Picture now, with me, a very-slightly-alternate universe.
In this other universe, most things are just as they are in ours. Humans themselves are no different. But the other universe differs from ours in one respect -- the way that its scientists conceive of physical fitness.
In our universe, we know -- in a casual, non-scientific (if not unscientific) way -- that there is such a thing as physical fitness. Some people are more fit than others. We recognize different levels of fitness, and refer to them in conversation. And we know that there is biology behind this notion. The body is a set of interconnected systems, and improvements in the functioning of one will often spill over into improvements in the functioning of the others. Thus, it makes some biological sense to speak of a person's "general" or "all-around" fitness, without getting into specifics like how much weight they can lift, how fast they can run and how far. But we know, too, that a full biological account would pull apart this general concept into distinct systems and performance measures, able to vary independently from one another: VO2max and vV02max, stroke volume and heart rate, myofibril growth and sarcoplasm growth, and so forth. Scientists, for their part, concern themselves almost wholly with these details; they do not deny that one could build a picture of "general fitness" by summing all the details together, but they have not built such a theory, and (I imagine) would not really see the use of building one.
In the other universe, scientific theory took a different course. In the early twentieth century -- when many of the biological details were, in fact, understood poorly or not at all -- physical performance was studied from a purely statistical perspective, by measuring it (how much can this person lift? how far can they run?) and computing distributions and correlations. The statisticians discovered something which was either remarkable or obvious, depending on how you looked at it -- perhaps both. They discovered that all measurements of physical performance were positively correlated in the population.
That is: suppose one takes a large number of people, sampled representatively from the population, and gives a battery of tests, measuring things like "how much can this person lift?" and "how far can they run?" Then, across the entire sample, being good at any one physical task will be associated with being good at all the others. Those who can lift more can, on average, can also run further, and run faster. Those with stronger arms have, on average, stronger legs. And so on, for every pair of tests you could conceive. Of course, there are individuals who are very good at some tests but very bad at others, but this "washes out" when one averages over the whole population, always leaving a positive correlation behind.
On one level, this is not surprising. It accords with our intuitive notion of "general fitness," for one thing. And it is not hard to imagine how these population-level trends might emerge from things we all know. Some people like exercise and some do not; some people have illnesses or jobs or lifestyles that bind them to a bed or chair all day, where all their systems atrophy simultaneously; some people play team sports, which tend to develop a number of (aerobic and anaerobic) capacities at once; and on and on.
The statisticians, however, saw this as a stunning scientific discovery. And not without reason. Intuition suggests that performance on physical tests would tend to correlate, but it was downright eerie how every single pair of tests correlated positively with one another. This result held up despite the heroic efforts, by some statisticians, to contrive tests which wouldn't correlate positively with all the others. And we must remember, too, that these statisticians were working before there was a detailed biological understanding of the body's various distinct systems. It was plausible at the time -- indeed, the result seemed to actively suggest -- that there might be a single biological thing, a sort of "fitness fluid" circulating in the body perhaps, which alone explained the correlations, and which would vindicate in a precise sense the conventional notion of "general fitness."
In any event, the alternate-universe statisticians began to speak of "general fitness" very seriously. They called it "f" for short, and defined it in terms of the positive correlations, in a way that was perfectly natural mathematically. Tests that correlated especially well with all the others were said to be especially good measures of "f." If one gave a person a whole battery of tests, one could estimate their "f" quite well. A suitably scaled version of "f" was dubbed the "FQ," or "Fitness Quotient"; you could be tested, and discover yours. (The population average was 100, by definition; professional athletes almost always exceeded 130.)
What's more, "f" could be correlated with all sorts of other things, too. It turns out that "f" was highly associated (in the population-wide-correlation sense) with success in many, perhaps all areas of life. One could come up with mundane explanations for this fact -- that various illnesses decreased "f" and also made it harder to succeed, that "f" correlated with height and lack of obesity and other things that made positive impressions in job interviews and the like, etc. -- but the fact remained.
This being the early 20th century, eugenicists were all over this stuff, and this gave Fitness Research (the study of "f") a stigma which it later could not shake. Prominent Fitness Researchers these days find this very frustrating. Their field has so much to offer, explains so much. They are not eugenicists, not now. They just have many amazing correlations to tell you about.
Meanwhile (still in the alternate universe), biology and physiology developed much as they did in our universe, and quite independently from the field of Fitness Research. As biology discovered the Krebs and Cori cycles, as physiology delved into the various ways the heart and muscles and nervous system adapted to exercise, the Fitness Researchers asked excitedly whether this or that molecule, this or that process, might be the biological substrate of "f," the key linking their field to these others. But it was never that simple. On the statistical level, there were the positive correlations, yielding the gloriously simple "f"; on the biological level, there were a mess of distinct but related processes, none of which looked like "f" when taken alone.
This, of course, does nothing to undermine the existence of the positive correlations. They remain untarnished, even in the year Alterna-2017. And the Fitness Researchers are frustrated that the public is so disdainful of their field. They can take a battery of tests and predict so much from it -- and people turn them away, saying that "general fitness" is a simplistic myth.
In a mood of outraged chutzpah, one Fitness Researcher decides to write a short book for a public audience, with the audacious title "Fitness: All That Matters" . . .
•
This has all been, obviously, a heavy-handed analogy for the history and content of intelligence research.
The cases of fitness and intelligence are admittedly distinct. I don't actually know that all fitness tests are positively correlated -- it just seems prima facie plausible. But the fact that all tests of intelligence are positively correlated is the entire basis for the field of intelligence research, for IQ and "g" (the equivalent of "f"). If we did find that all fitness tests were positively correlated -- and like I said, it wouldn't be too surprising -- then we could immediately construct an entire field of "fitness research" on the exact model of intelligence research, and it would look like the picture painted above. It may in fact be a historical accident that we have, in the real world, one of these fields and not the other.
Well, okay, there is another key difference, and it probably explains why we don't have "Fitness Research." Fitness is quite obviously trainable (although not always as trainable as we imagine -- cf. all of the "actually, losing weight long-term is almost impossible" studies, and the "aerobic exercise doesn't help you lose weight" ones). And it is plainly apparent that training one aspect of fitness does not magically improve all the others. If you go the gym and lift things with your arms, this will make your arms strong; it will not make your legs strong, or make you a fast runner. Because one can actually do such things -- many people do -- it is obvious from direct experience that "general fitness" is not a natural category, not a concept that maps naturally onto the way the body actually works. It is, at most, a useful statistical shorthand for the differences between bodies, in some zoomed-out, averaged sense.
Statistical shorthands can be quite useful indeed. They can sum up the variation in a population in a descriptive sense, and this can be essential for a clear understanding of the population, even if they have no relation to any of the individual-level physical mechanics, or only a cartoonish relation, not to be taken literally. (Cf. statistical mechanics in physics.) A variable called "general fitness" might well do a very good job of summing up how the population varies in its performance on physical tests. But it would not be much of a scientific discovery. The man on the street already knows about "general fitness," intuitively; and he also knows that there is more to the story than that. An account of human physical performance that only includes "f," the general factor, will say strictly less than what the man on the street knows. To insist that we throw away our intuition, and replace it with an "f"-only theory of fitness, is to move backwards.
Anyway, about the actual book: I read it months ago and don't remember very much about it. I remember that it did a decent job of arguing against some actual common misconceptions about intelligence research. But I also remember -- I was thinking about this as I read it -- that it said nothing at all that would not apply just as well to the hypothetical field of "Fitness Research" and its "f."