JOURNAL OF VERBAL LEARNING AND VERBAL BEHAVIOR 17, 721--743 (1978)

Richard Semon's Theory of Memory

DANIEL L. SCHACTER, JAMES ERIC EICH, AND ENDEL TULVlNG

University of Toronto

In the first decade of the 20th century, Richard Semon put forward a theory of memory that anticipated numerous recent developments in memory research. The theory is discussed both in its historical context and with reference to modem ideas. Semon's theoretical concern for retrieval phenomena is particularly noteworthy. Several reasons are suggested why the theory is virtually unknown today.

Current research in the area of human Wolfgang Semon, a German scientist who memory owes many of its present orienting wrote two books on the subject of human attitudes and research techniques to pioneer- memory in the first decade of the 20th century. ing psychologists of the late 19th and early In fact, Semon's work has been almost com- 20th centuries. Among the most well known pletely ignored by mainstream psychologists and important of these early investigators are concerned with processes of memory and Hermann Ebbinghaus, who performed the learning; later in this paper we will explore first systematic laboratory studies of human some of the reasons why Semon's work has memory; William James, whose distinction been bypassed. Yet prominent anatomist J. Z. between primary and secondary memory is Young (1965, p.288) commented that still today the target of much research and "...many modern ideas on the subject [of theorizing~ G. E. Miiller, who performed memoryl go back to Richard Semon...;" important early research on grouping and Gestalt psychologist Kurt Koffka (1935, interference; and Sir Frederic Bartlett, whose p. 598), though refraining from discussing reconstructive approach to memory has be- Semon's theory in detail, noted that "... this come influential in recent years. It is likely that omission is not due to a lack of appreciation of most modern students of memory are familiar Semon's great achievement;" Nobel Prize with the writings of the above psychologists, winning physicist Erwin Schr6dinger (1964, and have probably been influenced, to varying p. 44) regretted that a physiological model of degrees, by their research and theories. It is Semon's theory of memory had not yet been much less likely that these same students of developed "... important though it would be memory are familiar with the work o f Richard for the advancement of our knowledge;" and This paper was written while Daniel Schacter and Bertrand Russell who, in a chapter in The Endel Tulving were visitors at the Department of Analysis of Mind, introduced Semon's work Experimental Psychology, University of Oxford. The to English readers, flatly stated that, "The best work was supported by a National Science Foundation writer on mnemic phenomena known to me is Predoctoral Fellowship to D. L. Schacter, University of Toronto Predoctoral Fellowship to J. E. Eich, and by the Richard Semon..." (1921, p.83). National Research Council of Canada, Grant No. A8632 What in Semon's work elicited the ac- to E. Tulving. We thank Margaret Boden, Douglas colades of the distinguished scientist- Hintzman, Gregory Jones, and George Mandler for philosophers mentioned above? It is the helpful comments on the manuscript. Requests for re- prints should be sent to Daniel L. Schacter, Department purpose of the present paper to explore in of Psychology, University of Toronto Toronto, Canada, some detail Richard Semon's analysis of M5S 1A1. human memory, place this analysis in its 721 (1022 537t/78/0176-0721502.00/0 Copyright !: 1978 by Academic Press, lnc, All rights of reproduction in m!y form reserved. Printed in Great Britain 722 SCHACTER, EICH, AND TULVING historical context, and elucidate some of the the thesis that the mechanisms of memory and reasons why Semon's work is virtually un- heredity are identical. As we shall argue later, known to present-day students of human advocacy of this thesis, that also led Semon to memory. It is our thesis that Semon's analysis support the highly controversial Lamarckian anticipated many current research problems doctrine of the inheritance of acquired charac- and approaches to the study of memory in a teristics, proved to be a key scientific error. At most striking fashion, and that his work Easter 1918, Semon's wife succumbed to a contains potentially valuable suggestions and long illness. Severly disturbed by this loss, implications for contemporary researchers. shattered by the collapse of his native Germany in World War 1, and deeply dis- turbed about the lack of recognition of his THE BACKGROUND work, Richard Semon ended his own life on Richard Semon was born in Berlin on December 27, 1918. August 22, 1859. His father Simon was a Before we move on to a consideration of stockbroker; his older brother Felix became a Semon's work and its historical context, it will prominent laryngologist in England, received be useful to clarify two points. First is the a knighthood in 1897, and was appointed problem of terminology. Semon believed physician to King Edward VII in 1901. Semon strongly that everyday terms commonly used was awarded his Dr. Phil. for zoological work to talk about memory had too many undesir- at Jena in 1883, and earned his Dr. Med. in able connotations to be of precise scientific 1886. During this period, Semon studied with value. Accordingly, he invented his own terms some of the most prominent scientists of the to correspond more exactly to his intended day, including the noted biologist Ernst meanings. One of Semon's terminological Haeckel; Haeckel's emphasis on the theoreti- creations, the word "engram" [defined by cal unification of diverse biological pheno- Semon as "... the enduring though primarily mena had a particularly strong influence on latent modification in the irritable substance Semon. After receiving an associate professor- produced by a stimulus..." (1921, p. 12)] has ship at Jena in 1891, Semon led a successful persisted in present-day usage, and is prob- biological expedition to Australia from 1891- ably most closely associated with the famous 1893 (Semon, 1899). He left Jena in 1897 for paper of Lashley (1960). Another of Semon's personal reasons, and established himself as a creations, the term "ecphory" [defined by private scholar in Munich. Semon as "... the influences which awaken the It was during this period that Semon pub- mnemic trace or engram out of its latent state lished his two books on memory: Die Mneme into one of manifested activity..." (1921, (1904) (translated into English as The Mneme p. 12)] has been used quite infrequently. Since in 1921) and Die mnemischen Empfindungen "engram" is roughly equivalent to the phrase (1909) (translated as Mnemic Psychology in "memory trace" and "ecphory" is roughly 1923). 1 Mnemic Psychology is devoted com- equivalent to "retrieval" or "recall," we will pletely to the analysis of human memory. use these terms interchangeably throughout However, in The Mneme, Semon examined this paper. 2 We will, of course, define all new not only the phemomena of human memory terms as they arise. but also advanced and attempted to support Second, when we speak of the "historical context" of Semon's theories, we refer specifi- 1 We will cite the English versions of these two books cally to the period from 1885-1935. Con- throughout this paper. Although we will quote most extensively from Mnemic Psychology, the reader should 2 It is possible to distinguish between ecphory and bear in mind the most of Semon's major ideas concerning retrieval, as noted in Tulving (1976). We will not concern human memory date to 1904. ourselves with the distinction in the present paper. RICHARD SEMON'S THEORY OF MEMORY 723 sideration of memory research during this time of storage need be present at the time of period will enable us to describe the theoreti- recall in order for retrieval of the original cal concerns of the field in both the years event in its entirety to occur. This view of preceding and following publication of retrieval (one of the very few such views that Semon's work. We chose the year 1885 to had been explicitly formulated in Semon's initiate our historical consideration because time) was further developed and utilized by this was the year in which Ebbinghaus pub- Semon in analyses of problems such as as- lished the first experimental studies of sociation by contiguity vs association by memory; we terminate our historical survey in similarity and the temporal organization of 1935 because in subsequent years interest in memory, and led Semon to formulate novel the general problem of memory gave way to positions concerning matters such as the the more restricted concerns of the verbal active role of ecphory in establishing new learning tradition before surfacing again in the engram-complexes and the role that ecphory early 1960s. plays in the storage of new engram-complexes. Also, by allowing for the representation of internal or "energetic" stimuli in engram- PRELIMINARY OVERVIEW OF THE THEORY complexes, Semon was able to offer sur- Semon's theory of memory was based upon prisingly modern statements concerning two fundamental postulates, which the author phenomena such as state-dependent retrieval. termed the "Law of Engraphy" and the "Law We will explore these points in greater detail of Ecphory." The first law was Semon's shortly. characterization of memory storage: "All A third notion that is part of the kernel of simultaneous excitations...within our organ- Semon's theory (although it was not granted isms form a connected simultaneous complex the status of a "law" by Semon) is the concept of excitations which, as such, acts engraphi- of homophony. Homophony may be most cally, that is to say leaves behind it a connected simply viewed as a resonance metaphor; and, to that extent, unified engram-complex" Semon used it to describe the mechanism by (1923, pp. 159-160). There are several points which information from different sources is contained in this law that subsequently combined, defining it as "...the concordant emerge as critical features of Semon's theory. action of closely allied mnemic and original First there is Semon's emphasis on the unitary, excitations, a consonance which I have found wholistic nature of engram complexes that he it convenient to call Homophony" (1921, p. 13). later applies to the analysis of various Homophony can exist between two "original mnemonic phenomena. Second there is the sensations," between "original and mnemic notion that each event, or corresponding sensations," or between two "mnemic sen- "simultaneous excitation-complex" leaves sations." This resonance principle was in- behind a separate engram-complex; this idea voked by Semon in constructing what we is elaborated upon and utilized in Semon's might want to call "retrieval explanations" of analyses of repetition effecfs and recognition. repetition effects and problems of recognition; The law of engraphy also sets the stage for he also applied it to various problems of Semon's law of ecphory, which represents his perception that will not concern us here. The view of memory retrieval: "The partial return major point that we wish to extract from this of an energetic situation which has fixed itself highly condensed overview of Semon's posi- engraphically acts in an ecphoric sense upon a tion, and which we will document more fully simultaneous engram-complex" (1923, p. 180). later in the paper, is that the