Over the past two decades, language policy has seen a significant shift as a field of inquiry, with the previous focus on historical-structural, largely theoretical analysis now largely displaced by a body of empirical work. While the renewed empirical direction of the field, represented by both ethnographic and discursive frameworks, has done much to account for the complexity of language policy processes in different settings (i.e. how policy is ‘done’) and to promote inclusivity within those processes, it has thus far focussed less on the effects of language policy (i.e. what it ‘does’). In this presentation, I will argue that such empirical approaches, read against the background of recent debates in applied and socio-linguistics, provide much possibility for a rescoping of language policy as discourse management. Such a rescoping is necessary in light of calls to denaturalize taken-for-granted concepts like ‘language’, and in particular the idea of distinct, bounded ‘languages’, both being foundational to the emergence of language policy as a modernization project and to its present-day identity.  As an alternative, I will present four key dimensions of language policy as discourse management: hierarchization of knowledge (how policies position some actors as ‘knowing’ and others as ‘un-knowing’), cultural codification (how interactional routines are planned to police social relationships), policing the public sphere (how voices, identities and ideologies are controlled by limiting space in public discourse) and policing of access (how language is instrumentalized to control the movement of people across social spaces). I will conclude by outlining how the critical, empirical study of these four dimensions can serve as a foundation for rearticulation of language policy as the study of how language is instrumentalized in the exercise of power in society.