Iron Age levels have yielded dates of 2970 + 105 bp (TF-570) 1255, 1240, 1221 cal. BC and 2820 + 100 bp (TF-573) 993 cal. BC from Hallur, and 2905 + … - George Erdosy

For example, Erdosy (1995) noted that cremation was pretty common in 2000 BCE Balochistan (Penano Ghundai II, Mughal Ghundai III, Dabar Kot, Mehi, Sutkagen-dor) but very rare in Central Asia: “If anything, on present evidence, cremations appear to have originated in the Indo-Iranian Borderlands and spread northwest (and southeast) thence, against the grain of postulated movements of Indo-Aryan speakers”.

Erdosy testifies that “the principal concern” of scholars (like Witzel) studying South Asian linguistics is to find “evidence for the external origins - and likely arrival in the 2nd millennium BC - of Indo-Aryan languages”;

The alternative thesis (Chakrabarti 1977), that iron smelting was developed in the subcontinent, rests on two principal arguments. First, iron ore is found across the length and the breadth of India, outside alluvial plains, in quantities that were certainly viable for exploitation by the primitive methods observable even in this century (Ball 1881; Elwin 1942). Ample opportunities thus existed for experimentation, although given the complexity or iron smelting this is not a conclusive point. The second argument, that the earliest evidence for iron comes from the peninsula and not from the northwest, is much more persuasive, even if better examples than quoted by Chakrabarti can be adduced in support of it. Briefly, while the dating of Phase II of Nagda (the earliest iron bearing level) depends on ceramic analogies, and the stratigraphy of Ahar (another site which is claimed to have produced evidence for iron) is hopelessly muddled, the testimony of radiocarbon dates is instructive.