohibition "na lumatā…" — this is the sense. "ime"
— in auṅ: a-substitution, pararūpa, m for d by "daś ca," śī-substitution by
"napuṃsakāc ca," guṇa — this is the sense. "imāni" — "jaśśasoḥ śiḥ,"
a-substitution, pararūpa, m by "daś ca," num by "napuṃsakasya jhalacaḥ,"
lengthening by "sarvanāmasthāne ca" — this is the sense. "enad vaktavyaḥ" —
"idametadoḥ" is to be supplied. And idam becomes [enad] only in am. And so in the
Bhāṣya, having read "enad iti napuṃsakaikavacane vaktavyam," the example "kuṇḍam
ānaya prakṣālayainat" is given. Considering that by "auṭ-śasoṣ ṭā-y-ām-osi ca,"
"dvitīyā-ṭauḥsv enaḥ," the ena-substitution alone [applies], he says "ene…":
"ene" etc. In truth, however, let it be formulated as "dvitīyā-ṭauḥsv enat." In
the neuter, on the luk of am, "enad" is established; but "enam, enau, enān,
enena, enayoḥ" are established by tyadādi a-substitution — so it stands in the
Bhāṣya. Truly, though, in the neuter singular am there is the enad-substitution;
for the sake of "enam, enau," etc., the ena-substitution is to be enjoined in
"dvitīyā-ṭauḥsv enaḥ." Because, by the maxim "yena nāprāpti-nyāya," the
enad-substitution being an exception (apavāda) to the tyadādi a-substitution, the
tyadādi a-substitution is impossible [there] — so it stands in the
Śabdendu-śekhara and is elaborated in the Bhāṣya-pradīpoddyota. — End of the
m-final.
Now the n-final [stems]. "brahma" — when denoting the Veda and the like, the word
brahman is neuter. "vedas tattvaṃ tapo brahma, brahmā vipraḥ prajāpatiḥ," says
Amara. On the luk of su/am, n-elision — this is the sense. "brahmaṇī" — śī for
auṅ, ṇatva by "aṭ-kupvāṅ…." But no a-elision by "vibhāṣā ṅiśyoḥ," because of the
prohibition "na saṃyogād vamantāt" — this is the sense. "brahmāṇi" — "jaśśasoḥ
śiḥ"; that being sarvanāmasthāna, the lengthening characteristic of n-final
stems. "he brahman" — the optional n-elision by "saṃbuddhau napuṃsakānām" — this
is the sense. Then from the word ahan, on the luk of su/am, when "ahan" [with ru
by "sasajuṣo ruḥ"] is obtained, he says "ro 'supi": [su] being elided by luk,
there is no pratyaya-lakṣaṇa — this is the sense. "aharbhāti" — here, on rutva, by
"haśi ca" the u-substitution and then guṇa would give "aho bhāti." In the
repha-rule [ro 'supi] there is no u-substitution for that [ahan], because in
"roḥ" the u is taken as an indicatory marker (anubandha) — this is the sense.
"ahāni" — "jaśśasoḥ śiḥ," being sarvanāmasthāna, upadhā-lengthening; but no
a-elision, because it is sarvanāmasthāna — this is the sense. In the vowel-initial
[endings] beginning with ṭā, there is a-elision: ahnā, ahne, and so on. Before
the consonant-initial [endings] beginning with bhyām he states the special case —
"ahan": from "sasajuṣo ruḥ" the [word] ru is carried forward, and from "skoḥ
saṃyogādyoḥ" the [word] "at the end (ante)" is carried forward. "padasya" is the
governing adhikāra. "ahan" is [a pratīka] with the genitive ending elided. He
states this with "ahann ity asya…": "ahobhyām" — rutva of the n, u-substitution
by "haśi ca," guṇa — this is the sense. Objection: in "ahaḥ, ahobhyām," since the
ratva/rutva are asiddha, n-elision would occur. Nor can one say the ratva/rutva
are exceptions to the n-elision; for [the sūtra] "na ṅi-saṃbuddhyoḥ" prohibiting
n-elision, in the locus "he ahar" and in "dīrghāṇy ahāni yasmin sa dīrghāhāḥ, he
dīrghāho nidāgha," is undertaken even where there is no n-elision — [so] the
ratva/rutva are [independently] undertaken. Anticipating this, he refutes it with
"iha…": "ekena" — by the first of the two occurrences [of] the sūtra "ahan,"
meaning "at the pada-end there should be only 'ahan,' not n-elision" — this is the
sense. "dvitīyena" — [the second] meaning "of ahan there is ru at the pada-end"
— this is the sense. And thus the vārttika "ahno nalopa-pratiṣedhaḥ" need not be
composed — this is the sense. [The source here repeats the same passage:] before
the vowel-initial endings beginning with ṭā, a-elision: ahnā, ahne, etc.; before
the consonant-initial endings beginning with bhyām he states the special case —
"ahan": from "sasajuṣo ruḥ" ru is carried forward, and from "skoḥ saṃyogādyoḥ"
"at the end"; "padasya" is the adhikāra; "ahan" is [a pratīka] with elided
genitive; he states this with "ahann ity asya…": "ahobhyām" — rutva of n,
u-substitution by "haśi ca," guṇa; objection [as above], refuted by "iha";
"ekena" — by the first "ahan" sūtra meaning "at pada-end only 'ahan,' not
n-elision"; "dvitīyena" — meaning "of ahan there is ru at pada-end"; thus the
vārttika "ahno nalopa-pratiṣedhaḥ" need not be composed. "padāntasyāpi" — the
ratva/rutva occur also of [a compound] ending in the word ahan, because [the
rule] stands in the pada-adhikāra — this is the sense. As for the prohibition
"grahaṇavatā prātipadikena tad-antavidhir nāsti," it applies only to affix-rules,
as is clear from the Bhāṣya and Kaiyaṭa on the sūtra "asamāse niṣkādibhyaḥ." He
shows the operation of rutva at the end of such [a compound] with "dīrghāhā
nidāgha": from the masculine word dīrghāhan, before su, due to paratva,
upadhā-lengthening being done, then su-elision by "halṅyādi…," rutva by "ahan";
then, since it is now non-original (apūrva), y-substitution by "bho-bhago…," and
y-elision by "hali sarveṣām" — the form [dīrghāhā] — this is the sense. Objection:
when su is elided, by ratva ["ro 'supi"] there is no y-substitution for it,
because in the y-substitution rule "roḥ" the u is taken as an anubandha; and so
"dīrghāhār nidāgha" alone would be right. To this he says "iha halṅyādi…"
[reply]. Objection: even though the n-final lengthening is para, by the paribhāṣā
"akṛtavyūha…" it cannot operate before the halṅyādi-elision, since the n is on the
verge of destruction by rutva. To this he says "tasyāsiddhatvāt": because the
operated rutva is asiddha, the n-final lengthening is unobstructed — this is the
sense. In truth, however, since the paribhāṣā "akṛtavyūha…" is baseless, here too,
before the halṅyādi-elision, upadhā-lengthening due to paratva is what is proper
— enough of this. "saṃbuddhau tu" — in the vocative, the su [being elided] and
relying on pratyaya-lakṣaṇa after halṅyādi-elision, because [the upadhā-lengthening
rule] "asaṃbuddhau" operates [only in the non-vocative], there is no
upadhā-lengthening; [so] rutva, u-substitution by "haśi ca," guṇa by "ād guṇaḥ,"
giving the form "he dīrghāho nidāgha" — this is the sense. Here, the fruit of the
