Tuesday, October 26, 2004

The Argument from Reason: Reppert Replies to Carrier

Victor Reppert sent me this paper to which I have added some corrections, some formatting, a few hyperlinks, and some comments of my own. It receives the Maverick's imprimatur and nihil obstat and should be carefully studied by all.

Reply to Carrier

Richard Carrier has written a very extensive critique of my book, C. S. Lewis’s Dangerous Idea: In Defense of the Argument from Reason. So extensive, in fact, that when I printed it out it was 127 pages, compared to 128 pages of my book. Considering the fact that two of my chapters are not defenses of the primary argument of the book, this is rather remarkable. I am glad that he has put such an effort into responding to what I have written, although he obviously differs sharply with me on the merits of the arguments.

A quick note to begin with. A more careful reading of my book would demonstrate that I do not employ an argument from motive against naturalists, but only against those naturalists who insist, at all costs, that we cannot let a Divine Foot in the door. I realize that the quotation from Lewontin is sometimes generalized to be a statement about what all evolutionists or philosophical naturalists conduct themselves intellectually, but I have not made that kind of generalization. What I said was that when anyone, a naturalist or a non-naturalist, padlocks their belief system against even possibly considering an opposing view, then I suspect ulterior motives are work. How Carrier translates that into an Argument from Motive against naturalists is beyond me.

I should say that desires, either that God should exist or that God should not exist, unavoidably operate extensively in the minds of even those people who do their best to be rational about the matter. We should be conscious of the presence of ulterior motives on our own parts as well as those of our opponents, and should not presume that all ulterior motives are on one side of the question. But I would not accuse someone else of believing for ulterior motives unless I were convinced that their minds were padlocked against the possibility of accepting a contrary position.

The Scope and Limits of My Book

I must first raise some objections to what I consider to me a lack of regard for the scope and limits of my book. The book is relatively short, and written so as to be accessible to an audience of non-specialists. As such, it is not designed to be the final word on the arguments it presents. A lot more needs to be said than what I have said in the book. Those familiar with my original article on the Secular Web know that I originally wrote about one Argument from Reason. I could have simply developed one of the arguments from reason and left it at that, and if I thought my book would be the end of the discussion I probably would have done just that. But I see my own work as one link in a chain, going from Kant to Balfour to Lewis to Hasker to me and on forward to others. I have heard from a number of philosophy graduate students interested in pursuing these arguments. So the question one needs to ask in response to my book is not “Does Reppert conclusively refute naturalism?” Rather, one should ask “Has Reppert shown some promising ideas for criticizing naturalism that might give at least some reasonable people a good reason to reject it?” (Carrier seems to have made no mention of the discussion of Critical Rationalism in the chapter 2). But even by those standards, I am quite sure Carrier will say that, in light of the scientific and philosophical arguments he presents, the versions of the AFR that I defend are not very promising. I of course don’t agree, and that’s why I am writing this response.

Carrier’s disregard for the scope and limits of my book is further illustrated by his discussion of Pyrrhonian skepticism. I said, “If any thesis entails the conclusion that no belief is rationally inferred, then it should be rejected and its denial accepted.” Now it seems perfectly reasonable to me to rule out any world-view that renders rational inference (and therefore science) impossible, even if we have no idea of what world-view to put in its place. Carrier seems to be saying that if on all other counts my argument is persuasive, it might still be rational to reject some kind of anti-naturalist worldview by becoming a Pyrrhonian skeptic and adopting no worldview at all. I suppose, while I was at it, I should have refuted the Cartesian evil demon, or refuted the brain-in-a-vat scenario. Or maybe I should have thrown in a refutation of Parmenides and Zeno. If the only worldviews that permit rational inference are anti-naturalist ones, then either the discussion has to come to a screeching halt or anti-naturalism must be accepted. If Carrier were to tell me that he was going to stop arguing about atheism and instead seek enlightenment in a Zen monastery, I would probably not want to use the Argument from Reason to talk him out of it. This in no way undermines the task I have assigned to the Arguments from Reason, and that is to provide good reasons for preferring non-naturalistic worldviews to naturalistic ones.

