Well, I figure this is as good a place as any to make a return post.
I think this is pretty much the synthesis of a sort of meta-argument between Pat and I, although I always got the impression Pat discounted the seriousness which the potential ir-reality presents (unfortunately Pat isn't visiting these days to defend his position). To extend the metaphor used in the article: Just because the icon on the pc desktop is only a representation of the file, which in no way describes or represents to me its true "itness" if you will, the underlying itness is gone if I delete it (or at another level of gone, incinerate the hard drive).
This is a great article, and one that I think everyone should read and grapple with. In fact, the issues he's expressing occupy a central position in the work I'm doing; some of it even serves as the topic of my article coming out in July. These findings have serious implications for the study of narrative and narrative theory (after all, if observation designates a dilemma of perspective on "truth," can there be any such thing as a reliable narrator?).
As far as my own position goes, I can't think of any issue I have with taking representations "seriously." My only reservation derives from the apparent faith in evolutionary development, which even Hoffman seems to be wary of, despite his admission that we have no choice but to trust it. In short, I'm not sure that consciousness even qualifies as an acceptable fitness function, at least if we're expanding our scales here. Consciousness may be adequate for the social environment at hand, in which technology assures relative security/safety; but even then, we can barely wrap our brains around the algorithms that dictate those technologies. More than anything else, more than providing representationally efficient images of the world around us, I would say that consciousness allows for security by constructing the impression of control, without requiring us to actually map the terrain.
Kant responded to the problems that Hume raised by introducing intersubjectivity. Although what Hoffman is talking about isn't completely in alignment with the way Kant described intersubjectivity, but it's pretty close. Basically that we can't know the "thing in itself", but that we can know what we can all agree on. It's a compromise between empiricism and Hume's problem of induction, and appears to be right on a certain level.
I think this is a scientific recapitulation of the Kantian dilemma, but I think there's a contemporary intellectual more appropriate for the current terminology, and that's Niklas Luhmann. Kant is like the specter we can't get away from - forget Hegel and Marx, the crucial dilemma really comes back to Kant, and this is where Luhmann comes into play. In a way, the dilemma of Luhmannian systems theory isn't strictly Kantian because I think Luhmann would say it isn't about individual choice or belief - it's about necessity, and I think that Luhmann and Hoffman both understand this as evolutionary necessity (Kant didn't have the evolutionary paradigm to work with). But the model of Luhmannian systems theory definitely owes its organizational structure to Kant.
