Epiphenomenalism: Cause for Concern

In remarking on a recent post, commenter Titus Rivas offered a link to a paper he and Hein van Dongen wrote in 2001, in which they launch an assault on the mind-body model known as epiphenomenalism. Epiphenomenalism is the view that the subjective, conscious mind is a causally impotent byproduct of the physical activity of the brain — that it only witnesses our cognitive processes, without having the ability to influence them in any way.

I’ll admit that I have found epiphenomenalism attractive myself; it squares nicely with, to pick one example, the experimental results of Benjamin Libet, which seem to show that we act on our decisions before we are conscious of them. But it has its difficulties, too, and I think Rivas and van Dongen have mounted a successful refutation, which I will summarize here, of this philosophical position. It’s also worth mentioning that they stand in agreement on the insupportability of epiphenomenalism with philosopher Daniel Dennett, and when you have two committed dualists in firm agreement with Dennett that a particular mind-brain model doesn’t work, you have good reason to be skeptical of it.

The core of their argument is a common objection to epiphenomenalism, which is that if subjective mental states can have no causal influence on behavior, then it is difficult to see how we can report on them. If my visual apparatus, in processing its input, creates a substrate-level brain state that induces a subjective perception (that is to say, a quale) of “red”, and that quale is only an epiphenomenon, with no causal power, then how can I ever tell you about it? How can the physical process of speech or writing (or semaphore, or mime, or whatever) be initialized with information that is by definition outside the causally connected world?

This is a good point, but it doesn’t go far enough. Even if the epiphenomenally conscious mind itself cannot report on its experiences, might not the physical substrate, which undergirds and provides the form of the conscious experience, do the reporting? If my experience of “red” always corresponds to a physical state or process in the brain, then why can’t the brain itself answer questions and offer descriptions? As Rivas and van Dongen put it:

This version of the argument, however, can still be refuted by epiphenomenalism. While talking about the contents of consciousness, one does not have to be talking, according to epiphenomenalism, about the contents themselves, but in fact only about the specific physiological substrates that constitute the supposed cause of any kind of subjective experiences. A proposition such as ‘I see the color red’ would thus be caused completely by the supposed physiological correlate of the content of the consciousness concerned. That there would be such physiological substrates for any conscious content that exists, is a basic principle of epiphenomenalism: All subjective experiences would be caused by cerebral structures or processes.

Exactly. But a more serious difficulty arises when we consider the source of the concept of consciousness itself. Why, if the subjective experience of consciousness is causally disconnected from the processes in the brain that report on our beliefs, would the brain even suspect that such subjectivity exists in the first place? It is one thing to imagine the possibility, but there are many things we can imagine in whose existence we don’t believe: a twenty-legged purple cow, for instance. But why, if subjectivity is strictly epiphenomenal, do we report that we believe, with absolute confidence, that we have subjective experiences?

This is, I think, a crushing objection, and good reason, I think, to let go of any lingering attraction I may have felt for the epiphenomenalist view.

So what are the other options? Here is where I think Rivas and van Dongen narrow the alternatives too sharply; they admit of only two remaining possibilities: interactionist dualism, a position that is susceptible to a great many very serious objections, some of which I have listed in this series of posts; and eliminative materialism, which simply denies that there is any mystery to be explained in the first place. I think this does not exhaust the options, and I will take this up in the next post in this series, in which we will look at their paper in some detail, and examine some of their terms and assumptions. Readers with an interest in these matters are encouraged to read the article, which is linked to at the top of the page. I also recommend that readers have a look at this paper by R. W. Sperry; it is a good introduction to another line of thinking, one that I think is closer to the truth than the dualist model Rivas and van Dongen are defending.

Related content from Sphere
  1. see Consciousness Explained, 1991, pp 398-405

18 Comments

  1. bob koepp says

    Hi Malcolm, and welcome to the fray, Titus. My own impressions of the article by Rivas & van Dongen echo what Malcolm has said. I do think the argument against epiphenomenalism is cogent, but I think the range of alternative hypotheses is too narrow. It is unfortunate that discussions of mind and matter and their relations (or lack of relations) so often assume that at least one of these must be a substance. I think that judgment is premature, since I think neutral monism, according to which neither mind nor matter is substantial, is a coherent alternative.

