Temporal inhibition: effects of changes in rate of reinforcement and rate of responding.

Born et al. (1974) worked with pigeons on a two-part schedule. In Part A the birds pecked on a fixed-interval 60-s schedule that paid 60 food per hour. In Part B the same interval stayed, but the alternate side paid either 120 or 30 food per hour.

The team then flipped the question. They held reinforcement at 60 per hour on both sides and only changed how fast the birds had to peck in the alternate part. They watched what happened to the first few pecks right after the schedule switched.

When the alternate side paid faster (120 per hour), the birds waited longer before the first peck back on the FI 60 side. When the alternate side paid slower (30 per hour), the birds started pecking sooner. Changing only the required response rate in the alternate part did nothing.

In short, timing at the start of an FI depends on the pay rate in the other component, not on how hard the bird just worked.

Gaucher et al. (2020) extends this idea to children with autism. Their kids also adjusted timing when the payoff for slow responding changed, but only if they had strong language and IQ scores. The 1974 animal rule still holds: reinforcement rate drives the pause.

Baum (2025) gives a fresh reason why. His molar model says interval schedules stay steady at lean pay rates because the effective unit of reinforcement is bigger than on ratio schedules. That theory backs the 1974 finding that FI timing bends when the molar pay rate across components shifts.

Duker et al. (1991) seems to disagree at first glance. They showed that lengthening the delay between sample and choice hurt memory more than delaying the reinforcer. Looks like "delay matters more than rate." But their task tested remembering, not timing. Once the task is timing, rate wins—so the papers talk past each other, no true clash.

If you use mixed schedules, DRL, or mult-component teaching, know that the richer side will make the learner wait longer on the lean side. You can use this: raise the rate in one activity to create a natural pause in the next. Check language and IQ if the pause does not appear—Gaucher et al. (2020) warns that skill level changes the effect.