On the discriminability of stimulus duration.

Researchers worked with pigeons in a lab chamber.

The birds had to peck one key after seeing a short light and another key after seeing a long light.

The team made the two lights more alike or more different across conditions.

They also checked if changing the payoff for each key changed the birds’ choices.

When the short and long lights were very different, the pigeons almost never mixed them up.

As the gap between the two times shrank, accuracy fell.

Surprise: keeping the grain payoff equal or unequal did not budge the accuracy curve.

Duration difference alone controlled the birds’ choices.

Cameron et al. (1996) extended this idea. They showed that a long sample light can work like a green light for food: pigeons match faster when the long light means “food soon.”

THOMAS et al. (1963) set the stage. They first proved pigeons need to master side-by-side matching before you can ask them to remember a stimulus for a few seconds.

Neuringer (1973) seems to clash. That study found flat generalization after presence-absence click training, while D et al. got steep accuracy jumps with bigger duration gaps. The gap is method: presence-absence gives one cue, while comparing two durations gives two clear anchors.

If you want a learner to tell two stimulus lengths apart, first widen the gap. Make the short one really short and the long one really long. Once accuracy is high, slowly shrink the gap. Don’t worry about keeping the reinforcers perfectly equal; the size of the difference matters more than the size of the payoff.