A Constraint-based Morphological Analyzer for Concatenative and Non-concatenative Morphology

A Constraint-based Morphological Analyzer for Concatenative and Non-concatenative Morphology Farrah Cherry Fortes-Galvan1 and Rachel Edita Roxas2 1 BitMicro Networks International, Inc., Net Square Center, Bonifacio Global City, Taguig City, Philippines 2 Software Technology Department, 2401 Taft Avenue, Malate, Manila, Dela Salle University-Manila, Philippines 1000 {cherry_fortes@yahoo.com, roxasr@dlsu.edu.ph} Abstract. Morphological analysis in the current methods, such as finite-state and unification-based, are predominantly effective for handling concatenative morphology (e.g. prefixation and suffixation), although some of these techniques can also handle limited non-concatenative phenomena (e.g. infixation and partial and full-stem reduplication). A constraint-based method to perform morphological analysis that handles both concatenative and non-concatenative morphological phenomena is presented, based on the optimality theory framework and the two-level morphology rule representation. Although optimality theory has been proven effective in handling non-concatenative phenomena, it has been applied for the generation process, and in this study, it has been shown to be effective also in morphological analysis. The method was tested on 1,600 Tagalog verb forms (having 3 to 7 syllables) from 50 Tagalog roots which exhibit both concatenative and non-concatenative morphology. The resulting method is able to produce the correct underlying forms of the surface forms of 96% of the test data, having a 4% error, which is attributed to d-r alternation. Keywords: morphological analysis, morphological generation, optimality theory. 1 Introduction Morphology is the area of linguistics that deals with the internal structure of words – how they are systematically formed from smaller units. Computational morphology deals with the processing of words and word forms in both written (graphemic) and spoken (phonemic) form. A morpheme is the smallest meaningful unit of a language. It could be a word stem or affixes. A free morpheme can stand alone, while bound morpheme functions only as part of a word. Affixation is the process of adding a bound morpheme to a free morpheme attached in front (prefix) or at the end (suffix) of a stem. The combination of a prefix and a suffix is called a circumfix. The infix is an affix where the placement is defined in terms of some phonological conditions. Reduplication requires copying of some portion of the stem. It can be complete or partial (prefixal, infixal, or suffixal reduplications). Infixation and reduplication are non-concatenative morphological processes. Much of the work in Philippine linguistics focused on the Tagalog language [1]. Tagalog language exhibits a rather complex verbal system. A single verb may contain reduplicated syllables, prefixation or suffixation and at the same time infixation. Reduplication in Tagalog can be partial or full. There could also be a combination of such phenomena in one word especially seen in Tagalog verbs. For instance, the word pinanglilibang-libang has root word libang and prefix pang (with infix in), partial reduplication of syllable li and full word reduplication of libang. Research on computational morphology has been predominantly on concatenative morphology and on finite-state models of morphotactics [2]. Although attempts were made to handle non-concatenative phenomena, it has been on a limited capacity only [3]. Optimality Theory is a phonological approach that …

Content truncated to 500 words