Tuesday, November 10, 2015

Pavlov's Dogs

Stephen Lovell reviews Daniel P. Todes, Ivan Pavlov: A Russian life in Science in the TLS. Everybody knows Pavlov as the pioneering behaviorist who trained dogs to salivate at the sound of a bell. But maybe that's all wrong:
More interesting still is the extent to which these tensions were also present in Pavlov’s thinking. Like most other scientists of his generation, he professed the absolute authority of the “fact”. The task of the scientist was to accumulate a large amount of data from rigorously controlled and meticulously conducted experiments. The conclusions would then take care of themselves. Theorizing without data was worthless, and Pavlov was uncompromising in his critique of those he found guilty of this sin. . . .

So far, so neat and tidy. The truth, however, was that Pavlov did not stick to this set of intellectual procedures. Here Todes’s biography proves not merely definitive, but redefining. Pavlov was in fact a highly intuitive thinker, and an ambitious one: he was always striving to make the biggest claims his “facts” would allow. His co-workers conducted hundreds of experiments to his precise instructions; his role was to survey the data they gathered and pick out, as Todes puts it, “the signal amid the noise”. The accumulation of experimental data actually made underlying patterns harder rather than easier to discern, which is where the art of interpretation came in. Pavlov often used metaphor to make the jump from analysis to synthesis. He liked to characterize the digestive system as a “chemical factory” and the mind as a “machine”; later on, as his attention switched from the mechanistic particulars to the psychical whole, he referred to the cortex as a “grandiose mosaic, a grandiose signalizing switchboard”. . . .

One part of the stereotype is correct: Pavlov and his co-workers spent a great deal of their time getting dogs to salivate (though their preferred stimuli were the buzzer and the metronome, not the bell). But, as Todes shows in an exposition both lucid and nuanced, that was merely the starting point. Dogs for Pavlov were experimental subjects rather than machines for replicating results. He and his colleagues developed close working relationships with their animals (even as they tortured them). Control groups were unthinkable: all the dogs were individuals. This allowed Pavlov to conduct elaborate sequences of experiments on the same dog, with multiple combinations of stimuli, but quickly forced him to confront an inconvenient fact: dogs, like human beings, were different. Excitation and inhibition were not universal mechanisms, but varied in their intensity and interrelations from one animal to the next.
So Pavlov's dogs were all different, and they weren't conditioned with a bell. The things one learns about that turn out not to be true.

No comments: