Author . 
 
 ^f.^*^/« 
 
 Title 
 
 Imprint. 
 
 18 — 47372-2 ffPO 
 
RECITATION AS A FACTOR IN 
 MEMORIZING 
 
 BY 
 
 Arthur I. Gates, M. A. 
 
 Submitted in partial fulfillment of the requirements for 
 
 the degree of Doctor of Philosophy, in the Faculty 
 
 of Philosophy, Columbia University 
 
 REPRINTED FROM 
 
 ARCHIVES OF PSYCHOLOGY 
 
 No. 40 
 
 NEW YORK 
 I9I7 
 
RECITATION AS A FACTOR IN 
 MEMORIZING 
 
RECITATION AS A FACTOR IN 
 MEMORIZING 
 
 BY 
 
 Arthur I. Gates, M. A. 
 
 Submitted in partial fulfillment of the requirements for 
 
 the degreeof Doctor of Philosophy, in the Faculty 
 
 of Philosophy, Columbia University 
 
 REPRINTED FROM 
 
 ARCHIVES OF PSYCHOLOGY 
 
 No. 40 
 
 NEW YORK 
 I917 
 
L'ZB /<D 6 3 
 
 
 UNIVERSITY PRINTING OFFICE 
 COLUMBIA UNIVERSITY 
 
TABLE OF CONTENTS 
 
 Chapter Page 
 
 I. Introductory statement of the problem i 
 
 II. Summary of previous studies on the problem ... 4 
 
 III. The subjects, materials, and methods of procedure . 24 
 
 IV. Quantitative results 35 
 
 V. An analysis of reading and recitation as 
 
 factors in learning 65 
 
 VI. Conclusions and pedagogical implications .... 99 
 
PREFACE 
 
 The present investigation was begun in the Psychological Labo- 
 ratory of the University of California in the spring of 19 16. The 
 experiments in which children served as subjects were conducted in 
 a public school in Oakland, California, while those upon adults 
 were, for the most part, carried out in the Psychological Laboratory 
 of Columbia University. 
 
 The writer has been fortunate in having enjoyed, during the 
 course of the work, endless encouragement, suggestions, and assis- 
 tance from a large number of people. To Mr. N. Ricciardi, Prin- 
 cipal of the school visited, I am indebted for the privilege of con- 
 ducting the experiments upon his charges as well as for the ready 
 help in arranging details for the work. To the many teachers whose 
 class-rooms I invaded, I am indebted for the kindliest toleration and 
 for a great deal of valuable assistance. My debt of gratitude to 
 Professors G. M. Stratton and Warner Brown of the University of 
 California and to Professors J. McKeen Cattell, R. S. Woodworth, 
 E. L. Thorndike, H. L. Hollingworth, and Dr. A. T. Poffenberger of 
 Columbia University, is very great. To my friend Charles E. 
 Martin, I am indebted for valuable suggestions and criticisms in the 
 preparation of the manuscript. 
 
INTRODUCTORY STATEMENT OF THE PROBLEM 
 
 The process of learning as carried on by most adults depending 
 upon their native resources or practical experience, is frequently 
 interrupted by attempts at recitation or voluntary recall of what 
 has been learned. We tend to introduce an attempt at recitation at 
 the earliest possible moment, usually long before a perfect repro- 
 duction is possible. In that case, as a rule, we refer promptly to the 
 material being studied in order to complete the perusal. For 
 example, many years ago Francis Bacon observed, "If you read any- 
 thing over twenty times you will not learn it by heart so easily as 
 if you were to read it only ten, trying to repeat it between whiles, 
 and when memory failed looking at the book." ^ The spontaneous 
 methods of learning of many people resort so naturally to these 
 attempted reproductions that we can hardly refuse to believe that 
 they are helpful. Yet most of us would admit that the dominating 
 idea behind such a procedure is the fear of studying the lesson more 
 than is absolutely necessary, and it is by no means clear that intro- 
 ducing the recitation too early in the learning process may not 
 result in loss of time. This gives rise to several practical questions, 
 such as: — Is an attempted recitation of as much value in learning 
 as another perusal or reading, and is a recitation at one stage of the 
 learning as valuable as at another? 
 
 It is at once obvious that the solution of such questions is of 
 tremendous import for the work of the school. It is imperative that 
 recall or recitation, as a factor in learning, should be analysed and 
 its quantitative importance determined. Although several studies 
 of the problem have been made within the last decade, facts that 
 will permit indisputable application to the work of the school-room 
 are still wanting. The amount of experimentation required to solve 
 the problem adequately is much greater than would at first thought 
 appear, since different results might be expected according to the 
 age and training of the subject, the kind of material employed, the 
 length of the lesson or the purpose of the learner, i. e., whether the 
 material is to be 'learned by heart' or only partly learned. The 
 general condition of the problem is indicated by a recent statement 
 of Meumann, who, after summarizing the work in the field, con- 
 
 1 Novum Organum, 1620, translated by James Spedding, edition of 1863, p. 229. 
 
2 RECITATION AS A FACTOR IN MEMORIZING 
 
 eluded that^ "it is indispensable that such experiments be repeated 
 and confirmed before the results are applied to pedagogy." 
 
 That great variations in the methods of learning exist even among 
 adults with college training will be surprisingly evident to anyone 
 who will select at random twenty such individuals and carefully 
 observe the means employed in learning a stanza of poetry or a 
 series of nonsense syllables. Many cases of alleged 'poor memories' 
 may prove to be due to poor methods of study. One would even 
 more confidently expect to find among children a greater number of 
 inefTective methods of studying. In fact Miss M.J. Baldwin made 
 a study of this matter » and found that such was the case. She 
 undertook, by means of questionnaires and observations of pupils in 
 Grammar and High School, to determine the methods of study as 
 well as the methods employed in testing their attainment. The 
 methods of study, as one might expect, were various. Some em- 
 ployed one kind or another of attempted recall, such as trying to 
 say or write the main ideas, but more than one-fourth simply read 
 the lesson through time after time. In some classes from fifty to 
 sixty per cent, of the pupils came to the recitation without having 
 once attempted to test their mastery of the lesson in any definite 
 way. 
 
 Manifestly, when so many of our years are to be spent in studying, 
 it is imperative that some information concerning such broad func- 
 tions as reading and recitation as factors in learning should be 
 obtained and applied. While volumes have been written on methods 
 of study and on the economy of learning, so far as children are 
 concerned, no objective data are available demonstrating the rela- 
 tive value of these two functions which are fundamental in any 
 attempt to learn. Earlier investigators have found in the case of 
 many adults that the optimum combination of recitation and read- 
 ing may lead to the mastery of a given lesson in one-half the time 
 required to learn it by reading alone. If such findings should hold 
 for children, and if it is generally true, as Miss Baldwin found, that 
 twenty-five per cent, or more of the pupils in the schools rely 
 entirely upon reading in their learning, the loss of time and energy 
 is appalling. 
 
 The present study presents the results of an effort to answer a 
 practical problem of the school-room — namely, What are the relative 
 values of learning by reading as compared to learning by recitation in 
 the case of school children working under school conditions and with 
 the ordinary school-room methods of attack? It will be seen later 
 that all the previous work on this subject has employed adults as 
 
 • VorUsungen zur Einfilhrung in die experimentelle Pddagogik, vol. Ill, I9i4i P- I30. 
 
 5 'Studies in Development and Learning', Archives of Psychology, 1909, No. 12, pp. 65-70. 
 
RECITATION AS A FACTOR IN MEMORIZING 3 
 
 subjects. But as Meumann has said,* "We do not know whether 
 recitation is of the same value for children, nor whether the com- 
 bination of readings and recitations for optimum results is the same 
 as for adults." In most cases, moreover, the earlier researches were 
 conducted under rather rigorously controlled conditions. The 
 subjects were not permitted to study in their habitual manner; 
 sometimes the material was presented tachistoscopically with a 
 fixed tempo of presentation, sometimes articulation was prohibited 
 or other restrictions enforced. In the present work, so far as 
 practicable, conditions were made as nearly normal as possible. 
 The material selected is comparable to that with which the pupils 
 were accustomed to deal in their daily work. The children studied 
 in much the same manner that they would employ in learning a 
 vocabulary, a spelling lesson, or a history or geography lesson, with 
 the knowledge that at the end of the study period they would be 
 given a written examination. Details of material and methods, 
 however, will be reserved for a later page. 
 
 In addition to the experiments upon school children, adult 
 subjects were also tested with similar materials and methods. The 
 data thus obtained will make possible a more adequate comparison 
 of the present findings with those of other investigations and will 
 be of assistance in better interpretation of the results by virtue of 
 the more reliable introspective observations which would be ex- 
 pected from the more experienced learners. 
 
 From this study it is hoped that some information will be secured 
 on the following points: 
 
 1. The relative value of learning by reading as compared to learning by 
 recitation. 
 
 2. The differences in the functions involved in the two methods of learning. 
 
 3. The optimum time at which to introduce recitation into the learning 
 process. 
 
 4. The relation of the two methods of learning as dependent upon the age or 
 school status of the learner. 
 
 5. The relation of the two methods as dependent upon the kind of material 
 employed. 
 
 6. Incidental information concerning the learning methods of children and 
 adults. 
 
 In the next section a brief summary of the work previously done 
 on the problem will be presented. 
 
 * op. cit., vol. Ill, p. 130. 
 
II 
 
 SUMMARY OF PREVIOUS STUDIES ON THE PROBLEM 
 
 In 1908 M. Dimitre Katzaroff made a direct attempt'^ to obtain 
 evidence upon the relative value of reading as compared to recita- 
 tion as factors in the learning process. Series of eight or ten pairs 
 of nonsense syllables were presented serially on a drum which re- 
 volved so as to expose each pair for two seconds. A four-second 
 pause was made between perusals. Adult subjects were employed 
 exclusively, each learning three or four series at each sitting, a rest 
 of five minutes being given between series. After a certain number 
 of readings attempts were made at recall by exposing the first word 
 of each pair and calling for the second word. Each syllable was left 
 in view for twenty seconds unless its associate was recalled earlier, 
 and in cases of failure to recall the learner was prompted oralty. 
 Various combinations of study and recitation were tried as may be 
 seen from Table I. 
 
 The test of memory was made by the 'Treffer' method, the 
 original first members, however, being exposed in a new order after 
 intervals of twenty-four, forty-eight, or seventy-two hours. The 
 time required for each response was measured by a Muensterberg 
 chronoscope. Table I gives a summary of the results. 
 
 The general result of these experiments is that recitation, after a 
 certain number of original readings, is more valuable than additional 
 readings. In most cases, the advantage of recitation is very great, 
 measured by the amount correctly recalled, and usually the reaction 
 time is less. Tables D and E also show quite conclusively that 
 recitations grouped are more effective than recitations interspersed 
 with readings. 
 
 The greater effectiveness of recitation is explained by Katzaroff 
 as being due in the main: First, to a greater control over the condi- 
 tions of learning. Second, to a greater activity of the learner during 
 recitation, "in the readings, the subject is passive, calm, indifferent; 
 in recitations he is active, he has to seek, he rejoices when he has 
 found and is irritated at the syllables which evade his call."^ Third, 
 to a greater confidence with reference to the material learned which 
 is brought about by recitation, 
 
 ' 'Le Role de la recitation comme facteur de la memorisation', Archives de psychologic, 1908, 7 
 
 PP.224-2S9. 
 
 2 Op. cit., p. 257. 
 
RECITATION AS A FACTOR IN MEMORIZING 5 
 
 For several reasons, it would be unsafe to consider Katzaroff's 
 findings as typical. In the first place the results were obtained from 
 too few and highly trained adults. Individual peculiarities may 
 play too prominent a role. Moreover, practice effects were not 
 sufficiently taken into account, and finally the mode of presenta- 
 tion was not the same in the two methods. During the perusals by 
 reading the total presentation was visual, but during recitation oral 
 
 TABLE I 
 
 Showing a summary of results obtained by Katzaroff, op. cit. 
 
 Table 
 
 Number 
 of subjects 
 
 Number 
 
 of sittings 
 
 for each 
 
 2 Combination 
 
 L reading 
 
 R recitation 
 
 Test after 
 
 number 
 
 hours 
 
 Per cent, 
 correct 
 
 Reaction 
 time in 
 seconds 
 
 A 
 
 I 
 
 4 
 
 L lo R L5 
 L 10 R R5 
 
 48 
 
 43 
 50 
 
 5.6 
 4.0 
 
 B 
 
 3 
 
 4 
 
 L8 L7 
 L8 R7 
 
 72 
 
 6 
 20 
 
 8.0 
 6.7 
 
 C 
 
 2 
 
 4 
 
 L8 L7 
 L8 R L6 
 
 72 
 
 9 
 15 
 
 7.8 
 8.6 
 
 D 
 
 I 
 
 3 
 
 L4 L6 
 L4 R6 
 L4 RL RL RL 
 
 24 
 
 17 
 
 46 
 
 25 
 
 50 
 2.9 
 
 4-5 
 
 E 
 
 I 
 
 3 
 
 L4 L6 
 
 L4 R3 L3 
 
 L4 RL RL RL 
 
 24 
 
 4 
 62 
 
 54 
 
 2.9 
 31 
 2.9 
 
 presentation was added to the visual, since the prompts were made 
 by means of the experimenter's voice. 
 
 A more recent experiment by Thorndike'* has given very diff"erent 
 results. Twenty-eight adult students learned four vocabularies of 
 twenty pairs each, the second by attentive reading and rereading, 
 the first by reading the first members and trying to recall the second 
 members of the pairs. The third vocabulary was learned in the 
 same manner as the second, and the fourth in the same manner as 
 the first. The results are given in Table II. 
 
 There is no apparent superiority in the method involving recall; 
 in fact, the method of reading and rereading seems to give slightly 
 better results. Professor Thorndike explains that "This, however, 
 was partly due to the overlearning of the first vocabulary, there 
 
 ' For the sake of brevity, let L equal reading and R equal an attempted recitation. Thus Lio 
 R5 means ten original readings followed by five attempted recitations. 
 
 * 'Repetitions versus Recall in Memorizing Vocabularies', Journal of Educational Psychology, 
 1914, 5, pp. 596-597. 
 
6 RECITATION AS A FACTOR IN MEMORIZING 
 
 being a tendency to take profitable risks in the vocabularies after 
 the first." And moreover, "The experiment was too crude and too 
 slight to give numerical results worth presenting in detail." 
 
