Finite State Morphology and Sindhi Noun Inflections

Finite State Morphology and Sindhi Noun Inflections Mutee U Rahman, Mohammad Iqbal Bhatti Department of Computer Science, Isra University, Hala Road, Hyderabad Sindh 71000, Pakistan muteeurahman@gmail.com, iqbalbhatti@isra.edu.pk Abstract. Sindhi is a morphologically rich language. Morphological construction include inflections and derivations. Sindhi morphology becomes more complex due to primary and secondary word types which are further divided into simple, complex and compound words. Sindhi nouns are marked by number gender and case. Finite state transducers (FSTs) quite reasonably represent the inflectional morphology of Sindhi nouns. The paper investigates Sindhi noun inflection rules and defines equivalent computational rules to be used by FSTs; corresponding FSTs are also given. Keywords. Sindhi, morphology, noun inflections, two-level morphology, finite state morphology. 1 Introduction Morphology deals with word formation rules in a language. Word structures of a language are defined by its morphological constructions. Morphology defines that how smaller meaning bearing units called morphemes are combined to make larger meaning bearing units of a language called words. Morphology also deals with word formation by variations in already existing words. The morphological changes are mostly done by suffix addition, subtraction and replacement phenomenon. In few words morphology can be defined as syntax of word formation. Models for computational analysis of morphology always remained challenge for computational linguists until early 1980’s when 4Ks* discovered the two level morphology (Kaplan, R. M. and M. Kay. 1981) the first general model for morphologically complex languages. This two level morphology represents a word at lexical level and surface level. Morphotactics or morpheme ordering model is used in between these two levels to incorporate morphological changes. These morphotactics are implemented as Finite State Transducers (FSTs). Sindhi is one of the major languages of Pakistan and is spoken by approximately 40 million people(Sindhi Language Authority. 2009). Sindhi is an example of morphologically complex language. Nouns are marked by number, gender and case. Two level morphology can be used to model Sindhi noun inflections. Subsequent sections discuss Finite State Morphology, Sindhi morphological constructions, Sindhi noun inflections and role of finite state transducers in their computation. Section 2 discusses Sindhi noun inflections in detail. In section 3 finite state transducers for different noun inflection types in Sindhi are presented. Conclusions are discussed in section 4 and finally references are given in section 5. IPA Sindhi transliteration along-with Arabic script is also given. 1.1 Finite State Morphology Finite state transducers play an important role in language processing applications (Beesley, * Ronald M. Kaplin, Martin Kay, Lauri Karttunen and Kimmo Koskenniemi PACLIC 24 Proceedings 669 Kenneth R. and Lauri Karttunen, 2003) and computational studies of morphologically complex languages. Morphotactics (morpheme ordering rules) and orthographical rules (spelling rules) are represented by finite state transducers. Efficient morphological parsers can be implemented by combining these finite automata and computational lexicon (repository of words). Finite state transducers convert/translate lexical level constructs to surface level words by applying morphotactics and orthographical rules. Their reversible nature makes reverse conversion/translation possible. This two level (lexical and surface) morphology plays crucial role in implementation of …

Content truncated to 500 words