What does "to cause" mean?

What does it mean, strictly, for one event to "cause" another? If I throw a ball, does the movement of my arm cause the ball to move, or are they simply correlated events? If you say the arm caused the ball to move, how can you be sure?

I would like to think about this from basic concepts, in slightly less formal and less "according to the great thinkers" terms...

It seems obvious to me that if a set of conditions (C) necessarily came about because of an action (A) then A caused C.

If an event (E) happens after an action (A) (or conditions (C) that it created, as above) then it may have been caused by A. This is the interesting part...

If (E) must necessarily have happened if A (or C), is this sufficient to say A caused E?

Or should we add that E could not have happened without A (or C) as well?

What if E was only one of several possibilities, one (or many, but not all) of which must have happened once A had happened (or C had been created)? A caused something to happen, but not necessarily E: E is just the event that (randomly or by some decision/further event) ended up happening.

These questions have application in law as well as philosophy.

Example: X steals \$100 from Y (A). Y can now not pay all their bills (C), fails to pay the rent, and gets evicted (E). A caused C, there's little doubt about that. However, Y chose not to pay the rent, rather than their credit card bill. So did A cause E (the eviction) or was it the decision of what not to pay? In a perfect world, where X gets caught and is forced to repair all damage to Y, should they re-house Y as well, or is the causal relationship broken because of another conscious action (the decision) in between? If Y acted rationally, then we can assume that eviction was the least negative outcome of not paying a bill (maybe all his other services are supplied by the Mafia :P). In this instance, can we say that A caused E? (or "at least" E - not a particularly logical term, but one that has meaning in the legal sense here).

This is a line of thought I often find myself pondering, and I'm interested in people's own inputs and lines of thought, rather than a link to established questions on the matter. (BTW I'm not denying the usefulness of the previous response, just looking for more discourse on this.)

• Great answer, but is there any chance I might be able to persuade you to maybe show some directions for further exploration -- maybe reference some introductory articles to learn more? I really liked Michael's suggested SEP's entry on causation in metaphysics -- at any rate, I thought it was at least worth bumping here again in a comment. plato.stanford.edu/entries/causation-metaphysics Nov 15 '12 at 17:04

Causation is a very thorny philosophical area, and the literature on it is immense.
Here is an introductory article (from the SEP) on the subject.

If you are beginning your research, I'd suggest you begin by looking at Hume, as he raised issues that everyone since has had to grapple with.

The meaning of "cause" is very vague, it is a human term, not a scientific term. It supposes you are talking about a collection of very many analogous events, with a map between them which tells you how they are analogous, and with different attributes. When you say "A causes B", you mean that in those analogous cases where you find the bits describing the initial situation to describe A, you will reliably find the bits in the later state to describe state B, regardless of the other bits in the description. Then we say "A causes B".

It is always an analogy, and it's always complicated. The fundamental description in terms of physics does not include this complex notion of "cause and effect" that we attribute to macroscopic phenomena. It includes other things that are different that are called "causality", and this leads people to think causality is tied into physics more than it is.

• When you hit the "send" button to submit your answer here, were you merely a correlation and not a causative agent in the process? Sep 23 '12 at 21:53
• Is a scientific term a non-human term? Sep 24 '12 at 11:21
• @MarkJ: Honestly, I don't know. In the definition I gave, I am a causative agent, because the send-button pressing is A and the appearence of the message is B, and the definition works. But if I say "I know this guy, he's a bastard, he's going to walk out on the job half finished, so I'll hire someone else to do it." Is the cause of the hiring someone else the other person's walking-out-on-the job? Is it his "bastardness"? There are worse cases in the superrational examples, the cause-effect idea is hard to pin down, and it doesn't appear in physics, only in human world. Sep 26 '12 at 0:00
• @iphigenie: A scientific term ideally is something you can program a computer to understand. A human term is like the ultimate capcha. Like the statement "Beckett is an important writer" is hard to explain to a computer how to verify. Sep 26 '12 at 0:01
• @RonMaimon: Well, that's the thing, agency can really be only verified (and perhaps even understood) from the point of the agent him/herself, and never from a second-hand point-of-view. To deny the agent, is to partially deny the whole enterprise of science, since it's the scientist who is making the models of the world in which s/he is verifying. Sep 26 '12 at 0:52

This depends on who you're reading and what sort of perspective you want to explore. The four causes of Aristotle are not all the same. What modern scientists do with causality differs. From an empirical perspective, causes are not observable, and thus at best correlations or associations, at worst fictions.

"To cause" means that there is an unbroken chain, ordered in time, that links from cause to effect. To negate this would be to break reality itself (or at least its relationship to human consciousness).

The item which has caused all the confusion, is whether the human mind can be its own agent; other than that, despite what everyone may say, there is no controversy.

• Aristotle would take you to task for this. There still do exist people who believe in "final causes". Sep 23 '12 at 23:17

A causes B if all of the following are true:

• A exists when B exists.
• A can exist without B.
• B does not exist without A.

In reality though there is hardly ever just A and B. Normally there are also C, D, and so on which makes it impossible to determine the causes. For example:

Suppose

• A, B and C exist,
• C does not exist without A and B,
• and C exists with either A or B.

Then we don't know if C is caused by both A and B or if C is caused by either A or B but not both.

That's why only when simulating the 3 variations described in the first example can the causal relationship be understood.

• There's a literature of rival axiomatisations of the "is a cause of" relation. Where does yours come from? Nov 22 '12 at 13:58