“Given the inability of SDT-based models to account for blind insight, our data suggest that a more radical revision of metacognition models is required. One potential direction for revision would take into account the evidence, mentioned in the Introduction, that neural dynamics underlying perceptual decisions involve counterflowing bottom-up and top-down neural signals (Bowman et al., 2006; Jaskowski & Verleger, 2007; Salin & Bullier, 1995). A framework for interpreting these countercurrent dynamics is provided by predictive processing, which proposes that top-down projections convey predictions (expectations) about the causes of sensory signals, with bottom-up projections communicating mismatches (prediction errors) between expected and observed signals across hierarchical levels, with their mutual dynamics unfolding according to the principles of Bayesian inference (Clark, 2013). Future models of metacognition could leverage this framework to propose that both first-order and metacognitive discriminations emerge from the interaction of top-down expectations and bottom- up prediction errors, for example by allowing top-down signals to reshape the probability distributions of evidence on which decision thresholds are imposed (Barrett et al., 2013). We can at this stage only speculate as to whether such a model might provide the means to account for the blind-insight phenomenon and recognize that predictive coding is just one among a variety of potential frameworks that could be applied to that challenge (Timmermans et al., 2012).” Ryan B. Scott et al, “Blind Insight: Metacognitive Discrimination Despite Chance Task Performance,” 8
Just thinking in these terms renders traditional assumptions regarding the character and capacity of philosophical reflection deeply suspect. Is it really just a coincidence that all the old riddles regarding the human remain just as confounding? You need only consider the challenge the brain poses to itself to realize the brain simply cannot track its own activities the way it tracks activities in its environments. The traditionalists would have you believe that reflection reveals an alternate order of efficacy, if not being. So far, the apparent obviousness of the intuitions and the absence of any credible account of the work they seem to do has allowed them to make an abductive case. Reflection, they argue, discriminates autonomous/irreducible/transcendental functions and/or phenomena. Of course, they don’t so much agree on the actual discriminations they make as they agree that such discriminations can and must be made.
My bet is that the brain does a lot of causal (Bayesian) predictive processing troubleshooting its environments and relies on some kind of noncausal predictive processing to troubleshoot itself and other brains. You only need to look at the dimensions missing in the ‘mental’ or the ‘normative’ or the ‘phenomenological’ to realize they’re precisely the kinds of information we should expect an overmatched metacognition to neglect. Where the brain is able to articulate efficacies into mechanistic (lateral) relationships in certain, typically natural environments, it must posit unarticulated efficacies in other, typically social environments. My hypothesis is that the countless naturalistically inscrutable, ontologically exceptional, alternate orders of efficacy posited by the traditionalist amount to nothing more than this.
Either way, this research is killing traditional philosophy as we speak.