Metaphysical Naturalism and Consciousness

Frequent correspondent Dianelos Georgoudis recently responded to my request that he present three conceptual problems with metaphysical naturalism. I had planned to address all three in a single post, but the issues are tangled and multifaceted, so I really only managed to address his first one—and this is still a very long post. The other points will have to wait until later.

I think, as is so often the case, one source of our disagreement is that we have different understandings of the terms. Metaphysical naturalism (MN) is a hard term to define precisely. The metaphysical naturalist (hereafter, just “naturalist”) is most definitely not committed to the view that current science offers a complete and satisfactory description of reality. Clearly, it doesn’t. Hence, if quantum mechanics and relativity clash at some level, or if QM seems to imply irresolvable paradoxes, this is a problem for physicists—or rather, perhaps, philosophers of physics—but not for MN or the naturalist qua naturalist. The naturalist affirms that substantial (non-abstract) reality is in some sense physical or natural, that is, that it is constituted exclusively of entities that act and are acted upon in accordance with regular, impersonal, and universal laws. The naturalist is not committed to the view that we now have an exhaustive and accurate understanding of those entities or the laws that govern them. On the contrary, our understanding of the universe and its constituents is (and may always be) a work in progress.

At bottom, I think that what really motivates MN is the idea that nature is a closed system. That is, that whatever happens in the natural world is explicable in terms of natural entities, forces, and processes, and that there is no need to invoke supernatural beings, powers, or occurrences, such as gods, souls, or miracles to account for anything that happens in the universe. So, a naturalist might not really be averse to the idea of impersonal supernatural entities, such as Platonic ideas, so long as those purported entities are sequestered within “their own realm” and cannot act upon or influence the phenomena of the physical universe. In short, naturalists endorse the vision of the first philosophers of the Western tradition, the Milesians of the 6th Century B.C.E.: That the arche of things is to be found in physis itself, that is, that the origin, primordial stuff, and underlying principles of natural things is to be found in nature, not in myths about gods, ghosts, and gobbledygook.

Why would anyone accept MN? Speaking personally, I am extremely suspicious of any sort of positive assertions about “ultimate reality.” I think when we venture in that direction we are in extreme danger of having our thinking become perversely “dialectical” in Kant’s sense, that is, we begin to use concepts that have meaning in empirically constrained contexts, and try to apply them where there are no such constraints. The consequence is that we quickly fly off into conceptual cloud-cuckoo-land where nobody knows what he is talking about and philosophers become fools. For instance, I have no idea at all what it could mean to say, as some religious philosophers do, that some putative ultimate reality is more objectively probable than another. The reason is that it is meaningful and useful to speak of the objective probability of things or events where we have some sort of experience or theory to guide and constrain us, but when we have prescinded our probability judgments from any such background or context, such judgments can only be pretentious expressions of ignorance or bias.

Therefore, I am a naturalist by default. I know that physical entities exist. I know that science has made enormous progress with theories that postulate only natural entities, and that over the history of science supernatural hypotheses have been ruthlessly driven out in favor of naturalistic ones. I do not know that supernatural entities exist, and it is my judgment that the purported evidence for their existence ranges from the shoddy to the risible. Natural theology is a failed research program and, as for ghosts, demons, and supernatural bugaboos in general, for over thirty years the Skeptical Inquirer has had good fun debunking those claims. Therefore, if forced to place my bets about “ultimate reality” I’ll wager that it is physical, and I shall continue to look for physical explanations of physical phenomena. Also, I put the burden of proof on anyone who says that there is a problem with taking reality to be physical. Does Mr. Georgoudis meet that burden?

His first objection to MN is what he calls “the problem of consciousness”:

“Consciousness is a scientifically unnecessary concept, because it does not refer to something observable, nor is the hypothesis that it exists necessary for explaining anything observable. Thus consciousness lies beyond the scientific field of investigation. According to scientific naturalism nothing exists but the objects of science, which implies the non-existence of consciousness. But consciousness clearly does exist, and cannot be construed to be an illusory concept.”

What is consciousness? Nobody can define it and the whole confused notion is a conceptual minefield. We have to be very careful how we speak or we get inextricably mired in a quagmire of obfuscation. First of all, consciousness is not a thing, we must not reify it and try to think of it as some sort of concrete being or entity (I think such reification is at the core of Cartesian dualism). Consciousness is not a ghost in the machine, or a homunculus sitting inside of us watching and directing all that we do, as in the hilarious bit from Woody Allen’s Everything You Always Wanted to Know about Sex. We therefore do not expect that consciousness will be observable in the same way that a horse is, or even in the same way that a chromosome is. “Consciousness” is a term that, in ordinary usage, encompasses a very large number of phenomena, states, and processes. Phenomena like feeling, states like self-awareness, and processes like thinking through a problem all occur consciously, or, at least, some aspects of them occur consciously. Can any of these multifarious items be observed or detected in any sense?

Of course. What could be more observable? We enact states of consciousness constantly in our own first-person cases, and we have all sorts of ways of accessing other people’s conscious states indirectly. Is the supposed problem the fact that we can only observe other people’s conscious states indirectly, rather than immediately? I cannot literally feel your pain, but I have all sorts of ways of knowing when you are in pain (e.g., your report through gritted teeth that you are in pain). In fact, I can get quite a bit of detailed information about just what kind of pain you have. The fact that pain detection is indirect doesn’t mean that pain is not an appropriate object for scientific study. Most of the items that science studies are detected indirectly (e.g., neutrinos). It is a common mistake (often committed by creationists, BTW) to think that science deals with observables. What science requires is not direct observability but that we have reliable empirical access to phenomena, and that access is often quite mediated and roundabout.

