Sequential dependencies in free-responding.

Shimp (1973) watched pigeons peck a key for food on a variable-interval schedule. The team tracked every peck and the tiny gaps between them. They asked: does the pattern of recent pecks change the chance the next peck will occur?

The birds worked in standard operant chambers. No extra cues were added. The only data were the birds' own response times.

Short gaps raised the odds of the next response. Long gaps lowered those odds. The local sequence mattered as much as the overall rate.

In plain words, a bird that just paused was more likely to pause again. A bird on a roll kept pecking. Reinforcement did not just control how often the birds pecked; it shaped the rhythm of their responses.

Durand (1982) later mixed VI with fixed-ratio schedules. They still saw pigeons, but now choice, not rhythm, was the focus. Their dual-sensitivity model kept the idea that local contingencies drive behavior — the same core point Shimp (1973) made.

Grosch et al. (1981) used the same VI schedule and same species. They looked at signaled versus unsignaled delays instead of gap patterns. Both studies show that tiny schedule details change performance. The 1981 paper extends the 1973 finding into the delay domain.

Sainsbury (1971) came just before Shimp (1973) and also tinkered with local delays. S used changeover delays in concurrent schedules. Both papers agree: what happens right before a response weighs heavily on what comes next.

When you shape or reinforce, watch the micro-pattern, not just the total count. If a client stalls after a 3-s pause, the pause itself may be the cue for more pausing. Reset the rhythm with quicker reinforcement or a prompt. Use the bird data as a reminder: timing talks.