 A more extensive study has recently been reported by Alexander 
 Kiihn.^ Three kinds of material were employed: vertical rows of 
 
 TABLE II 
 
 Recall 
 
 Reading 
 
 Reading 
 
 Recall 
 
 Time 
 (seconds) 
 
 Number 
 correct 
 
 Time 
 (seconds) 
 
 Number 
 correct 
 
 Time 
 (seconds) 
 
 Number 
 correct 
 
 Time 
 (seconds) 
 
 Number 
 correct 
 
 18.6 
 
 13-9 
 
 16. I 
 
 16.9 
 
 15-8 
 
 15-7 
 
 14 7 
 
 14.6 
 
 twelve nonsense syllables, vertical rows of twelve one-syllable sub- 
 stantives, and short verses. Two methods of studying were em- 
 ployed, one in which the learner read and reread until he was 
 confident of his mastery of the material and another in which the 
 subject was permitted to employ recitation as soon as he desired. 
 
 TABLE III 
 
 
 Method of 
 learning* 
 
 Tempo Controlled 
 
 Tempo Free 
 
 Material 
 
 Number of 
 repetitions 
 
 Time 
 
 Number of 
 repetitions 
 
 Time 
 
 
 Average 
 
 Median 
 
 Average 
 
 Median 
 
 Average 
 
 Median 
 
 Average 
 
 Median 
 
 Verses 
 
 R 
 
 7-5 
 
 7-3 
 
 2 '09" 
 
 2 '06" 
 
 6.5 
 
 6.4 
 
 I '51" 
 
 I '49" 
 
 190 tests 
 
 L 
 
 8.8 
 
 8.8 
 
 2 '26" 
 
 2'25" 
 
 8.2 
 
 8.2 
 
 2'I9" 
 
 2'I5" 
 
 6 subjects 
 
 L-R 
 
 1-3 
 
 15 
 
 17" 
 
 19" 
 
 1-7 
 
 1.8 
 
 28" 
 
 26" 
 
 Words 
 
 R 
 
 153 
 
 14.6 
 
 2'35" 
 
 2'3l" 
 
 II .2 
 
 10.9 
 
 2 '04" 
 
 2'0I " 
 
 132 tests 
 
 L 
 
 20.8 
 
 20.4 
 
 3 '24" 
 
 3'2i" 
 
 18.8 
 
 18.7 
 
 2 '59" 
 
 2'58" 
 
 9 subjects 
 
 L-R 
 
 5-5 
 
 5-8 
 
 49" 
 
 50" 
 
 7.6 
 
 7.8 
 
 55" 
 
 57" 
 
 Syllables 
 
 R 
 
 22.7 
 
 22.7 
 
 3'56" 
 
 3'56" 
 
 18. 1 
 
 177 
 
 3 '33" 
 
 3'38" 
 
 132 tests 
 
 L 
 
 36.6 
 
 36.7 
 
 6'o6" 
 
 6 '09 " 
 
 31.8 
 
 317 
 
 5'22 " 
 
 5'i7" 
 
 9 subjects 
 
 L-R 
 
 13 9 
 
 14.0 
 
 2'l0" 
 
 2'I3" 
 
 137 
 
 14.0 
 
 I '49" 
 
 I '49" 
 
 In some experiments the tempo of presentation was controlled by 
 means of a metronome, the subjects being required to repeat a word 
 or a nonsense syllable each 0.65 of a second, while in the case of 
 poetry a syllable was repeated each 0.4 second. In other tests, the 
 subjects studied in their preferred tempo. In all cases, simultane- 
 
 ' 'iJber Einpragung durch Lesen und durch Rezitieren', Zeitschriftfiir Psychologic, 1914,68, pp. 
 396-481. 
 
 > R — recitation, L — reading. Sum of tests for all subjects. 
 
RECITATION AS A FACTOR IN MEMORIZING 7 
 
 ous visual presentaton was employed, the subjects prompting 
 themselves and correcting their own errors. For the learning of 
 syllables and words, trochaic rhythm was specified. The subjects 
 were university graduates and instructors. Each subject learned 
 a dozen or more lessons by each of the two methods. Table III 
 gives a summary of the results. 
 
 Table IV gives the relation of reading minus recitation to recita- 
 tion, computed from the data of Table III. 
 
 TABLE IV 
 
 The Relation of Reading minus Recitation to Recitation 
 
 
 Tempo controlled 
 
 Tempo free 
 
 Material 
 
 Repetitions 
 
 Time 
 
 Repetitions 
 
 Time 
 
 
 Average 
 
 Median 
 
 Average 
 
 Median 
 
 Average 
 
 Median 
 
 Average 
 
 Median 
 
 Verses 
 Words 
 Syllables 
 
 0.17 
 0.36 
 0.61 
 
 0.20 
 0.40 
 0.62 
 
 0.13 
 0.31 
 0.55 
 
 0.15 
 0.33 
 0.56 
 
 0.26 
 0.68 
 0.76 
 
 0.28 
 0.71 
 0.79 
 
 0.25 
 0.44 
 0.51 
 
 0.24 
 0.47 
 0.52 
 
 The general result is that for all materials recitation is a more 
 effective method of learning than reading only. Rather striking 
 individual differences, however, were found, ranging all the way 
 from certain subjects who required more than twice as long to learn 
 a series of nonsense syllables when no recitation was permitted, to 
 others (in all three subjects out of thirteen) for whom reading was 
 an equally or even more efTective method of learning than recita- 
 tion. The latter are representative of a peculiar type of mechanical 
 learners to whom we shall refer again. On the whole, however, the 
 advantage of recitation is clear. 
 
 It appears that the advantage of recitation differs considerably 
 according to the kind of material being studied ; the more senseless 
 and less connected the material, the greater the advantage of 
 recitation over reading. Thus, Table IV shows the superiority of 
 recitation to be rather small in the learning of verses, about twice 
 as great for learning series of words, and larger still for learning non- 
 sense syllables. The advantage of recitation also differs according 
 to the method of studying that is employed, being in all cases 
 greater when the learner is left to employ his own tempo than when 
 the tempo is controlled by the experimenter. It was found also 
 that in the reading method the subjects were unable to judge so 
 well when the material was mastered and often 'made too early an 
 attempt to recite'. In general, it was found that the controlled 
 tempo hampered the learning to a greater or less extent. 
 
8 
 
 RECITATION AS A FACTOR IN MEMORIZING 
 
 After a lapse of twenty-four, forty-eight, and ninety-six hours, 
 the material was relearned, the same methods being employed that 
 were used in the original learning. Table V shows the average 
 results for six subjects. 
 
 TABLE V 
 
 After tiventy-foiir hours. Total number of tests — 84 
 
 
 Tempo controlled 
 
 Tempo free 
 
 Method 
 
 Repetitions 
 
 Time 
 
 Repetitions 
 
 Time 
 
 
 Average 
 
 Median 
 
 Average 
 
 Median 
 
 Average 
 
 Median 
 
 Average 
 
 Median 
 
 R 
 L 
 L-R 
 
 41 
 0.9 
 
 4.0 
 I .0 
 
 I '06 " 
 
 I '2 1 " 
 
 15" 
 
 I '05" 
 
 I '20" 
 
 15" 
 
 3-8 
 4.6 
 0.8 
 
 3.8 
 4.6 
 0,8 
 
 I '00 " 
 
 I ' 1 2 " 
 
 12" 
 
 o'58" 
 
 I'I2" 
 
 14" 
 
 
 
 After 
 
 forty-eig 
 
 hi hours. 
 
 Total niimher of tests — 68 
 
 
 R 
 
 
 3-9 
 
 3.8 
 
 I '03" 
 
 I '02 " 
 
 3-7 
 
 2, -7 
 
 o'57 " 
 
 o'55" 
 
 L 
 
 
 4-5 
 
 4-4 
 
 I '14" 
 
 i'i3" 
 
 4-5 
 
 4-3 
 
 I '12" 
 
 l'l2" 
 
 L- 
 
 -R 
 
 0.6 
 
 0.6 
 
 II " 
 
 II " 
 
 0.8 
 
 0.6 
 
 15" 
 
 17" 
 
 After ninety-six hours. Total number of tests — j8 
 
 R 
 
 3.8 
 
 3-9 
 
 I '03" 
 
 I '02 " 
 
 3-3 
 
 3-3 
 
 o'53 " 
 
 0'52 " 
 
 L 
 
 4.8 
 
 4-9 
 
 I '2 1 " 
 
 I '20" 
 
 4-4 
 
 4.6 
 
 l'l2" 
 
 i'i5" 
 
 L-R 
 
 1 .0 
 
 1 .0 
 
 18" 
 
 18" 
 
 I . I 
 
 1-3 
 
 19" 
 
 23" 
 
 The results indicate the superiority of recitation as a factor in 
 relearning but do not show that the material learned by the recita- 
 tion method is better retained than the material memorized by 
 reading alone. In the case of a few individuals, tests were made 
 after various intervals by the 'Treffer' method. From the data 
 given by Kuhn (p. 463) the following averages for ten subjects have 
 been computed: 
 
 TABLE VI . 
 
 Per cent, of material recalled after an interval of from five 
 to ten minutes. Test by ' Treffer' method 
 
 Method of 
 
 Lists of words 
 
 Lists of syllables 
 
 learning 
 
 Tempo controlled 
 
 Tempo free 
 
 Tempo controlled 
 
 Tempo free 
 
 R 
 L 
 
 L-R 
 
 66.8 
 
 59-3 
 
 7-5 
 
 72.3 
 56.7 
 15.6 
 
 36.4 
 
 254 
 II .0 
 
 32.6 
 
 22.7 
 
 9-9 
 
RECITATION AS A FACTOR IN MEMORIZING 
 
 It is apparent that the material learned by recitation is better 
 retained. In the case of words, material learned by means of a 
 fixed tempo is not so well retained as material learned by a free 
 tempo, but this result is not clear for nonsense material, for which 
 the opposite, if anything, is true. 
 
 Three subjects were each given twelve tests by the 'TrefTer' 
 method first after ten minutes and again with other material after 
 twenty-four hours. The results are given below. 
 
 TABLE VII 
 
 Per cent, of material recalled after ten-minute or twenty-four-hour lapses 
 Studying done with free tempo 
 
 After ten minutes 
 
 Reading 
 Recitation 
 
 33-5 
 
 44.1 
 
 After twenty-four hours 
 
 Reading 
 Recitation 
 
 6.2 
 
 II 9 
 
 From this data, Kiihn concludes "the advantage of learning with 
 recitation for retention is much greater after a pause of a day than 
 after a pause of a few minutes." "^ 
 
 Kuhn found a great deal of interesting information concerning 
 individual peculiarities in learning. Many individuals had a con- 
 stant tendency to begin to recite too soon, or too late, for the best 
 results; some were unable to limit themselves to 'pure' reading, 
 more or less recitation unintentionally creeping in; some subjects 
 were found who seemed to derive no benefit whatever from con- 
 tinued readings beyond a maximum of four or five; and others 
 obtained better results under the reading method than when 
 recitation was a factor. Kiihn found the latter subjects to employ 
 a peculiar form of 'automatic' or 'mechanical' method of learning, 
 in which the usual method of building up associations between 
 items and binding them into some form of compact 'schema' or 
 structure was not employed. Simple visual imprinting was the 
 most effective procedure. 
 
 Kiihn came to the general conclusion, "that recitation is more 
 effective because it leads to a more fundamental, many-sided work- 
 ing over of the material" (p. 443). In recitation the items are more 
 attentively observed, the list is more carefully analysed, striking 
 words are picked out, and a better 'schema' of reconstruction is 
 employed. In the case of those individuals who rely upon the 
 various sorts of associative aids in learning, recitation is very 
 helpful, but the few who learn mechanically can do as well or better 
 by merely reading. 
 
 ' Op. cit., p. 466. 
 
10 
 
 RECITATION AS A FACTOR IN MEMORIZING 
 
 So far no information has been cited with regard to thie stage at 
 which it is best to introduce the first recitation. This question was 
 taken up and answered, in a measure, by Witasek.^ Rows of ten 
 pairs of nonsense syllables were exposed successively at the rate of 
 one per second by means of a Wirth memory apparatus. The 
 subject studied aloud both in reading and recitation, and all cor- 
 rections and promptings were made orally by the experimenter. 
 
 TABLE VIII (from Witasek, p. 267) 
 
 Showing the superiority of a group of recitations over a 
 group of readings, absolutely and relatively ^ 
 
 Number of 
 preliminary 
 
 Superiority in 
 number of repetitions 
 
 Superiority in 
 number of seconds 
 
 Superiority in 
 number of prompts 
 
 readings 
 
 Absolutely 
 
 Relatively 
 
 Absolutely 
 
 Relatively 
 
 Absolutely 
 
 Relatively 
 
 Of five recitations over five readings 
 
 6 
 II 
 16 
 
 2.3 
 2.0 
 
 1-4 
 
 90 per cent. 
 80 percent. 
 70 per cent. 
 
 87 
 73 
 54 
 
 90 percent. 
 80 percent. 
 70 per cent. 
 
 80 percent. 
 
 100 percent. 
 
 60 per cent. 
 
 Of ten recitations over ten readings 
 
 6 
 II 
 
 1-5 
 2.0 
 
 70 percent. 
 60 per cent. 
 
 62 
 61 
 
 70 per cent. 
 60 per cent. 
 
 5 
 3 
 
 80 per cent. 
 60 per cent. 
 
 
 
 Of fifteen recitations over fifteen readii 
 
 igs 
 
 
 6 
 
 1.8 
 
 60 percent. 
 
 62 
 
 60 percent. 
 
 4 
 
 60 per cent. 
 
 In the case of recitation, however, the first syllables of the pairs 
 were spoken by the experimenter. An interval of six to seven 
 seconds was given between repetitions. Three series of syllables 
 were learned each day, a three-minute rest being allowed between 
 lessons. The method of computation used by Witasek is, in many 
 cases, somewhat confusing. He frequently makes use of the term 
 'imprinting value' ('Einpragungswert') which means the value of 
 
 ' 'fiber Lesen und Rezitieren in ilirer Beziehungen zum G^dikchtms' , Zeitschrif If iir Psychologic, 
 1907, 44, pp. 161-18S, 246-278. 
 
 9 To make the meaning of the table clearer the procedure may be described more in detail. To 
 begin with, the lists were read six, eleven, or sixteen times as indicated. One hour later, the lists 
 were in one case read, and in the other case, recited, five, ten, or fifteen times as indicated, and 
 immediately after the learning of the lists was completed by further recitations. From these data, 
 the saving in the total time required to learn was computed for the groups of readings and for the 
 groups of recitations. The table above presents, absolutely and relatively, the differences between 
 the savings brought about by reading and by recitation — the differences always being in favor of the 
 groups of recitations as shown. 
 
