Praudhamanorama p28

p28

On the words "by the terms *guṇa* and *vṛddhi*": And this is obtained by continuing the words *guṇa* and *vṛddhi* from the preceding two rules with the construe: "those which are *guṇa* and *vṛddhi* [as denoted by] *guṇaḥ* and *vṛddhiḥ*". Therefore, it does not apply here: "diva aut" (Pā. sū. 7.1.84), *dyauḥ*; "tyadādīnāmaḥ" (Pā. sū. 7.2.102), *saḥ*, *imam*. On "ending in the genitive case": The intent is that because an item ending in the genitive is imitated in the rule, it denotes its own linguistic form, and so the case ending *su* is elided after a neuter. On "word": And that [word *ikaḥ*], when grammatical agreement is possible, is connected in the manner of "of a base ending in *ik*", etc. But in "miderguṇaḥ" (Pā. sū. 7.3.82), "mṛjervṛddhiḥ" (Pā. sū. 7.2.114), etc., since grammatical agreement is impossible, it is connected as "of an *ik* that is a part of *mid-* and *mṛj-*". As for what the older grammarian explained, namely that "they occur in place of *ik* alone", and what the commentator on his work said, namely that "this [rule] is for the sake of restriction when there is a contingency of non-restriction", etc.—all that is to be disregarded because it contradicts the Mahābhāṣya. For they say that the Mahābhāṣya refutes the view that it is an adjunct to that [genitive ending], and supports only the view that a word is present. However, the refutation of the objections even under the view that it is an adjunct to that [genitive ending] has been clarified by us in the *Śabdakaustubha*.
midaco'ntyātparaḥ (Pā. sū. 1.1.47)