The "New Look" in perception, particularly the perceptual defense-vigilance effect, is reformulated in information-processing terms following an evaluation of the major criticisms that have been leveled against it. These criticisms, including the logical paradox, homunculus, frequency, expectancy-set, and various response bias positions, are shown to be conceptually flawed or empirically incomplete. The reformulation of the phenomenon conceives of the perceptual defense-vigilance effect as a special instance of selectivity in cognitive processing. It is specifically argued that selectivity-and therefore perceptual defense and perceptual vigilance-is a multiprocess complex of phenomena under central regulative control. Selectivity-and consequently defensive selectivity-is brought into play through varied mechanisms at multiple loci of the information-processing sequence. Thus, selectivity is pervasive throughout the cognitive continuum, from input to output, and no single site is likely to provide exhaustive explanation of any substantial selective phenomenon."The verdict is not suddenly arrived at, & McGinnies, 1948) which suggested that the proceedings only gradually merge into the perception of external stimuli is not free the verdict [Franz Kafka]."°f the shackles of internal events: attitudes, values, expectancies, needs, and psycho-The proceedings, at least formally, started dynamic defenses all impinge upon percepsome quarter century ago with a series of t ion. This view became loosely known as publications (Bruner & Postman, 1947a, the "New Look." 1947bMcGinnies, 1949; As with most things new, this was not an .altogether novel departure. Past work, thei The author is very much indebted to a number °re , tical aS Wel1 aS empirical, had anticipated of individuals for their incisive reading of an and even preceded the movement. Howearlier draft of this paper. Special thanks are due ever, the New Look, as a concerted experito Donald Broadbent, Ralph Haber, Richard Lore, mental effort, launched a frontal assault upon Ulric Neisser John Santa, and Peter Suedfeld for ^ interface bet ween psychodynamiCS and their truly helpful suggestions, criticisms, and . , . ... .. ,. . encouragement. It goes without saying that their psychophysics, delineating the issue as a help need not imply agreement with any or all coherent subject area in psychology, substantive points developed in this paper.After two and one-half decades and prob-«Requests for reprints should be sent to aM more than j QQQ research publications,

Repression has become an empirical fact that is at once obvious and problematic. Fragmented clinical and laboratory traditions and disputed terminology have resulted in a Babel of misunderstandings in which false distinctions are imposed (e.g., between repression and suppression) and necessary distinctions not drawn (e.g., between the mechanism and the use to which it is put, defense being just one). "Repression" was introduced by Herbart to designate the (nondefensive) inhibition of ideas by other ideas in their struggle for consciousness. Freud adapted repression to the defensive inhibition of "unbearable" mental contents. Substantial experimental literatures on attentional biases, thought avoidance, interference, and intentional forgetting exist, the oldest prototype being the work of Ebbinghaus, who showed that intentional avoidance of memories results in their progressive forgetting over time. It has now become clear, as clinicians had claimed, that the inaccessible materials are often available and emerge indirectly (e.g., procedurally, implicitly). It is also now established that the Ebbinghaus retention function can be partly reversed, with resulting increases of conscious memory over time (hypermnesia). Freud's clinical experience revealed early on that exclusion from consciousness was effected not just by simple repression (inhibition) but also by a variety of distorting techniques, some deployed to degrade latent contents (denial), all eventually subsumed under the rubric of defense mechanisms ("repression in the widest sense"). Freudian and Bartlettian distortions are essentially the same, even in name, except for motive (cognitive vs. emotional), and experimentally induced false memories and other "memory illusions" are laboratory analogs of self-induced distortions.

Experimental research on memory since Ebbinghaus has predominantly focused on the fact of forgetting, creating the impression that memory inevitably decreases with time or time-correlated interpolated events. Recent laboratory work on the recall of pictures, however, has produced a contrary pattern, suggesting that memory for certain classes of stimuli may be hypermnesic rather than amnesic, increasing over time and recall attempts. Up to now, such hypermnesias have been obtained in single, 1-hr laboratory sessions. The purpose of the present studies was to determine the magnitude of memory growth over more significant time intervals. Tests of memory up to 1 week indicated substantial growth of recall for pictures, but not usually for words. The outcomes are discussed in terms of (a) their bearing on the Ebbinghaus experimental tradition, (b) the relation of this work to other hypermnesia literature, including Ballard's reminiscence, hypnotic hypermnesia, memory recoveries in therapy, and the Penfield effect, and (c) the implications of hypermnesia for psychodynamics and unconscious processes.In a recent series of experiments, we have Instead of the classic Ebbinghausian funcobtained an unusual memory phenomenon, tion in which memory decreases with time,we have been able to produce the opposite of forgetting-hypermnesia instead of am-We wish to thank the six anonymous subjects nesia--such that recall progressively in-Zftr^orftr 2 7 tsTStLtimt -eases with^time over successive recall atnights) of testing. We also thank John Santa tempts (Erdelyi & Becker, 1974; Erdelyi, for reading and commenting on an earlier draft Finkelstein, Herrell, Miller, & Thomas, of this article, and Jaime Perez for his assistance 1976; Shapiro & Erdelyi, 1974). When the "T'hrrtir* 6 S 1 stated by Public Health *imuli * * remembered are pictures (but Service/National Institute of Mental Health Grant not, apparently, when they are words) the 1 RP3 MH2S876-01 and by the City University effect is both highly reliable and powerful, of New York Faculty Research Award Program llr .,, ,• , . .« " a, "{Grants RF-10261 and FR-10704 to the senior Wlth °Ptlmal procedures, the growth of author. recall from initial test may average as high Requests for reprints should be sent to Matt as 15% to 20% in less than 1 hr. Since our

No abstract