But it gets even worse. Consider the claim
P) If any thesis entails the conclusion that no belief is rationally inferred, then it should be rejected and its denial accepted.
From this Carrier somehow thinks that this entails
P1) We should deny any premise that is not rationally inferred.
And he then argues that I must have a rational inference to support P, otherwise P is self-defeating. But this is preposterous. I never said you had to be able to rationally infer everything you believe; I said that we couldn’t accept a worldview that entailed that no one makes rational inferences. P1 in no way follows from P, nor does Carrier provide any argument to suggest that it does. This is symptomatic of a problem with Carrier’s review throughout: he puts such a heavy burden on someone advancing a case against naturalism that it can be no surprise that the case fails. But with more realistically calibrated requirements for success, perhaps my arguments will turn out a little more successful.

Carrier’s Three Underlying Problems

The first and most serious problem with my arguments is that I commit what he calls the Possibility Fallacy, that is, I assume that having no explanation is equivalent to not being able to have one. I mention this objection on p. 118, in the context of discussing Nicholas Tattersall’s critique of Lewis’s Miracles and Darek Barefoot’s response. I quote Barefoot’s reply in my book as follows:

Tattersall here confuses logical absurdity with phenomena incompletely known. To
learn why grass is green simply involves gathering more information. To learn how non-rational processes give rise to rational thought is like learning how a three-dimensional object can be created by arranging lines on a two-dimensional surface. We need not draw lines all day long in every geometric pattern imaginable to realize that the task is impossible. It is true that by means of perspective drawing we can usefully represent a three-dimensional shape, such as a cube, in two dimensions, just as human reason can be represented and communicated usefully by computer programs and even by humbler devices such as multiplication charts and slide rules. Nevertheless we can identify a set of lines in two dimensions as representing a cube only because we occupy three-dimensional space, and similarly we can appreciate that the blind functions of a computer have been so arranged as to accomplish a rational purpose only because, unlike the computer, we possess genuine rationality.

Carrier gives me two options for developing my argument. Either I prove conclusively that a naturalistic account of reasoning is impossible, or I conduct an exhaustive study of the finding of brain science and find that reasoning probably cannot be accounted for in terms of brain function. It seems to me that there is a third option available. I can show we are dealing with a conceptual chasm that cannot simply be overcome by straightforward problem-solving. An example would be the attempt to get an “ought” from an “is”. Moore argued that for any set of “is” statements concerning a situation, the question of whether this or that action ought to have been done is left open. To generate any confidence that you can get an “ought” from an "is," it simply won’t do to come up with one theory after another to show how you can get an "ought" from an "is." We need to be given some idea that these theories can surmount the conceptual problem Moore and others have posed.

Another way of putting my point is to say that reason presents a problem analogous to what David Chalmers called the hard problem of consciousness. When we consider seriously what reasoning is, when we reject all attempts at “bait and switch” in which reasoning is re-described in a way that makes it scientifically tractable but also unrecognizable in the final analysis as reasoning, we find something that looks for all the world to be radically resistant to physicalistic analysis.

So I maintain that there is a logico-conceptual chasm between the various elements of reason, and the material world as understood mechanistically. Bridging the chasm isn’t going to simply be a matter of exploring the territory on one side of the chasm.

BV: Very nice way of putting it, Victor.

Now someone might perceive the chasm and either think that some kind of paradigm shift in our thinking will bridge the chasm, or that it while it’s a mystery to us how all this is possible, that somehow there is a bridge over the chasm, even if we can’t see one that’s consistent with materialism. In pointing out the chasm, I do not necessarily claim that no possible considerations could persuade us to think that the chasm has been bridged. However, we have no reason to believe that the problem can be dissolved by doing just a little more science. Without necessarily demonstrating that the problem is insoluble, I can try to show that the problem is deep and intractable, and that an alternative to naturalism would resolve the problem.

BV: You seem to be distinguishing between insolubility and intractability. What's the difference? I see none.

And I should point out that lots and lots of naturalists, like Colin McGinn, think that there is a deep and intractable problem. The arguments from reason, in many cases, suggest that the descriptive discourse of physics cannot capture the normative discourse of reason. This presents a logico-conceptual gap, which is a very different kind of problem than pointing out something for which we don’t currently have a naturalistic explanation, and saying it must be supernatural because science can’t explain it now.

