Praudhamanorama p149

p149

As for what some say—that the mention of āp is for clarity—that is refuted by this. In reality, let there be no mention of ṅī and āp at all. The established conclusion is that the origination of a taddhita affix occurs from that which ends in sup. And ṅī and āp operate prior to sup, because it is accepted that operations conditioned by self-sense and so on are sequential; this should be kept in mind.
Regarding "of su and ṅasi": the u in su is for the sake of differentiation in "arvaṇastrasau". For if as were said, the meaning would be: when beginning with an a-sakāra; and thus "vājamarvatsu paya usriyāsu" would not be established. The i in ṅasi is so that its comprehension does not occur in "yuṣmadasmabhyāṃ ṅaso'ś". If it is argued that by the continuation of "pañcamyāḥ" in "ekavacanasya ca", by residual elimination, the other ṅas will be comprehended here, then it is for the sake of "ṭāṅasiṅasām—". The letters j and ś of jas and śas are for the purpose of demarcating their domains in "jasaḥ śī" and "tasmācchaso naḥ". If it is asked: let either one just be devoid of an indicatory letter, no; in the absence of j in jas, there would be āv for au, and consequently the affix might even be assumed to end in v. Similarly, in the absence of ś in śas, auṭ might be assumed to have the indicatory letter ṅ. Therefore, to prevent ambiguity, making them so is proper. The ṭ in auṭ is for the sake of the abbreviatory designation suṭ. But the ṭ in ṭā is for differentiation in "ṭāṅasi". The ṅ of ṅe and the like is for differentiation in "gherṅiti". The p, however, is for the sake of the abbreviatory designation. And thus, although all are fruitful, the statement of the older grammarian that "they are for the sake of technical names" is altogether incongruous. Even if one interprets it tortuously as: "saṃjñā means proper cognition, [hence] they are for that purpose, that is to say, for the sake of differentiation," it has been disregarded precisely because of its forced nature.
On "supaḥ" (Pā. Sū. 1.4.103): he explains that the phrase "trīṇi trīṇi" continues from the rule "tiṅastrīṇi trīṇi" [1.4.101], and the entire rule "tānyekavacana..." [1.4.102] continues as well, with the words: "supastrīṇi trīṇi..." and so forth. "ekaśaḥ" means: one by one. The affix śas occurs after that which ends in a nominative case ending in the sense of repetition (vīpsā) by "saṅkhyaikavacanācca" [5.4.43]. As for what was said by the older grammarian, "ekaikaśaḥ"—that is incorrect, because since repetition is already expressed by śas itself, reduplication is inappropriate; for the established conclusion is that by the maxim "yenanāprāpte" [an exception bars that in whose absence it would not apply], śas is an exception to reduplication. As for what they say: that both are correct, because of the Bhāṣya's usage "ekaikaśo dadāti" under "sarvasya dve" [8.1.1], and because of the Sūtrakāra's direction "ekaśaḥ" in the rule "tānyeka..."; that is not so, because the cited Bhāṣya has a different scope. For, having introduced [the topic with], "Where repetition is not expressed by a taddhita affix, there reduplication takes place," the Bhāṣya cites the example: "As in: ekaikaśo dadāti." And Kaiyaṭa explained: "There is also [the affix] śas denoting a small quantity; when there is reduplication in the sense of repetition of that, the affix śas is applied by 'bahvalpārthāc chac kārakādanyatarasyām' [5.4.42]." And this meaning is not possible in the present context; so let that be for now.
"rāmaḥ": in the view that this name-word is non-derived, it receives the technical name nominal stem (prātipadika) by the rule "arthavad..." [1.2.45]; whereas in the derivation where it is analyzed as "they delight in him" (ramante'smin) by the suffix ghañ under the governing rule "karaṇādhikaraṇayośca" [3.3.117] by "halaśca" [3.3.121] as an exception to gha, [it receives the name nominal stem] by the rule "kṛttaddhita..." [1.2.46]—such is the distinction.