76 State power and identification schemes in Rakhine Network on Statelessness and Institute on Statelessness and Inclusion 2019; Kyaw 2019). The colonial categorisation process is assumed in some scholarship and policy analyses relating to the 1982 Citizenship Law to be neutral, objective, and rigorous, rather than part of a flawed economic and governance project (reflected in work such as Leider 2018; Tonkin 2018). Ethnic groups within such accounts are assumed to exist largely unchanged until today. The accuracy and validity of these categories, as well as the essentialist premise on which they have been based, are challenged in Rohingya narratives. For example: If it is true that the Rohingya do not exist, then it is also true that the Rakhine do not exist. If we go by the British records as they do, we will see the British referred to Mohammedans, Arakanese and Moghs—a term which is rejected. The Moghs and the Arakanese became Rakhine, and the Mohammedans became Rohingya.22 This comment reflects the fact that British categorisation processes, established as part of an approach to make colonised populations in Burma ‘legible’ and thus ‘governable,’ were imperfect. Linguistic, ethnic, religious, national, and residence categories were inaccurately and inconsistently recorded over time even by administrators own admissions (Bennison 1933). The conceptual confusion and inaccuracies between categories caused considerable resentment by those who were wrongly recorded. Thus, the 1931 census, towards the end of British rule when the separation of Burma from India was being considered, attempted to enumerate which side of the India/Burma border individuals were domiciled. Collecting census data at different times during the rice season produced different results, which failed to capture the nature of circular and seasonal migration. Earlier censuses categorised people differently from the 1931 census and produced wide margins of error (Bennison 1933). ‘Arakan Mohammedans,’ the category within which many Rohingya fell was moved from the inaccurate category of ‘Indian races’ to ‘Indo-Burman races’ between the 1921 and 1931 censuses in an attempt to rectify past inaccuracies that conflated language and ethnicity with territorial belonging (Bennison 1933). These categories later provided inaccurate proxies for national belonging that failed to capture the long-standing demographic realities of the borderlands. Census data provided shaky ground on which to retrospectively establish one’s ethnic existence in modern Myanmar under the 1982 Citizenship Law. Campaigning on the part of Burmese Muslims more generally to be recategorised occurred particularly in the lead-up to independence. The categories conflated ethnic/linguistic/religious identities with national/residence categories and assumed them to belong to India, Chittagong, or Bengal. As Yegar (1972:63) noted, The Burmese Muslims resented the fact that the government (British India) considered them part of the Indian community and lumped them together in one category in census and other statistical reports.