This article aims at deconstructing the conception of multilingualism developed in mainstream sociolinguistics by critically examining the assumptions underlying this trend of research, which is grounded in the scholarship of Labov (1972), Fishman (1984) and even Gumperz (1972). In order to engage in that discussion, we use the Mauritian sociolinguistic landscape, as described by researchers following that tradition, as a case. We, thus, carry out a meta-analysis of existing sociolinguistic research conducted in Mauritius, which serve to illustrate the extent to which knowledge produced bear the influence of the structuralist approach. Then, we critically discuss and reflect upon the assumptions underpinning such research, and in so doing, challenge key concepts such as language and diglossia. Finally, we open a discussion on the need to adopt an alternative epistemological position in order to construct a different type of interpretation of the phenomenon following the ground-breaking work of scholars such as Makoni and Pennycook (2007), Herdina and Jessner (2002), Blackledge and Creese (2010), Garcia (2009) and de Robillard (2005, 2007).