Concurrent generalization gradients for food-controlled and shock-controlled behavior.

Pigeons pecked a disk for food or to turn off shock.

The birds saw colored lights that ranged from red to green.

After each bird learned which color meant food and which meant safety, the team tested how the birds responded to in-between colors.

They ran the same test again and again to see if the gradients changed.

Food-reinforced pecking dropped off quickly when the color shifted.

Shock-avoid pecking stayed high across more colors.

Extra discrimination training made both gradients steeper.

By the last test the food and shock curves looked the same.

Harrison et al. (1975) later showed the same steepening happens with avoidance baselines, so the effect is not tied to food or shock.

Touchette (1971) pushed training to 64 sessions and the gradients stayed steep, proving the change lasts.

HONIBOWER et al. (1964) seems to disagree: punishment with shock gave steep gradients, while HEARST (1962) says shock gives flat ones.

The difference is procedure: K used direct punishment; E used avoidance.

Avoidance keeps responding strong across colors, punishment shuts it down.

When you teach a new skill, the reinforcer you pick shapes how tightly the learner sticks to the target stimulus.

Food or praise gives narrow stimulus control, good for precision tasks.

Avoidance or escape gives broader control, useful for safety skills.

If you need sharper boundaries, add clear S-delta stimuli and extra discrimination trials; the lab shows the gradient will tighten.