David Barrettthe same in either case. The same amount of (diminished) processing power will be available
to either kind of invalidly cued target. Yet this is not what Kentridge and his colleagues find.
Though the invalidly cued targets are the same distance away from the bad cue, they appear in
separate invisible rectangles. One of the targets appears in the same invisible rectangle as the inva-
lid cue; the other target appears in a different invisible rectangle from the invalid cue. What they
find is that performance for the cases where the invalid cue is in the same rectangle is enhanced,
compared to cases where the invalid cue is not in the same rectangle as the target. They interpret
these results as showing that object-based attention is captured in these cases. Attention is cap-
tured by the invisible rectangles and this accounts for the facilitating effect when the invalid cue
appears in the same rectangle as the cue. Prinz’s gaze-shifting explanation cannot account for
these results. No matter where the invalidly cued target appears, it is the same distance from the
invalid cue; so, the same amount of processing power should be available in both cases. It appears
that attention in the absence of consciousness (in this case, of rectangles) is possible.
Claim 3One would have noticed, I think, the comparative lack of evidence that Prinz presents for his
view of what attention is compared to the evidence he provides for (1) and (2). Unsurprisingly,
then, claim (3) can come under heavy attack. Since Prinz could, in principle, abandon his view
about what attention is and still maintain the details of his theory—whether we would still want
to call it a true AIR theory or not—I will be somewhat brief in my remarks here.
To begin, there is a vast literature about attentional effects in V1. These so-called ‘early stage’
effects seem to show that attention is not what makes information available to working memory,
since the information that is being manipulated at these early stages is never anything that enters
working memory (it cannot be reported on, stored for later use, etc.). Prinz responds that these
modulations might not properly count as attentional modulations because they might merely
reflect back-propagations from intermediate level processes. It could be, then, that attention does
only affect processing at the intermediate level, which then itself alters the early stage processing.
Moving on, there is also much evidence for ‘late stage’ (post-perceptual) attentional modulation.
As Awh et al. (2006) describe it, there is even perhaps attentional modulation for items already
encoded in working memory. Additionally, Awh and colleagues show evidence that spatial atten-
tion is the mechanism for maintaining information in visual working memory. Much like covert
rehearsal is thought to maintain information in phonological working memory, spatial attention
might refresh visual traces in storage.
If any of these hypotheses capture something important about attention, then clearly it is not
correct to say that attention is simply the mechanism that makes information available to work-
ing memory. In the first case, the information never makes it to working memory; in the other
two the information that is modulated is already in working memory. Yet Prinz is free to drop
this claim (3) about attention. His main focus is on theorizing about consciousness, not atten-
tion. So, what ultimately counts is the evidence in favor of, or against, claim (4). If it turns out
that Prinz has no grand, unifying account of attention to offer, it need not show that his theory
of consciousness is unacceptable. It is to this theory that we now turn.
Claim 4I am aware of no evidence against the idea that information that is conscious must also be
information that is available to working memory. I will instead focus on the idea that there is
information available to working memory that is not conscious. Accordingly, I am attacking