Classical Conditioning

Classical Conditioning

There is something puzzling and not quite believable about Pavlov and his dogs. The dogs instinctively salivate to the sight and smell of food—this is an inborn reflex reaction. The claim is that if you sound a bell a number of times at the same time as presenting food you will get the salivation reflex to the bell alone: that is, the bell will elicit salivation. The dog will learn to salivate to a bell. But this is not generally true of instinctive reflexes: suppose you sound a bell whenever you tap someone’s knee—can you elicit the patellar reflex with the bell alone? Definitely not: you need the stimulus of the hammer. The reflex is hardwired and geared to the physical stimulus striking the knee. You might expect to see your leg go up soon after the bell sounds, because you expect it will be tapped; but it won’t actually go up without the tap. Similarly for the blink reflex: it won’t activate to the sound of a bell, no matter how many repetitions you perform. So, why would the salivation reflex be activated by the mere sound of a bell? It is different with a voluntary action: suppose you sound a bell just before you electrify the ground on which a dog is standing; the dog will soon learn to anticipate the electric shock and jump off at the sound of the bell.[1] But salivating isn’t a voluntary action; the dog can’t decide to salivate. The dog might think, “It would be good if I start salivating now, because food is about to be given to me”, but this can’t actually activate its salivary gland; it’s a reflex that acts independently of the will. Similarly, you couldn’t condition someone to sweat when a neutral stimulus is presented, after associating this with a sweat-inducing stimulus—a bell will not cause a person to sweat even if previously paired with vigorous exercise or a sauna. Nor have I ever heard anyone claiming that it can. So, what is going on with Pavlov and his dogs? Was the experiment ever replicated? I assume it was, but the conclusion apparently violates basic laws of reflexology. And it certainly doesn’t generalize to other sorts of instinctive reflex. How much did the dogs salivate, according to Pavlov? Was it the same as with a food stimulus or considerably less?

I have a theory, but I have never heard it stated before, and it is at odds with behaviorist psychology. It assumes that Pavlov did observe the effect he claims, incredible as it seems—and we should always be careful about the claims of psychologists. The theory is that the dogs imagined the food when they heard the bell—and this activated their salivary glands. Perhaps the imaginative act was quite vivid, approximating an actual perception; even an imagined smell entered their hungry consciousness. We might say they were under a kind of illusion of food presence—it seemed to them that food was under their nose or on the way. The reflex might then be activated by the imaginative act. Consider a dog in the wild hunting for food: even a quite neutral stimulus such as a certain type of bush might cause the hungry dog to imagine its prey hidden in the bushes, as so often in the past; it then begins to salivate in anticipation of food, on the principle that it’s good to get the juices flowing in order to consume the desired prey. In other words, a sequence of mental processes led to the salivation; it wasn’t just an unmediated response. So, the dog hears the bell and knows food is about to be delivered; it imagines the food and this acts on its nervous system like a perception of food; so, it salivates. The reflex is still geared to a mental presentation of food not to the bell as such. By contrast, there is no point in reflexively kicking up your leg at the sound of a bell, or reflexively blinking, or sweating. This theory at least makes sense of an otherwise incomprehensible experimental result. But it is unlikely that the salivation will be quantitively the same as the unconditioned salivation response; there will just be an incipient salivation response (no point in wasting good saliva when no food is forthcoming). In general, I find myself quite skeptical of Pavlov’s alleged result, despite its canonical status in the field. Did Pavlov really just discover that imagining food can cause an animal to salivate? That is not what we are usually told. And the finding is difficult to accept unless under this interpretation. Instinctive reflexes are stimulus-bound—tied to a specific type of stimulus. They can’t be triggered by associated arbitrary stimuli.[2]

[1] This is a completely unethical experiment, but it serves to make the point as a thought experiment.

[2] I first heard about Pavlov’s experiments on so-called classical conditioning nearly sixty years ago and yet this objection has only recently occurred to me. This is very typical of psychological experiments: there is always a gap between the observable results and the interpretation of these results. Pavlov’s experiments shaped scientific psychology, but if I am right they were badly misinterpreted (even if they were valid experimentally). I suspect their appeal lay in the fact that they were surprising and contrary to common sense—they made psychology seem interesting. But properly interpreted, they are either unsurprising or invalid (at least for reflexes generally). This is rather alarming.