ratva-rule "ro 'supi" operating also at the end of such [compounds] is to be
understood [in examples] like "dīrghāhar nidāghajālam" in the neuter. "daṇḍī" —
in the sense "he has a staff," by "ata iniṭhanau," the ini. From the word daṇḍin,
luk of su/am, n-elision — this is the sense. "daṇḍinī" — auṅ → śī; and no
lengthening, by the restriction "in-han…," since [auṅ/śī] is not a sarvanāmasthāna
— this is the sense. "daṇḍīni" — "jaśśasoḥ śiḥ," lengthening by "in-han" — this is
the sense. "sragvī" — by "asamāyām edhāsragjo viniḥ," the possessive vini after
the word srag. Because the entity "srag," relying on the internal vibhakti, is a
pada, kutva of its [final] j; from the word sragvin, sup is derived; the forms are
like daṇḍin. Here, although the ini is meaningless [i.e. redundant], it is indeed
included in "in-han," because of the statement "aninasman…" — this should be
understood. "vāggmī" — by "vāco gminiḥ" the gmini. Because it is a taddhita, the
g-kāra is not an it; jaśtva and kutva of the c [vāc → vāg]; from the word
vāggmin, sup is derived; the forms are like sragvin. In the bahuvrīhi "bahavo
vṛtrahaṇo yasmin manvantare" ("the age in which there are many Vṛtra-slayers"),
from the word bahuvṛtrahan, on the luk of su/am, n-elision, the form
"bahuvṛtraha." For auṅ, śī; a-elision; kutva by "ho hanteḥ" [h → gh]; the form
"bahuvṛtraghnī." No ṇatva, by the restriction "at-pūrvasya." Having treated this
much as established, he states jas/śas — "bahuvṛtrahāṇi": since śi is a
sarvanāmasthāna, there being no a-elision before it, upadhā-lengthening by
"in-han," and ṇatva by "ekāj uttarapade ṇaḥ" — this is the sense. "bahupūṣāṇi" —
"bahavaḥ pūṣaṇo yasmin," a bahuvrīhi: bahupūṣa, bahupūṣṇī; ṇatva by "raṣābhyām";
in jas/śas, śi, lengthening before śi, ṇatva by "aṭ-kupvāṅ…." "bahvaryamāṇi" —
"bahavo 'ryamaṇo yasmin," a bahuvrīhi. This he sums up in substance — "bahūrji":
in the word bahūrj, above the ū-kāra which is the final vowel, the prohibition of
num is to be stated; rather, num optionally [comes] before the final phoneme —
this is the sense. "bahūrji" — the form when num is absent in jas/śas. "bahūrñji"
— the form to be understood: num being made above the repha before the j-kāra,
since ścutva is asiddha, by "naś ca…" it → anusvāra, and by parasavarṇa → ñ-kāra.
Here the first vārttika "bahūrji pratiṣedhaḥ" need not be composed, because on the
interpretation of the sūtra "napuṃsakasya jhalacaḥ" as "num is for a neuter ending
in a jhal that comes after a vowel," num is not occasioned here at all. Nor should
under-application (avyāpti) be suspected in "vanāni" etc. on this reading,
because — carrying "aci" forward from "iko 'ci vibhaktau" — [there is] another
statement, "num is for a neuter ending in a vowel (ajanta) when a sarvanāmasthāna
follows" — so it stands in the Bhāṣya. And thus, in the word ūrk, before śi, the
[Kaumudī-]mūla "nara-jānā saṃyogaḥ" is to be disregarded as contrary to the
Bhāṣya, because there, there being no jhal after a vowel, num is not occasioned.
— End of the j-final.
Now the d-final [stems]. "tyat" — of tyad, tad, yad, etad, because [su/am] is
elided by luk, the tyadādi a-substitution, pararūpa, and the s-substitution by
"tado saḥ sau" do not occur. But elsewhere, everywhere, on the tyadādi
a-substitution and pararūpa, the forms are like an a-final [stem], and the
sarvanāman-operations [apply] — this should be understood. "anvādeśe tv enat" —
in an anaphoric reference (anvādeśa), in the neuter singular, that [enad] is
enjoined — this is the sense. In auṅ and jas/śas: ene, enāni. "bebhidyateḥ" — this
is a designation by [the root with] śtip. It means: from the yaṅ-final root
"bebhidya." "bhidir vidāraṇe" [the root bhid, to split]; from it, by "yaṅi
sanyaṅoḥ" doubling; halādiḥ śeṣaḥ; by "abhyāse carca…" the b of the reduplicant →
v by jaśtva; by "guṇo yaṅlukoḥ" guṇa; the form "bebhidya." From that, being a
dhātu by "sanādyantāḥ," kvip; a-elision (ato lopaḥ); y-elision by "yasya halaḥ."
