A finite-state morphological processor for Spanish

Abstract

A finite transducer that processes Spanish inflectional amt derivational morphology is presented. The system handles both generation astd analysis of tens of millions inflected ibrms. Lexical and surface (orthographic) representations of the words are linked by a program that interprets a finite directed graph whose arcs are labelled by n-tuples of strings. Each of about 55,000 base forms requires at le~t one arc in the graph. Representing the inflectional and derivational possibilities for these forms imposed an overhead of only about 3000 additional arcs, of which about 2500 represent (phonologicallypredictable) stem allomorphy, so that we pay a storage price of about 5% for compiling these form~ offline. A simple interpreter for the resulting automaton processes several hundred words per second on a Sun4.

Tagging French without Lexical Probabilities — Combining Linguistic Knowledge and Statistical Learning

“…Finite-state transducers (FST) are used to achieve morphological analysis. The FST is built on the model developed for Spanish morphology (Tzoukermann and Liberman, 1990) and handles mainly inflectional morphology with some derivational affixes, such as "anti-" in "anti-iranien" (anti-iranien), and "arrière-" (great) in "arrière-grand-père" (great-grandfather). The arclist dictionary -dictionary of finite-state transitions -was originally built using several sources, including the Robert Encyclopedic dictionary (Duval et al, 1992) and lexical information from unrestricted texts.…”

Multilingual text analysis for text-to-speech synthesis

“…expressions into finite-state lexicons; a tool (paradigm) to construct inflected morphological forms out of inflectional paradigm descriptions and lexicons that mark the paradigm affiliation of stems; a tool for constructing finite-state word grammars (arclist -the name being inspired by [18]); and a rewrite rule compiler [5], based on the algorithm described in [10]. The tool numbuilder constructs the transducer L 1 numbers , which converts strings of digits up to a user-defined length into numbernames appropriate for that string.…”

Reducing Information Variation in Text

“…Alternatively, the inflectional and morphological analysis can be based on a dictionary, in which the construction of inflectional and/or derivational forms is explicitly described for each root, like in the word-based approach [20], in the concatenative approach [126], or in the DELAS system [27]. An important dictionary-based approach to computational morphology is the two-level approach [81], which has been applied to numerous languages and stands out as particularly well adapted to morphologically complex, e.g.…”