Edwin Martin

Grammatical factors in sentence retention

A rationale for indexing the structural complexity of sentences was introduced and an experiment reported that demonstrated the relationship between this index and sentence retention. The proposed measure entails a phrase-structure analysis of the sentence and a counting of the grammatical commitments incurred by each word of the sentence. A word is said to be structurally embedded in a sentence to the extent that it determines the structure of those parts of the sentence that follow. In a six-trial free-learning experiment where sentence complexity and sentence kind were manipulated independently and sentence length held constant, sentences of lesser indexed complexity were recalled significantly more frequently than sentences of greater complexity. The role of sentence kind was found to affect recall, but not in the systematic way predicted by the transformation-grammar model.

Stimulus recognition in aural paired-associate learning

Aural paired-associates were learned by the study-test method. On test trials, the study-trials stimuli were randomly intermixed among new stimuli and the Ss, prior to responding, had to indicate whether the presented stimulus was old or new. It was found that correct responding was contingent upon stimulus recognition, and that correct responding given nonrecognition of the stimulus was at the chance level.

Short-term memory for sentences

Short-term memory for active and passive sentences at two levels of grammatical complexity was tested at four retention intervals, 0, 10, 20, and 40 sec. Sentence forgetting was analyzed in terms of differential word-class forgetting. It was hypothesized that S s selectively focus on key word classes, with grammatical structure as the guide to selection, and generate recall sentences around retained elements.

Independence of A-B and A-C associations in retroaction

After learning an A-B list to a criterion where every pair was correctly anticipated at least once, an A-C, C-D mixed list was introduced for 12 trials. Degree of A-C learning was varied within the list. Contingency analyses on the recall of B and C responses in a subsequent MMFR test were done separately for each A-C condition and for all A-C conditions combined. Either way, the B and C responses were recalled independently of each other. The results provide generality for the DaPolito independent retrieval phenomenon and affirm McGeochs independence hypothesis.

Sentence length and sentence retention in the free-learning situation

The hypothesis that sentence length is not a significant factor in sentence recall was tested and rejected. The lengths studied were 5, 7, and 9 words. All sentences were of the active-affirmative type. A structural-complexity factor was found to play no role in recall.

Stimulus component independence

After paired-associate learning to one of three degrees of learning with compound stimuli, the stimulus components were presented in isolation and S required to respond with the other components and the response. Response recall failure precluded recall of other components. This result was shown not traceable to a small number of such events, to item selection, or to the possibility that S forms intercomponent associations only after forming component-response associations. When the response was recalled, recalls of other components were frequent but stochastically independent of each other. Nevertheless, standard stimulus selection phenomena occurred. These analyses deny the general assumption that stimulus components become interassociated during paired-associate learning.

Serial learning: A multilevel access analysis

In Experiment I the lists were 36 and 48 unrelated words. Each was divided into successive groups of four words and learned to a perfect criterion. In Experiment II the lists were made up of six categorical groups of five exemplars each. Degree of learning was varied. In both experiments serial anticipation learning was followed by ordinary free recall and free recall under speed stress. Analyses of acquisition and of both recall tests indicate that group access is a marked function of serial position but that within-group retrieval given group access is constant over serial position. It is argued that serial learning proceeds simultaneously at different levels of representation.

Memory codes and negative transfer.

A view of transfer phenomena in terms of composite memory codes, code independence, and modification of sampling probability over a stimulus-defined retrieval search space is discussed.

Retroaction, Recovery, and Stimulus Meaningfulness in the A-B, A-Br Paradigm

The role of stimulus meaningfulness in paired-associate transfer, retroaction, and recovery was examined. Stimulus meaningfulness did not affect performance in second-list learning or in tests for retroaction in the A-B, A-Br paradigm. Relative to an A-B, C-B control paradigm, however, both transfer and retroaction were more negative with higher stimulus meaningfulness. Marginal absolute recovery was observed in the A-B, A-Br paradigm over repeated retroaction tests.

Stimulus pronunciability in aural paired-associate learning

An experiment was conducted in an attempt to assess the merits of a stimulus availability interpretation of stimulus pronunciability (PR) effects in aural paired-associate learning. Stimulus PR was varied within lists of 8 trigram-adjective pairs. The study-test technique was used. The intra-SR interval of study trials was either 2, 4, or 6 sec, thus allowing periods of different length in which to manipulate stimulus availability at the time of the response event. Subgroups of S s either counted backwardly during this interval, repeated the stimulus aloud, or were left uninstructed. Stimulus recall was given at the end of learning without forewarning. The results indicate that stimulus repetition efficiently serves to maintain availability over the intra-SR interval. Counting backwardly during this interval, however, markedly impedes acquisition, but not differentially from the stimulus-repetition condition over levels of PR as would be expected from short-term retention studies involving pronunciability as a variable. The failure of the three intra-SR activities to interact over levels of PR both during acquisition and in post-acquisition stimulus recall makes differential availability (in the sense of recallability) according to PR an untenable explanation of stimulus-PR effects.