A Pāniniān Framework for Analyzing Case Marker Errors in English-Urdu Machine Translation

Panini's Kāraka Theory is solely based on the syntactico-semantic approach to understanding a natural language which takes into consideration the arguments of the verbs. It provides a framework for exhibiting the syntactic relations among constituents in terms of modifier-modified and semantic relations with respect to Kāraka-Vibhakt̪i (semantic role and postposition).In this paper, it has been argued that Pāniniān Dependency Framework can be considered to deal with the MT errors with special reference to case. Firstly, a corpus of approximately 500 English sentences as input have been provided to Google and Bing online MT platforms. Thereafter, all the output sentences in Urdu have been collated in bulk. Thirdly, all the sentences have been evaluated and errors pertaining to case have been categorized based on the Gold Standard. Finally, Pāniniān dependency framework has been proposed for addressing the case-related errors for Indian languages.