LAUSR

Tony Movshon’s (2021) admirable obituary made me recall with appreciation Horace Barlow’s role in the mid-20th century revision of the place occupied by neurophysiology in the study of perception. No apology is offered for using these terms rather than the current “neuroscience” and “cognition:” The tale started at a time when Sherrington was still alive, when Shannon had recently formulated information theory, and when the gist of the discipline could be contained, for the profession, in Fulton’s moderate-sized volume “Physiology of the Nervous System,” (Fulton, 1943) or, for the student of experimental psychology, in 80 of the 1400 pages of Stevens’ (1951) handbook. Perception had, of course, been an enduring topic. At the time, as an escape from the conceptual constraints of positivism and behaviorism on one side, and underdefined formulations of Gestalt on the other, reductionism offered a program. But it was only that, a program. No path charted the course between Fechner’s K€ orper and Seele. So when experimental science emerged newly energized after its success in helping win WWII, it tackled the problem at its edges. The first target was the front-end—the eye, its optics and the retina, the ear and cochlea, the muscle spindle. We knew that the nervous system, often called the brain, was interposed between the physical changes that acted as a stimulus and the resultant changed state of the organism, and we were not deterred by the philosophical tangle that one would expect on the journey’s completion. The strategy we adopted in the early 1950s can be well described by reference to the Artificial Neural Nets of today’s Artificial Intelligence: thorough characterization of the input, stipulation as to the output layer, and a multiplicity of intermediate interconnected nodes, whose activation and the traffic between them we accepted as proceeding in an unknown manner. Attempts to tap into the system were not too promising, even in the best of hands, in part because of limitations in technology, of restrictions to anesthetized preparations, and of the sheer magnitude of the number of involved elements In our laboratories, we either lumped