The second fallacy Carrier says I commit is the Causation Fallacy. He thinks that I, along with C. S. Lewis, endorse the argument that “the presence of a cause and effect account of belief is often used to show the absence or irrelevance of a ground and consequent relationship,” and that therefore all cause and effect accounts prove the absence or irrelevance of ground and consequent relationships. However, he claims in arguing thus I commit the fallacy of Affirming the Consequent, Hasty Generalization and Red Herring. (Quite a lot of fallacies to commit in one argument!) But while this type of argument was used in Lewis’s original chapter, in the revised edition this is not the primary reason for thinking that reason cannot be accounted for naturalistically. And I don’t argue that way either. The argument against the relevance of ground and consequent relationships has to do with the causal closure of the physical. If a physical account of the process is causally complete, and physics is mechanistic, how do reasons come into play? Remember, at the most basic level of analysis physics, in order to play the role of physics in the kind of physicalism that is under attack in my book, must be mechanistic. This being the case, if we apply the Principle of Explanatory Exclusion and the Principle of Causal Exclusion defended by Kim and others, a case can be made that a comprehensive physicalistic causal explanation excludes a mentalistic account of rational inference.

The third fallacy he thinks I commit is the Armchair Science fallacy. That, he says, is my failure to interact with the extensive philosophical literature and the findings of cognitive science. Here I would first like to reiterate that no book, especially a book that is supposed to be readable by non-specialists, can interact with every opposing thinker. Well, perhaps I should at least interact with the really important stuff. But what is that? I see no interaction in Carrier with John Searle, in spite of the fact that he is advocating positions that Searle has quite famously criticized. Nor do I see any discussion of Thomas Nagel, who maintains that an evolutionary account of our capacity to reason has always seemed to him to be laughably inadequate. Nor do I see any interaction with Lynne Rudder Baker’s and William Hasker’s critiques of eliminativism, which I explicitly reference in the book. Nor do I see any references to Jaegwon Kim’s work on mental causation and the Principle of Explanatory Exclusion. The literature is enormous on this subject, and a comprehensive treatment of the relevant issues would require a 12,800 page book with a 128 page bibliography. Carrier and I are bound to differ as to what is “really important.” Of course the argument needs to respond to what Hasker calls the “sensible naturalist.” But naturalists and anti-naturalists might very well disagree on exactly who the sensible naturalists are.

I do, in my book, say that naturalistic analyses of mind “invariably fail,” largely because they “sneak in” the very concepts they are trying to explain through the back door. They also tend to re-describe what they are trying to explain in terms that will make such things as consciousness and reasoning more tractable to naturalistic analysis, but this produces what I call a “subtle changing of the subject.” Instead of explaining their subject matter, they explain it away. Since I did say these things, it simply won’t do for Carrier to merely assert that there’s all this philosophical and scientific literature out there. He needs to show evidence that these analyses of mind don’t commit the two errors listed above. Carrier’s own treatment of intentionality is repeatedly guilty of the first defect, as we shall see.

Further, scientific work on cognition can be illuminating without actually solving the fundamental problems of the philosophy of mind that my book is concerned with. I never denied that the mind is not in many important ways dependent on the brain, it is just that the brain story cannot be comprehensive, if I am right. So I have no problem with the idea that scientists like Cattell can help us understand how different mental functions are linked with different areas of the brain. As I put it on pp. 114-115:

But again I would reiterate the claim that the arguments from reason require only that the account of mental functioning in terms of blind physical processes, operating in accordance with the laws of physics, rather than in accordance with the laws of logic, cannot be comprehensive. It seems to me to be perfectly compatible with an extensive dependence of the mind on the physical brain; it only says that if mechanistic accounts of rational inference are the only accounts you can get out of brain science, then a neurophysiological account cannot be complete.

Now I should make perfectly clear that the more defenders of AFR interact with naturalistic philosophical and scientific literature, the better. Here I would strongly recommend Angus Menuge’s recent analyses of the Churchlands and Dennett, along with the critiques of the Churchlands by Hasker, Baker, and myself that I endorse in my book. But if we have what David Chalmers would call a “hard” problem of intentionality, then simply knowing about correlations between intentional states and physical states will not solve the problem. What science would have to provide is a successful intertheoretic reduction, and this is something over and above what straightforward neuroscience provides. What we need are answers to questions that are fundamentally philosophical rather than scientific, questions that are perhaps better addressed in armchairs rather than in laboratories.

