Next Steps in Harappan Decipherment

[Steve Farmer]

Michael Witzel writes:

> Rather, what needs to be done, is not to engage in  lofty speculations
> but to start from square one: a proper list of characters and of sign
> combinations. 800, not 400 signs.  The Indus matter *may* indeed
> be unsolvable with present materials in hand.

Can I add that a separate catalog is needed that focuses exclusively
on *datable* characters found at individual sites? All the
concordances that I've seen clump everything into an undifferentiated
mega-period. The recent study by Bryan Wells (1999) lists important
regional variants, but again passes over chronological issues. Some of
this may arise from poor archaeological controls in early IVC
excavations that yielded the most "seals," but I have the sense from
reading the old reports that more raw data is available on dates
than show up in existing concordances.

Maybe Parpola et al. are right in claiming that evolution in the
script is minimal. I accepted this claim at the start of this
thread in concluding -- based on the apparent lack of normal
scribal pressure forcing script change -- that no extensive 'lost'
Harappan texts ever existed. But that conclusion can't be verified
unless study of a reasonable subset of datable examples shows that
change in the script was in fact minimal -- at least in the mature
Harappan period.

Steve Farmer