An Identity Theory of Causation

Let causal dualism be the doctrine that cause and effect are separate things, one following the other in time: first the cause, then the effect.[1] There is no overlap between the two; they are quite distinct events. The short circuit caused the fire: one event (or state of affairs) caused another event (or state of affairs), period. The billiard ball striking another billiard ball made it move away–the striking and the moving being different occurrences. This is opposed to the view that talk of cause and effect is really talk of the same thing: we have a dualism of descriptions not a dualism of objects. Logically, the case is like mind and body: do we have a dualism of objects or a dualism of descriptions (senses, perspectives)? Granted we have causal interactions, but do these consist of a cause-effect pair, discrete and asymmetric, correlated but not identical?

One billiard ball strikes another billiard ball and that ball moves away—the former causes the latter. These seem like distinct existences. But consider: the second ball causes the first ball to slow down, perhaps stop in its tracks; it impedes the motion of that ball, which would have gone on in a straight line if not obstructed. We can imagine someone putting the impeding ball there on purpose so as to secure the result in question—to get that effect. One ball causes the other to move away (acceleration), but that ball causes the other to slow down (retardation). The causation was reciprocal. How many things happened? Can’t we say there was just one, a collision of billiard balls? There are two ways of describing it, but only one thing was done—sending one ball into another. The moving away and the stopping are aspects of the same event. What we call the effect (moving) actually caused a change in the cause (stopping). The effect had an effect on the cause. The stopping and the moving were aspects of the same basic event—the collision. Collisions affect both parties—the impact is mutual. There was an event of collision and it can be described as causing a ball to move or causing a ball to slow down. Something happened between the two balls, and that is the fundamental causal reality. That something was highly localized: momentary contact between parts of the surfaces of the two balls. The rest of the balls is strictly irrelevant to the causal nexus. There was a localized touching event and that is what led to the two effects of moving away and slowing down. The touching event was the core of the causal interaction; it can be described in cause language or effect language—a touched b or b touched a. The fact is that a and b touched each other. This symmetrical touching is the causal reality. It isn’t that the cause is a separate thing from the effect—as if the moving ball were merely an effect and not itself a cause. The touching is both cause and effect; it is both what caused the moving and what caused the stopping. It isn’t that causes are active and effects passive; both are active and passive. Every causing is a case of being affected and every being affected is a case of causing. It isn’t that some things are intrinsically causes and some things are intrinsically effects, and never the twain shall meet; all causation is both. One man’s cause is another man’s effect, and one man’s effect is another man’s cause. There could be a game of billiards in which the sole purpose is to block the other player’s balls: you score points when you force your opponent to make contact with your balls, while he tries to sink his balls into the pocket without touching your balls. The effect you are seeking is his ball making contact with yours. What we call the effect is relative to the point of the game. As far as nature is concerned, everything is both a cause and an effect; you are talking about the same thing from different points of view. Objectively, there are just symmetrical collisions—mutual touching. A cause is a touching and an effect is a touching—the same touching. Hence, the identity theory.

Suppose you press your finger into a piece of soft rubber. You cause it to deform into a dent; the dent, we say, was an effect of your finger pressure. But the rubber also put up some resistance to your pressure and caused your flesh to shift a bit; the shift was an effect of the rubber’s resistance. The effect caused an effect on the cause. Yet there was only one thing going on—finger and rubber being pressed together. You pressed and it pressed. The cause was the effect: the finger pressing down was the rubber pressing back. The purpose of the act could have been to exercise the finger, in which case we would have described this as the effect and the rubber the cause; while if we wanted to make a dent in the rubber, we would have said that the finger was the cause. There is causing and effecting, and these are the same. There are causal nexuses, and that’s all. Causal relations are not one damn thing after another, causes asymmetrically preceding effects; they are simultaneous reciprocal interactions, symmetrical links. When Newton said that for every action there is an equal and opposite reaction, he was expressing much the same point: causation always cuts both ways and is essentially an action-reaction nexus. It is better to speak of causality not of cause-effect relations. Effects arecauses, as the mind is the body (for an identity theorist). Our customary talk of cause and effect is pragmatically shaped; metaphysically, there is no such sharp division. Does the Sun cause plants to grow via photosynthesis? You can talk this way, but remember that plants cause the Sun’s rays to be absorbed and used in photosynthesis (they affect the Sun’s light). These are part and parcel of the same interaction: photon absorption. Causal dualism imposes a spurious dichotomy on causality, forcing causes and effects into exclusive categories; in reality, causality consists of multiple causal nexuses that can be variously described. Thus, I prefer to speak of causal monism—a causal identity theory.