From the word bebhid, luk of su/am, jaśtva-cartva, the form "bebhit" — this is the
sense. "bebhidī" — the form for auṅ → śī. In jas/śas, before śi, anticipating the
num characteristic of jhal-final stems, he says "śau…": "sthānivattvāt" — "by
'acaḥ parasmin'" is to be supplied. Nor can one say: here the d-kāra alone, not
the i-kāra, is prior to the vowel that is the sthānin of the a-elision, since it
[the i] is separated from it [the vowel] by the d-kāra, so the num-rule [would
fail as] not being the rule for what is prior to the sthānin-vowel — because in
"acaḥ parasmin" the grasping of even a non-contiguously prior [element] is stated.
As for "kvau luptaṃ na sthānivat," that is not universal, as is clear in Kaiyaṭa
on "dīdhī-vevī-ṭām." Objection: even if, by the sthānivadbhāva of the a-elision,
the num characteristic of jhal-final stems is absent, the num characteristic of
vowel-final (ajanta) stems is hard to avert. To this he says "ajanta-lakṣaṇas tu
num na": Why? To this he says "sva-vidhau sthānivattvābhāvāt" — relying on the
sthānivadbhāva of the a-elision, the num, being mit, operating as posterior to the
final vowel, ought to operate above the region marked by the a-kāra that is above
the d-kāra; and so, the num-rule being for the very [vowel] that is the sthānin of
the elision, in effecting it the sthānivadbhāva of the a-elision is impossible,
because "acaḥ parasmin" operates only for a rule concerning what is prior relative
to the sthānin. And the sthānivat-sūtra too does not operate for a rule concerning
the sthānin itself, because of the prohibition "analvidhau" — this is the sense.
The vārttika "in the sūtra 'acaḥ parasmin,' removing 'pūrvavidhau' and stating
'aparavidhau,' for the sake of sthānivadbhāva in a rule concerning oneself
(sva-vidhi)" is rejected in the Bhāṣya — so there is no fault. — End of the
d-final.
Now the c-final [stems]. [The verses posing an old riddle:] "Nine [forms] arise
in su, and nine likewise in am; six each in the gathering of bhyām, bhis, bhyas;
nine in sup; then three in jas, likewise in śas; four in the other
case-terminations. Whose [word's forms are these]? Let the wise ones know the
forms of that word, if they can; if [anyone] can declare them, the time-limit here
is six months." By these two verses he answers a certain ancient riddle —
"gavākchabdasya…": "arcā-gati-bhedataḥ" — by "aniditām" the n of añc is elided in
the sense of motion (gati); but in the sense of worship (pūjā), by "nāñceḥ
pūjāyām" the prohibition [means] the n is not elided — so [the difference arises]
from the difference of operation, namely n-elision and its absence, conditioned by
the difference of the meanings motion and worship. By "asandhi…" the prakṛti-bhāva
(non-application of sandhi) is intended. After "asandhy-avaṅ-pūrvarūpaiḥ" the word
"ca" is to be supplied; and after "śatam," the word "iti." And thus, by the
n-elision-and-its-absence conditioned by the difference of the senses motion and
worship, and by "asandhi (prakṛti-bhāva), avaṅ, and pūrvarūpa," the forms of the
word gavāc are one hundred and nine (109) — this is the accepted view. By this the
[alternative] is set aside: that, taking "śata" (hundred) as denoting a number per
the lexicon "viṃśaty-ādyāḥ… saṃkhyā-saṃkhyeyayoḥ," "the forms of gavāc" [in the
plural is required], while on taking it as denoting the numbered (saṃkhyeya) the
singular "matam" would be inconsistent. Because, taking [śata] as denoting number
and supplying the word "iti," the meaning "the view (matam) is that the forms are
one hundred and nine" is adopted. Or else, even taking śata as denoting the
numbered, the singular "matam," intending the oneness [of the view], is faultless.
