Stimulus-reinforcer contingencies and local behavioral contrast.

Dews (1978) worked with hungry pigeons in a small chamber. A red key stayed on for three minutes. Food sometimes followed a peck, but the rate never changed.

A second green key blinked for five seconds. When it blinked, food became twice as likely on the red key. The birds never pecked the green key; it only signaled better odds.

Pecking on the red key jumped each time the green light blinked. The rise stayed inside the three-minute period. This is local positive contrast.

The boost happened even though the green key itself paid nothing. Pure stimulus-reinforcer pairing was enough.

Parsons et al. (1981) ran the same birds and boxes three years later. They saw the jump only after they removed accidental food that followed green-key blinks. The later paper says contrast needs clean procedures; the 1978 data may include accidental reinforcement.

Cicerone (1976) showed pigeons already peck a keylight that predicts food. Dews (1978) adds that this Pavlovian link can push responding on a separate working key, creating contrast.

Reed (1991) zoomed in closer. Contrast shows up in the first quarter of the three-minute stretch, not evenly across it. The 1978 finding still holds, but its timing is tighter than first thought.

You now know that brief cues paired with richer payoff can lift ongoing operant rates. Check if ‘bonus’ stimuli in your session accidentally predict better odds. If they do, you may be creating local contrast instead of steady performance. Try inserting a neutral stimulus before high-rate reinforcement components and watch if skill bursts appear; if they do, consider spacing or masking that cue to keep behavior level.