THE EFFECTS OF TEMPORALLY DISTORTED SENTENTIAL ST IMUU UPON PERFORMANCE OF NORMAL AND MENTALLY IMPAIRED INDIVIDUALS Dissertation for the Degree of Ph. D MICHIGAN STATE UNIVERSITY STEPHEN E. MOCK 1977 L [B R A R Y Michigan State University .- —‘v ‘r This is to certify that the thesis entitled THE EFFECTS OF TEMPORALLY DISTORTED SENTENTIAL STIMULI UPON PERFORMANCE OF NORMAL AND MENTALLY IMPAIRED INDIVIDUALS presented by Stephen E. Mock has been accepted towards fulfillment of the requirements for Ph.D. Audiology degree in Major professor Date AUgUSt 5, 1977 0-7 639 ABSTRACT THE EFFECTS OF TEMPORALLY DISTORTED SENTENTIAL STIMULI UPON PERFORMANCE OF NORMAL AND MENTALLY IMPAIRED INDIVIDUALS By Stephen E. Mock The purpose of this study was to investigate the effects of temporally distorted sentential stimuli on the performance of a group of normal subjects and a group of mentally impaired stubjects. The subject groups were composed of 30 normal hearing university students and 30 normal hearing mentally impaired individuals. The mentally impaired group showed an age range of 17-36 years and an Intelligence Quotient (I.Q.) range of 40-75 as measured by various psychological bat- teries. I The experimental speech stimuli consisted of ten three-word and ten five-word normal sentences, ten three— word and ten five-word second order sentential approximations, and ten three-word and ten five-word first order sentential approximations. These stimuli were time-compressed by 0%, 30%, and 60%. Each subject was presented all sentential orders of either the three-word or the five-word sequence at all three levels of time compression. All stimuli were Stephen E. Mock presented in a sound field at an intensity level of 75-80 dB SPL. The results of this study demonstrated that an inter- action of time compression, order of sentential approxima- tion and sequence length was capable of differentiating mentally impaired subjects from normal subjects. This dif- ferentiation was accomplished by viewing the number of recall errors between groups. The normal control group showed significantly better recall than did the mentally impaired subjects. In addition, an analysis of error type showed that more than 70% of all recall errors of the men- tally impaired group could be classified as reflecting memory constraints, while 25% of the recall errors of the normal group fell into these classifications. Thus, these results suggested that memory deficiencies play a role in mental impairment. Based upon the results of this study, speculations were made about possible memory strategies employed by the men- tally impaired and their possible relationships to past hypothetical findings in the area were explored. Suggest- ions for future research efforts that may provide additional information about the perceptual abilities of the mentally impaired were provided. THE EFFECTS OF TEMPORALLY DISTORTED SENTENTIAL STIMULI UPON PERFORMANCE OF NORMAL AND MENTALLY IMPAIRED INDIVIDUALS By Stephen E. Mock A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Audiology and Speech Sciences 1977 Accepted by the faculty of the Department of Audiology and Speech Sciences, College of Communication Arts and Sciences, Michigan State University, in partial fulfillment of the requirements for the Degree of Doctor of Philosophy. Guidance Committee: \.,\_.....<§~<‘._. k §M¥XQ Linda Lou Smith, Ph.D. Director of Thesis Leo V. Deal, Ph.D. Charles A. Tait, Ph.D. ACKNOWLEDGEMENTS The author wishes to express his deep appreciation to Dr. Linda Smith, dissertation director, Dr. Leo Deal, and Dr. Charles Tait, members of the dissertation committee. The help, encouragement, stimulation, and friendship they have offered, not only on this dissertation but throughout my Doctoral program, are things that I will always remember and cherish. I would also like to express appreciation and special thanks to Ms. V. A. Goodrich, Director of E. B. I. Breakthru, Inc., for all her help in providing subjects and testing fac- ilities so that this study could be completed. I further thank the supervisors and employees of E. B. I Breakthru who gave their valuable time and cooperation for the benefit of this study. I wish to acknowledge my parents for their encourage- ment and stimulation. Also I wish to thank Gary Lawson and Mike Flahive for all their ideas, assistance, critiques, etc., but perhaps more importantly for their friendships over these last few years. Lastly, but perhaps most importantly, a special thank you to Chris and Eric, my wife and son. They alone know the extent of hardships and sacrifices that a family can endure in order that a ”goal” may be attained. To them I offer my deepest love and gratitude. TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES INTRODUCTION Short Term Memory: Background . Short Term Memory: Temporal Aspects Short Term Memory Aspects of Mental Impairment . . . . . . . . Time Compressed Speech Short Term Memory: Sentential Aspects Statement of the Problem EXPERIMENTAL PROCEDURES Subjects . Design and Stimuli Recording and Stimulus Generation Procedures Presentation Procedures Analysis of Data RESULTS Effect of Subjects . Effect of Time Compression Effect of Sentential Order Effect of Sequence Length Type of Error Response DISCUSSION Subject Differentiation . . Short Term Memory: Mentally Impaired Summary and Conclusions Implications LIST OF REFERENCES iii Page .vi MN .10 .19 .21 .22 .24 .24 .25 .26 .28 .38 .38 .46 .52 .54 .56 APPENDICES Appendix A. Revised Word List Basal Vocabulary On We Go (Houghton-Mifflin) I1966I . . . . . . . . . . B. Sentential Word Lists C. Results of Individual Subjects iv Page .64 .70 Table II. III. IV. VI. VII. LIST OF TABLES Analysis of Variance Mean Mean error response of subjects per level of time compression (0%, 30%, 60%) and sen- tence order (Normal sentences, 2nd Order, lst Order) for three-word sequence error response of subjects per level of time compression (0%, 30%, 60%) and sen- tence order (Normal sentences, 2nd Order, lst Order) for five-word sequence . Number error responses of normal subjects per type of error (discrimination, omission, addition, reversal, distortion) level of time compression (0%, 30%, 60%) and senten- tial order (Normal, second order, first order) for sequence length of three words Number error responses of mentally impaired subjects per type of error (discrimination, omission, addition, reversal, distortion), level of time compression (0%, 30%, 60%), and sentential order (Normal, second order, first order) for sequence length of three words . . . . . . . . . . . . . . Number error responses of normal subjects per type of error (discrimination, omission, addition, reversal, distortion), level of time compression (0%, 30%, 60%), and sen- tential order (Normal, second order, first order) for sequence length of five words. Number error responses of mentally impaired subjects per type of error (discrimination, omission, addition, reversal, distortion), level of time compression (0%, 30%, 60%), and sentential order (Normal, second order, first order) for sequence length of five sords . . . . . . . . . Page .31 .33 .35 .35 .35 .36 .36 LIST OF FIGURES Figure Page 1. Mean percentage error scores of normal and mentally impaired subjects for three and five word sequences . . . . . . . . . . . . . . . . .41 2. Mean percentage error scores of normal and mentally impaired subjects for three levels of time compres- sion (0%k 30%, 60%) . . . . . . . . . . . .42 3. Mean percentage error scores of normal and mentally impaired subjects for three levels of sentential order . . . . . . . . . . . . . . . . . . .44 4. Mean percentage of error response of normal and mentally impaired sub- jects according to type of error (Discrimination, Omission, Addi- tion, Reversal, Distortion) . . . . . . . .45 vi INTRODUCTION Since the publication of Galton's "Notes on Prehension in Idiots" (1887), researchers and scholars have been inter- ested in investigating how deficiencies in perceptual aspects of learning are related, at least in part, to behavioral inadequacies of the mentally impaired. Although such com- ponent areas as attention, experience, motivation, and memory have all been explored in relation to the mentally impaired, a dearth of information still exists. Several researchers (Calearo and Lazzaroni, 1957; Beasley, Schwimmer and Rintel- mann, 1972; Calearo, 1975) have indicated the important role played by temporal aspects in speech perception. One of the most recent developments in this area of time alteration is time compression. Time compressed stimuli reduce both the redundancy and the duration of the speech signal by discard- ing random segments of the verbal stimuli. In addition to these effects, time compression serves to reduce both semantic and syntactic content (Beasley, Schwimmer and Rin- telmann, 1972). This reduction can be further increased by the use of sentential order approximations. Previous stud- ies in the area of time compression have obtained response data for normal young adults (Beasley, Schwimmer and Rintel- mann, 1972; Beasley, Forman and Rintelmann, 1972), children 2 (Beasley, Maki and Orchik, 1976), geriatric populations (Konkle, Beasley and Bess, 1974), hearing impaired (Kurdziel and Noffsinger, 1973; Kurdziel, Rintelmann and Beasley, 1975), and less disabled children (Freeman, 1976). In view of these findings, an attempt to utilize time compressed stim- uli with the mentally impaired seemed warranted. Therefore, the purpose of this investigation was to study, clarify, and enhance some perceptual aspects of mental impairment by means of the utilization of time compressed sentential approxima- tions to full grammaticality. Short Term Memory: Background The process of short term memory has been described by several researchers within the past three decades (Hebb, 1949; Broadbent, 1958; Pollack, 1959; Aaronson, 1967; Atkinson and Shiffrin, 1971; Fisher and Zeaman, 1973). These descriptions were divided into two categories: a physiological approach and a behavioristic approach. The physiological category is primarily based upon an assumption proposed by Hebb (1949). This assumption states that memory is derived from a struc- tural change within the organism that occurs over time. In order for this structural change to occur, however, it was necessary to theorize how memory information could be main- tained prior to the structural change. Hebb postulated the presence of a memory system that was based upon past experi- ences. This memory basis was termed an activity trace. This trace could be interrupted by external events and thus was subject to decay. The activity trace was thought to 3 represent a short term memory system. If the trace were not allowed to be interrupted by outside events and if it were thus stabilized over time, a structural memory trace could be established. This structural memory trace was comparable to a long term memory system in which information could be retained and brought again to the conscious level if necessary. The behavioristic hypothesis is primarily based upon Broadbent's (1958, 1963) distinctions between temporary and long term memory stores. The information flow through the memory system was thought to begin with environmental inputs that were picked up by a sensory register (sensory receptors plus neural processes). This information was then transfer- red to a short term memory channel. Short term memory has been described as a working memory system wherein decisions are made, problems are solved, and information flow is con- tinued.. The important aspects of this system could, there- fore, be described as the perception, retention, and retrie- val of information. Following the completion of short term processing, information transfer to the long term memory system may be completed. Aaronson (1967) reported perceptual aspects to occur in two stages for short term memory recall tasks. Stage I pro- cessing involved the presence of a large capacity perceptual storage system in which unidentified sensations or attri- butes of physical stimuli were subject to relative instabil- ity, rapid decay, and parallel processing. Stage II, which 4 was thought to occur at a higher cortical level than Stage I, was characterized by slower decay and a limited storage capacity having a "series" input. Miller (1956) suggested the limits of capacity available within short term memory to be 712 units of information. This hypothesis suggested that whenever new material entered the short term or primary memory channel, old information had to be either transferred into a long term memory channel or forgotten. It should be noted, however, that the short term memory capacity could be enhanced by the process of "chunking." This process permitted new informational units to be coded into groups prior to entry into the primary memory channel. According to Atkinson and Shiffrin (1971), once the information had been processed into the short term memory channel, it was