A nonshocked, experimentally naive dog, when placed in a shuttlebox, typically behaves in the following way: at the onset of the first electric shock, the dog defecates, urinates, howls, and runs around frantically until it accidentally scrambles over the barrier and escapes the shock. On the next trial, the dog, running and howling, crosses the barrier more quickly. This pattern continues until the dog learns to avoid shock althogether.
But our dogs were not naive. While in a harness from which they could not escape, they had already experienced shock over which they had no control. That is, nothing they did or did not do, affected their receipt of shock. When placed in the shuttlebox, these dogs reacted at first in much the same manner as a naive dog, but not for long. The dogs soon stopped running and howling, settled down and took the shock, whinning quietly. Typically, the dog did not cross the barrier and escape. Instead, it seemed to give up. On succeding trials, dog made virtually no attempts to get away. It passively took as much shock as was given.
After testing alternative hypotheses, we developed the theory that it was not trauma per se (electric shock) that interferred with the dog's adaptive responding. Rather it was the experience of having no control over the trauma. We have found that if animals can control shock by any response - be it an active or passive one - they do not later become helpless. Only those animals that receive uncontrollable shock will later give up. The experience in the harness thought the dog that its responses did not pay, that his actions did not matter. We concluded that the dogs in our experiments had learned that they were helpless.
In our animal experiments, we knew that only when the dog learned to escape the shock, only when it learned that it could control its environment, would a cure for its learned helplessness be found.
At first, we could not persuade the dog to move to the other side of the box, not even by dropping meat there when dog was hungry. As a last resort, we forcibly dragged the dog across the barrier on a leash. After much dragging, the dog caught on, and was eventually able to escape the shock on its own. Recovery from helplessness was complete and lasting for each animal. We can say with confidence, that so far only "directive therapy" - forcing the animal to see that it can succed by responding - works reliably in curing learned helplessness."
(Martin P. Seligman, "Fall Into Helplessness", PT June 1973.)