Proceedings of the IndoWordNet Workshop 2012, Indian Institute of Technology, Kharagpur, India, 8 th December 2010, pp. 01-07 Polysemy and Homonymy: A Conceptual Labyrinth Niladri Sekhar Dash Linguistic Research Unit Indian Statistical Institute, Kolkata Email: ns_dash@yahoo.com Abstract Multifunctionality of a language generates multi- semanticity of words. If different senses of word are systematically related, then how these senses are derived from each other and how they should be organised to reflect their regularity in sense denotation? Before this question is addressed, in this paper an attempt is made to identify salient traits of distinctions between the polysemous and the homonymous words in a language, which, if addressed properly, will lead towards formation of methods to overcome the problems of word sense disambiguation. 1 Introduction The study of polysemy of a language has often been associated with the study of homonymy because distinction between the two has often not been very clear. In a piece of text, one can come across a set of words, which may appear either homonymous or polysemous. Since both types of word are often similar in surface representation (i.e., spelling and orthography) with no special mark for their distinction, one is easily misled to assume homonyms as polysemes or vice versa. However, there is a need to draw a clear line of distinction between the two, because these forms differ from each other not only in their nature, but also in function and implication. In this paper an attempt is made to identify the clues and strategies that can be adopted for tracing the differences between the two types of words. Since there is no well defined process for doing this, one has to use traditional knowledge from linguistics, semantics and cognitive science. Making perceptible distinction between the two types of words is a prerequisite for developing tools, systems and resources for natural language processing, language engineering, word sense disambiguation, machine translation, information retrieval, machine learning, cognitive linguistics, and applied linguistics. After identifying the theoretical and practical relevance of the phenomenon in various domains of human knowledge (Section 2), effort is made to understand polysemy (Section 3), identify the factors behind sense variations (Section 4), to explore the nature of homonymy (Section 5), and their conceptual relational interface (Section 6), the lack of which may cripple an investigator in the task of word sense disambiguation. 2 Relevance of the Phenomenon In recent years, the multi-semanticity of words in a natural language has been discussed at length in semantics, applied linguistics, psychology, philosophy, literature, and artificial intelligence. Following this trend, the study of polysemy and homonymy has arrested considerable attention in lexical semantics (Ullmann, 1962; Cruse, 1986; Palmer, 1995; Nida, 1997), cognitive linguistics (Cuyckens and Zawada, 2001; Deane, 1988), computational linguistics (Schütze, 1998; Ravin and Leacock 2000), discourse analysis (Leech, 1974; Kreidler, 1998), language teaching (Todd, 1987), psycholinguistics (Pinker, 1995; Gibbs et al., 1994), stylistics (Lyons, 1963; Allan, 2001), language and literature (Firth, 1957; Yule, 1985; Cruse, 2000) and many other fields. Information obtained from analysis of various multi-semantic lexical units has made remarkable contribution in understanding nature, and process of language cognition and acquisition, designing tools and systems for language processing, and developing strategies for language teaching. In this light, urgency arises for investigating the interface between polysemy and homonymy with an expectation that information obtained from this study will help in gaining insight about the phenomenon to overcome the hurdles of sense disambiguation of words.