acted upon by a series of control processes. These processes included rehearsal, coding, decision, and retrieval strategies. Rehearsal, an overt or covert repe- tition of information, increased the strength of the infor- mation in short term memory so that it could be held for a longer period of time in the primary memory channel or trans- ferred into long term memory. Without rehearsal, information would decay quickly and thus be forgotten (Waugh and Norman, 1965). Following the rehearsal stage, the information had to be either used immediately, stored in long term memory, or forgotten (Norman, 1966). If the information were to be stored in long term memory, however, it had to be coded to go along with the information that had previously been stored there. If the information were to be used immediately, the 5 previously coded information necessary to make a decision must first have been pulled from long term memory. Again, if the new information were not relevant to the decision, it would be forgotten (Lindsey and Norman, 1972). The final control process proposed by Atkinson and Shiffrin concerned retrieval processes. These processes have been related to certain effects that might occur. Feig- enbaum and Simon (1962) reported the presence of the recency and primacy effects within short term memory retrieval. In the recency effect, the latest information entered into the primary memory channel would be best recalled. If, on the other hand, the information needed involved earlier presented units, the recency effect would diminish and be replaced by a primacy effect. This effect involved the initially pre- sented informational units. Brown (1958) found recall to be inversely proportional to the amount of information to be recalled. In comparison to short term memory, the long term memory channel involves a permanent memory store of unlimited capa- city (Lindsay and Norman, 1972). In order for a unit of information to be entered into long term memory, it must initially be rehearsed and coded in short term memory. Re- call from long term memory involves a retrieval process that may allow thoughts to be brought to consciousness only after a sometimes laborious memory search. Short Term Memory: Temporal Aspects According to Broadbent (1958), the short term memory channel is affected by such temporal factors as stimulus 6 duration, interstimulus interval, and rate of presentation. In a review of these variables that may affect information processing, Aaronson (1967) reported discrepancies within the findings of previous researchers. Fraser (1958), Conrad and Hille (1958), and Posner (1963) reported that a reduc- tion of the presentation rate allowed more time for stimulus decay and thus resulted in lower recall accuracy. Studies by Pollack, Johnson, and Knaff (1959) and Pollack and Johnson (1963) were in disagreement with these findings. Their results showed a higher recall accuracy when a slower rate of presentation was utilized. This increased accuracy with decreased presentation rate was thought to result from in- creasing the time utilized for perceptual activities. Aaron- son (1967) reasoned that these conflicting results probably were not due to such things as experimental procedures or type of stimulus material, rather that the discrepancy resulted from the duration of the stimuli utilized in the aforementioned research efforts. That is, experimental results were probably secondary to the stimulus durations and interstimulus intervals utilized within the experiments. Aaronson reported that stimulus duration determined the amount of information entering the short term store, whereas the interstimulus interval affected the time available for identification or encoding of the stimuli. The importance of the interstimulus interval had previously been reported by Aaronson and Sternberg (1964). These authors, in holding word duration constant and varying the interstimulus interval, 7 found an increase in the presentation rate over time to result in a decrease in recall accuracy. This decrease was thought to be related to a reduction in the time available for information processing. In a follow-up to the Aaronson (1967) report, a study by Aaronson, Markowitz, and Shapiro (1971) attempted further to clarify the effects on perception resulting from changes in stimulus duration and interstimulus interval. These auth- ors varied the "speech-to-pause time" in three different experiments and compared subject response to seven digit sequences for recall, item, and order information. Results of the study indicated that when 33% of the speech signal was removed and substituted by an equal amount of silent interval, recall accuracy was significantly higher. These results suggested that cumulative perceptual delays in encod— ing item and order information may result when adequate pause time is not available. Shriner and Daniloff (1970) provided additional infor- mation concerning the importance of the interstimulus inter- val in auditory perception. Using both meaningful and nonmealingful stimuli with first and third grade children, Shriner and Daniloff demonstrated that correct responses significantly decreased when the silent interstimulus inter- vals were increased. This finding became more apparent when meaningless stimuli were used. These results thus demonstra- ted that the more syntactical and semantic system components that were available, the more facilitated was short term 8 memory recall. Optimum recall was found to occur when a short interstimulus interval was utilized with meaningful speech material. The relationship between word duration and interstim- ulus interval was further investigated by Beasley and Shriner (1973). Using temporally manipulated first and second order sentential approximations with normal young adults, these researchers showed that the number of items correctly re- called increased as stimulus duration increased. The size of the interstimulus interval and the grammatical aspects of the stimulus materials were also judged to be important fac- tors. The authors pointed out, however, that although stim- ulus duration was thought to play a more significant role in perception than interstimulus interval, a complex interaction between the two factors should discourage their being studied on an independent basis. In view of the suggested interaction between auditory perceptual processing and short term memory, a study by Beasley and Flaherty-Rintelmann (1976) further attempted to determine the importance of the silent interstimulus inter- val in an auditory recall task with second and fourth grade children. Because digits and word lists were felt to be inadequate for assessing functions of short term memory (Bocca and Calearo, 1963; Pollack, 1967), semantically mean- ingful materials were utilized. The stimuli were ten normal sentences and ten three-word and ten five-word sentential approximations of both the first and second order. Results 9 indicated that recall performance decreased in all sentential conditions as the interstimulus interval was increased. Gram- matical approximation, sentence length, and grade level of the child were also found to be important considerations in auditory processing. These results thus served to confirm the findings of previous research efforts of Beasley and Shriner, 1973; Speaks and Jerger, 1965; Miller, 1956; Panta- los et al., 1975; and Smith, 1972. Pantalos, Schuckers, and Hipskind (1975) studied sen- tence recall with preschool children. Using four- to eight- word sentences that were either of normal duration of 30% time compressed with varied interstimulus intervals (200 or 1000 msec), the suthors found recall performance decreased as the number of words in the sentence increased. Also, the limit of the preschool child's memory capacity appeared to be six or seven words. The processing of any sentential material exceeding this length was though to be related to the grammatical approximation of the material and to the child's language competence. A study to determine the ability of children to recall unrelated items in a sequence was completed by Bisset and Koshey (1975). These authors utilized two, three, four, and five word sequences at 0% and 30% time compression in con- junction with three different (normal, compressed, expanded) interstimulus intervals. Results of the study indicated the highest scores were obtained at 30% time compression for normal and expanded interstimulus intervals. These results supported both Aaronson's suggestion that the interStimulus 10 interval is important for perceptual activity and Kirk's (1966) contention that a child's performance in sequencing unrelated items will be enhanced by an increased stimulus presentation rate. In summary, it is apparent from previous research that temporal factors such as stimulus duration, interstimulus interval, and rate of presentation play key roles in the processing and perception of informational units. Although these research efforts have shown some disagreement relative to the individual importance of each factor, all seem to agree that recall accuracy can be both affected and influ- enced by temporal alteration within the presented material. The greated the amount of temporal alteration, the greater the effect upon accuracy of response. Short Term Memory Aspects of Mental Impairment Deficiencies within the short term memory channel have frequently been utilized as an explanation of behavioral inadequacies within the mentally impaired. Jensen (1964) reported that the ability to hold information within short term memory appears to be a critical skill in many, if not all, learning and problem solving situations. Thus, an indi- vidual possessing an impaired primary channel may lose a large amount of incoming information prior to its processing into the long term memory channel. The result would be to greatly reduce the probability that a specific informational unit had become permanently stored. Experimental studies involving short term memory systems of the mentally impaired 11 have primarily been based upon five theoretical interpreta- tions relative to the nature of perceptual deficiency. l. Broadbent (1958) proposed a model of memory in which the senses are constantly bombarded by environmental stimuli of which only a limited amount may be processed. Some of this incoming information enters a short term store known as the S system where it may be held for a matter of seconds before being passed by a selective filtering system into a limited capacity channel called the P system. The greatest factor that determines which information is passed into the P system is the individual's past experience. From the P system the filtered information can be either passed dir- ectly into long term memory, or it can be rerouted via a feedback loop into the S system in order that the amount of rehearsal necessary for retention may be completed. It has been hypothesized by Scott and Scott (1968) that any defic- iency within either the S or P system would result in a gen- eral learning deficiency that could be applied to mentally impaired populations. Therefore, Broadbent's theory seems to predict that either the mentally impaired possess an im- paired 5 system or that the capacity of the P system may be more restricted than in normals, such that mentally impaired individuals are able to attend to less incoming information than are persons of normal intelligence. Broadbent's idea was further expanded upon by Zeamon and House (1963) and Fisher and Zeamon (1973) who proposed that the mentally im- paired's attention to a stimulus might well be a critical factor in determining primary memory capacity. 