Statistics about X are not causal factors determining X
Mistaking statistics, which are merely our summaries of goings-on in the world, with causal factors in the world, is a confusion that pops up too regularly. For instance, here is Gregory Clark suffering from it. Mary Morgan understands this point:
"such statistical or probabilistic laws can be said to govern the behaviour of our population. We individual people know better -- we know that the births, marriages, and debts are determined by a whole realm of social, economic, medical, physiological, and other laws, which determine whom we fall in love with, whether we have children, why we die, and when any of these happen to us." -- The World in the Model, pp. 336-337
In fact, I think Morgan hasn't gone far enough here: the "laws" she cites are just our names for the regularities produced by concrete causal factors, and don't themselves cause anything.
(The proposition in the title of this post admits of exceptions, such as when a statistic about housing starts is interpreted by builders and changes their inclination to build houses.)
"such statistical or probabilistic laws can be said to govern the behaviour of our population. We individual people know better -- we know that the births, marriages, and debts are determined by a whole realm of social, economic, medical, physiological, and other laws, which determine whom we fall in love with, whether we have children, why we die, and when any of these happen to us." -- The World in the Model, pp. 336-337
In fact, I think Morgan hasn't gone far enough here: the "laws" she cites are just our names for the regularities produced by concrete causal factors, and don't themselves cause anything.
(The proposition in the title of this post admits of exceptions, such as when a statistic about housing starts is interpreted by builders and changes their inclination to build houses.)
I agree completely. Statistical trends often reflect underlying causal conditions, which is why it's dangerous to bet against them. But as you say they are never causes themselves. That's just confused thinking.
ReplyDeleteYes, I think it comes from working with a model so long that you start to confuse the model with what it models.
DeleteCommon enough.
DeleteIn fact, I think Morgan hasn't gone far enough here: the "laws" she cites are just our names for the regularities produced by concrete causal factors, and don't themselves cause anything.
ReplyDeleteDo you think the same thing for things like the laws of physics or the laws of chemistry?
Yes, the "law of gravity" does not cause anything: what causes gravitational attraction is the actual relationship between concrete bodies.
DeleteOf course, stating the law is VERY useful: I am not criticizing that. But it is just a summary of what happens in many concrete interactions.
I'm split on this: one side of me wholeheartedly agrees, but the other thinks that this is nitpicking and that most people would say when pressed "Well, of course it's not the laws of gravity that cause planetary, but you know what I mean.". What do you say on that? Would those people be fooling themselves?
Delete