Mr. Georgoudis says that consciousness is a “scientifically unnecessary concept.” What does this mean? Does it mean that the phenomena we explain by that concept could be explained more economically by other concepts, or that scientific inquiry has invalidated that concept, or that the concept reduces to a more basic concept? Of course, the concept of consciousness is logically unnecessary, but so are the past and an external world. The phenomena we explain in terms of consciousness (except in our own first-person case) could conceivably be due to unconscious causes. Likewise, our apparent perceptions of external objects could conceivably be due to delusions induced by a Cartesian evil genius. Similarly, there is no contradiction in supposing that everything, including every apparent trace of a past, could have been created ten seconds ago. It is conceivable that my friend’s expressions of outrage or empathy are not due to feelings of outrage or empathy, but are instead the programmed output of a cleverly constructed but unconscious android. Hence, the famous philosophical problem of other minds. But that the nonexistence of something is a logical possibility does not mean that it is something that should be neglected by science or common sense. You might be nothing but a brain in a vat, but you still had better get your doctor to check that suspicious looking mole.

Are there any episodes from the history of science where inferences to the best explanation have led to the postulation of conscious phenomena, states, or processes? Sure there are. Plenty of them. In their terrific book Baboon Metaphysics, Dorothy L. Cheney and Robert M. Seyfarth give some quite fascinating examples (pp. 7-9). In the early heyday of behaviorism, psychologists Edward C. Tolman and Otto L. Tinklepaugh (real name; bet he had a tough time in junior high) did some studies of monkeys that challenged the behaviorist tenet that mental states are private and unmeasurable and so not amenable to scientific study. Tinklepaugh had monkeys watch as either lettuce or banana was placed under one of two cups, the other cup being left empty. The monkeys were taken away for a few minutes and then returned and allowed to choose a cup. They all chose the one with food, though with much greater enthusiasm when they had seen the banana placed under the cup.

Suppose the monkey sees lettuce placed under the left-hand cup before being taken out of the room. Most non-behaviorists would say that the monkey has learned that there is lettuce under the left-hand cup. Behaviorists, though, denied that the monkey had a mental state of “knowing that there was lettuce under the left-hand cup.” Their claim was that the lettuce merely served to reinforce the link between the stimulus (the sight of the cup) and the response (looking under the cup). They denied that the monkey had learned anything about the lettuce at all. To test between the behaviorist and the “common sense” interpretations, Tinklepaugh did another experiment. This time he let the monkey see banana being placed under the cup, but when the monkey was out of the room, he replaced the banana with lettuce. When the monkey was returned to the room, here is what happened:

“Subject rushes to the proper cup and picks it up. Extends hand towards lettuce. Stops. Looks around on floor. Looks in, under, and around cup. Glances at other cup. Looks back at screen. Looks under and around self. Looks and shrieks at any observer present. Walks away, leaving lettuce untouched on floor (p. 8).”

I’d say that was one pissed off and disappointed monkey. After citing other more rigorous tests of this nature, Cheney and Seyfarth conclude:

“The results of these experiments challenge the more extreme behaviorists’ view that mental states like knowledge, beliefs, or expectations cannot be studied scientifically and may even be an illusion. Instead, they support Tolman’s view that learning allows an animal to form a mental representation of its environment. Through learning, animals acquire information about objects, events, and the relation between them. Their knowledge has content, and this content can be studied scientifically (p. 9).”

In fact, there are many books that report on scientific studies of consciousness. Adam Zeman’s Consciousness: A User’s Guide is a good one. There is even an Association for the Scientific Study of Consciousness. Check their website. Maybe Mr. Georgoudis means to claim that such studies can only address what David Chalmers calls the “easy problem” about consciousness, viz., how brain processes and conscious states relate, but cannot address the “hard problem,” that is, to explain why there should be phenomenal consciousness at all. This is the famous “explanatory gap” between information about brain states and processes, however complete, and the fact that things feel a certain way. We might explain in complete detail the underlying neurological processes involved in feeling pain, but this does not tell us why it hurts when such processes are occurring. Similarly, Frank Jackson tells the story of one hypothetical “Mary” who has complete scientific knowledge about the physiology of color perception. She can tell you precisely what occurs in your head when you have the experience of seeing a red rose. Yet Mary herself has been raised in a purely black-and-white environment. Then, one day, she has the experience of seeing a red rose. Only then does she discover what it is like to see a red rose; all of her scientific knowledge could not tell her that. Then there is the question Thomas Nagel put to us: What is it like to be a bat? We could have complete knowledge of the bat’s faculty for echolocation, but that would still not tell us what it phenomenally feels like to echolocate like a bat. Don’t all these puzzles show that natural science can never explain science and that we have to look beyond nature for an adequate explanation?

Let’s ponder this question a bit more: Why do so many thoughtful people (many of them metaphysical naturalists) think that neuroscience can never explain consciousness? Because it seems to them that even a completed neuroscience cannot tell us what we really want and need to know. It seems that neuroscience can only succeeded in displaying deep correlations between brain states and states of our phenomenal consciousness without explaining how anything like, e.g., pain, ideas, or blueness could come from anything like brain processes. At some point in a ground-up scientific explanation, consciousness will be seen to emerge within a physical system, and some philosophers will still see a vast, unbridgeable explanatory chasm here. Science, it appears, may someday tell us how states of consciousness occur by giving us an exhaustive list of their physiological correlates, but we cannot expect an account of why conscious states would be expected to arise from those functions of physical systems. In short, even at the end of the day neuroscience seemingly can only give us correlations, not explanations, only hows, not whys.

However, as William Lyons observes (Lyons, Matters of the Mind, p. 204), there is no sharp dividing line between “how” explanations and “why” ones. As he notes, the answer to the query why, given that metal expands when heated, railway lines do not buckle in hot weather, will explain why by detailing how the lines are constructed and laid so as not to buckle as they expand (Lyons, 205). In this case, explaining how is also explaining why, and there seems to be no reason why, in principle, this should not also be the case with the explanation of consciousness:

“It seems likely that if scientists could explain how, neurophysiologically and biochemically, the parts of our brain causally relevant to consciousness are put together and function in relation to one another, then ipso facto the scientists will have explained how those parts function together so as to generate consciousness (Lyons, 206; emphasis in original).”

