Abstract
With negative and positive evidence from Urdu/English intra-sentential code-switching data, the paper attempts to determine the empirical adequacy of two contrasting proposals regarding the placement of complements in mixed sentences. Although Mahootian and Santorini (1996) and Chan (2008) employ different theoretical frameworks, they uniformly reject the possibility of a ‘third’ grammar and attempt to account for the data in terms of existing apparatus. Employing Tree Adjoining Grammar as framework, Mahootian and Santorini (1996) claim that lexical categories as heads of elementary trees determine the position of their complements whereas Chan (2008), working in Principle and Parameters Theory, proposes that functional categories being associated with a particular value of head-parameter determine placement of their respective complements. The paper rejects both the proposals with counter-evidence from the corpus of Urdu/English code-switching. The pre-head placement of complements selected by English Vs and Ns in mixed sentences contradict Mahootian and Santorini’s (1996) proposal regarding lexical heads whereas post-head placement of complement TPs and placement of complement PPs/PosPs in projections without overt functional heads pose challenges to Chan’s proposal regarding the distinction between lexical and functional categories. Thus, with empirical evidence, the paper rejects both the proposals regarding placement of complements in code-switched data.