John F. Hall

Backward conditioning in Pavlovian type studies

Authors of almost all learning texts have indicated that unconditional/ stimulus-conditional stimulus (UCS-CS) trials (backward conditioning) utilized in classical conditioning experiments will not result in establishing a conditional response (CR) (excitation). Recently, it has been proposed that backward conditioning can take place although whether UCS-CS trials result in excitation or inhibition has been left unspecified. It is proposed that the diversity of findings can be attributed to inadequate methodology and the use of a classical-instrumental transfer paradigm. When only traditional Pavlovian conditioning studies are examined, the experimental findings suggest that UCS-CS trials will result in inhibition.

Adaptation, Sensitization, Forward and Backward Conditioning, and Pseudoconditioning of the GSR

Considering the degree to which the GSR is influenced by verbally induced sets (Cook & Harris, 1937; Silverman, 1960) , there has been a notable lack of concern about the degree to which the specific pairing of CS and UCS in the GSR conditioning situation actually modifies b e h a v i ~ r . ~ It is the purpose of this smdy to compare the performances of Ss exposed to the usual classical conditioning methodology with the performances of Ss given pseudoconditioning, backward conditioning, and sensitization training procedures. Essentially, this constitutes an experiment which tests for conditioning as something over and above the behavior modifications that accompany other modes of stimulus presentation.

Recall versus recognition: a methodological note.

Some current investigators have made direct comparisons of recognition performance with cued recall. In other instances, indirect comparisons of recognition and recall have been made by observing the difference in recognition performance with different types of target material, for example, high- and low-frequency words; this difference has been then compared with the difference obtained when recall is used to measure retention of the same type of target materials. Finally, some experimenters have examined how recognition performance is related to the type of target material that is used, for example, words, pictures, and so on. The results obtained from all of these studies have been used to make inferences about the nature of the memorial process. We have attempted to call attention to the generally accepted but neglected finding that performance on any recognition test can be manipulated by changing the characteristics of the distractors, thus casting doubt on the generality of the experimental findings as well as their theoretical implications.

The reliability of free association responses

Free association responses were obtained from high and low frequency stimulus words for 61 Ss. Seven or 21 days later, Ss responded to the same stimulus material. Results indicated that neither word frequency nor interval between sessions influenced the number of identical responses wade by the S. Findings indicated that an S will make the same response to the same stimulus word presented on two different occasions, about 50% of the time.

Reconstructive and reproductive models of memory

The adequacy of reconstructive and reproductive models of memory has been in dispute ever since Bartlett’s (1932) publication of his volume Remembering. Experimental as well as anecdotal evidence suggests that neither model can adequately account for all of the findings. We would suggest that to be adequate, an explanation of verbal memory should focus on a memory classification based on behavioral measures. Thus, it is suggested that verbal memory consists of at least two basic types of remembered material: (1) verbatim memory and (2) memory for gist. The latter category involves a frequently encountered process that we have identified as abstractive congruence; a reconstructive process, abstractive incongruence, is also found, but it is found much less frequently.

Free recall as a function of type of encoding and word frequency

The manipulation of strategies provided subjects learning a list of high-frequency and rare words was expected to result in an interaction between word frequency and type of presentation. If the hypothesis was correct, then an associative strategy (e.g., linking or associating list words in a sentence, phrase, etc.) would maximize recall for high-frequency words, whereas a repetitive procedure (e.g., repetition of each individual list word) would be best for the recall of rare words. Experimental findings examining these two types of presentation (i.e., repetitive and associative) and three levels of word frequency (i.e., high frequency, low frequency, and rare) supported the hypothesis.

List organization and recognition memory

Some investigators have been unable to find that list organization aids recognition, although it does aid recall, thus lending support to a model of memory that assumes that the memorial processes involved in recall differ from those operating in recognition. Other experimenters, however, have found organization to have a facilitating effect on recognition. The present study examined recognition performance using organized and unrelated word lists, with two types of distractors being used. Results indicated that organization aided recognition, regardless of the type of distractor employed.

PROACTIVE, RETROACTIVE, AND COACTIVE INHIBITION WITH MEANINGFUL MATERIAL

In examining the contribution of inhibitory effects to forgetting, practically all investigators have been interested in investigating the operation of either proactive or retroactive effects. Most nonlaboratory forgetting situations, however, appear to involve the simultaneous action of both. Thus, in contrast to the proactive (learn A, learn B, recall B) , or retroactive (learn A, learn B, recall A) paradigm, a paradigm for the study of simultaneous action would involve the sequence: learn A, learn B, learn C, recall B. The learning of A provides proactive inhibition (PI) while the learning of C provides retroactive inhibition (RI) for the recall of B. For convenience, we have termed this paradigm, coactive inhibition (CI). Just as PI effects have been frequently compared with those of RI, it seemed appropriate to compare the contribution of this type of inhibition with the other traditional two. The type of material with which we chose to investigate this relationship was meaningful adjective-noun combinations. This was in keeping with the plea made by some investigators (1, 7 ) to enlarge our experimental materials in the area of human learning to include those judged meaningful. An unpublished study using meaningful combinations (short answer, clumsy dancer) indicated that such combinations were capable of being ranked by judges on a meaningless-meaningful continuum and, furthermore, that such learning was a function of the ranked meaningfulness of the combination. An added advantage of such material is that standard paired-associate presentation can be used. The primary purpose of Exp. I was an attempt to obtain and compare PI, RI, and CI effects with the type of meaningful material which we have described. EXPERIMENT I Sabjects.-Ss were 64 undergraduates, randomly assigned to one of four groups as indicated in Table 1. They were not told the actual purpose of the experiment but their general instructions indicated that Es were interested in the relationships existing among a number of learning tasks. Material, upparatas, and procedare.-Three lists (A, B, and C) of 15 meaningful adjective-noun combinations were used as the antecedent, original, and interpolated learning tasks. An attempt was made to maximize inhibitory effects during the final recall period by selecting similar stimulus but neutral response words in accordance with the model proposed by Osgood (9 ) . This

The conditional emotional response as a model of pavlovian conditioning

Many investigators have assumed that the conditional emotional response (CER) with its attendant response measure, response suppression, can be used as a model for Pavlovian conditioning; that is, that the experimental results obtained with the CER will yield findings similar to those obtained with the direct conditional stimulus (CS)-conditional response (CR) paradigm,e.g., conditioning of the eyeblink, electro-dermal (EDR), finger withdrawal,etc. Several problems, however, have emerged with this assumption. The first is a methodologic-statistical one. Measurement of response suppression using the familiar ratio measure assumes that different base rates of responding should play no role in determining the size of the suppression ratio, and operant responding from one pre-CS period to another will remain stable. These assumptions have been found to be in error. In addition, the statistical assumptions of normally distributed data and homogenous variances of the experimental and control groups also have been called to question. Some experimenters, however; have taken the position that if CER and direct CS-CR experimental findings are in general agreement when a particular conditioning variable has been manipulated, the methodologic-statistical problems that we have identified can be ignored. The experimental evidence suggests that although such comparisons frequently indicate an identity of findings, there are sufficient exceptions in which caution is urged in assuming that the two paradigms are measuring the same learning correlate.

Memory for surnames

Two experiments were undertaken to examine if surnames and words that were orthographically identical differed in terms of (1) association value and pronounceableness and (2) characteristic associative structure. A third experiment examined free recall for these two types of materials. Results indicated that surnames and words were equally easy (or difficult) to pronounce and did not differ in rated number of associations. Free association responses to both types of items suggested that surnames were responded to as words. Lists of these items, identified as either surnames or words, were recalled with equal facility.