Why is there consciousness? Because there are brains that are organized and function in these particular ways; this is how they work and that is why there is consciousness. The feeling that in this case, as opposed to the railway tracks case, a how-explanation still leaves us wondering “But why?” may really be no more than that—a feeling. The neuroscientific account, when and if it is completed, may leave us with a sense of mystery, but that may be a feeling that we will eventually get over. On numerous occasions in the history of human inquiry, certain questions have been dropped, not because they were answered, but because there no longer seemed to be any point in asking them.

But the whole discussion so far has failed to ask two most pertinent questions: (1) Why should MN have the burden of explaining consciousness? (2) Is there any prospect that any ontology inconsistent with MN, say Cartesian dualism, could do a better job of explaining consciousness? Let’s take the second question first. To say that we think because we possess a soul, a “thinking substance” is most emphatically not to explain thought or any other aspect of consciousness. Such an “explanation” is really just like the old joke from Moliere about explaining the sleep-inducing effects of opium by saying that it has a “dormative potency.” You don’t explain the phenomenon; you just pretentiously and tendentiously re-name it. Souls explain consciousness no better than matter does; a lot worse, in fact. By definition, we can have no inkling about how an incorporeal posit like a soul is supposed to operate. Its functioning, in principle, is entirely occult. At least with matter we can, in principle, have “how” explanations of consciousness.

Well, why not just postulate consciousness as a primitive, as the idealists did? That is, consciousness is taken as a, maybe the, basic reality with “matter” only an epiphenomenon of mind. Because one of the most obvious things we know about consciousness is that, so far as we can tell, it occurs only in association with the physiological functioning of very complex organic entities that appear quite late in the evolution of the universe. We know this with as much certainty as we know any scientific fact. We know it as a scientific fact that at one time, probably quite recently in the geological sense, nothing on earth was conscious. Were trilobites conscious? Dinosaurs? Homo erectus? Further, even if consciousness has its own causal powers, and is not merely epiphenomenal, clearly its existence and nature are dependent, wholly and in detail, on quite subtle physical conditions. This takes us back to question (1).

In response to question (1), MN does not have the burden of explaining consciousness; it is quite sufficient to know (which we do) that consciousness is wholly dependent upon the quite precise functioning of certain physical processes. In fact, all of the evidence indicates that particular physiological changes in the central nervous system are sufficient for changes in consciousness. Quite subtle changes in your brain chemistry can turn you (yes, you) into a raving lunatic, sex addict, or gun-totin’ goober who thinks that Barack Obama is not a natural born U.S. citizen. The books of Oliver Sacks are quite instructive on these points. He shows with great eloquence how brain disease and trauma can change consciousness in quite remarkable and sometimes quite terrifying ways. One of his stories tells of a man, an artist, who suffered a head injury and could no longer see colors. He not only could no longer see colors, he could not even form a mental image of colors. He was like Frank Jackson’s “Mary.” Less frightening is Sacks’ story about the 90 year-old woman who began to feel young again, even with a renewal of libido. Turns out that she had contracted syphilis when young, and, after seventy years of dormancy, the spirochetes had become active in her brain. Sacks killed the spirochetes with penicillin and left her with her perky feelings. When anything, from spirochetes to speed, changes your brain functioning, it changes YOU.

Paul Churchland has a fascinating section in The Engine of Reason, the Seat of the Soul about the physiology of taste perception. The qualitative difference between two very different tastes, say limburger cheese and ripe peach, comes down to different levels of stimulation of a quite small number of different kinds of taste receptors on the tongue. Somewhat simplifying, if there are ten different levels of excitation for each of four different kinds of receptors—sweet, sour, bitter, and salty—then there will be 104 or 10,000 different states of excitation, which we can represent by ordered quadruples from (1,1,1,1) to (10,10,10,10). These will correspond to 10,000 distinct and discernable qualitative differences in “taste space.” Thus the taste of limburger cheese would be something like a (2,2,6,7) state of stimulation for the sweet, sour, bitter, and salty receptors respectively, and ripe peach would be (8,6,2,1) or something like that. Now, we might still insist that such correlations tell us nothing about why those particular tastes, or any taste at all, would always be associated with those, and only those, levels of cellular stimulation. But they are. This is just as MN would expect, and that is the crucial point. By the way, Cartesian dualism has no way of explaining why the features of consciousness should be correlated in such a comprehensive and fine-grained way with the detailed operations of wetware. As far as Cartesian dualism or idealism can say, we should be capable of having the same sort of consciousness we presently enjoy if our skulls were full of lime Jell-o. I cannot think of anything the dualist or the idealist could say here that would not be dreadfully ad hoc (or, less politely, ass-covering).

The upshot is that I do not see that the phenomenon of consciousness presents any special problem for MN. There may be aspects of consciousness that will always be a mystery, or at least will feel mysterious for us, but I do not see how any alternative to MN helps in that regard. I’ll end with a challenge: Show me any instance in which two identical states of an organism’s central nervous system are correlated with two distinctly different states of consciousness, and I’ll have to admit that the phenomena of consciousness constitute a legitimate challenge to my worldview.

About Keith Parsons
  • smijer

    As far as Cartesian dualism or idealism can say, we should be capable of having the same sort of consciousness we presently enjoy if our skulls were full of lime Jell-o. I cannot think of anything the dualist or the idealist could say here that would not be dreadfully ad hoc (or, less politely, ass-covering).

    Is it ass-covering to suggest that lime Jell-o is incapable of activating motor neurons and cueing muscular phenomena?

    Not that I am a dualist. I really appreciate this article.. I'm just nit-picking. Whenever I see "I cannot imagine" an answer from the other side, I always try to imagine to see if I can do something the other fellow couldn't… ;)

    P.S. I get really aggravated that Blogger doesn't allow the blockquote tag. I cannot imagine a good reason for them disallowing the use of it.