RECITATION AS A FACTOR IN MEMORIZING 
 
 II 
 
 a repetition in reducing the time or repetitions needed to complete 
 the learning. Table VIII shows the superiority of a group of recita- 
 tions over an equal group of readings, after a given number of 
 preliminary readings, in reducing the time required after an inter- 
 val of an hour, to complete the learning so that the entire list could 
 be recited without prompts in ten seconds or less. The table is 
 based upon the work of seven university graduates and faculty 
 members, tested three or four times, a total of twenty-four tests. 
 The table shows in summary form a finding which is demonstrated 
 in more detail by Witasek, e. g., that the imprinting value of 
 successive readings declines very rapidly after the first few. That 
 
 TABLE IX (from Witasek, pp. 184-185) 
 
 Number of 
 original 
 readings 
 
 Number of 
 additional 
 readings 
 
 Number of 
 recitations 
 
 Total 
 repetitions 
 
 Speed of 
 
 recitations 
 
 after one hour 
 
 Number of 
 prompts 
 
 6 
 
 
 
 
 
 6 
 
 78" 
 
 7-9 
 
 6 
 
 5 
 
 
 
 II 
 
 75" 
 
 7 
 
 2 
 
 6 
 
 
 
 5 
 
 II 
 
 63" 
 
 6 
 
 3 
 
 6 
 
 
 
 10 
 
 16 
 
 69" 
 
 5 
 
 8 
 
 6 
 
 10 
 
 
 
 16 
 
 74" 
 
 7 
 
 5 
 
 6 
 
 5 
 
 5 
 
 16 
 
 66" 
 
 6 
 
 
 
 6 
 
 
 
 15 
 
 21 
 
 66" 
 
 5 
 
 5 
 
 6 
 
 15 
 
 
 
 21 
 
 73" 
 
 6 
 
 7 
 
 6 
 
 5 
 
 10 
 
 21 
 
 65" 
 
 5 
 
 9 
 
 6 
 
 10 
 
 5 
 
 21 
 
 66" 
 
 5 
 
 7 
 
 6 
 
 10 
 
 10 
 
 26 
 
 69" 
 
 5 
 
 7 
 
 6 
 
 5 
 
 15 
 
 26 
 
 65" 
 
 6 
 
 2 
 
 is to say, readings are pronouncedly subject to a law of diminishing 
 returns. In this respect, recitations proved to be a better form of 
 repetition. It is apparent, from the table, that recitations intro- 
 duced into the learning at almost any point are of more value than 
 continued readings. However, the superiority of recitation seems 
 to be somewhat greater when introduced after six than when 
 introduced after eleven or sixteen readings. This would seem to 
 indicate that recitations, too, are subject decidedly to the law of 
 diminishing returns. But Witasek explains that in these particular 
 tests with a large number of repetitions "the readings unconsciously 
 become very similar to recitations." The learner, finding the read- 
 ings to become more and more fruitless, is unable to restrain a 
 natural inclination to partially recite. 
 
 So far experiments have merely confirmed the current opinion 
 that recitations, if not introduced too early in the learning process, 
 
12 
 
 RECITATION AS A FACTOR IN MEMORIZING 
 
 are of more value than continued readings. It remains to enquire 
 into the combination that will yield the richest returns in propor- 
 tion to the outlay of time and energy. Table IX shows the rela- 
 tive effectiveness of several combinations as measured by the speed 
 of the first recitation after an interval of an hour, together with the 
 number of prompts. The procedure in this recitation was as follows: 
 The first member of each pair was exposed, the subject responding 
 with the second member, whereupon the first member of the next 
 pair was exposed and so on. If the subject responded incorrectly, 
 he was corrected by the experimenter, and if the subject could not 
 
 TABLE X (from Witasek, p. i84f) 
 
 Number of 
 original 
 readings 
 
 Number of 
 additional 
 readings 
 
 Number of 
 recitations 
 
 Total 
 repetitions 
 
 Speed of 
 third recitation 
 after one hour 
 
 Number of 
 prompts 
 
 6 
 
 O 
 
 O 
 
 6 
 
 37 
 
 1.8 
 
 6 
 
 5 
 
 O 
 
 II 
 
 34" 
 
 1-7 
 
 6 
 
 o 
 
 5 
 
 II 
 
 22 " 
 
 G.8 
 
 6 
 
 o 
 
 10 
 
 i6 
 
 20" 
 
 -0.8 
 
 6 
 
 10 
 
 
 
 i6 
 
 32" 
 
 1.6 
 
 6 
 
 5 
 
 5 
 
 i6 
 
 19" 
 
 0.5 
 
 6 
 
 
 
 15 
 
 21 
 
 15" 
 
 0.5 
 
 6 
 
 15 
 
 o 
 
 21 
 
 27" 
 
 0.5 
 
 6 
 
 5 
 
 10 
 
 21 
 
 18" 
 
 0.7 
 
 6 
 
 10 
 
 5 
 
 21 
 
 19" 
 
 0.7 
 
 6 
 
 lO 
 
 10 
 
 26 
 
 20" 
 
 0.9 
 
 6 
 
 5 
 
 15 
 
 26 
 
 16" 
 
 0.4 
 
 respond at all within ten seconds, the experimenter gave the response 
 orally and exposed the next syllable in the series. 
 
 The results shown in this table are not very clean cut and in some 
 respects are rather surprising. Within the various groups showing 
 an equal number of repetitions, it is quite clear that a combination 
 of recitations with readings leads to a more successful recitation 
 after an hour than when reading only was employed in the study 
 period. The advantage does not appear to be very great, however. 
 What is quite surprising is that a small number of repetitions of any 
 sort (six or eleven) leads to nearly as effective a recitation after an hour 
 as a larger number (sixteen, twenty-one, or twenty-six). From this 
 it would appear that repetitions beyond eleven are pretty largely wast- 
 ed, and accordingly recitation, contrary to Witasek's earlier conten- 
 tion, must be subject also to a law of greatly diminishing returns. 
 
 The above table (Table X) which was computed from Wita- 
 sek's original data, shows that the results for the third attempted 
 
RECITATION AS A FACTOR IN MEMORIZING 
 
 13 
 
 recitation after the interval of an hour are quite difTerent from those 
 based on the first recitation after the interval. The second, fourth, 
 or fifth recitation would have shown a similar difTerence. 
 
 In the case of these later repetitions, the advantage of recitation 
 as a factor in the original learning is quite pronounced. It is 
 apparent also, that although the law of diminishing returns is still 
 seen to operate, its influence is very much less marked than appeared 
 in the results for the first recitation after the interval. 
 
 Table XI exhibits the results in terms of the total time required 
 to learn the series in two sittings separated by an hour. 
 
 TABLE XI (from Witasek, p. 274) 
 
 Work of the first sitting 
 
 Work of the second sitting 
 
 Combina- 
 tion 
 
 Time of 
 reading 
 
 Time of 
 reciting 
 
 Pauses 
 
 between 
 
 repetitions 
 
 Time in 
 recitations 
 
 Pauses 
 
 between 
 
 recitations 
 
 Sum with 
 pauses 
 
 Sum 
 without 
 pauses 
 
 L6 Ro 
 Lii Ro 
 L16 Ro 
 L21 Ro 
 L6 R5 
 L6 Rio 
 L6 R15 
 Lii R5 
 Lii Rio 
 Lii R15 
 L16 R5 
 L16 Rio 
 
 60 
 
 no 
 
 160 
 
 210 
 
 60 
 
 60 
 
 60 
 
 no 
 
 no 
 
 no 
 
 160 
 
 160 
 
 
 
 
 
 
 
 
 
 96 
 
 166 
 
 206 
 
 80 
 
 119 
 
 142 
 
 69 
 
 121 
 
 35 
 70 
 105 
 140 
 70 
 105 
 140 
 105 
 140 
 
 175 
 140 
 
 175 
 
 262 
 236 
 228 
 202 
 
 163 
 133 
 
 124 
 109 
 142 
 153 
 
 56 
 49 
 
 42 
 
 42 
 28 
 
 35 
 28 
 28 
 21 
 
 14 
 21 
 
 28 
 
 413 
 465 
 535 
 594 
 397 
 529 
 567 
 468 
 
 514 
 
 550 
 532 
 637 
 
 322 
 346 
 388 
 412 
 299 
 389 
 
 399 
 335 
 353 
 361 
 371 
 434 
 
 Table XII shows the data of Table XI rearranged, the combina- 
 tions being arrayed in the order of their effectiveness with the 
 percentages of time devoted to reading and to recitation. 
 
 The findings indicate that a small amount of work at the first 
 sitting pays better than a large amount; that is to say, the series 
 can be learned more quickly in the end if only a small proportion 
 of the total time is devoted to the first study while the larger portion 
 is saved for the review an hour later. There is considerable evi- 
 dence that better results are obtained if the original study period 
 is partly devoted to recitation; for example, 6 Ls plus 5 Rs gives 
 better results than 11 Ls; 11 Ls plus 5 Rs is much better than 16 
 Ls and so on. However, the most potent factor is the distribution 
 of the recitations. The best results are obtainable when the original 
 period includes about twenty-five per cent, of the total learning 
 
14 
 
 RECITATION AS A FACTOR IN MEMORIZING 
 TABLE XII (based on Table XI; 
 
 
 First sitting 
 
 Second sitting 
 
 Combination 
 
 Per cent, of 
 time for 
 reading 
 
 Per cent, of 
 time for 
 recitation 
 
 Per cent, of 
 time for 
 recitation 
 
 Sum with 
 pauses 
 
 Sum without 
 pauses 
 
 L6 R5 
 
 24 
 
 33 
 
 43 
 
 397 
 
 299 
 
 L6 Ro 
 Lii Ro 
 
 23 
 39 
 
 
 
 
 
 11 
 61 
 
 413 
 465 
 
 322 
 346 
 
 Lii R5 
 
 39 
 
 24 
 
 37 
 
 468 
 
 335 
 
 Lii Rio 
 
 35 
 
 37 
 
 28 
 
 514 
 
 353 
 
 L6 Rio 
 
 18 
 
 45 
 
 37 
 
 529 
 
 389 
 
 Li6 R5 
 Li6 Ro 
 
 50 
 50 
 
 20 
 
 
 30 
 50 
 
 532 
 
 535 
 
 371 
 
 388 
 
 Lii R15 
 L6 R15 
 
 33 
 17 
 
 45 
 55 
 
 22 
 28 
 
 550 
 567 
 
 361 
 399 
 
 L21 Ro 
 
 59 
 
 
 
 41 
 
 594 
 
 412 
 
 L16 Rio 
 
 42 
 
 30 
 
 28 
 
 637 
 
 434 
 
 time, allotting about half of this time each to reading and to recita- 
 tion. Beyond this amount, recitations introduced into the review 
 are much more effective than recitations in the first period of study. 
 
 On the whole, so far as the matter of the relative values of reading 
 and recitation as factors in learning are concerned, Witasek's 
 method of attack is subject to several defects. In the first place, 
 too many variable factors are introduced. The influence of various 
 divisions of the lesson, without regard to the methods of study 
 employed, makes interpretation difficult. Moreover, as Witasek 
 himself points out, the conditions of the experiment were such that 
 the readings, especially after a number of perusals, became, unin- 
 tentionally, very much like recitations. Witasek's procedure may also 
 be charged with most of the defects found in Katzaroff's experi- 
 ments; in reading, the subject prompted himself by looking at the 
 forgotten syllable, but during recitation, promptings were made 
 orally by the experimenter; the subjects were few and all were 
 trained adults, whose habitual mode of studying may have been 
 seriously interfered with by the particular conditions of the experi- 
 ment; and finally practice effects were by no means fully eliminated. 
 
 From his study, Witasek drew the general conclusion that recita- 
 tion, as compared to reading, is a much more effective method of 
 study. The difference in favor of recitation was attributed in the 
 main to a 'higher, degree of attention' which was made possible by 
 virtue of the opportunity afforded the subject to gauge his progress 
 in the learning and apply himself to the portions that offered dififi- 
 
RECITATION AS A FACTOR IN MEMORIZING 1 5 
 
 culty. The higher grade of attentiveness is closely correlated with 
 an apparently 'greater activity' shown during recitation. In read- 
 ing the subject is likely to relax into a state of passive receptivity, 
 in recitation, the attitude is one of alert, searching {'sich besinnen') 
 activity. 
 
 In an experiment by Miss Abbott,^" the problem has been attacked 
 from a somewhat different point of view. Miss Abbott endeavored 
 to determine the learning types of a limited number of individuals 
 and to utilize this information in the interpretation of the numerical 
 results. As material, lists of thirty nonsense syllables and sixty 
 English words were used. An apparatus was provided such that 
 the words or syllables could be exposed singly for any time 
 desired. A fixed time (sixteen minutes) was allowed for the 
 study period, this time being divided up into various combinations 
 of reading and recall. 
 
 The groups of words and syllables were presented in various ways 
 as shown in Table XIII. 
 
 TABLE XIII 
 
 Series 
 
 First eight minutes 
 spent in 
 
 Exposure 
 
 time per 
 
 item 
 
 Interval 
 
 between 
 
 items 
 
 Interval be- 
 tween first and 
 second learning 
 period 
 
 Second eight min- 
 utes spent in 
 
 a 
 
 visual imprinting 
 
 I" 
 
 O 
 
 l' 
 
 visual imprinting 
 
 b 
 
 " 
 
 I " 
 
 O 
 
 15' 
 
 " 
 
 c 
 d 
 
 u 
 
 I " 
 I" 
 
 o 
 o 
 
 45' 
 I' 
 
 Recall 
 
 e 
 
 " 
 
 I " 
 
 o 
 
 15' 
 
 " 
 
 f 
 
 " 
 
 I " 
 
 o 
 
 45' 
 
 u 
 
 g 
 
 " 
 
 I " 
 
 
 
 
 
 h 
 
 visual imprinting 
 and recall 
 
 I" 
 
 3" 
 
 I' 
 
 visual imprinting 
 and recall 
 
 i 
 
 visual imprinting 
 and recall 
 
 I" 
 
 3" 
 
 15' 
 
 visual imprinting 
 and recall 
 
 J 
 
 visual imprinting 
 and recall 
 
 I " 
 
 3" 
 
 45' 
 
 visual imprinting 
 and recall 
 
 k 
 
 visual imprinting 
 and recall 
 
 I " 
 
 3" 
 
 
 
 All series were allotted a sixteen-minute study period except 
 series g and k which received but eight minutes. In series a, b, c, 
 and g no opportunity is given for recall, the whole time being spent 
 in 'Einpragung'; in series d, e, and f the first eight minutes is 
 
 1" 'On the Analysis of the Factor of Recall in the Learning Process', Psychological Review 
 Monograph, 1909, 11, pp. 159-177. 
 
i6 
 
 RECITATION AS A FACTOR IN MEMORIZING 
 
 Spent in imprinting followed by eight minutes of recall; while in 
 series h, i, and j, three-fourths of the time is devoted to recall, which 
 is interspersed with the 'Einpragung' occupying the three-second 
 intervals indicated in the table. 
 
