Predictive Processing and the Epistemological Hypothesis: Solving the Hard Problem of Consciousness by Simulating a Brain Facing It

When we say that a theory is able to account for our subjective experience, we simply mean that if this theory were true—e.g., if our brain were as this theory stipulates—then our subjective experience would be indeed as we experience it. A scientist would formulate this idea using the notions of observation and prediction: A theory is able to account for our subjective experience if and only if it is able to predict first-person observations. Several thought experiments suggest that current theories are unable to make such predictions. For example, if we had never seen colors in our lives and therefore did not know what the color blue looks like, these theories would not allow us to deduce (i.e., predict) what it is like to see blue. This well-known problem is often called the hard problem of consciousness (HPC). Here, we address this problem through the lens of the epistemological hypothesis. Under the epistemological hypothesis, the HPC no longer reflects the inability of our theories to predict first-person observations; it reflects our inability to deduce what these theories imply regarding first-person observations. The HPC then becomes an epistemological problem that can be formulated as follows: If we are unable to deduce what a theory implies regarding first-person observations, how can we know whether this theory is able to account for first-person observations? In this paper, we outline a method to test experimentally the epistemological hypothesis and to solve the HPC. Notably, this method makes it possible to test any identity hypothesis. We then highlight the remarkable compatibility between this method and the theoretical framework of predictive processing. We show that a theory of consciousness based on predictive processing implies the epistemological hypothesis—this theory predicts that we are unable to deduce its own implications regarding first-person observations. Finally, this work suggests that the theoretical framework of predictive processing may already have the resources to simulate a brain facing the HPC.