na tisṛcatasṛ — When "tisṛ + nām" stands and the lengthening (dīrgha) by
"nāmi" would apply, [Pāṇini states] "na tisṛcatasṛ." "tisṛcatasṛ" is a word with
an elided genitive ending (lupta-ṣaṣṭhīka, i.e. "of tisṛ and catasṛ"). From
"ḍhralope" the word "dīrgha" is carried forward (anuvṛtti), and the sūtra "nāmi"
is [also] carried forward. This he states with "tisṛ…" etc.
"tisṛṇām" — the retroflexion (ṇatva, n → ṇ) is by [the vārttika] "ṛvarṇān nasya."
Objection: How can the r-substitution (ratva) by "aci ra" be an exception
(apavāda) to the u-substitution (uttva) by "ṛta ut," given that uttva applies
only in the ṅasi/ṅas [abl./gen. sg.], and that for the words tri and catur —
being always plural-ending [as numerals] — the ṅasi/ṅas do not occur? Response:
No; because [uttva, hence ratva as its exception] does obtain in a bahuvrīhi such
as "priyatisṛ" [which admits the singular]. And this will be stated just below
(anupadam).
Objection: In the bahuvrīhi "priyatri" (analyzed "priyās trayas trīṇi vā yasyāḥ
sā"), why should the tisṛ-substitution (tisrādeśa) not [also] apply? In
anticipation of this he [Bhaṭṭoji] says "striyām…" etc. For the word "priyatri"
[in this analysis] is feminine, but the word "tri" [within it] is not [the locus
that the rule requires].
"striyām" — this qualifies tri and catur [themselves], not [stems] ending in them
(tad-anta), since there is no authority (pramāṇa) [for extending it to tad-anta].
Nor should it be said that the tad-anta [sense] is obtained through [the
substitution's being] an aṅga-operation; for even so, since only tri and catur
are directly heard (pratyakṣa-śruta) [in the rule], it is fitting that "striyām"
qualify precisely those. This is the purport.
Considering the optional nadī-designation (nadītva-vikalpa) by "ṅiti hrasvaś ca,"
he says "like the word mati." "But in the ām" — in the genitive plural, since the
traya-substitution (trayādeśa) by "tres trayaḥ" [tri → traya] is an aṅga-operation,
it operates even in a [stem] ending in [tri] (tad-anta) [→ priyatrayāṇām]. This is
the purport. And thus, if "striyām" qualified an aṅga ending in tri/catur, there
would be over-application (ativyāpti) in a bahuvrīhi "priyatri" containing a
masculine or neuter "tri"; therefore it is established that "striyām" qualifies
only tri and catur.
Now, considering that "striyām" qualifies only tri/catur (not tad-anta) also in
order to remove non-application (avyāpti) in a bahuvrīhi containing a feminine
"tri," he says "priyās tisro…" etc.
"priyatisā" — In the compound, because the internal (antarvarti) vibhakti is
elided by luk, the tisṛ-substitution is set aside [at the compounding stage];
[then] a su is [added] to the word "priyatri." Here, although the word "priyatri"
is masculine, because the word "tri" [in it] is feminine, the tisṛ-substitution
[applies]. By "ṛd-uśanas[-puru-daṃśo 'nehasāṃ ca]" the anaṅ [ṛ → an]; by
"sarvanāmasthāne ca" the lengthening (dīrgha, an → ān); [then] na-lopa
[→ priyatisā]. But if "striyām" qualified an aṅga ending in tri/catur, there would
here be non-application (avyāpti). This is the purport.
Objection: Even if [striyām] qualifies an aṅga ending in tri/catur, there is no
avyāpti here, because the tisṛ-substitution, having operated in the analytic
sentence (vigraha-vākya) "priyās tisro yasya," can carry over (anuvṛtti) into the
compound as well. Response: Not so. For the ordinary-language sentence
(laukika-vākya), being [already] complete (pariniṣṭhita), is not the basis
(prakṛti) of the compound; rather, only the technical process-sentence (alaukika
prakriyā-vākya) [is]. And therefore, in the technical sentence "priyā as tri as,"
when the compound operation applies, by the paribhāṣā "antaraṅgān api vidhīn
bahiraṅgo lug bādhate" (a bahiraṅga luk sets aside even antaraṅga operations), the
tisṛ-substitution is set aside [at that stage]; then, the vibhakti-luk having
occurred, when a su is introduced upon the compound word "priyatri" — since
"tricaturoḥ" is aṅga-conditioned and so applies even in the tad-anta operation —
by the paribhāṣā "nirdiśyamānasya ādeśā bhavanti" (a substitute replaces [only]
what is [directly] specified), the tisṛ-substitution [applies] to the word "tri"
[within it]. And that [substitution] would not occur if "striyām" qualified an
aṅga ending in tri/catur, because the aṅga "priyatri" is masculine.
