Linearization of Nonlinear Lexical Representations.
George Kiraz · 1997
This paper presents a new schema for handling nonlinear morphology. The schema argues for linearizing nonlinear representations before applying phonological and morphological rules. 1 Introduction and Problem Statement Languages which exhibit templatic morphology have been lately treated using multi-tape finite state transducers, with one tape representing surface forms and the remaining tapes representing lexical forms (see (Kay, 1987; Kiraz, Forthcoming)). There are a number of advantages for using this multi-tape model. Not only does it accurately represent the linguistic insights behind the templatic nonlinear nature of these languages, it also allows the computational linguist to compile efficient, relatively small morphological lexica as opposed to lexica containing millions of entries. However, maintaining a nonlinear lexical representation has its own inconveniences and computational complexities. Firstly, the writer of multi-tape rules must keep track of multiple representat...