Muddy Waters

Causation is one of those philosophical topics that drive you up the wall. As soon as you start to think about it you draw a complete blank. As Hume observes: “There are no ideas, which occur in metaphysics, more obscure or uncertain, than those of power,force, energy, or necessary connexion” (Enquiry, p.45).[1]The cement of the universe is so much muddy water. Of course philosophers have done everything in their power to hide this fact from themselves, even going so far as to try to reduce causality to mere regularity. Hume’s own view was that causation is real but incomprehensible (by us). It is neither an affair of the senses nor of reason: we have no sensory impression of necessary connection (which is definitive of causation) but neither is causation grasped by reason (like logic, arithmetic, and geometry). It is a real relation between things but it is not revealed by perceiving them or by merely thinking about them. It fits neither empiricism nor rationalism. It sits uncomfortably between the two, awkwardly and inscrutably. No matter how much you gaze at an object or reflect on it you will never discover causation (as it exists in that object). But perception and reason are our only faculties of knowledge, so the mind draws a blank on the nature of causation. Yet we constantly refer to it, rely on it, and assume its reality. Evidently we can know that it obtains, relying on the observation of regularities of nature, but we can’t fathom its inner nature—or even fathom our lack of fathoming.

The problem concerns not just particular causal relations but also the notions of law and power (disposition, capacity, potential). Objects fall under causal laws and have causal powers: this is why they have the effects they have. But laws and powers are at least as inscrutable as causal relations between particulars; this is an interconnected knot of problems. Perhaps the notion of power concentrates the problem most acutely (as Hume intimates): how are the potential effects of a cause contained in it? Are the effects somehow already present in the cause? Does the cause “refer” to the effects? Are there shadows or signs of the effects lurking silently in the cause? But you can’t discern anything like this if you examine the cause, even going down to its elementary constituents. When a moving object imparts motion to another object by collision is the other object’s motion somehow prefigured in the moving object? It had the potential of creating that effect, so doesn’t it already contain it in someway? What ispotential? It’s a bit like the way the meaning of a word “contains” its uses: they are implicit in it, packed into it—but what does this way of speaking amount to?[2]The problem of causation is how an object can contain what it does not contain. If we think of a cause as a conscious being for a moment, it is as if it knows what effects it can bring about, but only unconsciously: it doesn’t have these effects before its consciousness, but it is subliminally aware of them—they are implicitly known not explicitly known. But that looks like a dodge: they aren’t anticipated in any way that we can discern—the mind of a cause is blank about its future effects. Yet it has the power to bring precisely these effects about, and this power is internal to it, so… Thus the waters fill with mud.

What can we say positively about causation? The logic, semantics, and conceptual analysis of “cause” are not so baffling. Thus “xcaused y” expresses a relation that is irreflexive, asymmetric, and transitive: nothing can cause itself, effects can’t cause causes, and the effects of effects are caused by the initial cause. Semantically, it is plausible to suggest that “cause” generates a transparent context and expresses a relation between events (though this view is not without its critics). The concept may also be analyzable in terms of counterfactuals or other necessary and sufficient conditions. So it is not that we can say nothing about the word and what it means—and much philosophical energy has been expended on these worthy tasks. But they don’t touch the underlying metaphysical and epistemological questions, the ones so memorably raised by Hume. What exactly is causation, and how do we know about it? Specifically, what is it to have a causal power, and how can we know causal powers? There are suggestions—there are always suggestions. One suggestion is that a causal power is identical to a structural property of the object, as it might be molecular structure. But this just pushes the question back: how is the power present in the structure? Isn’t it as invisible as ever? Nor can it be excogitated by pure reason. It can’t be seen and it can’t be deduced—it flouts both empiricist and rationalist epistemology. It isn’t a posterioriand it isn’t a priori. It is a peculiar kind of fact, being neither perceptible (even by extended perception: microscopes etc.) nor rationally apprehended. As Hume would say, it is neither a “matter of fact” nor a “relation of ideas”; it hovers ambiguously between the two. No wonder there has been a marked tendency towards elimination: causation must either be reduced to facts less problematic (regularities, dispositions to project) or eliminated outright. To accept it as it is runs into insurmountable metaphysical and epistemological difficulties. Indeed, it threatens to bring down the most fundamental structures of philosophical thought.