analysis of Thus Semon's view of retrieval is redinte- retrieval was one of Semon's principal grative. Only part of the total situation at the theoretical concerns. More specifically, we will 724 SCHACTER, EICH, AND TULVING argue that Semon's focus upon the conditions, parts covered theories of memory from the functions, and processes of retrieval 3 was one Greeks into the 19th century. Some of the of the few systematic attempts to elucidate the important, recurring issues during this period role of retrieval in memory during the period of time were whether memory is physical, under consideration, that his ideas about psychical, or "of the soul;" the relative import- retrieval anticipated much modern research, ance of different kinds of assocation; the role of and that his emphasis on retrieval phenomena habit in memory; and so on. More recent at a time when few were interested in this topics of interest, discussed by Burnham in the problem may well have contributed to his latter two parts of his paper, include the subsequent obscurity. physiological basis of association, the effects of retention interval on forgetting, the usefulness of memory span as an indicator of educability MEMORY RESEARCH BETWEEN 1885 AND 1935: in normal and retarded children, and the A BRIEF REVIEW nature of "memory illusions." Of course, In this section we sample the memory Ebbinghaus' pioneering work had just been literature between 1885 and 1935 in order to published. This initiated serious experimental convey a general idea of the problems and interest in the study of memory and brought theoretical issues that concerned memory the study of repetition effects and the quantifi- researchers of the time. We will consider in cation of forgetting curves to the fore of the somewhat more detail research that examined field. Interest in problems of retrieval, how- specific problems of interest to Semon in the ever, was negligible in the years covered by next section of the paper. Burnham's review. With the exception of Sir One useful indicator of research concerns in :-William Hamilton (1859), who expiicitly a given period of time is a listing of the topics divided memory into three stages of acqui- considered in major review papers of the area. sition, retention, and reproduction, and Accordingly, we have examined the topics offered an early redintegrative theory of covered in major reviews of memory pub- retrieval, and some passages from William lished in the American Journal of Psychology, James (1890), these early writers had little Psychological Bulletin, and Psychological direct interest in either the functions, condi- Review during the late 19th and early 20th tions, or processes of retrieval. centuries. First we will examine papers pub- Kennedy's (1898) review reflects the lished before and during the time that Semon emerging concerns of the young experimental wrote his two books; then we will examine science of memory. After discussing methods papers published after this time and up until and materials used in memory research, 1935. Kennedy outlined the problems of interest to The first major review of theory and re- contemporary researchers: measuren~ent of search on memory was Burnham's (1888- the depth of "initial impressions," the role of 1889) classic four-part article. The first two attention and repetition, the qualitative change in the memory "image" over time, individual differences in memory ability, and the nature of the to-be-remembered material 3 The phrase processes of retrieval refers to the mech- were all major issues of the day; and the initial anisms by which retrieval is carried out" the phrase investigation of grouping and organization conditions of retrieval refers to those properties of the had just been reported by M/iller and cognitive environment of the rememberer which affect Schumann (1894). In Kuhlmann's (1908) the retrieval process: and the phrase functions of retrieval refers to the effects that the act of retrieval has upon the review (which appeared just after Thf Mneme subsequent state of the memory system. and just before Mnemic Psychology) we RICHARD SEMON'S THEORY OF MEMORY 725 observe continued concern with many of the Miiller and Pilzecker (19001, are now well topics reviewed by Kennedy, and observe under way; the problem of affective tone and heightened interest in problems that had been retention, owing largely to Freudian in- barely touched at the time of Kennedy's fluences, has become a major research topic: review. For instance, the problem of massed vs and the problem of reminiscence, first distributed practice, first attacked by addressed in Ballard's (1913) monograph, has Ebbinghaus, was now a major research con- been given experimental attention. We also cern; the introspective analysis of the find reference to studies comparing the effect "memory consciousness," concerning the form of recalls vs extra study presentations on re- and content of memory imagery and indi- tention: that is, studies exploring the functions vidual differences in this imagery, had of retrieval. The third section of Robinson's "...come to the foreground of memory in- review, entitled "Recall," indicates just how vestigation" (p.285); and the analysis of sparse research in this area was. A few studies recognition was becoming a full-fledged ex- that examined various aspects of legal testi- perimental problem. But again there was a mony are cited, and one study (Laird, remarkable absence of attention paid to Remmers, & Peterson, 1923) that was con- problems of retrieval. In the years before and cerned with retrieval conditions is mentioned; during which Semon published his two but the problem of recall is still clearly of volumes on human memory it is extremely minor experimental and theoretical interest at difficult to find any theory or research that the time of Robinson's review. The remaining raises questions specifically directed at the sections of his review cover qualitative studies conditions, functions, or processes of retrieval. of memory, and memory in the "insane and Association was assumed to be the mechanism defective." of the retrieval process, an assumption the The final reviews we will consider were Semon criticized elegantly, and that we will published by McGeoch (1928, 1930). The examine shortly; there was interest in retrieval topics reviewed in these papers, as well as conditions only insofar as the laws of associ- many of the studies we have already made ation required some specification of effective reference to, are brought together in associative stimuli; and the possible mne- McGeoch's (1933) massive, 1200-item biblio- monic functions of the act of retrieval itself graphy of learning and memory research. were simply not considered by theorists of the These reviews reflect heightened interest in time. (For a more detailed account of memory serial postion effects and problems of transfer; research during this period, see the paper by a new concern with the establishment of Murray, 1976). general laws of learning; and continued in- We now move ahead to the 1920s and terest in traditional problems such as quantifi- consider the lengthy review paper by cation of the curve of retention, the most Robinson (1924). Robinson's paper was economical methods of practice, the role of divided into five major sections. The first sensory modality in memory, etc. As regards section, on "memorizing," considers many of analyses of retrieval, several studies of the the topics previously mentioned, as well as the similarity of "stimulating conditions" (i.e., role of intention in learning and studies of context) at storage and retrieval are con- associative inhibition, for example, the sidered, as are several studies of successive Ranschburg effect. The second section, con- recalls; but the largest number of studies cerned with retention, reviews several topics concerned with recall are found under the that had not previously been given much heading of "Relationships between different attention. Studies of retroactive interference, measures of retention" (McGeoch, 1933, p. 57). which originated in the classical work of These studies, best exemplified by Luh's (1922) 726 SCHACTER, EICH, AND TULVING work, reflect the commoh attitude of investi- search interests between 1885 and 1935, has gators at the time towards the issue of been to demonstrate the relative lack of retrieval: It is viewed as a problem of theoretical concern for the problem of measurement, rather than as a psychological retrieval throughout this period. process that forms a crucial part of the memory system and that requires systematic SEMON'S THEORY IN theoretical analysis. Serious theoretical con- HISTORICAL PERSPECTIVE: sideration of retrieval and its role in the A DETAILED ANALYSIS memory system was to emerge shortly in the work of Bartlett (1932), K6hler (1930), and We now proceed to a more fine-grained Koffka (1935), before disappearing for the next exposition of Semon's theory. As we consider 30 years; but with few exceptions (such as the various aspects of Semon's theoretical Hollingworth, 1926, 1928; Meumann, 1913; postion, we will relate his formulations to the Selz, 1913, 1922; and to some extent, Myers & theories of his contemporaries, and will also Myers, 1916) the problem was treated in an juxtapose Semon's views with those advanced almost uniformly atheoretical fashion in the by modern students of human memory. years preceding and following Semon's work. Additional evidence on this point can be Biological Perspective adduced by sampling general theoretical We initiate the discussion by considering statements made about memory in the period the substrate of engraphy: how did Semon under consideration. For instance, there was conceptualize the representation of engrams? William James' (1890, p. 653) view that "... the It may be advisable first to remind the reader cause both of retention and of recollection is the that, in the early part of the 20th century, the law of habit in the nervous system, working as it question of whether memories are represented does in the 'association of ideas'." Kennedy, in in the brain (that is, physically) or in the mind his 1898 review, reduced memory to two (that is, psychically) was a hotly debated issue. factors: "... the conditions which govern the Eminent thinkers such as Bergson (1911) and chances that a certain object be remembered McDougall (1911) argued at great length that depend, first, upon the depth and clearness of memories are not physically represented in the the impression which that object made on me brain but rather in some nonphysical, in my experience of it, and second, upon the "psychical" form. Semon's view on the matter transformation which my image undergoes in reflected his biological training; he unequivo- the temporal flow" (p. 485). Thorndike's (1913) cally took the postion that engrams are stored proclamation provides a typical textbook via physiochemical processes in the brain. view of memory around Semon's time: However, Semon declined to hypothesize "Goodness of memory depends upon the about the precise form of this biological permanence of impressions, the permanence storage, arguing that in the limited state of of connections, their number and their nature then contemporary physiology, such specu- or arrangement" (p. 25). The neglect of lation was unwarranted. Semon did go on to retrieval factors in the above formulations is raise the question of whether memory storage clear, and these statements are quite typical of is neurologically distributed or localized, and memory theory at the time. We will refer to offered a surprisingly modern view (see, for some of the individual experiments and ideas example, Luria, 1973): that were directed at retrieval phenomena in We seem ...to be placed in the dilemma of having the next section of the paper. The major either to reject altogether a localization theory purpose of the present section, in addition to which imagines that each single engram can be providing a general characterization of re- stored up in a single cell--or in a comparatively RICHARD SEMON'S THEORY OF MEMORY 727