If we switch our attention to objects (substances) and causality, we get a clearer picture of the metaphysics of causation. The ontology of events in discussions of causality is only part of the causal picture. For it is clear that there are no intrinsically cause objects and intrinsically effect objects; there are just interacting objects. The Sun warms the Earth and the Earth absorbs the Sun’s energy. Objects interact with their environment all the time, acting and being acted on. If you cut down a tree, it blunts your blade; the saw is both a cause object and an effect object. There is no sense in trying to classify it as one or the other. An identity theory of object causation is the indicated position. Neutral monism about causation is the way to go. The world is not inherently divided into causes and effects; it isn’t the totality of these. The causal cement of the universe is not a compound of causes and effects (like H2O), i.e., distinct ontological categories; it is a totality of causal nexuses. We carve these up in various ways, using the concepts of cause and effect; but it is a mistake to project this division onto reality, as if we had hold of distinct metaphysical natural kinds. What we call cause and effect is merely the appearance of causal reality.[2]

[1] I think causal dualism is endemic to philosophy, especially analytical philosophy, and perhaps to human thought generally. It may help the reader to keep in mind Davidson’s treatment of causal statements, which crystalizes the doctrine in question; also, the positivist way of thinking about causality, biased as it is towards epistemological considerations (how we experience causality). I am trying to peer into the deep metaphysics of the subject, not the logical form of causal statements or how we experience and know causal truths. I want to grasp the ding an sich.

[2] This paper was difficult to write. It required me to think hard about matters that are difficult to grasp and even more difficult to express; I felt always on the brink of saying something strictly false, but essentially true. I won’t be surprised if my readers react with bemusement. It really isn’t easy to see into a reality that differs from our ordinary ways of talking and thinking. Causality has always been a very tricky subject. As Hume realized, it taxes our intellectual resources. I don’t think I’ve ever read anything on the subject I found intellectually satisfying. Most philosophers wisely steer clear of the subject. It ought to be easier.

Identity, Causality, and Caterpillars

Suppose you don’t know the true relationship between caterpillars and butterflies: you don’t know that the former turn into the latter by metamorphosis. You notice the two are often found in the same vicinity, but you don’t know that they are really the same organism at different stages of its life. You assume they belong to different species. You decide to investigate these creatures: you place five butterflies and five caterpillars in a sealed container to see what happens. Lo and behold, when you return some time later you find there are now ten butterflies and no caterpillars—but you never witnessed the transition through the chrysalis stage. You form the hypothesis that caterpillars cause butterflies—without being butterflies. For some reason the one causes the other, as if a flea could cause an elephant; it seems miraculous but you feel driven to accept it. Things cause other very different things all the time. Now suppose you conduct more diligent observations and come to the conclusion that actually they were the same organism all along, but with a different appearance at different times; you have discovered metamorphosis. Question: would you conclude that you were wrong earlier to suppose that caterpillars cause butterflies? I think not: you would hold that the causal claim is true despite the identity—that the two propositions are logically (metaphysically) compatible. And you would be right to do so, in my opinion. Isn’t it just straightforwardly true that caterpillars are the cause of butterflies (not flowers or bees or hummingbirds)?

Now suppose you are an alien visiting Earth and you observe babies and adults, but you don’t know that babies grow into adults while retaining their identity. You think these are creatures of different species that are often found in the same vicinity. After some experiments (see above), you conclude that babies cause adults but are not the same as them (they look so different). Then later, after more careful observation, you reach the startling conclusion that babies and adults are the same creatures at different stages of a single life. Do you revise your earlier causal judgment? No, you conclude that there is a causal relationship between them and that they are identical. Identity does not preclude causality. You would be right to do so, in my opinion.