 In all cases the subjects worked under certain restrictions. 
 During the presentation of the material in series a, b, c, d, e, f, and 
 g, the subject was not to form any associative links between the 
 items and while one item was before him, he was not to think of 
 another. During the recall period in the d, e, f series, while the 
 subject was permitted to image the items and form such associations 
 
 TABLE XIV (from Abbott, p. 173) 
 Percentages of words and syllables correctly recalled after four hours 
 
 Series 
 
 a 
 
 b 
 
 c 
 
 d 
 
 e 
 
 f 
 
 g 
 
 h 
 
 i 
 
 j 
 
 k 
 
 Subject 
 
 Words 
 
 23 
 
 23 
 
 24 
 
 23 
 
 33 
 
 25 
 
 23 
 
 41 
 
 59 
 
 42 
 
 8 
 
 V 
 
 Syllables 
 
 42 
 
 38 
 
 40 
 
 35 
 
 40 
 
 17 
 
 22 
 
 68 
 
 92 
 
 70 
 
 47 
 
 
 Words 
 
 8 
 
 8 
 
 13 
 
 29 
 
 12 
 
 8 
 
 5 
 
 97 
 
 98 
 
 97 
 
 62 
 
 w 
 
 Syllables 
 
 20 
 
 15 
 
 12 
 
 20 
 
 30 
 
 27 
 
 8 
 
 97 
 
 97 
 
 98 
 
 62 
 
 
 Words 
 
 15 
 
 28 
 
 23 
 
 28 
 
 15 
 
 18 
 
 13 
 
 31 
 
 34 
 
 30 
 
 18 
 
 X 
 
 Syllables 
 
 13 
 
 13 
 
 22 
 
 22 
 
 17 
 
 18 
 
 5 
 
 30 
 
 27 
 
 48 
 
 10 
 
 
 Words 
 
 67 
 
 78 
 
 83 
 
 48 
 
 42 
 
 43 
 
 28 
 
 53 
 
 55 
 
 17 
 
 23 
 
 Y 
 
 Syllables 
 
 67 
 
 53 
 
 70 
 
 67 
 
 63 
 
 50 
 
 57 
 
 63 
 
 67 
 
 53 
 
 43 
 
 
 Words 
 
 54 
 
 
 
 61 
 
 
 
 
 56 
 
 
 
 
 Z 
 
 Syllables 
 
 60 
 
 
 
 50 
 
 
 
 
 55 
 
 
 
 
 
 as he wished, he was not allowed to pronounce them or to write 
 them down. 
 
 Five students of psychology acted as subjects in the tests. The 
 image type of each was determined as follows: Subject F employed 
 inner speech and concrete visual imagery; Subject W was of a 
 motor-auditory verbal type; X, motor-auditory with some visual 
 imagery; Y was strongly visual, never pronounces a word, just lets 
 it 'soak in'; and Z was of mixed type, employing different kinds 
 of imagery at different times. 
 
 The test of memory consisted in requiring the subjects to write, 
 four hours after the test, all the words or syllables they could 
 remember. Table XIV shows the results in the form of percentages 
 of the total lists that were correctly reproduced. 
 
 Table XV was derived from Table XIV by subtracting the 
 results obtained in series g, from the results obtained in series a, b, 
 c, d, e, f, respectively. This table, consequently, presents the gain 
 
RECITATION AS A FACTOR IN MEMORIZING 
 
 17 
 
 brought about by the second eight minutes of reading or recall as 
 compared to the results obtained by the first eight minutes imprint- 
 ing alone. 
 
 TABLE XV (from Abbott, p. 173) 
 Showing the advantage of sixteen minutes study over eight minutes 
 
 Series 
 
 a 
 
 b 
 
 c 
 
 d 
 
 e 
 
 f 
 
 Subject 
 
 Words 
 
 
 
 
 
 I 
 
 
 
 10 
 
 2 
 
 V 
 
 Syllables 
 
 20 
 
 16 
 
 18 
 
 13 
 
 18 
 
 —5 
 
 
 Words 
 
 3 
 
 3 
 
 8 
 
 24 
 
 7 
 
 3 
 
 w 
 
 Syllables 
 
 12 
 
 7 
 
 4 
 
 12 
 
 22 
 
 19 
 
 
 Words 
 
 2 
 
 15 
 
 10 
 
 15 
 
 2 
 
 5 
 
 X 
 
 Syllables 
 
 8 
 
 8 
 
 17 
 
 17 
 
 12 
 
 13 
 
 
 Words 
 
 39 
 
 50 
 
 55 
 
 20 
 
 14 
 
 15 
 
 Y 
 
 Syllables 
 
 10 
 
 —4 
 
 13 
 
 ID 
 
 6 
 
 7 
 
 
 Table XVI gives the results for words and syllables combined 
 together with the averages for a, b, c ; d, e, f ; and h, i, j, respectively, 
 based on the data from Table XIII. 
 
 TABLE XVI (from Abbott, p. 174) 
 Showing the combined results for words and syllables 
 
 Seriep 
 
 a 
 
 b 
 
 c 
 
 Aver- 
 age 
 
 d 
 
 e 
 
 f 
 
 Aver- 
 age 
 
 li 
 
 i 
 
 j 
 
 Aver- 
 age 
 
 g 
 
 k 
 
 Sub- 
 ject 
 
 
 29 
 
 28 
 
 29 
 
 29 
 
 21 
 
 35 
 
 22 
 
 29 
 
 50 
 
 70 
 
 51 
 
 56 
 
 23 
 
 21 
 
 V 
 
 
 12 
 
 II 
 
 12 
 
 II 
 
 26 
 
 18 
 
 14 
 
 19 
 
 97 
 
 98 
 
 97 
 
 97 
 
 6 
 
 62 
 
 w 
 
 
 14 
 
 23 
 
 23 
 
 20 
 
 26 
 
 16 
 
 18 
 
 20 
 
 31 
 
 32 
 
 36 
 
 35 
 
 10 
 
 16 
 
 X 
 
 
 67 
 
 70 
 
 79 
 
 72 
 
 54 
 
 49 
 
 46 
 
 49 
 
 57 
 
 59 
 
 29 
 
 48 
 
 38 
 
 30 
 
 Y 
 
 From Table XV it appears that with the exception of three 
 cases the additional eight minutes of reading or recall results in a 
 greater amount of material recalled, and for all subjects except Y, 
 the value of the additional study is more pronounced in learning 
 nonsense syllables than in learning words. 
 
 The most significant comparisons appear in Table XV. Subject 
 W, of auditory-motor type, does much better in series h, i, j, than 
 in d, e, f, which in turn gives better results than a, b, c. That is, the 
 methods restricting learning to visual imprinting alone are the 
 poorest of all; the method giving eight minutes of imprinting 
 followed by eight minutes recall is much better; while the method 
 
1 8 RECITATION AS A FACTOR IN MEMORIZING 
 
 giving three-fourths of the time to interspersed recall gives results 
 about nine times as good as the first. Subjects V and X agree in 
 showing sixteen minutes of visual imprinting to be as effective as 
 eight minutes of imprinting followed by eight minutes of re- 
 call, but each shows to better advantage when three-fourths 
 of the time is spent in interspersed recall (series h, i, j). 
 These subjects ordinarily employed auditory-motor imagery or 
 inner speech and were undoubtedly greatly hampered by some 
 of the restrictions placed upon them in the d, e, f series. 
 Subject Y, who possessed strong visual imagery, learning by sim- 
 ply allowing the items to 'soak in', does very well in the method 
 of visual imprinting and very poorly in either method employ- 
 ing recall. 
 
 From this study, Miss Abbott draws the following conclusions: 
 
 1. That the factor of recall is always an aid in the learning process. 
 
 2. That when recall comes after the Einprdgung of the material, immediate 
 recall is of more value than delayed recall and its value decreases as the delay 
 increases in length. 
 
 3. That the recall is of greater value when it is interspersed with the Ein- 
 prdgung. 
 
 4. That localization is one of the factors which go to make recall an aid to 
 memory, but that the relative importance of this factor is determined by indi- 
 vidual type. 
 
 5. That the relative value of recall and Einprdgung depends on individual 
 type. 
 
 To the present writer, it seems that the third conclusion, e. g., 
 "That recall is of greater value when it is interspersed with the 
 Einpragung," is not entirely borne out, at least not in such form as 
 to be applicable to every-day, non-restricted methods of learning. 
 In the first place, the methods employing the interspersed recall 
 devote twenty-five per cent, more time to it than do the methods 
 in which the recalls are grouped. Again, the severe restrictions 
 placed on the first eight minutes of learning by reading in the series 
 a, b, c, d, e, and f are avoided in the series in which recall is inter- 
 spersed with reading. It will be recalled that Katzaroff in experi- 
 ments in which restrictions were less severe, and employing a 
 larger number of subjects, found that recitations grouped gave 
 better results than recitations interspersed with readings. (See 
 
 P- 5-) 
 
 A study by Clemens Knors,^^ although not primarily concerned 
 
 with the present problem, contributes some information concerning 
 
 three different methods of memorizing paired material. Method A 
 
 is similar to the 'reconstruction method' introduced by Miss 
 
 11 'Experimentelle Untersuchungen uber den Lernprozess", Archive f. d. g. Psychologic, 1910, 
 17, pp. 297-362. 
 
RECITATION AS A FACTOR IN MEMORIZING IQ 
 
 Gamble. ^^ The series was first read through once, the subject then 
 attempted to recite both members of the pairs; the series was then 
 read again, followed by another attempt at reconstruction and so 
 on until learned. Method B was the same except that, in recitation, 
 the first members of the pairs were exposed, the learner attempting 
 to recite the second members only. In Method C the subject read 
 and reread the series until he felt that they were mastered ; where- 
 upon he was tested by exposing the first members of the pair as in 
 Method B. It will be noted that Method A permits the recitation 
 of both members of the pairs, Method B of but the first member, 
 while Method C permits no recitation during the learning. 
 
 In all methods the subject read or recited aloud, the number of 
 perusals and the number of promptings being recorded. The scores 
 are given in the form of the total amount of material that was read 
 by the subject plus the amount supplied him in the form of prompt- 
 ings or corrections by the experimenter. The following sample will 
 show how the score was computed. Suppose a series of fourteen 
 syllables is learned by eight readings plus seven attempted recita- 
 tions, during which fifty-one syllables were supplied by the experi- 
 menter. Then the total score would be eight (the number of read- 
 ings) plus fifty-one (the total number of prompts) divided by 
 fourteen (the number of syllables in the list). That is, score = 
 8 + 51/14 = 11.64. 
 
 From the original data given by Knors, the results shown in 
 Table XVII have been computed. Unfortunately Knors did not 
 print all of the raw data that he collected, so that some of the tables 
 are incomplete. The subjects are three adults (A, B, C) and four 
 children (a, b, c, d) eleven to thirteen years of age. The table 
 presents the average score of three or four tests for each individual. 
 
 Although the results are somewhat irregular, a few points can 
 be made out. Sections H and / indicate that, for adults, Method 
 A, which requires the recitation of both members of the pairs, is 
 superior to Method B. in which but the second member is recited. 
 For Subjects A and C the differences are very great. The same 
 subjects, however, show but a slight superiority of Method A over 
 Method C in which reading alone was involved. 
 
 Although the findings for the children are very irregular, some 
 differences between the methods seem clearly to appear. When 
 the series of nonsense syllables to be learned is long (Section L), 
 Methods A and B are both superior to C, which permits reading 
 only; but when the series is short (Section M) the differences are 
 very small. The differences between Methods A and B in either 
 
 " 'A Study in Memorizing Various Material by the Reconstruction Method', Psychological 
 Review Monograph, 1909, 10, No. 4. 
 
20 
 
 RECITATION AS A FACTOR IN MEMORIZING 
 
 case arc so small as to be negligible. In the case of senseful words, 
 the advantage of Method A over C is very great, the two standing, 
 for different individuals, in various ratios ranging from seven to 
 five up to four to one. It appears in general, then, that children, 
 as compared to adults, profit much more through the employment 
 of recitation in learning. 
 
 However, but little reliability can be placed upon the scanty 
 findings of these experiments. The number of subjects is too small 
 and the quantitative results are too meager. The time of the various 
 readings and recitations not being kept, there is some doubt whether 
 any of the methods would show a distinct advantage with respect 
 
 TABLE XVII (from Knors) 
 
 Sub- 
 jects 
 
 H 
 
 Fourteen nonsense 
 syllables. Methods 
 
 I 
 
 Eighteen nonsense 
 syllables 
 
 J 
 
 Ten sense 
 words 
 
 K 
 
 Eighteen sense 
 words 
 
 
 A 
 
 B 
 
 c 
 
 A 
 
 B 
 
 C 
 
 A 
 
 B 
 
 C 
 
 A 
 
 B 
 
 c 
 
 B 
 C 
 
 6.9 
 10.4 
 
 9 5 
 
 12.3 
 
 145 
 13,2 
 
 
 7-9 
 II . I 
 
 7-9 
 
 14.8 
 13.2 
 12.2 
 
 
 4-5 
 
 3-9 
 
 41 
 
 
 4.8 
 3.8 
 41 
 
 5-2 
 
 4 7 
 
 3.8 
 
 
 7-2 
 6.8 
 56 
 
 Sub- 
 jects 
 
 L 
 
 Fourteen nonsense 
 syllables. Methods 
 
 M 
 
 Eight nonsense 
 syllables 
 
 N 
 
 Ten sense 
 words 
 
 
 
 Fourteen sense 
 words 
 
 
 A 
 
 B 
 
 c 
 
 A 
 
 B 
 
 c 
 
 A 
 
 B 
 
 C 
 
 A 
 
 B 
 
 C 
 
 a 
 b 
 c 
 d 
 
 17.9 
 II . I 
 III 
 10.5 
 
 12.2 
 10.8 
 
 13.3 
 13.6 
 
 23.7 
 29.3 
 30.7 
 
 19.6 
 
 14. I 
 
 9-1 
 
 7.8 
 
 10.6 
 
 8.9 
 
 II. 3 
 
 8.7 
 9-9 
 
 12.9 
 
 12.0 
 
 131 
 II .2 
 
 5.7 
 
 4-3 
 5.0 
 6.0 
 
 
 7-3 
 10. 1 
 19.2 
 11-5 
 
 6.5 
 5-9 
 6.1 
 7.8 
 
 
 12.5 
 19 5 
 
 22.4 
 17.8 
 
 to the total amount of time required to learn. Unfortunately 
 Knors did not print all of his raw data, but from what does appear, 
 it is clear that the variability of the performances, especially those 
 of the children, is very great. Subject A, for example, in three tests 
 of learning series of nonsense syllables by Method A, shows an 
 average score of 9.0 with a P. E. of 2.2. On the whole, it would not 
 be safe to consider Knors's results as more than suggestive. 
 
