Date of Award


Document Type


Degree Name

Doctor of Philosophy (PhD)




Improvement in learning successive lists has been found in both serial and paired-associate verbal learning tasks. Both intrasession and intersession effects have been observed, although it is well established that intrasession effects are greater in magnitude. Intrasession improvement in learning successive verbal tasks has been called warmup. Earlier investigators (e.g., Heron, 1929; Irion, 1949; Thune, 1950) favored an explanation of warmup similar to that used in motor skills learning. They viewed warmup as postural and attentive changes taking place within a learning session. These changes were seen as nonassociative and temporary. Warmup was regarded as separate from intersession improvement, usually called learning to learn, the latter being taken as a relatively permanent associative process.

The conventional view of warmup seems particularly inappropriate when applied to free recall verbal learning. A number of investigators have failed to find warmup in free recall of words. In a free recall study, Raffetto and Koeppel (1968) verified the fact that warmup does not occur with words, but demonstrated that warmup occurs with trigrams. An associative process, response integration, was suggested to explain the occurrence of warmup in the free recall learning of trigrams. In a subsequent study (Koeppel & Raffetto, 1969) the response integration interpretation was tested in a two-stage analysis of paired-associate learning of word pairs. As predicted, no warmup was found in the response learning stage and a significant warmup effect was observed in the associative learning stage of the task.

From these studies it was hypothesized that if either the task or the materials required associative integration, warmup would occur. Three experiments were conducted to test the hypothesis. In Experiment I a paired-associate task and trigrams were used. The paradigm involves two stages of learning (response and associative) and with trigrams both stages would require integration; therefore, a warmup effect was predicted for both stages of learning. Experiment II consisted of a serial learning task with trigrams. Since the associative characteristics of this experiment were identical to those of Experiment I a warmup effect was predicted for both response and associative stages. Experiment III involved serial learning of words. Since words necessitate little integration, no warmup effect was predicted for the response learning stage, although warmup was expected in the associative learning stage. For all three experiments the data clearly support the predictions.

Warmup is not a phenomenon that always occurs in verbal learning. That its occurrence can be predicted knox^ing the integrative characteristics of task and materials indicates that the warmup process is associative in nature. It is concluded that postural and attentive wTarmup changes are unimportant in verbal learning and that warmup and learning to learn are probably the same process. An alternative way of categorizing nonspecific transfer, based on the temporal spacing of learning sessions, is suggested.