Now imagine you are an astronomer investigating the planets, about whose nature you know very little; you vaguely suppose they are big fireballs in the sky. One you call “Hesperus” and another you call “Phosphorous”; you get it into your head that Hesperus causes Phosphorus—one fire in the sky ignites another fire in the sky (you don’t know how). Then you find out these two objects are actually the same hunk of rock hurtling through space. Well, I never, Hesperus is Phosphorus! Do you abandon your earlier causal judgment? You see no reason to; it has merely turned out that Hesperus and Phosphorus stand in both relations—identity and causality. The thing you see in the evening is the cause of the thing you see in the morning (what it does during the day you have no idea). Nothing else does—not the Moon or the Sun or Mars.

You have seen Superman and Clark Kent around and you begin to suspect a connection. Is one of them the cause of the other? It seems unlikely that Clark is the cause of Superman, given the former’s rather unimpressive persona, so you conjecture that Superman is the cause of Clark. No way they can be identical, given their very different abilities and sartorial style. Then one day you see Superman actually changing into his Clark Kent clothes and you decide, rightly, that the two are identical. Not only does Superman cause Clark Kent but the latter is the former, surprisingly enough. Again, you feel no inconsistency—correctly, in my view. You also note that causal contexts are not always referentially transparent (you can’t say “Clark Kent is the cause of Superman”, even though the two names have the same reference).

You are a budding physicist and chemist interested in heat, light, and water. You have been reading about molecules, photons, and gases, and you form some hypotheses: heat is caused by molecular motion, light is caused by streams of photons, and water is caused by molecules of oxygen and hydrogen. These hypotheses have notable explanatory virtues; you are hoping for a Nobel Prize. On reflection and further experimentation, you arrive at the view that heat is molecular motion, light is a stream of photons, and water is H2O. Does that mean you were wrong in your causal claims? I don’t think so: identity and causality are compatible. Water is causally explained by H2O and water is identical to H2O.

Finally, you are a neuroscientist interested in the brain-mind connection. You are convinced that the brain causes the mind and you assume they are distinct things (you might even be a Cartesian dualist). Later, you become persuaded that the mind-brain identity theory is true—should you give up your causal claims? Again, why should you—the two things are compatible. Believing the identity theory does not rule out believing the causal theory, both of which seem eminently defensible. We can have theoretical identification and causal relations. Pain can be an effect of C-fiber firing as well as being identical to it.

It is true that these doctrines are contrary to the way philosophers have talked of identity and causality for many years, but these ways of talking seem not to fit the intuitive facts. Causal relations don’t always require numerical distinctness.[1] We have to get used to new ways of talking.

[1] See my “Causality and Identity”. The present paper is intended to provide some intuitive data in support of that earlier paper. It doesn’t substitute for it.

Identity and Causality

It is generally supposed that an object cannot cause itself; causation only holds between distinct things. I will question this orthodoxy. The argument will take the form of nudging intuitions in the direction of the target thesis. The claim will be that every object causes itself (among other things); every object is an effect of itself (inter alia). The causal relation is (or can be) reflexive. I am both a cause and effect of myself. For any x, x causes x. First consider what Aristotle calls the material cause of a thing: the stuff that composes it. This is contemporaneous with the object composed, yet they are not identical objects (according to most views). I am caused (materially) by the cells in my body and the atoms in the cells: they produce me, constitute me, bring me into existence. We are not stretching the concept of cause very far in accepting Aristotle’s notion of material cause. There is more to causation than efficient causation by events spread out in time. We can also add causation by parts: a whole object can be said to be caused by its parts; together they produce or yield the object. The parts are the reason the whole exists, as the material of a thing is. But this doesn’t yet imply self-causation. How can we take the next step?