small complex of cerebral cells.., or to admit that Relation of Perception to Memory in the human organism a special interdependence Returning to Semon's view of engraphy, an exists between definite regions of the cerebral important point to note is that Semon's cortex and the ecphory or, as perhaps we ought to say, the possibility of the ecphory of distinct analysis of memory was directly related to his individually-acquired engrams. The latter admis- analysis of perception: sion, implies, however, the recognition of a certain localization, although it need not be the kind •.. the very expression "mnemicsensation" necess- which makes each nerve-cell of the brain a arily implies that such a sensation has been repository for a specific engram (1921, p. 120). preceded by an original o'ne. The nature of this It is interesting to note the manner of dependence will be fully explained later. The bare criticism directed toward these views. On the fact of its existence, however, makes it a pre- condition for the study of mnemic sensations that one hand, Semon was grossly misinterpreted we should closely follow the orientation of certain and accused of one of the worst scientific sins aspects of original sensations, because the former of the day: the advocacy of vitalism. Since he depend on the latter as inevitable predecessors did not construct a model which specified the (1923, pp. 69-70)... precise nature of biological memory storage, and because he did not adhere to a strict Indeed, the first part of Mnemic Psychology localizationist view, various critics (e.g., is devoted to an analysis of sensation and Kostyleff, 1911) attached the vitalistic label to perception, upon which Semon's analysis of Semon's position. To these critics, Semon gave memory is based. In Semon's time, the an incisive reply which merits close attention analysis of perception was largely divorced even today: from the analysis of memory, and with few I sl,~ould be as able as anyone else to turn out some exceptions (e.g., K~lpe, 1895), explicit sort of schematic representation on the model of theoretical realization of the interrelatedness the diagram of Mendelian determinants in which of perception and memory was not manifested engrams would be naively represented, schema- tized as tiny particles and conveniently packed until the appearance of the work of Bartlett together. This would meet the views of those (1932), Gibson (1929), and the Gestaltists readers whose thirst for causality requires such (K6hler, 1930; Koffka, 1935). The inter- schematic representation, and who cannot resign relatedness of perception and memory is, of themselves to leaving such questions open for the time being. My own conception of inductive course, a major theme in contemporary re- science is a different one,-and I attribute more search, forming, an important part of value to an honest note of interrogation than to numerous theories. constructions which are only representable In order to understand the relation between through an effort of imagination (1923, p. 329). perception and memory in Semon's theory, we On the other hand, avowedly vitalistic must first briefly consider his analysis of writers such as Bousefield (1928~-who ad- perception. In contrast to atomistic concep- vanced the theory that memories are not phy- tualizations of sensation and perception that sically represented in "protoplasm" but rather were dominant in experimental psychology are psychically represented in "psycho- around the turn of the century (e.g., Kfilpe, plasm"--accused Semon of erring by revert- 1895; Titchener, 1911; Wundt, 1902), Semon ing to crass mechanistic reductionism. Thus, stressed the unity and wholistic nature of Semon was criticized by both mechanists and "sensation-complexes': "...what we experi- vitalists for espousing general hypotheses con- ence immediately are not single sensations cerning the biological nature of memory but connected complexes of sensation, form- storage that today seem both perfectly ing at any given moment the whole content of reasonable and appropriate to the state of consciousness" (1923, p. 65). Semon preferred physiological knowledge at the time. to speak of "fields of sensation" rather than 728 SCHACTER, EICH, AND TULVING

"sensation-elements," emphasizing the recording of perceptual experience; rather "primary unity" of Such fields, and he ascribed "...nearly every complex of original sen- an important role to the "reciprocal influence" sations has grouped around it numerous of the components uf a field of sensation. In so mnemic sensations which are evoked by it and doing Semon anticipated Gestalt analyses of work engraphically in the grouping...every perception and the context and contrast simultaneous complex of sensations is com- effects that the Gestalt theory was based upon. posed of original and mnemic sensations Similar views concerning "reciprocal in- which are closely connected with one another, fluence" can be found in H6ffding (1891, and thus form a whole; and this whole-- p. 114), and other early wholistic, Gestalt-like regarded from its energetic side--works en- analyses of perception were offered by Mach graphically" (1923, p. 168). (1959); but these were exceptions to the pre- Semon also considered the problem of why vailing atomistic views (Boring, 1942). only fragments of this whole (however dif- Semon's conceptualization of the engram ferent from a literal "snapshot" it might be) follows directly from his wholistic analysis of can be remembered. Semon argued that these perception. The memory trace is to be re- memory fragments are not isolated links in a garded as a unified complex, reflecting the semi-intact associative chain; rather, reflecting unitary nature of perceptual experience. The his multicomponent orientation, he suggested essence of this idea is captured in Semon's law that "... a conception much more in accord- of engraphy. However, Semon stressed that ance with otir meaning leads us to regard such the engram is comprised of "emergent com- fragments not as associated, but as integral ponents" that could to some extent be components, as emergent points of a connected dissociated from each other. This view, similar simultaneous complex of sensations" (1923, to recent multicomponent theories of the p. 164). Semon's concern with memory frag- memory trace (Bower, 1967; Underwood, ments foreshadowed the recent systematic 1969), was elaborated further by Semon and work of Jones (1976), whose "fragmentation will be discussed shortly. At this point the hypothesis" of memory holds that memories reader may rightly ask if all Semon has done is are stored as fragments of a perceived to offer another "literal-copy" theory of situation. Semon offered three explanations of memory stated in Gestalt-like terms: engram- fragmentation, two of these reflecting his view complexes are simply faithful recordings of that fragmentation arises at retrieval. First, perceptual experience, and when we remem- Semon argued for the importance of retrieval ber we have access to this stored "snapshot" of conditions in fragmentation. The direction of the world. It turns out that Semon was saying attention at the time of ecphory influences no such thing; he in fact directly addressed the which fragments of the trace are "noticed:" question of why the output from the memory "The fixing of attention on specific points in system is so different from the input, and his the simultaneous complex acts as a dissolvent answers to this question form some of the and dissociates these parts from the rest of the most interesting parts of his theory. connection" (1923, p. 165). Second, Semon First, Semon argued that distortion is intro- suggested that homophony, or resonance, duced into the memory trace at the time of between the ecphoric stimulus and certain storage. This is because every act of storage, in components of the retrieved engram-complex Semon's view, involves some retrieval: the new might accentuate some fragments at the ex- input acts as a retrieval cue which operates pense of others. Third, he noted that mnemic ecphorically on the principle of "partial sensations are "less vivid" than original sen- return" described in Semon's second law. sations; hence, only the "peaks of sensation" Hence, what is stored is not just a faithful may emerge during ecphory. RICHARD SEMON'S THEORY OF MEMORY 729

In outlining Semon's position on the input- each new memory trace; however, the mech- output discrepancies, it is clear that in most anisms of this process postulated by Semon respects his approach to the problem was and by Underwood are quite different. Thus, it quite different from that of his contem- seems clear that the thrust of Semon's position poraries. The major theoretical emphasis in on the problem of input-output discrepancies explanations of input-output discrepances at was more in accord with the modern approach the time Semon wrote was on the role of than was the popular theory of his era. unconscious processes in changing, distorting, or weakening the memory trace over time (Kennedy, 1898); this notion also played a 7he Acoluthic Phase and Temporal large role in subsequent Gestalt analyses of Organization memory (c.f., Koffka, 1935). Semon's notion Closely related to the perception-memory that the fading of mnemic sensations contri- issue is Semon's distinction between the butes to fragmentation is similar to these synchronous and acoluthic phases of sensation. traditional views. Exceptions to the prevailing The synchronous phase lasts only as long as a hypothesis are found in the work of Bentley physical stimulus is present, whereas the (1899) and Kuhlmann (1906), who attached acoluthic phase persists for some time after the importance to associations formed at storage cessation of the stimulus. Semon distinguished as major determinants of input-output dis- two components of the acoluthic phase: a crepancies. This position is in some ways period of short-lived oscillating activity (less similar to Semon's point that each new than a second) which "...manifests itself engram-complex is comprised of "original" regularly in sensations above the threshold of and "mnemic" sensations, although Semon's consciousness..." (1923, p. 140); and a longer conception of this process is quite different lasting activity of unspecified duration that is from the mechanisms envisaged by Bentley or not always manifested in consciousness. At Kuhlmann. It was not until the work of first glance, it is tempting to say that Semon Crosland (1921) and the later monograph by was simply talking about after-images that Bartlett (1932) that serious consideration of were the subject of some research during this retrieval conditions as causes of input-output time (Boring. 