Oscillatory behavior during approach-avoidance conflict.

Fantino (1967) put hungry rats in a box with one lever. Pressing produced both food and electric shock at the same time. The schedule mixed reward and punishment in the same moment.

The researcher varied how often shock came with the food. He then counted how often rats started the stimulus and how long they stayed in it.

Rats did not freeze and did not feast. They pressed at a middle speed and quit the stimulus faster as shock grew more frequent. The behavior swung back and forth—approach, then retreat—like a pendulum.

The pattern is called oscillatory approach-avoidance. It peaks when reward and aversive strength are nearly equal.

Shahan et al. (2023) later showed that shock can suppress behavior without becoming a clear signal. Their data say punishment works through raw aversive control, not through neat stimulus control. The two papers sit side-by-side: 1967 maps the tug-of-war, 2023 names the rope as shock itself.

WINOGRAD (1965) had already shown that unavoidable shock wrecks an old food discrimination. The 1967 study moves one step forward: it traps the rat between the same shock and food in real time, revealing the moment-to-moment vacillation instead of long-term breakdown.

Burgess et al. (1971) also used shock to punish fighting. When they removed the shock, the fighting returned. Together these papers show that shock can both create wobbly conflict and stamp out other behaviors, yet the change is reversible once the shock stops.

If you run concurrent reinforcement and punishment programs, watch for vacillation, not steady suppression. The learner may approach, back off, then approach again. Use data paths, not just rate, to see the swing. When the aversive side grows stronger, expect quicker exits and shorter bouts. Balance the two contingencies or the behavior may wink out entirely.