The Atlas of Reality
Page 94
Hume's claims have not gone unchallenged. Elizabeth Anscombe (1974), for example, has argued that we do have direct empirical knowledge of causal connections. We see things push, pull, break, hit, and dent other things (to pick just a few examples), and each of these actions involves a causal connection. George Berkeley (1710/2009) argued that, at the very least, we have direct experience of causation in our own case, when we move our own bodies or generate ideas in our minds at will.
26.1.2 Arguments for Causal Realism
Causal Realists have offered at least five arguments for their view.
1 APPEAL TO COMMON SENSE. We believe that many ordinary events and situations are caused. As Anscombe pointed out, our experience of the world includes the experience of events like boys' kicking balls, rocks' breaking windows, horses' pulling carts, and so on. In addition, we seem to have implicit knowledge of our own agency. We do things, and what we do makes a difference to what subsequently happens. That is, our actions cause things to happen. If this were not so, then most of our beliefs about our everyday lives would be radically mistaken.
The process of making decisions involves a tacit commitment to the causal efficacy of our choices. Here is an example. Suppose that scientists discover that people who say ‘Good morning’ at least six times every day are likely to live ten years longer, not because saying ‘Good morning’ has any propensity to make one healthier or safer, but just because people with healthy genes are more likely to say ‘Good morning’ than those with unhealthy genes. It would be irrational, upon learning about this study, to begin saying ‘Good morning’ more often, even though doing so might raise the probability of long life. In order to make the right distinctions here between rational and irrational choices, we have to bring into account the effects of one's possible actions. An action is rational when it is believed that it would cause favorable results, not merely when it is correlated with favorable future events (see Skyrms 1980, Lewis 1981).1
2 SCIENTIFIC PRACTICE, ESPECIALLY IN THE SPECIAL SCIENCES LIKE MEDICINE, ASTRONOMY, BIOLOGY, AND GEOLOGY. As we mentioned above, Russell pointed out that the laws of physics don't mention causation. Nancy Cartwright (1983, 1994) has argued that the laws of physics “lie” in a certain sense. According to Cartwright, when we abstract the laws from actual contexts and turn them into abstract, mathematical formulas, we enter a fantasy world that is cut off from the actual practice of science. The laws of physics describe the tendencies or propensities of things. When we apply them to any actual setting, we must add various phenomenological “fudge” factors, as well as hedges and exceptions. We can say that two bodies will tend to accelerate toward one another at a rate proportional to the product of their masses and inversely proportional to the square of their distance, except insofar as other factors and forces (like friction or viscosity) interfere. Thus, the application of the laws of physics always involves reference to a network of actual and potential causes.
Moreover, as Cartwright and others have pointed out, the identification of causal factors is a central and indispensable part of all of the so-called special sciences, that is, of all of the sciences except for fundamental particle physics. In medicine, we seek the causes of health and disease, in economics the causes of inflation or unemployment, in sociology the causes of social conflict or harmony, and so on.
Finally, the problem of the interpretation of quantum mechanics has brought causal questions to the fore, even in pure, theoretical physics (Healey 1989, Bub 1997). Quantum systems seem to evolve in two quite different ways, sometimes in accordance with the linear dynamics of the Schrödinger equation, and sometimes by way of the collapse of a quantum wave in accordance with the probabilities associated with its instantaneous state. We seem to be forced to ask what causes the collapse of a quantum wave. Is it interaction with a macroscopic measuring instrument, interaction with a conscious observer or is collapse a spontaneous event that occurs with a certain probability? In addition, the Bell inequalities reveal that causal interaction in quantum mechanics is non-local. That is, the occurrence of a measurement in one location can bring about a simultaneous collapse in an associated quantum wave at any distance, no matter how great. This non-locality seems to conflict with the constraints of Einstein's relativity theory, according to which no signal or influence can travel faster than the speed of light.
Whatever the ultimate solution may be to the problems of measurement and non-locality in quantum mechanics, the mere fact that these have become central problems for physicists and philosophers of physics demonstrates that questions about causation have not become obsolete (see Cushing and McMullin 1992 and Albert 1994).
3 CAUSATION NEEDED IN AN ADEQUATE ACCOUNT OF EMPIRICAL KNOWLEDGE. If nothing is caused, then sensory knowledge and memory are impossible. This fact was not always recognized. For a very long time, going back as far as Plato's Theaetetus, it was thought that knowledge could be defined as justified true belief. None of these three elements, whether justification, truth, or belief, involve causation in an obvious way. However,Edmund Gettier (1963) crafted a persuasive refutation of the justified true belief theory of knowledge. Here are some Gettier-inspired examples of justified true belief that are not cases of genuine knowledge:
Smith has good evidence that Jones owns a Ford. On the basis of this evidence, Smith infers that Jones owns either a Ford or a Chevrolet. In fact, Jones owns a Chevrolet. Smith's belief that Jones owns a Ford or a Chevrolet is true and justified, but Smith doesn't know that Jones owns a Ford or a Chevrolet.