1942). However, closer con- discrepancies is found. These authors placed sideration of Semon's treatment of the major emphasis on the subjects' "attitude" at acoluthic phase reveals that he assigned it an the time of recall in developing accounts of important functional role in the overall distorted and fragmentary remembering. In memory process. As we shall see shortly, modern studies of input-output discrepancies, Semon's concern for the mnemonic functions the "unconscious transformation" hypothesis of the acoluthic phase sharply distinguishes so popular in Semon's time has been aban- this conception from that of the static after- doned, and research has been directed at image that was not studied as a functional processes of generalization, fragmentation, component of the memory system during this and abstraction occurring at both storage and time. retrieval (Bransford & Franks, 1971; Before describing the functional use to Frederiksen, 1975; Jones, 1976; Loftus, Miller, which Semon put the acoluthic phase, we & Burns, 1978). Additionally, the notion of must first outline his conception of the tem- "implicit associative responses" advanced by poral arrangement of the memory store, for it Underwood (1965) to account for false recog- is in his analysis of temporal relations that nition data resembles Semon's idea that both Semon makes use of the acoluthic phase. incoming stimulus information and informa- Semon hypothesized that the engram-store is tion from the memory store are represented in organized primarily along temporal dimen- 730 SCHACTER, EICH, AND TULVING sions, and that engram-complexes are view this process as one of "horizontal" as- deposited in "chronological strata:" sociation between successive events. Rather, the simultaneous conjoining of synchronous Every simaltaneous complex which may be figur- and acoluthic phases establishes a unique atively described as one "layer" of an engram-store is joined to the layer immediately preceding it and, engram-complex. The nature of this dis- in its turn, bears the same relation to the next most tinction is best illustrated by Semon's ex- recent layer. Owing to the uninterrupted laying planation of Ebbinghaus' finding that remote down of these 'layers' the components of each layer associations exist between nonadjacent are in immediate contact with those of its nearest nonsense syllables in his classic study. Rather predecessor and nearest successor (1'923, p. 327). than positing a direct "horizontal" as- Again it is worth noting that Semon con- sociation between, say, Event X and Event Y, cerned himself with an issue that was not of Semon suggested that X and Y form a new prime importance to his contemporaries. engram-complex by virtue of the simul- Galton (1879), James (1890), and Ribot (1882) taneous occurrence of the synchronous phase did talk about the temporal organization of of Y and the fading acoluthic phase of X. Then, memory, but the problem was clearly not of when there occurs partial return of this major concern in Semon's time, as can be complex (let us say Event X as a retrieval cue) verified by examining the review papers we memory for the whole complex follows. Thus, cited earlier. Semon's interest in temporal in contrast to the traditional associative organization anticipated modern theories account of remote association, Semon offered such as Landauer's (1975), in which time is a novel explanation based on the functional viewed as the principal dimension of organi- role of the acoluthic phase and on his red- zation in memory, and experimental work integrative principle of ecphory through such as that reported by Crovitz and partial return of the conditions of engraphy. Schiffman (1974), Guenther and Linton (1975), This theory led Semon to adopt the view and Underwood (1977), in which various that all association is simultaneous: apparent aspects of temporal factors in memory are cases of successive association arise through explored. Semon's ideas on temporal organi- the simultaneous occurrence of the acoluthic zation also bear a striking resemblance to the phase of Event X and the synchronous phase general theory of temporal organization put of Event Y. Semon's provocative position on forward in Murdock's (1974) conveyer belt this matter led directly to some of the earliest model. experimental work explicity concerned with Having exposed Semon's views on the simultaneous vs successive association in "chronological stratification" of memory, we human memory, reported by Wohlgemuth now return to a consideration of the func- (1915). Although Wohlgemuth claimed that tional role of the acoluthic phase in the genesis his results strongly supported Semon's theory, of temporal organization. Semon proposed inspection of his method and data suggest that memory traces of successive events are extreme caution in interpretation of his temporally linked by the co-occurrence of the results. It is interesting to note, though, that synchronous phase of event N with the this is the only instance we have found in acoluthic phase of event N-1. Thus, the per- which one of Semon's theoretical positions on sisting acoluthic excitation provides the human memory was put to direct experi- "temporal glue" permitting the establishment mental test. There are, clearly, serious defi- of unique engram-complexes comprised of the ciencies in Semon's conceptualization of the synchronous phase of one stimulus and the acoluthic phase that hinder meaningful ex- acoluthic phase of a preceding one. The perimental investigation of it, for example, it is critical point to note is that Semon did not unclear how long the acoluthic phase lasts, RICHARD SEMON'S THEORY OF MEMORY 731 and how one obtains independent evidence of to Sir William Hamilton, who argued for a its existence. Accordingly, it would be fruitless redintegrative position in 1859. Semon did to attempt to evaluate Semon's hypothesis in not cite Hamilton in either The Mneme or the light of the subsequent experimental Mnemic Psychology, and was most likely un- literature (Carr, 1919; Froeberg, 1918: see also aware of Hamilton's redintegrative position. Robinson, 1932). Somewhat similar redintegrative views can be However, there are two senses in which found in H6ffding (1891) and Selz (1913, 1922). Semon's conception of the acoluthic phase The most prolific exponent of redintegration anticipated modern research and theory. in the period just following publication of First, Semon's ideas would fit well with re- Semon's work was Hollingworth (1926, 1928)~ cent experimental demonstations of long- more recently, Horowitz and Prytulak (1969) persisting visual memory traces (e.g., Kroll, have refamiliarized modern students with the Parks, Parkinson, Bieber, & Johnson, 1970). notion of redintegration. Second, the conception of the acoluthic phase Although the basic redintegrative position is in some ways quite similar to modern taken by Semon was similar to the positions of conceptions of short-term memory. The his contemporaries H6ffding, Hollingworth, acoluthic phase constitutes a preliminary and Selz, there are two distinguishing charac- stage of processing which temporally precedes teristics of Semon's approach to redintegra- a more permanent engraphic representation, tion. First, Semon's elaboration of his position and it plays an important functional role in the sounds much like modern theories of retrieval memory system, as does the short-term in which feature overlap between retrieval cue memory envisaged in recent theories (e.g., and memory trace is granted a critical role in Atkinson & Shiffrin, 1968: Baddeley & Hitch, the retrieval process (Kintsch, 1974; Tulving, 1974). Of course, there are numerous ways in 1976): "Resemblance, that is to say, partial which Semon's conception of the acoluthic coincidence between the components of an phase has little in common with modern actual group of excitations and those of any conceptions of short-term memory; but at a previous engram-complex, causes ecphory of time when the major interest in short-term the latter through the former" (Semon, 1923, memory concerned individual differences in p. 326). Although Selz advanced similar memory span (Binet & Henri, 1894: Hawkins, notions, none of the other redintegrationists 1897) and the effects of varying materials on have taken such a position. Second, Semon span length (Kennedy, 1898), the similarities directly applied his law of ecphory to a variety are impressive. of specific problems in the study of memory. Hollingworth did apply his redintegrative The Law of Ecphory principle to various problems in psychology, We have already described Semon's law of especially those involving pathology, but he ecphory, and have outlined his application of did not relate it specifically to memory. it to two problems of memory, namely, input- output discrepancies and simultaneous vs suc- Contiguity vs Similarity and the Engraphic Role cessive association. We now discuss Semon's of Ecphory conception of ecphory in greater detail and One of Semon's most striking and in- describe further applications of this concept to novative applications of redintegrative problems of memory. ecphory to a problem of memory concerned As noted earlier, a key notion embodied in the question of association by contiguity vs the law ofecphory is that of redintegration, the association by similarity. This problem has a reinstatement of a whole via one of its parts. long history in psychology and philosophy, The classical historical reference is, of course, with some authors arguing that all association 732 SCHACTER, EICH, AND TULVING is by similarity, others that all association is think of association as a descriptive by contiguity, and others that both forms of concept which should be logically dis- association occur (see Warren, 1921; tinguished from the mechanism of retrieval. Robinson, 1932). Semon took the position By disentangling these two concepts, Semon that all association is developed through was able to offer a novel analysis of as- contiguity, which in itself was nothing new; sociation by similarity: "... the essential gain but his manner of reaching the conclusion was from our investigation is that the notion of quite interesting, and merits close attention association through likeness was based on a for two reasons. First, it again highlights the confusion of two concepts: association and important role of retrieval processes in ecphory" (p. 189). Semon's analysis is closer in Semon's theory; second, it brings to light an spirit to modern conceptions such as encoding important distinction that was largely over- specificity (Tulving & Thomson, 1973) than to looked at the time. the theories of his time. In order to account for apparent associ- The second point that is brought into bold ation by similarity (that is, the case in which relief by this analysis concerns one of the Stimulus X evokes a semantically, visually, important functions of retrieval in Semon's acoustically, etc., similar Memory Trace Y) theory: the establishment of new engram- Semon invoked his principle of partial return. complexes. Semon argued that every act of Owing to shared components between ecphory results in the establisment of a unique Stimulus X and Trace Y, Y is ecphorized in the engram-complex comprised of the retrieved presence of X, just like in any other ecphory information and information in the present via partial return. It is only at this point that context: "...each ecphory of an engram- the two events are associated, through complex produces not only a mnemic