    I’ll be watching with interest to see how this discussion develops.

    Posted April 22, 2007 at 4:21 pm | Permalink
  2. Hello Malcolm and Bob,

    Thanks for your interesting comments.

    The main question is whether you accept that consciousness (in the sense of subjective, qualitative awareness) is real and irreducible to something non-conscious. If you do, neutral monism obviously will not work because it always seems to include a parallelism of subjective experiences and non-subjective physiological processes. What we claim to have shown in our article is that acknowledging the irreducible reality of both subjective experiences and non-subjective physiological processes simply must imply that you also acknowledge the non-parallelist, interactional impact of subjective experiences on physiology and vice versa. Parallelism would exclude such a mutual impact and therefore it cannot be part of the range of logically tenable options. In other words, if neutral monism implies parallelism between subjective and non-subjective aspects of a monistic underlying reality rather than interaction of subjective and non-subjective processes, it is not a tenable option either.

    As we say, only the rejection of the irreducible nature of subjective experiences (be it in a reductive or eliminative sense), dualist interactionism, and idealism (which would reject the irreducible nature of non-conscious processes in the brain, and in fact the brain itself) remain. Epiphenomenalism, parallellism and the identity theory are a priori untenable.

    Concerning the position held by Roger Sperry: what he seems to propose is a kind of emergent materialism which does acknowledge the irreducible reality of subjective experiences. His emergent materialism may be regarded as a kind of ‘holistic materialism’ in that he proposes that subjective aspects of the brain have an irreducible, real impact (“top down causation” if I’m not mistaken) on the physiological aspects of the brain. In this context his views are just as ‘magical’ and ‘heretical’ as our own full-blown dualist solution. It really does not matter that he is a very respected neurologist.

    The point is that consciousness as such and not just its supposed underlying physiological processes would have to be causally efficacious.

    We have not explicitly included Sperry’s position in our paper, but there is a endnote that refers to his and similar views, namely note 9 (please read it in the corrected version – see website linked here):

    We hope it is clear that apart from eliminationism we also discard the various types of identity theory, functionalism and emergence materialism. In practice all of these positions can from an ontological point of view be seen here as forms of materialism, as all of them hold that the mind does not constitute a separate domain of reality, but that it can be seen – and this sense be reduced to – an “interior side”, “pattern” or “level” of matter.
    However, matter can by definition never be subjective, neither in a special manifestation of it nor as some kind of mysterious level. As Karl Popper has shown (p. 81, etc.) the negation of this fact leads to a pseudo-materialism which really is a kind of idealism, or to a definitory confusion (a form of obscurantism) in which the term “matter” [be]comes something like “reality” so that it cannot fulfill a distinctive function in the debate any longer.

    In other words, emergentism or holism do not go far enough in their rejection of reduction. They either claim that subjective awareness consists of a ‘higher level’ of non-subjective processes which would imply that subjective awareness would really consist of such physiological processes. It would constitute an irreducible, specific kind of pattern of such non-conscious processes but apart from this (formal) pattern there would be nothing mysterious about it (in the sense of subjectivity or qualia).
    This is obviously nonsense, as subjective experiences are always subjective experiences and they cannot be seen as some mysterious and abstract holistic level of wholely non-subjective processes.

    Or they claim that subjective awareness is something that exists in an ontological domain of its own, but then it we see a type of emergent dualism rather than emergent materialism, of the kind proposed by Karl Popper.

    Consciousness cannot be regarded as simply a (formal) functional or structural peculiarity of brain physiology.

    In this particular context it does not matter how radical one’s dualism is — mine is very radical by the way — but simply whether one acknowledges that consciousness is irreducibly real and that it does not consist of a ‘higher level’ of anything else. In other words, as in epiphenomenalism and parallelism, whether you accept that consciousness is ontologically different from non-conscious processes and not simply different in a formal holistic sense.

