Learning and Motivation 42 (2011) 13–25 Contents lists available at ScienceDirect Learning and Motivation journal homepage: www.elsevier.com/locate/l&m Short-term memory for temporal intervals: Contrasting explanations of the choose-short effect in pigeons Carlos Pinto ∗ , Armando Machado Minho University, Portugal article info Article history: Received 4 March 2010 Received in revised form 7 May 2010 Available online 9 June 2010 Keywords: Choose-short effect Short-term memory Timing Subjective shortening model Coding model Confusion hypothesis Delayed matching-to-sample Pigeon abstract To better understand short-term memory for temporal intervals, we re-examined the choose-short effect. In Experiment 1, to contrast the predictions of two models of this effect, the subjective shortening and the coding models, pigeons were exposed to a delayed matching-to-sample task with three sample durations (2, 6 and 18 s) and retention intervals ranging from 0 to 20 s. Consistent with the coding model, the results suggested a sudden forgetting of memories for duration. In Experiment 2, to test the confusion hypothesis, the characteristics of the ITI and the retention interval differed. Contrary to the confusion hypothesis, a choose-short effect was obtained. In both experiments, a test with only two of the three comparison keys was performed. The results suggest three effects that may be controlling the birds’ responses: stimulus generalization when no retention interval is present; an increase in random responding at longer retention intervals; and, similarly, an increase in preference for the “short-sample” key at longer retention intervals. © 2010 Elsevier Inc. All rights reserved. In a study aimed at understanding short-term memory for temporal intervals, Spetch and Wilkie (1982) exposed pigeons to a matching-to-sample procedure with time intervals as samples. When the houselight was turned on for 2 s (short sample), the pigeons were rewarded for choosing one comparison key (say, Green), and when the houselight was on for 10 s (long sample), they were rewarded for choosing another comparison key (say, Red). After the animals learned this conditional discrimination, the authors introduced a retention interval, ranging from 0 to 20 s, between the offset of the sample and the presentation of the comparison stimuli. They found that accuracy on the long-sample trials decreased with the retention interval (i.e., the pigeons were more likely to choose the short-sample comparison), whereas accuracy on the short-sample trials remained high and did not vary with the retention interval. They labelled the phenomenon the “choose-short effect”. To explain the effect, Spetch and Wilkie (1983) proposed the subjective shortening hypothesis. According to it, the memory for a sample is represented quantitatively on a subjective scale of duration such that the longer the sample, the higher the memory value. During training, the animal learns to choose the long-sample comparison when the memory has a high value and the short-sample comparison when the memory has a low value. To explain the choose-short effect, the authors also assumed that, when the sample stimulus is turned off, the memory decays or “shortens”. Hence, on long-sample trials, as the memory for the long sample decays, its value becomes increasingly closer to the value of the short sample and therefore more responses to the short-sample comparison occur. In contrast, on short-sample trials, the decay of the memory does not result in incorrect choices because the shortened memories remain more similar to the memories for the short sample than to the memories for the long sample. An alternative account of the choose-short effect is Kraemer, Mazmanian, and Roberts (1985)’s coding hypothesis. This account considers that, based on the samples’ relative durations, each sample is coded categorically in a non-temporal ∗ Corresponding author at: Escola de Psicologia, Universidade do Minho, 4710 Braga, Portugal. E-mail address: carlos.arop@gmail.com (C. Pinto). 0023-9690/$ – see front matter © 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.lmot.2010.05.001