I feel safe to make the strong claim: Analysis IS constrained, circumscribed, guided and suggested when one starts from activity systems framework (or any other). It's quite interesting, in a Hofstadter's G.E.D. sort of sefl-reference, that CHAT's mediational element can illuminate the role of theory in observation and analysis.
A preprint of what Mike suggests is at the url below but it does not have an explicit language element to it, which is one of it's weaknesses:
xmca mailing list
This archive was generated by hypermail 2b29 : Tue Sep 05 2006 - 08:13:16 PDT