Hume on Causation

I have heard from multiple sources that all conversation concerning causation starts with Hume.

In An Inquiry Concerning Human Understanding Hume defined causation as follows:

We may define a cause to be an object followed by another, and where all the objects, similar to the first, are followed by objects similar to the second. Or, in other words, where if the first object had not been, the second never would have existed.

The first half of this quote is quite similar to the idea of pragmatic causation that I have been suggesting in the previous posts: A cause is that which we have reason to manipulate in virtue of the fact that, through it, we can bring about that which we desire or prevent the realization of that to which we have an aversion.

Note, however, that this account of practical causation differs from Hume's definition in that it includes an element whereby manipulating the cause is a means of manipulating the effect. Hume's definition is binary; A either is or is not followed by B. The account of practical causation says that we can manipulate B by manipulating A. Now, either bringing into existence or denying existence of something is one form of manipulation. However, the account of practical causation at least makes clearer the fact that, for example, we can turn a dial to make the music louder or quieter, or make the light in a room brighter or dimmer.

One of the criticisms of Hume is that his theory of constant conjunction seems to identify certain things as causes that we typically do not accept to be causes. A standard example is the fact that night always follows day. Indeed, all "objects" (it would make more sense to use the term 'state of affairs') similar to a day are followed by objects similar to night. Yet, we do not say that day causes night or night causes day. Rather, we say that day and night have a common cause - the fact that the earth spins on its axis.

The pragmatic account of causation, in that it includes an element that looks at what we can manipulate in order to change the effect, avoids this objection. If we look at that which we can manipulate to bring about a change in the effect we discover that there are things that we can manipulate that will bring about a similar effect in both day and night. We can increase the length of the day (and the night) by slowing the spin of the Earth. We can eliminate the day and create a permanent night by eliminating the sun.

So, the practical view of causation respects the fact that the day and the night have a common cause - that which we can manipulate in order to influence both of them.

Off of the top of my head, I fear that this account of causation is going to be considered circular. After all, "that which we can manipulate" seems to be assuming causation, and thus it is assuming that which we are trying to explain. How do we account for the fact that the manipulation of A brings about a change in B? This "bringing about a change in" is the very subject of causation that we are trying to understand.

If we look to Hume for an answer to this question, we acquire this knowledge from experience. We learn that the conscious manipulation of A brings about a change in B in regular ways that we can observe and understand. By experience we learn that the manipulation of A brings about a change in B.

Note that this still contains a significant difference with respect to Hume's account of causation. Hume describes causation in terms of the regular connection between events, while the pragmatic view of causation looks at the conscious manipulation of A to bring about a change in B. Or, at the very least, it is something where we have reason to expect that if we could manipulate A (e.g., slow down the rotation of the earth), then we could bring about a change in B (the length of the sun). This is information we acquire through observation.

This also provides us with a way of acquiring a more precise definition of "objects similar to" the cause and the effect. Objects "similar to" the cause are those that we can manipulate that will bring about a change in the effect. Objects "similar to" the effect are those that are changed by manipulating the cause.

Experience tells us that by changing the color of a light switch we do not change the capacity of the light switch to turn the light on and off. Thus, the color does not cause the light to go on or off. It is not something that we can change to manipulate the effect. Similarly, something is "similar to" the light going on if it is something that will illuminate when switch is turned on. A different light bulb, a heating lamp that fits into the same socket, or a heater plugged into the same socket, are all objects similar to the light in that they are things we can manipulate by turning the switch on and off.

By defining "similar to" as that which we learn through observation that we can manipulate in order to bring about changes in the effect, then this is not circular. This is empirical. It fits in with Hume's basic empiricist assumptions.

Note that I have written this entire post concentrating on the first half of Hume's quote. I have said little about the second half - "where if the first object had not bee, the second never had existed". Though it follows from what I have said that this is a poor restatement of the first premise. The first statement allows for the manipulation of the properties of the second object by manipulating the first. The second statement is entirely black and white, 'exists' or 'does not exist'. I do not think that causation is as binary as this second statement seems to require.