Psychological Research © Springer-Verlag 1991

PsycholRes (1991) 53: 162-168

Haptic information processing in direct and indirect memory tests Werner Wippich Fachbereich I - Psychologic,Universit~tTrier, Postfach3825, W-5500Trier, FederalRepublicof Germany ReceivedOctober 17, 1990/AcceptedMarch 18, 1991

Sulmnary. Effects of learning can show up in a direct, i.e., an explicit, way or they can be expressed indirectly, i. e., in an implicit way. It was investigated whether motor processes underlie effects of repetition priming in haptic information exploration. In the test phase, blindfolded subjects had to handle objects in order to answer questions as fast as possible about their properties (e. g., temperature, texture, weight, or form), exploring the object with their hands. Some of the objects were old ones (presented in a study phase); others were added as new objects to the test phase. In addition, recognition judgements were required. The results demonstrated reliable effects of repetition priming in terms of reaction times to old, in comparison with new, objects for subjects who had been treated the same way in the study phase (active touch). Passive touch at encoding or studying the names or the visible objects themselves did not lead to effects of repetition priming in the test phase. On the other hand, performing adequate hand movements during the study improved recognition memory. The role of motor processes in indirect and direct tests of memory is discussed and related to research on memory of action events.

Introduction Information acquired during a single episode can facilitate performance on a number of tests that do not make explicit reference to the study episode, such as word-stem and fragment completion (Graf & Mandler, 1984; Tulving, Schacter, & Stark, 1982), word identification (Jacoby & Dallas, 1981), and lexical decision (Scarborough, Cortese, & Scarborough, 1977). The facilitation of performance in indirect memory tests has been labeled repetition priming, and it occurs without the deliberate intention of recollecting the past episode. Graf and Schacter (1985) have used the descriptive terms explicit and implicit memory to denote the forms of memory involved in recall/recognition as

examples of direct memory tests and priming performance, respectively. Measures of performance in direct and indirect memory tests have been shown in some tasks to be stochastically independent (Eich, 1984; Hayman & Tulving, 1989; Jacoby & Witherspoon, 1982) in the sense that performance in the two tasks is uncorrelated at the level of individual items. The two types of memory measure have also shown functional independence in the effects of other variables such as levels of processing (Jacoby & Dallas, 1981), delay (Tulving et al., 1982), or retroactive and proactive interference (Graf & Schacter, 1987). Furthermore, priming effects can be surprisingly tong-lasting (Mitchell & Brown, 1988). Finally, the distinction between direct and indirect memory measures is relevant to the understanding of age differences in memory performance (Light & Singh, 1987; Wippich, Mecklenbrfiuker, & Brausch, 1989) and the cognitive capabilities preserved by patients suffering from organic amnesia and other disorders (Shimamura, 1986; Parkin, 1987). The growing body of evidence on the stochastic and functional independence of direct and indirect memory measures is based primarily on research with verbal materials (for an exception, see Schacter, Cooper, & Delaney, 1990). Another important trend in recent memory research, the discovery of motor memory or memory of simple action events, has relied primarily on measures of explicit memory and also has strong roots in verbal-memory paradigms. Typically, the content of the items to be learned are simple actions such as to light a cigarette, provided verbally by the experimenter. The core finding, which has been observed repeatedly (Cohen, 1983; Engelkamp, 1990b; Helstrup, 1989), is that performing the actions during encoding (either symbolically or with the denoted object) improves explicit memory, compared with learning the phrases under a standard learning instruction. Further research has shown that the performance of an action during encoding is decisive for recall or recognition. Observing how a model performs the actions or imagining somebody performing the actions leads to memory performance that is significantly worse than that under motor

