This article reports evidence that a verbal association is symmetrical if its units are equally available. An item's availability was increased by having S produce it from memory. By the PA method S learned pairs of disyllables, which together formed an associative structure. The stimuli of some pairs were responses in other pairs; these stimuli thus became available. S was then asked for 2 free associations to every disyllabic. 4 conclusions were drawn from the response hierarchies: (a) Available items occurred more often as responses in all hierarchies, (b) A backward association occurred as readily as a learned forward association if the PA stimulus was available, (c) Associations occurred between items that had shared a common associate, even though these items had not coappeared. Such associations, however, only manifested themselves when their response was available, (d) The data did not support a simple principle of mediation as an explanatory concept.

Sixteen male college students participated in a semantic conditioning experiment which consisted of five successive daily sessions of 40 conditioning trials each preceded by the presentation of a list of common words. The conditioned stimulus (CS)-unconditioned stimulus (UCS) interval during conditioning was 10 sec, and the UCS was a ,5-sec 110 dH (A) white noise. Neutral unrelated filler words were interspersed between conditioning trials. On the fourth and fifth days, the room lights were dimmed for .5 sec cither 5 sec prior to the CS or simultaneously with the CS on 5 different trials. A second kind of stimulus change was examined following 40 conditioning trials on Day 5, omission of the UCS, and the effects of different instructions. The palmar galvanic skin response (GSR) and cephalic vasomotor responses were recorded throughout the experiment. During conditioning these responses were scored .5-5.5 sec and 5.5-10.5 sec following the CS, selected control words, and the UCS. Also scored were UCS omission responses during extinction.There was little evidence of differential semantic conditioning of either cephalic vasodilation or vasoconstriction, presumed measures of orienting and defensive reflexes, respectively. There was extensive habituation of the GSR within and between days in the sessions preceding conditioning. Extensive inhibition of reinforcement, decreased conditioned responses with continued reinforcement, occurred during conditioning of the GSR. Superior differential conditioning of first-as compared to second-interval GSRs was obtained each day with no evidence that second-interval responses anticipate the UCS. The two kinds of stimulus change produced different effects. Dimming of the lights simultaneously with the CS resulted in a large augmentation of the GSR at the time, with no effect on differential semantic conditioning on subsequent trials. Omission of the UCS, in contrast, produced augmented differential conditioning on the following trial. The results were interpreted in terms of voluntary and involuntary orienting responses reflected in the GSRs induced in the various phases of the experiment.

No abstract