  • Philip

    Whether consciousness is a problem for metaphysical naturalism (MN) obviously depends on how “MN” is defined. Keith Parsons says such a naturalist is committed to saying “that substantial (non-abstract) reality is in some sense physical or natural, that is, that it is constituted exclusively of entities that act and are acted upon in accordance with regular, impersonal, and universal laws.”

    This doesn’t seem sufficient to me to rule out ghosts and other putatively supernatural entities. As long as ghosts emerge from physical interactions and their ghostly properties depend on physical ones, and as long as ghosts operate in ways that follow some laws, that is, as long as there could be ghosts given supervenience physicalism, saying they exist is consistent with MN, as stated.

    (Saying that the laws have to be “universal” is redundant. Even if the laws are ceteris paribus, the laws must be universal in that were the conditions met anywhere, at any time, the nomic relation would be instantiated. And saying they have to be “impersonal” is confusing. Laws about people, such as psychological or social laws, would be personal with regard to the laws’ content. Moreover, all laws, as opposed to the nomic relations themselves, are personal in that they’re statements made by scientists who are persons. What matters is that the laws be descriptive rather than prescriptive, that they be generalizations made by scientists rather than commandments of a creator of the phenomenon.)

    Obviously, saying that substantial reality has to be “natural” doesn’t help define “MN,” since the question is about the meaning of “MN.” Also, saying that this reality has to be “physical” doesn’t help, since if the claim is that the physics isn’t present-day physics, then once again ghosts may be physical, given future physics. So ghosts, goblins, and so forth would come out as natural.

    I think “MN” had better mean, in part, that nature is objective, that what really, not just “substantially,” exists is what can be scientifically explained and measured, in which case the subjective viewpoint of consciousness would be, not just an emergent phenomenon, nor a higher-level, irreducible but supervenient one, but something illusory, apparent, and thus supernatural. This is because consciousness couldn’t be scientifically explained. Oh, the brain and whatever objective process is correlated with consciousness could be publicly observed and thus scientifically explained, but the private, subjective viewpoint seems precisely what can’t be explained using known scientific methods. Of course, science can acquire radically new methods, in which case “MN” is vacuous. And if the reality-appearance distinction is brought in, the subjective viewpoint seems ruled out by definition, since this viewpoint is precisely what deals with appearances (qualia) rather than with objective reality.

  • Keith Parsons

    Philip's objections to my post are captious and shallow at best. He thinks that ghosts could still exist given my characterization of MN because, after all, ghosts might somehow emerge from physical interactions. Hey, Philip, Maybe God does too, huh? Why not? Of course, somebody determined to obfuscate can always play the Humpty Dumpty game and arbitrarily redefine terms. For instance, we could redefine "ghost" or "god" so that they might somehow come out physical. This is a very old trick of philosophers when they are more interested in being a pain in the ass than in having an intelligent debate.

    Philip also objects that my characterization of the laws governing physical entities as "universal" is redundant, because laws, by definition, are universal. Universality may be, and in my opinion is, a defining characeristic of the laws of nature. Universality is therefore a feature that can be used to pick out laws of nature and distinguish them from other types of (non-nomic) regularities like accidental generalizations or correlations. The universality requirement is that a law of nature must be expressible in the form of a counterfacutal conditional e.g., "If condition C were instantiated, event E would follow." Saying that the laws that govern physical entities are "universal" is merely a way of unpacking the concept a bit and making sure that the distinction between nomic and non-nomic regularities is put front-and-center. Of course, someone so deeply conversant with the philosophy of science as Philip might not need such a reminder, but perhaps a less erudite reader could use a nudge.

    Philip also is confused by the fact that I said that such laws must be "impersonal." There could be regularities that are universal, but personal in the sense that they depend upon the choices of conscious beings. For instance, God might choose to perform a miracle every time certain conditions are met. We can say that were those conditons met, then God would perform a miracle. Maybe theologians could even identify such "laws of supernature." By specifying that the laws of nature be "impersonal" I am ruling out regularities, even universal ones, that depend upon the acts, choices, or decisions of personal entities.

    Finally, Philip objects that I use terms like "natural" and "physical" to spell out what I mean by "MN." Uh, yeah. Then I immediately flesh out what I mean by those terms. Maybe Philip means that I haven't defined them to his satisfaction. Dare I say that satisfying the likes of Philip is the least of my concerns?

  • Philip

    Keith Parsons,

    You defined, or at least explicated, “MN” in terms of some type of physicalism. But my point about the compatibility of ghosts with physicalism wasn’t about definitions. Granted, at present there’s no known physical basis of ghosts. But since physicalism needn’t entail that all types are physical, there can be nonphysical types, such as biological, psychological, or even spiritual ones–whatever those are. As long as these higher-level types supervene (depend) on physical ones and there are higher-level generalizations to be made about the behavior of ghosts, supervenience physicalism is satisfied and thus so is MN on your account. Again, currently there’s no known way of showing that ghosts do supervene on physical processes, but that’s OK too, since you appealed to future physics, not to present-day physics.

    So my point wasn’t that you haven’t taken into account an arbitrary redefinition of “ghost.” Instead, I suggested that your account of MN, in terms of physicalism, doesn’t make ghosts supernatural. Obviously, if ghosts are defined as supernatural, they might be *called* supernatural, but who cares about that? The question is whether, on metaphysical grounds, ghosts really would be supernatural, not whether they’re called supernatural because of a choice of words. Were ghosts to supervene on some physical process, they would be natural even though ghosts themselves wouldn’t be physical. What with your explicating “MN” in physicalistic terms, which allow for levels of nature, your account of MN allows for emergent properties, nonreducible laws, and so forth. That might be enough to make ghosts natural, and that was my point.

    You say you use terms like “natural” and “physical” but then “immediately flesh out” what you mean by them. Well, you talk about causal closure and say “whatever happens in the natural world is explicable in terms of natural entities, forces, and processes.” The “natural world” is explicable in terms of “natural forces”? No dice. You must mean, instead, your appeal to physicalism, where you add, “that is, that it [reality] is constituted exclusively of entities that act and are acted upon in accordance with regular, impersonal, and universal laws.” But that’s just the sentence I criticized in my last message as failing to exclude ghosts as natural.