12 2. Ellis (1963) proposed a "Stimulus Trace" theory as a means of explanation of "sub—normal intelligence." Ellis described a stimulus trace as "an explanatory mechanism to account for immediate memory." His ancillary hypothesis was that the apparent learning deficiencies of the "retarded" were the result of a noncontinuity of events that were sec- ondary to dysfunction within the short term memory channel. As the basis for his theory, Ellis primarily cited the words completed by Kohler (1929) and Hull (1952). Kohler invoked a "fading trace" theory as a means of accounting for negative time error in psychophysical judgment. He reported that if two stimuli of equal dimensions were compared one after the other, the second would be judged as greater. This effect was thought to result from a comparison of the trace of the second stimulus to the fading trace of the initial sensation. The longer the time interval between the two experiences, the greater the error. In addition, Hull felt the stimulus trace acted as a basis of learning in that behavioral acts of the present are related to the traces of preceding events. On this basis, behavioral sequences that are considered inap- propriate can be eliminated because of a lack of reinforce- ment, whereas appropriate or reinforced behaviors are conditioned within the organism. Because of the lack of ingrained past experiences within mentally impaired popula- tions, Ellis proposed that the strength and duration of the stimulus trace were probably diminished within this group. He further speculated that short term memory function showed a developmental trend in that the stimulus trace increased 13 in both strength and duration as a function of age and intel- ligence. Ellis was thus able to predict a difference in retention both as a function of Mental Age (MA) and Intel- ligence Quotient (IQ). By pointing to the probable role of memory in a number of tasks, Ellis reported the performance of the mentally impaired to be below that of normals of com- parable chronological age in such areas as delayed response, reaction time, paired-associate learning, and simple reten- tion. 3. One of the most prominent theories of forgetting in general psychology is the Interference theory (Underwood, 1957; Postman, 1963). The basic premise behind this theory is that one or more stimuli may be associated with one or more responses. When this association occurs, recall of the multiple associations may result in a response competition from which errors of recall and forgetting may arise. The theory also predicts that errors will vary as a function of the similarity of the competing material and as a function of the number of successive items to be recalled. Develop- mentally, it might be expected that children would become more tolerant to interference as a function of both age and intelligence. 4. A Neural theory presented by Spitz (1966) proposed that the mentally impaired are characterized by deficits within their nervous systems that include underlying brain structures and processes. The basis of the theory was pre- sented via four postulates: (l) in the mentally impaired, 14 more time is required to induce both temporary and permanent changes in stimulated cortical cells; (2) once temporary changes are induced, more time is required to return to home- ostasis; (3) it is more difficult to form new, or different, behavior patterns following permanent cortical cell change; and (4) in the mentally impaired there is less spread of electrochemical activity between stimulated cells and the surrounding cortical field. The implications from this theory are twofold. Since the mentally impaired are less likely than normals to organize efficiently, less total information can be permanently stored. On the other hand, because the cortical system of the mentally impaired is less resistant to change than the normal, any incoming information that has reached long term memory should be less susceptible to extinc- tion or interference. It was therefore proposed that, rela- tive to the amount of information processed into long term memory, the information retrieval system of the mentally impaired should be superior to that of the normal. 5. The last major theory relating to short term memory deficits in the mentally impaired is the Verbal Dysfunction theory of Luria (1963). This theory of learning disability in the mentally impaired differs in some important aspects from previously mentioned theories. In Luria's view mental retardation is thought to be the consequence of nervous sys- tem pathology, although in many instances the nature of the lesion has been admittedly difficult to specify. Luria pro- posed that the presence of such a lesion would produce a pathological weakness of the basic nerve processes such that 15 connections between nerve fibers would be acquired slowly, would respond inappropriately, and/or would be easily dis- rupted by extraneous stimuli or fatigue. The extent of the mental handicap was thought to be directly related to the size and location of the lesion. Luria expressed little opti- mism for remediation of the individual showing profound involvement. However, when the individual's verbal system remained intact, it was felt that a great deal could be done to rectify the mental defect. For those individuals, treat- ment consisted of attempting to compensate for the neurody- namic defects by bringing the motor reactions under the regulation of the verbal reactions. Therefore, according to Luria's thinking, the key to educating the mentally retarded lies in (l) the diagnosis of the defect, (2) enlargment of the vocabulary, and (3) the forced verbal coding of materials. Although much has been written in regard to agreement or disagreement with the previously mentioned theories (Kel- las and Butterfield, 1971; Waugh and Norman, 1965), few studies have been successful in adequately comparing the memory performance of the normal versus the mentally impaired (Butterfield, Wambold, and Belmont, 1973). Also, with the possible exception of Ellis and his associates (1968, 1968, 1969, 1969) who have attempted with great endeavor to justify the stimulus trace rationale, few practical efforts to prove or disprove the theories have been attempted. O'Connor and Hermelin (1965) attempted to differentiate between the Broadbent and Ellis theories by presenting l6 simultaneous or successive visual displays of digits at different rates. Both a normal and a sub-normal group were utilized. Results showed that normals performed better than the mentally impaired group for the simultaneous condition at fast rates, whereas both groups were comparable at slow simultaneous rates and for the successive condition. The authors interpreted their results as support for Broadbent's hypothesis of input restriction because of limited capacity. Two predictions derived from Ellis' stimulus trace theory were tested by Hayes and Routh (1972). Because Ellis proposed that both the amplitude and the duration of the stimulus trace were diminished in the retarded, it was assumed that by increasing the intensity and the duration of the environmental stimulus, retardate memory should improve. Ellis (1963) had previously stated that "...any state of affairs (drugs, increase in intensity of stimuli, etc.) which increases the duration of the stimulus trace should facilitate learning." Some evidence of performance enhance- ment of the retarded under high intensity stimulus presenta- tion had also previously been reported (Kouw, 1968; O'Connor and Hermelin, 1963). Hayes and Routh (1972), therefore, varied the length of recall interval (0 vs. 8 seconds) and the intensity level of the aurally presented items (55 vs. 90 dB H.L.) on a task requiring oral recall of five non-re- peated 4-digit messages. Results found that the two predic- tions considered to be crucial to the support of the Ellis Theory were not confirmed. That is, the retardate's perfor- mance was neither more adversely affected than that of the II 17 normal's by increasing the retention interval nor differ- entially facilitated by an increase of the stimulus intensity. Other challenges to the Ellis Theory include Ebenholts, 1963; Jensen and Rohwer, 1965; and Butterfield, 1968. Butterfield, Wambold, and Belmont (1973) reported that the poor short term recall of the mentally retarded could be the resultant effects of deficiencies in such areas as imper- fect learning, poor retention, or incomplete retrieval. If this were the case, these short term memory deficiencies would therefore be related to difficulties in acquisition, storage, and retrieval of information. The authors conducted three different experimental tasks in order to evaluate each of these processes of short term memory in the retarded. The results indicated that although retarded individuals can com- petently utilize all of the individual component processes that are necessary for accurate recall of stimulus events, there is a lack of expertise in the area of correct sequenc- ing of these processes. Unless training methods can be implemented to "teach" the retarded adequate sequencing pro- cedures, there could be little chance of substantial improve- ment in short term memory performance. These findings are perhaps best related to the Neural Theory of Spitz (1966), who reported little capacity for systematic change within the memory system of the mentally impaired individual. Spitz (1966) reported that spatial groupings of visually presented digits had been found to increase the channel cap- acity for information processing in retarded adolescents. 18 The positive effects of groupings were thought to be related to the retarded individual's failure to spontaneously organ- ize material as efficiently as non-retarded subjects. In view of Miller's (1956) findings that subjective organiza- tion of stimuli (chunking) increases channel capacity for memory, Spitz proposed that his research evidence indicated that this organizational strategy was not efficiently em- ployed by the retarded. However, a study by MacMillan (1972) failed to replicate Spitz's findings in that no significant improvement in recall of retarded subjects was found via the spatial grouping of digits. Despite the discrepancy of the findings in the two studies, MacMillan felt that his results only further clarified the idea of inefficient use of stra- tegy development in the retarded. As previously presented, Luria (1963) indicated that much emphasis must be placed upon language development and upon relating verbal behavior to motor behavior. While he did not suggest that an enlarged language system alone was sufficient to overcome the abnormality of mental function, he did imply that it appeared to be a necessary condition for cognitive development. In short, Luria felt the men- tally defective's possession of a reduced capacity to use symbols could be greatly enhanced when the retarded subject could be taught to mediate his motor responses verbally. Luria's thinking was challenged on two points by O'Connor and Hermelin (1963) and Rosen, Kevits, and Rosen (1965). The former study pointed out that not only do retardates 19 usually possess a small vocabulary but they are also some- what reluctant to use what they do possess. Rosen et al. presented an even greater assault to the Luria hypothesis when they failed to find any substantial "dissociation" between the verbal and motor systems of a mentally impaired population. Time Compressed Speech Inview of the role that temporal factors play in the perception of speech, time compressed stimuli have been in- creasingly utilized as a means of examining aspects of audi- tory processing. As presently employed, the method of time compression randomly discards segments of speech stimuli in order to reduce both the redundancy and the duration of that stimuli. Jerger (1960) prOposed a rationale for the presen- tation of time compressed stimuli based on a ”Subtlety Prin- ciple.” Jerger stated that "the site subtlety of the auditory manifestation increases as the site of lesion pro- gresses from peripheral to central" and therefore that it would be necessary to increase the difficulty of the auditory task in order to evaluate adequately the higher auditory pathways. Although such procedures as filtered or interrup- ted speech have previously been utilized in central auditory testing, Calearo (1975) reported that time c6mpressed speech is presently the most effective means of evaluating auditory processing at the cortical level. Although early studies assisted in the development of new and more efficient methods of time compression and 20 provided some insight into its uses, it was not until 1972 that adequate normative data were obtained. In that year, a study by Beasley, Schwimmer, and Rintelmann provided data about 96 normal hearing adults at six different levels of time compression (0, 30, 40, 50, 60, 70%) and four different sensation levels (8, 16, 24, and 32 dB SL). The four word lists of Form B of the Northwestern University Auditory Test #6 (N.U. #6) were utilized as the speech stimuli. These data were further expanded by Beasley, Forman, and Rintelmann (1972), who gathered additional normative information at a 40 dB sensation level. These investigations provided a basis for comparative study with adult subjects suffering from auditory pathology. Since publication of these studies, additional data on time compressed speech have been provided for children (Beasley, Maki, and Orchik, 1976; Maki, 1975; Shoup, 1975), geriatric populations (Konkle, Beasley, and Bess, 1974), organic pathologies of the peripheral and cen- tral auditory pathways (Kurdziel and Noffsinger, 1973; Kurd- ziel, Rintelmann, and Beasley, 1975), and learning disabled children (Manning, Johnson, and Beasley, 1975; Freeman, Beas- ley, and Overholt, 1975; Freeman, 1976). Also, studies by Bratt (1975) and Konkle (1976) have provided further norma- tive data concerning the use of time compressed stimuli in sentential approximations and as a contralateral masker. A review of the results of these studies indicates support of the notion that time compressed speech stimuli are diagnos- tically important for delineating central auditory dysfunction. 