IV.—Second Thoughts on Causation, Dualism, and Interaction

Howard D. Roelofs

Mind, New Series, Vol. 56, No. 221 (Jan., 1947), 60-71.

Stable URL: http://links.jstor.org/sici?sici=0026-4423%28194701%292%3A56%3A221%3C60%3ASTOCDA%3E2.0.CO%3B2-K

Mind is currently published by Oxford University Press.

In this essay I shall present as clearly and as concisely as I can, certain conclusions about causation, dualism of substances, and interaction. These conclusions are the result of reflections, at times sustained and systematic, at times intermittent and perhaps unconscious, extending over a period of years. I shall indicate the primary data and the chief arguments supporting my conclusions, but I shall present neither in any detail. Two considerations have persuaded me to this procedure and each is of sufficient importance for my present purpose to warrant statement.

The first is that my primary data, although of two sorts, are thoroughly familiar to those who may reasonably be expected to read this essay. One sort of data to which I shall refer, is found in the daily, ordinary experience of every one of us. I mean no more and no less than our ordinary use of, and reliance upon, particular causal connexions; our daily encounters with bodies and with minds, our own included; and finally the prima facie response of mind to body, of body to mind, which is a stable feature of human life both individual and social. The other source of primary data is the chief arguments of Hume and Berkeley, in the one case against causation, in the other against matter. Certainly if we, as philosophers, have a working knowledge of any material, it is of this. It may be that Hume's arguments are more securely in our grasp than the standard pattern of daily life, for some of us are better readers than we are observers. But since we are ourselves living and active, at least I so assume, the actuality can make good any deficiency in memory and in previous observations.

The second consideration impelling me to omit detailed exposition arises from what I take to be the nature of reasoning. There is a tradition that in philosophy the appeal is to reason. I accept that tradition and endeavour to act in accordance with it. But reasoning is essentially individual. There can be mass suggestion, there are methods of group persuasion, but reasoning is carried on by each of us not only for himself but by himself. The principles of reasoning are indeed common to us all, and when we reason correctly our reasoning is the same; yet the particular reasonings each one of us actually does are nonetheless individual; in fact, they are private. A simple, formal argument, such as a single syllogism, can be put in front of us by another, our attention directed to it, and we can then reason along with him. But when premises are many and varied, when reasoning is complex, not completely formalised, and more often indicative of a probable conclusion than a demonstrative, then to accompany the first presentation of a debatable thesis with detailed arguments is more likely to produce confusion than persuasion. That is the second reason why I shall restrict myself simply to indicating data, suggesting arguments, and stating conclusions. I do not aim at demonstrating anything in this essay, and I entertain little hope of persuading any except those already convinced. My purpose is much more modest. It is to direct attention, to re-awaken concern for supposedly settled situations, to occasion new reflections on familiar themes. In a word, I hope either to encourage or to provoke some of my readers to think again on causation, dualism, and interaction.

That which started me on these second thoughts may or may not be equally effective with my readers, but I can think of no better stimulus and I shall begin with that. It was the chance conjunction of a number of circumstances, not one particularly notable in itself. Here they are: (1) a re-reading of Berkeley and Hume resulting in an enhanced sense of the cogency of their reasoning; (2) a re-reading of Moore's Refutation of Idealism and a glance at a few sentences in the introduction to Sabine's History of Political Theory in which Sabine states his acceptance of Hume's destructive analysis, these two events resulting for me in the enhanced sense that Moore did not refute Berkeley and that Sabine outdid Hume in making his practice contrary to his professed convictions; (3) the insistence of a scientist of my acquaintance that, although constancy of correlation is an important criterion, what science seeks to discover is not that, but causal connexions; and (4) the realisation that I myself regarded physical bodies as certainly real and so did most of my acquaintances, philosophic and scientific.

