The purpose of Experiment 2 was to repeat the procedure used in Experiment 1 but to equate the number of reinforcers that could be obtained on the fixed ratio and improving progressive schedules. Thus, in Experiment 2, independent of the pigeons choice, four reinforcers could be obtained for choice of either schedule. If the pigeon chose the fixed interval schedule, it could obtain four reinforcers per trial, each requiring 23 pecks. If the pigeon chose the improving progressive schedule, it would also obtain four reinforcers, the first after 32 pecks, the second after 16 pecks, the third after 8 pecks, and the fourth after 4 pecks. If for any reason, the pigeon switched between schedules, it would obtain a total of four reinforcers on each trial.
Although the number of pecks required for the reinforcer on the improving progressive schedule were regularly decreasing, theories of delay discounting (Mazur, 1989) should give greater value to the shorter delay associated with the first reinforcer. According to delay discounting, at the time of initial choice, later reinforcers should be discounted to the point that they should have greatly reduced value, especially for pigeons, a species known to have very steep delay discounting functions (Green & Meyerson, 2004).
Method
Subjects
The subjects in Experiment 2 were nine unsexed White Carneau pigeons purchased from the Palmetto Pigeon Plant, Sumter, SC. All the subjects had experience learning to discriminate colors. All of the pigeons were housed and were treated as were the pigeons in Experiment 1.
Apparatus
The apparatus was the same as that used in Experiment 1
Procedure
The pretraining in Experiment 2 was the same as in Experiment 1. The training in Experiment 2 was similar to that of Experiment 1 with the exception that choice of fixed ratio schedule allowed the pigeon to obtain four successive fixed ratio reinforcers (that required 24, 24, 24, 24 pecks), the same number of reinforcers as choice of the progressive schedule (that required 32, 16, 8, 4 pecks). The design of Experiment 2 is presented in Figure 4. If a pigeon switched from one schedule to the other during a trial, it could receive only four reinforcers on that trial (in fact, that never happened). Once again, each session consisted of 48 trials, with a 5-s intertrial interval. Because there was little evidence of change in preference over training in Experiment 1, there were only 15 sessions of training in Experiment 2. Again, sessions were conducted six days a week.
Results
Once again, the pigeons quickly showed a preference for the progressive improving schedule. The training results are presented in Figure 5. A one-way repeated measures analysis of variance performed on the preference scores as a function of training session indicated that the effect of sessions was not significant, F(1,14) = 1.35, p>.05. When the data were pooled over all 15 training sessions and tested for schedule preference, the preference for the progressive improving schedule (.67) was significantly different from chance (.50), t(11) = 5.20, p = .0003, d = 3.14.
Discussion
The results of Experiment 2 indicate that in spite the fact that the number of reinforcers were equated for the fixed-ratio 24-peck alternative and the improving progressive schedule starting with 32 pecks, the pigeons showed a significant preference for the progressive decreasing schedule. These results are inconsistent with what one would expect for pigeons based on their steeply hyperbolically decreasing delay discounting functions (Green & Meyerson, 2004; Laude et al., 2014).
General Discussion
The results of the present experiment confirm and extend the result reported by Chandel et al. (2021) who found that pigeons left a progressively increasing ratio schedule earlier than would have been optimal. They left earlier than would have been optimal as indicated by the fact that they chose a larger number of pecks to reinforcement (on a signaled fixed ratio schedule) over a smaller number of pecks to reinforcement (on a progressive schedule with an increasing number of pecks to reinforcement). They did so presumably because the smaller number of pecks on the progressive schedule was going to be followed by an increasing number of pecks to the next reinforcers.
In the present experiments, we tested whether pigeons would prefer to choose to make a larger number of pecks (32) to reinforcement over a smaller number of pecks to reinforcement (23), if the larger number of pecks to the first reinforcer led to a smaller number of pecks to the second reinforcer (16), etc. Delay discounting theory might predict that the pigeons would weigh the first reinforcer obtainable from choice of an alternative more heavily than successive reinforcers.
It is possible that delay discounting theory would predict that because the second reinforcer on the improving progressive (decreasing requirement) schedule would come earlier than the second reinforcer on the fixed ratio schedule (16 pecks vs. 23 pecks), the progressive schedule might be preferred. However, the added delay from the initial schedule choice to the second reinforcer should make the value of the difference between 23 pecks and 16 pecks required to obtain the second reinforcer quite small. Furthermore, the total number of pecks to the second reinforcer following choice of the fixed ratio schedule, 46 pecks (two times 23 pecks), is still fewer than the total number of pecks to the second reinforcer following choice of the progressive decreasing requirement schedule, 48 pecks (32 plus 16 pecks).
The present research confirms and extends the finding that pigeons are capable of serial pattern learning. That is, they are capable of taking into account not only the effort and delay to the next reinforcer but also the general pattern of increasing or decreasing effort and delay to successive reinforcers. More generally, this finding has implications for the ability of animals to make choices based on their prospective memory (Cook et al., 1985; Zentall et al., 1990) or future planning (Raby et al. 2007). That is, they have the ability to integrate the pattern of responses and reinforcers to result from their choices.