    If by “natural,” you mean, as you say in this immediate fleshing out, just law-governed, that surely doesn’t make ghosts supernatural. Any wacky psychic can come up with pseudoscientific, unfalsifiable generalizations about the workings of ghosts. So nomic relations aren’t enough. Like I said, what seems needed is an appeal to scientific methods, so that not just any “regular, impersonal, universal laws” will do.

    Also, when you appeal to future physics, you make “physical” vacuous, assuming you allow for radical changes between present-day and future physics. If you don’t allow for such changes, there’s no point in appealing to future physics.

  • Keith Parsons

    Philip's latest missive is no help at all and just sinks further into a morass of confusion. He seems badly confused about what it means to "supervene." He thinks my characterization of MN does not rule out ghosts because ghosts might supervene on physical processes. Philip: Entities do not supervene. Properties might supervene; states like consciousness might supervene; moral values might supervene, but beings (things, objects, substances) do not supervene. Speaking this way is to commit a gross category mistake, and the only result can be to muddy the waters.

    Ghosts and God are normally understood as supernatural entities. If "supernatural" is to have any significant meaning it must mean that supernatural things are fundamentally and in principle different from natural things (not just incompatible with natural things as they are presently understood). Supernatural beings allegedly are not part of the physical universe, though they supposedly can act in that universe. Once you start binding the actions of ghosts or God by regular, universal, impersonal law, you do in fact blur the distinction between the natural and the supernatural.

    Might there be an ectoplasmic physics in the future that studies the regular, universal, and impersonal laws that govern the manifestations and behavior of ghosts and poltergeists? Sure! It seems most unlikely now, but I can see no reason in principle why not. In that case, we would come to understand ghosts as natural beings. So, what's your point, Philip?

  • Jim Lippard

    Does that mean it's wrong to say that words supervene on sounds and writing? Or that our concepts supervene on brain states?

    What's the appropriate way to describe that? It doesn't seem right to me to say that words, concepts, and our self-symbols are properties and not things.

  • Daniel A. Wang

    "It doesn't seem right to me to say that words, concepts, and our self-symbols are properties and not things."

    It's certainly wrong to say that words are things. Words stand for things, and they do so arbitrarily. There's nothing about "dog", for instance, which would lead one automatically to the conclusion that one is talking about the canis lupus familiaris.

  • Daniel A. Wang

    Or as it has also been said: "the map is not the territory".

  • Jim Lippard

    Why do you say that words are not things? I'm not confusing a word with its meaning or reference, but aren't words, meaning, and reference all things? Does "thing" mean only concrete particulars, not abstractions?

    The map is not the territory, but both maps and territories are things. So are nations, ideas, words, sentences, thoughts, beliefs, and concepts.

    Aren't they?

  • Keith Parsons


    I think you meant to say "meaning," not "word," right? Word-tokens are physical things–sounds, written marks, Morse Code dots and dashes, etc. Meanings, however are abstract entities. Concepts have two aspects. In one sense they are states or maybe modifications of my consciousness. But concepts also have propositional content, and a concept's propositional content is also an abstract entity.

    I think the main metaphysical difference between naturalists and supernaturalists is over the kinds of substantial, non-abstract, entities we are willing to admit into out ontologies. Naturalists want to say that all substances are physical or natural substances. Supernaturalists also say that there are natural or physical substances but add that there are also entities like ghosts, souls, or gods that are conceived as supernatural but substantial.

    Ghosts, souls, and gods are not abstracta. Ghosts, for instance, supposedly can haunt graveyards, occupy old houses, have visual apparitions, upset furniture, moan weirdly, rattle chains, and, in general, scare the bejesus out of the living. The number seven or the meaning of the word "cat" cannot do such things. Theists most definitely do not consider God to be an abstraction. For them, God is a, no THE, most paradigmatically substantial (but supernatural) entity.

    I agree, Jim, that meanings and propositonal contents seem to be the kinds of abstract "things" that can supervene (the word "thing" does a lot of mischief). Ghosts and gods cannot, or, at least, that seems a very odd way of talking that needs much explanation and unpacking. Now, ghosts and gods could conceivably turn out to be physical in some sense (Don't Mormons in fact say that God has a body?). But discovering this would not be a matter of finding out that ghosts or gods supervened on some other physical entities. Rather, it would be, for instance, the discovery that they do possess some sort of body (e.g., ectoplasm) and that those bodies, like other bodies, are law-goverened in their functions and operations.

    So, do we agree?

  • Philip

    Keith Parsons,

    My point is that I’d like a better account of metaphysical naturalism than the one you gave. The problem with the one you gave is that it seems internally inconsistent. You spoke of physicalism and of causal closure. Supervenience physicalism allows for emergent properties and for nonreducible levels of nature. In a way, such a natural world, with patterns and nomic relations emerging unpredictably from other ones, isn’t causally closed in the old, mechanistic sense. On the contrary, such a world is creative and open-ended.

    You said that what the naturalist cares about is that the universe be closed in the sense that there is no need to refer to supernatural entities like ghosts. But with respect to causal closure, the practical difference between substance dualism and supervenience physicalism is negligible. On substance dualism, as you said, there could be ghosts that are “fundamentally” part of a difference universe, but that intervene in the natural one. Thus, ghosts wouldn’t have any physical properties, and their ghostly properties wouldn’t even depend on physical ones.

    On supervenience physicalism, ghosts would either have physical limitations or else their emergent, ghostly properties would have to depend on some physical properties. These emergent properties needn’t be epiphenomenal, and so these properties, about which only nonreducible generalizations could be made, might affect the physical level, just as if ghosts were intervening from a different universe. The difference is only between the idea of there being two separate universes, allowing for some mysterious interaction between them, and the idea of there being levels of one universe, each level perhaps emerging from another one, again allowing for mysterious interaction between them.

