Many people are sceptical of Mokyr's idealist take on the Industrial Revolution, but even if you are suspicious, The Enlightened Economy is still a tremendous book. I think this review in JEL gives the correct flavour. It says the comprehensiveness is a curse, but the 'curse' teaches you a lot !!!
In light of the genetic evidence about very long-run jati endogamy, what is the best case for the arguments put forth by Bayly and Dirks? I consider Bayly superior, but they both assume, as a matter of course, that caste identities had always been fluid in the pre-modern period without elaboration
( Neither argues that caste was 'invented' in modern times, which is a strawman argument used by critics. They both acknowledge its antiquity. But they both argue caste identities were hardened / made less flexible in modern times )