I agree entirely with Razib Khan's case that the Harappan language of the Indus River Valley people was neither a Dravidian nor a Munda language. (There is also overwhelming evidence that it was not Indo-European.)
UPDATED August 9, 2018
Michael Witzel has described the linguistic substrate in Rig Vedic Sanskrit evidenced by about 300 words, with the unfortunate appellation "Para-Munda" which is misleading at best, although he does reference the historically correct name for the language of the Indus Valley Civilization's people, Meluhha, which we know from Sumerian accounts of its IVC ex-patriot neighborhood that captured the name phonetically.
He is right, I am very comfortable in concluding, however, that Meluhha was not a member of the Dravidian language family or the Indo-European language family.
To the extent that there are any commonalities between the Meluhha language and the Munda languages of India, they are probably either mere coincidences or derive from a Meluhha adstrate/substrate or Meluhha borrowings into the Munda languages.
I am also inclined to think that Witzel's timeline for Indo-Aryan migration into South Asia is probably a bit late. I tend to see the evidence as favoring a date closer to 1950 BCE (based upon evidence from Cemetery H of cremation, upon traces of new metallurgy technologies and upon historical climate factors), while he tends to favor a date as late as 1300 BCE.
On the other hand, there is something to Witzel's inference based upon the kind of vocabulary that ended up having substrate influences, that Meluhha's substrate influences on early Vedic Sanskrit arose at a post-collapse phase of Harappan civilization and reflects rural agricultural village life rather than the grand urban civilization involved in long distance international trade found in the Indus River Valley that preceded it.
His conclusion that Harappan seals were probably not a full fledged alphabet or written language and were instead a system of logos or symbols (albeit with some very limited additional information in a small character set) is probably accurate, however.
I also tend to agree that there is some merit in Witzel's claim that the Vedic religion was predominantly Indo-European rather than having a great deal of continuity with the Harappan religion, although I would be inclined to think that he underestimates the extent to which indigenous Harappan and Dravidian religious elements may have subsequently resurfaced and influenced the Hindu religious tradition that evolved and emerged from the early Vedic religion as recounted in the Rig Veda.