And when the tisṛ-substitution has occurred here, the kap [samāsānta affix] by
"nadyṛtaś ca" does NOT [apply]. For that [kap], being a samāsānta, deserves to
operate — like the compound itself — in the technical (alaukika) analytic
sentence; and at that stage, since the final member (uttarapada) "tri" does not
end in ṛ, there is no kap. Moreover, because the internal vibhakti is elided by
luk, there being no pratyaya-lakṣaṇa, the tisṛ-substitution occasioned by jas
[nom. pl.] is [also] absent [there]. Nor should it be suspected that, the
samāsānta kap not yet having been made, when a su is introduced upon "priyatri"
and the tisṛ-substitution occurs, the kap [might then apply]; for, the samāsānta
kap not being made, the very compound-hood (samāsatva) would be unaccomplished,
and so the introduction of a su thereupon would be impossible. So let this rest.
"priyatisrau" [nom. du.] — setting aside guṇa, the r-substitution (ratva)
[tisṛ + au → tisr-au]. "priyatisraḥ" [nom. pl., jas] — setting aside the
pūrva-savarṇa-dīrgha, the ratva. "priyatisram" [acc. sg., am] — setting aside both
the pūrva-rūpa (single substitution as the prior form) and guṇa, the ratva. [The
maxim] "[ratva is] an exception to guṇa, dīrgha, and uttva" is [meant] also as an
indication (upalakṣaṇa) that it is an exception to pūrva-rūpa. "and so forth" —
priyatisrau [acc. du.], priyatisraḥ [acc. pl.], priyatisrā [instr. sg.],
priyatisroḥ [gen./loc. du.]. In the ṅasi/ṅas [abl./gen. sg.]: priyatisr-[aḥ]
itself — setting aside the uttva ("ṛ → ut") by "ṛta ut," [there is] ratva.
priyatisroḥ [gen./loc. du.]. In the ām [gen. pl.], setting aside the trayādeśa,
since [tisṛ] is subsequent (paratvāt) the tisṛ-substitution takes effect; then,
setting aside ratva, the nuṭ [augment applies] by "numaciratṛjvadbhāvaguṇebhyo
nuṭ" (by pūrva-vipratiṣedha) [→ nāmi lengthening → ṇatva → priyatisṛṇām]. priyatisṛṇām
[gen. pl.]. priyatisr-[i, loc. sg.] — by "ṛto ṅi" the ratva as an exception to
guṇa. priyatisroḥ [loc. du.].
Objection: How [do we get] "priyatri" [neuter] for "that family (kula) for which
there are three (feminine) things (priyās tisro yasya tat)"? Since, the word "tri"
being feminine, the tisṛ-substitution (tisṛ-bhāva) would apply, [in anticipation]
he says "sv-am-or luk." Meaning: because su and am are elided by luk by
"sv-am-or napuṃsakāt," and, by "na lumatā[ṅgasya]," there being no
pratyaya-lakṣaṇa, there is no [following] vibhakti [to condition it], the
tisṛ-substitution does not occur. "Because of non-obligatoriness (anityatva)" —
that "na lumatā[ṅgasya]" is anitya will be stated in the neuter section
(napuṃsaka-adhikāra), [as inferred] from the mention of "ac" in "iko 'ci
vibhaktau." Anticipating that, before a vowel-initial vibhakti, the r-substitution
by "aci ra ṛtaḥ" — being subsequent (para) relative to the num by "napuṃsakasya
jhalacaḥ" — [might apply], he says "[rādeśāt]…," [which] is a fifth-case with
elision of lyap (lyab-lopa-pañcamī, i.e. "having set aside the r-substitution").
The meaning is: by prior-superordination (pūrva-vipratiṣedha), setting aside the
ratva, [there is] num.
"priyatisṛṇī" [neut. du.] — setting aside ratva, upon the num [there is] ṇatva.
"priyatisṝṇi" [neut. pl.] — by "jaś-śasoḥ śiḥ" [śi replaces jas/śas in the
neuter]; setting aside ratva, the num. Because śi is a sarvanāmasthāna, [there is]
the lengthening characteristic of an n-ending stem (nānta-lakṣaṇa dīrgha) [and]
ṇatva. It should be understood that, on the strength of the examples
"priyatisṛṇī, priyatisṝṇi" [given] in the Bhāṣya on the sūtra "pratyayottarapadayoś
ca," the ratva is set aside by the num by resorting to pūrva-vipratiṣedha.
"priyatisṛṇā" [instr. sg.] — in the ṭā, on the alternative of there being no
puṃvadbhāva, the form [is] with num. "priyatisrā" — on the [alternative of]
puṃvadbhāva, since there is no num, [there is] ratva. "and so on" — by "ādi," forms
such as "priyatisre, priyatisṛṇe" [dat. sg., the two alternatives] etc. should be
understood.
