ALIGNING ENGLISH AND PUNJABI WORDNETS: A CORPUS-BASED STUDY

Author(s) : Muhammad Farukh Arslan, Prof. Dr. M. Asim Mahmood and Hira Haroon

Abstract:

The Punjabi language is primarily spoken in the Punjabi region of the subcontinent. In Pakistan, the variant spoken is known as Shahmukhi Punjabi, and it is written in the Perso-Arabic script, while in India, it is written in the Gurmukhi script. Punjabi serves as the mother tongue for 38.77% of Pakistan's population. Despite its historical significance, literary wealth, and a large speaker base, Punjabi has not attained the status of a prestigious, digitally advanced language. To address this, our study aimed to develop a Punjabi WordNet and align it with the English WordNet. The analysis process involved collecting a corpus of nine million (9M) words, cleaning and tagging it to extract grammatical categories, and morphologically analyzing vocabulary items to understand word formation patterns, placing them under specific grammatical tags. The morphological analysis was conducted using Distributed Morphology (DM), revealing the patterns of combining derivational and inflectional morphemes with the root. During this research, a variety of software and digital resources were developed. These include a spell checker, a morphological analyzer, a parts of speech (POS) tagger, a semantic tagger, the Punjabi WordNet, and an alignment tool.
Keywords: Alignment, English and Punjabi Wordnet, Corpus based study