He elaborates "nava-adhika-śatam" with "svam-supsu nava": "prati-ekam" (each) is
to be supplied; "rūpāṇi" construes throughout. "ṣaḍ bhādau" — six each in bhis,
the three [cases with] bhyām, and the two [cases with] bhyas — this is the sense.
"trīṇi jaśśasoḥ" — "each" is to be supplied. "catvāri śeṣe daśake" — "each" is to
be supplied [in the remaining ten endings]. He establishes "svam-supsu nava" with
"tathā hi": "gām añcati" — meaning "goes to a cow" or "worships a cow." With kvin,
in the upapada-compound, on the luk of the sup, "go añc" stands. He shows the
derivation — "gatau nalopaḥ": in the sense of motion, by "aniditām" the n is
elided — this is the sense. "gavāk-gavāg" — when "go añc s, go añc am" stands: luk
of su/am; because by the paryudāsa "aklībasya…" there is no sarvanāmasthāna-hood,
no num by "ugidacām"; avaṅ-substitution for the o-kāra; savarṇa-dīrgha;
jaśtva-cartva — this is the sense. "sarvatra" — because the avaṅ-substitution is
optional, on the alternative of its absence, prakṛti-bhāva by "sarvatra vibhāṣā,"
and there being also no pūrvarūpa by "eṅaḥ padāntād ati" — this is the sense [→
go-ak / go-ag]. "pūrvarūpe" — in the absence of avaṅ and prakṛti-bhāva, when there
is pūrvarūpa by "eṅaḥ padāntāt" — this is the sense [→ go'k / go'g]. Thus, in the
sense of motion, six (6) forms. "pūjāyām" — because of the absence of n-elision by
"nāñceḥ pūjāyām," from "go añc": on the luk of su/am, saṃyogānta-lopa of the
c-kāra, kutva of the n-kāra by "kvin-pratyayasya kuḥ" → ṅ-kāra; then avaṅ,
prakṛti-bhāva, and pūrvarūpa → three (3) forms [gavāṅ, go-aṅ, go'ṅ]. Adding the
previously cited six, nine (9) forms in su. "auṅaḥ śī" — in the sense of motion,
when n-elision and śī-substitution for auṅ are done, by the paryudāsa
"aklībasya…," being non-sarvanāmasthāna it is a bha, so by "acaḥ" a-elision,
[giving] "gocī" — one (1) form only, because the a being elided, avaṅ etc. are
impossible. "pūjāyām tu" — because the n is not elided, no a-elision by "acaḥ" —
this is the sense. And so in auṅ, three (3) forms [gavāñcī, go-añcī, go'ñcī].
Adding the previously cited one form, four (4) forms each. "śeḥ sarvanāma…" — in
the sense of motion, on n-elision, before śi, num by "ugidacām," its anusvāra by
"naś ca…," parasavarṇa — this is the sense. But in the sense of worship, because
the n is not elided, no num; rather, the anusvāra of the natural n, parasavarṇa —
this is the sense. He states that in śi the three forms are undifferentiated for
motion and worship — "trīṇy eva": "goci" — in the sense of motion, because the n
is elided, on a-elision by "acaḥ," one (1) form only — this is the sense; in the
sense of worship, because the n is not elided, no a-elision by "acaḥ," so avaṅ,
prakṛti-bhāva, pūrvarūpa — three (3) forms — he says "gavāñca-go-añca-go'ñca." In
bhyām: in the sense of motion, on n-elision, jaśtva and kutva of the c,
avaṅ/sandhi/pūrvarūpa — three (3) forms; in the sense of worship, because of no
n-elision, saṃyogānta-lopa of the c, ṅ-kāra by "kvin-pratyayasya…," kutva of the
n, avaṅ/sandhi/pūrvarūpa — three (3) forms — intending this he says "gavāgbhyām"
etc. "ityādi" — gavāgbhiḥ, go-agbhiḥ, go'gbhiḥ; gavāṅbhiḥ, go-aṅbhiḥ, go'ṅbhiḥ;
goce, gavāñce, go-añce, go'ñce. In bhyām and bhyas, as before, six each. ṅas —
gocaḥ-gavāñcaḥ-go-añcaḥ-go'ñcaḥ; gocoḥ-gavāñcoḥ-go-añcoḥ-go'ñcoḥ;
gocām-gavāñcām-go-añcām-go'ñcām; goci-gavāñci-go-añci-go'ñci. In os, as before.
