Spontaneous recovery of memory during posthypnotic amnesia.

Repeated testing of posthypnotic amnesia indicates that some Ss, initially responsive to the suggestion, show appreciable recovery of memory before the pre-arranged signal is given to cancel the amnesia. Comparison of Ss who received 2 successive memory tests during amnesia with others who received only a single test preceded by a distracting activity indicated that the recovery effect was attributable to the passage of time rather than to prior testing. There were wide individual differences in the extent of recovery, with some Ss maintaining a fairly dense amnesia on the second test. Those Ss who maintained amnesia were more hypnotizable, and showed a denser initial amnesia, than those who breached it. An analysis of subjective reports lent credence to the notion of partial response among some hypnotizable Ss who fail to meet a standard criterion of complete amnesia, and pseudoamnesia among some insusceptible Ss who appear to pass it. Some Ss reported voluntarily engaging in cognitive activity designed to induce forgetting, but these reports were related to neither the Occurrence of initial amnesia nor its persistence. A failure of memory which reflects momentary disorientation upon transition from one mental state to another should be conceptually distinguished from a reversible amnesia initiated by hypnotic suggestion. In conventional practice, suggestions for posthypnotic amnesia consist of two components: the suggestion that the subject will have difficulty, Manuscript submitted March 16, 1982; final revision received January 24, 1983. 'This research was supported in part by Grant MH-35856 from the National Institute of Mental Health, United States Public Health Service. prhe authors thank Heather A. Brenneman, Richard H. Bashner, Beverly R. Chew, Patricia A. Register, Judith Scharff, and Leanne Wilson for their assistance in conducting the experiment and analyzing the data, and Daniel M. Baer and Lance K. Canon for providing access to Ss. 3R0nald E. Shor died January 26, 1982. 'Reprint requests should be addressed to John F. Kihlstrom, Ph.D., Department of Psychology, University of Wisconsin, W. J. Brogden Psychology Building, 1202 West Johnson Street, Madison, WI 53706. D ow nl oa de d by [ 71 .1 98 .1 84 .1 63 ] at 0 6: 03 2 4 M ay 2 01 6 310 KIHLSTROM ET AL. upon awakening, in remembering the events and experiences that transpired while he or she was hypnotized; and the establishment of a signal by which the amnesia suggestion may be cancelled and the memory restored. Despite the explicit inclusion of a reversibility cue in conventional suggestions of amnesia, initially amnesic subjects may show some degree of recovery of memory before the signal is given. In a recent study by Kihlstrom, Evans, E.C. Orne, and M.T. Orne (1980), for example, hypnotized Ss received a standard suggestion for amnesia followed by two successive recall tests before administration of the reversibility cue. Four types of special instructions concerning the manner of recall (demands for honesty, extra effort, and strategic organization, as well as a simple retest) were interpolated between the two amnesia tests. These special instructions had no observable differential impact on the recall of moderately or highly hypnotizable Ss who met a criterion for initial posthypnotic amnesia. In each of the four treatment conditions, however, these same Ss showed a significant overall improvement in memory on the second amnesia test, compared to their previous performance. This improvement was not typically large enough to abolish the initial amnesia entirely, and many Ss maintained a fairly complete amnesia across both tests. Interestingly, there were no differences in general hypnotizability between those Ss who maintained amnesia and those who breached it. As Kihlstrom et al. (1980) noted, their failure to find‘any effect of changing instructional demands on posthypnotic amnesia is problematic, as it amounts to accepting the null hypothesis. Perhaps more powerful demands would have had a discernible effect on these Ss. Failing that, perhaps some individual-difference variable would predict which Ss would breach amnesia and which would not. A series of studies by Coe and his colleagues has addressed both points. Howard and Coe (1980) employed Ss who had proved to be both highly hypnotizable and amnesic on an initial screening session. In a second session, posthypnotic amnesia was tested in one of three contexts: while Ss were relaxed, after receiving