21 Short Term Memory: Sentential Aspects The relationship between language skills and memory processes seem closely linked. In this regard, several researchers have proposed that syntactic and semantic com- ponents of an utterance facilitate the recall process. Conrad (1962) and Miller and Isard (1963) were probably the first to point out that intrusion errors in the short term memory channel may be reduced when speech and languagelike inputs are utilized. Schulman (1971), on the other hand, postulated that semantic coding in short term memory, though possible, was a time consuming process and consequently employed only when the task demanded it. Schulman further pointed out, however, that "the effects of semantic and lin- guistic factors on immediate recall have not been extensively studied." Despite the controversy, several authors (Savin and Perchonock, 1965; Craik and Lockhart, 1972; Baddeley, 1972; Craik, 1973; Wetherick, 1975) have supported the contention that short term memory utilizes syntactic and semantic cues. Perhaps some of the soundest and most practical findings came from an observation of Savin and Perchonock (1965) who stated that "English sentences are much easier to recall after a single hearing than are equally long strings of random English words." The authors further suggested that the use of syntax and semantics was related to the process of "chunking" and that the capacity of short term memory would thus be enhanced through their utilization. 22 Statement of the Problem Evidence from the literature gives strong support to the idea that auditory perception is a temporally based phenomenon and that the utilization of temporally distorted speech material shows great potential for the measure of central auditory processing abilities. Research also shows that deficiencies associated with the mentally impaired have often been characterized as the resultant effects of inad- equate utilization of the short term memory channel (Broad- bent, 1958; Ellis, 1963; Spitz, 1963; Scott and Scott, 1968). In view of Smith's (1972) statement that "Behavior associ- ated with language and perception may be developmentally bound and (that) such developmental behavior may very well be related to performance characteristics of short term memory," it can be seen that the employment of a temporally distorted stimulus is a justifiable means of measuring short term memory capacity. This is especially true if semantic, syntactic, and intonational cues remain relatively intact (Tejerian, 1968; Scholes, 1969; Carrow and Mauldin, 1973). Because of the dearth of knowledge that is presently available concerning the processing capabilities and capa- cities of the mentally impaired, there exists a need for an investigation into this area. With this in mind, the pur- pose of this study was to determine the comparative perfor- mance between mentally impaired individuals and individuals of normal intelligence through the utilization of temporally distorted speech stimuli. Specifically, this investigation will attempt to: (1) (2) (3) (4) 23 Determine whether time compressed sentential stimuli can be utilized to differentiate normal and mentally impaired populations. Determine whether short term recall in normal and men- tally impaired populations is affected by changes in temporal aspects of speech stimuli. Determine whether short term recall of sentences in normal and mentally impaired populations is affected by semantic and syntactic aspects of speech. Determine whether differences in short term recall between normal and mentally impaired individuals are the result of discrimination errors or memory constraints. EXPERIMENTAL PROCEDURES Subjects Two groups of subjects participated in this study: (1) 30 normal hearing university students enrolled at Mich- igan State University; and (2) 30 normal hearing mentally impaired individuals selected from E. B. I. Breakthru, Inc., a sheltered workshop for the mentally impaired located at Lake Odessa, Michigan. The mentally impaired group showed a chronological age range of 17-36 years and in Intelligence Quotient (I.Q.) range of 40-75 as measured by various psy- chological test batteries. Prior to participation in the study, each of the subjects was required to pass a hearing sensitivity screening via pure tone air conduction tech- niques. This screening was administered at the level of 20 dB H.L. (ANSI, 1969) at octave intervals of 250 through 4000 Hertz. In addition, each of the subjects within the mentally impaired population was required to pass a picture articulation test. Passing criteria were defined as no more than three articulation errors. Individuals showing more severe speech problems such as stuttering, voice disorders, etc. were categorically eliminated prior to commencement of the study. 24 25 Design and Stimuli The experimental stimuli consisted of three experimen- tal tapes: 1. Ten three-word and ten five-word first order sentential approximations, time compressed by 0%, 30%, and 60% (Appendix B). 2. Ten three-word and ten five-word second order sentential approximations, time compressed by 0%, 30%, and 60% (Appendix B). 3. Ten three-word and ten five-word normal senten- ces, time compressed by 0%, 30%, and 60% (Appendix B). The sentential approximations to full grammatical sen- tences were ordered in a manner described by Beasley and Flaherty-Rintelmann (1976) and Speaks and Jerger (1965) and later modified by Freeman (1976). These sentential approxi- mations were composed of 100 monosyllabic words selected from the Basal Vocabulary of On We Go (Teacher Eddition, Houghton Mifflin, 1966), a primary reading workbook (Appen- dix A). The selection of monosyllables was based upon the findings of Massaro (1972) who reported the syllable to be the most important unit for auditory processing. For the first order sentences, the words were randomly chosen for placement within three and five-word approximations. The second order sentential approximations were compiled in the following manner: (1) the first word of the approximation was chosen at random; (2) a second individual was asked to 26 choose a word from the master list of monosyllables that could follow the first word; and (3) a third individual, having no knowledge of the initial monosyllable selected, was asked to choose a word that might follow the second word. This process was continued until all the second order sen- tential approximations were constructed. A similar process was used to deve10p the normal sentences. In this case, however, all previous individual selections were made known to those selecting the monosyllables. Recording_and Stimulus Generation Procedures All of the stimuli used in this study were recorded onto a master magnetic tape via a tape recorder (Ampex, Model AG 440B) at a speed of 7% inches per second following the procedure utilized by Freeman (1976). A complete des- cription of the stimulus generation procedure is provided by Beasley and Flaherty-Rintelmann (1976). Following completion of the master tape, a cassette copy was manufactured from the same recorder through the utilization of an Advent, Model 201, cassette tape recorder. This cassette tape was then processed through an electrical time compressor/expander (Lexicon, Model Varispeech I) at compression speeds of 0%, 30%, and 60%. This tape was pro- cessed to the desired levels of time compression through the utilization of a method proposed by Konkle, Freeman, Riggs, Riensche, and Beasley (1975). Final experimental tapes were recorded via the Ampex AG 440B tape recorder that was con- nected to the time compressor/expander. 27 Copies of the experimental tapes allowed a ten second interval between each of the three-word and five-word sen- tential approximations. The carrier phrase, "Number _____J" preceded each of the sentential stimuli. Presentation Procedures Subjects were divided into the two experimental groups. Each subject was then presented either ten three-word or ten five-word sentential approximations of the first order, ten three-word or ten five-word sentential approximations of the second order, and ten three-word or ten five-word normal sen- tences. Each of these stimulus conditions were time compres- sed by 0%, 30%, and 60%. In order to avoid possible order effects, stimuli presentation was counterbalanced with regard to rate of time compression. All subjects were tested individually in a quiet room (Ambient noise = 60 to 68 dBC using a Bruel and Kjaer Type 2203 sound level meter) via a high quality tape recorder (Ampex 600). The intensity level for the presentation of the stimuli was set at 75 to 80 dB SPL. Prior to the begin- ning of the study, a calibration check was carried out on all equipment in order to insure that ANSI (S 3.6 - 1969) specifications were met. These specifications were period- ically monitored throughout the study in order to insure instrument stability. Upon seating a subject within the test room, the follow- ing verbal instructions were given: "I want you to listen to what the man is saying on the tape recorder. He will say 28 a number and then some words. When the man stops speaking, I want you to tell me exactly what you heard him say. If you can't remember everything, say as much as you can remem- ber. Do you have any questions?" These instructions are similar to those employed by Beasley and Flaherty-Rintelmann (1976). Prior to the presentation of each time compressed condition, two practice items of either three-word or five- word first order sentential approximations were presented. Analysis of Data Subject response was recorded both manually and on tape during the experimental session. These responses were then evaluated with regard to accuracy of recall. Accuracy of response determination was further enhanced through the util- ization of an additional judge. This judge, a Ph.D. student in Audiology with more than five years of professional exper- ience, re-checked the tape recorded response. In cases of disagreement between the judges, the tape was replayed and a joint decision was made. Determination of error was made with regard to word discrimination, word order, additions, omissions, and distortions according to the following criteria: 1) A discrimination error occurred when a subject's res- ponse was incorrect but Shared similar acoustic proper- ties with the word stimulus. 2) An omission error occurred when one or more words within the stimulus word string was not reported by the subject. 29 3) An error of addition occurred when a word not present within the stimulus word string was reported by the subject. 4) A reversal error occurred when the subject reported an incorrect word order. 5) A distortion error occurred when memory constraints resulted in the subject's reporting only a portion of a word within the word string. Discrimination errors were categorically interpreted as errors of the perceptual system, whereas omissions, addi- tions, reversals, and distortions were thought to reflect memory constraints (Conrad, 1962). The number of items incorrectly recalled was the score for each subject. The data were then subjected to a four factor (2x3x3x2) analysis of variance with repeated measures. A Duncan Multiple Range Test was tuilized as a post hoc statistical measure. RESULTS The total number of incorrect items reported for each condition was determined and a computerized analysis of variance was performed. Statistically significant effects at the 0.05 level were found for the four main factors of subjects, sequence length, time compression, and sentential order. In addition, several significant interactions among these factors were demonstrated. The complete results of the analysis of variance can be found in Table l. A Duncan Multiple Range Test was utilized to investigate further the significant main effects and interactions. The results of this study thus demonstrated that populations possessing normal intelligence and mentally impaired populations can be differentiated by imposing semantic, syntactic, and temporal constraints upon sequential stimuli. Effect of Subjects The analysis of variance showed a significant main ef- fect of subjects (F = 245.2, df = l, p =«<0.0005) in that the mentally impaired individuals exhibited greater difficul- ties in recalling sequential stimuli than did the subjects of the normal group. These difficulties were further exhib- ited by significant interactions between subjects and se- quence length (F = 49.8, df = l, p = (0.0005), time 30 31 TABLE 1 ANALYSIS OF VARIANCE SOURCE Subjects Sequence Length Time Compression Sentential Order Subjects X Sequence Length Subjects X Time Compression Subjects X Sentential Order Sequence Length X Time Compression Sequence Length X Sentential Order Time Compression X Sentential Order Subjects X Sequence Length X Time Compression Subjects X Sequence Length X Sentential Order Subjects X Time Compression X Sentential Order Sequence Length X Time Compres- sion X Sentential Order Subjects X Sequence Length X Time Compression X Sentential Order MEAN S UARE 19608. 7676. 