What if William Gull were to rise from the grave?

Using a historical thought experiment grounded on the perspective of William Gull (d. 1888), this manuscript critically examines the progressive reconceptualization of anorexia nervosa (AN), contrasting contemporary diagnostic criteria with the original clinical descriptions. Historically, AN was defined by observable signs such as severe weight loss, food restriction, hyperactivity, and amenorrhea. Over time, however, DSM criteria have increasingly privileged unobservable psychological symptoms-particularly fear of weight gain and body image disturbance-while de-emphasizing or eliminating clinical signs. This shift has culminated in the introduction of "atypical anorexia nervosa" (AAN) in DSM-5, allowing a diagnosis of AN in individuals of normal or above-normal weight. We argue that this development reflects a broader epistemic move from signs to symptoms, largely driven by the widespread acceptance of Hilde Bruch's psychopathological model. Evidence from non-Western populations and Western clinical samples suggests that fear of fatness and body image disturbance are not universal features of AN. We further highlight the marginalization of hyperactivity, an early and prevalent clinical sign, in both diagnostic systems and research. Furthermore, progress in understanding underlying psychopathological processes in AN has not been mirrored in improved clinical outcomes. Finally, we propose a reorientation of AN diagnosis leveraging DSM-5 alternative criteria emphasizing observable behavioural phenotypes, with the aim of reducing diagnostic heterogeneity and enhancing the translational validity of animal models-specifically the activity-based anorexia paradigm linking starvation, thermoregulation, and excessive activity.