This piece from Cosma Shalizi about IQ/g was going around a few years ago. Not being a statistician, I spent several days attempting to understand it.

Eventually, I did.

I was not impressed.

After all that reading and looking things up, though, I grew too bored with it to write about it.

His argument, like many academic arguments with some hidden political agenda, reduces to attempting to define the problem or quality in such a way that no one ever does, and then “proving” this ridiculous definition of the property does not exist.

It’s all obfuscated of course in fancy talk that few people can understand, but that’s what it boils down to.

Luckily I’m clever enough to understand it and then realize it’s mostly bullshit.

To clarify what I am talking about, is there a thing called “love?” Or the “mind?” Or “trust?”

If so, where are they? How do you locate them statistically?

Hint: you cannot.

Therefore in Shalizi’s world they do not exist.

Obvious balderdash like this is what occurs when academic discourse is strained through politics.

Personally, I don’t think g is nearly as important or as “real” as many make it out to be. But I do believe it exists and is measurable, that something is being measured there.

“Proving” that it’s not any one thing you can point to in the real world is about as good as proving that bicycles aren’t airplane pupae – utterly worthless.

For a more nuanced and thorough (and technical) refutation of Shalizi’s hack piece, read here.

Shalizi and his piece are beloved for political reasons.

I don’t care about political reasons.

I care about whether g is useful and what it can tell us. I believe it can tell us something, though it’s over-emphasized.

Defining it in bizarre ways and making category mistakes about it, and then proving that it doesn’t exist using specious methodology, is not really that helpful.

(See also this piece.)