Modeling case marking system of Urdu-Hindi languages by using semantic information

Rule based modeling of the grammar of a language is important for various natural language processing chores. In this paper, formulation of Urdu/Hindi case marking system based on lexical functional grammar (LFG) is presented. In this paper, semantic information associated with nouns is incorporated to better classify grammatical roles adopted by each case. Especially, the versatile case marker ´sey´ has found to adopt different roles depending upon semantic information associated with nouns. The agent role of ´sey´ found by such classification helped identifying subject and indirect subject roles of tetravalent causative verbs found in Urdu/Hindi languages.