3796. 6381. 3979. 433. 2668. 299. 1112. 335. 44. 610. 21. 19. 4 5 0 0000 ['11 245. 96. 195. 523. 49. 22. 218. 15. 91. 39. 50. 4-0 ((1 40. 4.0 4-0 <-0. <.0. (.0. (CL <(L <-0 PO .0005 .0005 0005 .0005 .0005 0005 0005 0005 0005 0005 .109 .0005 .041 .063 .514 32 compression (F = 22.3, dr = 2, p = 40.0005), and sentential order (F = 218.7, df = 2, p = 0.0005). As can be seen from Tables 2 and 3, the subjects of the mentally impaired popu- lation showed a greater mean error response for all circum- stances of presented material under both three word and five word conditions. Significant three-way interactions were found among subjects, sequence length, and sentential order (F = 50.0, df = 2, p = (0.0005) and among subjects time com- pression, and sentential order (F = 2.5, df = 4, P = 0.941). These results demonstrated that l) the subject's error re- sponse increased significantly as sentential order was decreased and temporal redundancy was reduced; and 2) recall accuracy was significantly affected by grammaticality and length of utterance. Effect of Time Compression There was a significant main effect associated with the rate of time compression (F = 195.0, df = 2, p =740.0005). Reference to Tables 2 and 3 show that the number of items incorrectly recalled increased as rate of time compression increased. Furthermore, time compression interacted signi- ficantly with both sequence length (F = 15.4, df = 2, p = (0.0005) and sentential order (F = 39.3, df = 4, p =1<0.0005). These interactions indicated that recall accuracy is affected by the covariance between time compression and these two factors. Thus, as stimulus complexity was increased as a function of increasing sequence length and decreasing gram- maticality, the mean error response of both subject groups was also increased. 33 TABLE 2 Mean error response of subjects per level of time compres- sion (0%, 30%, 60%) and sentence order (Normal sentences, 2nd Order, lst Order) for three-word sequence. . . Mentally Order/Time Compre551on Normals Impaired Normal/0% 0.0 1.1 Normal/30% 0.1 1.3 Normal/60% 0.6 4.9 2nd Order/0% 0 5.3 2nd Order/30% 2 4 7.6 2nd Order/60% S 7 16.5 lst Order/0% 0.5 8.6 lst Order/30% 2.2 12.2 lst Order/60% 6.3 19.9 TABLE 3 Mean error response of subjects per level of time compres- sion (0%, 30%, 60%) and sentence order (Normal sentences, 2nd Order, lst Order) for five-word sequence. . . . Mentally Order/Time Compre551on Normals Impaired Normal/0% 0.0 2.2 Normal/30% 0.6 4.1 Normal/60% 2.0 10.1 2nd Order/0% 1.7 19.1 2nd Order/30% 4.9 27.1 2nd Order/60% 11.5 44.9 lst Order/0% 1.2 22.0 lst Order/30% 2.3 30.3 lst Order/60% 12.6 41.0 34 Effect of Sentential Order The significant main effect of sentential order (F = 523.1, df = 2, p = (0.0005) is illustrated in Tables 2 and 3. These tables show that order of sentential approximation affected performance of both the normal and the mentally impaired groups under three and five word conditions. For all conditions the highest recall accuracy was obtained with the normal sentences, whereas the first and second order sentential approximations showed considerably lower recall accuracy. This trend is especially apparent with the men- tally impaired subjects. The data thus support the conten- tion that an increase in sentential order will provide more cues for perception, thereby aiding recall. The significant sentential order X sequence length interaction (F = 91.2, df = 2, p = (0.0005) shows that the multiple cues provided by sentential material may allow a facilitated "chunking" process and thus aid in the enhancement of both storage capacity and recall performance. Effect of Sequence Length The main effect of sequence length (F = 96.0, df = l, p = (0.0005) as it interacts with the other variables has been discussed in the above text. These interactions closely followed the general trend for both subject groups: as length of utterable sequence increased, mean error re- sponse increased (Tables 2 and 3). 35 TABLE 4 Number error responses of normal subjects per type of error (discrimination, omission, addition, reversal, distortion), level of time compression (0%, 30%, 60%), and sentential order (Normal, second order, first order) for sequence length of three words. (Maximum Error = 30 per subject). NORMAL 3-WORD 0% 30% 60% Type Of Error Normal lst 2n? Normal lst 2nd Normal lst 2nd Discrimination 0 7 1 l 31 36 4 83 69 Omission O 0 0 0 2 0 5 12 16 . Addition 0 0 O 0 0 0 0 0 0 Reversal 0 0 0 0 0 0 0 0 0 Distortion 0 0 0 0 0 0 0 0 0 TOTAL 0 7 1 1 33 36 9 95 85 TABLE 5 Number error responses of mentally impaired subjects per type of error (discrimination, omission, addition, reversal, dis- tortion), level of time compression (0%, 30%, 60%), and sen- tential order (Normal, second order, first order) for sequence length of three words. (Maximum Error = 30 per subject). MENTALLY IMPAIRED 3-WORD 0% 30% 60% Normal lst 2nd Nonmal lst 2ndI‘Normal ISt 2nd“ Type of Error Discrimination 8 61 40 11 108 70 30 103. 121 Omission 8 33 30 7 42 30 41 126 88 Addition 1 4 l 6 3 Reversal 0 0 5 0 0 2 Distortion 0 28 0 24 12 2 62 33 TTTTAI. 17 129 79 19 183 114 74 298 248 36 TABLE‘6 Number error responses of normal subjects per type of error (discrimination, omission, addition, reversal, distortion), level of time compression (0%, 30%, 60%), and sentential order (Normal, second order, first order) for sequence length of five words. (Maximum Error = 50 per subject). NORMAL S-WORD 0% 30% 60% Type Of Error Normal lst an Normal lst 2nd Normal lst 2nd Discrimination 0 13 17 9 26 58 18 121 120 Omission 0 l 2 0 l 7 11 51 44 Addition 0 0 0 0 O l 0 l Reversal 0 4 7 0 5 7 1 16 S Distortion 0 0 0 0 2 0 0 'TOTAI. 0 18 26 9 34 73 30 189 173 TABLE 7 Number error responses of mentally impaired subjects per type of error (discrimination, omission, addition, reversal, dis- tortion), level of time compression (0%, 30%, 60%), and sen- tential order (Normal, second order, first order) for sequence length of five words. (Maximum Error = 50 per subject). MENTALLY IMPAIRED S-WORD Type of Error N 0% 30% 60% ormal lst 2nd Normal lst 2nd Normal lst 2nd Discrimination 19 111 106 34 81 109 49 110 164 Omission 10 102 107 21 226 193 90 316 388 Addition 1 ll 12 2 13 11 2 11 6 Reversal 3 39 16 3 11 ll 3 4 21 Distortion 0 67 46 l 124 82 7 174 95 'FOTAI. 33 330 287 61 455 406 151 615 674 37 ' Type of Error Response Tables 4, S, 6, and 7 illustrate the number of each type of error response (discrimination, ommision, addition, reversal, distortion) for normal and mentally impaired sub- ject groups under the varied conditions of time compression, sentential order, and sequential length. An overall analy- sis of these errors showed errors of discrimination to account for greater than 75% of all errors of the normal p0pu1ation, whereas the mentally impaired group showed the vast majority (68%) of error response to fall within the omission, addition, reversal, or distortion classifications thought by Conrad (1962) to reflect memory constraints. It can also be noted from tables 4-7 that as the degree of time compression increased, a greater preponderance of memory- type errors resulted. This finding was enhanced when the five-word sequences and sentential approximations were used. DISCUSSION The results of this study indicated that normal and mentally impaired populations can be differentiated on the basis of various semantic, syntactic, and temporal interac- tions of verbal stimuli. Although it was evident that, in some instances, specific variables took precedence over others in affecting recall accuracy, no variable was found to be particularly dominant. Subject Differentiation On the basis of the data provided by the normal sub- jects, it can be seen that these subjects experienced rela- tively little difficulty in performing the designed tasks. Although some decrease in recall performance was evident as the task difficulty increased, the results obtained from this group concurred with those of previous investigations (Freeman, 1976; Beasley and Flaherty-Rintelmann, 1976; Wingfield, 1975; Bratt, 1975). Thus, it was demonstrated that normal individuals who possessed an intact auditory per- ceptual system were capable of overcoming the syntactic, sem- antic, and intonational limitations imposed by time compressed tasks of varied sentential length and order. The mentally impaired group showed findings that were similar to the normal group in that recall performance 38 39 decreased as the task difficulty increased. However, the magnitude of performance decline among the mentally impaired was much greater than that of the normal population subjects. Only when normal sentential order was utilized at the lowest rate of time compression (0%) did the mentally impaired achieve scores that reflected adequate processing ability. These results thus seemed to substantiate the findings of earlier investigators (Savin and Perchonock, 1965; Beasley and Shriner, 1973; Craik and Lockhart, 1972; Craik, 1973; Wetherick, 1975) who suggested that syntax and semantics played a key role in auditory perception and memory. It might also be noted that the type of errors made by the men- tally impaired group during normal sentence repetition seemed to lend additional support to these earlier findings. Des- pite the occurrence of errors, an evidence of semantic util- ization was present. For example, the sentence "I was a good boy" might have become "me good boy". Furthermore, the mentally impaired often attempted to structure the first and second order sentential approximations into grammatical utterances possessing sentential meaning. The fact that peak recall accuracy was obtained with the full grammatical sen- tences was thought to be reflective of a multiple-cueing process which enhanced perceptual processing capabilities. The fact that recall accuracy decreased as sequence length increased supported the findings of Aaronson (1967) and Pantalos et a1. (1975). Again, the deterioration of per— formance in the mentally impaired group was much more readily 40 observable than that of the normal group as sentence length was increased from three to five words. In this study, a greater error response occurred with five word than with three word sequences (Figure 1). In view of the findings of Miller (1956), who reported short term memory capacity to be seven plus or minus two units of information, the results of the present study seemed to suggest that the short term capa- city described by Miller had thus been exceeded under certain experimental conditions. When this exceeding of short term capacity occurs, errors of order, omission, addition or dis- tortion may result. The effect of time compression was also found to play a significant role in this study. Figure 2 shows that the num- ber of recall errors increased as the recorded material was time compressed from 0% to 60%. It can be seen from Figure 2 that the 0% condition showed the highest recall accuracy, whereas the 60% condition showed the lowest recall accuracy. The recall accuracy of the 30% time compressed condition fell between the two extremes. Results such as these seemed to be indicative of the fact that time compression affects the extrinsic redundancy of recorded sequences and thus re- duces the number of intonation, stress, semantic, and syntac- tic cues that would be available for perception. Several investigators (Beasley et al., 1972; Freeman, 1976; Beasley and Flaherty-Rintelmann, 1976; Pantalos et al., 1975) have demonstrated that an individual possessing an intact auditory system should be able to compensate for this reduction in 41 100 90 normal 80 mentally impaired N 70 60 50 40 Percent Error Response 30 20 10 \\\\\\\\ \\ 3-word S-word Figure 1. Mean percentage error scores of normal and mentally impaired subjects for three and five word sequences. 42 100 90 normal 80 /// mentally impaired 70 60 50 40 Percent Error Response 30 20 10 \\\\\\\\\\\ I \\\\\\ ‘” \\\\\\\\ 0 30 60% Figure 2. Mean percentage error scores of normal and mentally impaired subjects for three levels of time compression (0%, 30%, 60%). 43 perceptual cues and thus overcome the limitations imposed by time compression, up to a level of 60%. The results of this study were in agreement with these findings in that the normal subject group showed relatively little difficulty in recalling the recorded sequence items. This finding, how- ever, was not true of the mentally impaired individuals with the exception of fully grammatical sentences. These subjects demonstrated significantly greater difficulty in the recall of time-compressed stimuli than did the normal group. The finding that highest recall accuracy for both sub- ject groups was found for normal three and five word senten- ces may be linked to the "chunking" process described by Miller (1956). This process permits new informational units to be coded into groups prior to their entry into the pri- mary memory channel. When normal sentential order was utilized, it appeared that the presented material was enhan- ced by the multiple cueing effect and the chuncking process was thus facilitated (Figure 3). If the presented material contained a high informational content (such as that found in a normal sentence), variables such as increases in time compression or sequence length did not exert as great an effect upon the material to be recalled. Thus, recall accur- acy was enhanced. Differentiation between the subject groups was also apparent when type of recall error response was observed (Figure 4). As can be seen from the figure, some variabil- ity within categorical groupings was present. However, in 100 90 80 70 60 50 40 Percent Error Response 30 20 10 44 \\\\\\\\\N normal .