In “Why and how does consciousness seem the way it seems?” Dennett gives an argument for why philosophers and scientists should abandon Ned Block’s distinction between access consciousness and phenomenal consciousness, zombies, and qualia altogether. The argument is twofold: first Dennett lays down his argument for why the assumption of phenomenal consciousness as a second medium whose states are conscious experiences or qualia is “scientifically insupportable and deeply misleading” (Dennett this collection, section 1). It is insupportable because there is simply no need to posit such entities to explain any of our behavior, so for reasons of parsimony they should not be a part of scientific theories (see also Dennett 1991, p. 134). The assumption is deeply misleading because it makes us look for the wrong things, namely, the objects our judgments are about, rather than the causes of these judgments, which are nothing like these objects.
In a second step Dennett shows why creatures like us must be convinced that there are qualia, that is, why we have such a strong temptation to believe in their existence, even though there are no good reasons for this (Dennett this collection, section 2 and 3; other places where Dennett acknowledges this conviction, the zombic hunch, are Dennett 1999; Dennett 2005, Ch. 1; Dennett 2013a, p. 283). The following sections are exclusively concerned with this part of the target paper.
After completing the second step, Dennett explains why we ascribe qualia their characteristic properties—simplicity and ineffability (Dennett this collection, section 4 & 5). Although I also say something about this point (see section 3), Section 6 is an intuition pump (cf. e.g., Dennett 2013a) that will help the reader to apply Dennett’s alternative view to the experience of colors.
Before I present a short outline of Dennett’s second step, I want to briefly describe the predictive processing framework. This is necessary since both Dennett’s argument as well as my reconstruction make use of this framework. I shall not go into details of hierarchical predictive processing (PP) accounts here, since at least three papers in this collection (Clark, Hohwy, and Seth), as well as the associated commentaries (Madary, Harkness, and Wiese), are concerned with this topic and also offer ample references for introductory as well as further reading. I will instead give a very short description of the points that are most relevant to Dennett’s argument and recommend the above-mentioned papers and the references given there to the interested reader.
In the PP framework, the brain refines an internal generative stochastic model of the world by continuously comparing sensory input (extero- as well as interoceptive) with predictions continuously created by the model. The overall model is spread across a hierarchy of layers, where the sensory layer is the lowest and each layer tries to predict (that is, to suppress) the activation pattern of the layer beneath it. The whole top-down activation pattern might be interpreted as a global hypothesis about the hidden causes of ongoing sensory stimulation. The difference between predicted and actual activation (prediction error) is what gets propagated up the hierarchy and leads to changes in the hypothesis. To be exact, this is only one possibility. Another is that this leads to an action that changes the input in such a way that the prediction is vindicated (active inference, see e.g., Friston et al. 2011). However, although this aspect of PP—that it provides one formally-unified approach to perception and action—is a strength of the framework, it is not important here, given the context of this commentary. These changes are supposed to follow Bayes’ Theorem, which is why one might speak of Bayesian prediction (cf. e.g., Hohwy 2013).
The higher the layer in the hierarchy the more abstract the contents and the longer the time-scales or the predictive horizon. One example of a very abstract content is “only one object can exist in the same place at the same time” (Hohwy et al. 2008, p. 691, quoted after Clark 2013, p. 5).
One point to keep in mind is that, according to Hohwy (2014), this framework implies a clear-cut distinction between the mind and the world. That is, there is an evidentiary boundary between “where the prediction error minimization occurs” and “hidden causes [of the sensory stimulation pattern] on the other side” (Hohwy 2014, p. 7). I will come back to this point later in this commentary.
Our own dispositions, expectations, etc. are part of the generative self-model instantiated by our brains. “We ought to have good Bayesian expectations about what we will do next, what we will think next, and what we will expect next” (Dennett this collection, p. 5)
When our brains do their job (described in (1)) correctly, i.e., there are no prediction-error signals, we misidentify dispositions of the organism with properties of another object. For instance, instead of attributing the disposition to cuddle a baby correctly to the organism having the disposition, our brain attributes “cuteness” to the baby.[3] Color qualia and other types of qualia also belong to this category.[4]
This means, under a personal level description, that we believe that there are properties independent of the observer, such as the cuteness of babies, the sweetness of apples, or the blueness of the sky, etc.
