Causation vs correlation in the context of physics

Question

How to know whether A and B is causal or correlated? Is it correct to say that physicists have always been concerned about causation? The laws of physics are stated in terms of equations that have nothing to say about causation. They are correlational laws.

Please give me some reference if possible.

Answer 1

Physics is concerned with causality in the framework of natural law. Causality there is deeply embedded in both the two main fundamental theories: QM & GR. In QM, because time is open and things happen whilst in GR, its proveable that the very structure of spacetime is causal. This is a theorem of Malament who showed that a spacetime could be rebuilt (upto a conformal factor) from a knowledge of its causal structure. This has prompted the development of the causal set programme in Quantum Gravity.

More broadly speaking, as al-Ghazali pointed out, physics is merely correlational, requiring the intervention of the Absolutely Real to make what is wholly unreal, real. But this is outside what is now understood to be physics - if not what was understood traditionally as metaphysics. Of course some physicalists take physics to be the ground but this is not primarily because they are physicists or even interested in physics but because they are a certain tribe of materialists.

Answer 2

Causation in physics goes like this:

• We have some equations that describe how the state of a system changes over time. These equations typically give the time-derivative of different variables of the system, as a function of the current values of those variables.
• At time t, we set up some initial conditions for the system.
• We use the equations of the system to trace how the system changes, based on its initial conditions, to some later conditions at a later time.
• Then we say that the initial conditions caused the later conditions.

It is true that the equations typically don't inherently say which is the cause and which is the effect, and in this sense you might say they are describing mere correlation. We might use the same equations to trace how the state of the system changes backwards in time. However, by convention, the positive time direction is considered the direction of causation.

Here's another way you can use the equations to talk about causation:

• You first deduce that initial conditions X at time t lead to later conditions Y at time t2 > t, according to the equations of the system
• You change the initial conditions, so that instead of X you have X2, slightly different from X. Then you trace from X2 at time t, to later conditions Y2 at time t2 > t.
• Then you say that the change in initial conditions from X to X2, caused a change in later conditions from Y to Y2. This lets us talk about "but-for" causation: what would have happened, if initial conditions had been different.