contiguity, and the new engram-complex is sensation ...but through this creates a new then stored. Thus Semon argued that associ- engram which adheres to the new engram- ation by similarity is due to: stratum" (1923, p. 178). He utilized this notion in his explanation of association by similarity ... an ecphory based on the partial recurrence of presented above, and also invoked it in his certain components of an excitation-complex. consideration of how engrams from different When departing, it leaves behind a new engram- "chronological strata" are combined. Semon complex in which the two images are associated, suggested that when a particular engram- but this consecutive association is a typical simul- complex has been ecphorized, it can then act taneous association (1923, p. 189). as a cue for engrams in other chronological Semon went on to state that, "In fact 'associ- strata with which it shares common com- ation' through resemblance does not exist. ponents, and hence can be retrieved via parital What is taken for it is ecphory due to the return; this new juxtaposition is then stored partial return of a complex which has pre- as a unique engram-complex. viously left its engram" (1923, p. 189). Semon's conceptualization of ecphory as a There are two critical points to be extracted generator of novel engram-complexes was from this analysis. First, it led Semon to make unique in its time. There was some concern an important distinction between ecphory and with the functions of recall in the years association. Association can be revealed following publication of Semon's two books, through ecphory, but it cannot be equated but little before. Abbott (1909) compared the with the process of ecphory. Here Semon was memorial effects of extra recall time and study challenging the common assumption of his era time, finding that time spent recalling is more that association is the mechanism of recall. beneficial than additional study time; Trow Semon rejected this notion, preferring to (1928) and Raffel (1934) came to similar RICHARD SEMON'S THEORY OF MEMORY 733 conclusions. In a slightly different vein, Internally-Generored Stimuli Bartlett (1928) and Whitely and McGeoch We noted earlier that Semon specifically (1927) investigated the effect of one recall on a allowed for representation of internal stimuli subsequent recall. With these few exceptions, in the formation of engram-complexes, and however, the functional significance of the act suggested that these internal stimuli function of retrieval was rarely treated in a theoretical as potent ecphoric cues. First let us consider manner during Semon's time. Indeed, this the notion that internally-generated stimuli characterization can be accurately applied to are stored as components of an engram- the period extending into the 1960s at which complex. This notion, which forms part of the point serious theoretical concern with the law of engraphy, is one that very few investi- function of the act of recall began to emerge. gators of Semon's time explicitly formulated. This concern has manifested itself in several The notable exception is Hollingworth (1926, ways. The memorial consequences of recall 1928) who in his various expositions of red- trials vs study trials have been pursued in a integration ascribed considerable mnemonic systematic manner (e.g., Izawa, 1969: Tulving, import to internally-generated stimuli. 1967); the facilitating effects of retrieval on Research and theory explicitly concerned with subsequent retrieval have been the object of the representation of internal stimuli in both experimental and theoretical attention memory is still scant in modern times. (e.g., Darley & Murdock, 1971; Modigliani, Anderson and Bower (1973) specifically 1976); and inquiries into the inhibiting effects allowed for the representation of internally- of the act of recall have been made (e.g., generated stimuli in HAM, the mnemonic Roediger, 1974; Rundus, 1973). Thus it would significance of inner or experiential contexts appear that Semon's early interest in the has been suggested by McGeoch (1939) and functions of ecphory anticipated a number of Reiff and Scheerer (1959), and experimental current research trends. It should be noted, investigation of memory for internally- though, that in the above studies, facilitating generated stimuli has recently been reported effects of recall are almost invariably attri- by Dosher and Russo (1976). It isclear, buted to some sort of"strengthening" process: though, that serious research on this topic of The accessibility of an existing memory trace concern to Semon is just getting under way. is increased via the strengthening effect of COnsider next Semon's position on internal recall. Semon, on the other hand, suggested stimulias ecphoric cues: '°... ecphory can arise that what emerges after recall is not a without any recurrence of an original stimulus strengthened version of an already existing through the mere partial return of the inner trace, but rather a new, unique constellation of energetic situation which was present at the information in the present context plus in- formation of the engram-complex" (1923, formation in the retrieved trace. The modern p. 180). It is exceedingly difficult to find any research closest in spirit to Semon's approach experimental or theoretical work bearing on is the recent work on memory for remembered this issue in Semon's time: exceptions are events reported by Gardiner and Klee (1976: found in Hollingworth (1926); and to some Klee & Gardiner, 1976). These authors posit extent, in Ribot~s (1882) book. Also, concern that "Each act of remembering itself ... con- for the relation between internal states and stitutes a new event in episodic memory" (Klee memory was manifested in the clinical work of & Gardiner, p. 471), and provide evidence to Freud (1913), Janet (1928), and Prince (1916); support this hypothesis. Further research in but the emphasis here was on exploiting the this area might well profit from serious con- relation between internal states and sideration of Semon's ideas on the engram- mnemonic processes for therapeutic purposes establishing capacities of the act of recall. rather than on developing a theoretical 734 SCHACTER, EICH, AND TULVING analysis of memory per se that would account At the ecphory of a combination of engrams for the ecphoric efficacy of internal states. ... what is given is not a single indissoluble blend of One particularly interesting manifestation mnemic excitations--"coalescence" some physio- logists call it--but a unisonant chorus in which the of Semon's interest in internal states as single components of an apparently uniform ecphoric stimuli is his anticipation of modern combination of engrams, distinct indeed from each research on state-dependent retrieval (Eich, other as to their time of origin, may be individually 1977). Although one can spot the outlines of discerned (1921, p. 165). this anticipation in the quotation cited above, Semon developed this position far more Semon emphasized that the contributors to explicitly, noting that "... in cases where the homophony (be they original or mnemic energetic condition has greatly changed.., not sensations) run a "side-by-side" course and are even the recurrence of the original stimulus superposed much in the manner of individual suffices for the ecphory of the corresponding transparencies containing different infor- engram" (1921, p. 144). This led Semon to mation that are placed on top of each other. suggest that, "Alcoholic intoxication may, He went on to distinguish two subtypes of under certain circumstances, create an homophony, a "nondifferentiating" homo- energetic condition whose engrams are phony, in which there is a combination of the ecphorable in the next state of intoxication, components, and a "differentiating" homo- but not in the intervening state of sobriety" phony, which is "...always the result of an (1921, p. 144). Semon followed this remark- antagonism between two components or two ably prescient statement with one that even groups of components" (1923, p.248). modern students of state-dependence would Whether a differentiating or non- have difficulty in addressing satisfactorily: differentiating homophony occurs depends "Only in cases where by virtue of the ex- largely on the conditions of ecphory. perience of years the engrams are deeply fixed In order to further clarify Semon's con- and frequently ecphorized may we expect ception of homophony, let us briefly consider ecphory independent of abnormal or con- his use of the idea in an analysis of generic vs trasting conditions" (p. 145). Does repetition temporally specific memory imagery. Semon attenuate state-dependent effects? Can we find argued that the ecphory of generic images (e.g., evidence of state-dependence in semantic a particular house) occurs when "...all the memory? Are frequently retrieved memories engrams belonging to my view of the house are less susceptible to state-dependent effects than allowed to act homophonously" (1923, p. 277). rarely retrieved memories? These issues, However, with different retrieval demands, barely touched upon in present-day research, specific temporally dated images of the house are intriguing and important questions about may be ecphorized: "When I wish to ecphorize state-dependence that emerge directly from the images of this house I can do so by fixing consideration of Semon's analysis. my attention on a definite, temporally deter- mined engram of the same..." (p. 276). Thus, the information from the indiviudal engrams is combined via homophony; and whether or Homophony, Repetition and Recognition not homophonous resonance occurs depends Earlier in the paper we briefly described on the intentions of the rememberer and the Semo~a's conception of homophony as a conditions ofecphory. It would be tempting to resonance metaphor that Semon employed to suggest that Semon was here anticipating the describe the way in which information from distinction between episodic and semantic different sources is combined. The following is memory (Tulving, 1972) as a utilization a general characterization of homophony: phenomemon; but since he did not explicitly RICHARD SEMON'S THEORY OF MEMORY 735 discuss such a distinction, such a statement vailing trend is found in Ward's (1893) paper, would probably reveal more about our own in which he explicitly distinguished between theoretical dispositions than about Semon's "functional" (strengthening) and "atomistic" thoughts on the matter. In any case, Semon's (multiple-trace) views of repetition in the conception of homophony, although admit- context of a discussion concerning recognition tedly somewhat ,~ague, again reflects his and association. However, debate examining intimate concern with the nature of retrieval the relative merits of strengthening and processes, and stands out as a unique con- multiple-trace points of view did not emerge in ception in its time. There are few ideas in the the subsequent experimental or theoretical memory literature of his day that bear even the literature. remotest resemblance to Semon's conception It is against this background that we of homophony, as noted by Becher (1910) in a introduce Semon's theory of repetition effects: review of Die mnemischen Emfindungen: "Every repetition of a stimulus and, con- perhaps Selz's notion of"pattern completion" sequently, of an original excitation deposits a (Kir~tsch, 1974) or Loeb's idea of association