    One response to this is that if the ghostly facts supervene on the physical ones, anything the ghosts do must be mirrored by something done just at the physical level, making the physical level causally closed. This leads to Kim’s problem of the overdetermination of the lower-level effects. Still, the possibility of top-down causation is as mysterious as that of the interaction between substances or universes. So the bottom line is that when you explicate “MN” in these terms (physicalism and causal closure), you wind up with an unclear picture.

    How would I improve on this account of MN? By connecting ontology to methodological naturalism, that is, to a consideration of scientific methods. To make it into naturalistic ontology, an entity should be objective, in that the entity should be scientifically knowable. A ghost isn’t, but neither is what Thomas Nagel calls the subjective viewpoint. So on what I take to be an improved account of MN, there seems a problem of accounting for consciousness.

  • Philip

    Oh, and about the supervenience nonsense: I was speaking about types, not objects. Thus, I said, “But since physicalism needn’t entail that all types are physical, there can be nonphysical types, such as biological, psychological, or even spiritual ones–whatever those are. As long as these higher-level types supervene (depend) on physical ones…” Speaking of a type is a loose way of speaking about the properties in virtue of which some objects are instances of the type. So when I said that ghosts supervene on physical processes, I was speaking in a rough, shorthand way about ghosts in general, about a feature they have in common that makes them instances of a type.

    As for whether objects, facts, and anything else can supervene, in any case: of course they can. “Supervene” means generally “to come on top of, as something additional, extraneous, or unexpected.” In philosophy, there’s the additional claim that there can be a change in what is added only if there’s a corresponding change in that on which the added thing depends. In the context of metaphysics and philosophy of mind, the things that supervene are indeed properties, or sets of properties. But to say simply that “Entities do not supervene” is just to confuse the philosophical use of the word with the general use.

  • Jim Lippard

    Keith: I meant word, not meaning–but not word-token. I did mean the abstract object word, the same thing Daniel Dennett uses as an example of a prototypical meme that everyone agrees exists. The word "cat" may be tokenized as a sound or as markings on a page, but those tokens have in common that they are all instances of the same word.

    I'm not sure if we're entirely in agreement, but we're not far off. I agree that ghosts are not supposed to be abstracta. I'm not sure that abstracta can't have causal effects, even if those causes are mediated through their physical instantiations. I'm thinking of numerous examples in Douglas Hofstadter's _I am a Strange Loop_, as well as examples of putative "downward causation" described here.

  • Jim Lippard

    BTW, at that last link the paper worth reading is the Carl Craver/William Bechtel paper, "Top-down causation without top-down causes" from Biology and Philosophy, which argues that downward causation is cached out without reference to inter-level causes.

  • Jim Lippard

    Make that "cashed out."

  • Keith

    I think a ghost can be characterized as either an unknown natural entity, or as a supernatural entity, and that the inability to distinguish the two is an epistemological issue rather than something conceptually problematic for naturalism.

    There is the conceptual problem about what counts as natural, and then there is a diagnostic problem about how we identify what counts as natural.

    If you think of naturalism as a kind of empirical hypothesis, in the way that "an all-powerful, all-knowing, and all-good God exists" can be an empirical hypothesis (because it can be "falsified" by evidence of gratuitious evil), then naturalism has survived without being falsified because nothing even remotely like a ghost (however you characterize it) has been shown to exist. That is, not a single supernatural thing (or "supernaturally looking" natural thing) has ever been demonstrated to exist. Hence I take it that naturalism is probably true until some contrary evidence is offered.

  • Daniel A. Wang

    An explanation of what one means to say by "word" would also be useful. It is by no means obvious, from a linguistic point of view. Is "go" one word and "goes" another? What about the context of the sentence? What about polysemy?

  • Jim Lippard

    Daniel: The point of Dennett's statement that words are real is a defense of the concept of meme, with words as an example of memes that have actual existence in the world. Since the analogous item in biology is the gene, which shares some of the same ambiguity as a word, I'm not sure the questions you raise are an objection to the reality of words–the proper inference to draw from your questions isn't "Well, I guess words don't really exist," is it?

  • Keith Parsons

    Philip takes me to task for holding him to the philosophical meaning of the word "supervene" when he only meant it in the "general" sense of “to come on top of, as something additional, extraneous, or unexpected.” Silly me. I thought we were having a philosophical discussion where one should use philosophical terms with some degree of strictness. OK. So what is Philip saying after all? Is he speculating that ghosts and other supposedly supernatural entities might turn out to be "additional, extraneous, or unexpected" vis-a-vis the physical processes and entities we now recognize? So what? Actually, I have no idea what he means, and, as I said before, his use of the word "supervene" only serves to muddy the waters.

    The only other thing Philip says that merits a reply is this:

    "You spoke of physicalism and of causal closure. Supervenience physicalism allows for emergent properties and for nonreducible levels of nature. In a way, such a natural world, with patterns and nomic relations emerging unpredictably from other ones, isn’t causally closed in the old, mechanistic sense. On the contrary, such a world is creative and open-ended."

    Can there be causal closure of the universe given supervenience physicalism? Of course. For supervenience physicalism, causal closure of the universe with respect to the physical means that all caused occurrences in the universe either have physical causes or causes that are efficacious in virtue of the physical. Supervenience physicalism in the philosophy of mind, for instance, speaks of mental events as causal. For instance pain causes wincing. As Jaegwon Kim explains it (Philosophy of Mind, 1st edition, p. 150-151), "This approach [supervenient causation] views the pain's claim as a cause of the wincing as consisting in its supervenience on neural state N, which causes a certain physiological event on which the wincing supervenes. More generally, it takes causal processes at the microlevel as fundamental and considers causal processes at the macrolevel as dependent, or supervenient, on those at the microlevel." If all causes are physical or in virtue of the physical, then causal closure is achieved.