an honesty demand, and after being attached to a “lie detector.” The two experimental conditions did lead to an increase in recall compared to control Ss, but only for those Ss who reported that their remembering was under voluntary control. Those Ss who reported that they had no control over remembering the material showed no effect of the manipulation. This differential effect was replicated by Schuyler and Coe (1981), comparing only the “lie detector” and relaxation conditions. Apparently, individual differences in perceived involuntariness of response successfully predict who will respond to contextual changes by breaching their amnesia. This very reasonable solution is clouded somewhat, however, D ow nl oa de d by [ 71 .1 98 .1 84 .1 63 ] at 0 6: 03 2 4 M ay 2 01 6 RECOVERY OF MEMORY DURING AMNESIA 311 by a recent study by Ham, Radtke, and Spanos‘. They replicated Kihlstrom et al. (1980) in finding no overall effect of an honesty demand compared to a retest control, but they failed to replicate the findings of Coe and his colleagues (Howard & Coe, 1980; Schuyler & Coe, 1981) with respect to the importance of perceived involuntariness. In the study of Kihlstrom et al. (1980), a significant improvement over the initial memory test was observed even in those hypnotizable, amnesic Ss who received a simple retest, without any change in instructional demands. Whatever the effects of contextual shifts may prove to be, this apparently spontaneous recovery of memory is of some interest because it suggests that posthypnotic amnesia may be somewhat fragile and transitory. From a theoretical point of view, the spontaneous recovery from amnesia observed in these Ss may be a function of two processes: a remission or decay of amnesia due to the passage of time, or a reminiscence effect stemming from the successful retrieval of some memories on the initial test of amnesia. Within the domain of memory, analogous phenomena are found in cases of posttraumatic retrograde amnesia (Whitty & Zangwill, 1977; Williams, 1977), which shows appreciable “shrinkage” over time. Somewhat more prosaic examples of spontaneous recovery of forgotten memories have long been studied under controlled conditions in the psychological laboratory. It has long been recognized that a single test of recall may not fairly represent all the items that are available in S’s memory (Brown, 1923). When Ss are allowed repeated recall attempts, as in Tulving’s (1967) Recall-Test-Test (RTT) paradigm, they will often remember on later trials items that appeared to be forgotten on earlier ones. In many conventional verbal-learning experiments, such recovered items are traded for others that are recalled on earlier trials but forgotten on later ones, such that overall levels of recall remain fairly stable. Under other circumstances, however, intertrial recovery can exceed intertrial forgetting, so that there occurs a net increment in recall over time, without any further study of the items themselves. For example, Ballard (1913) observed that children who had partially memorized stanzas of poetry improved their recall of the material over a period of days, even though they were given no opportunity to study the material again. Interest in the phenomenon has been revived by a series of investigations by Erdelyi and his colleagues (e.g., Erdelyi & Kleinbard, 1978). The present experiment was intended to explore the spontaneous recovery of memory during posthypnotic amnesia observed by Kihlstrom ‘Ham, M.L., Radtke, H.L., & Spanos, N.P. The effects of suggestion type and the experience of involuntariness on the breaching of posthypnotic amnesia. Manuscript in preparation, 1983. D ow nl oa de d by [ 71 .1 98 .1 84 .1 63 ] at 0 6: 03 2 4 M ay 2 01 6 312 KIHLSTROM ET AL. et al. (1980). Aside from replicating the earlier findings, it sought to reexamine the contribution to the effect of individual differences in hypnotizability and response to the amnesia suggestion. Another condition was included in the experiment as well, in order to evaluate two competing accounts of the recovery effect: simple remission of the amnesia with the passage of time, or reminiscence due to the administration of the initial recall test. In this condition, a distracting activity was substituted for the test of initial amnesia. This variation controlled for two major factors in reminiscence by eliminating the initial test of amnesia and preventing the period that would be consumed by that activity from effectively extending the recall period assigned to the later test.