163 encoding (Engelkamp & Krumnacker, 1980; Engelkamp & Zimmer, 1983). This result, however, is not a general one. Cohen has reported several experiments showing comparable memory performance after acting or after seeing someone else performing the actions (e. g., Cohen, Peterson, & Mantini-Atkinson, 1987). The discrepancy in the results may be due to differences in the experimental procedures (such as the provision of imaginary or real objects; for further discussion, see Engelkamp, 1990 a). Nevertheless, it seems justifiable to conclude that the performance effect is not reducible to imagery processes, but that performance (or underlying motor processes) is (are) decisive. This conclusion is corroborated by experimental findings in the paradigm of selective interference. Motor learning decreases with motor distraction as compared with imaginal distraction (Saltz & Donnenwerth-Nolan, 1981; Zimmer & Engelkamp, 1985). If motor processes underlie the performance effect, then it will be useful to study the role of motor processes in memory performance more directly. Indirect measures of memory may be suited to this task. There are at least four reasons for justifying this proposal. First, some researchers have pointed to interrelations between the rich body of literature on motor skills (e. g., Schmidt, 1988) and indirect memory measures (Roediger, 1990). Second, performed action events are multimodally encoded and comprise a variety of features (B~ckman, Nilsson, & Chalom, 1986). Indirect measures of memory are particularly well suited for isolating specific consequences and encoded features of a learning episode. For example, modality congruence during study and test is a potent variable when the test is indirect, but a weak one when the test is direct (for a review, see Richardson-Klavehn & Bjork, 1988). Third, there is some evidence that motor encoding provides the memory system with more item-specific information than does standard learning, but it does not add anything to relational information (Engelkamp, 1990b). Studies with indirect memory tests, on the other hand, have shown that manipulations influencing item-specific information will be more evidenced on such tests, whereas most direct tests are especially sensitive to relational information in memory (e.g., MacLeod & Bassili, 1989). Fourth, there has been some speculation (but no direct evidence) that the mere enactment of verbally provided action phrases adds a unique memory component to study encoding that may be revealed with indirect memory measures (Nilsson & B~ckman, 1989). A first attempt to study motor processes with indirect measures of memory was reported by Wippich, Mecklenbr~iuker, and Sidiropoulos (1990) within a developmental context. Children from 7 to 10 years old had either to perform simple actions such as to catch a ball or to observe how a model announced and performed the same actions. At the indirect testing stage, some of the verbs were presented again, together with new ones. Without any reference to the study phase, the children had to perform an adequate action and to associate objects that came to mind ("What can you catch?"). Finally, the children were required to recall the action events that had been performed or seen. The results showed that age-related differences were confined to direct memory measures. Moreover, free

recall benefitted from performance of the actions during encoding. On the other hand, the object-association scores surpassed the baseline, thus demonstrating effects of repetition priming, but were not influenced by the motor-encoding manipulation. However, the movements performed were more likely to match actions previously performed than actions seen. Thus, motoric consequences of the learning experience showed up with an indirect motor memory test, but could not be revealed by the performance on a verbal indirect memory test. Thus far, this study seems to show that multimodally encoded features of the learning episode (e.g., motor and verbal or conceptual attributes) can be disentangled at the level of priming performance. In order to study motor components more precisely, it may be useful to exert more control over the actions to be performed. Within the context of haptic information processing, Wippich and Wagner (1989) set blindfolded subjects to handle objects in order to answer questions about their distinct properties as quickly as possible. More specifically, the subjects had to make adequate hand movements to examine the object's temperature, texture, weight, or form. Other research has shown (e.g., Lederman & Klatzky, 1987), that the examination of these properties is associated with distinct hand movements. At the testing phase, the task was repeated with the same objects and questions and with further control items. The results demonstrated reliable effects of repetition priming: reaction times in answering questions about old objects were reliably shorter than those in answering the same questions about new objects. Is the motoric component within this task of haptic information processing really decisive for effects of repetition priming? Making adequate hand movements to an object (as well as symbolically enacting an action event) not only involves motor processes, but delivers kinesthetic and sensory information, too. Recently, Wippich (1991) has shown that the effects of repetition priming in haptic information processing depend primarily on a match in the hand movements. Subjects who had to wear plastic gloves in the study phase demonstrated an advantage in answering questions about old, as against new, objects, comparable with that of subjects who did not have this treatment at encoding. Since this manipulation should dampen the registration of sensory information about the objects, Wippich concluded that priming does not depend crucially on the retrieval of, and match with, previously encoded sensory features. On the other hand, changing the questions - and thus the hand movements - from the study to the test phase abolished priming effects, indicating a strong role for motor components in haptic information processing. Explicit memory was also hampered in subjects who wore gloves during the learning phase, as was revealed by their performance in a recognition test. This indicates an important role for sensory information with direct memory testing. Thus, subjects had built up sensory-motor records in the study phase that were accessed differently under the two task conditions. The current research investigates the extreme possibility that one need not encode sensory-motor records in the first place in order to reveal priming in haptic information processing in the test phase. All of the subjects were blind-