    Past exchanges with Philip quickly reached a point of diminishing returns, i.e. they were more trouble than they were worth. I think I've reached that point again and I shall ignore any further messages from him.

  • Philip

    I pointed out that Keith Parsons forgot or is ignorant about the general use of “supervene,” but that I was all along using the word here in the philosophical sense, not the general one, speaking about properties in a roundabout way by speaking about types (not tokens). He thinks I was saying I was using the word in the general sense. So he couldn’t have read my response with any care.

    Parsons points out there “can” be causal closure in a world satisfying supervenienc e physicalism. Yes, on one interpretation there can, as I said when I referred to Kim’s problem of overdetermination in the case of top-down causation. On another interpretation, though, on which top-down causation is possible, there’s a conflict here, which is why I said this way of talking about MN is “unclear” and can be improved on.

    As it turns out, I have criticized some of Parsons’ other posts in this forum, which appears to explain why he immediately used what would otherwise be a perfectly inexplicable hostile tone towards my messages right off the bat on this topic (not that any of my previous responses merited this tone).

    But I wonder how he knew the Philip writing responses to his piece on MN is the same Philip who wrote responses to his posts on earlier topics. There’s no content at all in my user profile. So he seems to have just assumed there’s only one Philip in the world interested in philosophy who writes responses to him here at the secular outpost. And that assumption was all he needed to write as though I were assaulting his family members.

    It’s that kind of laziness on his part, accounting also for why he couldn’t take the trouble to read my response with any care, and for the deficiency of his account of MN, that I thought I’d try to counteract with my comments here about MN. Of course, I assume there’s virtually no one reading all of these messages. It just feels good, sometimes, to try to improve things, no matter what the other consequences.

  • Dianelos Georgoudis

    Before commenting on the main point about whether consciousness is compatible with scientific naturalism, I would like to discuss the issue of “top-down causation”. In the otherwise terrible “Quantum Gods” Victor Stenger makes the valid argument that even though we may discover high level order which we did not expect and even though the mathematical modeling of this high-level order may not be analytically derivable from the models that describe the low-level order, there is no known case where the top-level order does not emerge in computer simulations of the low-level order. In other words there is nothing to the high level order than the consequence of the low-level order applied to complex systems. So the appearance of “top-down causation” is illusory.

    Jim Limppard suggested some reading on this topic, but I can’t make heads or tails of it. For example one case of top-down causality suggested is that we can’t explain the movement of an individual molecule within a wheel that runs downhill unless we take into account the shape of the wheel. That’s correct; after all the wheel defines the whole system and to explain (or predict) the movement of a single molecule we must take into account the other molecules around it. By the same measure we can’t explain the movement of a single billiard ball without taking into account other billiard balls around it which it may strike. In a mechanistic understanding of reality the behavior of one part of the system depends on the state of the entire system, for the trivial reason that any part of the entire system may affect that one part. In conclusion, as nobody has found a single example where the whole is more than the sum of its parts it’s unreasonable to believe that the whole can be more than the sum of the parts.

  • Dianelos Georgoudis

    Keith Parsons:

    Thanks for the comprehensive analysis of the relation between MN and consciousness.

    I agree with the definition that the naturalist is committed to the premise that reality “is constituted exclusively of entities that act and are acted upon in accordance with regular, impersonal, and universal laws”. (I will henceforth call such entities “naturalistic entities”). I also agree with the observation that the naturalist is not committed to the view that current science exhaustively and accurately describes these entities as science is a work in progress, while noticing that it was via science and the kind of entities which science describes that the naturalist came to naturalism in the first place. Perhaps a way to put it would be that according to scientific naturalism science (or, to be precise, the reification of scientific concepts) will always represent the currently best possible description of reality.

    I have two observations with the idea that according to MN the universe is a causally closed system. I agree that there is overwhelming evidence that all physical phenomena can be explained using only impersonal naturalistic entities, causes, and mechanisms – specifically excluding any supernatural personal entities such as gods, ghosts and souls. My first observation is that physical phenomena are by far not all the data we have, for there is also our subjective experience of life. My second observation is that the fact that the universe appears to be causally closed does not imply that it is. To see this one can use the metaphor of Plato’s cave: the shadows may be causally closed, while the reality that throws these shadows is not. Similarly the shadows may be amenable to mechanistic (impersonal) modeling while the reality that throws them is not. Interestingly enough quantum mechanics opens an avenue for God to massively interfere with the universe’s history up to and including the evolution of homo sapiens to God’s specifications without the phenomenal causal (mechanistic) closure of the universe being compromised. My only point here then is that the apparent causal closure of the universe does not imply that the universe is causally closed, and if the naturalist believes the latter they do so not based exclusively on scientific knowledge but on some additional metaphysical or epistemological principle such as Occam’s razor. (Of course Occam’s razor only applies to viable alternatives, and my larger claim is that MN is not viable.)

    Keith says that he is “extremely suspicious of any sort of positive assertions about ‘ultimate reality’”. But that’s exactly what the difference between naturalists and theists is about: Does reality ultimately consist of entities that act and are acted upon in accordance with regular, impersonal, and universal laws (as naturalism has it), or does reality ultimately consist of entities that act and are acted upon in accordance with free, personal, and universal will (as theism has it)? Or alternatively: Is knowledge about reality ultimately based on the former premise or on the latter premise? Are the most comprehensive explanations ultimately based on mechanical causality or on agent causality?

    [continued in the next post]

  • Dianelos Georgoudis

    The fact that consciousness is difficult to define is not I think especially relevant, but only evidences that consciousness and related concepts refer to the most basic concepts there are. To avoid circularity definitions must stop somewhere, and conscious experience is where the chains of definitions must in the end stop. In any case the fact that from ten year old children to eminent philosophers all normal people can and do discuss consciousness in a consistent and intelligible manner evidences that we all mean the same when we use this term. I agree that consciousness encompasses a large number of “phenomena, states and processes” (or "experiences" for short), but would like to point out that consciousness cannot be identified with the set of the latter but is rather what makes the latter possible: consciousness is the space in which the latter occur. It’s true that consciousness is not “a thing” but then again physical spacetime is not a thing either. Another way to put it is that “consciousness” is a property and refers to the capacity to experience: to say “this being is a conscious being” or “this being has consciousness” means the same as “this being has the capacity to experience”.