For philosophers this is a pretty kettle of fish. Philosophy appeals to reason, Berkeley and Hume are outstanding among philosophers for valid reasoning, and we who call ourselves philosophers refute neither Berkeley nor Hume, yet hold contrary beliefs. In this there is something amiss. What is it? Our practice? The arguments of Berkeley and Hume? Our estimate of what precisely those arguments prove? At that point, with those questions, my second thoughts began.

My procedure in this paper will now consist not in a detailed narrative of the course of those second thoughts but, as was explained, in an indication of the data and arguments considered, and a statement of the conclusions reached. In the matter of causation the primary data are the facts of practice found alike in ordinary living and in science. I suggest that those facts exhibit and substantiate the following. As regards causality, we possess a working understanding of the causal relation as involving agency, transeunt production or determination. As regards particular causal connexions, we possess reliable knowledge of some of these. These two conclusions go together; the same facts illustrate and substantiate both. It is the blow of the hammer that drives in the nail; the waves rock the boat; the increased flow of gasoline to the carburettor produces, when vaporised, mixed with oxygen, introduced into the cylinder, compressed and ignited, the increased force which moves the car faster. Those relations are correctly termed causal for the reason that in them there is transeunt control; and in those specific cases we know what is the cause and what the effect.

Having stated this much it is important that I make precise and explicit what is not claimed. No account is given of the origin of our knowledge of what a cause is. No explanation is offered of how agency operates. The generality claimed is in each case specific and limited : given, and within the complex of, waves, boat, and rocking, it is then said that the waves rock the boat.

What about Hume's arguments? They still stand. But care is needed in stating what they prove. He demonstrates for us, as for himself, that the principle of causality in its universal form, every event has a cause, has neither a priori nor a posteriori proof; he still proves that, antecedent to experience, we have no knowledge of what specific effect will result from any given cause; his exhibition is still conclusive that in all alleged cases of causal connexion between events external to ourselves, the nexus itself, even if present, is not the object of a simple perception as colours and sounds are. Further, Hume's analysis of voluntary activity as a source of our knowledge of causation is as inconclusive now as it was when he made it; and his own account of that origin in terms of habit and expectancy, is as mistaken. But what Hume proves does not touch what has been stated as to the knowledge of causation possessed and used by all of us. For to say that a genuine cause involves agency and that there are such, is not to say that every event has a cause; and to assert that the waves rock the boat is not to claim that this was known antecedent to the empirical evidence on which the knowledge is based. In general, all that I have said of normal practice is coherent with and involves no contradiction of Hume's genuine demonstrations.¹

¹At the present time it is fair to assume at least this much agreement about Hume: (a) His contention that, as regards causation, we perceive only a sequence of impressions, is an incorrect statement of the facts. (b) When such sequences are observed, they are not accepted as examples of causal connexion—cf. Kant's second analogy and the standard fallacy, post hoc, ergo propter hoc. (c) When perceptions of causal connexions are made, the epistemic claim is that the connexion is not between 'impressions' but between things or events. If to this last it is objected that it begs the question by assuming that we perceive objects, the proper reply is that this objection is not against the perception of causal connexions, but against the perception of objects in general. But in this essay, the existence of an external world and our knowledge of it are not in general under discussion.

Putting together the two types of evidence thus far presented, I find certain further statements are warranted. Hume's proof that the causal nexus is not the object of a simple perception, the debatable cases of voluntary actions being excluded, and our actual knowledge of particular cases of the causal nexus, provide, when taken together, the basis for the conclusion that this knowledge is an interpretation of perceived data. We perceive certain things, and by an interpretation we perceive also that there is a causal connexion of this with that. I hope that this use of the word 'interpretation' will not itself be misinterpreted. I need a generic term to cover without prejudice cognitions ranging from intuitions or spontaneous inferences to perceptions guided and developed by innate ideas. I say 'without prejudice' because in this essay I am not concerned with the details of how we know there is a causal nexus, but with the fact that we do know what such a nexus is, and with certain very important aspects of that knowledge. These are: as Hume proved, the nexus is not the simple object of a simple perception; as experience proves, our knowledge arises in perceptual activity. Therefore I call it an interpretation. Further, the genuineness of that knowledge is not to be confused with the validity of specific cognitions of this cause of that effect. We can and do make mistakes in particular cases; perhaps any instance selected for examination may be open to question; but even if that be granted, it does not prove either that all particular instances are mistaken, or that our knowledge of the nature of the causal nexus is illusory. It may also be true that the knowledge we have of causation is not complete. Even so, that does not prove that the partial knowledge we do have is spurious.

