Here, however, arises a discussion [as to the true application of the alleged rule of Páṇini], for we hold that, by ii. 3, 66, wherever an object and an agent are both expressed in one and the same sentence in connection with a word ending with a _kṛit_ affix, there the object alone can be put in the genitive and not the agent;[309] this limitation arising from our taking _ubhayaprápti_ in the sútra as a _bahuvríhi_ compound.[310] Thus we must say, "Wonderful is the milking of cows by an unpractised cowherd." We may, however, remark in passing that some authors do maintain that the agent may in such cases be put in the genitive (as well as the object); hence we find it stated in the Káśiká Commentary: "Some authors maintain that there should be an option in such cases without any distinction, and thus they would equally allow such a construction as 'the exposition of words _of_ the teacher' or '_by_ the teacher.'" Inasmuch, however, as the words of the phrase in question really mean that the "exposition" intended relates to _words_ and not to _things_, and since this can be at once understood without any mention of the agent, _i.e._, the teacher, any such mention would be plainly superfluous; and therefore as the object and the agent are _not_ both expressed in one and the same sentence, this is not an instance of the genitive of the object (coming under ii. 3, 66, and ii. 2, 14), but rather an instance of quite another rule, viz., ii. 3, 65, which directs that an agent or an object, in connection with a word ending with a kṛit affix, is to be put in the genitive [which in this instance is expressed by the _tatpurusha_ compound]; and the compound in question will be strictly analogous to such recognised forms as _idhma-pravraśchana_, _paláśa-śátana_, &c.[311] Or we might argue that the genitive case implied in this _shashṭhítatpurusha_ is one of the class called "residual," in accordance with Páṇini's rule (ii. 3, 50), "Let the genitive be used in the residuum," [_i.e._, in the other constructions not provided for by special rules];[312] and in this way we might defend the phrase against the opponent's attack. "But," it might be replied, "your alleged 'residual genitive' could be assumed everywhere, and we should thus find all the prohibitions of composition in constructions with a genitive case rendered utterly nugatory." This we readily grant, and hence Bhartṛihari in his _Vákyapadíya_ has shown that these rules are mainly useful where the question relates to the _accent_.[313] To this effect are the words of the great doctor Vardhamána-- "In secular utterances men may proceed as they will, "But in Vaidic paths let minute accuracy of speech be employed. "Thus have they explained the meaning of Páṇini's sútras, since "He himself uses such phrases as _janikartuḥ_ and _tatprayojakaḥ_."[314]