// mentally impaired . 7/ Figure 3. Normal 2nd Order \\\\\\\\\\\ Mean percentage error scores of normal and mentally impaired subjects for three levels of sentential order. 45 100 90 normal 80 //’ mentally impaired 70 60 50 40 Percent Error Response 30 20 ? / .—_l71 L 10 \ \\\\ \\\\\\\ \\\\X\\\ Discrimination Omission Addition Reversal Distortion Figure 4. Mean persentage of error response of normal and mentally impaired subjects according to type of error (Discrimination, Omission, Addition, Rever- sal, Distortion). 46 the main area of differentiation, that of discrimination error versus memory-type error, results were readily obser- vable. Errors of discrimination which were thought to rep- resent a breakdown within the perceptual system were found to account for nearly 75% of all recall errors of the normal group. Conversely, this classification was found for only 32% of the total recall error of the mentally impaired group. Therefore, nearly 70% of all recall errors of the mentally impaired fell into the classifications (omission, addition, reversal, distortion) that have historically been thought to reflect memory constraints (Scholes, 1969; Conrad, 1962). Only 25% of the errors of the normal group seemed reflective of these memory capacity limitations. It should be noted, however, that the number of memory errors increased for both groups as task difficulty increased. However, this rate of increase of memory-type errors was significantly more drama- tic with the mentally impaired than with the normal group. Short Term Memory: Mentally Impaired The data regarding short term memory aspects of mental impairment have not been conclusive in the past. Rather, a number of researchers have speculated, through their experi- mental results, about the interrelationship between behavioral inadequacies of the mentally impaired and deficiencies within the primary memory channel. Broadbent (1958) suggested that the mentally impaired may possess a short term channel capa- city that is restricted to an evan greater degree than that of the normal. This capacity limitation would result in a 47 breakdown of processing abilities if the amount of incoming information exceeded the channel capacity. Ellis (1963) proposed the "Stimulus Trace" Theory in which he reported the performance of the mentally impaired to be below that of the normal because presented informational units were usually forgotten prior to the completion of processing action. The interference theory (Postman, 1963) associated mentally impaired performance with a response competition among pre- sented stimulus items that resulted in recall errors or for- getting of information. Luria (1963) felt mental impairment was a consequence of nervous system pathology which resulted in a description of incoming information either by the pres- ence of simultaneous extraneous stimuli or by fatigue. The results of the present study seemed to confirm the role that short term memory deficiencies play in relation to mental impairment. The error analysis of the sentential stimuli showed that greater than 70% of all recall errors of the mentally impaired were reflective of memory constraints. By contrast, memory type errors were evident for less than 25% of the error response of the normal group. This finding would thus seem to confirm that a short term memory defi- ciency is indeed present within a group of mentally impaired individuals. In regard to why the short term memory channel may be deficient in the mentally impaired, a number of aspects must be considered. First, when the sentential stimuli were time compressed and increased in rate by 0%, 30%, and 60%, the 48 deterioration of recall performance of the mentally impaired was significantly greater than that of the normal subjects. This result thus seemed to indicate that as the load placed upon the primary memory channel of the mentally impaired was increased, group performance deteriorated. Aaronson et al. (1971) had previously suggested that high levels of time compression can produce an overload of the perceptual system and can thus result in an increase in memory errors. While individuals possessing intact auditory pathways are often able to develop strategies to overcome this overload, the mentally impaired usually cannot. This may be due to two factors: 1) the mentally impaired may possess a short term memory of more limited capacity than the norm; and 2) the mentally impaired may not have the linguistic foundation necessary to overcome the limitations imposed by time com- pression. These memory limitations of the mentally impaired became even more obvious when sentential order was decreased and sequential length was increased. When these variables were added, the overload of the system was further increased to the point where additional breakdown of the short term memory channel was evident. In this case the lack of the multiple cues of a fully grammatical utterance seemed to restrict the "chunking" process and thus resulted in an in- crease in memory errors. It is also possible that an inter- ference aspect, similar to that discussed by Postman (1963), may have been interacting with the limited capacity of the mentally impaired individual to result in an additional component for error response. 49 The significant difference in error scores between the normal and the mentally impaired seemed to be related to the assumption that the linguistic competence necessary to process distorted temporal, syntactic, and semantic cues may be deficient in the mentally impaired. This finding, when coupled with the possibility of a limitation of primary mem- ory capacity, shows that the mentally impaired were thus required to alter the strategies employed in a short term recall task. The strategies employed by the mentally im- paired subjects of this study showed a considerable amount of variation and individual difference. One might speculate that the utilization of these strategies was an attempt on the part of the mentally impaired to develop a feeling of success, to increase the linguistic structure of the utter- ances and thus enhance the multiple cueing effect, to gain additional time for processing, or to inhibit the forgetting process. One of the recall strategies employed by many individ- uals within the mentally impaired group was to attempt to limit the amount of information to be processed. This lim- itation of information was seemingly accomplished by the concentration upon and isolation of one word within the pre- sented word string. This single word was then reported by the subject during the allotted response period. The iso- lated word was usually found in either the initial or the final position and was thought to be related to the primacy- recency effect discussed by Feigenbaum and Simon (1962). 50 These authors reported that the first and last informational units of a group of stimuli would be best recalled. The utilization of this strategy by mentally impaired individuals became more evident as task difficulty was increased. Thus, this strategy imployment was most evident in five word first and second order sentential approximations at 60% time com- pression. Perhaps the greatest single strategy utilized by the mentally impaired was to attempt to structure the sentential approximations into a normal sentence having sentential meaning. As discussed earlier, this strategy would increase the number of linguistic cues that could be utilized for per- ception and thus enhance the multiple cueing effect. It was also noted that particular difficulty was experienced when the mentally impaired were required to recall an utterance in which a verb was not present. (For example, "Tree up yes me word”.) When this was the case, the mentally impaired often showed increased confusion. Also, some mentally im- paired subjects attempted to add grammatical constraints to the utterance. This finding may lend additional credance to the research of Foder, Garrett, and Bever (1968) who reported the verb phrase, around which the remainder of the sentence revolved, may be the basic unit of perception. As with the strategy of single word isolation, the greater the difficulty of the recall task, the greater the employment of the strat- egy. Another observable strategy utilized by a number of the mentally impaired involved their repeating of the word string 51 with a shorter than normal latency following its presenta- tion. This attempt at "quick" recall often resulted in a running together of words such that individual word inter- pretation by the tester was impossible. Results such as these were interpreted as "distortion" errors. The appli- cation of this strategy was thought to be reflexive of an attempt by the mentally impaired group to process informa- tion into the primary memory channel prior to the occurrance of informational extinction by forgetting. Thus, it was speculated that the utilization of this strategy may be related in part to the "Stimulus Trace" Theory presented by Ellis (1963). In this theory Ellis presented data that showed the mentally impaired possessed a steeper "forgetting curve" than the normal. Hence, quick repetition might be used as a means of counteracting the forgetting process. Conversely, the utilization of this strategy might also be reflective of the limited capacity theory of Broadbent (1958). Because this increase in vocal speed also seemed to be direc- tly related to task difficulty, it is possible that this strategy was utilized when the short term memory channel was so bombarded with information that a capacity "overload" of the system occurred. An "overload" such as this might very well be reflected in an increase in vocal rate such that phonemic components of one word would overlap or interact with phonemic components of a following word. This effect would result in the garbled indistinct response that charac- terized this type of recall strategy. 52 The final observed recall strategy of the mentally impaired involved a longer-than-normal latency period between completion of the recorded utterance and the verbal response by the subject. In some instances this latency period was so long that the verbal recall of the subject was not ini- tiated prior to the commencement of the following recorded stimulus. If the subject did not respond within the allotted interstimulus interval of ten seconds, errors of "omission" were charged. It was speculated that the use of this recall tactic may have indicated an attempt by the mentally impaired to focus greater attention or concentration upon the stimuli to be recalled. Zeamon and House (1963) and Fisher and Zeamon (1973) reported that the attention factor may be defi- cient within mentally impaired populations. Furthermore, this attempt by the mentally impaired to enhance attention may be in some cases a detrement to recall. This hypothesis is based on the Ellis (1963) finding of the steeper than nor~ mal "forgetting curve" within the mentally impaired. Thus, a strategy employment that involves an additional amount of processing time prior to recall may have the effect of in- creasing the potential for informational forgetting. If this occurs, additional memory-type errors should result. Summary and Conclusions The overall results of this study demonstrated that an interaction of time compression, order of sentential approx- imation, and sequence length was capable of differentiating mentally impaired subjects from normal subjects. The fact 53 that a time-compressed stimulus was used provided additional support to the notion of previous research findings that time-compressed speech may be an important diagnostic tool for the delineation of central auditory dysfunction. How- ever, the limited population and the limited past utilization of time-compressed sentential stimuli make it impossible to draw any definite conclusions as to general clinical appli- cation. The results of this study also suggested that informa- tion processing capabilities differed between the normal and the mentally impaired groups. This differentiation of sub- ject groups was evident when reviewing the number and type of recall errors. The normal control group showed signifi- cantly better recall than the mentally impaired subjects. In addition, an analysis of error type showed that