To say that something is an interpretation is neither to eliminate it nor to establish that it is false. But it does follow that our cognitions of causal connexions, being interpretations, are inseparable both as regards significance and truth or falsehood, from the specific situations which they interpret. The general statement, 'one event causes another', has little significance, and its truth in strictness cannot be determined until it is given a specification such as 'hammer blows cause nails to penetrate wood'. That is the first result of these second thoughts : judgments of causal connexions are interpretations of empirical data, and their significance and truth are tied to the situations they interpret.

A complementary conclusion, negative in form, is obtained by considering together our knowledge of specific causal connexions and Hume's demonstration that no study of an alleged cause can disclose, prior to its performance, what its effect will be. From this I draw the conclusion that no restrictions and no specifications based simply on our knowledge of the nature or character of the factors being considered, are ever legitimate as regards possible causal relations. Given that X is a leopard and Y is a leprechaun, nothing can be said about possible causal relations between X and Y simply on the basis of our knowledge of what a leopard is and what a leprechaun is. It is possible that the sight of a leprechaun makes a leopard change his spots. I know of no empirical evidence on this matter. Lacking such evidence, it is idle to argue the matter one way or the other on the ground that the kind of thing the leopard is and the kind of thing the leprechaun is, make this or that causal relation possible or impossible.

This negative conclusion, and the earlier positive one that judgments of causal connexion are inseparable from the specific situations they interpret, do not rule out inductive generalisations about causes and effects. What is established is that no inductive generalisation about a particular type of causation can itself be generalised into a universal rule or metaphysical principle about causality überhaupt. Every inductive generalisation stating a causal rule is limited in its application to the type of situation out of which it has been derived and for which it is the generalisation. Within that restriction, the reliability of the inductions is a function of the appropriate evidence.

Here are two examples of the application of this conclusion, one positive, the other negative. It is asserted in the science of physics that in a closed system where there are exchanges of energy, energy is conserved. I am aware that there are competent men who question the conclusiveness of the evidence supporting this rule. That does not concern me. What does concern me is that it be seen that this rule is not a principle of causality with universal application. For we have no knowledge of causality such that we can say that causation is possible only in terms of exchanges of energy in closed systems. So far as I can learn, we have reliable knowledge of a specific type of causal relations involving energy in both cause and effect, and in such situations, energy is conserved. The empirical evidence suggests the generalisation and supports it. If it does, it does. Period. Whether or not there are other types of causal relation, for example, one where energy is found in the effect but not in the cause, or vice versa, is another and an empirical question. That is the positive illustration.

The negative one is equally simple. Descartes, it is commonly agreed, held that an imperfect being could not of itself produce the idea of a perfect being, appealing to the principle that there must be as much reality in the cause as in the effect. Much as I admire Descartes, I can find no warrant for such a general rule about causation. What causes can produce what effects, can be settled only by performance. If I may be permitted an obiter dictum on Descartes' problem, I should say that in it only two things are clear, all else obscure. The first thing that is clear is that we are imperfect; the other is that we men have in our possession no a priori knowledge of causation prescribing what we can and cannot produce in the way of ideas—or of anything else for that matter. Reliable knowledge of causal relations can be and is established in the actual situations in which they operate. What is then found, is found; what not, not.¹

¹It might be added that there is adequate empirical evidence that imperfect beings such as we men are cannot by their own efforts make themselves perfect. It is our causal capacity with reference to the production of ideas of perfect beings which still seems unsettled.