So why not just bring them down? Because we have nothing to put in their place, that’s why. It is not as if we have some otherway to think about causation that we can substitute for the old inadequate dichotomies: the waters are thick with mud and our vision fails us. It is really a horrible problem. Best not even to go near it; just leave it alone to fester. But maybe we can articulate the problem better, gain a better sense of its dimensions and density. Can we at least pinpoint whyit is so difficult? It is in some ways worse than the problem of consciousness, because in that case at least we know what we are talking about—we don’t just referto consciousness, we experienceit. But we don’t experience causation (as opposed to its symptoms), despite our readiness to refer to it. We refer to a we-know-not-what. Appealing to mental causation won’t help, despite our immediate acquaintance with the mental phenomena between which causation holds: for mental causation is as opaque as physical causation (as Hume noted). We can say that physical causation is no less activethan mental causation; the will is not somehow a livelierform of causation. Nor is causation by physical contact more transparent than causation-at-a-distance, since its operation is as obscure there as it is in the remote case. In this respect old-style mechanism offers an illusory paradigm of transparency (this was Hume’s central insight, in effect): it isn’t that causation by contact is quite clearly grasped while causation-at-a-distance must be deemed “occult”. Neither is really intelligible to us, not when you get right down to it. Hume’s billiard balls hit each other, unlike orbiting planets, but their causal powers are no more evident to sense or reason than gravity. For some people this was taken as a reason to eliminate causation altogether from physics, and one can appreciate the motivation.

Can we be more constructive? I think we can say two positive things, though nervously. The first is that nature must be more tightly interlinked than we tend to suppose going by the appearances: causation connects things because any effect of a cause must be somehow written into the cause (though not in a way we clearly conceive). The colliding billiard balls don’t appear to sense perception as having any intelligible connection; nor can human reason discern any such connection: but they must somehow be intelligibly linked. The laws of nature essentially relate separate things, because causal powers are essentially powers to bring about certain specific effects: an object xhas the power to make an object yhave a property P. There is thus more “holism” at work in nature than is apparent to our epistemic faculties. We could introduce the idea of the “causal boundary” of an object to signify the class of objects that fall within its causal reach—for instance, the class of solids a given liquid can dissolve. This class falls within its causal boundary but not its spatial boundary. Then nature will be said to consist of the totality of such causal boundaries—these are the true units of nature.

The second thing we can say is that whatever causal powers are they must be very different from their manifestations in observable phenomena. This is because the manifestations never add up to a causal power, as it exists in objects. It can’t be mere regularity and it can’t be a “categorical base” (e.g. molecular structure)—these are not what the power is or else we wouldknow what it is. Powers must be as different from their manifestations as mental states are from behavior—perhaps more so. Potentiality must be different from actuality; yet the two must be intimately related. I can’t tell you how potentiality differs from actuality because of its obscurity, but it evidently does differ, dramatically so. (Or else actuality is merely the way potentiality looks to our senses and doesn’t go deep ontologically speaking.[3])

Would other things become clearer if we had a better grip on causation? Anything in which causation directly figures would be—laws of nature, the origin of the universe, the operation of fundamental particles. But it might also help with the mind-body problem and the free will problem: How is the mind caused by the brain? How are free actions caused? Certainly this is an enormous gap in our understanding of nature, what with the ubiquity of causation, but a gap there seems little prospect of filling. The water may remain forever muddy.[4]

[1]I will be accepting the “skeptical realist” view of Hume in what follows, according to which causal powers are real existences that defy our limited understanding, not the positivist interpretation of Hume according to which the concept of causal power is incoherent and should be rejected.

[2]I am alluding to Wittgenstein’s discussion of meaning and use in Philosophical Investigations. He explicitly connects this issue to that of causal powers in the sections on machines (193, 194).

[3]Such a view is suggested by the thesis that all properties consistof causal powers. Then there is nothing to nature butpowers.

[4]This essay is intended to reflect the inadequacy of our understanding of causation, containing little in the way of genuine illumination. But perhaps it serves to scratch the depths.