 SUMMARY OF RESULTS 
 
 Without doubt, this brief enumeration of the results of such a 
 medley of experiments has left but a vague impression with regard 
 to present status of the problem under consideration. Perhaps an 
 
RECITATION AS A FACTOR IN MEMORIZING 21 
 
 effort to summarize the findings will assist somewhat to a better 
 understanding. Such an effort, however, is fraught with difficulty. 
 When one considers the individual differences possible among the 
 subjects, the variations in materials, in method of presentation, 
 and in the methods of scoring and the like, it can be readily under- 
 stood that direct comparison of many studies is quite out of the 
 question. Perhaps it will be worth while, first, to review the 
 methods of attack employed in the several studies and, by throwing 
 the differences into relief, pave the way for a concise summary of 
 the outstanding results that will then be presented and for an 
 understanding of the relation of the present study to those which 
 have gone before. 
 
 1. Differences as regards materials. Most of the studies have 
 employed the method of paired associates; as material, nonsense 
 syllables in pairs, senseful words in pairs, digits paired with non- 
 sense syllables, and foreign words paired with the vernacular have 
 been used. Single series of senseless or senseful words of various 
 lengths have also been used, and, in some cases, connected sense 
 material such as prose or poetry. It is possible that the results 
 might differ considerably according to the kind of material used; 
 in fact, Kiihn and Knors found that this was decidedly the case. 
 
 2. Subjects. With the exception of the few experiments with 
 four boys, conducted by Knors, well educated adults have been 
 employed. In nearly every case, moreover, the number of subjects 
 has been entirely too small to eliminate differences which might be 
 due to the influence of previously acquired habits of study as well 
 as the more innate differences such as those considered by Miss 
 Abbott. 
 
 3. Methods of presenting the material. In most cases, the material 
 has been presented visually, but as was noted above (p. 5) some- 
 times the method of presentation changes within the lesson. Both 
 Witasek and Katzaroff presented their material visually when the 
 subject was reading, but during the recitation the material was 
 presented orally. In some cases, the material is printed large, in 
 some, small; in some it is held in the hand; in others, it is at a 
 distance or thrown on a screen; sometimes the material is pre- 
 sented simultaneously, sometimes serially. The tempo of presenta- 
 tion is an important matter also. In nearly all cases, the tempo 
 was controlled and varies greatly from experimenter to experimenter. 
 Kiihn presented syllables at the rate of one each 0.4 second, Abbott 
 one per second, and Katzaroff one every two seconds. No one 
 knows how closely these rates corresponded to the habitual tempo 
 of the learner, and, what is more, as was shown b}^ Kiihn and as we 
 shall see again later, the natural tempo of recitation is considerably 
 
22 RECITATION AS A FACTOR IN MEMORIZING 
 
 slower, on the average, than that of reading. In short, the results 
 are affected somewhat by the rate of presentation, influence of which 
 is likely to be different upon reading than upon recitation. 
 
 4. Methods of reciting or recalling. Aside from employing a 
 fixed rate of presentation of the material, the recitation or recall 
 often worked under other restrictions. For example, Miss Abbott 
 in some tests restricted the learning to mere inner visualization of 
 the data. In other cases, Knors for example, the subject was 
 required to read and recite aloud. As a result of these various 
 controls, the methods of learning became highly artificial; seldom 
 was a subject permitted to study in the manner that he would 
 spontaneously adopt, and too often the restrictions were not the 
 same for recitation as for reading. 
 
 5. Testing the learning and computing the results. Sometimes the 
 lesson consisted of a certain number of repetitions, in which cases 
 the learning was never complete. The success attained might be 
 measured by the rate of the next recitation following immediately 
 or after an hour (Witasek), or simply by the amount of material 
 that could then be reproduced immediately or after an interval, or 
 by the time required to complete the learning then or later. Some- 
 times, a certain amount of time was given for study, and the amount 
 that could be reproduced immediately or after an interval (Abbott) 
 was taken as a measure of the learning. Sometimes, the assignment 
 was learned at a sitting (Kiihn), the score being based on the time 
 or repetitions required to learn. Other things, such as the number 
 of prompts required (Witasek), or the recitation time (Katzaroff), 
 have been introduced as a measure of success. Add to these 
 differences the highly ingenious yet anything but clean-cut methods 
 of computation, such as those introduced by Witasek and Knors, 
 and it is clear that to adequately compare the results of these 
 studies one with another, is next to impossible. 
 
 Certain other sources of error, such as neglect of practice effects, 
 fatigue, diurnal variations in efficiency add to the uncertainty. So 
 it is only with all these differences and sources of error in mind that 
 an attempt will be made to give a brief summary of the general 
 status of the problem. 
 
 First. A predominance of evidence points to a greater effective- 
 ness of recitation, compared to reading, as a factor in learning in 
 the case of adults, at least. 
 
 Second. This rule holds true only after the learning has advanced 
 somewhat by virtue of preliminary readings, but the exact point at 
 which it is best to introduce recitation into the learning, or the 
 optimum distribution of readings and recitations within the lesson, 
 has not been satisfactorily demonstrated. 
 
RECITATION AS A FACTOR IN MEMORIZING 23 
 
 Third. The more reliable experiments, such as those of Kiihn, 
 indicate that the advantage of recitation over reading is greater in 
 learning senseless, non-connected material than in learning senseful, 
 connected material. 
 
 Fourth. The matter of the relative value of recitations grouped, 
 as compared to recitations interspersed with the readings, is still 
 an open question. 
 
 Fifth. No satisfactory evidence is at hand indicating that the 
 general results found for adults will hold in the case of children of 
 grammar or high school age and training. 
 
 Sixth. A considerable, but not thoroughly convincing amount of 
 evidence indicates that the efficacy of the two methods of study 
 depends entirely upon the learning or imagery type of the individual. 
 
 Seventh. The two broad functions, learning by reading and learn- 
 ing by recitation, have not as yet been adequately analyzed into 
 their constituent functions. 
 
Ill 
 
 THE SUBJECTS, MATERIALS, AND METHODS 
 OF PROCEDURE 
 
 As was mentioned earlier, the present study was devised to 
 answer a practical question of the school-room — namely. What are 
 the relative values of learning by reading as compared to learning 
 by recitation in the case of school children working under school 
 conditions and with the ordinary school-room methods of attack? 
 So far as practicable, everything was done to secure normal condi- 
 tions for the work. The details concerning subjects, materials, 
 methods of study, and computation of results will now be con- 
 sidered. 
 
 Experiments were conducted with adult subjects as well, the 
 data from which will be used for comparative purposes and for 
 purposes of determining more exacth^ the functions operative in 
 the two methods of study. For the sake of convenience, the experi- 
 ments upon adults will be described in a later section where the 
 results are presented. 
 
 The Subjects 
 
 The subjects used were pupils of a grammar school of Oakland, 
 California. The members of the first, fourth, sixth, and eighth 
 grades acted as subjects for the experiments in which the nonsense 
 syllables were used, and the third, fourth, fifth, sixth, and eighth 
 grades for the tests with sense material. Each class consisted of 
 from forty to forty-five pupils. 
 
 The school in which the experiments were conducted is situated 
 in a residential suburb of Oakland and draws its pupils from the 
 homes of business men and artisans of moderate means. In general 
 the school stands in the first class. 
 
 As will be explained later in detail, the pupils were grouped by 
 grades rather than by age for the tests. The following table sum- 
 marizes the distribution of the members of the several grades 
 according to age. 
 
 Materials Used 
 
 The materials were of two sorts, senseless, non-connected material 
 and connected, sense material in the form of biographies. The 
 nonsense syllables were constructed in a manner similar to that of 
 
RECITATION AS A FACTOR IN MEMORIZING 
 
 25 
 
 Age in years 
 
 6 
 
 7 
 
 8 
 
 9 
 
 10 
 
 II 
 
 12 
 
 13 
 
 14 
 
 IS 
 
 16 
 
 17 
 
 Grade i 
 
 24 
 
 13 
 
 4 
 
 
 
 
 
 
 
 
 
 
 Grade 3 
 
 
 I 
 
 II 
 
 21 
 
 8 
 
 I 
 
 
 
 
 
 
 
 Grade 4 
 
 
 
 5 
 
 16 
 
 II 
 
 4 
 
 I 
 
 
 
 
 
 
 Grade 5 
 
 
 
 
 4 
 
 17 
 
 12 
 
 4 
 
 I 
 
 I 
 
 
 
 
 Grade 6 
 
 
 
 
 
 2 
 
 7 
 
 13 
 
 9 
 
 2 
 
 I 
 
 I 
 
 I 
 
 Grade 8 
 
 
 
 
 
 
 
 I 
 
 8 
 
 13 
 
 12 
 
 7 
 
 
 Miiller and Schumann.^ The sense material was constructed by 
 the writer from material found in J. McKeen Cattell's American 
 Men of Science and Who's Who in America. Samples are appended. 
 While this material is senseful and connected, the organization of 
 different parts of the whole is not so complete and systematic as 
 would be generally found in poetry or prose, in which the ideas are 
 more closely related and the material more closely unified by 
 rhythms, accents, and natural pauses. This biographical form of 
 material was used because it was desirable to approximate the kind 
 of material that the pupils were accustomed to study in their 
 regular history, geography, or grammar lessons. 
 
 The nonsense syllables were mimeographed in vertical columns on 
 cards and were handed out one to each student. The sense material was 
 mimeographed on sheets which were likewise distributed to the pupils. 
 
 Preliminary tests were conducted in order to determine the 
 amount and difficulty of the material to include in the lesson as 
 well as to give the subjects some preliminary practice in the tests 
 before the actual experimentation began. The kind and amount of 
 material was arranged so that the lesson was somewhat too large 
 for the best students to master in the time allotted. 
 
 In the case of nonsense syllables, the series contained for the 
 eighth grade sixteen syllables; for the sixth, fifteen; and for the 
 fourth, fourteen. The pupils of the first grade were unable to read 
 or write these syllables, so the teacher kindly constructed series of 
 twelve syllables of a kind they were accustomed to manipulating, 
 such as ad, en, ig, op, ot, etc. These syllables were written with a 
 black crayon by the teacher on large strips of heavy paper. 
 
 The sense material was also arranged to suit the capacities of 
 the different classes. For the eighth grade the biographies of five 
 men served as a lesson; for the sixth and fifth grades, the same 
 biographies for but four men were used. For the fourth grade 
 easier biographies of four boys were used ; while for the third grade, 
 the biographies of three boys sufficed. Samples of the material are 
 appended. 
 
 1 Described by Meumann in The Psychology of Learning, pp. 365-368. 
 
26 RECITATION AS A FACTOR IN MEMORIZING 
 
 The following is a sample of the material used by the fifth, sixth, 
 
 and eighth grades: 
 
 James Church, born in Michigan, February 15, 1869. Studied in Munich, and 
 later studied Forestry and Agriculture. Director of Mt. Rose Weather 
 Observatory in 1906. Studied evaporation of snow, water content, and frost. 
 
 John Clark, born in Indiana, June 4, 1867. Studied Surgery and became a 
 doctor in Philadelphia. Taught at Johns Hopkins. Has visited Italy and 
 Russia. Has a brother in Vancouver. 
 
 Morton Clover, born in Ohio, April 25, 1875. Studied Chemistry at Michigan. 
 Worked in Manila for eight years. Wrote articles on the content of dog- 
 wood, of sugar, and acids. Now lives in Detroit. 
 
 Clarence Cory, born in Indiana, September 4, 1872. Studied in Purdue and 
 Cornell Universities. Now lives in Berkeley. Is Professor of Engineering 
 and Dean of Mechanics. Since 1901 has been Consulting Engineer of San 
 Francisco. Is a member of the British Institute. 
 
 George Curtis, born in Massachusetts, July 10, 1872. Studied at Harvard on 
 Geography. Won Gold Medals at Paris in 1900. Member of Boston Scien- 
 tific Society. Went on the Dixie Expedition in 1902. 
 
 The following is a sample of the material used by the third and 
 fourth grades: 
 
 Harry, is 14 years old. His father is a farmer. Around the farm are red stones, 
 black-berry bushes, red clay, green clover, and small trees. Harry is in the 
 eighth grade, and is tall and slender. He likes dancing and singing. 
 
 James, was born in June, 1905. He is going to be a carpenter. He can make a 
 chair, a stool, a box, a gate, and a window. His mother has white hair and 
 wears a black dress. His father is fifty-five years old. 
 
 Harold, was born in New York. He came to California when six years old. He 
 is now fifteen years old and has a gun, a bicycle, a kite, a pair of skates, and 
 a baseball suit. He is going to be a lawyer and live in Seattle. 
 
 Fred, was born in March, 1898. He lives on 31st and Parker Streets. He 
 goes to business college. He is tall, has black hair and blue eyes, wears a 
 gray suit and brown necktie. His home is made of brick and granite. 
 
 Since, as will be seen later, the same sort of tests were repeated 
 several times, it was necessary to construct different texts equal in 
 number to the tests given. ^ An attempt was made, of course, to 
 make the various texts of equal difficulty, but as is usually the case, 
 they probably vary considerably. That such differences in diffi- 
 culty as may exist will not invalidate the results to any considerable 
 extent, will be made clear later. 
 
 Method of Conducting the Tests 
 
 Several very conspicuous sources of error are to be contended 
 with in experimental work of the present sort. That such errors 
 have found their way into the work of previous investigators on this 
 topic, has already been pointed out. The more important sources 
 of error are as follows: 
 
 2 In the case of nonsense material five tests and five texts were used; in the case of sense ma- 
 terial six tests and an equal number of texts. 
 
RECITATION AS A FACTOR IN MEMORIZING 27 
 
 (a) Practice effects. In a series of five or six practice periods of 
 from five to ten minutes each, it would be expected that practice 
 efifects would be considerable. Some of the earlier studies have not 
 taken this sufficiently into account. 
 
 (Jb) Unequal difficulty of texts. Since one individual must repeat 
 a similar test with many different texts, any inequality in their 
 difficulty will affect the results. Even series of nonsense syllables 
 may differ greatly in difficulty for different individuals. 
 
 (c) Individual differences. In the case of most of the earlier 
 investigations the subjects were so few that individual peculiarities 
 may have played a large role. 
 
 {d) Diurnal variations in efficiency and fatigue. It is imperative 
 that comparative experiments should be conducted at the same 
 hour of the day with subjects as nearly as possible in the same state 
 of physical fitness, unless some adequate estimate of these influences 
 be introduced as a check. In this respect nearly all of the earlier 
 investigators have been negligent. 
 
