“Phenomenology is, as I have argued elsewhere, striated with “darkness” – experiencing it only affords a partial and very fallible insight into its nature. We are not normally aware of this darkness because, as Scott Bakker writes, it “provides no information about the absence of information.” However, this opacity can be exhibited from a third-person perspective in cases of “anosognosia” – conditions where patients are unable to access the fact that they have some sensorimotor deficit, such as blindness, deafness or the inability to move a limb. Sufferers from Anton’s syndrome or “blindness denial,” for example, are blind as a result of damage to visual areas in the brain. But, when questioned they deny that they are blind and attempt to act as if they were not.  This shows not only that the people can be radically mistaken about the contents of their conscious experience but that a standard Cartesian impossibility claim – that we cannot make a perceptual judgment without having a corresponding perception – is false. Minds assumed impossible on the basis of armchair reasoning turn out to be quite possible
The blindness of the mind to its true nature is also exhibited among unimpaired agents. We regularly assume that we are authoritative about the reasons for our choices. Yet studies into the phenomenon of “choice blindness” by Petter Johansson and Lars Hall suggest that humans can be gulled into attributing reasons to themselves that they did not have. In one case, subjects in a supermarket were asked to rate jams and teas, following which they were apparently presented with samples of the tea or jam they had chosen earlier and asked to explain their choice. In manipulated trials the samples were sneakily switched with samples of different products. Remarkably, less than a half the experimental participants noticed the switch, despite striking differences between the substituted pairs of flavours. The remainder sought retrospective justifications for choices they had not made.
Lars and Hall have been able to exhibit choice blindness in moral reasoning. In another experiment, subjects were asked to rate their agreement with controversial moral claims in a survey form. Unbeknownst to the experimental subjects, the pages with the original rated statements were switched for subtly altered sentences expressing contrary moral claims. However, when asked to review and discuss their rating, a majority of experimental subjects confabulated reasons for moral positions opposing the ones that had earlier embraced.
Phenomena such as choice blindness and anosognosia suggest that our insight into subjectivity depends on a fallible process of self-interpretation that is subjectively “transparent” and immediate only because we are not aware that it is a process at all. Thomas Metzinger calls this constraint “autoepistemic closure.” By virtue of it, the vivid world “out there” and our vital, rich “inner” life appear not to be models or interpretations only because we are not aware of concocting them.
Metzinger argues that phenomenology is systematically misleading about what phenomenology really is because it needs to be. A system that modeled itself and attempted to model that modeling process in turn (and so on) would require infinite representational resources. Phenomenological darkness thus prevents the self-interpreter from becoming entangled “in endless internal loops of higher-order self-modeling.” It is thus reasonable to argue that the anti-reductionist intuition that subjective experience is inexplicable in terms of non-subjective physical or computational processes is an artifact of this phenomenological darkness.” references @ http://enemyindustry.net/blog/?p=6411