Dealing with the debates that surrounded the preparation and the organisation of the census, this study by a member of the Commission responsible for the composition of the list of nationalities for the first post-Soviet census, focuses primarily on the “Tatar” question loudly expressed at that time. Recognising that the overall operation was depending on not only scientific, but also administrative and political considerations, the author provides a useful historical outline of Russian censuses. According to him, the reduction of the number of nationalities officially recognised by Soviet officials from 190 in 1926 to 109 in 1959 was not due to any processes of ethnic assimilation but to administrative injunctions. Thus, the 194 nationalities listened for the 2002 census stayed as the sign of a liberalisation based on the recognition of the principle of free choice of their national identity by all citizens of Russia. Arguing in favour of a open list of nationalities, the author demonstrates the validity of his liberal orientation through the Tatar example. Analysing how the term “Tatar” became an ethnic denomination, he remembers how Crimean Tatars were excluded from “Tatar” classification before the 1989 census. In 2002, the exclusion from the “Tatar” category of the “Kriashens”, Orthodox Tatars claiming for the recognition of their national status, was debated by the authorities of Tatarstan as allegedly undermining the unity of the Tatar nation. The author sheds light on the primacy of political logics in these Tatar complains against the official list of nationalities. Because the local “ethnocratic regime” would lose part of its legitimacy in case of a reduction of the number of Tatars living in Tatarstan, its elites harshly criticised Moscow’s imperialism. Despite some exaggerations (when denouncing the “mini-imperialism” of Tatars towards the Kriaschens), this article is to stay as a reference for social scientists interested in the modern and contemporary evolution of collective identity-building processes.