The blacklisting of me continues, irrational and foul though it be. When will Americans ever learn?

I am tired of hearing people say, as if it were uncontroversial, that they are “Humeans” about laws and causation, where they mean to defend the “regularity theory”. That was not Hume’s view, as has been amply demonstrated. They are claiming the authority of a great philosopher for a position he did not hold (in fact, I know of no great philosopher who held this view). What they are following is Ayer’s version of Hume not the real Hume. So they should declare themselves not Humeans but positivists. But that doesn’t sound so laudable since the fall of positivism. They should at least acknowledge that many people reject this version of Hume as well as the doctrine falsely attributed to him.

Can There Be Subjective Facts?

I invite you, intrepid reader, to accompany me on a journey into the heart of darkness—into a region of utter obscurity. It will test us both, but it should be worth the effort, including the bouts of sickness and insanity. No, it won’t be that bad—just a bit of a headache and some mild nausea.[1]For I propose that we explore the question of whether reality could include facts that can only be grasped subjectively. By that last phrase I mean grasp that depends on occupying a particular point of view. The background is familiar from bats and blind people and alien forms of consciousness: we can’t grasp what it’s like to be a bat because we don’t share bat experience (we don’t occupy the bat’s “point of view”) and blind people can’t grasp the nature of color experience because they don’t have color experience (they don’t occupy the “point of view” of color perceivers). Some facts (properties, meanings) can be grasped without oneself instantiating them—for example, being an elephant or being a dodecahedron—but others (allegedly) can’t be grasped without being an instance of them oneself. You can’t know what it is like to see red without seeing red yourself, though there is no difficulty about knowing what a mountain is without oneself being a mountain. This suggests that there are subjective facts—facts that can only be grasped subjectively, i.e. from a specific point of view, i.e. by instantiating those facts oneself. It suggests that it is in the natureof certain facts that they can be apprehended only by beings equipped with specific types of experience, viz. those they are endeavoring to apprehend. Such facts might be called “intrinsically subjective”. By contrast, there are other facts that are “intrinsically objective”, i.e. they have a nature that allows them to be grasped from many points of view (or from nowhere). Thus some facts are subjective and some are objective, and this classification is written into them, part of their essence. The facts dictate the conditions of their being known (grasped, apprehended). My question is whether this is the right way to look at the matter.

Facts about consciousness are supposed the primary kind of subjective fact. If we define consciousness as consisting of the realm of facts there is something it is like to obtain, then we can say that facts about what states of consciousness are likeare subjective facts, i.e. they can only be grasped by sharing them. It will be convenient to introduce some abbreviations here, so let us call facts about what it is like “W-L facts” and let us call knowledge that depends on sharing the facts in question “S-I knowledge” (short for “self-instantiation-dependent knowledge”). Then we can say that W-L facts are intrinsically known in an S-I manner: that is, we can only grasp facts about what an experience is like by ourselves having experiences of that phenomenological type. In brief: W-L entails S-I. This is a move from the metaphysical to the epistemological: it is built into the nature of the facts (metaphysical) that they are grasped or known subjectively (epistemological). This thesis figures in an argument against materialism: mental facts can’t be physical facts because the former are S-I while the latter are not, i.e. the former are subjective while the latter are objective (understood in the epistemic way defined).[2]One response to this argument is to maintain that so-called objective facts are not as objective as you might think; in fact, they are subjective. This is because what we call the “physical” really consists of facts about consciousness, though consciousness of a primitive and alien type (I am talking about panpsychism and its spiritual cousins). The thought is that the problem of emergence can be solved by supposing that matter is shot through with mind, perhaps ismind at a fundamental level. That certainly looks like a promising strategy: straddle the mind-body gulf by edging the body in the direction of the mind. Mind arises from mind—not so inconceivable a feat. The point I want to make about this sort of approach is that it sharpens a more fundamental problem about consciousness, viz. how it is that consciousness has being at all. If reality consists of consciousness, and consciousness is W-L, and W-L requires S-I, then reality is intrinsically such as to be knowable only from a specific point of view. If the basic mental level of reality consisted of elementary consciousness of the type exemplified by bats, then only beings equipped with this type of consciousness could grasp the nature of reality. Reality would be intrinsically subjective in the sense that only certain types of being could have an adequate conception of it—and that could be a small subclass of the class of all possible knowing beings. Reality would (could) exist only forthis privileged class of being—for everyone else it would be strictly inconceivable. The question is whether that is intelligible: could reality consist of such essentially subjective facts, partially or totally? Could there be facts that could onlybe grasped from a particular point of view? Could there be facts that rule out any objective conception? Specifically, could W-L facts rule out a non-S-I mode of conception? In yet other words, could facts about consciousness resist any attempt to include them in an “absolute conception” of reality? Are they inconceivable objectively? My thesis will be that this is wrong—such facts cannot be intrinsically subjective. Facts about consciousness are not necessarilygrasped subjectively: even W-L facts must be graspable objectively. To put the point as strongly as possible, everythingabout consciousness must be conceivable objectively—and this is a fundamental metaphysical truth. There cannot, as a matter of metaphysical necessity, be purely subjective facts. There can be purely subjective conceptions, and we might well be confined to conceiving consciousness in this subjective way, but no fact(property, states of affairs) can be subjective in the intended sense. It is a category mistake to describe facts as subjective of objective; those appellations belong only to conceptions of facts—thoughts, representations, or ideas. The subjective-objective distinction is an epistemic distinction not an ontological one. But this leaves us in a very precarious position with respect to our understanding of what consciousness is, as will emerge.