164 folded in the test phase and had to answer questions about the objects' properties as quickly as possible, exploring the object with their hands to verify the requested property. Some of the objects were old ones (presented in the study phase), others were added as new objects to the test phase. In addition, immediately after each answer subjects had to make a yes/no recognition judgement (direct m e m o r y test). The subjects were treated differently in the initial and the encoding phases. All of them had to answer questions about the objects' properties. But some o f them were given only the name of the objects (name group), some of them were shown the objects (seeing group), and others were blindfolded and provided with the objects themselves (touching group). In addition, some subjects within each group were instructed to make sufficient hand movements to examine the requested property on an imaged object (name group) or a seen object (seeing group) or to examine the property (touching group). Other subjects had to make their judgements from knowledge (name group) or from the seen or touched object without any hand movements. As can be seen, the design of the study phase combined procedures f r o m Wippich (haptic information processing) with aspects o f experiments looking for m e m o r y of action events. For example, subjects within the name group instructed to make sufficient hand movements to answer the question: "bottle: rough or smooth?" are treated like subjects instructed to enact a phrase such as "touch a smooth or a rough bottle". The results will show whether the performance effect reported in m a n y experiments on m e m o r y of action events will generalize to recognition judgements of blindfolded subjects provided with the objects themselves. In general, we expected that recognition m e m o r y would be enhanced by hand movements at encoding. The performance effect was expected to be strongest under conditions of impoverished encoding, that is, with blindfolded subjects, contrasting passive touch (no hand movements) with active touch. More importantly, we were interested to see whether the indirect m e m o r y test demonstrates any effects of repetition priming - that is, shorter reaction times to repeated questions about the old than about the new objects. If priming in haptic information processing can be reduced to motor components and does not depend on the encoding of sensory-motor records, one might expect that repetition priming would show up under all encoding conditions in which subjects had to perform adequate hand movements. It m a y be conceded that hand movements in the touching group during encoding will match better with similar movements at testing, whereas in the seeing group, and especially in the name group the match will be lower, because of a greater uncertainty in the properties of the objects to be touched, and thus in the hand movements. This leads to the prediction that priming will be most pronounced in the touching group, diminishing gradually from the seeing group to the name group. In any case, effects of repetition priming should not show up under encoding conditions without hand movements. On the other hand, priming m a y presuppose the establishment of sensory-motor records. According to the results of Wippich (1991), the motor component of the record m a y simply be more rapidly accessible than the sensory corn-

portent. This view presupposes that repetition priming in haptic information processing will be restricted to encoding conditions that provide subjects with sensory-motor information (the active touching group).