    Best wishes,

    Titus

    Posted April 23, 2007 at 3:29 am | Permalink
  3. Malcolm says

    Hi Titus,

    Thanks for joining us again. There is much I would like to examine more closely here, but I do not wish to do so casually or in haste. These issues are very subtle, and terminology — such as what we mean by “reduction”, and even the meaning of such common terms as “matter” — is of central importance. Something has to give, somewhere, and with all due respect to Popper, a flat declaration of what “matter” is and isn’t capable of may be premature, and might be the crux of the problem. After all, you must admit that it is not as if interactionist dualism doesn’t have serious difficulties of its own, as summarized briefly in these posts.

    Would that the circumstances of my life permitted me to devote more time to quiet contemplation, but alas it is not so, especially right at the moment, and it will probably be a week or more before I am able to offer a worthwhile continuation.

    Thank you for parsing the issues so clearly. Readers are encouraged to join in.

    Posted April 23, 2007 at 10:46 am | Permalink
  4. bob koepp says

    Titus – As I understand it, neutral monism rejects causal interactions between mind and matter, since these are treated as “aspects” of a neutral ground rather than as “substances” capable of acting as causal agents. Parallelism, if it exists, would presumably indicate that mind and matter are correlated as common effects of an underlying cause, with the causal chains _terminating_ in mind and matter rather than _issuing from_ them.

    Posted April 23, 2007 at 2:54 pm | Permalink
  5. Hi Malcolm,

    The distinction between subjective experiences and non-subjective physiology (as a manifestation of what’s usually called matter or the physical world) seems plain enough I would like to say.

    To my knowledge there is one major current that denies that this basic distinction is valid, namely (the non-idealist variants of) panpsychism. However, as far as I know, panpsychism is almost by definition parallelist and if so, it should be considered disqualified as well, for reasons already mentioned.

    Concerning the problems of dualism: the problems of reductive materialism and idealism are obviously a whole lot greater. Dualism accepts both subjective experiences and the brain, whereas reductive materialism simply denies the very consciousness that entertains it, and idealism denies the physical world as more than a mirage. In my view, reductive (and eliminative) materialism is an intellectual monstrosity and idealism is logically viable, but very unlikely, at least starting from my personal intuition, though I admit that — unlike reductive materialism — it is a respectable, rational position.

    Besides, a unilateral impact on consciousness is a lot more problematic than bilateral interaction between consciousness and the brain. I think the problems of dualism are usually gravely exaggerated and turn out to be less impressive if you really consider all the options.

    The main problem about dualism is that it runs against the mainstream scientific consensus and I believe this may well be an echo of an ancient intellectual revolt against dogmatic Christianity (which may be even more relevant in the States than in Europe). By embracing materialist monism scholars have tried to emancipate themselves from the obsolete world view, but they have sometimes been overly zealous. For instance, notions like a soul or spirit are a lot more ancient than the dogmas of te Church and can be found in Greek and Indian philosophical traditions. I think it is about time that we take them seriously again.

    Thanks for taking this issue seriously.

    Titus

    P.S.: You might also like this related paper of mine: Metasubjective Cognition Beyond the Brain: Subjective Awareness and the Location of Concepts of Consciousness

    Posted April 23, 2007 at 3:05 pm | Permalink
  6. Hi Bob,

    Yes, I think you’re right about neutral monism and parallelism.

    My point is of course that if we have a good reason to believe in the physical aspects of the proposed neutral ground it can only be that our minds have been touched by them somehow. Also, if we wish our physical expressions about (the concept of) consciousness to be meaningful they must be based somehow on an impact of consciousness on our ‘cognition’ (be it physical or mental) and from there on the areas of the brain correlated with speech or writing.

    Therefore, it does not make sense to believe in irreducibly physical aspects or talk about consciousness if we also wish to believe that the physical aspects never causally reach the subjective aspects nor vice versa.

    Therefore, parallelism and neutral monism are logically self-defeating, i.e. fatally incoherent. Their proponents (implicitly) claim to have good reasons to believe in physical phenomena whereas they also claim they cannot have any such good reasons.
    And they (implicitly) claim they can physically express meaningful notions about conscious experiences whereas they also claim that physical expressions can never have been causally affected by conscious experiences, but only by physical causes.

    Also see: Why the efficacy of consciousness cannot be limited to the mind.