 In order to eliminate, as far as possible, the effects of such sources 
 of error, the method described below was employed in the work.^ 
 
 A class, consisting of forty or more pupils on the average, was 
 divided into a number of sections or squads,^ the number of squads, 
 for reasons which will be evident, being made equal to the number 
 of methods of study that were tested. Each squad thus consisted 
 of seven or eight pupils, the personnel remaining unchanged through- 
 out. Different texts, of as nearly equal difficulty as possible, were 
 of necessity used. A particular squad was tested but once on a 
 single day, and to complete the series for each squad required five 
 or six days. The accompanying table shows in detail the manner 
 in which the tests were conducted. The procedure was as follows: 
 At nine a. m. of the first day, squad one was given its first test 
 under method one,^ using text one. Immediately after, squad two 
 studied the same text, according to method two; then squad three 
 worked under method three and so on. On the next day, squad 
 three was taken out at the first hour and studied text two according 
 to method two ; at the next hour squad four worked under method 
 three with the same text and so on. Thus the squads progressed, 
 during the five days, through all the trials, texts, methods, and 
 hours. The outcome, as shown under the column indicated 'Total' 
 is that from the point of view of the methods employed, which is 
 the only factor with which we are concerned; all other influences 
 are balanced or neutralized. 
 
 ' The first grade was handled as a whole and not by squads as were the others. To be taken into 
 new surroundings under the charge of a stranger proved to be too disturbing for these little children. 
 * Five for the learning of nonsense material, six for the learning of sense material. 
 5 'Method' refers to the manner in which the material was studied, see p. 30. 
 
28 
 
 RECITATION AS A FACTOR IN MEMORIZING 
 
 Differences in practice effects are neutralized because the sum 
 total of practice for any one method is the same as for all others. 
 Individual differences are neutralized because each subject has 
 studied under each method, and no one more than once, ^he errors 
 arising from differences in the difficulty of the texts are avoided, 
 because each method has to its credit one group working with each 
 of the six texts. The influences of diurnal variations in efficiency or 
 
 
 Day I 
 
 Day i 
 
 
 Day 3 
 
 Day 4 
 
 Day 5 
 
 Totals 
 
 Method I 
 
 Squad i 
 
 Squad 
 
 2 
 
 Squad 3 
 
 Squad 4 
 
 Squad 5 
 
 All squads 
 
 
 Trial i 
 
 Trial 
 
 2 
 
 Trial 3 
 
 Trial 4 
 
 Trial 5 
 
 All trials 
 
 
 Hour A 
 
 Hour 
 
 E 
 
 Hour D 
 
 Hour C 
 
 Hour B 
 
 All hours 
 
 
 Text I 
 
 Text 
 
 2 
 
 Text 3 
 
 Text 4 
 
 Text 5 
 
 All texts 
 
 Method 2 
 
 Squad 2 
 
 Squad 
 
 3 
 
 Squad 4 
 
 Squad 5 
 
 Squad i 
 
 All squads 
 
 
 Trial i 
 
 Trial 
 
 2 
 
 Trial 3 
 
 Trial 4 
 
 Trial 5 
 
 All trials 
 
 
 Hour B 
 
 Hour 
 
 A 
 
 Hour E 
 
 Hour D 
 
 Hour C 
 
 All hours 
 
 
 Text I 
 
 Text 
 
 2 
 
 Text 3 
 
 Text 4 
 
 Text 5 
 
 All texts 
 
 Method 3 
 
 Squad 3 
 
 Squad 
 
 4 
 
 Squad 5 
 
 Squad i 
 
 Squad 2 
 
 All squads 
 
 
 Trial i 
 
 Trial 
 
 2 
 
 Trial 3 
 
 Trial 4 
 
 Trial 5 
 
 All trials 
 
 
 Hour C 
 
 Hour 
 
 B 
 
 Hour A 
 
 Hour E 
 
 Hour D 
 
 AH hours 
 
 
 Text I 
 
 Text 
 
 2 
 
 Text 3 
 
 Text 4 
 
 Text 5 
 
 All texts 
 
 Method 4 
 
 Squad 4 
 
 Squad 
 
 5 
 
 Squad i 
 
 Squad 2 
 
 Squad 3 
 
 All squads 
 
 
 Trial i 
 
 Trial 
 
 2 
 
 Trial 3 
 
 Trial 4 
 
 Trial 5 
 
 All trials 
 
 
 Hour D 
 
 Hour 
 
 c 
 
 Hour B 
 
 Hour A 
 
 Hour E 
 
 All hours 
 
 
 Text I 
 
 Text 
 
 2 
 
 Text 3 
 
 Text 4 
 
 Text 5 
 
 All texts 
 
 Method 5 
 
 Squad 5 
 
 Squad 
 
 I 
 
 Squad 2 
 
 Squad 3 
 
 Squad 4 
 
 All squads 
 
 
 Trial i 
 
 Trial 
 
 2 
 
 Trial 3 
 
 Trial 4 
 
 Trial 5 
 
 All trials 
 
 
 Hour E 
 
 Hour 
 
 D 
 
 Hour C 
 
 Hour B 
 
 Hour A 
 
 All hours 
 
 
 Text I 
 
 Text 
 
 2 
 
 Text 3 
 
 Text 4 
 
 Text 5 
 
 All texts 
 
 fatigue are neutralized, since each method has been tried by one 
 squad working at each of the different hours. 
 
 Almost ideal arrangements were made for conducting the tests.' 
 In a well lighted and well ventilated room about twenty-two by 
 fourteen feet in size, a library table large enough to seat about a 
 dozen people was provided. The situation of the room was such 
 that practically all noise and distractions of whatever kind were 
 avoided. Care was taken to keep the physical conditions of the 
 room as constant and comfortable as possible. Fresh air was kept 
 in circulation, an abundance of light was admitted, and the tem- 
 perature was kept constantly between fifty-seven and sixty degrees 
 Fahrenheit. 
 
 " In the case of sense material, six methods, squads, texts, etc., were used instead of five. 
 ' For this I am greatly indebted to the school principal, Mr. N. Ricciardi. 
 
RECITATION AS A FACTOR IN MEMORIZING 29 
 
 Since there were but seven or eight pupils undergoing a test at a 
 time, the experimenter who stood at the head of the table could 
 easily keep an eye on the work of each individual. Any attempt on 
 the part of a pupil to copy from another, to loaf, or use improper 
 methods of any sort, could be instantly detected. Such policing 
 was quite unnecessary and such violations of rules as did occur were 
 in most cases unintentional. However, such factors which might 
 result in the unreliability of the data were urgently sought, and in 
 cases where such an unreliability was known or suspected, the 
 entire data of that child were thrown out. In addition to the obser- 
 vations of the writer, the opinion of the teacher, especially with 
 reference to doubtful cases, was sought and freely obtained. Each 
 teacher listed the pupils in her room according to the following 
 request, "Please list your estimates of the intelligence of the pupils 
 in your room, in order of rank, putting the most intelligent as 
 
 Number one, . Use your own methods of estimating and your 
 
 own conception of what intelligence is. Please do not, however, 
 make it a mere record of class standing according to grades received, 
 and mere maturity should not be considered." The teachers also 
 fulfilled a request to give the names "of such pupils that you think 
 on account of feeble intelligence or inattentiveness', lack of persis- 
 tence, indolence or inclination toward dishonesty in work, etc., 
 would be unreliable subjects for experimental purposes." The teach- 
 ers were consulted also in particular cases when the occasion arose. 
 
 As a result of these precautionary measures the work of a few 
 pupils was discarded. The following were the chief factors which 
 seemed to justify discarding a pupil's data: 
 
 First. Absence from one of the tests. In case a pupil missed one or 
 more of the tests, his entire work was discarded. This was necessary 
 because in succeeding tests he would be one or more stages behind 
 in practice. The absentees on return were allowed to continue the 
 work without being told that the data would not be used, as a pre- 
 caution against creating any ill feeling among the pupils. 
 
 Second. Copying from others or using unfair methods of any sort. 
 Intentional or unintentional disregard of rules was very rare. 
 
 Third. Lack of interest or loafing. Occasionally a pupil from lack 
 of interest or less worthy motives, felt inclined to be balky or to 
 'quit' for a moment in the midst of a test. The data of such were 
 discarded. 
 
 Fourth. Mental defectives. A few pupils were found to be marked- 
 ly below the average in the test work. Consultation with the teacher 
 confirmed the suspicion of sub-normality and the data of such were 
 discarded, although they went through the work with the other 
 pupils. 
 
30 RECITATION AS A FACTOR IN MEMORIZING 
 
 Fifth. Physical defectives. Bad cases of eye defects, weakness 
 from previous illness, and school-yard accidents occasionally inter- 
 fered with maximal performance to such an extent that the data 
 were excluded. 
 
 On the whole, however, such cases were very rare and the spirit 
 among the pupils was of the finest. A keen spirit of competition 
 arose with reference both to an individual's own previous record 
 and to the records of other individuals, such that in nearly every 
 case the results were the products of the pupils' very best endeavors. 
 The number of pupils who completed all of the tests in a satisfactory 
 manner ranged from thirty-seven to forty-one in the various grades. 
 
 Methods of Studying 
 
 A single squad having been seated at the table in the separate 
 room, a copy of the material was passed out face downward before 
 each pupil, and the following instructions were given: "On each of 
 these cards is a list of nonsense words [show a sample]. They are 
 called nonsense words because in English they have no meaning. 
 Now the object of the test toda}^ is to see how many of these words 
 you can learn in a certain short time. 
 
 "We will proceed like this. I will give you two signals to start. 
 At 'Ready' you take the card at the corner like this and at 'Go' 
 you turn the card over and begin to study. 
 
 "Now you are going to study for a while in one way and then 
 later you are going to study in a very different way. To begin with 
 you are to study by reading this list of words over and over from 
 beginning to end [illustrate]. Remember you are to read only. 
 You should never look away from the paper; never close your eyes 
 to see if you can say the words; in fact never say a single word 
 unless you are actually looking at it, actually reading it. Remember 
 you are to read through from the first to the last every time. 
 
 "After you have read the words through and through in this way 
 for a while, I am going to give you a signal 'Recite'. When I say 
 'Recite' you are to hold your paper in front of you so that when 
 you are looking straight ahead, you look over the top of it and you 
 can see it by glancing downward a little like this. Now you are to 
 try to say to yourselves as many of the syllables as you can without 
 looking at the card. When you cannot remember the next word 
 look down at your card and then go on saying as many of them as 
 possible without looking. Glance at the card again whenever you 
 cannot remember. Go through the list from the first word to the 
 last in this way and continue until the word 'Time' is given. 
 Remember you are not to look at the words unless you absolutely 
 have to. 
 
RECITATION AS A FACTOR IN MEMORIZING 
 
 31 
 
 "When the learning period is over I am going to ask you to write 
 as many of these words as you can." 
 
 It should be remembered that every class had received previous 
 practice in the learning. The first grade had been given two trial 
 tests of five minutes each, and every other grade one or two trials 
 of eight minutes each, the data from which were not used. 
 
 Following is a table showing the absolute and relative amounts 
 of time devoted to reading and to recitation in each method. 
 
 NONSENSE MATERIAL 
 Grade one 
 
 Method 
 
 Time of reading 
 
 Time of recitation 
 
 Per cent, reading 
 
 Per cent, recitation 
 
 I 
 
 5' 
 
 0' 
 
 100 
 
 
 
 2 
 
 4' 
 
 l' 
 
 80 
 
 20 
 
 3 
 
 3' 
 
 2' 
 
 60 
 
 40 
 
 4 
 
 2' 
 
 3' 
 
 40 
 
 60 
 
 5 
 
 l' 
 
 4' 
 
 20 
 
 80 
 
 Grades four, six, and eight 
 
 I 
 
 9' 
 
 
 
 100 
 
 
 
 2 
 
 7'l2" 
 
 I '48" 
 
 80 
 
 20 
 
 3 
 
 5'24" 
 
 3'36" 
 
 60 
 
 40 
 
 4 
 
 3'36" 
 
 5'24" 
 
 40 
 
 60 
 
 5 
 
 I '48 " 
 
 7'I2" 
 
 20 
 
 80 
 
 SENSE MATERIAL 
 
 Grade three 
 
 Method 
 
 Time of reading 
 
 Time of recitation 
 
 Per cent, reading 
 
 Per cent, recitation 
 
 I 
 
 I'io" 
 
 
 
 100 
 
 
 
 2 
 
 6' 
 
 I '30" 
 
 80 
 
 20 
 
 3 
 
 4'30" 
 
 3' 
 
 60 
 
 40 
 
 4 
 
 3' 
 
 4'30" 
 
 40 
 
 60 
 
 5 
 
 I '30" 
 
 6' 
 
 20 
 
 80 
 
 6 
 
 45" 
 
 6 '45" 
 
 10 
 
 90 
 
 Grades four, five, six, and eight 
 
 I 
 
 9' 
 
 
 
 100 
 
 
 
 2 
 
 7'I2" 
 
 I '48 " 
 
 80 
 
 20 
 
 3 
 
 5'24" 
 
 3'36" 
 
 60 
 
 40 
 
 4 
 
 3'36" 
 
 5'24" 
 
 40 
 
 60 
 
 5 
 
 I '48" 
 
 7'I2" 
 
 20 
 
 80 
 
 6 
 
 54" 
 
 8 '06" 
 
 10 
 
 90 
 
32 RECITATION AS A FACTOR IN MEMORIZING 
 
 The study period was made somewhat shorter for the first and 
 third grades, because it was found that steady apphcation for longer 
 periods was quite fatiguing. 
 
 At the end of each study period the pupils promptly placed the 
 text papers face downv/ard and began at once to write the material 
 upon sheets that were provided. They were instructed to give the 
 material in the original order as far as possible. In the case of 
 nonsense syllables, the recall was pure reproduction, but when the 
 sense material was used, the names of the individuals whose biog- 
 raphies were studied were written on the board in proper order. 
 This was the only aid that was given. Ample time was allotted in 
 which to write the material remembered. 
 
 Three or four hours later, tests for retention were given. The 
 test consisted in simpb/ asking the pupils to write, as before, all 
 the material they could remember. No aids were given except 
 that the names, in the case of sense material, were written on the 
 board as in the immediate test. 
 
 Notes were kept of all manifestations of the children's work 
 such as movements of the lips, whisperings, rhythmical move- 
 ments of the head, or hands or feet, tappings of the fingers, 
 directions of the gaze, etc., in fact, of all appearances which 
 might be of later service in interpreting the results. The judg- 
 ments of the pupils were frequently called for upon such mat- 
 ters as the methods which they liked or disliked, why the non- 
 sense syllables were hard to learn and the like. These will be 
 dealt with later. 
 