Method Subjects. There were 48 subjects, all undergraduate students from the University of Trier, with 8 in each between-subjects condition. Design. The experiment consisted of a study phase and a test phase. During the study phase, subjects were assigned randomly to the six groups resulting from crossing study modality (name, seeing or touching) with movements (with or without). The four types of question (temperature, texture, weight, or form) were associated with specific objects and were varied within subjects. In the test phase, all of the objects from the study phase were repeated with the same questions (old) and other objects were added as new ones. The latter variable (old vs. new) was manipulated within subjects. Thus the complete design conformed to a 3 (study modality) x 2 (with or without movement)x 4 (type of question) x 2 (old/new objects) mixed factorial. Materials. One of two series of items was used with equal frequency in the study phase for different subgroups of subjects. Each series consisted of 16 concrete objects (or names of these objects). During the test phase, both series were combined in a random order. Thus, each item was used with equal frequency as an old or a new one within each group of subjects. Four names or objects in each series were associated with one type of question. The association was maintained for repeated items in the test phase. The type of question is thus confounded with specific questions. This confounding was judged to be adequate to ensure a meaningful relation between objects and properties to be examined. The order of the objects within a series (or within both series) was such that the same question was never repeated immediately. The items were selected from normal household objects (e. g., bracelet, bulb, apple, nailfile). Reaction times were measured with an electronic timer. Procedure. At the beginning of the session, subjects were told that they would participate in an experiment concerned with phenomena of imagination and touch. Subsequent memory testings were not mentioned by the experimenter (incidental learning). In the study phase, the experimenter announced the names of the objects, showed each object separately, or placed each object on the palm of the nondominant hand. Subjects in the two touching groups were blindfolded and had to stretch their hands out on a table so that both palms were visible. All subjects had been familiarized with the four types of question. The two possible answers for each question were made clear. The subjects could answer "warm" or "cold" (temperature), ,,rough" or "smooth" (texture), "heavy" or "light" (weight), "angular" or "round" (form). Subjects in the conditions without any hand movements were told to respond as quickly as possible on the basis of their predominant impression (the passive touching group), image (the seeing group), or knowledge (the name group). Subjects who were instructed to use their hands to examine the requested property of the named, seen, or touched object were urged to come to a decision after they had made adequate exploratory hand movements. Each trial started with the presentation of the object (or object's-name), followed immediately by one of the four questions (temperature, texture, weight, or form). At this moment, the timer was started and stopped when the subject made his/her response ("warm"/"cold", "rough"/"smooth", "heavy"/"light", or "angular"/"round"). In the test phase which followed immediately, all the subjects were blindfolded and treated like the active touching group at encoding. That is, each object was placed on the palm of the nondominant hand. The experimenter announced one of the four questions. The subjects were instructed to use both hands in order to answer the question as quickly as posible. The timer was started by the experimenter by hand with the question and stopped with the response of the subject. With 16 subjects, a second person was present at the testing phase, timing the subjects without any

165 knowledge of their encoding conditions. Correlations between the two measures of reaction times were extremely high (r = .99). After each response to the property requested, all the subjects had to make a yes/no recognitionjudgement. To facilitate this decision, further explorations of the object were permitted.

Table 1. Mean reaction times (ms) to questions about weight or form

(WF) and temperature or texture (TT) for named, seen, or touched objects (name, seeing, and touching groups) with or without movements at encoding.

The encodingphase In agreement with previous research (Wippich, 1991), reaction times to questions about more global properties (weight and form, WF), on the one hand, and about more local properties (temperature and texture, TT), on the other hand, proved to be comparable and were combined. Thus, mean reaction times were calculated for each subject on the basis of eight WF questions and eight TT questions, respectively. An analysis of variance, with study modality and movement as between-subjects variables and type of question (WF or TT) as within-subjects variable on the mean scores, revealed main effects for movement, F (1/42) = 16.53, 052 = .21, and for type of question, F (1/42) = 9.24, 052 = .01. In addition, the triple interaction proved to be significant, F (2/42) = 3.92, 052 = .01, whereas all other effects can be neglected, F _< 1.50. Group means are shown in Table 1. As can be seen from the table, subjects who had to perform an adequate hand movement generally required more time to answer the questions, and TT questions were associated with longer :reaction times than those for WF questions. The triple interaction, however, shows that this effect varies with the conditions at encoding. An analysis of simple main effects showed that the effect of question type is most pronounced and significant for the active touching group, F (1/42) = 18.28, whereas all other comparisons within the groups did not reach the level of significance, F_< 1.20.

TT

No movement

name seeing touching

714 641 1040

883 806 1049

Movement

name seeing touching

1793 1519 1156

1908 1550 !815

Results

The results of the encoding phase and the two different memory tests (indirect and direct respectively) will be reported in different sections. The level of significance for all the results was set at .05. Estimates of effect size (052) are based on total variances. Preliminary analyses showed that the two series of materials were equivalent in both recognition and reaction times.

