Wednesday, April 22, 2009

The interaction problem, Part II

In an earlier post, I suggested that one of the advantages of hylemorphic dualism over Cartesian dualism is that its notion of formal causation allows it to sidestep the interaction problem. For if the soul is the form of the body, rather than a distinct substance in its own right, then there is no question of two substances having to “interact” in the order of efficient causes on the model of two billiard balls. There is rather just the one substance, a human being, having (as every other material substance has) two constituents, its form (or soul) and its matter (or body). The “interaction” between them is no more problematic than the “interaction” between the form of a tree and the matter that makes up the tree. For soul and body do not “interact” in the first place the way two distinct things do; they together constitute a single thing. My intention to raise my hand is not one event which has somehow to get into causal contact with another, physical event. It is rather the formal-cum-final cause of a single event of which the activity in my nervous system and arm is the efficient-cum-material cause. The solution to the “interaction problem” is to break out of the conceptual Procrustean bed of the mechanical picture of the world and return to a philosophy of nature informed by Aristotle’s four causes.

(Actually, saying that this “sidesteps” the interaction problem is misleading and anachronistic, since it conveys the false impression that hylemorphic dualism was motivated in part by a desire to solve the interaction problem. In fact there was no interaction problem until early modern philosophers like Descartes abandoned hylemorphism and redefined matter, mind, and causation in an explicitly anti-Aristotelian way. As I show in The Last Superstition, the “mind-body problem,” like the “traditional” philosophical problems of induction, personal identity, causation, and many others, is largely a consequence of the early moderns’ mechanistic revolution.)

Some modern dualists have suggested that the interaction problem is oversold in the first place. And they are right to complain that materialists fling it around much too glibly. To be sure, the interaction problem really is a problem for Cartesian dualism, but it is not (by itself, anyway) a refutation of it. Let’s briefly consider why – before going on to see why it is nevertheless a serious enough problem that any dualist is well-advised to consider opting for hylemorphic dualism over the Cartesian variety (especially given that, as I would argue, there is already ample independent reason to adopt hylemorphism as a general metaphysics).

One reason why the interaction problem does not strictly refute Cartesian dualism is that the Cartesian dualist could always simply deny that mind-body interaction is real in the first place, and opt for occasionalism, or parallelism, or epiphenomenalism. Of course, the extreme oddity of these views leads many critics of Cartesian dualism to regard recourse to them as little better than an admission of defeat, a desperate appeal to a deus ex machina. As Bill Vallicella notes (without necessarily endorsing the judgment) both Malebranche’s occasionalism and Leibniz’s pre-established harmony have been accused of deploying a deus ex machina strategy, especially since both literally appeal to God to resolve the question of the mind-body relationship.

But though the charge is common, it is unfair. Malebranche and Leibniz both had independent philosophical reasons for believing in God, and both also had independent reasons for denying that there could be causal interaction between created substances (any substances, not just mind and body). And given their respective specific understandings of the nature of substances, Malebranche had good reason to think that God continuously mediates between them, and Leibniz to think that God does not do so but instead established a universal harmony between them at creation. Hence, Malebranche quite naturally concluded that (for example) when you decide to have a beer your body moves towards the fridge, not because the decision causes the bodily movement, but because God, on noting that you have made that decision, causes the body so to move. And given his different conception of substance, Leibniz quite naturally concluded instead that the decision and the bodily movement in question were each simply the natural unfolding of what was pre-programmed into each substance at their creation. These views of the mind-body relationship were not developed simply to deal with the interaction problem, but flowed naturally from two sophisticated and independently defensible metaphysical positions.

Defensible, but still bizarre, rarely actually defended, and subject to various objections of their own. And most modern dualists would agree with materialists that it would be preferable to avoid occasionalism and pre-established harmony if one can manage it. Hence the greater popularity of epiphenomenalism, according to which mental events do not cause physical events but are rather merely the ineffectual byproduct of the flux of physical events. When you decide to have a beer, the decision itself (or at least the conscious awareness of it) is not what causes your body to walk over to the fridge. Rather, entirely unconscious physical events caused your body to do so, and in the process also caused the conscious experience of making the decision in question, which event itself had no causal efficacy at all.