Smith is seeing a very lifelike hologram, and she believes that she sees a bright red apple in front of her. In fact, there is a bright red apple in front of Smith, of exactly the right kind and in exactly the right place, but behind an opaque screen. Smith justifiably believe that a red apple is there, and her belief is true, but she doesn't know that the apple is there.
Smith has completely forgotten a childhood visit to Disneyland, but a memory-like image of the Disney Matterhorn is placed in her mind by hypnosis. Smith justifiably believes that she remembers how the Disney Matterhorn looked, and her pseudo-memory does match exactly the appearance the Disney Matterhorn had at the time, but she does not know that it looked like that.
Smith justifiably believes that the stock market went up 2% yesterday on the basis of what looks like a copy of the Wall Street Journal delivered to her house. In fact, the copy Smith has is a fake produced as a prank by her neighbor. Coincidentally, the market did go up 2% yesterday, but she doesn't know that it did.
In each case, it seems that what is missing is a causal connection of the right kind between Smith's belief and the corresponding fact. Smith's belief that Jones owns either a Ford or a Chevrolet is not caused in any way by the fact that Jones does own a Chevrolet. Smith's red-apple-ish visual impression is not caused by the red apple, her memory-like image of Disneyland was not caused by an actual past experience of hers, and the accurate report about the stock market in the fake newspaper was not caused by the actual rise in the market (see Goldman 1977).
4 CAUSATION NEEDED FOR AN ADEQUATE ACCOUNT OF SEMANTICS. In his classic work on the philosophy of language, Naming and Necessity (Kripke 1980), Saul Kripke put forward a strong case for the thesis that what we refer to when we use a name depends at least in part on the actual causal connections between our use of the name and the origination of the name, when the first user “baptized” some object as the name's bearer. This Kripkean or causal/historical theory of reference has been further developed and applied to certain uses of descriptions (Donnellan 1966) and demonstratives, like ‘this’ or ‘that’ (Kaplan 1989, Evans 1982).
Kripke argues, for example, that one can refer to the great German-American mathematician Kurt Gödel by using the name ‘Kurt Gödel’, even if one knows virtually nothing about him, and even if the descriptions that you associate with the name really apply to someone else (like the mathematician David Hilbert, for example). What ties one's use of a name with a particular name be
arer is a connected history of using the name in a certain way. The end of the historical chain is connected with the bearer by the intention and knowledge of some initial user, and each of the links in the chain involves an intentional connection between the earlier and later use: the later user intends to use the name as part of the ongoing practice as instantiated by the earlier use. Causation seems to play an indispensable role in this theory, both at the beginning and at each intermediate link. The bearer of the name impresses itself somehow upon the senses of the original user of the name, so that he or she is in a position to intend for this name to refer to that particular object, and each use of the name impresses itself in a similar way upon the senses of the next user.
Kripke's picture can be generalized to apply to giving the semantics (in a sense) of our mental ideas or concepts. RCK's concept of Dan Bonevac is something like a mental symbol that has Dan as its reference, by virtue of causal connections between Dan's features at various times and the mental picture that RCK associates with his concept of Dan as an individual. The causal connection needn't be knowledge-conferring: RCK might think that Dan is taller than he really is, if his sense impressions of him all occurred under misleading circumstances (in the context of some optical illusion, for example). What matters is that RCK thinks of Dan in a certain way because Dan (that very person) affected RCK's senses in the corresponding way and in the appropriate contexts. Thus, a full account of mental reference or intentionality, of the aboutness of ideas, seems to require the reality of causation.
5 CAUSATION NEEDED FOR AN ADEQUATE ACCOUNT OF MODALITY AND MODAL KNOWLEDGE. If there is no such thing as causation, then there can certainly be no such thing as causal powers. If there are no causal powers, then an Aristotelian account of potentiality or real possibility, like Aristotelian Modality (15.2T.7), cannot be correct. Without the knowledge of the causal powers of things, we would have no reliable basis for beliefs about counterfactual possibilities. We can have no knowledge of how the world as a whole could have been without knowledge of how particular things might have acted or responded in actual past circumstances.
There is a simple reason for this: the world is a very big, very complex place. If the world were much smaller and simpler, we might experience a world that simply cycles over and over through some set of finite states. In such a world, we might notice that world-state A is sometimes followed by world-state B and at other times it is followed instead by world-state C. Observations like this would give us the basis for constructing models of alternative, counterfactual possibilities, like the possibility that the world actually passed from state A to B to D, but it might have passed instead from A to C to E.
However, it is impossible for us to observe anything like that in a world as large and complex as ours. Instead, what we can actually observe is that the world is made of a large number of separate things that fall into a relatively small number of natural kinds. We can observe that the members of each natural kind have certain powers and dispositions in common, and on the basis of this knowledge we can reliably construct models of how things might have gone counterfactually. The reliability of this method requires real powers, or, equivalently, causal laws of nature, which, in turn, require real causation.
