Wilson's Point in Contrasting Kanzi and Rush Hour

In my previous post, I commented on the way the Kanzi and Rush Hour cases seem to distinguish Clark's version of EC from some other versions of EC (in particular one that Wilson seems to prefer.)  Where Wilson goes with this, however, is to invoke it as a distinction between technological cases of extended cognition (such as Kanzi's) and sociological cases of extended cognition (such as Rush Hour).  (It seems to me to be a bit strained to see the Rush Hour case as a sociological instance, but set that aside.)  Here's the point:

This isn't to say that technological development and appropriation is unimportant in thinking about extended cognition, or that it is never a primary shaper of the extended mind. Rather, it is to claim that we need also to take the nonsymbolic environment seriously in articulating the extended mind thesis, and that the most important place to look in doing so is the social realm. The social and the technological are both significant aspects of extended cognition-evolutionarily, historically, biologically, culturally-and I suspect that it is in tandem that they have sculpted human cognition over tens of thousands of years to its present level. (Wilson, 2010, p. 182).

The symbolic/non-symbolic seems to me to be an odd way to draw a distinction between the technological and the sociological.  Some years ago, in personal communications, I found that Sandy Goldberg and Deb Tollefsen seemed to be interested in cases of extended cognition that involve interpersonal interactions.  [Wilson: "These social features include the interpersonal relations found in human social groups-from dyads through to face-to-face communities" (p. 182)]  They thought that these cases were more plausible cases of extended cognition than some of the standard cases, such as Otto and his notebook or the three modes of Tetris play, since other people's mind are plausibly construed as bearers of (non-derived) representations.  What is odd is that Wilson seems to think that humans are not symbol users.  I'm not sure I'm getting the picture here.