greater than 70% of all recall errors of the mentally impaired group could be classified as reflecting memory constraints. Only 25% of the recall errors of the normal group fell into these classifications. These findings were speculated to be rela- ted to the idea that the mentally impaired may possess a short term memory capacity that is more restricted than the norm. These capacity restrictions became even more evident as task difficulty was increased. It was also apparent from the results of this study that the grammatical structure of an utterance plays an impor- tant role in the recall of that utterance. These results support the findings of Miller and Isard (1963) and others 54 who reported the ease with which sentences can be perceived depended to a large extent upon their degree of grammatical- ity. While the normal subjects showed little additional difficulty with recall of sentential approximations, this was not the case with the mentally impaired. Thus, the significant differences observed among the sentential orders implied that the mentally impaired may not have the linguis- tic foundation necessary to overcome the limitations imposed by less grammatical utterances. Implications Inview of the response of the mentally impaired subjects to the time-compressed stimuli, it might also be of interest to observe the response of this subject group to other types of temporally-altered stimuli. For example, variations in interstimulus intervals and stimulus durations may provide additional information concerning processing abilities of the mentally impaired. The use of time-expanded stimuli might also be helpful in examining these abilities. In addi- tion, research efforts in the areas of phonemic synthesis, auditory closure, and speech-in-noise may furnish additional data about learning deficiencies within the mentally impaired and thus should also be investigated. The results of this study may also suggest a possible implication for the educational habilitation of the men- tally impaired. It was readily apparent from the results of this study that the mentally impaired achieved greatest re- call accuracy under conditions in which fully grammatical sentences were utilized. This finding was speculated to be 55 related to a multiple cueing process that was associated with normal sentence recall. Although no concrete data was obtained that could be directly related to the "education" of the mentally impaired, the evidence that grammatic struc- ture played an important role in auditory perception cannot be denied. Therefore, it can be speculated that the educa- tional achievement of the mentally impaired might be enhanced through the utilization of a teaching framework that stresses the presentation of grammatical utterances. However, further investigation in this area is definitely warranted. LI ST 0]? REFERENCES LIST OF REFERENCES Aaronson, D., Temporal factors in perception and short term memory. Psychological Bulletin, 67, 130-140 (1967). Aaronson, D., Markowitz, N., and Shapiro, H., Perception and immediate recall of normal and "compressed" auditory sequences. Perception and Psychophysics, 9, 338-344 (1971). Aaronson, D., and Sternberg, S., Effects of presentation rate and signal-to-noise ratio on immediate recall. Paper presented to the Eastern Psychological Associa- tion (1964). Atkinson, R., and Shiffrin, R., The control of short term memory. Scientific American, 244, 82-90 (1971). Baddeley, A., Retrieval rules and semantic coding in short term memory. Psychological Bulletin, 78, 379-385 (1972). Beasley, D., Forman, B., and Rintelmann, W., Perception of time compressed CNC monosyllbles. Journal of Auditory Research, 12, 71-75 (1972). Beasley, D., Schwimmer, S., and Rintelmann, W., Intelligi- bility of time compressed CNC monosyllables. Journal of Speech and Hearing Research, 15, 340-350 (1972). Beasley, D., and Shriner, T., Auditory analysis of tempor- ally distorted sentential approximations. Audiology, 12, 262-271 (1973). Beasley, D., and Flaherty-Rintelmann, A., Children's percep- tion of temporally distorted sentential approximations of varying length. Audiology, 15, 315-325 (1976). Beasley, D., Maki, J., and Orchik, D., Children's perception of time compressed speech on two measures of speech dis- crimination. Journal of Speech and Hearing Disorders, 41, 216-225 (1976). Bisset, J., and Koshy, B., Effect of interstimulus interval on children's ability to sequence unrelated items. Un- published Manuscript, Michigan State Univeristy (1975). 56 S7 Bocca, E., and Calearo, C., Central hearing processes. In Modern Develppments in Audiology, Ed. Jerger, J. (1963). Bratt, 6., Performance of subjects on time compressed sen- tential stimuli. Unpublished Master's Thesis, Michigan State University (1975). Broadbent, D., Perception and Communication, Oxford: Per- gamon Press (1958). Broadbent, D., Flow of information within the organism. Journal of Verbal Learning and Verbal Behavior, 2, 24- 39 (1963). Brown, J., Some tests on decay theory of immediate memory. Quarterly Journal of Experimental Psychology, 10, 12-21 (1958). Butterfield, E., Serial learning and the stimulus tract theory of mental retardation. Americal Journal of Mental Deficiency, 72, 778-787 (1968). Butterfield E., Wambold, C., and Belmont, J., On the Theory and practice of improving short term memory. American Journal of Mental Deficiency, 77, 654-669 (1973). Calearo, C., Verbal tests of central audition. Audiology, 14, 300-311 (1975). Calearo, C., and Lazzaron, A., Speech intelligibility in relation to the speed of the message. Laryngoscope, 67, 410-419 (1957). Carrow, E., and Mauldin, M., Children's recall of approxima- tions to English. Journal of Speech and Hearing Research, 16, 201-212 (1973). Conrad, R., An association between memory errors due to acoustic masking of speech. Nature, 193, 1314-1315 (1962). Conrad, R., and Hille, B., The decay of immediate memory and paced recall. Canadian Journal of Psychology, 12, 1-6 (1958). Craik, F., "Levels of Analysis" view of memory. In Communi- cation and Affect, Lan ua e and Thou ht, Eds. P. Pliner, L. Krames, and T. Allaway, (19735 Vqume II. Craik, F., and Lockhart, R., Levels of processing: A frame- work for memory research. Journal of Verbal Learning and Verbal Behavior, 11, 671-684 (1972). 58 Ebenholtz, S., Serial learning: Position learning and sequen- tial associations. Journal of Experimental Psychology, 66, 353-362 (1963). Ellis, N., The stimulus trace and behavioral inadequacy. In Handbook of Mental Deficiency, Ed. N. Ellis (1963). Ellis, N., A behavioral research strategy in mental retarda- tion: Defense and critique. American Journal of Men- tal Deficiency, 73, 557—566 (1969). Ellis, N., Memory processes in retardates and normals. In International Review of Research in Mental Retardation, Ed. N. Ellis (1971). Ellis, N., and Anders, T., Effects of interpolated recall upon short term memory. Journal of Experimental Psychology, 79, 568-569 (1969). Ellis N., and Dugas, J., The serial position effect in short term memory under E- and S- paced conditions. Psycho- nomic Science, 12, 55-56 (1968). Ellis, N., and Hope, R., Memory processes and the serial position curve. Journal of Experimental Psychology, 77, 613-619 (1968). Feigenbaum, E., and Simon, H., A theory of the serial posi- tion effect. British Journal of Psychology, 53, 307- 320 (1962). Fisher, M., and Zeamon, D., An attention—retention theory of retardate discrimination learning. In International Review of Research in Mental Retardation, Ed. N. Ellis (1973). Fodor, J., Garrett, M., and Bever, T., Some syntactic deter- minents of sentential complexity, II.: Verb structure. Perception and Psychophysics, 3, 454-461 (1968). Fraser, D., Delay of immediate memory with age. Nature, 182, 1163 (1958). Freeman, B., Performance of reading impaired and normal read- ing children on temporally altered monosyllables and sentential stimuli. Unpublished Doctoral Dissertation, Michigan State University (1976). Freeman, B., Beasley, D., and Overholt, 6., Performance of learning disabled children on temporally altered speech stimuli. Unpublished manuscript, Michigan State Uni- versity (1975). 59 Galton, F., Supplemental notes on "prehension" in idiots. Mind, 12, 79-82 (1887). Hayes, C., and Routh, D., Short term memory in normals and retardates as a function of stimulus intensity and recall interval: A test of Ellis' S+ theory. Psycho- logical Reports, 30, 831-836 (1972). Hebb, C., The Organization of Behavior. New York: John Wiley and Sons (1949). Hull, C., A Behavior System. New Haven: Yale University Press (1952). Jensen, A., Role learning in retarded adults and normal children. American Journal of Mental Deficiency, 69, 828-834 (1964). Jensen, A., and Rohwer, W., The effect of verbal mediation on the learning and retention of paired associates by retarded adults. American Journal of Mental Deficiency, 68, 80-84 (1965). Jerger, J., Audiological manifestations of lesions in the auditory nervous system. Laryngoscope, 70, 417-425 (1960). Jerger, J., Carhart, R., Tillman, T., and Peterson, J., Some relations between normal hearing for pure tones and speech. Journal of Speech and Hearing Research, 2, 126- 140 (1959). Kellas, G., and Butterfield, E., Effect of response require- ment and type of material on acquisition and retention performance in short term memory. Journal of Experi- mental Psychology, 88, 50-56 (1971). Kirk, S., The diagnosis and remediation of psycholinguistic disability. Urbana, Illinois: University of Illinois Press (1966). Kohler, W., Gestalt Psychology. New York: Liveright (1929). Konkle, D., Intelligibility of time-altered CNC monosylla- bles as a function of contralateral masking. Unpub- lished Doctoral Dissertation, Michigan State University (1976). Konkle, D., Beasley, D., and Bess, F., A study of time com- pressed speech on an elderly population. Presented at the convention of the American Speech and Hearing Association (1974). 60 Konkle, B., Freeman, B., Riggs, D., Riensche, L., and Beasley, D., Calibration procedures for time compres- sed/expanded speech. Proceeding of Third Louisville Conference on Rate Controlled Speech (1975). Kouw, W., Effects of stimulus intensity and duration upon retardates short term memory. American Journal of Mental Deficiency, 72, 734-739 (1968). Kurdziel, S., and Noffsinger, D., Performance of cortical lesion patients on 40% and 60% time compressed speech materials. Presented at the convention of the American Speech and Hearing Association (1973). Kurdziel, S., Rintelmann, W., and Beasley, D., Performance of noise induced hearing-impaired listeners on time compressed consonant-nucleus-consonant monosyllables. Journal of the American Audiology Society, 1, 54-60 (1975). Lindsey, P., and Norman, D., Humap_Information Processing. New York: Academic Press (1972). Luria, A., Psychological studies of mental deficiency in the Soviet Union. In Handbook of Mental Deficiency, Ed. N. Ellis (1963). MacMillan, D., Facilitative effect of input organization as a function of verbal response to stimuli in EMR and non- retarded children. American Journal of Mental Defic- iency, 76, 400-411 (1972). Maki, J., Speech discrimination and response latency of nor- mal hearing and hearing impaired children as a function of time compression. Unpublished Doctoral Dissertation, Michigan State University (1975). Manning, W., Johnson, K., and Beasley, D., The performance of children with auditory perceptual disorders on a time compressed speech discrimination measure. Pre- sented at the convention of the American Speech and Hearing Association (1975). Massaro, D., Pre-perceptual images, processing time, and per- ceptual units in auditory perception. Psychological Review, 79, 124-145 (1972). Miller, G., The magical number seven, plus or minus two: Some limits on our capacity for processing information. Psychological Review, 63, 81-97 (1956). 61 Miller, G., and Isard, 8., Some perceptual consequences of linguistic rules. Journal of Verbal Learning and Verbal Behavior, 2, 217-228 (1963). Norman, D., Acquisition and retention in short term memory. Journal of Experimental Psychology, 72, 369-381 (1966). O'Connor, N., and Hermelin, B., Recall in normal and subnor- mals of like mental age. Journal of Abnormal Social Psychology, 66, 81-84 (1963). O'Connor, N., and Hermelin, B., Input restriction and immed- iate memory decay in normal and subnormal children. Quarterly Journal of Experimental Psychology, 17, 323- 328 (1965). Pantalos, J., Schukers, G., and Kipskind, N., Sentence length-duration relationships in an auditory assembly task. Journal of Communication Disorders, 8, 61-74 (1975). Pollack, 1., Message uncertainty and message reception. Journal of the Acoustical Society of America, 31, 1500- 1508 (1959). Pollack, 1., Language as behavior. In Brain Mechanism Underlying Speech and Language, Ed. C. Milliken and F. Darley (1967). Pollack, 1., Johnson, L., and Knapf, P., Running memory span. Journal of Experimental Psychology, 65, 137-146 (1959). Pollack, I., and Johnson, L., Continuing memory span for digits. Perception and Motor Skills, 17, 731-734 (1963). Posner, M., Immediate memory in sequential tasks. Psycholo- gical Bulletin, 60, 333-349 (1963). Postman, 1., Does interference theory predict too much for- getting. Journal of Verbal Learning and Verbal Behavior, 2, 40-48 (1963). Rosen, M., Kivitz, M., and Rosen, B., Uni-molded and cross modal "coding" in the mentally retarded. American Journal of Mental Deficiency, 69, 716-722 (1965). Savin, H., and Perchonock, E., Grammatical structure and the immediate recall of English sentences. Journal of Ver- bal Learning and Verbal Behavior, 4, 348-353 (1965). 