I don't understand what you're trying to say. Are you saying that 'g' might conceivably be real, but that you believe that it happens not to be real? Or are you saying that it's not real by definition?
I'm saying I've seen no evidence that it's real. Personally, I believe that it's not real
because I've seen no evidence that it's real, and it ends up getting filed in the same drawer as dragons, unicorns, and the Easter Bunny. But I'm willing to believe in the reality of the Easter Bunny under appropriate epistemological circumstances that haven't yet happened.
Factor analysis can certainly find real things. But it's also well-known as being able to "find" things that don't really exist.
We can say "correlation doesn't imply causation", but it would be nice to understand why the correlation exists.
It would indeed. And for nearly a century of research, no one has even investigated the question.
Maybe this is getting too philosophical, but what is "causation" anyway, if not "correlation whenever we check"?
Hume answered this one long ago -- "correlation whenever we check" doesn't work. The cock's crowing does not cause the sun to rise no matter how many times we check. There are a lot of formal philosophical definitions of "causation," none of which are entirely satisfactory.... but they're a lot better than naive "correlation."
[Phlogiston] predicts some things incorrectly. Otherwise, we'd still believe in it. No?
Yes, but only because someone started to actually think about the implications of reifying phlogiston
as a substance. Since "substances" have positive weight, an unburned piece of wood should weigh more than the products of combustion. Doing the experiment was rather tricky, but when they finally found out how to do it, they found that "phlogiston" had negative weight, and a better description of the actual substance involved would be "lack of oxygen."
Another nail in the coffin was when people realized that the substance phlogiston would be exhaustible, but found that "heat" could be generated (via friction) more or less without limit, without exhausting the "phologiston."
As a result, we now know of two separate aspects of combustion -- the chemical process of oxidation, and the revision of "heat" as a kind of energy, not a substance, that is produced by many processes including oxidation. From one factor ("phlogiston") comes two, because we actually thought about what we were doing.
There's an obvious, if somewhat facetious, analogy to the g-model of "intelligence." Perhaps what we measure as IQ is in fact a
failure of mental process; the actual human mind/brain is ideally capable of infinite performance, but various aspects limit it. Perhaps we should be measuring a "stupidity quotient" and searching for ways, not to enhance intelligence, but to limit (aspects of) stupidity.
Similarly, perhaps the neurologists are wrong (and the esoteric idealists were right) all along, and "intelligence" is a dualistic process tapping like a radio receiver into the "mind of God," where IQ is simply a measure of how little interference is present.
Either of these cases (which I admit I find improbable) would suggest that the g-factor model is accurate, but ontologically incorrect.
Similarly, there may be some process ("oxidation") that produces behavior ("heat"). This is one of the central questions behind much theoretical IQ work -- does the g-factor model measure the process capacity for intelligence, or does it measure the behavioral results of the processs? But of course, if the g-factor itself does not exist, then the question itself reveals a fundamental misapprehension
caused by the reification of 'g.'
But to cut directly to the chase, what test could you perform to falsify the first hypothesis above in favor of a 'real' g-factor model?
)... but: