Controlling and Predicting Unpredictable Behavior.

de Souza Barba (2015) wrote a theory paper. He asked why behavior sometimes looks random. He looked at two ideas about how reinforcement makes acts switch around.

The paper used math models, not people. It stayed in the world of pure theory.

The two ideas can live together. One says animals learn a special 'vary' rule. The other says choices stay in balance because each act blocks the next.

Lourenço says both forces can work at once. Unpredictable behavior can pop out when these forces tug against each other.

Kuroda et al. (2018) gave real data that backs this up. They showed pigeons work harder when response rate and reward rate line up. Their experiment gives legs to Lourenço's balance idea.

Hammond (1980) did the same trick years earlier. He slowly cut the link between lever presses and food. The rats slowed their pressing step by step. That early lab work foretells Lourenço's point that the response-reinforcer link itself drives the pattern.

Cowie et al. (2016) seem to push back. Their review says stimulus cues, not the reinforcer, steer choice. But the gap is only skin deep. Lourenço talks about how contingencies create variation; Cowie talks about what animals notice. Both can be true at once.

When a client's behavior looks random, do not jump to 'sensory seeking' or 'automatic reinforcement.' First check the schedule. Mixed or thin rewards can make response strength seesaw. Try tightening the contingency or adding a clear stimulus cue. You might see the 'random' acts settle into a cleaner pattern.