WF

Table 2. Mean reaction times (ms) to questions in the test phase for old

and new objects as a function of different encoding conditions. Old

New

No movement

name seeing touching

1289 1007 1363

1407 1039 1344

Movement

name seeing touching

1869 1223 1159

1763 1268 1529

conditions of encoding. Group means are shown in Table 2. An analysis of the simple main effects for object status confirmed what Table 2 demonstrates. A strong effect of repetition priming (with a mean difference of 370 ms per answer) is seen for the active touching group, F (1/42) = 15.39, thus corroborating previous results. All other comparisons within the groups failed to reach the level of significance, F < 1.52. In addition, a significant effect was associated with study modality, F (2/42) = 4.30, 052 = .07, with the shortest reaction times for those seeing the objects and the longest for the name group. Furthermore, reaction times to WF quetions were generally faster than those to TT questions, F (1/42) = 27.17, 052 = .03, with means of 1,246 ms and 1,464 ms, respectively. In contrast to the study phase, interactions with this variable could not be detected, F _< 1.84.

Recognition Repetitionpriming Mean reaction times were calculated for the eight TT questions and the eight WF questions about old and new objects, respectively. An analysis of variance with study modality and movement as between-subjects variables and type of question and object status (old vs. new) as withinsubjects variables did not show a general significant effect of priming, F (1/42) = 3.62, p <.10. Most importantly, the triple interaction between the between-subjects variables and object status proved to be significant, F (2/42) = 5.35, c~2 = .01. Thus, effects of repetition priming depend on the

For explicit judgements of memory, a three-factor analysis of variance (same variables as at encoding) was performed for hits, false alarms, and for an index of discrimination (Pr) that estimates the sensitivity of memory performance independently of decision bias. The discrimination index (hits-false alarms) was recommended by Snodgrass and Corwin (1988). As the results proved to be very similar (see Table 3 for the means of each dependent variable), we shall discuss the discrimination scores only. It should also be mentioned that the type of question was influential only in an interaction with the movement variable, F (1/42) -- 6.45, 052 = .01, overridden by the triple

166 Table 3. Meanrecognitionperformance(%Hits, %FalseAlarms,and Pr)

as a functionof differentencodingconditions. Hits

False alarms

Pr

No movement

name seeing touching

89.8 91.4 60.9

7.8 3. t 15.6

82.1 88.4 45.3

Movement

name seeing touching

88.3 93.8 88.3

1.6 0.8 4.7

86.9 93.1 83.8

interaction, F (2/42) = 3.88, 6O2 = .02. Further analyses showed that in the passive touching group recognition was more precise with questions focussing on local (TT) than on global features (WF), F (1/42) = 4.76, with means of 51.6 and 39.0, respectively. All other comparisons within the groups did not reach the level of significance, F _< 1.22. More importantly, there was a strong movement or performance effect, F (1/42) = 20.73, (52 = .11, an effect of study modality, F (2/42) = 20.21, 6O2 = .22, as well as an interaction between both variables, F (2/42)= 10.19, 6o2 = . 11. As can be seen from Table 3, recognition performance was - with one exception - generally excellent. The performance effect - that is, better recognition for subjects with hand movements at encoding - was pronounced and significant within the touching modality, F (1/42 = 39.88, contrasting active and passive touching conditions. The other two study modalities also showed a slight performance advantage. But, probably owing to ceiling effects, the differences turned out to be insignificant. Looking at the interaction in the other direction, study modality was uninfluential in the movement condition, F (2/42) = 1.21, but had a significant effect in the groups without hand movements at encoding, F (2/42) = 29.20, owing to the low discrimination scores of the passive touching group.