To return to Hume, there remains for consideration that the principle of universal causality lacks both a priori and a posteriori proof: we are not in any position to say, 'Every event has a cause'. Yet we do know that some events have causes, and that others as causes produce certain other effects. This knowledge is limited, but it is knowledge. Hume was right about the principle, wrong about us. We are neither as ignorant nor as bamboozled as he tried to make out. Yet his being right about the principle has a significance for us we have largely failed to grasp. Granted both that we know some causal connexions and that the universal principle lacks proof, we should then be ready to find in experience not merely two possibilities but at least three. The first is the ordering of events in terms of causation involving exchanges of energy. And we do find some of these. The second is the occurrence of events with a de facto orderly pattern yet lacking causal origin and connexion. The existence of such events is debatable. The third is the connexion of events in a settled, reliable fashion yet the connexion is not interaction. This third possibility readily permits verbal juggling, not all of it perverting. One way of describing this possibility is to say that while all cases of compulsive or energy determination are cases of cause and effect, not all cases of cause and effect need be cases of compulsive determination. Personally, I prefer to say that there may be a relation in which there is a settled and reliable causal connexion and yet no interaction—i.e., no exchange of energy. In any case, whether this possibility, however named, is actual, remains to be seen. And that is an empirical question.

I now come to my second thoughts on Berkeley. As I have already said, the so-called refutations of Berkeley on careful examination are found to be no refutations at all. On the other hand, an equally careful re-examination of Berkeley disclosed that his own proofs against matter are likewise inconclusive. Berkeley does prove a number of important things—he proves, for example, the inseparability and equal epistemic standing in perception of primary and secondary qualities—but he does not disprove the existence of material substance. The inconclusive-ness alike of the argument against materialism and against idealism, both terms being used in the sense appropriate to Berkeley, indicates that physical objects, or alternatively God, as the independently existing cause of our external perceptions, are interpretations of the data of experience. Berkeley admits as much when he argues that God, being active, is a more likely and comprehensible cause of the content of perception than inert matter could be. This is to say, in Berkeley's terms, that neither God nor material substance is the object of a simple perception: we perceive tulips or strawberries as the case may be. Physical objects or God, as the independently existing causes of our perceptions, have epistemic standing as interpretations for which arguments pro and con are offered. Berkeley's own arguments in favour of God appeal to causation in two distinct ways, first, to the general principle of causality, second, to specific considerations relevant to the appropriate cause in a particular kind of causal relation. These two arguments must be kept distinct.

Starting with the first, we find that Berkeley recognises, as completely as any realist, that when we open our eyes, we see not what we will, but what is present in our visual field. To account for this, he tells us, there must be a cause other than what is perceived. What we perceive has its being in being perceived, yet that we perceive this specific 'what' rather than another, is not in our power to determine, and therefore must be determined by something else. This is, at least in part, an appeal to the general principle of causality, 'every event has a cause'. Such an appeal is, as we should now know, invalid, because of Hume's demonstration that this principle has itself no independent standing. Of course, as Hume usually forgets, the contradictory principle, some events have no causes, has no independent standing either. But I shall not go further into this question, for it is in fact the general problem of the justification, if any, of the belief in an external world. One point only I wish clearly noted. The view that our perceptual experience is of an independently existing source is, in its generic form, a view on which Berkeley, Dualists, and modem Realists agree. They agree there is some sort of independent reality—physical things, God, or something else.

What particular kind of external reality we are to believe in, is the next question. Berkeley rejects a material world and proposes God. His reason, in brief, is that God is by nature active, matter inert; God as mind or spirit is like unto us in whom the perceptions as effects are to be produced, matter is different in kind. In this Berkeley directs our attention to the differences between matter and mind, physical things and ideas, and then he takes these differences to be evidence for the impossibility or improbability of any causal activity of mind with matter, matter with mind.