    Best wishes,

    Titus

    Posted April 23, 2007 at 3:20 pm | Permalink
  7. Malcolm says

    Hi Titus,

    No need to thank me for taking this seriously. I am consumed by the desire to understand consciousness, and have for many years been pursuing a three-pronged effort to get to the bottom of it: through inner work as guided by my 30-year study of Chinese martial arts, along with other systems, in particular the “Fourth Way” system of G.I. Gurdjieff, who my father knew personally; through following the progress of neuroscience; and though philosophical readings and discussions.

    I join you in rejecting idealism, but am inclined to disagree that the problems of interactionist dualism are greater than those of a suitably constructed materialism. I think that your assumption that it is simply the fact of the matter (no pun intended) that the distinction between subjectivity and physiology can be taken for granted, except by assuming panpsychism, may be premature. But I don’t want to develop this argument here in this comment thread: I have scant time today, and I want to give this the careful attention it deserves. It may take me a little while to write a proper response, as I am very busy at the moment, unfortunately, but I don’t want to go off half-cocked. I appreciate your patience, and your willingness to examine these matters with me.

    Posted April 23, 2007 at 3:26 pm | Permalink
  8. Please take your time. I usually am quite occupied myself.

    By the way, your personal background sounds pretty interesting.

    Titus

    Posted April 23, 2007 at 3:54 pm | Permalink
  9. Malcolm says

    Thanks!

    I will say that my life hasn’t been dull, at least so far. I consider myself very fortunate.

    M

    Posted April 23, 2007 at 4:00 pm | Permalink
  10. bob koepp says

    Titus –
    I don’t follow your reasoning when you say, “… if we have a good reason to believe in the physical aspects of the proposed neutral ground it can only be that our minds have been touched by them somehow.” From the perspective of neutral monism, if a cognizer (a part of the neutral ground) is influenced by some other part of the neutral ground x, that influence will have both material and mental aspects. Is there some a priori reason why that influence couldn’t constitute a good reason for belief about some material aspect of x?

    Posted April 23, 2007 at 7:57 pm | Permalink
  11. bob koepp says

    Titus –
    Sorry for posting again, but it just occurred to me that the sort of parallelism that could be accepted by a neutral monist is very different from the sort of parallelism you object to here and in your article. As you describe it, parallelism involves parallel chains of causal influence — mind causally influencing mind and matter causally influencing matter, but with no causal influences passing between mind and matter. For the neutral monist, on the other hand, I think neither mind nor matter influences anything. The only source of any causal influence would be the neutral ground itself. So neutral monism requires that we rethink the nature of causality at the same time we reconceptualize mind and matter as aspectival rather than substantial features of reality.

    Posted April 23, 2007 at 8:21 pm | Permalink
  12. Titus Rivas says

    Hello Bob,

    If you’re right about neutral monism, namely that causality would be completely absent on the level of phenomena this would be just as bad for the position!

    It would imply that we have no reason whatsoever based on experience to believe either in mental phenomena or in physical phenomena. Our conceptualisations of both mental phenomena and physical phenomena would never be brought about by experiences with those phenomena themselves. There would be no causal impact on the conceptualisation process of either the mental aspects or the physical ones. Concepts of the mental and physical aspects of the neutral ground would be directly caused by the neutral ground itself.

    In other words, neutral monism takes away any good empirical and introspective reasons to believe in any aspects of reality and therefore it can have no empirical or introspective reason to believe in the suppose neutral ground of those aspects either. It would reduce us to blind believers who’d simply follow the dictates of the neutral ground about the reality of mental and physical aspects rather than our own experiences, because the latter would never affect us cognitively and we don’t have any memories caused by the experiences themselves. Any ‘knowledge’ would be a matter of blind faith in the wisdom of the supposed neutral ground. Thus, it would be completely incoherent to claim that we have experiences that point to mental and physical aspects of the neutral ground. It would never be those experiences we would build neutral monism on. The concept of neutral monism would be revealed by the neutral ground rather than deduced from our experiences. So neutral monism would have nothing to do with formulating a theory to account for the occurence of mental and physical aspects of what we experience as (phenomenal) reality. Neutral monism asks us to believe in it whereas it can point to no specific experience on which to found such a belief. The only ‘reason’ to believe it is true is precisely that the neutral ground ‘says so’. Now, unless the neutral ground would exclusively give us valid concepts, why should we trust it? Of course, it does not exclusively give provide us with valid concepts, because otherwise there would be no alternatives to neutral monism around. As there are, why should we simply believe in neutral monism whereas it can give no good reason why it should be true, other than ‘it is true’?

    Best wishes,

    Titus

    Posted April 24, 2007 at 2:05 am | Permalink
  13. bob koepp says

    Titus –
    Echoing your earlier remarks you say that neutral monism “would imply that we have no reason whatsoever based on experience to believe either in mental phenomena or in physical phenomena.” So, I’ll repeat my earlier question. Is there some a priori reason why the influence of one part of the neutral ground on another part of the neutral couldn’t constitute a good reason for belief about some material or mental aspect of x?

    Your remarks about “good empirical and introspective reasons to believe” strike me as very odd. First, if you are assuming a causal theory of “good reasons” or, more generally, a causal theory of knowledge, then that account quite obviously needs to be revised when we move from a metaphysical view of mind and matter as substances to one where they are viewed as aspects (maybe “modes” would be better, a la Spinoza) of an underlying substance. Second, let us not forget that we are talking here about alternative metaphysical theories, where empirical and introspective phenomena are not so much evidence for a theory as terms of the problem we are trying to resolve. Let’s at least admit that kicking a table provides no reason to adopt materialism as opposed to idealism — or neutral monism for that matter.

    Posted April 24, 2007 at 9:49 am | Permalink
  14. Titus Rivas says

    Hi Bob,

    My point is not neutral monism itself would hold that its concepts should be based on the causal impact of physical aspects on the mental aspects and that the physical expression of concepts about the mental aspects should be based on a causal impact of those mental aspects on the physical aspects. What I’m trying to say is that once we discard a causal theory of conceptualisation of aspects of the supposed neutral ground, the only alternative turns out to be blind faith, based on nothing but itself.

    In other words, in my view there may indeed be an alternative to a causal theory of knowledge, but I simply cannot accept its claimed rationality. Could you give a valid reason to take neutral monism seriously which would not be based on any kind of experience?

    I hope this clarifies things a bit.

    By the way, I accept that idealism is viable, even though I’m not a proponent of this alternative metaphysical theory. So this is not a matter of discarding any alternatives out of hand, but just of discarding alternatives that I can’t regard as rational.

    Best wishes,

    Titus

    Posted April 24, 2007 at 2:50 pm | Permalink
  15. bob koepp says

    Hi Titus –
    I think we are probably approaching these issues with different assumptions, and probably operating with different definitions of some key concepts.

    If I understand the first paragraph in your last post, however, what you are saying is that if non-substantial aspects of a substantial neutral ground do not exert causal influences, then there can be no rational account of how they could be objects of knowledge. I appreciate the problem here, but I don’t think it arises only in this context.

    I firmly believe that we don’t have any reasonable account of any kind of causal influence — just labels that we use to cover gaps in our understanding. We don’t know, in metaphysical terms, how one bit of matter could cause anything in another bit of matter. We are similarly ignorant about how mental events could causally influence other mental events. And virtually everybody admits that we have no understanding at all of how minds and material objects can interact causally.

    Finally, you press again for reasons favoring neutral monism that are not based on any kind of experience. How about, “Neutral monism doesn’t generate a puzzle about the mode of interaction between substantive mind and substantive matter.”

    I still suspect that you are using notions like ‘experience,’ ‘reasons’ and ‘based on’ in a way that makes sense only if we assume that mind is substantive. But such assumptions must be reconsidered and their terms reinterpreted if we are going to be able even to consider alternatives to the idea of mind as substance.

    Posted April 24, 2007 at 4:23 pm | Permalink
  16. Titus Rivas says

    Hi Bob,

    I don’t exactly know what you mean by substantive in this context. I’d prefer the term real instead. How can we know that there really are subjective experiences?

    I for one do not doubt the existence of subjective experiences and I’m also pretty sure that there are non-subjective physical phenomena as well. For instance, I fail to see why the problem of causality should be replaced by a theory that would deny the reality of my consciousness.

