Lakṣaṇā as Inference

Abstract

This paper questions a few assumptions of Gaṅgeśa Upādhyāya’s theory of ordinary verbal cognition (laukika-śābdabodha). The meaning relation (vṛtti) is of two kinds: śakti (which gives us the primary referent of a word) and lakṣaṇā (which yields the secondary referent). For Gaṅgeśa, the ground (bīja) of lakṣaṇā is a sort of inexplicability (anupapatti) pertaining to the composition (anvaya) of word-meanings. In this connection, one notices that the case of lakṣaṇā is quite similar to that of one variety of postulation, namely, śrutārtāthāpatti, where the subject hears only a part of a sentence and immediately grasps the words that are needed to render the sentential meaning complete. Unless he does that, sentential meaning, i.e., the composition (anvaya) of word-meanings shall suffer from the same inexplicability that characterizes instances of lakṣaṇā. In fact, in the ‘Śaktivāda’ section of Tattvacintāmaṇi, Gaṅgeśa himself draws a parallel between the cognition of sentential meaning in a śrutārthāpatti-like case and the cognition of sentential meaning in an instance of lakṣaṇā. However, Gaṇgesa himself treats Śrutārthāpatti as a piece of inferential cognition. If there is no fundamental difference between cases of śrutārthāpatti and cases of lakṣaṇā, then the cognition of sentential meaning in instances of lakṣaṇā must also be inferential in essence. In that case, we must admit, against Gaṅgeśa’s view, that such cognition of sentential meaning cannot be accommodated within the framework of verbal cognition (śābdabodha). Therefore, I conclude that some revision is needed in Gaṅgeśa’s theory of verbal cognition with respect to lakṣaṇā.