Comments on Derek Ball’s “The New New Mysterianism”

Hello my name is James Dow. I’m an ABD doctoral student at the CUNY Graduate Center and an Adjunct Faculty member at Drew University. In the following video, I will be commenting on Derek Ball’s paper “The New New Mysterianism.” Derek’s paper begins with a summary of the anti-physicalist argument that consciousness cannot be explained in physical terms. The anti-physicalist (Derek uses Chalmers as his main target...) has argued that reductive physicalism about consciousness requires armchair (or a priori) deducibility. If physicalism is true, then it must be in principle possible to infer the truths of consciousness from the truths of physics on the basis of merely conceptual competence and armchair reasoning. The anti-physicalist, however, maintains that there is an explanatory gap— the facts of consciousness do not appear among the facts of the complete physics. So, the anti-physicalist infers, physicalism is false.

Derek is willing to grant that physicalists are committed to armchair deducibility, but he does not account for the explanatory gap by maintaining that there are facts about consciousness that cannot be explained. Instead, he wants to argue that even if we could in principle deduce the consciousness truths from the physical truths, the explanatory gap might arise from another source— a limitation on human reasoning abilities. So, he reasons, an armchair physicalist is free to insist that there are reasoning abilities that could deduce the consciousness facts from the physical facts, but that the explanatory gap is unbridgeable for human reasoning abilities as they are now. If Derek’s argument can be made good, then the existence of this kind of explanatory gap cannot be used to argue against physicalism. As he suggests in his concluding sentence, his view is like the type-A physicalist’s view because he denies that there is an explanatory gap and commits to armchair deducibility, but it is unlike that view because he provides a substantive account of why the explanatory gap is unbridgeable.

The argument begins by appealing to the idea that there are undecidible propositions. Derek discusses undecidability by considering the halting problem for Turing machines. Given any Turing machine, that Turing machine cannot solve the halting problem. It cannot determine whether at any point of its calculation, it will get a result given the algorithm or it will halt. Derek argues that since human reasoning is captured by the Turing machine, then although it would be in principle a priori deducible (given the program) whether at any point its calculation, a machine will get a result or will halt, we human beings (again assuming our reasoning is limited to the Turing machine computable) cannot determine whether it will get a result or will halt.

Derek’s argument seems to depend upon the idea that there is a deep analogy between the explanatory gap and the halting problem. Although the consciousness facts may be in principle a priori deducible from the physical facts (like the halts are in principle a priori deducible from the program facts), we cannot make that deduction because of a limitation to our reasoning abilities (like the Turing machine cannot deduce whether it will get a result or halt). So, Derek reasons, the anti-physicalist’s argument is invalid, because it does not consider the possibility that the explanatory gap is really a gap between types of reasoning abilities, those that can in principle bridge the gap and those that as a matter of psychological fact cannot bridge the gap.

I want to start by considering Derek’s reason to take the anti-physicalists argument to be invalid. He suggests that the argument is invalid because it does not follow from the idea that there is an unbridgeable gap in human reasoning abilities, that there aren’t reasoning abilities that suffice to bridge the gap. But, an immediate question is, “Why should the explanatory gap be a gripping problem for anything other than human beings or beings with human reasoning abilities?” What other reasoning abilities are there than terrestrial ones? He suggests that the explanatory gap might be in principle bridgeable.

The argument seems to require that we make a strict distinction between on the one hand, being in principle deducible and on the other hand, something like being merely terrestrially deducible. Derek points out, “perhaps the physicalist must hold that God could bridge the gap, but she need not grant that we could bridge it; it is bridgeable in principle, but not bridgeable with the tools we have not even in principle.” I would like to pause here to ask a few questions.

Supposing that such a transcendent being could bridge the gap, what type of inference or deduction would he rely upon? One might think that Spinoza’s God (to take a familiar example) would analytically infer or deduce rather than a priori infer or deduce the consciousness facts from the physical facts, given that a priori reasoning is usually explained in terms of entertaining human intuitions or insights. These might seem like strange questions to ask, but, it gets at a worry that the anti-physicalist might express in another way.

The anti-physicalist might wonder why the abstract possibility of undecidable propositions that Derek considers should matter to issue of the explanatory gap for accounts of consciousness. There are other occasions in the philosophical literature where thinkers have attempted to use results in the meta-theory of logic to drive conclusions in other areas. Similar questions have been raised about whether the Lowenheim-Skolem theorem has import for ontological relativity. While Quine and Putnam may have been comfortable with its import, the relationship between such theorems and metaphysical conclusions has come under attack. So, my comment here is that there is some methodological work to be done in showing that we can and should draw metaphysical conclusions from logical results, and in particular that the undecidability results have import for a discussion of the explanatory gap. This is problem that I think Derek can deal with in the abstract, but I actually think that it is a symptom of a more pervasive problem in the argument.

As I mentioned, the argument in general depends upon an analogy between the explanatory gap, on the one hand, and what we might call the undecidability gap, on the other hand. The analogy is something like the following: the gap between consciousness facts and physical facts is analogous to the gap between Turing machine halts and Turing machine physical facts. But, is there an analogy between the explanatory gap and the undecidability gap? That depends partly on what your standards are for analogies. I want to consider some problems for the analogies that Derek employs.

