ルドラタの詩論書に見られる〈直喩〉の詩的欠陥について

In Sanskrit poetics, theorists lay down the rule that a simile expressed in the form of a sentence (vākyopamā) must be constructed in such a way that the subject of comparison can grammatically agree with the object of comparison. However, it should be noted that in poetry, there are many examples of vākyopamā not conforming to this rule, which makes it necessary for theorists to modify the rule. Focusing on the poetic treatise Kāvyālaṃkāra, written by the theorist Rudraṭa (ca. ninth-century CE), this paper aims to answer the question of what grammatical requirements are necessary to construct vākyopamā.
According to the classification in Kāvyālaṃkāra 8.4–16, vākyopamā has six subtypes: (a) those in which the subject and object of comparison, the common property between them, and an indicator of comparison are explicit; (b) those in which the subject and object of comparison and an indicator of simile are explicit; (c) those in which the subject and object of comparison are interchangeable (ubhayopamā); (d) those in which the subject of comparison is identical to the object of comparison (ananvayopamā); (e) those in which the common property between the subject and object of comparison is inferred from the modifiers that modify the subject and object of comparison (kalpitopamā); (f) those in which the subject of comparison is compared to the unreal object of comparison, assumed to be real in a subjunctive clause (utpādyopamā). In the verse used as an example of subtype (f) Rudrat.a compares a piece of clothing that covers breasts (stanāvaraṇam) to a skin (tvak). It is worth noting that the former is neuter in gender, whereas the latter is feminine.
In Kāvyālaṃkāra 11.25–26, Rudrat.a states that in vākyopamā, the subject of comparison must agree with the object of comparison in gender (liṅga), number (vacana), tense (kāla), participant of an action (kāraka), and nominal or verbal ending (vibhakti), so that the modifier denoting the common property between the subject and object of comparison can modify both without changing its grammatical form. Observation of this fact leads us to the following conclusion: Rudraṭa allows a poet to construct vākyopamā in which the subject of comparison does not grammatically agree with the object of comparison, as long as the modifier denoting the common property is implicit. Therefore, it is reasonable to say that the vākyopamā adduced as an example of subtype (f) is not regarded as grammatically defective, because the theorist does not use any modifiers denoting the common property between the subject and object of comparison.