"supi tu" — in the sense of worship, because of the prohibition of n-elision,
saṃyogānta-lopa of the c, ṅ-kāra by "kvin-pratyayasya…," kutva of the n; and then,
for the three ṅ-final forms made with avaṅ/sandhi/pūrvarūpa, there is the
kuk-augment — this is the sense. But in the absence of kuk, gavāṅṣu-go-aṅṣu-go'ṅṣu,
three (3) forms, not stated because easy. In the sense of motion, on n-elision,
ka-kāra by kutva of the c, he shows three (3) forms with avaṅ/sandhi/pūrvarūpa —
"gavākṣu" etc. And thus in sup, nine (9) forms. Objection: among these, in the six
forms having ka-kāra, by "cayo dvitīyāḥ," on the alternative of kha for ka, there
would be six (6) additional forms with kha. Anticipating, he refutes — "na
ceha…": "cartvasya…" — in the sense of motion, on n-elision, kutva of the c, that
[k] → g by jaśtva, that [g] → k by cartva ["khari ca"], the three forms
gavākṣu-go-akṣu-go'kṣu stand. There, on the theory of "cayo dvitīyāḥ," because the
cartva-rule is asiddha, there is no ka-kāra at all, but only a g-kāra; that being
not cay, "cayo dvitīyāḥ" does not apply — this is the sense. And thus the
[supposed] surplus of three in motion is set aside. But in worship, on the
kuk-alternative, in the three gavāṅkṣu-go-aṅkṣu-go'ṅkṣu, when the ka has its second
[substitute kha], the surplus of three with kha is indeed desired — he says
"kukpakṣe tu." Nor should contradiction with "nava-adhika-śatam" be suspected,
since "nava-adhika-śatam" is the view of the sūtra-author.
"ūhyam eṣām" — of the [forms] shown, which are one hundred and twelve (112) —
this is the sense. "aśvākṣi-bhūtāni" — meaning five hundred and twenty-seven
(527). For the word "aśva" denotes [the number] characterized by seven-ness, from
"saptāśvā haritaḥ sūryasya" (the sun's horses are seven). The word "akṣi" denotes
[the number] characterized by two-ness, since among humans etc. the eyes are
generally two. The word "bhūta" denotes [the number] characterized by five-ness,
since the bhūtas — earth, water, fire, air, ether — denoted by the word bhūta are
five. There, by the word "aśva" mentioned first, seven-ness alone is intended; by
the word "akṣi" mentioned second, the twenty (two tens) indicated by the two-ness
it signals is intended; by the word "bhūta" mentioned third, five hundred is
denoted by the five-ness it signals. And it is said in astronomy: "eka, daśa,
śata, sahasra, ayuta, lakṣa, prayuta, koṭi in order; from that arbuda, abja,
kharva, nikharva, mahāpadma, śaṅku; jaladhi, and antya; madhya, parārdha — these
are the names, increasing tenfold." Here one should write first the seven-ness
signaled by "aśva," then above it the two-ness signaled by "akṣi," then above that
the five-ness signaled by "bhūta," per the dictum "of digits [the reckoning]
proceeds from the right" — and so on, to be known from the works of the founders
of the reckoners' tradition such as the Līlāvatī. And thus "seven and twenty and
five hundred (527) forms arise," as reckoned by the wise — this is the sense. To
wit: in su, of the nine forms, on the doubling of the final phoneme by "aco 'nci
ca" and its absence, eighteen (18) forms. In the nom. dual, of the four, of the
three in the sense of worship, on the doubling of the m by "aco 'nci ca" and its
absence, six forms — but not "aṇo 'pragṛhyasya," because of its being pragṛhya —
and in motion only one. By addition, seven (7). In jas, of the three [in worship]
with doubling of the ñ and without, six forms; and of these six, of the final
i-kāra, on the nasal alternative by "aṇo 'pragṛhyasya…," six, and on the non-nasal
alternative the same six stand; by addition, in jas twelve (12). And thus in the
first (nominative) vibhakti, thirty-seven (37). Likewise in the second
(accusative) vibhakti, thirty-seven (37). In the third-case singular (ṭā), of the
four, of the three in the sense of worship, on the doubling of the ñ and without,
six forms; and in motion only one. By addition, fourteen (14). In bhyām, of the