Here again I would just reiterate what I said earlier, that the problem with naturalistic analyses of reasoning has to do with a logico-conceptual gap between our ordinary conception of what goes on when we reason, as opposed to what has to be given in a properly mechanistic analysis of that same process. If you give an analysis of how the brain produces some output or other, that may not do the job if when we get does the thing that has been analyzed physicalistically cannot plausibly be described as reasoning. Further, many people in the philosophy of mind, even some who would be described as card-carrying physicalists, often maintain that we have no real understanding of how the mental and physical are related to one another. Physics describes things from a third-person, non-normative, mechanistic point of view, while we describe our own thought processes from a first-person, normative perspective. Hence, they maintain, mental events are irreducible to physical events, nevertheless they are either token-identical to physical events or supervene upon physical events. However, they also often say that they perceive the relation between the mental and the physical as being deeply mysterious.

Non-Reductivism and Beyond

One strategy in responding to the versions of the argument from reason that I have presented is to admit that these realities are profoundly mysterious in a naturalistic universe, but that providing a supernatural explanation for them is unacceptable. Keith Parsons sums it up when he says:

Physicalists may have to admit that some mental phenomena are mysteries and
likely to remain so. Consider consciousness. How consciousness can exist in the physical world remains the “world-knot,” as Schopenhauer called it, and the expressions of despair quoted by Reppert are apt. But the honest thing to do when we confront an insoluble mystery is to admit that we do not know. It is obscurantist to “explain” the mystery in ways that only deepen our ignorance.

In other words, yes, we can’t reduce reason to the non-rational activities of physical particles, but to accept a theistic account of these is obscurantist. His quarrel is going to be with chapter 6 of my book, entitled “The Inadequacy Objection.”

Now Carrier, to be sure, is not a non-reductive materialist. He thinks that mental-physical reductions are indeed possible. But I think his discussion of the relevant literature underestimates the sizable group of people in the philosophy of mind who are philosophical naturalists but who eschew reductionism. The “sensible naturalist” of Hasker’s reply to me is clearly a non-reductive materialist, who tries to stay within the naturalist fold while not holding out any promise for, for example, a causal analysis of mental states.

Then there are other people who think mind’s being what it is requires a change in metaphysics, such the mind is fundamental to the universe and is not an evolutionary by-product. However, these thinkers are reluctant to accept theism as the solution to the problems they pose. Such people include Daniel Hutto, who has proposed Bradleyan absolute idealism as the solution to the hard problem of consciousness, and of course Thomas Nagel. In fact Nagel’s The Last Word is pretty much a defense of the Argument from Reason, the difference being that he does not offer God as his solution of choice. Reason for Nagel, is fundamental to the universe, but not the reason of God. The general thrust of Nagel’s book is to show the importance of rejecting accounts of reason that make reason relative, but, once you do that, you end up having to accept a metaphysics that, as he puts it “makes us more at home in the universe than is secularly comfortable.” However, Nagel has traveled outside the realm of what people like Carrier would consider to be naturalistically acceptable and adopted what I call explanatory dualism, the idea that in addition to mechanistic explanations at the most basic level, we must have rational explanations as well.

Now Carrier’s strategy is very different. He spends little energy trying to show that even if naturalistic explanations of reason are unsuccessful, offering a theistic account of the phenomenon of reason does nothing to alleviate our ignorance. He instead maintains, not that I have overestimated the power of theistic explanations, but that I have underestimated that power of naturalistic explanations. (Of course one can argue both that naturalistic explanations are adequate and that theistic explanations are inadequate, but neither Carrier nor Parsons actually do both.) He thinks I have failed to pay sufficient attention to analyses of mind that are on offer.


Consider his treatment of intentionality. Carrier’s task is to show that you can build and intentional brick wall out of non-intentional bricks. Just as a brick wall can be six feet tall even though none of the bricks are, a state can be intentional even though the fundamental, underlying states are non-intentional, as is required by the understanding of naturalism that both of us accept.

But what does Carrier say about intentionality? He says that a material state A is about material state B just in case “this system contains a pattern corresponding to a pattern in that system, in such a way that computations performed on this system are believed to match and predict behavior in that system.”

Unfortunately, this analysis of intentionality is simply loaded with intentional concepts, so if we didn’t know what intentionality was before we heard from Carrier, we wouldn’t know now.