Discussion

The repetition priming data are consistent with previous results (Wippich & Wagner, 1989; Wippich, 1991). Reaction times in answering questions about old objects were reliably faster than answering the same questions about new objects for subjects who encoded the objects with the same procedures as in the test phase (active touching with hand movements). Wippich (1991) has provided evidence that repetition priming in haptic information processing depends crucially on the retrieval of motor components. In agreement with this evidence, subjects of the current investigation who had made their decisions about objects' properties without any hand movements at encoding did not show any effects of repetition priming. Even in the passivetouching encoding condition, in which the objects were placed on the palm of the hand, effects of repetition priming could not be detected, although subjects should have encoded some of the objects' relevant sensory features. Thus, the retrieval or the reenactment of hand movements in the test phase seems to play a prominent role in priming. On the other hand, subjects in the name or in the seeing

encoding conditions did not show priming effects, even when they had been instructed to perform adequate hand movements at encoding. We had speculated that haptic exploration in the seeing group, and especially in the name group, might be somewhat inadequate, owing to a greater uncertainty in the properties to be touched, and, thus, in the hand movements to be performed. Unfortunately, we could not find a gradual decline in repetition priming from the touching to the name groups in subjects instructed to use their hands at encoding. Thus, the radical proposal that priming depends only on the retrieval of encoded hand movements in the test phase cannot be substantiated with firm evidence. Priming may presuppose the establishment of sensory-motor records at encoding as in the active touching group. On the other hand, it remains possible that more adequate hand movements at encoding under either name or seeing conditions could be sufficent to induce priming effects. The extremely delayed responses within the name group, with hand movements at encoding, may indicate some sort of interference between exploratory movements at encoding and in the testing phase, corroborating the need to induce adequate movements at encoding. That is, name and seeing subjects should be trained in movements that are very similar to the exploratory hand movements to be expected with real objects when subjects are blindfolded. In support of the latter claim, attention may be paid to the data from the encoding phase. In the active touching group, reaction times to questions were faster for more global (weight or form) than for more local (temperature and texture) properties of objects. No other group showed this pattern of performance at encoding. In the test phase, however, the type-of-question effect was a general one. Thus, subjects in the name and seeing groups who had to perform movements failed to show the typical effect at encoding characteristic of active object exploration (Wippich, 1991). This substantiates the conjecture that adequate hand movements at encoding without any registration of sensory information about the objects may suffice to induce repetition priming. By the way, longer reaction times to questions about local properties in haptic information processing can be attributed to a stronger involvement of detailed finger movements in object exploration (Lederman & Klatzky, 1987). Up to now, we have talked about the effects of repetition priming in discussing the results and have avoided the term implicit memory, which was introduced by Graf and Schacter (1985) to denote a form of memory that can occur in the absence of recollective experience. Priming is the more neutral term and describes a specific facilitation in test performance that occurs in a situation that does not make explicit reference to a study episode. But does the observed repetition priming occur without any deliberate intention of recollecting the past episode and is it, thus, an expression of implicit memory? Since our subjects had been asked to recognize the objects handled, one may have doubts as to the implicit nature of the decision test about the objects' properties. Subjects may have made their decisions by remembering the answers to old (and recognized) objects. There is some evidence contrary to this conjecture. First, the proposed strategy of remembering previous an-