new engram which, if by nothing else, is by "esonance (Loeb, 1901) are the closest. distinguishable from all its predecessors by the ]r Als0.the notion of homophony bears a strong important difference of its being an integral resemblance to the resonance metaphors of element of an engram-complex belonging to a retrieval recently adopted by Lockhart, Craik, new layer" (1923, p.254). In contrast to the and Jacoby (1976), Moeser (1977), and by dominant strengthening theory of the time, Ratcliff (1978). Semon hypothesized a mechanism of What is most interesting about Semon's repetition effects much closer to the recently conception of homophony is the use to which advanced multiple-trace views of repetition he put it in the analysis of various memory (Bernbach, 1969; Bower, 1967; Hintzman & problems. Let us specifically consider Semon's Block, 1971). Like Semon, these theorists have postion on the problem of repetition. The argued that each repetition of a stimulus dominant theoretical approach to repetition creates a separate, unique memory trace. effects during Semon's time was the strength- Recent critical reviews contrasting multiple- ening view: Repetition exerts its beneficial trace and strength theories have concluded effects on memory by strengthening the repre- that the bulk of the experimental evidence sentation of the repeated item. This paradig- currently favors the multiple-trace hypothesis matic theory of repetition effects was stated (Hintzman, 1976: Howell, 1973). However, clearly by Ebbinghaus: "... as the number of while most modern multiple-trace theories :epetitions increases, the series are engraved suggest some sort of read-out of the number of more and more deeply and indelibly..." stored traces as the vehicle of repetition effects, (1885, p. 53). In almost all studies of repetition Semon had something quite different in mind. during the time period we have considered, He ventured that although multiple traces are strengthening is assumed to be the mechanism stored, the output from the memory system of repetition effects. Of all the questions asked could be in the form of separate traces or of about repetition at the time [which include the some amalgam of the separate traces via problem of massed vs distributed repetitions homophony, depending upon the conditions (Browning, Brown, and Washburn, 1913; of retrieval. So while repetitive input to the Perkins, 1914), rate of repetition (McGamble, memory system is always in the form of 1916), and number of repetitions (Calkins, multiple traces, output varies, depending 1894)l questions and hypotheses concerning upon retrieval conditions and consequent the mechanism of repetition effects are con- homophony. Semon was quite explicit in spicuously absent. One exception to this pre- juxtaposing his theory with the strength 736 SCHACTER, EICH, AND TULVING theory, noting that "... if the formation of an which he developed his own theory of engram through excitation were a question of recognition, a comparison theory which relied mere facilitation of channels, the repetition of heavily on the process o f differentiating homo- an excitation would, at best, only enlarge the phony. engram, but could not create a new, distinct, Semon distinguished two kinds of isolatedly ecphorable engram..." (1923, recognition. Simple recognition of a previous p. 255) and stating emphatically that: event is "...the manifestation of differ- REPETITION OF A STIMULUS DOES NOT entiating homophony between an original and STRENGTHEN AN ALREADY EXISTING a mnemic sensation..." (1923, p.283) and ENGRAM, BUT GENERATES A NEW does not entail conscious awareness of the ENGRAM, AND THE MNEMIC EXCI- TATIONS RESULTING FROM ANY SUB- difference between the original and newly SEQUENT ECPHORY ARE IN HOMO- perceived stimuli. To the contrary, in recog- PHONY (1921, p. 169). nition accompanied by the "sensation of dif- These statements leave little room for doubt ference" the differentiating homophony gives about Semon's position on repetition effects; rise to a "differential of sensation" (that is, a they also illustrate the functional use to which discrepancy between original and re- Semon put the concept of homophony. encountered stimuli that is large enough Fro be In Semon's analysis of recognition, we find consciously perceived by the rememberer). the notion of homophony being put to a Semon characterized both of these modes of slightly different use, and again encounter recognition as products of "homophonous Semon's distaste for "strengthening" views of comparison." Consistent with his earlier memory. By the time Semon published his two position that new engrams are established by books, there already existed a substantial each ecphory, Semon posited that both kinds literature, both experimental and theoretical, of recognition, which he viewed as subtypes of on the problem of recognition (Woods, 1915). ecphory, deposit new engrams in the memory Many different analyses of recognition had system. been put forward in this literature [Woods Semon professed no special concern with (1915) was able to distinguish between 13 the problem of recognition, commenting that theories] and debate among theorists of dif- "...it is interesting to us only as the mani- ferent persuasions often reached highly festation of differentiating homophony and in emotional levels. Among the popular theories its capacity of a differential of sensation" of the time were notions that recognition (1923, p. 288). Semon did make a special effort depends upon the reactivation of images to criticize H6ffding's (1893) theory that re- associated with the recognized object cognition results from the "greater ease" of (Hollingworth, 1913; Lehmann, 1889); the neural transmission of a second encounter theory that recognition is a subprocess of with a stimulus relative to the first, noting that recall (Miiller, 1913); the view that recognition this theory was in direct opposition to his own results from a comparison of image and "homophonous comparison" view. In terms of percept (Wolfe, 1886; Foucault, 1911); and the modern theories of recognition, Semon's ideas quite popular idea that recognition and the seem close in spirit to those of modern re- accompanying "feeling of familiarity" are searchers who posit a continuity between caused by a facilitation in underlying neural recognition and recall (Lockhart, Craik, & processing when perceiving a stimulus for a Jacoby, 1976; Tulving & Watkins, 1973). second time (Allin, 1895; Dearborn, 1899; Semon, like these theorists, drew no sharp H6ffding, 1893). distinctions between recognition and other It was this last view that Semon emphati- forms of ecphory. However, Semon did not cally disagreed with, and in contrast with offer specific comparisons of recognition and RICHARD SEMON'S THEORY OF MEMORY 737 recall; hence we must be cautious in con- cessing in memory. Semon noted that the sort trasting his ideas on this topic to modern of competition between perceptual and theories. mnemonic information he had previously dis- cussed exemplified spatial interactions. He Miscellaneous Phenomena then initiated his analysis of serial vs parallel We have now presented the body of processing by remarking that, "Instead of two Richard Semon's theory of human memory. simultaneous states of sensation, we are now While there are numerous other points in dealing with two temporally parallel chains of Semon's work that we will not discuss, two sensation in competition..." (1923, p.313). ideas do merit at least some mention. The issue of interest to Semon had to do with The first concerns Semon's analysis of the question of what takes place when two ~'competition" between stimulus input and engram-complexes are related equally to a output from the memory store. Semon was particular ecphoric cue. In such cases, are the interested in whether original and mnemic two candidate complexes retrieved in parallel, sensations are processed in the same channels or only through "alternately ecphorizing," or regions of sensation-fields. Semon based his that is, serially processing the two complexes? discussion upon an analysis of binocular In The Mneme, Semon concluded that in all rivalry (that is, competition between two such cases, serial processing is the rule; original sensations) and concluded that apparent parallel processing could be attri- original and mnemic sensations do in fact buted to rapidly alternating ecphory. How- compete with each other, and hence must ever, in Mnemic Psychology, Semon modified share common mechanisms. Semon specified this position: four consequences of such competition, but It is not altogether impossible to produce side by deferred their further discussion to future side and simultaneously two manifestations of work: ecphorized verbal engrams and to let two series of such engrams run their course simultaneously side (1) Mnemic sensations already present lose in by side.., side by sideness undoubtedly exists, if for vividness: (2) or are extruded; (3) the ecphory of a very short time, and therefore I have to abandon new sensations is hampered: (4) or altogether my first contention as to the impossibility of any prevented according as other original or mnemic simultaneous manifestation of two different chains sensations are already present or simultaneously of excitation ecphorized from verbal engrams ecphorized (1923, p. 312). (p. 315). Here we find Semon clearly anticipating the Here again we find Semon concerned with a issues arising from Sternberg's (1969) pace- problem of little interest to his contem- setting work that initiated interest in serial vs poraries. With the exception of Baxt's early parallel processing in memory. This question work on masking effects (see Murray, 1976), was simply not part of memory research in there is little to be found in the literature of Semon's day; but for Semon, interest in serial Semon's time concerning interference or com- vs parallel ecphory is just another mani- petition between perceptual and mnemonic festation of his theoretical concern with the information. Several modern experimenters process of retrieval. have investigated various aspects of this prob- lem (e.g., Chow & Murdock, 1975; Doost & Turvey, 1971; Johnston, Griffith, & Wagstaff, FLAWS IN THE THEORY 1972), but research on this topic is just getting under way. We have now outlined Semon's theory of The second idea, related to the first, is memory and have contrasted it with the Semon's interest in serial vs parallel pro- theories of his time and with those of today. 738 SCHACTER, EICH, AND TULVING

We have been impressed by Semon's original organism is fundamentally a passive one, and incisive thinking and by the remarkable incapable of changing or directing the flow of degree to which his various theoretical postu- information into the memory system. His lates accurately forecast many problems and theory would have great difficulty accounting theories of modern memory research. How- for modern research on the coding and trans- ever, we do not wish to imply that Semon's formation of sensory input. Of course, almost theory was free of problems or inconsistencies: every theory of his time would have an as with any psychological theory of his time, identical problem; but this is clearly one area or of the present time, Semon's analysis was in which Semon's work did not anticipate incomplete in several important respects. modern trends. The most conspicuous omission from Semon's theory is any attempt to specifically WHY IS SEMON'S WORK UNKNOWN TODAY? deal with the problem of forgetting; in fact, it is extremely difficult to find the word "for- In the light of the points made in the getting" in either of Semon's books. A second previous section, it would perhaps be egsiest problem, related to the first, is that Semon did to answer the above question by simply noting not assign a functional role to interference that Semon's theory was clearly imperfect, and phenomena in his account of memory. As that its subsequent obscurity befits such a mentioned earlier, he did offer some dis- flawed theory. However, such a line of argu- cussion of "competition" and interference as ment would not take us very far, since all other regards perceptual and mnemonic infor- theories of memory put forward in the period mation: but he did not systematicallyintegrate we have considered had at least as many gaps this into his theory of memory. Given his lack as Semon's theory. Indeed, comprehensive of concern with forgetting, it is hardly sur- theories of memory were conspicuously prising that interference phenomena did not absent from the literature during the time that play a functional role in Semon's theory. Semon wrote; it is no exaggeration to say that A third conspicuous gap in Semon's Semon's theory attempted to unify and ex- analysis is his failure to specify the role of plain more phenomena and problems than attention in memory storage. Semon did attri- almost any other theory of his time. Why, bute some importance to attentional pro- then, has such a theory, so close in spirit to cesses occurring at retrieval, as noted several many modern positions, remained virtually times, but he never considered its possible role unknown to contemporary researchers? as a determinant of memory storage. As it We will suggest four factors that may have stands, Semon's theory suggests that all per- contributed to the obscurity which character- ceptual events are given permanent engraphic ized Semon's theory then and has continued to representation; there is no mechanism for this day. First, consider the heavy emphasis emphasizing relevant and for ignoring irre- placed on the conditions, functions, and pro- levant information at the time of study. The cesses of retrieval in Semon's theory. As noted role of attention in memory had been given earlier, there were few studies of retrieval in experimental and theoretical consideration in Semon's time compared to the large number the years preceding Semon's work (e.g., of studies directed at other features of human Gordon, 1903): it is difficult to understand memory. More important, the few studies of why he did not address this issue. retrieval were divorced from any theoretical A final noticeable drawback of Semon's superstructure that would have accounted for position is that it leaves little room for active retrieval phenomena in a systematic manner. encoding processes, that is, transforming and With the exception of Selz and possibly recoding of input. Semon's remembering Hollingworth, systematic theoretical concern RICHARD SEMON'S THEORY OF MEMORY 739 for retrieval phenomena was rare in Semon's A third and more important reason why time. Why, then, should anyone have paid Semon's theory of memory has not received attention to Semon's views? Semon's prob- its due acknowledgment is related to the lems and those of his contemporaries were memory-heredity issue mentioned earlier. very different; retrieval phenomena were of the Semon's Lamarckian views, and his thesis that greatest importance to Semon and of the least the mechanisms of heredity and memory are importance to his contemporaries. Toulmin identical, received considerable harsh critic- (1961, p. 57) captured the essence of such a ism (e.g., Weismann, 1906). Consequently, his state of affairs in the domain of science: name became closely tied with the "wrong" views on the memory-heredity problem which Men who accept different ideals and paradigms have really no common theoretical terms in which lessened considerably the impact of his views to discuss their problems, They will not even have on human memory in general. Numerous the same problems: events which are "phenomena" authors cite Semon for his ideas on heredity in one man's eyes will be passed over by the other but nowhere mention the bulk of his theory of as "perfectly natural." human memory; Brett (1921), Edgell (1924), Given Semon's "aparadigmatic" stature Moore (1939), and Rignano (1926) exemplify with'-respect to the rest of the field, it is not this tendency. surprising that his theories were never noticed A fourth factor that we propose to account and passed down to succeeding generations. for the lack of recognition of Semon's theory is A second, and probably less important, that he provided no original experimental factor contributing to Semon's obscurity is his evidence to support 'his theory. Semon often invention of terminology. As discussed earlier, cited the experimental literature of the time, Semon created his own scientific terms in and attempted to incorporate experimental order to avoid the misleading connotations findings into his analysis of, various memory inherent in the everday language typically problems: he commented explicitly that used to describe human memory. Others had "...the results of experimental psychology commented on the problem (Bentley, 1899: must be reckoned with" (1923, p. 57). How- Ebbinghaus, 1885; Hamilton, 1859), but only ever, Semon offered no original evidence other Semon did something about it. Unfortunately, than his own introspection~ in support of his his usage of admittedly strange terms such as theory. Accordingly, one is in no way driven to "engraphy" and "ecphory," which led to agree with Semon or take his radical, strange- potentially intimidating chapter titles such as sounding notions very seriously at all. "Ecphoric Quantivalence of Components" These conjectures may or may not be valid: (1923, Chap. 11), may have served to create a they are, admittedly, educated speculations. barrier to the uninitiated reader. Additionally, We also do not know if progress in memory it became easy to focus upon Semon's termi- research would have been accelerated had nological creations rather than concentrating Semon's ideas been accepted in their time. But on the substance of his work. For instance, we do know that Semon's work on memory Campion and Smith (1934, p. 105) noted has been ignored for over half a century, and rather harshly that, "It is held by many that there must have been some reasons for it. Semon's 'engrams' have no neurological signi- In her introduction to Mnemic Psychology, ficance and should be dismissed with others of Vernon Lee lamented that Semon devoted so the uncouth terms in which he enshrined his much of his time to the "hopeless" psychological tenets." Thus, in his quest for Lamarckian thesis, but expressed optimism purity and precision of expression, Semon concerning Semon's psychological theory of may have unwittingly contributed to the memory: isolation of his own work. 740 SCHACTER, EICH, AND TULVING

The psychological part of his work remains, BERGSON, H. Matter and memory. New York: The however highly elaborated, a fragment--a frag- Macmillan Company, 1911. ment, however, whose shape and substance are so BERNBACH, H. A. Replication processes in human suggestive that I cannot but think that a part of memory. In G. H. Bower & J. T. Spence (Eds.), The Semon's importance may consist in what will be psychology of learning and motivation (Vol. 3). New added by others to the work he left unfinished York: Academic Press, 1969. (1923, p. 27). BINET, A., & HENRI. V. La memoire de roots. L'Annee It is perhaps sad that Semon's work never Psychologique, 1894, 1, 1-23. BORING, E. G. Sensation and perception in the history of did exert the influence that Vernon Lee hoped experimental psychology. New York: Appleton it would. But it is ironic that modern students Century Crofts. 1942. of memory have been concerned with so many BOUSEFIELD. W. R. The basis of memory. London: Kegan of the issues dealt with in his theory without Paul, 1928. being aware of it. Sixty years after his death, BOWER, G. H. A multicomponent theory of the memory trace. In K. W. Spence & J. T. Spence (Eds.), The Semon's rich theoretical constructs and novel psychology of learning and motivation (Vol. 1), New conceptualizations not only deserve full re- York: Academic Press, 1967. cognition; they are also potential sources of BRANSFORD, J. D.. & FRANKS, J. J. The abstraction of insight to those of us who continue to be linguistic ideas. Cognitive Psychology, 1971, 2, intrigued by the phenomena of memory that 331-350. * BRETT, G. S. A history of psychology. London: George Richard Semon studied with such penetrating Allen & Unwin, 1921. vision. BROWNING, M., BROWN, D. E., & WASHBURN, M. F. The effect of the interval between repetitions on the speed of learning a series of movements. American REFERENCES Journal of Psychology, 1913, 24, 580-583. BURNHAM, W. H. Memory, historically and experimen- ABBOTT, E. E. On the analysis of the factor of recall in the tally considered. American Journal of Psychology. learning process. Psychological Monographs, 1909, 1888-9, 2, 39-90; 225-270; 431-464; 568-622. 11, 159-177. CALKINS, M. W. Association. PsychologicalReview, 1894, ALLIN, A. The recogr~ition theory of perception. 1, 476-483. American Journal of Psychology, 1895, 7, 236-248. CAMPION, G. G., & SN1TH, E. G. The neural basis of ANDERSON, J. R., & BOWER, G. H. Human associative thought. London: Kegan Paul, 1934. memory. Washington, DC: Winston, 1973. CARR. H. Length of time interval in successive associa- ATKINSON, R. C., & SHIFFRIN, R. M. Human memory: A tion. Psychological Review, 1919, 26, 335-353. proposed system and its control processes. In K. W. CHOW. S. L., & MURDOCK, B. B. The effect of a subsidiary Spence & J. T. Spence (Eds.), The psychology of task on iconic memory. Memory and Cognition, learning and motivation (Vol. 2). New York: 1975, 3, 678-688. Academic Press, 1968. CROSLAND, H. A qualitative analysis of forgetting. BADDELEY, A. D., & HITCH, G. Working memory. In Psychological Monographs. 1921, 29, Whole No. G. H. Bower (Ed.), The psychology of learnin9 and 130. motivation (Vol. 8). New York: Academic Press, CROVITZ. H. F., & SCHIFFMAN, H. Frequency of episodic 1974. memories as a function of their age. Bulletin of the BALLARD, P. A. Obliviscence and reminiscence. British Psychonomic Society, 1974, 4, 517-518. Journal of Psychology Monograph Supplements, DARLEY, C. F., & MURDOCK, B. B. Effects of prior free- 1913, 1, 1 81. recall testing on final recall and recognition. Journal BARTLETT, F. C. An experiment upon repeated repro- of Experimental Psychology. 1971, 91, 66-73. duction. Journal of General Psychology. 1928, 1, DEARBORN, G. ¥'. Recognition under objective rehearsal. 54-63. Psychological Review, 1899, 6, 395-406. BARTLETT, F. C. Remembering. Cambridge: University DOOST, R., & TURVEY, M. I. Iconic memory and central Press, 1932. processing capacity. Perception and Psychophysics, BECHER, E. Review of Die mnemischen Empfindunoen. 1971, 9, 269-274. Archiv fiir Gesamte Psychologie, 1910, 17, 165-172. DOSrtER, B. A., & RusSO, J. E. Memory for internally BENTLEY, I. M. The memory image and its qualitative generated stimuli. Journal of ExPerimental fidelity. American Journal of Psychology. 1899, 11, Psychology." Human Learning and Memory, 1976.2, 1-48. 633-640. RICHARD SEMON'S THEORY OF MEMORY 741

