The argument as I will state it is somewhat different from anything either Popper or Putnam has said, though it is in the same spirit. Before stating the argument, it is worthwhile recalling the “mechanistic” conception of the natural world which, as I have emphasized in earlier posts in this series, implicitly or explicitly informs materialism. On this conception, the world is devoid of what Aristotelians call formal and final causes: there are in nature no substantial forms or inherent powers of the sort affirmed by the medieval Scholastics, and there is no meaning, purpose, or goal-directedness either. The physical world is instead composed entirely of inherently purposeless elements (atoms, corpuscles, quarks, or whatever) governed by inherently meaningless patterns of cause and effect. All the complex phenomena of our experience, from grapes to galaxy clusters, from mudslides to minds, must somehow be explicable in terms of these elements and the causal regularities they exhibit.
But in fact there can be no such explanation of the mind, not even in principle. In particular, there can be no such explanation of intentionality, the mind’s capacity to represent the world beyond itself – as it does, say, when your thought that the cat is on the mat represents the cat’s being on the mat.
The reason is this. As already indicated, any materialistic explanation of intentionality is bound to be a causal explanation. That is to say, it is going to be an attempt to show that the intentionality of a mental state somehow derives from its causal relations. The causal relations in question might be internal to the brain (as they are according to “internalist” theories of meaning); they might extend beyond the brain to objects and events in a person’s environment (as they do according to “externalist” theories); they may even extend backwards in time millions of years to the environment in which our ancestors evolved (as they do according to “biosemantic” theories). An adequate description of the relevant causal relations may require any number of technical qualifications (such as an appeal to Fodor’s notion of “asymmetric dependence”). In every case, though, a materialist is bound to appeal to some pattern of causal relations or other as the key to explaining intentionality. He’s got nothing else to appeal to, after all; the basic elements out of which everything in the physical world is made are by his own admission devoid of any meaning (“intentionality simply doesn’t go that deep,” as Fodor insists in Psychosemantics) and anything other than these elements exists only insofar as causal interactions between the elements generates it.
Now, specifying the relevant causal relation entails specifying a relevant beginning point to the series and a relevant end point. We have to identify some physical phenomenon as that which does the representing, and some other physical phenomenon as that which is represented; or in other words, we have to pick out one thing as the thought, and another thing as that which is thought about. To take a simple example, if we imagine that a certain brain process is associated with the thought that the cat is on the mat because it is caused in such-and-such a way by the presence of cats on mats, then we will have to take the cat’s presence on the mat as the beginning of the relevant causal chain (call it A) and the occurrence of the brain process in question (call it B) as the end. (Of course, specifying exactly what the “such-and-such a way” involves can get pretty complicated, as anyone familiar with the contemporary literature knows, but the complications are irrelevant for our purposes here.)
But what objective reason is there to identify A and B as “the beginning” and “the end” of a causal sequence? Consider what happens in a situation like the one in question. Someone flips on a light switch, which causes electrical current to flow through the wires in the wall up to a ceiling lamp. Light from the lamp travels to a cat sitting on a mat below, is reflected off of the cat, and travels to the retinas of a nearby observer. This in turn causes signals to be sent up the optic nerves to the brain, which results in the firing of a certain cluster of neurons, which in turn results in the firing of another cluster, which in turn results in the firing of yet another cluster, and so on and so forth. All this neural activity ultimately results in a behavioral response, such as walking over to the refrigerator to get the milk bottle out so as to give the cat a snack. And this is followed, say, by an accidental dropping of the milk bottle, which results in broken glass, a cut to the ankle, a yelp of pain, and the kicking of the cat.
Now, again, what is it about this complex chain of events that justifies picking out A and B specifically and labeling them “the beginning” and “the end” respectively? Why is it the cat’s presence on the mat that counts as “the beginning” – rather than, say, the flipping of the light switch, or the flow of the current to the ceiling lamp, or the arrival of such-and-such a photon at exactly the midpoint between the surface of the cat and the observer’s left retina? Why is it brain process B exactly that counts as “the end” of the causal chain – rather than, say, the brain process immediately before B or immediately after B, or the walk over to the refrigerator, or the motion of such-and-such a shard of glass from the broken milk bottle as it skips across the floor? Of course, we have an interest in picking out and identifying cats and not in picking out and identifying individual photons, and an interest in brain processes and their associated mental states that we don’t have in shards of glass. But that is a fact about us, not a fact about the physical world itself. Objectively, as far as the physical world itself is concerned, there is just the ongoing and incredibly complex sequence of causes and effects, which extends indefinitely forward and backward in time well beyond the events we have described. Objectively, that is to say, there is no such thing as “the beginning” or “the end,” and nothing inherently significant about any one event as compared to another.
Popper’s point, and Putnam’s, is that what count as the “beginning” and “end” points of such a causal sequence, and thus what counts as “the causal sequence” itself considered in isolation from the rest of the overall causal situation, are interest relative. These particular aspects of the overall causal situation have no special significance apart from a mind which interprets them as having it. But in that case they cannot coherently be appealed to in order to explain the mind. It is no good saying that the representational character of our mental states derives from their causal relations when the causal relations themselves cannot be specified except in terms of how they are represented by certain mental states. A vicious circularity afflicts any such “theory” of intentionality.
Now it is important to emphasize that the point is not that causation per se is interest relative or mind-dependent; the argument is not an exercise in idealism or anti-realism. The overall complex ongoing sequence of causes and effects is entirely mind-independent. The claim, again, is just that something’s counting as a “beginning” or “end” point within the series is interest-relative and mind-dependent. Still, even this much might seem to be too close to idealism or anti-realism for comfort. It might seem to make causal explanations somehow subjective and arbitrary. (Indeed, Putnam attributes something like this sort of objection to Noam Chomsky.) But to fear that the Popper/Putnam argument we’ve been considering might entail that causal explanations are somehow subjective or arbitrary doesn’t show that the argument is wrong.
Is there any way to reconcile the argument with the objectivity and non-arbitrariness of causal explanations? Absolutely. The way to do it is to show that certain physical phenomena really can objectively count as the beginning or end points of a causal sequence after all – that they can indeed be picked out in a way that is not mind-dependent or interest-relative. But how can that be done? By showing that natural objects and processes are by their natures inherently directed towards the generation of certain other natural objects and processes as an “end” or “goal.” That is to say, by showing that natural objects and processes have what Aristotelians call substantial forms and final causes. In short, the way to explain how causal explanations can be objective and non-arbitrary as opposed to subjective and interest-relative is to acknowledge that the mechanistic conception of the world is mistaken, and that the Aristotelian-Scholastic conception that it replaced is correct after all.
So, the Popper/Putnam argument shows that the mechanistic conception of nature to which materialists are explicitly or implicitly committed entails that there can be no materialistic explanation of the mind. (As we have seen in earlier posts in this series, other arguments tend to show the same thing.) And the only way to sidestep the argument is to abandon the mechanistic conception of nature, which entails rejecting materialism anyway. Either way, materialism is refuted.
What positive view results? That depends. If one holds on to the mechanistic conception of nature, the result would seem to be some broadly Cartesian form of dualism – either substance dualism or property dualism. (Popper himself opted for the former. Putnam does not consider what consequences his view might have for the dualism/materialism debate.) If instead on opts to return to an Aristotelian conception of nature – the right choice, in my view – then one is on the path toward hylemorphic or Thomistic dualism. (I examine these options in my book Philosophy of Mind and defend the latter at length in The Last Superstition.)