Lieberman et al. (1979) and Thomas et al. (1983) found an enhancement of learning in a two-choice, delayed-reward, spatial discrimination task with rats if every choice response was followed by a salient event. They proposed that the salient event marked the preceding choice response in memory so that the subject was more likely to recall it when given reinforcement.
The experiments reported here tested whether a marking effect could also be found with a visual discrimination using different subjects – pigeons – and different apparatus – the operant chamber. Experiments 1 to 8 involved a discrete-trial procedure and a variety of parameters. Following choice responses with a marker did not facilitate learning. Two explanations for this outcome are proposed. One, early experience with the stimuli used as markers might have reduced their effectiveness. Two, the onset of the discriminative stimuli on the response key might have attracted the subjects’ attention at a level such that marking provided no additional benefit.
Experiments 9 to 11 involved an invisible—trials, free—operant procedure. The discriminative stimuli remained available throughout each session and the onset of a trial was not indicated in any way.
The first response following the start of a trial was designated the choice response. A clear marking effect was found, and the results could not be explained in terms of either arousal or generalisation decrement. This result indicates the generality of the marking phenomenon, and provides an automated procedure for its investigation.
Marking shows that an added stimulus during the gap between a response and reinforcement can sometimes facilitate learning. The implications of these results for models of information-processing in animals are discussed. Marking is examined in relationship to two areas where additional stimuli have interfered with learning — response overshadowing, and retroactive interference in delayed matching-to-sample.