Though not much less bizarre than occasionalism and pre-established harmony, epiphenomenalism at least has this advantage over them as a way for Cartesian dualists to deal with the mind-body problem: Materialists too seem led into it, so that they can’t plausibly use it as a stick with which to beat dualists. For materialist theories of mind have a notorious problem explaining the efficacy of mental content. If (as materialists tend to hold) it is only the physical properties of mental states which give them their power to cause other physical states, then their mental or intentional content seems epiphenomenal. For example, if we suppose, as a materialist might, that my decision to have a beer is identical with or at least supervenes upon some event in my nervous system, then if it is only the physiological properties of that event that enter into the explanation of how it caused my bodily movements, the fact that it involved a representation of beer, specifically, or indeed had any representational content at all, drops out as causally irrelevant.

So, if materialists as well as Cartesian dualists are faced with the possibility of having to swallow epiphenomenalism, the former cannot accuse the latter of having a special difficulty in accounting for mind-body interaction. Still, this is more a rhetorical victory for Cartesian dualism rather than a substantive one. For epiphenomenalism is notoriously unsatisfactory, and not just because it is odd to say that your decision to have a beer is not what caused you to go to the fridge. If our mental states can have no causal influence whatsoever on our bodies, it would seem to follow that we cannot even talk about them. Indeed, the epiphenomenalist himself could not even talk about his thoughts about epiphenomenalism. For those thoughts would be as inefficacious as any other mental state or event. When he says “Epiphenomenalism is true,” the fact that he thinks it is true has absolutely nothing to do with his saying so. This is bizarre at best and incoherent at worst. And though epiphenomenalists have tried to find various ways around the problem, it would be better not to have to deal with it in the first place.

So, a Cartesian dualist is well-advised not to deny that mind and body interact. And this brings us to the second reason why a Cartesian dualist has a right to complain that his critics’ appeal to the interaction problem is often too glib. As Bill Vallicella has pointed out in several past posts, whether a Cartesian dualist can account for mind-body interaction depends on what view of causation one is assuming. And there is at least one view of causation – a regularity theory – on which no interaction problem arises at all for Cartesian dualism. As Bill has suggested:

Suppose we say that:

Event-token e1 causes event-token e2 if and only if (i) e1 temporally precedes e2, and (ii) e1 and e2 are tokens of event-types E1 and E2 respectively such that every tokening of E1 is followed by a tokening of E2.

On this Hume-inspired theory (sans the contiguity condition), causation is just regular succession. If this is the correct theory of causation, then there is nothing problematic about mental events causing physical events, and vice versa.

About this, Bill is absolutely right. If such a regularity analysis is correct – and there are philosophers who would defend such an analysis on grounds independent of their position on the mind-body problem – then the interaction problem is solved. At the very least Cartesian dualists can plausibly hold that objections to their position based on the interaction problem are less conclusive than their critics often let on.

But the “if” in question is a very big one. Is such a regularity theory of causation really plausible in the first place, or at least plausible enough to show that Cartesian dualism really can account for mind-body interaction after all? I think not. One reason why is that apart from its use of the word “cause,” the proposed analysis is perfectly compatible in substance with occasionalism, parallelism, and epiphenomenalism. For on each of those views, it is perfectly possible to say that a mental event of type M is always followed by a physical event of type P, in which case, on Bill’s suggested regularity theory, M will count as the cause of P. But an “interactionist” theory which differs in substance not at all from occasionalism, parallelism, or epiphenomenalism – all of which deny interaction – is an “interactionist” theory in name only.

Another problem with the proposed regularity analysis is that it simply doesn’t capture what we mean by “cause.” As Hume himself recognized, the connection we take to hold between a cause and its effect is not just a regular one, but also a necessary one. We don’t just think A was in fact followed by B, but that in some sense it had to be followed by B. Of course, Hume thinks there is no objective source for this idea of necessity, that it has to be traced to a purely subjective expectation on our part. For he holds that there is nothing in our ideas either of a cause or of its effect that necessarily links them together. Objectively speaking, causes and effects are “loose and separate,” and any effect or none could in theory follow upon any cause.

This Humean result is what makes “regularity” theories of causation seem at all plausible. But what they really give us is not causation, but rather only some replacement for causation. (The same holds true, I would say, for counterfactual analyses of causation.) So, no appeal to such a theory really solves the interaction problem at all. Rather, it simply adds one mystery to another, saying, in effect: “Causation in general is already mysterious, so why shouldn’t mind-body interaction be?”