First, let us note that experiences admit of several types of objective characterization drawn from within our current conceptual scheme. We can bring conscious experiences under causal descriptions (“the cause of this”, “the effect of that”), and we can express their relations with physical facts (such as correlated brain states). It is also possible to characterize them in abstract structural terms as well as mathematical terms (as in psychophysics). There is more to a conscious experience than what it is like for the subject: we can conceive such states in many different ways. The hard question is whether the W-L aspect can in principle be conceived in a way that doesn’t depend on sharing the point of view implicit in it. Is there a way of conceiving it that is available to beings that do not themselves instantiate the property in question? We naturally think of God here, but that raises questions it would be best to avoid (Does God instantiate all points of view? What is his knowledge like anyway?). One possibility is that this property exists but is completely ineffable (recall my warning about dark obscurity at the beginning): there are objective W-L facts, but no being could ever conceive them. All beings must conceive these facts subjectively or not conceive them at all: if the latter, then reality contains ineffable ingredients. A less alarming position is that possible beings might be able to grasp W-L facts without reliance on their own point of view: they might be able to form a conception that captures these facts and yet is independent of how they view things.[3]For example, they might be able to forge structural descriptions that encapsulate the facts and yet can be grasped by anyone whether they instantiate these facts or not. Then the nature of the facts would not dictateadopting a subjective conception of them; an objective conception would be available too. Of course, we find it hard to understand what such a conception would look like, given our own concept-forming faculties, but the general idea is intelligible enough—other beings (or us in the future) might be able to conceptualize consciousness in ways that transcend our current self-centered subjective conceptions. Then reality would not contain properties that can only be conceived subjectively; it would not be populated by properties that inherently favor one way of viewing the world over another.

Here it is important to make a distinction between two ways in which a fact might resist being conceived objectively. We have the modal truth: Necessarily it is impossible for us (fully) to conceive consciousness otherwise than subjectively. That truth might, however, have two possible sources: it might stem from the very nature of the fact in question—itdecides how it can be known; or it might be that ourfacultiesrestrict us to subjective conceptions—not the facts themselves. And we must not assume that just because we can’t conceive of W-L properties in objective terms that it is in the nature of those propertiesthat this is so—the reason might stem from our own cognitive make-up. Wenecessarily can’t conceive them in any other way, but theyimpose no restrictions on how they can be conceived; it is how we interact with them not what they are in themselves that gives rise to the necessity. It seemsto us that it is in the nature of color experiences to preclude a blind man from grasping them, but it may be that the difficulty arises rather from our distinctive human way of forming concepts of experience. Perhaps the blind man in other types of cognitive being has no trouble conceiving of experiences that he happens to lack. What is a necessary truth for us need not be true for everybody; in particular, it may not be the property itself that is the source of the problem. Thus the facts in question can be grasped objectively as a matter of principle even though we can never so grasp them. Then we would not have to admit that reality contains facts that are intrinsically tied to a specific point of view.

