My theme is causality. I would like first to sketch the philosophical problem in general, and then proceed to what I call the hard problem of causality.
Any discussion of causality in Western thought must begin from Aristotle’s theory: “these are the . . . ways of apprehending the ‘how and why’ of things: we may refer it either (1) to the essential nature of the thing in question, in the sphere of unchanging objects (as in mathematics, where the conclusions ultimately depend on the definitions of straight line, or incommensurability, or whatever it may be); or (2) to that which first initiated the movement (as in: ‘Why did they go to war? Because the others had raided them’); or (3) to the result aimed at (as: ‘To gain an empire’); or (4) in the case of anything that comes into existence out of something already there, to the material” (Aristotle, Physics ii.7). These “ways of apprehending” later became known as the formal, efficient, final, and material causes respectively. The formal and material causes have for the most part dropped out of the discussion, not because they are unimportant, but because we no longer call them causes.
Aristotle appealed to final cause not only with reference to conscious human purpose, but as pertaining to any change (Aristotle, Metaphysics 983a). The final cause is out of favor among scientists, but still deserves discussion because it is constantly appealed to in ordinary speech in the explanation of human actions. One major reason for this idea’s fall from grace is that final cause is associated in the minds of many with divine purpose, now a disreputable conception in certain important circles. There are at least two reasons why this consideration carries little weight. One is that the reservation is misconceived. As Aristotle’s doctrine exemplifies, final cause may be inherent rather than dependent on planning by an external intelligence. Another reason we need not be impressed with the agnostic or atheist disdain for final cause is that these ideologies amount to little more than fads, given that the decline of theism in mainstream European culture was not brought about by a series of experiments telling against God’s existence – though admittedly an argument for atheism could be carved out with Ockham’s razor.
The adoption of the theory of evolution by means of natural selection constituted a scientific revolution partly because it was such a decisive dislodging of final cause. Where Aristotle had thought it self-evident that organisms exhibit teleology, Charles Darwin’s writings opened up a vista on a world in which events and living structures could be explained by efficient cause (as that which initiates movement or change) alone (formal and material causes not being at issue in the 19th century).
But the notion of final cause is not to be rejected on the ground that it is no longer the mode. There is a stronger reason than any I have mentioned. This is that the appeal to final cause sometimes puts us in the situation, awkward if not untenable, of explaining a real event as being due to an unreal event. Imaging a philosopher asking, “Why does this chick embryo have an egg tooth – as is revealed by this X-ray image?” A teleologist participating in the dialogue responds, “The egg tooth is there so that the chick can use it in hatching.” Let us then suppose, if the hypothesis be not too savage, that one of our group of natural philosophers picks up the egg, dashes it to the ground and crushes it under his heel. The hatching of the egg has become a non-event. Therefore to explain the presence of the egg tooth by a story involving the hatching is as if I were to say, “Last week’s blackout was the doing of space aliens – not that there are any such things, of course”. To explain the real as an effect of the unreal is no explanation.
These remarks are exactly what they appear to be: an assault on instrumentalism. I can apply no other label than hypocrisy to the practice of offering an explanation for a phenomenon and then disavowing belief in the terms used to explain. As Wilfred Sellars said, “to have good reason for holding a theory is ipso facto to have good reason for holding that the entities postulated by the theory exist” (in van Fraassen 1065). It may be objected that this principle opens a floodgate to a spate of hypothetical entities that we would be required to believe in, given that to account for any set of phenomena an endless series of complicated theories may be created (see Hempel 453-457). The appropriate prophylactic device to apply here is the time-honored criterion of simplicity as applied to hypotheses. Scientists shall believe in as little as they can get away with and still explain appearances.
Three kinds of cause down, one to go. Do we need the idea of efficient cause? Bertrand Russell sounded its death knell in 1921. He begins confidently: “The traditional conception of cause and effect is one which modern science shows to be fundamentally erroneous” (Russell 93), going on to explain that chopping up the world into hunks called cause and effect is an outmoded caricature of a world that in reality undergoes continuous change. Unfortunately for the effect of his rhetoric, he finds himself compelled to slip in a disclaimer in a footnote: “The theory of quanta suggests that the continuity is only apparent. If so, we shall be able theoretically to reach events which are not processes” (94 n). Russell does not mention - what may not have been apparent in 1921 - that quantum theory, for quite different reasons, blows much bigger holes in causality than even he is attempting to make. In any case, undaunted, he presses on with his gospel of continuous change, arguing that, since in reality any number of disturbances may intervene between an event and a second one that occurs a finite time later, to achieve absolute accuracy we must shorten the interval between cause and effect to zero. “The result is that we merely have, as the embodiment of our causal law, a certain direction of change at each moment. Hence we are brought to differential equations as embodying causal laws” (95).
Russell has a point. However, though differential equations may provide, for some realms of science, a vastly more precise tool than any traditional idea of cause, this mathematical approach may be inapplicable to other kinds of dialogue. For example, what if the question is: “What was the cause of the Crusades?”
David Hume was right that necessary connections between events cannot be established empirically. If it is a question of a singular cause and a singular effect, and if we do not allow ourselves to use subjective and vague notions (such as two I shall mention later), we may find ourselves at a loss as to how to test any such singular assertion. And yet some explanation, if not justification, is required for the fact that we utter, believe and use such propositions all the time, e.g., “It was a leak in the power steering system that made the wipers stop working” or “The shrimp she got at that place made her sick”.