62 Scholes, R., The role of grammaticality in the initiation of word strings by children and adults. Journal of Verbal Learning and Verbal Behavior, 8, 225-228 (1969). Schulman, H., Similarity effects in short term memory. Psy- chological Bulletin, 75, 399-415 (1971). Scott, K., and Scott, M., Research and theory in short term memory. In International Review of Research in Mental Retardation, Ed. N. Ellis (1968). Shriner, T., and Daniloff, R., Reassembly of segmented CVC syllables by children. Journal of Speech and Hearing Research, 13, 537-547 (1970). Shoup, J., Performance of children to the time compressed WIPI Test. Unpublished Master's Thesis, Michigan State University (1975). Smith, A., Short term memory and imitation: Effects of sen- tence length, sentence type, word type, stress, and grade. Unpublished Master's Thesis, Michigan State University (1972). Speaks, C., and Jerger, J., Method for measurement for speech identification. Journal of Speech and Hearing Research, 8, 185-194 (1965). Spitz, H., The role of input organization in the learning and memory of the mental retardates. In International Review of Research in Mental Retardation, Ed. N. Ellis (1966). Tejirian, E., Syntactic and semantic structure in the recall of orders of approximations to English. Journal of Ver- bal Learning and Verbal Behavior, 7, 1010-1015 (1968). Underwood, B., Interference and forgetting. Psychological Review, 64, 49-60 (1957). Waugh, N., and Norman, D., Primary memory. Psychological Review, 72, 89-104 (1965). Wetherick, N., The role of semantic information in short term memory. Journal of Verbal Learning and Verbal Behavior, 14, 471-480 (1975). Wingfield, A., Acoustic redundancy and the perception of time-compressed speech. Journal of Speech and Hearing Research, 18, 96-104 (1975). Zeamon, D., and House, B., The role of attention in retar- date discrimination learning. In Handbook of Mentgl Deficiency, Ed. N. Ellis (1963). APPENDICES APPENDIX A REVISED WORD LIST BASAL VOCABULARY ON WE GO HOUGHTON-MIFFLIN (1966) REVISED WORD LIST BASAL VOCABULARY ON WE GO HOUGHTON-MIFFLIN (1966) a red pan sand girl zoo high stop cry with eye was all ten dog late boy set feet jar but is like come had in my six her to not men I we milk ring dot pop that said man take toy they out will wish lock on me sat see put hot good snow tell the it food see eat ride dish sun do play go two call wood egg of cat yes know here him no work box up name word big us end tree car you door this five nest run time day store one would 63 APPENDIX B SENTENTIAL WORD LISTS SUBJECT l. I 2. the 3. I 4. stop 5. you 6. I 7. I 8. I 9. I 10. lock will cat know the work am eat like like the NORMAL SENTENCES play played you car late big here milk her door 64 CONDITION SUBJECT 1. men 2. that 3. snow 4. dog 5. wish 6. good 7. like 8. up 9. man 10. do will man time nest with girl good and zoo hot SECOND ORDER do of will girl her work one to food wish 65 CONDITION SUBJECT 10. go take food jar red wish egg UP nest name girl hot jar sand toy P0P call jar ring will FIRST ORDER zoo girl wish man us milk ten said sun jar 66 CONDITION SUBJECT NORMAL SENTENCES CONDITION 1. we went to the zoo 2. that girl is not good 3. she ran to the store 4. we play in the snow 5. you do like the snow 6. that boy is with me 7. I was a good boy 8. that boy will play here 9. five men came to work 10. you take that cat out 67 SUBJECT 10. here stop feet nest man late is eat put all 988 is milk sun time said milk egg SECOND ORDER CONDITION ”P of zoo big one eat to five P0P take like stop with ten sun with wish feet come UP 68 tree good good said milk you P0P zoo ten SUBJECT 1. men 2. red 3. snow 4. name 5. time 6. snow 7. p0p 8. tree 9. me 10. dog ZOO play take milk all pep 200 UP play cat FIRST ORDER CONDITION pop nest is late I call play big man good girl sun sun pop tree dish is name cry food milk yes me word wish milk nest high man hot 69 APPENDIX C RESULTS OF INDIVIDUAL SUBJECTS .3223. u E £33262 u :2 53320 u 3.6 .8323me u 8:: .8 u .55 HmHeHsHEH .583: 253.883 ... SE .cofimmopgsou mags we we moocosvom whoz-oouzh How muuomnsm Hashoz ammonmom uouum mo :oHumoHMHmmmHu o o o o a o o o o H o o o o o HEB. o o o o o o o o o o o o o o 0 H o o o o H c o o o o o o o o 9 H o o o o H o o o o o o o o o o 2 o a c o H o o o o o o o o o 0 NH o o o o o o o o o o o o o o o HH o o o o o o o o o H o o o o o S o o o o o o o a o o o o o o o m o o o o o o o o o o o o o o o w o o o o o o o o o o o o o o o H o o o o o o o o o c o o c o o o o o o o H o o o o o o o o o o m o o o o o o o o o o o o o o o e o o o o H o o o o o o o o o o m o o o o H o o o o o o o o o o N o o o o H o o o o o o o o o o H GE E a: was 88 EB Ex 84H 25 HS .55 smm SHE £5 88 «memo .5: $95 58% g a 70 u .88 .Hmmumsmm u E .8332 n 92 53825 u 320 .eoflmfifibms u 8an .3. u .stm HmseHsHHeH 55682 303888 .conmngeou oEHH Non um moocosvom pHoz-eoH:H How muoonnsm Hmeuoz HmspH>HucH mo omzommom Hopum Ho :oHumonHmmmHu o o o N Hm o o o o om o o o o H .7:»:. o o o o m o o o o e o o o o o mH o c o H m o o o o m c o o o H «H o o o o H o o o o N o o o o o H o o o c o o o o o m o o o o o NH o o o o m o o o o H o o o o o HH o o o o o o o o o N o o o o 0 OH o o o o m o o o o N c o c o o m o o o o N. o o o o H o o c o o m o o o o N o o o o N o o o o o n o o o o H o o o o N o o o o o o o o o o H o o o o m o o o o o m o o o o o o o o o o o o o o o q o o o o N o o o o m o o o o o m o o o H H o o o o m o o o o o N o o o o e o o o o H c o o o o H .HMHQ >mm and. £35 omHQ «memo 5E .HmHQ E and. mHHB umHo mmamo onumm .HmHQ >mm DE mHsG umHHH .2382 .HUmHHSm 71 .8 u 8.5 H8331: Ea: 938.888 a BS .3823 u E .8332 a a3. 53325 u 820 .coHuEHsHDmE u HHS .aonmoHeBu «EH woo am 36533 Hu.H03-oo.HH:. SH 3838 Hgfioz 3:33ch we momcommom .82.".— .Ho :oHuonhHHmmeu o o o NH mm o o 0 0H me o o o m e .25 o o o H w o o o m m o o o N H m H o o o o m o o o c m o o o o 0 H o o o o e o o o N m o o o o 0 MH o o o a HH e o o o 0H c o o o o NH o o o o m o o o o m o o o o 0 HH o o o o v o o o o v o o o o O OH o o o o o o o o m m o o o o N m o o o o N. o o o o o o o o o o w o o o o o o o o m o c o o m o N. o o o N m o o o m N o o o o o o o o o o N o o o o m o o o o o m o o o o o o o o o N o o o o o v o o o H S o o o o OH o o o o H m o o o H H o o o o v o o c o o N o o o H v o o o o m o c o c o H .HMHQ >mm 99H. £20 88 .53 >mm an? wHE 88 .53 >mm 99H mHyB umHa mmomo .HmmHm mmnmo mzoumm 12502 .858 72 .8 u 8.5 H§H>HEH Sana: HeoHthHHH u .85 .Hmmugom u >mm .HHoHuHHOBH u 93. .5335 u 920 .HHoHumHHHEHHumHQ u 8:: .HHonmoHHHEou as: No um 36 .5563 303-635. How 3833 ©0335 NHHmuHHoz 3:33ch we momconHmom Sam .3 :opruHmHmmaHu N n v mm Ho H m m cm ow o o H w w H58 0 H H H v H o o o N o c o o o H H o o o m o o o H N o o o o o «H o o o o o o o o o H o o o o o mH m o H o m o o c c m o o o o 0 NH o o o o H o o o o o o o o o 0 HH o o o o m o o o m N o o o o o S H o o o c o N o o N o o o H H m o o o H N o o H mH H o o H o o w H N o H w o N o m o o o o o H N o o o o H o o o o H o o o o N o o o o N N o o o o N o o o H o m o o H o m o o o o v o o o H o v v c o m N o o o m m o o o H N m m o H H N o o N v m o o o m N N H c o H m o H o H m o o o H o H .HmHa >mm an? 2:5 88 .53 >mm and. mHsU Q30 .85 >mm 99H £5 85 El mmamo Fwd: mmomo @2086 .2552 73 .3 u “chum H§H>HHEH .55sz 303.833 n 53 .me»o>om u >mm 50333 u 92 53325 u 35 503955283 u 8an 5333980 05H won um m8 .533 203-085. how 3835 @8395 N335: H3359: mo mmmcommmm Sham .Ho :oHumuHmHmmmHu N m w Nv 2: NH c N on ON 0 o H N HH 452 H H H o m o o o o N o o o o o mH m o o o o o o o o v o o o o o QH o H o c m o o H o m o o o o o MH H c o o o o c o o H o o o o 0 NH 9 o o o m .o o o H m o o o o o HH H o o o c o o o o m o o o o o OH N o o O OH N o o m N o o o N H m H H o aH m o o o m m o o o o N w o o H v HH H o H m m o o o o m N o o o o v o o o o v o o o N N o o o o N v o o o o o o o o o H m o N N H w o o o o m o o H H N v o o o m o H o o H v o o o o o m o o o H NH o o o v o o o o N o N N o o H S N o o H N o o o o o H 5H9 >mm Q3 mHSO 85 [BE >mm and. mHa 83 .55 >mm an? mH—S ume mmmmo E: mmnmo mzoumm .2502 E88 74 .3 u Sham H§H>HHEH gnu: 983.835 u .53 .Hmmugom u >mm 53533 u 99» 5332.0 u szo .coHumaHEHpumHo u 038 .conmmeEou 0E2. woo pm mou -coscmm 303635. you 383%.. 8:35 NHHmucoz H33>HEH mo momcommmm notm mo :oHumunHmmmHu No N o ONH NOH mm m m mm HNH N o H H« om H468. m H N o « o o o m m o c o m o mH « H H H m m H H o m c o o H H «H H o o mH m o o H mH N o o o N N mH OH o o o o o o o o N o o o c H NH « o o o N o H o H m o o o H o HH « o H NH « m o o o N o o o H 0 0H NH o o N m NH H o H m o o o o N m o o 0 MN N o o o NN o o o o HH H w « o H m «H m o o m NH 0 o H o m N m o o H N o o o H «H o o o H « o o o o NN o o o o wH m o o o H N m o o H « OH o o o H OH o o o o N « o o c oN o N o N m m N o o 3 m. m o o o OH « m o o «H m o o o « o N o o o H mH N o o o «H o o o H m H HmHQ >mm Q3 35 0mg 55 >mm H54. 25 umHa Ema >5. 9?. mHsU uwHQ amaze Em H m mug—o; 9200mm £02 5; 7S .8 u 8.5 3:3:«5 .5534 253.838 u as .Hmmumim n >mm .5332 u OO< 53325 u 320 53355933 u 033 .conmmumEou 05:. NO am 305ng 6.33-9»5 How 38.33 H9502 3333?: mo momcommmm .Stm mo coHumuHmHmmmHu O « O H «H O N O N NH O O O O O .289. O O O O H O O O O N O O O O O mH O O O O O O O O O O O O O O O «H O H O O H O H O O H O O O O O «H O H O O O O O O O O O O O O O NH O O O O H O O O O H O O O O O HH O O O O O O H O O O O O O O O OH O O O O N O H O O N O O O O O O O O O O H O N O O H O O O O O O O O O O « O H O O « O O O O O N O O O O O O H O O O O O O O O o O H O O N O O O O O O O O O O m O H O O O O O O O O O O O O O « O O O H O O O O N « O O O O O m O O O O H O O O O N O O O O O N O O O O O O O O O O O O O O O H HmHQ >mm Gm? 220 85 EH9 >mm an? mHsG 85 .55 >mE an? mHZO umHQ mmamo EH m mmamo Ozoumm ‘ZZmOZ 553m 76 .8 u 8.5 RBEHEH €598: Jagged u E .8382 u 92 63825 u was .cofiéflfibma u 085 €03,833 u ems .conmanou 9:3. wow 3 mmucoscow $83-3: pom 300.3% HQEoz 3:33qu mo mmmcommmm 8.5m .«o :oHumonHmmmHu N m O H ON O N H N am O O O O O 12,89 O O O O n O O O O n O O O O H mH O O O O N O O O O n O O O O H «H O H O O H O H O O H O O O O O «H O H O O O O H O O O O O O O O NH O O O O H O O O H O O O O O H HH O O O O O O O O O m O O O O O OH O O O O o O O O O o O O O O O O O O O H H O N O m o O O O O H O O H O O m O O O O m O O O O H N O H O O N O H O O N O O O O O o N O O O m . O O H H o O O O O H m O O O O O O O O O N O O O O H « O H O O O O N O O o O O O O H n O O O O N O O O O « O O O O H N O O O O O O O O O N O O O O O H .HMHQ >mm and. £5 umHQ amaze 9mm: Ema >mm an? mHE umHQ mmmmo 9200mm HmHQ >mm and. szo 83 £02 Umazm 77 .8. u 82m 33235 5.3%: 20:83.8 u as .Hmmpo>om u >mm .coHpHOO< u OO< .conmHEO u OHzo .coHumcHEHpumHO n OOHOO .conmmumeou mEHN OOO um moucoscom Opoz-m>Hm pow muuomnsm Hmspoz HmSOH>chH mo momcommom “Chum mo coHumuHmHmmmHu O OH H Hm HNH « O O «« ONH O H O HH OH 4mm OQ< OHZO UOHQ POHQ >mm DQ< OHZO UOHQ HOHQ >mm QQ< OHZO UOHO mmmmo HmmHm mmmmo 9200mm Aom u >mm .8333 u OO< £039.25 u OED .coUmNNNECumNO u 8:: LEN—cow Ouozéfim Now 3835 @2395 N335: H§N>NO=H mo mmmcommom .No.N.Nm mo 20383338 203833 a BS .Nmm .conmoNNNNBO 9:: NO um moo No S NN NoN NNN ON 2 NN NS 2: o m N ON 3 #98 N N o N N o N o N m o o o o N mN N N N N NN o N N o S c N o N N NN N NN N N 2 m N m N m o o o N N 2 N o o N m m o c N o o o o N o NN 2 N N m NN NN N N m N o o o o N NN N m o m N o N o m NN o o o o o S o o N o ON N N o N NN o o o o N a o o o o N o o o o m o o o o o N o N o N N N o N NN N o o o o o N o m N NN m o o o NN m o o o o N o S N o N m o N N m N o N o o N N. N m N o N o N o m N o N o o o N N m o N m N o N o o o o o o o N NN N o 2 o S o o N N o o N m m N o N N NN m o m o 2 N o o o N N N BE RE 92 2:0 85 .58 RE 92 25 BE .83 Ex 92 2:0 88 $95 ENE $95 228% a a 79 .om u nowhm_Hm:va>Nn:H_s:an§£z L33. u E .aoNNNB< u EN. 53320 u 25 .coNNSNESmNNN ... 8N3 .5563 303-03» now 3835 3:35 NHHmuNNoz H3235 mo momcommmm Baum mo :oHumuGNmmmHo NSNNNBNE u NmNNN .Nam 5338950 9:; New um m3 «NH HH mH oNN Hm NN HH HH mmH moH H m N HN em Hmm ao< wH—é umHQ mmamo HmmHm HmHQ >mm GEN. mHsU um:— mmnmo 9200mm NmHQ >mm Qn<. mHZD umHQ HHNEH mo momcoammm Hohm mo :oHumuNNHHmmmHu .8 u SEN N§N>NBN 55$: NaoNthNNN u .58 .N8 .NNonmoNNNNBu 9:2. NOO um 30 «NH « HH OHm OHH OHH O O NON «OH N m N OO O« Hmm OQ<. mHZU uwHQ mmamo NmmHm NmHo >mm GEN 335 um:— mmnmo mzoumw .55 >5— 9?. mHyU umHQ q¢zaoz El 81 IHJIMIJHI‘IH 1 196 3 1293 O3 1 llHlWlll