The thing is, the reason causation in general is mysterious is the same reason mind-body interaction in particular is: the mechanistic revolution that displaced the Aristotelian-Scholastic model of explanation, throwing out formal and final causes and trying to make do with bastardized versions of material and efficient causes. As I have noted in earlier posts and discuss at length in TLS, one of the main arguments the Aristotelian tradition gives for formal and final causes is that without them efficient causation becomes unintelligible. Unless there is something in the nature (or “substantial form”) of a thing by virtue of which it “points to” or is “directed at” the generation of a certain effect (as its final cause) then there is no way to account for why exactly it produces that effect as opposed to some radically different effect, or none at all. Hume was merely drawing out the inevitable consequences of the mechanistic revolution. (And even here Hume is, as always, overrated, since the skepticism vis-à-vis causation implicit in the rejection of formal and final causes was already foreshadowed in Ockham and the late medieval nominalist tradition.) The way to solve both the interaction problem and the problem of causation is, accordingly, the same: a return to the Aristotelian metaphysics early modern philosophy displaced.


  1. Epiphenomenalism actually seems on par with parallelism to me. I know the difference between them, I think, but that's still the first impression.

    As always, great post.

  2. An excellent philosophical critique of epiphenomenalism is available in this paper by philosopher and psychologist (and substance dualist) Titus Rivas:

    In the abstract, you can read "This article examines the background, implications and merit of the position of epiphenomenalism. Most of all, the authors systematically present an analytical argument against epiphenomenalism, the argument from the justification of the assertion of the existence of consciousness. It is shown that whereas epiphenomenalists claim to know that consciousness exists, they implicitly deny the possibility of knowing consciousness, since (according to their position) consciousness cannot have any influence on our knowledge. Similarly, the authors examine and reject the position of parallellism. Parallellism implicitly states it knows of the existence of an unknowable physical world. Consequences are mentioned for philosophy and empirical science"

    On substance dualism, some argue that the "interaction problem" only is a problem if we taked for granted classical (newtonian) physics, which is essentially mechanistic and localistic. In this framework, only material things can interact with material things by local interactions (like billard balls contacting to each others)

    But quantum mechanics doesn't entail local causation as the only possible way of interaction between particles. So, the "spooky action at distance" (as Einstein called the quantum pehnomenon known as "Entanglement") shows that causation doesn't necessarily requires local mechanical interactions.

    It doesn't proves that the interaction of consciousness with the brain is based in that quantum phenomenon; but it can't be ruled either.

    In fact, quantum physics Henry Stapp has published a paper presenting a model of dualistic interaction based on quantum mechanics, entitled "Quantum Interactive Dualism: An Alternative to Materialism" published in the Journal of Consciousness Studies:

    The model is controversial, but I think Stapp have defended very well (especially in his book Mindful Universe), where he responds to many objections to it (e.g. the objeciton that QM applies only to microscopic things, and not the brain)

    Whatever are the merits of Stapp's theory, the point is that the interaction problem is entailed in the mechanistic conception of nature (and classical physics), but not necessarily in quantum mechanics. It's probably insoluble in a superceded classical physics framework, but not necessarily in a QM one. in this point, it's an open question.

    It adds another line of defense of substance dualism, in addition to other philosophical and empirical arguments in support of it.

  3. Sorry, the link to Stapp's paper is this:

    The other link is to Stapp's website (he has many good papers there)

  4. One question and one comment:

    (1) I've always wondered: Granted that hylomorphism avoids the Cartesian mind-body problem, why doesn't a different version of the interaction problem nonetheless arise for Aristotelians and Thomists in the relation between the intellect and the body? There's no interaction problem between a tree's form and matter, because the tree's form isn't spiritual/immaterial. But man's substantial form is, which is what allows for continuity and bodily resurrection after death. So why doesn't an interaction problem arise between man's immaterial cognitive capacities and his animal capacities?

    (2) I don't think that the mark of genuine causation is necessitation. As Anscombe showed, Hume was wrong to assume this. X can cause Y without X necessitating Y. Rather, causing is a kind of making, and the mark of causation is the derivativeness of an effect from its cause.

  5. Hello Crude,

    One difference is that epiphenomenalism at least allows one-way causation, from the physical to the mental. Why this is supposed to be less mysterious than mental to physical causation is not clear, but the advantage is presumably supposed to be that it makes it possible both to give an entirely materialistic account of behavior without denying an irreducibly mental element to human nature.

    Hello Jime, thanks for the references.

  6. Hello Anonymous,

    1. The reason is that though the rational soul is a subsistent form, it is still a form, so that its relation to the body is that of formal rather than efficient causation. The fact that intellect is immaterial doesn't entail that we suddenly have to think of it as a distinct substance that somehow interacts with the body, understood as a separate, independent substance. The body just is not a body at all without its form, the soul, which just happens to have an immaterial power but nevertheless functions as a form rather than as an independent substance.

