From All Possible Worlds to Small Worlds: A Story of How We Started and Where We Will Go Doing Semantics

Copyright 2012 by Kiyong Lee 26th Pacific Asia Conference on Language,Information and Computation pages 1–18 From All Possible Worlds to Small Worlds: A Story of How We Started and Where We Will Go Doing Semantics Kiyong Lee Korea University, Seoul ikiyong@gmail.com Abstract This is a short story of how we have evolved over the last 40 years, doing semantics. could partially overlap with a history ofIt PACLIC which is commemorating the 25th year or a quarter of a century of its founding. The story tells how we semanticists of natu- ral language moved from all possible worlds to small worlds, now living in and with a tiny mobile world. 1 Introduction The story goes back to the early 1970s with gener- ative semantics and the dawning of Montague se- mantics. The beginning was concerned with big open worlds, all possible worlds, for truth meant, in the eyes of philosophers, being true in all pos- sible worlds. And linguists inherited their notion of truth in constructing a formal theory of natural lan- guage semantics. In the 1980s, however, the focus of linguistic semantics changed from necessary or possible truth to something more contingent or in- formative, namely various sorts of information ob- tainable from small worlds, called situations. A new trend developed in the 1990s towards the computa- tional modeling of semantic theories, based on so- called real language data or large corpora such as BNC (the British National Corpus). This required various situations of language use to be constrained with an idealized set of conditions. Then around the turn of the second millennium, semanticists have followed a data-driven approach to the construction of their model-theoretic semantics, which requires a large amount of language resources or raw corpora tagged with a variety of information, both morpho- syntactic and semantic. As a result, some semanti- cists including myself have proposed doing seman- tics using annotated language resources, which was known as annotation-based semantics. My story will narrate how our colleagues have re- acted to all these changes. Not being an historian, however, the speaker dares not guarantee his view to be fair and objective. Instead, it will be very sub- jective and introspective. Hence, it will simply be head-driven without being data-driven. I, as an old member of the PACLIC community, justify this nar- rowly defined role of an invited speaker because I trust that other PACLIC founding members, Ben- jamin T’sou and Akira Ikeya, will balance whatever might be one-sided in my talk. 2 The 1970s: Truth and All Possible Worlds Every decade has its own exciting moments. To me, the 1970s must have been the most exciting decade in my life. In the summer of 1971, I went as a Ful- bright student to the University of Texas at Austin, hoping to specialize in machine translation and the theory of translation. When I got there, there was no trace of such a thing as machine translation with all the projects and the people gone away and with nothing left of …

Content truncated to 500 words