One of the challenges in characterizing non-native varieties of English is accounting for variant uses of ostensibly standard English forms. The present corpus study examines both quantitative and qualitative aspects of pluperfect use in Indian English (IndE), British English (BrE), and American English (AmE). IndE is found to differ from native usage by associating had + V-ed with present perfect and preterite meanings. Licensing of pluperfect contexts by time adverbials is also found to be significantly lower in IndE. AmE shows the lowest overall use of the pluperfect and the highest use of disambiguating adverbials. Thus, AmE and IndE show distinct patterns of divergence from BrE. Variation within IndE exhibits a tendency for greater non-nativeness in regional (vs. national) press and in bureaucratic (vs. press) registers, suggesting a multidimensional distribution of IndE nonstandardness in India. These nonstandard uses are shown to convey new pragmatic meanings deriving from ambiguity in the native system and reinforcement from substrate languages. Finally, these changes are evaluated in relation to the broader tense–modality–aspect system of IndE as well as those of other non-native Englishes which exhibit similar characteristics.