 The method thus far described applies only to the work with the 
 school children. DifTerent methods were employed upon the adult 
 subjects and they can most conveniently be described on a later 
 page where the results are presented. 
 
 TREATMENT OF THE DATA 
 
 Method of scoring the nonsense syllables. The nonsense syl- 
 lables were scored by giving three points for a syllable correct in 
 form and position; two points for a syllable correct in position with 
 one letter incorrect; two points for a syllable correct in form but 
 not in correct position; one point for a syllable with two letters 
 correct but in wrong position. For example: 
 
 Correct list 
 
 Reproduced list 
 
 pib 
 
 pib 
 
 bah 
 
 dah 
 
 rem 
 
 bug 
 
 lor 
 
 rem 
 
 Cllg 
 
 lag 
 
RECITATION AS A FACTOR IN MEMORIZING 33 
 
 Although more exact methods of scoring nonsense syllables are 
 available, it was thought that the additional precision that might 
 be obtained by their use would scarcely justify the additional labor 
 involved. 
 
 All of the nonsense syllables were scored by a person who had no 
 acquaintance with the nature of the experiment. In order to test 
 the personal equation as manifest in the scoring, a set of forty lists 
 were graded by two individuals, neither being aware that the lists 
 were to be, or had been, corrected by another. The variations were 
 found to be very small and due to variable errors so that the averages 
 were about the same. The average score for forty papers was for 
 one grader 22.81, for the other 22.88. From these figures the P. E. 
 
 A. D. dis. 
 was computed by means of the formula P. E. = .84435 — 
 
 V n 
 
 The P. E. thus determined is 0.021. The personal factor involved 
 in the scoring of results is thus too small to be of significance. 
 
 The material of the first grade pupils (two letter syllables) was 
 graded by simply counting the number of syllables that were 
 correct in spelling. These children had had but little experience in 
 writing on paper and as a consequence their syllables were mixed 
 up so badly that it was impossible in many cases to be sure what 
 order was intended. Consequently, correctness in form, only, was 
 considered. 
 
 MetJwd of scoring the sense material. The sense material was 
 scored by dividing the original texts into details, ideas, or facts that 
 were mentioned, to serve as a guide. One credit was given for the 
 correct reproduction of each of these 'details' when they fell under 
 the proper name. When a detail, such as a birthplace, was correctly 
 reproduced but applied to the wrong person, one-half a unit was 
 given. In some cases the credits of one-half or three-fourths were 
 given to details or facts partly correct, depending upon the judg- 
 ment of the reader. 
 
 Part of the sense material was scored by one individual and part 
 by another, neither of whom was acquainted with the experiments 
 in general. To test the reliability of the judgments, forty papers 
 were scored independently by each. Variations of small magnitude 
 but greater than for the nonsense material were found, but these 
 were due to variable errors that compensated each other in the 
 long run, producing on an average of forty scores very slight differ- 
 ences. The P. E., computed as above, is 0.015. This P. E. is so 
 small in comparison with the P. E.'s of the averages that it has not 
 been taken into consideration in the final computations of the 
 results. 
 
34 RECITATION AS A FACTOR IN MEMORIZING 
 
 METHODS OF COMPUTING THE RESULTS 
 
 The results show the average scores based upon the methods of 
 grading just mentioned. To be more accurate, the tables show a 
 grand average of the averages of the several squads for each method 
 of study. The work of several pupils, for various reasons that have 
 been cited, was rejected, with the result that the final number of 
 individuals in some squads is greater than that in others. Since, 
 from the point of view of any particular method, the practice efifects 
 of each squad differed from every other, to permit the results of a 
 squad to enter the final average with full weight, would distort the 
 figures in a degree amounting to the average difference in efficiency 
 due to the greater amount of practice of the one over the other. 
 This overweighting was avoided by averaging each squad separately 
 and then making an average of these figures. 
 
 For the same reason the P. E.'s could not be computed in the 
 
 regular manner but must be based upon the results of the individual 
 
 squads. Assuming that the averages of the several squads would be 
 
 equal except for differences due to practice, fatigue, and diurnal 
 
 variations, the deviations of the figures within each squad from the 
 
 average of that squad were computed. A sum of the deviations for 
 
 all individuals from the average of their squad was thus obtained 
 
 and divided by the total number of individuals in the class, thus 
 
 giving the Average Deviation. The P. E.'s were then computed 
 
 according to the formula: ^ 
 
 A. D. dis. 
 
 P. E. tr. av. - obt. av. = 0.8453 — 
 
 Vn 
 
 s See Thorndike, E. L., Mental and Social Measurements, New York, 191 2, pp. 186 ff. 
 
IV 
 QUANTITATIVE RESULTS 
 
 It was pointed out earlier that the amount of material given as a 
 lesson was slightly greater than the best students could learn in the 
 time allotted. Learning was never complete, although in the case 
 of many individuals it was nearly so. With nonsense syllables as 
 material, the average scores for the best methods are for different 
 classes from fifty to seventy-three per cent, of the highest possible 
 score. For the sense material, the best average scores are in the 
 neighborhood of forty per cent, of the highest possible scores. This 
 fact should be kept in mind during the consideration of the results 
 which follow. For convenience of expression, we shall speak of 
 'methods' in which there was a 'combination of twenty per cent, 
 reading with eighty per cent, recitation', etc., but it must be 
 remembered that such expressions have a strictly local meaning, for 
 several reasons. In the first place, such 'combinations' lead only 
 to partial learning of the data. Perhaps the same combination 
 would lead to very different results if applied to the time required 
 lo completely learn the lesson. A second consideration is that a 
 ''combination' has reference only to the particular kind and the 
 particular amount of material here used. The optimum combina- 
 tion would doubtless be different according to the difficulty and 
 length of the lesson. These matters will be given more considera- 
 tion on a later page. 
 
 RESULTS FOR THE LEARNING OF NONSENSE SYLLABLES 
 BY CHILDREN 
 
 Table XVII shows the results of the immediate test for nonsense 
 syllables in the form of average with P. E.'s computed in the 
 manner described in the previous chapter. Table XVIII shows 
 the same data transformed to relative scores in which the average 
 of each class for all five methods is considered loo, serving as a 
 basis for the other scores. The P. E.'s were changed to correspond. 
 Figure i shows graphically the data of Table XVIII, the average 
 being denoted by the heavy line, which is enclosed within two light 
 lines representing on either side the area including the P. E. ^ 
 
 For the fourth, sixth, and eighth grades the results are clear. The 
 results for Grade one were a disappointment and should be considered 
 
 • After the manner originally suggested by Professor J. McKeen Cattell. 
 
36 
 
 RECITATION AS A FACTOR IN MEMORIZING 
 
 apart from the others. The averages for this grade seem to indicate 
 that Method Five, in which the amount of recitation is greatest, pro- 
 duces the poorest results while the methods involving more reading 
 show to better advantage. The P. E.'s, however, show the averages 
 not to be highly reliable and their significance is slight. One reason 
 for this may lie in the fact that a less refined method was used in 
 the case of this grade (see p. 2^]). In all probability, moreover, the 
 inexperience of these beginning pupils accounts for the results in a 
 large measure. They were simply unable to adjust themselves to 
 
 TABLE XVII 
 
 Showing the average score for each grade for the various methods of study 
 
 Method 
 
 I 
 
 2 
 
 3 
 
 4 
 
 s 
 
 Combination in mins. and sees. 
 
 / Lp' Ro 
 
 Lv'iz' 
 Ri'48" 
 
 Ls'24' 
 R3'36' 
 
 L3'36' 
 Rs'24' 
 
 Li'48' 
 
 R7'i2' 
 
 Combination in per cent. 
 
 Lioo Ro 
 
 L80 R20 
 
 L60 R40 
 
 L40 R60 
 
 L20 R80 
 
 Grade eight Average score 
 
 16.92 
 
 23.86 
 
 25-79 
 
 27.28 
 
 35-51 
 
 P. E. 
 
 0.61 
 
 0,69 
 
 0.65 
 
 0.66 
 
 0.S6 
 
 Grade six Average score 
 
 13.21 
 
 20.18 
 
 22.64 
 
 2515 
 
 30.52 
 
 P. E. 
 
 0.61 
 
 0.84 
 
 0.60 
 
 0.91 
 
 1.07 
 
 Grade four Average score 
 
 9-45 
 
 12.00 
 
 16. 10 
 
 16.95 
 
 20.03 
 
 P. E. 
 
 0.57 
 
 0.46 
 
 0.56 
 
 0.75 
 
 0.79 
 
 Combination in mins. and sees. 
 
 Ls'Ro 
 
 L4'Ri' 
 
 L3'R2' 
 
 L2'R3' 
 
 Li'R4' 
 
 Combination in per cent. 
 
 Lioo Ro 
 
 L80 R20 
 
 L60 R40 
 
 L40 R60 
 
 L20 R80 
 
 Grade one Average score 
 
 6.2 
 
 6.1 
 
 6.2 
 
 5-6 
 
 4-7 
 
 P. E. 
 
 0.22 
 
 0.27 
 
 0.20 
 
 0.20 
 
 0.21 
 
 TABLE XVIII 
 
 Shoiving the data of Table X VII on a relative basis 
 
 Metliod 
 
 I 
 
 2 
 
 3 
 
 4 
 
 s 
 
 Grade eight 
 
 Relative score 
 
 65.40 
 
 92-23 
 
 99.69 
 
 105-45 
 
 137.26 
 
 
 P. E. 
 
 2.37 
 
 2 
 
 69 
 
 2.53 
 
 2 
 
 57 
 
 3 35 
 
 Grade six 
 
 Relative score 
 
 59-13 
 
 88 
 
 35 
 
 101.34 
 
 112 
 
 57 
 
 136.61 
 
 
 P. E. 
 
 2-74 
 
 3 
 
 78 
 
 2.70 
 
 4 
 
 09 
 
 4.81 
 
 Grade four 
 
 Relative score 
 
 63.42 
 
 80 
 
 53 
 
 108.05 
 
 113 
 
 75 
 
 134-42 
 
 
 P. E. 
 
 3-42 
 
 2 
 
 76 
 
 3-36 
 
 4 
 
 50 
 
 4-74 
 
 Grade one 
 
 Relative score 
 
 107.64 
 
 105 
 
 90 
 
 107.64 
 
 97 
 
 22 
 
 81.59 
 
 
 P. E. 
 
 3.80 
 
 4 
 
 67 
 
 3-46 
 
 3 
 
 46 
 
 3-63 
 
Method I 
 
 RECITATION AS A FACTOR IN MEMORIZING 
 FIGURE I 
 
 Based on the data of Tables XVIII and XIX 
 234 
 
 37 
 
 100% 
 
 100% 
 
 100% 
 
 100% 
 
 100% 
 
38 RECITATION AS A FACTOR IN MEMORIZING 
 
 the experiment. Many employed practically the same methods of 
 study throughout, as could be seen from observation in some cases. 
 Others made an effort to follow the prescribed directions which 
 often resulted in poor records, especially in the methods in 
 which recitation began very early. A great deal of time was lost 
 in fruitless endeavor to recall syllables that were not as yet forth- 
 coming. These young children were not skilled enough in testing 
 their knowledge and prompting themselves where needed, which 
 probably accounts for the apparent inferiority of the results obtained 
 in Methods Four and Five, in which the reading periods were short. 
 
 For the fifth, sixth and eighth grades the results are convincing. 
 The averages show a very great superiority of Method Five, in which 
 the most recitation is introduced, over Method One which employs 
 reading only. Twice as much is learned by the former as by the 
 latter method in the same time. The small P. E.'s indicate a high 
 reliability of these extreme differences, as well as the fact that the 
 same thing is shown by all three classes. The graphs (Figure i) 
 show that there is an increase in the amount learned as the relative 
 amount of recitation becomes larger, a . fact shown by all three 
 classes. The amount of this increase is not constant, however, 
 being marked by a particularly great difference between Method One 
 which permits no recitation and Method Two which gave twenty per 
 cent, of the time to recitation. The effect of a minute and a half 
 of recitation is very marked. The increase in effectiveness is fairly 
 constant from Method Two to Three to Four, but the step from Four 
 to Five is somewhat greater than any one of these. The most 
 probable explanation for this exceptional score in the case of 
 Method Five is that it was usually productive of a little more 
 enthusiasm than other methods. The children anticipated this as 
 the 'record breaking' method. 
 
 The reliability of. the differences between the methods has been 
 
 computed in a different way, as shown in Table XIX and displayed 
 
 graphically by the broken line curve in Figure I. This table shows 
 
 the averages of the three grades (eighth, sixth, fifth) for each 
 
 method with the P. E.'s of the averages computed according to 
 
 the formula: ^ 
 
 a dis. 
 P. E. tr. av. - obt. av. = .6745 — 1=^ 
 
 V n 
 
 The P. E.'s should be magnified to some extent in this table for 
 the reason that the number of cases is very small (three) and that 
 influence of any factor tending to create differences between the 
 groups considered, for example, the effects of maturity or length of 
 
 2 See ThOrndike, E. L., Mental and Social Measurements, p. i88 ff. 
 
RECITATION AS A FACTOR IN MEMORIZING 
 
 39 
 
 school training, would make the P. E.'s larger. However, the P. 
 E.'s are still very small. 
 
 TABLE XIX 
 
 Showing the average percentile scores with P. E.'s for Grades four, five, and eight 
 
 Method 
 
 I 
 
 2 
 
 3 
 
 4 
 
 5 
 
 Average score 
 P. E. 
 
 62.65 
 1 .01 
 
 87.04 
 1.88 
 
 103.59 
 1 .40 
 
 110.59 
 I 45 
 
 136.09 
 0.45 
 
 The following table shows more plainly the differences between 
 the various methods and the P. E.'s of those differences, the com- 
 putation being based on the preceding table. The formula em- 
 ployed for obtaining the P. E. of the differences is:^ 
 
 P. E. diff. = V(P. E. av.)2 + (P. E. av.)^ 
 
 TABLE XX 
 
 Showing the differences of the various methods in percentages with the 
 P. E.'s of the differences 
 
 Differences of methods 
 
 2 — I = 24.39 ± P. E. 2. 1 1 
 
 3— I = 40.38=*= P. E.I. 72 
 4 — I = 47.94=1= P. E. 1.76 
 5— I = 73.44=1= P. E.I. 10 
 
 Differences of methods 
 
 3— 2= 1 5.99=*= P. E. 2.34 
 4— 2= 23.55 =fc P. E. 2.37 
 5— 2 = 49.05 ± p. E.I. 93 
 
 Differences of methods 
 
 4—3= 7.56=tP. E.2.0I 
 5— 3= 33-06 ± p. E.I. 47 
 5— 4= 25.50=*= P. E.I. 52 
 
 The differences are all conspicuous and reliable. 
 
