Reinforced behavioral variability: Working towards an understanding of its behavioral mechanisms.

The team worked with pigeons in an operant chamber. Birds had to peck four times on each trial. The order of left and right pecks had to meet a variability rule to earn food.

Some birds got food only when their four-peck pattern was new. Other birds got food just for switching patterns. The study asked which rule made the birds vary more.

Pigeons produced far more varied four-peck sequences when the rule said 'be different' than when it only said 'change'. Direct reinforcement of variability beat simple changeover rules.

The result supports the idea that variability itself can be reinforced, not just the act of switching.

Moxley (1989) set the stage with a percentile schedule that raised or lowered variability on cue. Doughty et al. (2015) sharpen the tool by showing a simple threshold rule works just as well.

Nergaard et al. (2020) push back. They argue that 'reinforced' variability is really the product of quick extinction of old forms, not true reinforcement. The two papers seem to clash, but they test different questions. H et al. look at what the bird does next; Nergaard et al. look at why the old form stops. Both can be true.

Dugdale et al. (2000) and Hopkinson et al. (2003) extend the same logic to people. Adolescents with autism and college students with depression both varied more when the contingency told them to. The animal rule travels well to human clinics.

If you want a client to try new responses—new play schemes, new sentences, new problem-solving steps—reinforce the novelty itself, not just the absence of the old response. Use a lag schedule or a simple 'show me something you haven’t done' rule. Pair it with a signal as Reed (2023) shows and the effect gets even stronger. Start small, count response forms, and deliver praise or tokens only when the form is new.