I speculate that the idea of cause (Greek aitia, aition) derives ultimately from the feeling that our will is responsible (aitios) for moving our limbs; but this conjecture, whatever psychological interest it may possess, cannot supply us with an adequate philosophical definition of causality. We are looking for an objective definition instead.
Let us then address the hard problem of causality. What does it mean to say that event x caused event y? This problem is hard in the sense that a usable answer to it does not follow in any direct or obvious way from any of the commonly encountered notions of causation. Aristotle’s definition of efficient cause, “that which first initiated the movement,” may not help us, in any particular case, to determine whether a given event was the cause of a given effect. The example he gives, that a people went to war because they had been raided, takes (what one might call unfair) advantage of the fact that humans come equipped with a ready array of reasons for their actions. Ask someone why he was spurred to commit an action, and he will likely give you an answer. Your chief problem then is whether to believe him or not. With the events of inarticulate nature, we have no such proffered motives, and must hunt afoot.
In trying to apply this kind of definition of efficient cause to events that do not speak for themselves, some have embraced one of Hume’s definitions of a causal relation: “if the first object had not been, the second never had existed” (Hume 76). Like the idea that my will causes my limbs to move, Hume’s conditional sentence poetically conveys some of the connotation of cause; but it leads us further from a scientifically useful solution. Science ascertains facts, not counterfactuals. To invoke a realm of what is not but would be only enlarges the problem, as I argued above in connection with final cause. Lewis Carroll illuminated the situation for us thus: “if it was so, it might be; and if it were so, it would be; but as it isn’t, it ain’t. That’s logic” (Carroll 147).
Recalling more Hume affords us a clue. “Let an object be presented to a man of ever so strong natural reason and abilities; if that object be entirely new to him, he will not be able, by the most accurate examination of its sensible qualities, to discover any of its causes or effects” (27). This suggests that nothing about appearances, without their being conceived as belonging to known categories of events, could tell anyone whether or how they belonged in a causal scheme.
Our quest for a solution is similarly furthered but not achieved by a formula for cause such as Russell suggested at one point, “A is usually followed by B” (Russell 96). Note that though A and B here might at first glance be mistaken for individuals, they really stand for classes of events, as in “Jumping off the Golden Gate Bridge is usually followed by death”. If we want to apply the formula to a particular pair of events, we at once encounter an ambiguity, in that any single event can be placed in an indefinite number of different classes; for example, World War I was both a continuation of diplomacy by other means, and a disaster. In accord with this formula, in which classes shall we place events x and y, in order to determine whether they are cause and effect?
I propose that whenever we make an assertion that one event caused another, we do always in fact, implicitly if not explicitly, place the events in classes, silently invoking a general causal law that links the classes, and of which the sequence is believed to be an instance. If a scientist, speaking in her professional role, makes a causal statement about two past events, we expect her to be able to supply, if challenged, the general law that supports her assertion. But when ordinary folk make a singular causal claim, do we really think that they have a worked-out causal law in mind? What classes of events have they chosen to assign the given events to for the purposes of applying a sound general principle? My answer is that ordinarily no theorizing is required of the speaker because the classification of the events has already been performed for him by the language he speaks.
This process can be brought out in the open when there are multiple possibilities for how to word a description of phenomena. Imagine two eyewitnesses to a sequence of events, Ellie and Rachel, describing it to a third person, Steve:
STEVE. Why did Rosemary drop the teapot?
ELLIE. It was because Barbara walked into the room.
STEVE. Huh? So?
RACHEL. It wasn’t because Barbara walked into the room; it was because she walked into the room naked.
Here each witness knows that the other is describing the same events with empirical accuracy, but Rachel rejects Ellie’s statement because the simple narration of Barbara’s entrance fails to place the event in a class capable of causing Rosemary’s shocked reaction. Only if Steve knew that a chronic relation of extreme antipathy or anxiety obtained between Rosemary and Barbara could he have accepted Ellie’s version as causally adequate. What we look for is an implicit designation of two classes of events such that events in class A are usually followed by events in class B.
This paper has taken the path of analysis. But no connecting of the dots of outer experience can supply what thinkers, in long centuries of trying to define or prove necessary connection, have been attempting to grasp: the inner unity of the world. Plato spoke to some of our deepest intuitions when he called time “a moving image of eternity” (Plato, Tim. 37).
Aristotle. The Physics. Trans. Philip H. Wicksteed and Francis Cornford. Cambridge: Harvard University Press, 1957.
Carroll, Lewis. “Through the Looking-glass and what Alice found there” (1872). The Works of Lewis Carroll. Ed. Roger Lancelyn Green. London: Paul Hamlyn, 1965.
Hempel, Carl G. “Criteria of Confirmation and Acceptability” (1966). Philosophy of Science: The Central Issues. Ed. Martin Curd and J. A. Cover. New York: W. W. Norton, 1998.
Hume, David. Enquiries Concerning Human Understanding and Concerning the Principles of Morals (1777). Ed. L. A. Selby- Bigge. 3d ed., rev. P. H. Nidditch. Oxford: Clarendon Press, 1975.
Plato. Timaeus and Critias. Trans. Desmond Lee. London: Penguin, 1977.
Russell, Bertrand. The Analysis of Mind. London: George Allen & Unwin, 1921.
Sellars, Wilfred. Science, Perception and Reality. New York: Humanities Press, 1962.
Van Fraassen, Bas C. “Arguments Concerning Scientific Realism” (1980). Philosophy of Science: The Central Issues. Ed. Martin Curd and J. A. Cover. New York: W. W. Norton, 1998.