D D Kosambi offered remarkable insights into the history of ancient India. Does his archaeology measure up to his history? The approach in this paper is to view the internal logic of his hypotheses in archaeology, and to ask if Kosambi did justice to the data available in his time. Did he present a sound data analysis that could be emulated, enlarged, or reworked? The answer has to be "no". Kosambi's site locations were not precise; he was not interested in the typology of stone tools; and his correlations of tool occurrences with sacred sites, of the tribe with an absence of plough agriculture, and of iron technology with agricultural surpluses, were flawed. Perhaps Kosambi's archaeology does not measure up to his history because for him archaeology was only an extension of history. But in order to work with the entities of archaeology, typology and classification are indispensable: as indispensable as is the knowledge of an ancient language for the historian. Failure to engage in the grammar of these entities and an ignorance of site formation processes give rise to faulty generalisation.
