Robert Briscoe and I have been working for quite some time on a paper that synthesizes work that the two of us had been doing independently concerning the Two Visual Systems Hypothesis. Each of us is concerned (albeit with somewhat different motivations) to defend the idea that conscious visual awareness plays a direct role in controlling visually-guided behavior. However, that idea is widely supposed to have been refuted by Milner and Goodale’s influential theory, according to which direct visuomotor guidance is an entirely non-conscious affair. (See Andy Clark’s famous Phil. Review paper for the most forceful statement of this claim.)
Robert and I argue, however, that this common supposition is absolutely unfounded: we review the empirical evidence for TVSH very carefully, and argue at length that this evidence is entirely compatible with the supposition of a direct role for conscious vision in the guidance of action, and in deep conflict with the claim that visuomotor guidance is entirely the responsibility of the non-conscious dorsal system. How do we show this? A link to our paper is below. We plan to send it to a journal some time soon, but in the meantime we would be very grateful for feedback from anyone who’d like to read it. Feel free to post in the comments below, or to send one of us an e-mail.
I agree with your thesis. For visually guided behavior (as opposed to reflexive visuo-motor behavior) we need to know *what is where* in our egocentric space. The two visual systems model (what *and* where) only points to the separate preconscious brain systems (dorsal and ventral) that provide input to what I call retinoid space, where our various preconscious sensory patterns are integrated in spatio-temporal register to compose the current content of our egocentric phenomenal world. Visually guided behavior then depends on information parsed out of this conscious representation of the world to which we must adapt.
Hi Arnold,
Thanks for your comment. An important part of the account of TVSH we are challenging is a denial of the Ungerleider & Mishkin account of the dorsal/ventral distinction in terms of “what” and “where” pathways. Instead, Milner and Goodale would say that both streams represent information relevant to location as well as object identity (though high-level categorizations are the responsibility of the ventral stream), and the main differences are between (1) the way these spatial properties are represented (egocentrically in DS, allocentrical in VS) and — most importantly for our purposes — (2) the uses to which these representations can be put (VS is for conscious vision, DS is for online visuomotor guidance). So I think that M&G would concede your first point — that visually guided behavior requires knowledge of both “what” and “where” — while denying your last one: they would insist that the visual information drawn on in this guidance has nothing to do with our conscious representation of the world. And that is what Robert and I dispute.