Propositions entail other propositions, but they also entail themselves—p implies p. So, entailment is compatible with identity; indeed, it might be viewed as the primary case of entailment. Propositions are self-entailing. But isn’t entailment a type of causation? It is productive, generative, a type of power; a proposition is a logical consequence of itself—a conclusion of which it is also the premise.[1] We could call entailment “logical causation”. The premise is sufficient for the conclusion; it guarantees the conclusion. Yet they are identical. But why not extend this way of thinking to objects? An object is sufficient for itself, a guarantor of itself, an upshot of itself. Like a proposition, it acts as a premise to itself as conclusion. If you know that x exists, you can infer that x exists: every object suffices necessarily for itself. Every object causes itself. It causes other things too, but among them is itself. It is a kind of limit case. Why not say this? Surely, it is reasonable to say that a property suffices for itself: if F is satisfied, then that property is satisfied. It will cause other properties to be satisfied, but it also causes itself to be satisfied—trivially, one might say. It is causally sufficient to satisfy the property of being red by being red—being red is enough to make a thing red, obviously. What is this “making” relation if not a type of causation? We don’t normally mention this case because it is so painfully obvious—it adds nothing to the conversation—but it is true nevertheless. If every object is trivially self-causing, we won’t have any need to assert that fact; but it is still a fact—as is the identity of every object with itself. Self-identity is a truism scarcely worth noting, and so is self-causation. If someone were to assert, “This table is a cause of itself”, we would reasonably (and sarcastically) respond, “So, what else is new?”

If your intuitions are now moving in the desired direction, let me politely nudge them further. Can’t we say that type identity is compatible with causation? If an object replicates itself, producing an identical copy, doesn’t it cause the same type to exist elsewhere? True, there are two tokens, but doesn’t the type do the causal work? The type causes itself to appear elsewhere. When a cell divides, the new cells come to have the type they have in virtue of the type of the original cell, but now in a different place. The causal relation operates at the level of the type, and it is the same type in another location. The type is causing an identical type in another token. Types cause themselves to replicate. We say, “This type came to be because of that type”, referring to two token instances of the same type. Now consider identity through time: an object at one time causes itself to have certain properties at a later time—say, in the development of an organism (e.g., a butterfly). The object is causing a change in itself not in some distinct object: x at t1 causes x at t2 to be F. The causation is reflexive and cross-temporal. There is no conceptual problem about this. So, why not allow that an object can cause itself at a time? It can cause itself to be F at a time, to exist at a time, and to be itself at a time. To be sure, it can’t efficiently cause these effects, i.e., by being antecedent to itself, but not all causation is efficient causation. Once we loosen the concept up a bit, we can see our way clear to self-causation. I am the reason for my own being—not the sole reason, but a reason. I am me because of me: how could that fail to be true? I don’t go around excitedly announcing this to all and sundry, because it is insultingly obvious, but it is a metaphysical truth (like “numbers are objects” or “events happen”). It is puzzling that Aristotle never included this in his list of causes, but he could have—call it “identity causation”. Just as we say that every object is identical to itself, so we can say that every object causes itself. Neither proposition will make the evening news (“Scientists have discovered that not only is every object self-identical but every object is self-causing. For more on this over to…”).

The metaphysics of causation has long debated the extension of the concept cause, with narrow and wide interpretations proposed. For some reason, so-called efficient causation (isn’t all causation efficient?) has recently held pride of place, with Aristotle’s four types reduced to one. I am suggesting we up the tally to five. In time it may even seem to people as if this was the right place to start: the first and most obvious category of causation is that between an object and itself; everything else is an extension of that basic idea. We thus have identity cause, formal cause, material cause, teleological cause, and efficient cause—in that order. The most internal cause is identity cause: causation begins at home, as a relation between a thing and itself. And isn’t it nice to think of oneself as one’s own cause? I am literally a self-made man (and I did it my way).[2]

[1] See my “A New Metaphysics”, “Causal and Logical Relations”, and “Because”.

[2] It is customary to think of one’s parents as the proximate cause of one’s existence, but on the present conception there is another link in the causal chain (or structure)—you yourself. And notice that you are indubitably sufficient for your own existence, but nothing else is—the causal chain might have been interrupted at any point before your glorious self popped into being. Only you guarantee you. Granted you are the product of other things too, but at least you play a causal role in bringing about your existence. Everything, in fact, is a self-caused cause—and effect.