Shalizi’s most important argument is that Flynn and others who attach a good deal of importance to g haven’t made much of an argument that it measures a single phenomenon.
After a century of IQ testing, there is still no theory which says which questions belongs on an intelligence test, just correlational analyses and tradition.
Flynn and others have good arguments that whatever g measures is important. But Shalizi leaves me with the impression that the only way to decide whether it’s a single phenomenon is to compare its usefulness to models which describe multiple flavors of intelligence. So far those attempts that I’ve looked at seem underwhelming. Maybe that means trying to break down intelligence into components which deserve separate measures isn’t fruitful, but it might also mean that the people who might do a good job of it have been scared away by the political controversies over IQ.
HT Kenny Easwaran.