Why is it so important that we have knowledge of any counterfactual possibilities? Couldn't we be content with knowing everything there is to know about the actual world? Who needs alternative possible worlds?
First of all, we saw in Chapters 4–6 and 14–16 that our knowledge of possibility is of central importance to both our common sense and our scientific view of the world. In addition, we would argue that we could not understand the actual world without knowing about the existence of other possible worlds. To understand the actual world, we must understand the natures of actual things. Understanding the nature of a thing, in turn, involves knowing, not only the details of its actual history, but also what alternative possibilities are consistent with its essence.
Finally, our knowledge of the semantics of general words and our knowledge of the content of our concepts both involve knowledge of when the word or concept would be true of something and when it wouldn't. To grasp the meaning of a word is to grasp its application-conditions, that is, what something has to be like in order for the word to be true of it. To understand the word ‘red’ is to know something about a range of possible situations, namely, which of them include something that would fall under the extension of ‘red’. Similarly, one has a grasp of the concept of redness only if one can recognize, across a wide range of possible situations, which do and which do not include something that falls under that concept. Merely knowing which things are red is not enough. Suppose there were only one red thing in the world. Knowing that that thing is red and nothing else is would not be sufficient for understanding the meaning of the word ‘red’ or for grasping the concept of redness. Some knowledge of possible situations is required.
26.2 The Scope of Causation
Given Causal Realism, we can move on to the second question, which is about the scope of causation. How many things are caused? Everything, or only some things?
26.1T.1T Universal Causation. Everything is caused.
26.1T.1A Special Causation. Some things are caused, and some things are not.
26.2.1 Arguments against Universal Causation
There are four objections to Universal Causation, one assuming Mereological Universalism (22.3T.1), a second employing a pluralized version of Universal Causation, a third involving a chain principle, and a fourth including an argument against infinite causal regresses.
OBJECTION 1: ASSUMING MEREOLOGICAL UNIVERSALISM If Mereological Universalism were true, then the sum of all reality, R, would have to have a cause, C. But causes of existing things must themselves exist:
Existence of Causes. All actual causes exist.
Consequently, C, the cause of R, would have to be part of R. This would make C the cause of itself. But nothing can be self-caused:
Impossibility of Self-Causation. It is impossible for something (wholly) to cause itself.
We've already seen some reason to deny the possibility of causal loops, namely, the grandfather paradox (see Section 20.5.6). As we saw, Neo-Humeists (4.4T) have a way out of the paradox that is compatible with self-causation, since they can deny the Intrinsicality of Powers (PMeta 2).
Another reason for denying self-causation has to do with causal asymmetry. If the direction of causation between two events, which determines which event is the cause and which is the effect, is always reducible to other facts about the two events, then the basis of causal direction must be an asymmetric relation. If we assume that causation is transitive, then any successful reduction of causal direction to an asymmetric relation will impose a “partial order” on the class of events, ruling out any cases of self-causation. However, if causal direction is an irreducible, metaphysically fundamental fact about cause-effect pairs, then it would be much harder to rule out cases of circular causation a priori.
OBJECTION 2: A PLURALIZED CAUSAL PRINCIPLE We can avoid the appeal to Mereological Universalism by simply stating Universal Causation using plural quantification (Boolos 1984):
Pluralized Universal Causation. If the x's are some things, then the x's have (collectively) some causes.
We can then apply Pluralized Universal Causation to the plurality of all things that exist. These things would then have to have (collectively) some causes. These causes would have to exist themselves, and so they would be among their own effects. Consequently, at least one thing would have to be (at least in part) a cause of itself, in contradiction to the Impossibility of Self-Causation.
We could make this point by stating Pluralized Universal Causation principle in a stronger form:
Strong Pluralized Universal Causation: If the x's are some things, then there are some other things, the y's, such that the y's collectively cause the x's, and each of the y's is distinct from each of the x's.
OBJECTION 3: MEYER'S CA
USAL CHAIN PRINCIPLE Robert Meyer (1987) provides a similar objection to Universal Causation. A causal chain is a chain of entities, each of which is a cause of its immediate successor in the chain. Meyer assumes that if there is a causal chain, then there is a single thing that is a cause of every member of that chain. If we also assume that there is a set of causes, and we employ the usual axioms of set theory (including the Axiom of Choice), then Universal Causation entails that something causes itself. This violates the Impossibility of Self-Causation.
OBJECTION 4: AGAINST INFINITE REGRESSES Even if Mereological Universalism were false, Universal Causation would entail the existence of infinite causal regresses. If everything were caused, then each cause would itself be caused by something else, assuming again that nothing can cause itself. If causation is transitive, then an endless chain of causation would have to be infinitely long. If the chain were to form a closed loop, then each thing in the chain would cause itself, via the other links in the loop. That situation would also violate the Impossibility of Self-Causation.