講演者：Malcolm Keating (Yale-NUS College, Singapore)
Title: Is Ellipsis Completion Knowledge? Linguistic Interpretation in Classical Indian Philosophy
Natural languages vary in how much information they encode into lexemes. Yet speakers can utter subsentential units which are syntactically or otherwise incomplete and still communi- cate successfully. Linguists and philosophers, in analyzing this widespread interpretive prac- tice of completing ellipsis, differ over whether such utterances constitute genuine speech acts, are disguised but complete syntactic/semantic units, as well as how the ellipsis is completed– syntactically, semantically, or pragmatically. The answers to these questions are significant since, for instance, they may challenge the thesis that languages are compositional, that is, with expressions being semantically determined by their syntax and lexical semantics.
Classical Indian philosophers, although committed to the compositionality thesis, gave vary- ing accounts of how interpretive practices allowed for ellipsis completion. The philosophers known as the Bhatta Mimamsa argued that an interpretive process, which they called arthapatti or “postulation,” could yield certain knowledge of what is elided. For instance, since the San- skrit language is highly inflected, someone who hears a speaker say “the door, the door!” can rely on syntactically-encoded information to help them recover a complete sentence, “Close the door, close the door!” In the 16th century, Narayana Bhatta discusses this process in the Manameyodaya, arguing that postulation requires the positing of words in order for there to be anvaya or “connection” within the expression. This argument is posed in response to opponents who argue that only the word meanings, and not the words themselves, must be posited.
I then draw connections between Narayanabhatta and contemporary Anglophone literature on the topic. In particular, I argue that the position of Narayana’s opponent (who is identified as belonging to another school of Mimamsa, the Prabhakara) is roughly analogous to that of pragmatic contextualists. In contrast, the Bhatta view could fruitfully be reconstructed as an abductive completion of lexical underspecification, along the lines of James Pustejovsky’s pro- posal. However, due to the ambiguity in the notion of connection, these reconstructions must be tentative, as Indian proposals maybe consistent with multiple formal analyses. The cru- cial implication to draw from their dialectic is the claim that ellipsis completion rises to the level of knowledge, and that it does so through a rational process grounded in the principle of compositionality.