167 swers should have been available to all subjects. Thus, the pattern of results in priming performance restricting effects to the active touching group is not explained. Second, given the results o f the recognition test, one should expect a different pattern o f priming effects correlating with recognition performance if explicit m e m o r y mediates priming effects. Third, we reanalyzed the priming data, restricting the evaluation of reaction times to those questions that had been answered identically at encoding and at testing. The pattern of results did not change in any way. Besides, the lowest level of congruence in answers to the same questions about objects' properties was found in the passive touching group and the highest level was observed in the active touching group, with means o f 81% and 98%, respectively. The relatively low level of correspondence within the passive touching condition corroborates the view that passive touch is an impoverished and unreliable encoding device for identifying or recognizing objects (Lederman & Klatzky, 1987). The impaired recognition performance of this group is not surprising. On the other hand, active touch is an excellent encoding condition, as is demonstrated by the results of the recognition test. Thus far, the performance effect expected in recognition m e m o r y comes as no surprise when the two touching conditions are compared. More importantly, subjects in the name and seeing groups also showed excellent recognition results. Exploratory hand movements at encoding could do little to improve recognition. So we cannot state definitely whether the performance effect found in m e m o r y o f action events generalizes to the test conditions in the investigation. In any case, the results make it clear that recognition can be based on a variety of features. Verbal or visual encoding can be transferred to, and used in, recognition decisions, cued primarily with sensory and motor information at the testing stage. This documents a substantial discrepancy between forms of implicit and explicit m e m o r y as assessed by indirect and direct m e m o r y tests, respectively. According to our results, priming is restricted to test conditions that maximally resemble the conditions at encoding. A n explicit and above-chance j u d g e m e n t about the m e m o r y status of an item, on the other hand, is possible across a wide spectrum of different encoding conditions, including cases of cross-modal changes in presentation format from study to test. This attests the flexibility of explicit m e m o r y expressions and confirms the specificity o f priming effects. Some investigators have speculated that the enactment of action phrases adds an implicit m e m o r y component to study encoding (Nilsson & B~ickman, 1989). The results o f the current research on haptic information processing do not contradict this view. Whereas Wippich et al. (1990) showed that the motor component at encoding is revealed by a motoric test, this study demonstrates that exploratory hand movements at encoding are necessary (but, perhaps, not sufficient) preconditions for the detection of effects o f repetition priming in making decisions about objects' properties when subjects are blindfolded in the test phase. W e cannot be sure, however, whether motor information as isolated in indirect m e m o r y testings is the same as that retrieved in direct m e m o r y situations. In any case, measures o f explicit m e m o r y are based on at least something

more (if not other) than indirect measures of memory. This may be even more important in situations consisting of more detailed and complex motoric components such as the retrieval o f previously enacted action phrases.

References B~ickman, L., Nilsson, L.-G., & Chalom, D. (1986). New evidence on the nature of the encoding of action events. Memory & Cognition, I4, 339 - 346. Cohen, R. L. (1983). The effect of encoding variables on the free recall of words and action events. Memory & Cognition, 11,575- 582. Cohen, R. L., Peterson, M., & Mantini-Atkinson, T. (1987). Interevent differences in event memory: Why are some events more recallable than others? Memory & Cognition, 15, 109- 118. Eich, J. E. (1984). Memory for unattended events: Remembering with and without awareness. Memory & Cognition, 12, 105 - 111. Engelkamp, J. (1990a). Das menschliche GediJchmis. Grttingen, Hogrefe. Engelkamp, J. (1990 b). Memory of action events: Some implications for memory theory and for imagery. In C. Cornoldi & M. McDaniel (Eds.), Imagery and cognition (pp. 183 - 219). New York: Springer. Engelkamp, J., & Krumnacker, H. (1980). Imaginate and motorische Prozesse beim Behalten verbalen Materials. Zeitschrififiir Experimentelle und Angewandte Psychologie, 27, 511-533. Engelkamp, J., & Zimmer, H. D. (1983). Zum Einflug von Wahrnehmen and Tun auf das Behalten von Verb-Objekt-Phrasen. Sprache & Kognition, 2, 117- 127. Graf, P., & Mandler, G. (1984). Activation makes words more accessible, but not necessarily more retrievable. Journal of Verbal Learning and Verbal Behavior, 23, 553 - 568. Graf, P., & Schacter, D. L. (1985). Implicit and explicit memory for new associations in normal and amnesic subjects. Journal of Experimental Psychology: Learning, Memory, and Cognition, 1I, 501 - 518. Graf, P., & Schacter, D. L. (1987). Selective effects of interference on implicit and explicit memory for new associations. Journal of Experimental Psychology: Learning, Memory, and Cognition, 13, 45 53. Hayman, C. A. G., & Tulving, E. (1989). Contingent dissociation between recognition and fragment completion: The method of triangulation. Journal of Experimental Psychology: Learning, Memory, and Cognition, 15, 228-240. Helstrup, T. (1989). Loci for act recall: Contextual influence on processing of action events. Psychological Research, 51,168-175. Jacoby, L. L., & Dallas, M. (1981). On the relationship between autobiographical memory and perceptual learning. Journal of Experimental Psychology: General, 110, 306- 340. Jacoby, L. L., & Witherspoon, D. (1982). Remembering without awareness. Canadian Journal of Psychology, 32, 300-324. Lederman, S. J., & Klatzky, R. L. (1987). Hand movements: A window into haptic object recognition. Cognitive Psychology, 19, 342-368. Light, L.-L., & Singh, A. (1987). Implicit and explicit memory in young and older adults. Journal of Experimental Psychology: Learning, Memory, and Cognition, 13, 531-541. MacLeod, C. M., & Bassili, J. N. (1989). Are implicit and explicit tests differentially sensitive to item-specific vs. relational information? In S. Lewandowsky, J. C. Dunn, & K. Kirsner (Eds.), Implicit memory (pp. 159-172). Hillsdale: Erlbaum. Mitchell, D. B., & Brown, A. S. (1988). Persistent repetition priming in picture naming and its dissociation from recognition memory. Journal of Experimental Psychology: Learning, Memory, and Cognition, 14, 213-222. Nilsson, L.-G., & B~ickman, L. (I989). Implicit memory and the enactment of verbal instructions. In S. Lewandowsky, J. C. Dunn, & K. Kirsner (Eds.), hnplicit memory (pp. 173-183). Hillsdale: Erlbaum. Parkin, A. J. (1987). Memory and amnesia. Oxford: Blackwell. Richardsohn-Klavehn, A., & Bjork, R. A. (1988). Measures of memory. Annual Review of Psychology, 39, 475 - 543.