BV: This is a very important point, muchachos; meditate on it carefully.

Moreover, on its face, it doesn’t even come close to being a physicalistically acceptable concept of intentionality that analyzes intentionality in non-intentional, physical, terms. The problem that should jump out of the page at you when you read this is that the definition is literally loaded with concepts that make sense only if you understand what it is for something to be about something else, in short, if you know what intentionality is. Consider the term “corresponds.” What does “corresponds” mean in this context? If I’m eating a pancake, and the piece of pancake on my plate resembles slightly the shape of the state of Missouri on the map, can we say that it corresponds to the state of Missouri; that it is a map of Missouri? I’m looking at bottle trees right now. Is each of the bottle trees about the other bottle trees because there is a “correspondence” of leaves, branches, bark and roots, one to the other? In order for “correspondences” to be of significance, doesn’t it have to be a “correspondence” recognized by somebody’s conscious mind as being “about” the thing in question? And if that’s the case, then are we anywhere in the vicinity of a naturalistic account of intentionality?

BV: I think it helps to distinguish between original and derivative intentionality. A map is about a chunk of terrain. Its intentionality or aboutness is derivative from original acts of intentionality whereby I assign certain marks on the map to certain geographical features. Thus I or we assign to contour lines that are close together on a topographical map the meaning: steep terrain. The map by itself cannot mean or intend or be about anything. The same goes for extremely complex systems like this Pentium IV sitting in front of me. It is not thinking in any serious sense at all.

What the naturalist cannot account for is original intentionality.

And then there’s more to the definition than that. The intentional state has to be believed to correspond. But how could we define belief if we didn’t have any idea what it was for a mental state to be about something? If I have to believe that brain state X is about object Y only if I believe it to correspond to Y, then how do we analyze my belief that there is a correspondence without throwing us into an infinite regress? Presumably this is explained by a “choice,’ though this doesn’t have to be a conscious choice. Decisions, presumably can be done by computers, even without intentionality. But do decisions generate beliefs? “The core engine of intentionality derives from the attentional centers of the brain. That’s why cats can keep track of their prey, for example—by the same means, we can track the image or thought of, say, our uncle, by attending to it cognitively, a process well understood in neurophysiological terms.” But surely one can “track” something without thinking “about” it. A heat-seeking missile tracks its object, but surely we don’t want to say that it’s activities are in any way intentional. Does a thermostat make decisions about what number to show? In dealing with questions of mind you have to be awfully careful to make sure that the words people are using really do mean what you think they mean, or whether they have been subtly re-defined to make them tractable to a physicalistic account.

BV: Reppert is right. It is easy to be misled by language. A thermostat can be described as a sensor. But obviously it is not sentient strictu dictu. It cannot feel anything. Clinton can feel your pain, but no thermostat feels the heat or the cold. Both intentionality and qualia are beyond the reach of naturalistic explanations.

Another highly ambiguous term, though quite popular in the philosophy of mind, which Carrier uses in his account of intentionality, is the term “computation.” But what does that mean? Chris Eliasmith suggest that there are serious problems with all definitions of intentionality currently on offer. He writes:

There are numerous competing definitions of computation. Along with the initial definition provided here, the following three definitions are often encountered:
Rule governed state transitions
Discrete rule governed state transitions
Rule governed state transitions between interpretable states
The difficulties with these definitions can be summarized as follows:
Admits all physical systems into the class of computational systems, making the
definition somewhat vacuous
Excludes all forms of analog computation, perhaps including the sorts of
processing taking place in the brain.
Necessitates accepting all computational systems as representational systems. In
other words, there is no computation without representation on this definition.

So how does Carrier want to define computation in his account of intentionality?

In short, I just don’t see it. C. S. Lewis wrote an essay in which he delineated the difference between “looking at” and “looking along.” When you look at something, you view it from a third-person, outside perspective. When you look along something, you view it from within. An attempt to come up with a physicalistic view of the mind invariably ends up looking “at” mental events, and always fails to capture what is going on when you look “along” those same events, as the thinking subject. But these are not merely the musings of a popular Christian apologist of the last century. Philosophers of the stature of Thomas Nagel, Frank Jackson, and John Searle have, in essence, said the same thing.