six, in motion of the g and in worship of the ṅ, on doubling: six; without: the
same six stand; by addition, twelve (12). Of these, on the doubling of the y by
"yaṇo mayaḥ": twelve — "maya" being ablative, "yaṇa" genitive [i.e. the y of yaṇ
doubled before the m of maya] — by that reasoning; without doubling of y, the same
twelve stand; by addition, twenty-four (24). Of these, on doubling of the m:
twenty-four; without: the same twenty-four stand; by addition, in bhyām
forty-eight (48). In bhis, of the six, in motion of the g and in worship of the ḍ,
on doubling: six; without: the same six stand; by addition, twelve (12). Of these,
on doubling of the visarga: twelve — because the doubling of visarga is enjoined
for the ayogavāhas before aṭ and śar; without its doubling, the same twelve stand;
by addition, in bhis twenty-four (24). And thus in the third vibhakti, eighty-six
(86). In the fourth-case singular (ṅe), of the four, of the three in worship, on
doubling of the ñ and without, six (6); and in motion only one (1); by addition,
in ṅe seven (7). Because the e is non-aṇ, no nasal. In bhyām, as before,
forty-eight (48). In bhyas, of the six, in motion and worship, of the three each,
on doubling of g/ṅ: six (6); without: the same six (6) stand; by addition, twelve
(12). Of these, on doubling of y by "yaṇo mayaḥ": twelve; without: the same
twenty-four stand; by addition, in bhyas forty-eight (48). And thus in the fourth
vibhakti, one hundred and three (103). In ṅasi: of the three [worship] forms, on
doubling of ñ and without, six; and in motion one stands; by addition, seven (7).
Of these, on doubling of visarga: seven; without: the same seven stand; by
addition, in ṅasi fourteen (14). In bhyām and bhyas, as before, forty-eight each
(48). And thus in the fifth vibhakti, one hundred and ten (110). ṅas: like ṅasi,
fourteen (14). In os, of the four, of the three in worship, on doubling of ñ and
without, six; and in motion one; by addition, seven (7); of these, on doubling of
visarga and without, fourteen (14). In ām, of the four, of the three in worship,
on doubling of the ñ in the three, six; and in motion one; by addition, seven (7);
of these, on doubling of m and without, fourteen (14). And thus in the sixth
[vibhakti], forty-two (42). In ṅi, of the three in worship on doubling of ñ and
without, six; in motion one; by addition, seven (7); of these, of the final
i-kāra, on nasality and without, fourteen (14). In os, as before, fourteen. In
sup, of the twelve, in worship, on the kuk-absence alternative, on doubling of the
ḍ, three; without, the same three stand; by addition, twelve (12); of these
twelve, on doubling of ṣ by "khayaḥ śaraḥ," twelve, without the same twelve stand,
by addition twenty-four (24); but in the kuk-absence the six stand; by addition,
in worship thirty (30). In motion, of the three, on doubling of the k, three,
without the same three stand; of these six, on doubling of ṣ by "khayaḥ śaraḥ,"
six, without the same six stand; by addition, twelve (12). And thus by addition,
in motion and worship, forty-two (42); of these, on the nasality of the final and
without, in sup eighty-four (84). And thus: su eighteen (18), au seven (7), jas
twelve (12) — total in the nominative thirty-seven (37); [likewise the accusative,
thirty-seven (37)]; ṭā fourteen (14), bhyām forty-eight (48), bhis twenty-four
(24) — total in the third eighty-six (86); ṅe seven (7), bhyām forty-eight (48),
bhyas forty-eight (48) — total in the fourth one hundred and three (103); ṅasi
fourteen (14), bhyām forty-eight (48), bhyas forty-eight (48) — total in the fifth
one hundred and ten (110); ṅas fourteen (14), os fourteen (14), ām fourteen (14) —
total in the sixth forty-two (42); ṅi fourteen (14), os fourteen (14), sup
eighty-four (84) — total in the seventh one hundred and twelve (112). And thence,
by addition, five hundred and twenty-seven (527) forms. — Such is the derivation
of the word gavāc.