    2. Actually, I agree, and I should have stated my point more cautiously.

  7. Professor Feser, do you think the cases of "split brain patients" present a problem to dualism?

    If, beyond any doubt, it could be shown that those patients have "two streams of consciousness", it would be a refutation of substance or cartersian dualism?

    Also, does it present a problem to the Thomistic conception of the mind and brain?

    On other topic, but relevant to discussions on consciousness and of possible interest for the readers here, psychologist Charles Tart has published a book titled "The End of Materialism"

    I just finished the reading of it, and I consider it a masterpiece.

    Tart is a first rate scholar with an encyclopedic knowledge of psycholohy, human abilities and consciousness literature.

    His book examines in detail the influence of materialism and scientism in contemporary science, and discusses some empirical evidence that conflicts with such paradigm.

    A must read in my opinion.


  8. ZC,

    I'm not Prof Feser, of course, but I don't see it as a refutation of Thomism, or even of dualisms in general. In both cases because it's not clear how what we'd be experiencing is anything more than a damaged single person.

    Just my thoughts.

  9. Hi ZC, the short answer is that I agree with Crude that such examples are really cases of a single person whose mind has become disordered, not cases of one mind splitting into two.

  10. Great post!!
    Dr. Feser,
    my teacher was going on about how the idea of a soul in a body is like a ghost in a machine. And how nonsensical this was. I mentioned the idea you have in here of the soul being the formal cause and the body being the material cause.
    She had no reply other than saying "I'm not familiar with that position".

    But! She was talking like an expert who knew for a fact that there is no such thing as a soul. Kind of weird for her to do that considering she really doesn't understand all of the positions.
    I wanted to say that to her, but i don't want to get treated differently in class.

    My friend goes to St. Norbert College, which is Catholic, .... and he said that not a single philosopher on staff has a theistic bent. At least the one's that outwardly mention it in class and he said it is most of them.
    Why would a Catholic school do that? Hire a bulk of staff that doesn't even believe?
    not that they shouldn't hire any. But, before coming across some Christian professors I was pretty close to ditching my faith.

  11. ZC:

    In support of Crude and professor Feser's position, I'll quote the paper "Neuropsychology and personalist dualism: a few remarks" by philosopher Titus Rivas:

    There is one subfield of neurology that seems less obviously irrelevant: experiments with split-brain patients which at first sight may appear to show that within an experimental context they possess two separate streams of consciousness. (A related phenomenon may be the so called alien hand syndrome.) If this interpretation were the only one possible, substance dualism would be in serious trouble (it would certainly not discredit personalist dualism in general by the way, as it might be reconciled with emergent semi-substantialist dualism as defended by Karl Popper).

    However, there is a lot of debate about this issue, and there is no general consensus that the data should inevitably be explained in this manner. As John Beloff (whose dualism might be less personalist than mine) remarked in a paper about dualism, during these experiments one of the apparently conscious streams of consciousness may simply be subconscious (i.e. mentally active, but at a non-conscious level). The phenomenon would presumably be caused by the peculiar experimental setting which makes the simultaneous conscious integration of all the data into one total picture impossible. The fact that subconscious processing is allowed to express itself through one of the hands may be due to the experimental setting as well, or else it might be related to a lower level of inhibitory control (caused by commissurectomy) by consciousness of motoric expressions based on mental, non-conscious processing (perhaps functionally comparable to the strange exclamations of patients suffering from the Gilles de la Tourette-syndrome, expressions which may be humorous and to the point but which are not usually seen as based on conscious thought either). The personal self would consciously be aware of only one part of the perceptual data and the rest of the information would reach its subconscious mind. The process would be comparable to other settings such as subliminal perception, blindsight and automatic writing, in that there would be a lot of parallel cognitive processing going on, but exclusively at a subconscious (or non-conscious) level rather than by (or through) a presumed co-consciousness. As we can also see in those other cases, this type of processing may be a lot more complex and intentional than what we would expect of a 'blind', mindless automaton, but there is no reason to suppose that it also has to be conscious (in the sense of subjectively experienced by a subject). It is certainly wrong to suppose that subjectivity has absolutely no impact on the brain (as is claimed in epiphenomenalism and other forms of physicalism) but it is also wrong to believe that all 'creative' cognitive processing and responses must inevitably be caused by consciousness, unless we wish to confuse two very different meanings of the word conscious, i.e. (cognitively, but not necessarily subjectively) aware and subjectively experienced (in other words: specifically subjectively aware).
    We know from psychological and psychiatric literature that people may develop subconscious personality structures and it is enough to conclude that part of the perceptual information presented in split-brain experiments cannot reach consciousness though it is still processed and acted upon on a subconscious level, by a subconscious part of the person's personality. So there would be no creation of a new, separate mind or self (in the sense of conscious subject), but it would simply be another (rather limited) example of a cerebral impact on the mental activity of a substantial personal self, namely a limited impact on the integration of perceptual data in consciousness during specific experiments. Results of split-brain experiments would remain special in other respects (e.g. regarding the localisation of parts of the brain that specifically interact with specific mental functions), but not in the sense that they would prove the supposed divisibility of the self.

