Extending construction morphology: typological and pragmatic challenges from Urdu compounding patterns

This study investigates the structure and productivity of Urdu endocentric compounds through the lens of Booij’s construction morphology (CxM). While compounding has been extensively studied in Indo-European languages, Urdu remains an underexplored domain despite its typologically rich blend of Indo-Aryan and Perso-Arabic morphological patterns. The research addresses this gap by examining whether CxM can adequately model the formal and semantic regularities of Urdu compounds, which often combine agglutinative and analytic features. Using schema-based head–modifier templates, the analysis covers several compound types such as Noun–Noun, Adjective–Noun, Verb–Noun, Adjective–Verb and Verb–Verb. The findings reveal that Urdu compounds are predominantly right-headed and endocentric, with the second constituent determining grammatical category and meaning. CxM effectively captures the productive and lexicon-based nature of these constructions, providing clear correspondences between form and meaning. At the same time, the analysis identifies typologically distinctive patterns in Urdu that challenge the predictive power of CxM. These include coordinate (dvandva) and exocentric compounds, reflexive and idiomatic expressions and metaphorical extensions with non-compositional meanings. Furthermore, Urdu’s phonological characteristics, such as stress alignment, sandhi assimilation, and Persian-Arabic loanword blending are insufficiently represented within CxM’s morphosyntactic framework. To address these gaps, this study proposes a theoretically integrative model that draws on Lexical Phonology (LP) for sound patterning, distributed morphology (DM) for argument structure in verbal compounds, and Lieber’s Lexical Semantic Framework (LSF) for cultural and idiomatic semantics. By focusing on Urdu as a typologically hybrid and theoretically informative case, this research extends the empirical reach of CxM and demonstrates the need to adapt construction-based models to underrepresented linguistic systems.