Intermediate Level Theory of Consciousnessthe sufficiency of working-memory-available information for consciousness. More specifically,
I think there is evidence of information already encoded in working memory that is not con-
scious. If I am right, then it follows that there can be information merely available to working
memory that is not conscious. Prinz’s theory would, then, be in trouble.
This evidence comes from Hassin et al. (2009) and Soto, Mantyla, and Silvanto (2011). Focusing
on the former study, they present subjects with disks that are either ‘full’ or ‘empty.’ The task is
simple: identify whether the disk that appears on the screen is full or empty. Disks appear in sets of
five that can either follow a pattern (like a zig-zag, pattern, for instance), follow a broken pattern,
where four disks follow the pattern and the fifth disrupts it, follow a control condition, or finally
follow a totally random condition. Subjects are not informed of the patterns. The variable of inter-
est is, of course, the reaction times for the discrimination task. Predictably, the researchers found
that reaction times were higher for control and random conditions versus the pattern condition.
The interesting result is that reaction times for the broken pattern condition were significantly
higher than in the pattern condition. This supports the hypothesis that subjects were extracting
information about the pattern and using it to aid in the discrimination task. By multiple post-test
measures, subjects were unaware of the presence of the pattern. So, it appears that subjects were
unconsciously using an extracted pattern to enhance their performance on the full-or-empty task.
Since this is the kind of executive work expected of working memory, this all counts as evidence
for encoded items in working memory that are unconscious. Hence it is evidence against Prinz’s
AIR theory of consciousness.
Prinz’s response to this evidence is to implicate another memory store besides working
memory. Instead of the information about patterns being stored there, he offers the alterna-
tive theory that the information is stored in what is called ‘fragile visual short-term memory,’ a
modality specific, high capacity storage unit. If the information is stored there, and can explain
the increased performance in the pattern condition, then there is no evidence from the Hassin
study that unconscious information can be encoded in working memory. There is a swift, and
to my mind decisive, response to Prinz’s alternative theory. One extra experiment run by Hassin
and colleagues utilized the same experimental setup, but altered the stimuli. In place of a visual
pattern, they substituted an algebraic pattern (e.g., 2, 4, 6, 8; or 1, 3, 5, 7). Again, the same results
from they pattern and broken pattern conditions obtained (and so, too, the unawareness of the
patterns in post-task examinations). Since the patterns here were algebraic rather than visual,
the fragile visual short-term memory store can offer no explanation for the performance. It
appears, then, that there can be unconscious information encoded in working memory. Since
that information is stored there, it must have been, at some point, available to working memory,
too. The result is that information (intermediate level representations, or whatever you like) can
be available to working memory without being conscious. This is exactly the wrong result for
Prinz’s theory of consciousness.
4 ConclusionWhat, in the final analysis, should we make of Prinz’s AIR theory? I think the empirical details
do not bear out his main contentions very well, including crucially the heart of the theory
(claim 4). Nevertheless, there is much of value in the evidence he presents for his theory. The
methodology he employs comparing cases of consciousness with cases of no consciousness is a
very natural idea. Pursuing that method, it might be of value to examine the neural processes
of the Hassin subjects and compare them with neural processing involving conscious patterns.
Perhaps there are differences that would offer clues about the nature of consciousness, much
like hemispatial neglect patients give us reason to implicate attention. It is also very tempting to