Coding in pigeons: Multiple‐coding versus single‐code/default strategies

To investigate the coding strategies that pigeons may use in a temporal discrimination tasks, pigeons were trained on a matching‐to‐sample procedure with three sample durations (2s, 6s and 18s) and two comparisons (red and green hues). One comparison was correct following 2‐s samples and the other was correct following both 6‐s and 18‐s samples. Tests were then run to contrast the predictions of two hypotheses concerning the pigeons’ coding strategies, the multiple‐coding and the single‐code/default. According to the multiple‐coding hypothesis, three response rules are acquired, one for each sample. According to the single‐code/default hypothesis, only two response rules are acquired, one for the 2‐s sample and a “default” rule for any other duration. In retention interval tests, pigeons preferred the “default” key, a result predicted by the single‐code/default hypothesis. In no‐sample tests, pigeons preferred the key associated with the 2‐s sample, a result predicted by multiple‐coding. Finally, in generalization tests, when the sample duration equaled 3.5s, the geometric mean of 2s and 6s, pigeons preferred the key associated with the 6‐s and 18‐s samples, a result predicted by the single‐code/default hypothesis. The pattern of results suggests the need for models that take into account multiple sources of stimulus control.