"tiryak" — "tiraḥ añcati" analysed, kvin by "ṛtvigādi." In motion, n-elision by
"aniditām." From "tiras-añc," sup is derived; luk of su/am; because of the absence
of pratyaya-lakṣaṇa and of sarvanāmasthāna-hood, no num. Because it is not a bha,
no a-elision by "acaḥ." tir-ādeśa for tiras; yaṇ;
because the kutva by "kvin-pratyayasya…" is asiddha, kutva of the c by "coḥ kuḥ,"
jaśtva-cartva — this is the sense [→ tiryak / tiryag]. "tiraścī" — when "tiras añc
au" stands, in auṅ → śī, being a bha, a-elision by "acaḥ"; because "a-lope…" is
stated [i.e. the tir-substitution is conditioned], there is no tir-ādeśa; the
s → ś by ścutva — this is the sense [→ tiraścī]. "tiryañci" — from "tiras añc," in
jas, "jaśśasoḥ śiḥ," being sarvanāmasthāna, num, anusvāra-parasavarṇa; tir-ādeśa
for tiras; yaṇ; because it is not a bha, no a-elision by "acaḥ" — this is the
sense. "tiryaṅ" — on luk of su/am, because it is not a bha and the n is not elided
[in the worship sense], no a-elision by "acaḥ"; tir-ādeśa; saṃyogānta-lopa of the
c; kutva of the n → ṅ-kāra — this is the sense. "tiryañcī" — the form in auṅ → śī;
because the n is not elided, no a-elision by "acaḥ"; tir[-ādeśa]. "tiryañci" — in
jas/śas, "jaśśasoḥ śiḥ." The rest like the masculine. — End of the c-final.
Now the t-final [stems]. "yakṛt" — a certain lump of flesh named yakṛt [the
liver], well-known to the ritualists. luk of su/am, jaśtva-cartva — this is the
sense. "yakṛtī" — auṅ → śī. "yakṛnti" — "jaśśasoḥ śiḥ," because it is jhal-final,
num, anusvāra-parasavarṇa — this is the sense. Before śas etc. he states the
special case — "paddan… iti vā yakan iti" [by "pad-dan-no-mās…," the substitute
yakan optionally]. "yakāni" — in śas, śi; the yakan-substitute; being
sarvanāmasthāna, the n-final [lengthening → yakāni]; in the absence of the n-final
yakan [i.e. when yakan is not taken], yakṛdbhyām and so on. "śakṛt" — the word
śakṛt, denoting excrement (viṣṭhā), like yakṛt. "dadat, dadatī" — this word dadac
[dadat], ending in the śatṛ affix, has been derived in the treatment of the
masculine. Its luk of su/am; auṅ → śī; but no num, because it is
non-sarvanāmasthāna and because of the prohibition "nābhyastāc chatuḥ."