"dver atve" — meaning: from the word "dvi," when a vibhakti follows, upon the
a-substitution belonging to the tyad-ādi class (tyad-ādy-atva, dvi → dva), the
ṭāp by "ajādy-ataḥ" [→ dvā]. "dve" etc. — when the ṭāp is present, upon the
savarṇa-dīrgha, the forms of the word "dvā" are like [those of] "ramā" [dvā + au
→ dve]. This is the purport. — Thus [end] the words ending in [short] i.
Now the words ending in [long] ī. "gaurī" — from the word "gaura," by the ṅīṣ
characterized by [the gaṇa] "gaura-ādi," with elision of the a-kāra by "yasyeti
ca," [comes] the word "gaurī." From that [comes] su; [then] the elision by
"hal-ṅy-ādi[-lopaḥ]" [of su → gaurī]. This is the purport. "gauryau" — the
pūrva-savarṇa-dīrgha being prohibited by "dīrghāj jasi ca," [there is] yaṇ [ī → y
before the vowel]. This is the purport. "gauryaḥ" — the pūrva-savarṇa-dīrgha being
prohibited by "dīrghāj jasi ca," [there is] yaṇ. This is the purport.
"nadī-kārya" (the nadī-operations) — meaning: the shortening by "ambārtha-nadyor
hrasvaḥ" [voc. sg.], the "āṇ nadyāḥ" [the āṭ augment of the ṅit endings], the nuṭ
by "hrasvanadyāpo nuṭ" [gen. pl.], [and] the ām-substitute enjoined by "ṅer ām
nadyām nī bhyaḥ" [loc. sg.]. The nadī-designation is by "yū stryākhyau [nadī]";
[gaurī declines] like "bahuśreyasī." "Likewise vāṇī, nadī, etc." — [vāṇī:] from
"vaṇa" (in the sense of sound), "that which is sounded (vaṇyate = śabdyate)" is
"vāṇī"; the affix iñ by "iñ vapādibhyaḥ," and the ṅīṣ by "kṛd-ikārād aktinaḥ."
[nadī:] because [it is derived] from [a base] read in the pacādi [gaṇa] with the
affix "nadaṭ," being ṭit [it takes] ṅīp. By "ādi" [etc.], words such as "kartrī,
daṇḍinī" are included.
[sakhī:] by "sakhy-aśiśvī iti bhāṣāyām," from the word "sakhi," with the ṅīṣ and
elision of the i-kāra by "yasyeti ca," [comes] the word "sakhī." For it,
anticipating the anaṅ by "anaṅ sau" and the ṇid-vadbhāva by "sakhyur asambuddhau,"
he says "prātipadika…" — [the vārttika] "vibhaktau liṅga-viśiṣṭāgrahaṇam." This
paribhāṣā is stated in the Bhāṣya on "yuvor anākau" and on "ṅy-āp-prātipadikāt."
Its meaning: when an operation occasioned by a vibhakti is to be performed, in the
mention of a prātipadika there is no inclusion of the gender-qualified [form]. And
thus neither the anaṅ nor the ṇid-vadbhāva occurs. This is the purport. "sakhī" —
because it ends in ṅī (ṅy-anta), [there is] elision of su. "sakhyai" etc. —
meaning: the forms are just like [those of] gaurī.
By "lakṣer muṭ ca," from the root "lakṣ," upon the ī-affix with the muṭ augment,
[comes] the word "lakṣmī." Its special feature he states with "[an]aṅyantatvāt"
[because it does not end in ṅī, there is no su-lopa]; but with the ṅīṣ [added] by
"kṛd-ikārād aktinaḥ" [in an alternative derivation], su-lopa does indeed occur. The
rest is like gaurī.
"strī" — from "styai [in the sense of] sound and conglomeration (śabda-saṃghāta)."
"she in whom the semen and blood (śukra-śoṇita) come together (styāyataḥ = saṅgate
bhavataḥ)" is "strī." From "styai" [the affix] "ḍraṭ"; the ḍa and ṭa are its
it-markers (anubandhas). By the force of its being ṅit (through the ḍ), there is
elision of the ṭi (final portion) even for a non-bha [stem]; by "lopo vyoḥ [vali]"
the elision of y (ya-lopa); because [ḍraṭ] is ṭit, [there is] ṅīp; [and in the
nominative singular] the elision by "hal-ṅy-ādi[-lopaḥ]" [→ strī]. This is the
purport. "he stri" [voc. sg.] — the shortening (hrasva) by "ambārtha[-nadyor
hrasvaḥ]." When "strī + au" stands, because [strī] is not a modification of a
dhātu (adhātvikāratva), the iyaṅ by "aci śnu-dhātu-[bhruvāṃ yv-oḥ iyaṅ-uvaṅau]"
not applying —