There are two elements in this argument against matter.¹

¹ In appraising this and the next few paragraphs it should be remembered that this essay does not claim to be a complete examination of Berkeley's philosophy. In particular, Berkeley's polemic against abstract ideas is not considered because such consideration is not necessary for the subject in hand. Yet even in a footnote this much can be stated. Berkeley's polemic proves that we cannot have either in perception or in imagination an image of figure in general: images are particular. Descartes had noted that fact before Berkeley. But Berkeley does not prove either (a) that we have no knowledge of universals—this is a commonplace—or (b) that we have no knowledge of, to be specific, figurateness whose specific character is not to be identified with any one of the sensory qualia, such as colour, found in our sensory images of particular figures. Seen figures, in perception or in imagination, are coloured figures. But no one has ever demonstrated that our knowledge of figurateness must be in visual terms, or tactual terms, or any other specific sensory qualia. There is evidence that the contrary is the case : that we can know, not sense, figure in terms of sheer boundaries in pure space.

The one is the contention that matter has a specific, essential character, inertness. The other is the reliance on an alleged causal rule: matter and mind, because unlike in kind, cannot causally interact. Let us first consider the rule. It may be understandable that Berkeley regarded that rule about causes as valid. But for us to do so after Hume's work, is to disclose a singular ignorance and stupidity. As has previously been brought out, we have no general rule about what kind of thing can interact causally with what other kind, or what cause can produce what effect. Prior to actual performance no man knows what can cause what. Matter cause an idea? Mind control body? Why not? The only proper question is neither why not, nor why, but simply, does it? And to get any answer we must reverse Berkeley's order. We must first determine whether or not there are physical bodies and mental ideas, whether or not mind is active, matter inert. If such questions are answered in the affirmative, we can then look for empirical evidence relevant to the question, does inert matter cause ideas in active minds?— do minds determine motions in bodies? That order of questions, and precisely those questions in that order, are the clues to reliable conclusions in this total situation. To discover that was for me an important achievement of my second thoughts.

I now direct attention to Berkeley's assertion that matter is inert. Strictly this was not his own view but that of those who believed, as Berkeley did not, that matter has real existence. He took it from them and then used it as an argument against that belief. In the taking he was certainly justified. The inertness of matter was a cardinal doctrine of the physical science of Berkeley's day, in fact, if not in word. It was given definitive expression in the Newtonian law "every body continues in its state of rest or motion in a right line unless acted upon by an external force". The use of the term 'motion' should not deceive us, nor should a shift to such terms as energy or electricity. The crucial phrase is "unless acted upon by an external force". If that holds, then matter is in itself perfectly inert. At rest, it cannot begin to move; in motion, it cannot stop, cannot swerve, cannot even slow down—of itself. Something external is required. Berkeley was completely right in saying that anything thus described is inert. I, myself, am not one of those philosophers who readily challenge the truth of the pronouncements of science. Towards science I am as a dutiful child to an eminently respectable adult, accepting with deference all pronouncements on matters of fact, no matter how extraordinary they may appear. It is not questioning its truth to make the following notations on this scientific account of matter as inert. The empirical evidence offered by science to confirm its law is derived from carefully selected and restricted situations, and the theoretical considerations are similarly restricted in their reference. In consequence, the inertness of matter is well established as regards the behaviour of physical bodies in a specific type of situation—exchanges of energy in a system containing only physical bodies—but there is no warrant for claiming that this inertness is an exhaustive account of the capacity of matter in all situations. Precisely as in the case of causation, we have a generalisation which is true in its proper reference; but the range of this reference may not be coincident with the entirety of the empirical world. Further, the inertness of matter in the scientific sense is strictly formal. That is one reason why it is so useful to science and so deceiving to ordinary people. This formal inertness assures science that no physical body will ever of itself upset an equation by showing up with a velocity value on one side underived from the other; that energy systems will maintain themselves as a whole if there is no interruption. This inertness is perfectly compatible with ceaseless jumpings, reboundings, even atomic explosions. Finally, I call it formal because this inertness is not to be identified with, although it has a relation to, perceptual or sensory inertness.