    By the claim that “consciousness is a scientifically unnecessary concept” (or, similarly, that “experience is a scientifically unnecessary concept”) I mean that the best explanation of physical phenomena need not make use of such concepts, the same way that the best explanation of fire need not make use of the phlogiston. Conversely we believe that atoms exist in as far as the best explanation of phenomena (of Brownian motion say) requires them. Similarly we know that a monkey’s behavior is best explained by the electrochemical processes in its brain. Now one can name a particular pattern of neuron firings in a monkey’s brain “disappointed” (or say that the behavior caused by such neuron firings evidences that the monkey is disappointed) but such naming conventions add nothing to the explanation; one might as well name the same pattern P342. Indeed, it’s easy enough to anthropomorphize phenomena in contexts where no consciousness is present. For example a commentator of a chess match by Deep Blue may say that the computer is worried about its queen. Characters in a virtual reality computer game may appear to be evil. As a child I used to observe live microbes under the microscope and they appeared self-aware to me. The fact that such anthropomorphizing is possible only evidences that there are probably certain similarities between the physical processes in our own brain which produce a particular kind of behavior in our case, and the physical processes which produce the respective behavior of these other entities. But the fact that such anthropomorphizing is possible does not in any way render the consciousness hypothesis scientifically necessary.

    Similarly that an entity “acquires information about objects, events, and the relation between them” (and that it in this sense creates a representation of its environment) does not imply thinking: after all the old and certainly non-conscious computer program SHRDLU did all of that.

    I agree that both the existence of the past and the existence of consciousness are logically unnecessary, but the existence of the past is scientifically necessary in the sense that many scientific models (e.g. in cosmology or in Darwinism or in thermodynamics) must make use of the concept.

    [continued in the next post]

  • Dianelos Georgoudis

    David Chalmers’ distinction between the easy and hard problems of consciousness is important, because one must not confuse objective behavior with subjective experience. Science can almost certainly explain objective behavior (whether a human’s or a microbe’s or a volcano’s or an apple’s) using only naturalistic entities. My point is that precisely for the same reason science does not need the consciousness hypothesis, i.e. one can always describe systems of such entities *without* using the concept of consciousness. So my argument is not about why or how a material system becomes conscious, nor about the explanatory gap between scientific and experiential knowledge; these are distinct problems on their own right. I simply observe that from science’s point of view the hypothesis that consciousness (including experiential phenomena, states, and processes) exists serves no purpose. If true my premise falsifies all naturalistic ontologies according to which only concepts which are required by science refer to elements of objective reality. At this juncture a scientific naturalist might argue that consciousness is not a “thing” or is not a “substantial entity”, but even though it’s trivially true that consciousness is not a physical thing it’s no less of a “thing” or a “substantial entity” than physical space. Consciousness exists in reality; that’s the one thing we may be completely certain. As does experience – for clearly our own experience of life does exist.

    In relation to Keith’s discussion about correlates between conscious experience and brain states, I would like to point out that science can only find correlates between objectively observable phenomena. So science can find out correlates between people asserting feeling pain (or seeing blue color) and their corresponding brain state. But science cannot possibly find a correlate between the pain (or blue color) and brain states. After all, science cannot even tell whether there is such a thing as “blue color” (see the inverted color paradox).

    Keith suspects that at some point science will have advanced so much in finding the relevant correlates that the hard problem will be dropped. After all, if through such correlates we develop the technology to control pain it won’t really matter in any practical sense that we haven’t explained how anything material could become conscious to experience pain in the first place. But what about other questions, such as which material systems are conscious? Such questions cannot be dropped because they are morally relevant. Correlates will not help us decide whether, say, cockroaches are conscious beings, or whether intelligent computers are conscious. Correlates will not help us decide whether our experience and personal identity can survive the destruction of our brain. Or whether it is reasonable to believe that our brain produces our consciousness in the first place (that’s the second conceptual problem I have pointed out.) So I don’t think naturalism’s problem with consciousness will ever go away.

    At the beginning of his analysis Keith appears to be hinting at a possible solution to this problem, namely that the naturalist may accept the existence of parallel and non-interfering realms of “impersonal supernatural entities”, for example the existence of a realm of Platonic ideas. This sounds similar to the suggestion made in “Naturalism in Question” that naturalism must evolve away from an ontology informed by the natural sciences (i.e. scientific naturalism) and towards some kind of a “multi-polar” view of reality. I do not deny that such a solution may work, but right now this is too vague an idea and naturalists will have to flesh it out before one can evaluate it. (For example I don’t quite see how a parallel realm of impersonal entities can possibly account for our personal experience.) Meanwhile I stand by the claim that one conceptual problem that naturalism as we know it (namely as scientific naturalism) suffers from is the problem of consciousness.

  • Keith Parsons

    D.G., Thanks for your extensive comments. I am not ignoring them. I am tied up in end-of-term busy work right now and it will be a few days before I can respond.

  • Curtis Edward Clark; Dean

    Your statement, "The metaphysical naturalist (hereafter, just “naturalist”) is most definitely not committed to the view that current science offers a complete and satisfactory description of reality," rings true with what I have been attempting to convey through my own blog, In my usage of the term MN, I not only accept naturalism, but the metaphysics that things exist beyond the "merely" physical. Tom Clark of and I disagree about the entities "mind", "soul", and "free will." So what if they are "caused" by the physics of the human anatomy. So is the hand caused by it, but those naturalists would not say that "handwriting" does not exist because it is "caused" by forces in the brain working with forces of the body.

    Glad to find your website. I'll be back.
    Curtis Edward Clark