The enumeration of caste and religion by the all-India Census was emblemati
c of a particular kind of imperial rule. Might a different procedures of co
mpilation and record on an alternative epistemological basis have developed
? Censuses had provided governments in India with statistical information a
t least from the 1820's and been used for the local taxation from much earl
ier. The possibility of a continually updates civil registry had been moote
d from time to tome but not put into effect. Another alternative, in which
a census register formed part of the village record similar to the land reg
isters, existed for a brief period in the 1850s in the Panjab. But this 'ru
le by record' was itself based on a theory of positive legislation and the
preservation of social difference, pannomain if not panoptic. A different k
ind of census could only have belonged to another kind of imperial rule.