The thinking is that if we cut back sufficiently on how much we eat, surely we will get that excess fat out of our bodies, regardless of how seemingly trivial the overeating might have been that produced that fat. In the 1960s and ’70s these calorie-restricted diets were often known even in the research literature as semistarvation diets. I’m going to use that terminology because it is entirely appropriate. This assumption that people will lose weight if they are starved sufficiently is certainly true. This is one reason clinical researchers and physicians from Newburgh onward were so convinced that we get fat because we eat too much. Cut back enough on the calories a fat person is allowed to eat, and the result is a less fat person. But as the Harvard psychologist William Sheldon observed in the late 1940s, starving a fat man (an endomorph, in his terminology) doesn’t actually turn him into a lean man (an ectomorph) or a muscular, athletic one (a mesomorph) any more than starving a mastiff turns it into a collie or a greyhound. For the dogs, you get an emaciated mastiff. For the humans, an emaciated fat man. So this thinking, too, has some serious problems that have to be ignored to embrace it. If you put a lean person on a semistarvation diet, you also get a less fat person—actually, an emaciated lean person. Starving or semistarving a growing child will result in an emaciated child whose growth is stunted, but no authority would ever assume, let alone state publicly, that children grow because they eat more than they expend. At least I hope not. Yet that’s always been considered the reasonable interpretation of the starve-a-fat-man observation. The important question, however, is why it is that some of us have to be chronically starved or semistarved—exercise portion control and be hungry for a lifetime—to be lean, or at least leaner, and others don’t. This is another question that is rarely asked.

Gary Taubes. The Case for Keto: Rethinking Weight Control and the Science and Practice of Low-Carb/High-Fat Eating (Kindle Locations 868-869). Knopf. Kindle Edition. 

Sheldon attended the University of Chicago, where he received a Ph.D. in psychology in 1926 and an M.D. in 1933. In 1951, after having worked at various universities, Sheldon joined the University of Oregon Medical School, where he became distinguished professor of medicine and director of the constitution clinic, which examined the relationships between physical characteristics and disease; he remained there until his retirement in 1970. Also in 1951 he became director of research at the Biological Humanics Foundation in Cambridge, Massachusetts.

Influenced by the pragmatism of American philosopher and psychologist William James and by his background as a naturalist who had also studied animals, Sheldon became convinced that the psychological makeup of humans had biological foundations. He constructed a classification system that associated physiology and psychology, which he outlined in The Varieties of Human Physique: An Introduction to Constitutional Psychology (1940), The Varieties of Temperament: A Psychology of Constitutional Differences (1942), and Atlas of Men: A Guide for Somatotyping the Adult Male at All Ages (1954). Sheldon classified people according to three body types, or somatotypes: endomorphs, who are rounded and soft, were said to have a tendency toward a “viscerotonic” personality (i.e., relaxed, comfortable, extroverted); mesomorphs, who are square and muscular, were said to have a tendency toward a “somotonic” personality (i.e., active, dynamic, assertive, aggressive); and ectomorphs, who are thin and fine-boned, were said to have a tendency toward a “cerebrotonic” personality (i.e., introverted, thoughtful, inhibited, sensitive). He later used this classification system to explain delinquent behaviour, finding that delinquents were likely to be high in mesomorphy and low in ectomorphy and arguing that mesomorphy’s associated temperaments (active and aggressive but lacking sensitivity and inhibition) tend to cause delinquency and criminal behaviour. Although his research was groundbreaking, it was criticized on the grounds that his samples were not representative and that he mistook correlation for causation.