    Causal connections could simply be fundamental, irreducible part of the structure of reality. They can’t be explained by anything more fundamental if (in this context) they are the most fundamental thing around.

    Denying the reality of causality would take away any reason to believe in anything whatsoever. How can one first accept the reality of subjective and non-subjective ‘aspects of a neutral ground’ and then deny that we can know them through the impact of our experiences with them? It is impossible to be a coherent neutral monist (in the sense you have defined) and hold on to a belief in subjective and non-subjective aspects of reality on the basis of our experiences.

    Either one accepts that our experiences give us reason to believe in the reality of subjective and non-subjective phenomena, but then it becomes impossible to maintain the neutral monistic position of ‘no causal efficacy within the phenomenal realm itself (i.e. there is only causality coming from the neutral ground)’.

    Or one does accept our experiences as a causal basis for our belief in the reality of subjective and non-subjective phenomena, but then one has to discard the neutral monistic tenet of non-causality.

    Either way, neutral monism turns out to be incoherent. Therefore, it cannot serve as a rational alternative for causal accounts. But then again, why should be so eager to find such an alternative, if we also can accept interaction as a basic fact of nature?

    Best wishes,

    Titus

    Posted April 25, 2007 at 3:31 am | Permalink
  17. bob koepp says

    Hi Titus –
    It’s very clear to me now that we really aren’t “on the same page.” I’ve used the term ‘substance’ in more or less the way it’s been used in traditional metaphysics, to refer to self-subsisting individuals with causal powers. I think there are things that are real that don’t have causal powers, i.e., the capacity to influence other things. I even take very seriously the idea that to be real is to be embedded in a network of causal relations. Some nodes in causal networks, however, might be termini rather than sources of causal influence — hence “real” but not “substantial.”

    With that in mind, I don’t doubt that there are subjective experiences. But I do question whether they are causally efficacious. And I certainly don’t deny the reality of causality. What I do claim is that ‘causality’ is a metaphysical notion par excellence, liable to reinterpretation depending on what other elements are incorporated into a particular metaphysical view. My position in this discussion has been that assumptions about the kinds of causal relations into which mental and physical phenomena can enter cannot simply be retained as we move from a dualistic to a monistic metaphysical framework.

    I also have questioned assumptions you seem to make about the causal basis of cognitive states. You seem to take the view that we could not cognize anything that did not exert a causal influence on us. I think this not at all obvious. Presumably, consciousness is never external to itself, and so never acts as an efficient cause on itself. Could you explain how, on your view, we come to consciously entertain the concept of ‘consciousness?’

    I hope that the above helps at least a little to put some of my comments in context. I began my part of this discussion by noting that I think the argument you presented in your article was cogent; meaning that given the assumptions you seemed to be making about metaphysical matters, your conclusions followed. But I also noted that there are alternatives to those metaphysical assumptions which might point toward possibilities you were rejecting as incoherent. I haven’t, however, made it my business to deny consciousness, subjectivity, causalilty, reality, etc.

    Pax

    Posted April 25, 2007 at 10:29 am | Permalink
  18. Titus Rivas says

    Okay, Pax, Bob!

    Let me limit myself to your question, as I suppose that by now we agee to disagree about this issue.
    For me, it is an epistemological matter as well as an ontological one, so that I believe I have some reason to criticize other ontological frameworks on epistemological and logical grounds. For me, assuming (to know) that any empirical concept refers to a real entity while denying that the entity in question has had a causal impact on our cognition (so that we can know the concept is realistic) is incoherent, regardless of the specific ontology that is at stake.

    But anyway, I’ll give you my answer, your question being

    Could you explain how, on your view, we come to consciously entertain the concept of ‘consciousness?’

    This is how I view it:

    – First we have subjective experiences.
    – These subjective experiences have a causal impact on our memory.
    – Our cognitive abilities abstract a concept from our stored memories of subjective experiences.
    – This concept may in turn become part of our subjective thoughts, though as a concept it also exists outside the conscious mind.

    In other words, the impact is indirect via a conceptualisation process based on stored memories of subjective experiences.

    Best wishes,

    Titus

    Posted April 25, 2007 at 8:06 pm | Permalink