    The subsequent processing of the data that are not accessed consciously does not have to be influenced directly by the condition of split-brain. It is more parsimonious to assume that as the data are not (rather than separately) perceived consciously, their further cognitive elaboration, and any motor response following from this, is shielded from consciousness as well. This would also explain unexpected emotional responses and the release of repressed wishes, attitudes, styles of communication, etc. through the (part of the) motor apparatus linked to the (non-dominant) hemisphere that is (during these experiments) not directly, interactively associated with consciousness. Similarly, data perceived consciously might not always be taken along during the subconscious processing of data shielded from consciousness. In that sense, during the split-brain experiments there may be two (or more) chains of relatively independent parallel processing going on within the mental life as a whole of one and the same self. It may well be this temporary independence of the chains of processing that leads scholars to the conviction that there are now two truly independent minds in one and the same skull. This mistake amounts to the confusion of functional dissociation within the self's mental life with ontological dissociation of the very self. Functional dissociation is compatible with substantialism.
    We already know very similar phenomena from the field of automatic writing in persons who have not undergone a commisurectomy. Parts of these subconscious characteristics do not need to have been created by splitting the brain, but they may have been present a lot longer. We do not have to embrace psychoanalytic theory to accept that a lot of our psychological structure normally does not reach the surface of consciousness.
    I think William James' trasmission theory of consciousness can account for these cases too. See James' lecture on human consciousness and immortality (a masterpiece):


    Regarding Tart's book, yes it's a very good book. It have been recommended by consciousness researchers like Dean Radin or Stanley Krippner.

    Another good book is "The mind and the brain" by psychiatrist and philosopher Jeffrey Schwarzt.

  12. The commentary on William James is mine, not of Rivas. (It appeared very close to Rivas' final sentence; a involuntary typo of mine)

  13. I forget to put the link to Rivas' paper on dualism:

  14. Hi Dr. Feser,
    you write; "the "interaction" between them is no more problematic than the "interaction" between the form of a tree and the matter that makes up the tree".

    If the soul is the form of the body... like the tree is the form of the material constituents that make up the tree.... does this entail accepting that trees have souls?

    It just seems like soul is being equated with form. Which might be right. But everything has a form as well as its material substance.
    But we don't say that everything has a soul because it has a form.

  15. Plants have the "vegetative/nutritive" soul. That is their proper form. Nutrition, tropisms, blossoming, pollinating, etc. are how plants seek their proper perfection (finality).

    Animals have the nutritive and "sensitive/risible" soul. That "something more" is what makes animals something more than plants, yet something organically just like them in other ways. I prefer to call this the semiotic soul.

    Humans have the lower two formal causes as well as the "rational" soul. This formal order accounts for our own "something more" as organic animals. The intellect is the keenest tool in our nature, but we are still basically semiotic responders seeking nutrition for growth and propagation of kind. Read David Braine's book on human nature.

    The doctrine of the soul is a subset of hylomorphism. Once you get there, you can get to "the soul". Once you see how powerful hylomoprhism is as a general metaphysic, you can easily brush off the scoffers about "ghosts in the machine" and other such canards. Feel free to search my blog for lenghtier comments of mine on these topics.

    Forms are the metaphysical seams in nature. read Machuga's book on the soul. Calling them souls is just a verbal fact, which needs to be clarified in a discussion, sort of like clarifying what "Holy Ghost" in a bygone age means vis-a-vis Holy Spirit in our day. One could in theory call all forms "souls" but then one would just be asking for headaches from the reigning scoffers.

    All the best,

  16. Hi Elliot,
    What you typed kind of made sense to me.
    I think part of my mistake was assuming that:
    form = form = form

    They might all be classified as forms, but that doesn't mean that they are all nondistinguishable from each other.
    Cat and humans are both placed in the category mammals.... but they aren't the same in every respect.

    Thanks for the explanation.

  17. Dr.Feser, how do we understand how angels/demons affect the minds of men? If the human intellect is formal cum final causality, how would a pure mental substance affect our embodied minds?