168 Roediger, H. L. (1990). Implicit memory: A commentary. Bulletin of the Psychonomic Society, 28, 373-380. Saltz, E., & Donnenwerth-Nolan, S. (1981). Does motoric imagery facilitate memory for sentences? A selective interference test. Journal of Verbal Learning and Verbal Behavior, 20, 322 - 332. Scarborough, D. L., Cortese, C., & Scarborough, H. S. (1977). Frequency and repetition effects in lexical memory. Journal of Ex-

perimental Psychology: Human Perception and Performance, 3, 1-17. Schacter, D. L., Cooper, L. A., & Delaney, S. M. (1990). Implicit memory for unfamiliar objects depends on access to structural descriptions. Journal of Experimental Psychology: General, 119, 5 -21. Schmidt, R. A. (1988). Motor control and learning: A behavioral emphasis. Champaign: Human Kinetics Press. Shimamura, A. P. (1986). Priming effects in amnesia: Evidence for a dissociable memory function. Quarterly Journal of Experimental Psychology, 38A, 619 - 664. Snodgrass, J. G., & Corwin, J. (1988). Pragmatics of measuring recognition memory: Applications to dementia and amnesics. Journal of

Experimental Psychology: General, 117, 4 - 50.

Tulving, E., Schacter, D. L., & Stark, H. A. (1982). Priming effects in word-fragment completion are independent of recognition memory.

Journal of Experimental Psychology: Learning, Memory, and Cognition, 8, 336-342. Wippich, W. (1991). Begreifen oder Ffihlen? Motorische und sensorische Prozesse bei Erinnerungen an haptische Erfahmngen. Archivfi~r Psychologie, 142, (in press). Wippich, W., Mecklenbr~uker, S., & Brausch, A. (1989). Implizites and explizites Ged~ichtnis bei Kindern. Bleiben bei indirekten Behaltensprfifungen Altersunterschiede aus? Zeitschrifi fiir EntwickIungs- und Piidagogische Psychologie, 21,294- 306. Wippich, W., Mecklenbr~inker, S., & Sidiropoulos, I. (1990). Kinder erinnem Handlungen: Bei impliziten Behaltensprfifungen bleiben Altersunterschiede aus. Schweizerische Zeitschrift fUr Psychologie, 49, 75-85. Wippich, W., & Wagner, V. (1989). Auch H~nde haben ein Ged~ichtnis. Sprache & Kognition, 8, 166- 177. Zimmer, H. D., & Engelkamp, J. (1985). An attempt to distinguish between kinematic and motor memory components. Acta Psychologica, 58, 81-106.