Sensory inertness brings us to what Berkeley did say for himself, that ideas are inert. Disagreement with his final conclusions or awareness that for us there is a persistent ambiguity in his use of the term idea, should not blind us to the correctness of Berkeley's observations. Resistance to our efforts, adverse space occupancy, what Dr. Johnson intended us to understand when he dashed his foot against a stone—and the Bishop certainly needed no instruction from the Doctor—these are the particulars of the inertness Berkeley and we certainly find in our commerce with things. This inertness is crucial in one of Berkeley's arguments. Inertness, accepted as an essential character, gives positive content to the term 'material substance', which otherwise is as empty as Locke admitted when he said it was a something he knew not what. But Berkeley believed that extended, inert matter could not be the causal source of our perceptions. It has been one of the primary purposes of this essay to indicate that there is no warrant for that view of matter. Let material substance be extended and inert, that settles nothing as to its causal capacity with reference to minds. Only the trial of performance can be decisive. With that seen clearly, my second thoughts, guided by Berkeley but attending strictly to the empirical evidence, brought me to the conclusion that material substance exists, with inertness as one of its chief attributes. This, for me, meant dualism. I have not felt it necessary in this essay even to state that I regard myself as certainly real, a thinking substance. First thoughts, guided on occasion by St. Augustine, Descartes and Berkeley, suffice for that. Only for the existence of material substance were second thoughts required. But assiduous reflection on the evidence did achieve that result. Mind and matter are different in kind, both are substantial, both are real.

But that is not all. The inertness of matter is bound up in fact with the type of causation involving energy exchanges, causation by the application of external force. Instances of such causation can be discovered. There are also situations in which something else is found. Attending particularly to inertness and causation in terms of forces, I discover a constant and crucial difference among bodies. Bodies are inert in relation to other bodies and to my body, and my body to them : any change of motion in one requires energy from an external source. But my body is not inert to my mind, nor my mind to my body. All the genuine empirical evidence I can get indicates that my body is responsive to my will, and my mind responsive to my body. The relation is causal, but not interactive. No motion, no energy is transferred. To a volition in the mind the body responds by what is appropriate to its mode of existence, namely movement. The energy involved is the body's own, none comes from the mind, for the mind strictly has no energy. To a motion or energy state in the body, the mind responds by what is appropriate to its mode of existence, namely feelings, sensations, perceptions. Then the activity is the activity of the mind, none comes from the body, for the body strictly cannot think or feel. Yet this relation is, causal. Volitions determine and control bodily motion. The body, inter alia, determines and controls what the mind feels and perceives. Mind and body, each without loss of essential differences, exercise causal control each over the other. The control is limited; it is specific in that the effects are always those appropriate to the kind of substance in which they occur, motions in matter, ideas in minds; but the causation is genuine, actual. Against this, the conclusion of this set of second thoughts, what evidence can be brought? That matter and mind because different cannot have causal connexion? There is no such rule or principle established on a priori, rational, or a posteriori grounds. The genuine empirical evidence is all the other way. The law of the conservation of energy is not relevant, and in any case it is not contravened. It is no objection that this imputes to matter a new quality, quite different from inertness. What if it does? It leaves untouched the inertness of bodies vis-a-vis other bodies, the only situation in which inertness is verified or important. This view asserts a character I have called "responsiveness" of both bodies and minds in the specific relation of living union. It does so on the basis of empirical evidence.

On that note my second thoughts rest. I care little for labels, but for those who like them I shall label this position. It is resolute empirical rationalism. It is resolute because it is not frightened by names. It is empirical because it holds to the evidence. It is rational because it is the result of reflections judged by the criterion of truth, and because it makes sense.