The general point here is that the terms “subjective” and “objective” are not descriptive of facts themselves but of ways of conceiving facts: no facts are subjective or objective considered in themselves. Physical facts can be conceived subjectively or objectively, relatively or absolutely, and the same is true of mental facts. If we are confined to subjective conceptions of mental states, that is a truth about us not about them: “subjective fact” is an oxymoron, unless it means “fact we happen to know about subjectively”. So consciousness exists neutrally between subjective and objective, despite our bias in conceiving it towards the subjective. But that means that we don’t grasp the manner of its existence: how we conceive it doesn’t tell us how it exists in reality, as an objective conception of it would. Thus we have an inadequate grasp of consciousness—in an important sense we don’t know what it is. What it is would be revealed from an objective standpoint, but that standpoint is inaccessible to us; we are limited to a standpoint that gives at best a partial conception of it. So we don’t understand what the existence of consciousness amounts to—what it isfor it to exist. That is the really hard problem: trying to figure out what we are talking about—trying to arrive at an adequate conception of consciousness itself. Even if we could solve the problem of emergence, say by invoking panpsychism, we would still be left with the more fundamental problem of discovering what the existence of consciousness really consists in. Granted, we know somethingabout consciousness, but we don’t know the most vital and basic thing—what its mode of existence is. We only grasp its existence-for-us not its existence tout court. It must, as a matter of metaphysical necessity, have an existence apart from our subjective mode of conceiving it, but we are blocked from discovering what this objective existence amounts to; at any rate, we presently have no conception of what it might be. This means, of course, that we don’t grasp the nature of our ownexperiences, let alone the bat’s, since they too must have an existence quaexperiences that is accessible from all points of view. We don’t grasp what it is like to be us(as that would be represented in an objective conception of what it is like to be us).

I have drawn a sharp line between two concepts: the concept of what an experience is like and the concept of a subjective conception. The former does not entail the latter: what it is like pertains to the nature of consciousness (its metaphysics) while the subjectivity of a conception pertains to how we think of a certain subject matter (epistemology). It is a substantive claim that the metaphysical property requires a subjective conception; and I have disputed this, suggesting instead that all facts must be accessible to an objective conception, even if not one of which we are capable. This logical point is lost if we carelessly use the world “subjective” to refer to a feature of consciousness itself andto the way we tend to conceive of it (by using our own case as model). Consciousness cannot be necessarilyconceivable only via a subjective conception, i.e. one that relies on self-instantiation, because that is an unintelligible idea of what reality is like. To repeat, reality cannot be relative to a particular point of view—only conceptions of reality can be. Facts are just facts: how we conceive them is another matter. This means that idealism cannot be restricted to how we now think of the mental: if the world is really the mind, then mental facts can’t be limited by the ways we generally conceive of them. Mental reality would have to be conceived objectively not subjectively, because its existence must be independent of the point of view a given type of creature brings to it. We may not be capable of acquiring the concepts needed to make sense of the objective existence of mental reality. At any rate, mental facts must have a nature that goes beyond what is contained in our present subjective conceptions. Those conceptions cannot even reveal what it is like to be a bat, though that latter fact is something that must be accessible in principle to different points of view. There must be a way of conceiving bat experiences that is accessible to beings that lack such experiences, on pain of making reality point-of-view-dependent. In other words, it is only a contingent fact that bat experiences can only be grasped by beings that share those experiences. There are no inherently subjective facts; subjectivity is in the eye of the beholder (which is not to say it is unreal).[4]

Colin McGinn

[1]Note to American readers: tongue is in cheek here.

[2]This argument is presented in Thomas Nagel’s “What is it like to be a Bat?” (1972), as is the general notion of subjective and objective I am working with.

[3]Nagel discusses the possibility (necessity) of objective descriptions of mental facts in The View From Nowhere(1986), chapter II.

[4]Of course there are subjective facts in other senses of the word: there are facts about conscious subjects, mental facts, facts about personal opinions.