E~mYGUA'dS, H. Uber das Gedhchmis. Leipzig: Duncker HOROWITZ, L. M., & PRYTULAK, L. S. Redintegrative & Humblot. 1885. (English translation. Dover memory. Psychological Review, 1969, 76, 519 531. Press. 1964). HOWELL, W. C. Representation of frequency in memory. EDGELL, B. Theories ojmemory. Oxford: Clarendon Press, Psychological Bulletin. 1973, 80, 44-53. 1924. IZAWA. C. Comparison of reinforcement and test trials in ElCH. J. E. State-dependent retrieval of information in paired-associate learning. Journal of Experimental human episodic memory. In I. M. Birnbaum & E. S. Psychology, 1969.81, 600-603. Parker (Eds.). Alcohol and human memory. JAMES. W. Principles of psychology. New York: Henry Hillsdale, N.J.: Lawrence Erlbaum Associates, 1977. Holt. 1890. FOUCAULT, M. Etude exp6rimentale sur l'association de JANET, P. L'evolution de la mbmoire et de la notion du resemblance. Archives de Psychologic, 1911. 10, temps. Paris: Chahine, 1928. 338-360. JOHNSTON. W. A., GR1FFITH, D., & WAGSTAFF, R. R. FREDERIKSEN. C. H. Acquisition of semantic information Speed, accuracy, and ease of recall. Journal of from discourse: Effects of repeated exposures. Verbal Learning and Verbal Behavior, 1972. 11, Journal of Verbal Learning and Verbal Behavior, 512-520. 1975. 14, 158-169. JONES, G. V. A fragmentation hypothesis of memory: FREUD, S, The interpretation of dreams. London: George Cued recall of pictures and of sequential position. Allen & Unwin, 1913. Journal of Experimental Psychology." General. 1976. FROEBERG, S. Simultaneous versus successive association. 105, 277-293. Psychological Review, 1918, 25, 156-d63. KENNEDY. F. On the experimental investigation of GALTON. F. Psychometric experiments. Brain, 1879. 2, memory. Psychological Review. 1898.5, 477 499. 148-162. KINTSCH, W. The representation of meaning in memory. GARDINER, J. M.. & KLEE, H. Memory for remembered Hillsdale, N J: Lawrence Erlbaum Associates, 1974. events: An assessment of output monitoring in free KLEE. H., & GARDINER. J. M. Memory for remembered recall. Journal of Verbal Learning and Verbal events: Contrasting recall and recognition. Journal Behavior. 1976, 15, 227-234. of Verbal Learning and Verbal Behavior. 1976. 15, GIBSON. J. J. The reproduction of visually perceived 471-478. forms. Journal of Experimental Psychology, 1929. KrHLER, W. Gestahpsychology. London: G. Bell & Sons. 12, 1-39. 1930. GORDON. K. Meaning in memory and attention. KorrKa, K. Principles of gestalt psychology. New York: Psychological Review, 1903, 10, 26~283. Harcourt. Brace, & Co, 1935. GUENTHER. R. K.. & LINTON, M. Mechanisms of tem- KOSTYLEFF, N. Review of Die Mneme. Review poral coding. Journal of Experimental Psychology: Philosophique, 1911, 71,749-753. Human Learnin9 and Memory, I975, 1, 182 187. KROLL, N. F. A., PARKS, T., PARK1NSON, S. R., BIEBER, HAMILTON. W. Lectures on metaphysics and logic, S. L, & JOHNSON, A. L. Short-term memory while Edinburgh: William Blackwood & Sons, 1859. shadowing: Recall of visually and of aurally pre- HAWKINS, C. J. Experiments on memory types. sented letters. Journal of Experimental Psychology, Psychological Review, 1897, 4, 289-294. 1970, 85, 220-224. H1NTZMAN, D. L.. Repetition in memory. In G. H. Bower KUHLMANN, F. On the analysis of the memory conscious- (Ed.), The psychology of learning and motivation ness: A study in the mental imagery and memory of (Vol. 10). New York: Academic Press, 1976. meaningless visual forms, Psychological Review, HINTZMAN, D. L.~ & BLOCK, R. A. Repetition in memory: 1906, 13, 316-348. Evidence for a multiple-trace hypothesis. Journal of KUHLMANN, F. The present statues of memory in- Experimental Psychology, 1971.88, 297-306. vestigation. Psychological Bulletin. 1908. 5, HOFFDING, H. Outlines of psychology'. London: 285-293. Macmillan. 1891. KOLPE, O. Outlines ofpsychology. New York: Macmillan. H&'FDING, H. Zur Theorie des Wiedererkennens. 1895. Philosophische Studien. 1893~ 8, 86-96. LAIRD, D. A. REMMEkS, H., & PETERSON, L, J. An HOLLINGWOR'rI-I. H. L. Characteristic differences between experimental study of the influences of organization recall and recogntion. American Journal of of material for memorizing upon its retention. Psychology, 1913, 24, 532-544. Journal of Experimental Psychology, 1923, 6, 6%8 l, HOLHNGWORTH, H. L. The psychology of thought, New LANDAUER. T. K. Memory without organization: York: Appleton & Co., 1926. Properties of a model with random storage and HOLLINGWORTH, H. L. General laws of redintegration. undirected retrieval. Cognitive Psychology, 1975, 7, Journal of General Psychology. 1928, 1, 79-90. 495-531. 742 SCHACTER, EICH, AND TULVING

LASI-ILEY, K. S. In search of the engram. In F. A. Beach, M~LLER, G. E., & SCHUMANN,F. Experimentelle Beitr/ige D. O. Hebb, C. T. Morgan, & H. W. Nissen (Eds.), zur Untersuchung des Gedfichtnisses. Zeitschriftfiir The neuropsychology of Lashley. New York: Psychologie, 1894, 6, 81-190. McGraw-Hill. 1960. MURDOCK, B. B. JR. Human memory: Theory and data. LEE, V. Introduction. Being notes on some applications of Potomac, Md: Lawrence Erlbaum Associates, 1974. mnemie principle iff~recent psychological literature. MURRAY, D. J. Research on human memory in the In R. Semon, Mnemicpsychology. London: George nineteenth century. Canadian Journal of Allen & Unwin, 1923. Psychology, 1976, 30, 201-220. LEHMANN, A. Ober Wiedererkennen. Philosophische MYERS, A. G., & MYERS. E, C. Reconstructive recall. Studien, 1889. 5, 96-156. American Journal of Psychology. 1916, 27, 493-506. LOCKHART, R. S., CRAIK, F. I. M.. & JACOBY, L. Depth of processing, recognition, and recall. In J. Brown PERKINS, N. L. The value of distributed repetitions in rote (Ed.), Recall and recognition. London: Wiley, 1976. learning. British Journal of Psychology, 1914, 7, LOEB, J. Comparative physiology of the brain and compara- 253-261. tive psychology. London: John Murray, 1901. PRINCE, M. The unconscious. New Yorki Macmillan, LorTus, E. F., MILLER. D. G., & BUR~qS, H. J. Semantic 1916. intergration of verbal information into a visual RAFFEL, G. The effect of recall on forgetting. Journal of memory. Journal of Experimental Psychology: Experimental Psychology, 1934, 17, 828-838. Human Learning and Memory. 1978, 4, 19-31. RATCLIrr, R. A theory of memory retrieval. Psychological LuH. C. W. The conditions of retention, Psychological Review, 1978, 85, 59-108. Monographs. 1922, 31,Whole No. 142. REIFF, R., • SCHEERER, M. Memory and hypnotic age LURIA, A. R. The working brain. London: Penguin, 1973. regression. New York: International Universities Press, 1959. RIBOT, T. Diseases of memory. New York: Appleton, McDOUGALL. W. Body and m&d. London: Methuen, 1882. 1911. RIGrqANO, E. Biological memory. New York: Harcourt, MACH. E. The analysis of sensations. New York: Dover, Brace. & Co., 1926. 1959. ROBINSON. E. S. Memory. Psychological Bulletin. 1924. MCGAMBI~E. E. A. Rate of repetition and tenacity of 21, 569-594. impression. Psychological Monographs, 1916, 22, ROBINSON, E. S. Association theory to-day. New York: 101-151. Century, 1932. McGEOCH, J. A. Memory. Psychological Bulletin, 1928, 5, ROEDIGER, H. L. Inhibiting effects of recall. Memory & 513-549. Cognition, 1974, 2, 261-269. McGEocH, J. A. Memory. Psychological Bulletin, 1930, RUNDUS, D. Negative effects of using list items as recall 27, 514-563. cues. Journal of Verbal Learning and Verbal McGEOCI~, J. A. The psychology of human learning: A Behavior, 1973, 12, 43-50. bibliography. Psychological Bulletin, 1933, 30, RUSSELL. B. The analysis of mind. London: George Allen 1 62. & Unwin, 1921. McGEocH, J. A. Learning. In E. G. Boring, H. S. Langfeld. & H. P. Weld (Eds.), Introduction to SCHg6D1NGER, E. My view of the world. Cambridge: psychology. New York: Wiley, 1939. Cambridge Univeristy Press. 1964. MEUMANN. E. The psychology of learning. New York: SELZ, O. ~)ber die Gesetze des 9eordneten Denkverlaufs. Appleton Company, 1913. Stuttgart: Spemann, 1913. MOD1GL1ANI, V. Effects on a later recall by delaying initial SELZ, O. Zur Psychologie des producktiven Denkens und recall. Journal of E:cperimental Psychology." Human des Irrtums. Bonn: Cohen, 1922. Learning and Memory, 1976, 2, 609-622. SEMON, R. In the Australian bush and on the coast of the MOESER, S. D. Recognition processes in episodic memory. Coral Sea. London: Macmillan, 1899. Canadian Journal of Psychology, 1977, 31, 41-70. SEMON, R. Die Mneme als erhaltendes Prinzip im Wechsel MOOgE. T. V. Cognitive psychology. New York: des organischen Geschehens. Leipzig: Wilhelm Lippincott, 1939. Engelmann, 1904. MOLLEg, G. E. Zur Analyse der Gedfichtnistfitigkeit und SEMON, R. Die nmemischenEmpfindungen. Leipzig: des Vorstellungsverlaufes, III. Teil. Zeitschrift fffr Wilhelm Engelmann, 1909. Psychologie, Erg"dnzungsband, 1913, 8, 1-567. SEMON, R. The nmeme. London: George Allen & Unwin, M/JLLER. G. E.. & P~LZECKER, A. Experimentelle Beitrfige 1921. zur Lehre vom Gedfichtnis. Zeitschrift fiir SEMON, R. The mneme. London: George Allen & Unwin, Psychologie, Erg?inzungsband: 1900, 1, 1-300. Unwin, 1923. RICHARD SEMON'S THEORY OF MEMORY 743

STERNBERG. S. Memory scanning: Mental processes re- U~qDERWOOD. B. J. Attributes of memory. Psychological vealed by reaction-time experiments. American Review, 1969, 76, 559-573. Scientist. 1969.57, 421-457. UNDERWOOD, B. J. Temporal codes for memories. THORNDIWE, E. L. The elements of psychology. New York: Hillsda]e, NJ: Lawrence Erlbaum Associates. 1977. A. G. Seller. 1913. WARD, J. Assimilation and association. Mind. 1893. 3, TITCHENER. E. B. A text-book of psychology. New York: 347 362. Macmillan. 1911. WARREN, H. C. History of the association psychology. TOULMIN. S. Foresight and understanding. New York: London: Cons~/able & Co.. 1921. Harper & Row. 1961. WEISMANN, A. Semon's "Mneme" und die' Verebung TROW, W. C. Recall vs repetition in the learning of rote erworbener Eigenschaften. Archly for Rassen- und and meaningful material. American Journal of Gesellschafts biologie. 1906, 3, 1-27. Psychology. 1928, 40, 112-116, WmTELY. P. L.. & MCGEOCH, J. A. The effect of one form TULVING, E. The effects of presentation and recall of of report upon another. American Journal of material in free-recall learning. Journal of Verbal Psychology. 1927, 38, 280-284. Learning and Verbal Behavior. 1967, 6, 175-184. WOHLGEMUTH, A. Simultaneous and successive asso- TULVING, E. Episodic and semantic memory. In E. ciation. British Journal of Psychology, 1915, 7, Tulving & W. Donaldson (Eds.), Organization of 434-452. memory. New York: Academic Press. 1972. WOLFE. H. K. Untersuchungen fiber das TongedS, chtnis. TVLVING, E. Ecphoric processes in recall and recognition. Philosophiche Smdien, 1886, 3, 534-571. In J. Brown (Ed.). Recall and recognition. London: WOODS, E. L. An experimental analysis of the process of Wiley, 1976. recognizing. American Journal of Psychology. 1915. TULVING. E., & THOMSON, D. M. Encoding specificity and 26, 313-387. retrieval processes in episodic memory. WUNDT. W, Outlines of psychology. Leipzig: Wilhelm Psychological Review. 1973, 80, 352-373. Engelmann. t902. TULVlNG. E., & WATKINS. M. J. Continuity between YOUNG, J. Z. The organization of a memory system. recognition and recall. American Journal of Proceedings of the Royal Society of London. 1965, Psychology. 1973, 86, 739-748. 163, 285-320. UNDERWOOD. B. J. False recognition produced by im- plicit verbal responses. Journal of Experimental Psychology. 1965, 70, 122-129. (Received July 6, 1978)