This is why it is so hard for us to doubt that qualia exist in the real world.
The crucial points seem to be (1) and (2). Before I lay out my interpretation I want to highlight some points that are not addressed in Dennett (this collection), but which are crucial if we are to have a complete picture. In the section Our Bayesian brains, I present a reconstruction that addresses these issues.
Why do we need to monitor our dispositions? As noted in Dennett (2010), self-monitoring, in the sense of monitoring of our dispositions, values, etc., isn’t needed unless one needs to communicate and to hide and share specific information about oneself at will. In his paper, Dennett does not address this issue, yet presupposes that “among the things in our Umwelt that matter to our wellbeing are ourselves”. This is obvious if one reads “ourselves” as the motions of our bodies, but not so obvious if one includes things like “what we will think next, and what we will expect next”, as Dennett does (Dennett this collection, p. 5). The next question is concerned with this latter form of self-monitoring:
How is self-monitoring accomplished? Hohwy (2014) refers to an evidential boundary in the predictive processing framework (see the section 21): there is a clear distinction between the mind/brain and the world (of which the body without the brain is a part), whose causal structure is yet to be revealed. Our expectations are part of our mind, which, if talk of the boundary is correct, does not have direct access to its own states as its own states—the mind is a black box to itself. So the prediction of its expectations needs to be indirect (just like the predictions of the causes of the sensory stimulation in general), and therefore the question arises how the self-monitoring of the mind is achieved according to Dennett. There is a further concern with self-monitoring, which one might call the “acquisition constraint” (cf. e.g., Metzinger 2003, p. 344):
How did this self-monitoring evolve in a gradual fashion? Large parts of Breaking the Spell are dedicated to making understandable how “belief in belief” could have evolved over the centuries, beginning long before the appearance of any religion. Dennett’s goal here is quite similar: the explanation aims to make understandable how we came to believe in qualia, etc. But a step-by-step explanation is missing. I consider this form of the acquisition-constraint one of the most crucial for any satisfying explanation of this sort: each single step has to be understandable as one likely to have happened. One reason for this is that it would support a more fine-grained and mechanistic understanding; another is that it would satisfy the gradualism-constraint of Darwinism, which says that minds (just like anything else) “must have come into existence gradually, by steps that are barely discernible even in retrospect” (Dennett 1995, p. 200, emphasis in original).
Once we know why and how our brains accomplish the task of monitoring our dispositions and how they came to do so, one might still wonder why (as claimed in point 2, page here) exactly these abstract properties of the organism would be misidentified as concrete properties of other things:
Why do we misidentify our dispositions? One of Dennett’s central claims is that we misidentify our own dispositions, which leads to belief in qualia.[5] Although misidentification seems to be ubiquitous (see superstition, religion, magic tricks, the rubber hand illusion—Botvinick & Cohen 1998; and even full body illusions—Blanke & Metzinger 2009) it nonetheless requires a special explanation in each case: is this a shortcoming of a system that has no disadvantages, or is it even something that benefits the system in some way (cf. McKay & Dennett 2009)? Keeping this last possibility in mind one might ask:
Why are we so attached to the idea of qualia? There seems to be something more that leads people to believe in qualia. There is the intuition that without qualia we would be very different—we would be “mere machines”, we could not enjoy things like a good meal or the smell of the air after it rains (a discussion of this characteristic of beliefs-about-qualia can be found in Dennett 1991, p. 383). Some might go further and say that our whole morality rests on the existence of qualia of pain and suffering (this worry is dealt with in Dennett 1991, p. 449). However, what I am concerned with here is not whether it is true that qualia are the basis of our morality, but why we should think them to be so. From the argument presented by Dennett it is not clear why we are so attached to the idea of qualia. It is not obvious why we do not react as disinterestedly to their denial as we did to the revelation that there is no ether.[6] But, as a matter of fact, we react differently: this is not like when any other entity, posited for theoretical reasons, is shown to not exist; it is as if without qualia we couldn’t possibly be us.