I forgot to mention in this video, but another critical part of adding credence to causation is the temporal dependence - does cause happen before effect?
I really enjoyed this video is there a community of people who I can chat to about these things. I'm autistic so I've found it hard to find people with interest in philosophy, especially in iowa.
While this is interesting from a statistical perspective, it seems like you (and a lot of other people tbh) are missing a key point of the claim "correlation isn't causation." No matter how high a certain correlation is, no matter how many experiments we perform to narrow down the number of alternative possible explanations of causation, we're still fundamentally making the claim "correlation is causation," but demanding a much higher bar of correlation. But this doesn't need to be the case for causation. After all, it might be that some X causes some Y irregularly, (as with "Drunk Driving causes Car Accidents") such that there are many cases with X without Y (some Drunk Drivers make it home without causing an accident), or it might be that Y is fully caused in multiple ways, such that Y is sometimes caused by X, but it is sometimes caused by Z as well (There are other causes of Car Accidents than just Drunk Driving, such as Road Rage). Likewise, there are many cases where correlation is one to one with causation, but isn't an instance of causation. For example, some X may cause some Y in all cases, but in all cases of X causing Y, there exists some Z which is present at the time of Y, but otherwise ineffectual on Y. This might most obviously be the case if X and Z are always simultaneously caused by some object A at some irregular interval, where X produces Y, but Z does not, though it always occurs at the same time as X. Thus, it seems as though you have given an insufficient method to prove causation. Now, it might be wondered, how, then, do we prove causation? Personally, I don't know. It's obvious that Causation occurs, and several instances of it are obvious as well, but providing a criterion for it or establishing it when it is a controversial case (such as the Tobacco-usage causing cancer case, man-made climate change, and all other divisive topics) seems difficult to me.
Flippin useful. Thanks very much for this.
I forgot to mention in this video, but another critical part of adding credence to causation is the temporal dependence - does cause happen before effect?
I really enjoyed this video is there a community of people who I can chat to about these things. I'm autistic so I've found it hard to find people with interest in philosophy, especially in iowa.
Doesn’t the random selection eliminate the confounding variables?
While this is interesting from a statistical perspective, it seems like you (and a lot of other people tbh) are missing a key point of the claim "correlation isn't causation." No matter how high a certain correlation is, no matter how many experiments we perform to narrow down the number of alternative possible explanations of causation, we're still fundamentally making the claim "correlation is causation," but demanding a much higher bar of correlation. But this doesn't need to be the case for causation. After all, it might be that some X causes some Y irregularly, (as with "Drunk Driving causes Car Accidents") such that there are many cases with X without Y (some Drunk Drivers make it home without causing an accident), or it might be that Y is fully caused in multiple ways, such that Y is sometimes caused by X, but it is sometimes caused by Z as well (There are other causes of Car Accidents than just Drunk Driving, such as Road Rage). Likewise, there are many cases where correlation is one to one with causation, but isn't an instance of causation. For example, some X may cause some Y in all cases, but in all cases of X causing Y, there exists some Z which is present at the time of Y, but otherwise ineffectual on Y. This might most obviously be the case if X and Z are always simultaneously caused by some object A at some irregular interval, where X produces Y, but Z does not, though it always occurs at the same time as X. Thus, it seems as though you have given an insufficient method to prove causation.
Now, it might be wondered, how, then, do we prove causation? Personally, I don't know. It's obvious that Causation occurs, and several instances of it are obvious as well, but providing a criterion for it or establishing it when it is a controversial case (such as the Tobacco-usage causing cancer case, man-made climate change, and all other divisive topics) seems difficult to me.
Empirically? No.