First, let’s consider the analogy of Turing machine physical facts and physical facts in general. When Derek discusses the nature of Turing machines in the paper, he sometimes considers Turing machines to be concrete physical machines and at other times considers Turing machines to be abstract machines.

Suppose we consider a Turing machine to be a physically realizable machine, then notice it is not the program that we turn to in order to deduce the halting truths, but instead, the truths about the vehicle that instantiates the program. And, if we consider the physical facts as the vehicle of the program, not of the program itself, then the physicalist, given a full enough picture of the physical truths, could predict from those physical facts whether the machine would provide a result or not, given that it would just be inferring the results from the mechanism. But, if this is the case, then Derek cannot infer that there are undecidable propositions or machine halts that are inaccessible.

Alternatively, if we consider a Turing machine as an abstract machine (as it is usually understood), then the question arises that came up before, namely, why should such facts about abstract machines tell us anything about the explanatory gap, since the armchair physicalist signs up for the task of inferring the consciousness facts from the concrete physical facts, rather than abstract machine facts. It seems to me that if we approach the explanatory gap, in a way that conceives of the physical facts in this abstract way, then although it might enable us to account for the explanatory gap, our conception of the physical facts lead us away from major tenets of physicalism.

And, I’m not sure whether it matters if we think of the physical in terms of Newtonian physics, or some other type of physics. If it does, Derek hasn’t told us why it does. So at least, it seems unclear whether the analogy holds between the Turing machine facts and the physical facts.

consciousness facts and turing machine halts

Does the analogy between the consciousness facts and the Turing machine halts fair any better? It seems to me a critique of this part of the analogy goes to the heart of the physicalist/anti-physicalist debate. Derek does not articulate what such consciousness facts would be such that they could be even considered to be analogous to the halting truths. The section entitled ‘Consciousness’ might have given a more precise account of Derek’s position on consciousness so that it would come into view how the consciousness facts could be considered to be analogous to the machine halting truths. What Derek does do instead is appeal to other philosophers’ notions of consciousness. But this poses problems for Derek’s argument from the anti-physicalists’ perspective.

I should say by way of a disclaimer that I am not an anti-physicalist under any description, nor do I think that physicalism entails a priori deducibility of consciousness facts from physical facts, but it seems that anti-physicalists like Chalmers, Jackson or Nagel, have a response to Derek’s argument.

Each of these thinkers suggests that the consciousness facts are intuitively accessible in experience. According to many anti-physicalist accounts of consciousness, when one entertains consciousness, one has accessible to oneself a consciousness fact. “You’ll know it when you feel it,” they say. These consciousness facts provide one with all the resources that one needs to capture consciousness demonstratively, whether that be articulated in terms of “qualia,” “phenomenal concepts,” or “what-it’s-like” properties.

According to this kind of story about consciousness, the consciousness facts are simply not the same types of facts as the facts about Turing Machine halts. The facts about Turing machine halts are inaccessible and inferential. The facts about consciousness, so the anti-physicalists might argue, are accessible and non-inferential. Derek may have a response to the anti-physicalist up his sleeve, but it is not in the presentation as far as I can tell. An answer to this question might come from considering the analogy between the explanatory gap and the undecidability gap, so let’s turn to that analogy.

the explanatory gap and the undecidability gap

I did not find a clear enough articulation of the analogy between the explanatory gap and the undecidability gap. In a discussion of inaccessibility, he suggests that the halting truths might be inaccessible to thinkers whose reasoning abilities are limited to the Turing machine computable. He then suggests that consciousness facts might be inaccessible to a central nervous system, because such consciousness facts were an aspect of inaccessible modules that exceed the central nervous system’s computational abilities. But, what does ‘inaccessible’ mean here? These two types of inaccessibility seem very different. Derek defines inaccessibility as “a proposition P is inaccessible to a thinker T if P is inferable from T’s evidence given some set of reasoning abilities, but T cannot infer P without undergoing a qualitative change in her reasoning abilities” (4). While it might make sense to ask what qualitative difference in reasoning abilities would be required to allow a Turing machine to deduce halting truths, it would NOT make sense, for the anti-physicalists, to ask what qualitative difference in reasoning abilities would make a difference in deducing consciousness facts. Presumably, the anti-physicalist does not have to come up with a transcendent reasoning ability that is qualitatively different, only an experience which is qualitatively different. And, the anti-physicalist might argue that because a being possesses consciousness, she can grasp the conclusion without relying on deductions, inferences, etc. The anti-physicalist might argue that Derek’s argument focuses on the reasoning abilities at the expense of the importance of a special kind of conceptual competence. The anti-physicalist argues that the possession of consciousness enables one to have phenomenal concepts that could not appear in a complete physics. For this reason, the anti-physicalists might argue that the explanatory gap is not analogous to the undecidability gap.

Overall, the paper that Derek presents is interesting and provocative, and I very much enjoyed thinking about the arguments. I also look forward to the continued dialogue with Derek and others this week. I will close by commenting that in order for the argument to go through, it seems Derek needs to bolster the main connections in the analogy between the explanatory gap and the undecidability gap. And, until this analogy is propped up, the anti-physicalist does not have to admit that his argument is invalid.