> I haven't seen any evidence to suggest that it is strong correlation with an innate, biological, causative metric that closely corresponds to a general, popular notion of "intelligence"--which is of course the use that pop psychology immediately attaches to it.
I don't think you need to correlate IQ with anything physical for it to exist. But it's been correlated with genes, amount of white matter, size of different regions in the brain etc...
Maybe we have a different idea of what IQ means.
I would say the IQ "theory" could be broken down into a couple of distinct hypotheses.
1. Problem solving is a skill/ability - People who are good at solving some types of problems, will be better than average as solving other types of problems.
2. We can test this ability - We can write down problems and ask them to solve them. Because of 1 this tells us how good they will be at solving other problems.
3. This ability helps with some real world tasks - The real world has lots of problems that benefit from problem solving ability.
When we put together this means we can predict a portion of someone's ability at a real world task we care about (like physics or software development) by estimating their problem solving ability through a test. Based on the evidence we currently have I don't know how this couldn't be true.
IMHO, the most biting criticism of IQ is that we (that is the psychology community at large) expect such a thing to exist, and so we treat anything that looks like it might fit that glove, no matter how poorly, as evidence of its existence. Our scientific methods aren't really capable of forcibly rebutting such a theory--see the existence of parapsychology, where psi is statistically known to exist (unless you're a skeptic). See https://slatestarcodex.com/2014/04/28/the-control-group-is-o... for a discussion of this phenomenon.