 Differences in results among classes 
 
 A glance at Figure i will show that the findings for Grades 
 four, six, and eight are very similar. In all grades Method Five is 
 about twice as good as Method One. About the only difference is 
 that the fourth grade does not do well, relatively, with very short 
 periods of recitation. The difference in percentages* of Method 
 Two (i' 48" Recitation) over Method One (all reading) is for the 
 eighth grade 26.83 per cent., for the sixth grade 29.22 per cent., for 
 the fourth grade 17. 11 per cent. Grade four shows the slightest 
 superiority of Method Two over Method One, but its value is ren- 
 dered somewhat doubtful since the sixth grade shows a slightly 
 greater superiority than does Grade eight. Computing the superiority 
 of the average score from all Methods over the score of Method 
 
 3 See Thorndike E. L.. op. cit. 
 
 < Based on Table XVIII. 
 
40 
 
 RECITATION AS A FACTOR IN MEMORIZING 
 
 Two, the results are: Grade eight, a superiority of 7.73 per cent.; 
 Grade six, 11. 12 per cent., and Grade four, 19.47 P^r cent. From 
 these figures it appears that the older children are able to do better, 
 relatively, with the short recitation periods. 
 
 Summary 
 
 To summarize, then, it may be stated: {First) that for the learn- 
 ing of nonsense material by children, recitation after a few initial 
 readings is of much greater value than continued readings. (Secojid) 
 That after preliminary readings for i' 48", the more quickly the 
 attempts at recitation are introduced, the better results will be 
 obtained. (Third) No conspicuous differences appear in the results 
 for the different classes with the exception of Grade one. (Fourth) 
 which for reasons mentioned on p. 27 must be treated as a distinct 
 case. 
 
 RESULTS FOR SENSE MATERIAL 
 
 Table XXI shows the results of the immediate test for sense 
 material in the form of average scores. Table XXII shows the same 
 data on a relative basis in which the average score for all methods 
 
 TABLE XXI 
 
 Showing the average score for each grade for the various methods of study ^ 
 
 Method 
 
 I 
 
 2 
 
 3 
 
 4 
 
 5 
 
 6 
 
 Combination in mins. and sees. 
 Combination in per cent. 
 
 ( Lq Ro 
 Lioo Ro 
 
 Lt'iz" 
 
 Ri'48' 
 
 LSo R20 
 
 Ls'24" 
 
 R3'36" 
 
 L60 R40 
 
 L3'36'' 
 
 R5'24'' 
 
 L40 R60 
 
 Li'48' 
 
 R7'i2" 
 
 L20 R80 
 
 L54'' 
 
 R8'o6' 
 
 Lio R90 
 
 Grade eight Average score 
 P. E. 
 
 Grade six Average score 
 P. E. 
 
 Grade five Average score 
 P. E. 
 
 Grade four Average score 
 P. E. 
 
 20. -n 
 0.72 
 
 15-13 
 0.75 
 
 11.79 
 0.40 
 
 14.61 
 0.77 
 
 22 . 39 
 0.87 
 
 16.55 
 0.59 
 
 13 95 
 0.43 
 
 16.91 
 0.78 
 
 24.84 
 0.70 
 
 18.01 
 0.69 
 
 15 21 
 0.48 
 
 16.36 
 0.86 
 
 24-95 
 0.69 
 
 17.70 
 0.68 
 
 15-96 
 0.56 
 
 18.81 
 0.77 
 
 25-28 
 0.50 
 
 17-77 
 0.82 
 
 15-33 
 0.50 
 
 17.62 
 0.70 
 
 23-75 
 0.82 
 
 16.63 
 0.68 
 
 15-74 
 0.55 
 
 17.20 
 0.71 
 
 Combination in mins. and sees. 
 Combination in per cent. 
 
 f L7'30' 
 \ Ro 
 Lioo Ro 
 
 L6' 
 Ri'30" 
 LSo R20 
 
 L4'30' 
 
 IR3' 
 L60 R40 
 
 L3' 
 R4'30° 
 L40 R60 
 
 Li'3o' 
 
 R6' 
 L20 R80 
 
 L4S" 
 R6'4S' 
 Lio R90 
 
 Grade three Average score 
 P. E. 
 
 8.66 
 0.39 
 
 10.34 
 0.49 
 
 11.18 
 0.49 
 
 14.12 
 0.46 
 
 13.10 
 0.56 
 
 12.09 
 
 0.54 
 
 5 The highest possible score, approximately, is for Grade eight, 60; for Grades six, five and four. 
 48; and for Grade three, 36. 
 
RECITATION AS A FACTOR IN MEMORIZING 
 
 41 
 
 for that grade equals 100. The P. E.'s are computed as described 
 above (p. 34). Figure 2 shows graphically the data of Tables 
 XXII and XXIII. 
 
 TABLE XXII 
 
 Showing the data of Table XXI on a relative basis 
 
 Method 
 
 I 
 
 2 
 
 3 
 
 4 
 
 s 
 
 6 
 
 Grade eight 
 
 Relative score 
 
 87.78 
 
 94.62 
 
 104.98 
 
 105-45 
 
 106.80 
 
 100.03 
 
 
 P. E. 
 
 3 
 
 01 
 
 3.64 
 
 2-93 
 
 2 
 
 89 
 
 2.09 
 
 3-43 
 
 Grade six 
 
 Relative score 
 
 89 
 
 21 
 
 97.58 
 
 106.19 
 
 104 
 
 36 
 
 104.77 
 
 98.06 
 
 
 P. E. 
 
 4 
 
 42 
 
 3.48 
 
 4.09 
 
 4 
 
 01 
 
 4-83 
 
 4.01 
 
 Grade five 
 
 Relative score 
 
 80 
 
 42 
 
 95-15 
 
 103-75 
 
 108 
 
 86 
 
 104.57 
 
 107.36 
 
 
 P. E. 
 
 2 
 
 72 
 
 2.93 
 
 3-27 
 
 3 
 
 81 
 
 3-41 
 
 3-75 
 
 Grade four 
 
 Relative score 
 
 86 
 
 34 
 
 99.94 
 
 96.69 
 
 III 
 
 17 
 
 104.13 
 
 101.65 
 
 
 P. E. 
 
 4 
 
 54 
 
 4.60 
 
 507 
 
 4 
 
 54 
 
 4-13 
 
 4.18 
 
 Grade three 
 
 Relative score 
 
 74 
 
 78 
 
 89.29 
 
 96-54 
 
 121 
 
 93 
 
 113. 12 
 
 104.40 
 
 
 P. E. 
 
 3 
 
 35 
 
 4.21 
 
 4.21 
 
 3 
 
 95 
 
 4.81 
 
 4.64 
 
 A glance will show that the results here obtained differ from those 
 received with nonsense material. In general the advantage of 
 reading with recitation as compared to reading alone is less great. 
 Moreover it appears that introducing the recitation too early proves 
 to be of no value; in fact, for the lower grades it may prove to be a 
 positive hindrance. This point will be taken up later. All grades 
 agree in showing reading alone to be a poor method of study, while 
 a combination of forty per cent, reading with sixty per cent, recita- 
 tion seems to give best results. 
 
 The following table (XXIII) shows the average results for all 
 classes combined, with the P. E.; the methods of computation 
 being the same as those previously described. 
 
 TABLE XXIII 
 
 Showing the average percentile {relative) score for all grades combined 
 
 Method 
 
 Relative score 
 P. E. 
 
 I 
 
 2 
 
 3 
 
 4 
 
 S 
 
 83-71 
 1.64 
 
 95-32 
 0.99 
 
 101.63 
 1.26 
 
 110.35 
 I .90 
 
 106.67 
 I. 01 
 
 102.30 
 
 0.74 
 
 In the average results, Method Four seems to be distinctly superior 
 to Method Three and possibly superior to Methods Five and Six. In 
 
42 
 
 RECITATION AS A FACTOR IN MEMORIZING 
 
 order to more accurately determine the reliability of the differences 
 between the methods, Table XXIV was computed after the method 
 earlier described (p. 38) using the data from Table XXIII. 
 
 TABLE XXIV 
 
 Showing the differences between the relative scores for the several 
 methods with the P. E. of the differences 
 
 Differences of methods 
 
 Differences of methods 
 
 Differences of methods 
 
 2 — I =12.61 ±P. E. 1. 91 
 3 — I =17.92 ± P. E. 2.06 
 4 — 1 =26.64 ± P. E- 2.50 
 5 — I =22.96 ± P. E. 1.92 
 6— i = i8.59±P. E. 1.79 
 
 3- 
 4- 
 
 5- 
 6- 
 
 -2= 6.31 ±P. E. 1.60 
 -2 = 15.03 ±P. E. 2.14 
 -2 = 11.35 =tP. E. 1. 41 
 -2= 6.98 ± P. E. 1.23 
 
 4 — 3 = 8.72=1= P. E. 2.27 
 5— 3 =5.04 ± P. E. 1. 61 
 6—3=0.67=1= P. E. 1.46 
 5—4 = 3.68* P. E. 2.15 
 6— 4 = 8.05 =t P. E. 2.03 
 6— 5 =4.37 =t P. E. 1.25 
 
 This table shows, that for the average results, every method is 
 clearly superior to Method One (all reading), the smallest difference, 
 that between Methods Two and One, being more than six times the 
 P. E.^ It is also certain that every method except Method One is 
 superior to Method Two, the smallest superiority being four times 
 the P. E. The difference between Method Three and Four is also 
 quite reliable, being four times the P. E. The superiority of Method 
 Five over Method Three is more than three times the P. E. ; while 
 there is no real difference between Methods Six and Three. There is 
 no evidence that Method Four is superior to Method Five, but Four 
 is superior to Six, and Five is also superior to Six by small but re- 
 liable differences. 
 
 We are safe in concluding then, that in general, Method One, 
 which includes no recitation, is the poorest method, while Method 
 Four or Five is the best. Method Two is considerably superior to 
 Method One and Method Three is better than Two. That is to say, 
 the best results are obtained when the recitation is introduced after 
 one and one-half to three and one-half minutes of preliminary 
 reading. Beginning earlier or later than this leads to poorer 
 results. 
 
 Differences in results among the classes 
 
 Such are the findings in general, but it was noted earlier that the 
 classes differ in certain respects. These differences appear quite 
 clearly in Figure 2. In the first place, the difference between the 
 
 • It should be repeated that the P. E. should be very large for the reason that all the apparent 
 differences in the curves for the various classes (see Figure 2) result in attenuation of the P. E. 
 
Method I 
 
 RECITATION AS A FACTOR IN MEMORIZING 
 FIGURE 2 
 
 Based on the data of Tables XXII and XXIII 
 ^345 
 
 43 
 
44 RECITATION AS A FACTOR IN MEMORIZING 
 
 best and the poorest method is greater for the lower grades. The 
 superiority of the best over the poorest method is: 
 
 For grade eight 19.02 per cent. Average for 8+ 6 = 18.00 
 For grade six 16.98 per cent. 
 
 For grade five 28.44 per cent. 
 
 For grade four 24.83 per cent. Average for 3 + 4 = 35.99 
 For grade three 47.15 per cent. 
 
 The differences do not increase uniformly with the grades, but if we 
 average the differences for the eighth and sixth grades, also for the 
 third and fourth, the latter figure is exactly twice the former, while 
 the fifth grade lies midway. The older children are doubtless not 
 so badly handicapped by the lack of an opportunity to recite; or, 
 stated in another way, the younger children are more dependent 
 upon the factors involved in recitation in their learning. 
 
 Another difference also appears. Optimum results may be obtained 
 by introducing recitation earlier in the period in the case of the 
 upper grades. For Grades five, six, and eight the differences between 
 Methods Three, Four, Five, and Six are nil or unreliably small, but in 
 the case of Grade four the difference between Methods Six and Four 
 is 9.52 per cent. =*= P. E. 6.08; the same difference for grade three 
 being 17.53 per cent. ± P. E. 6.0. The introduction of the recita- 
 tion period too early has a deleterious effect upon the learning of 
 the two lower grades. The probable explanation of this difference 
 between the grades is to be found in the better adaptation of the 
 older or more experienced learner to the conditions of the test. 
 When recitation is introduced too early, the younger pupils waste 
 time and energy in fruitless endeavor to recall the material. Posi- 
 tive errors of recall are probably numerous also and thus retard the 
 learning. The older pupils, on the other hand, realizing that so 
 early an attempt at recitation would be unprofitable, continue for 
 some time to read, or divide the repetitions between reading and 
 recitation, reciting those few sections which can be recalled, but 
 referring promptly to their paper when the material is not forth- 
 coming. Another explanation is possible, but less probable, e. g., 
 that the results are due to a real difference among the classes in 
 ability to make rapid headway in the first few minutes of study. 
 That this is not highly probable is indicated by the fact that under 
 optimum conditions all classes learned approximately the same 
 proportion of their respective lessons in the given time. 
 
 A final difference is that the upper grades, in comparison with 
 the lower, do better when recitation is not introduced until fairly 
 late, i. e., when the proportion of reading is greater. Table XXII 
 discloses the fact that for Grades six and eight Method Three (sixty 
 
RECITATION AS A FACTOR IN MEMORIZING 45 
 
 per cent, reading) is as good as any other, but for Grades three and 
 four, Method Three is considerably inferior to Methods Four, Five, 
 or Six. For Grade three the superiority of Method Four over Method 
 Three is 25.39 per cent. =t P. E. 5.65; for rade Gfour, 14.48 per cent. 
 ± P. E. 6.70. Grade five Hes between the extremes, showing a 
 small (5. II per cent.) but scarcely reliable superiority of Method 
 Four over Method Three. This difference cannot easily be accounted 
 for, precisely, with the evidence at hand. It is probably due to the 
 greater experience of the more advanced students in learning 
 material — history, geography, and other lessons — in which reading 
 plays a very important part. By virtue of this experience, the 
 older children were more skillful in employing the most fruitful 
 methods of attack in reading which virtually amounted to less pure 
 reading, i. e., reading which was in some degree recitation. The 
 younger children stuck more strictly to pure reading. These 
 matters, however, must be waived to a later consideration. It is 
 only necessary here to suggest that such class differences, whatever 
 the explanation for them may be, are of marked pedagogical 
 importance. 
 
 Summary of Results f or Sense Material 
 
 1. In general, best results are obtained by introducing re citation 
 after devotin g about forty per cent, of the time to reading. Intro- 
 .ducing recita tion too early or too late lead s to poorer results . 
 
 2. In general, the optimum combination of reading and recita- 
 tion, under the conditions of the present tests, shows a superiority 
 over reading alone by about thirty per cent. 
 
 3. The lower grades differ from the upper grades in three respects. 
 
 a. The advantage of the best combination of reading and 
 recitation over the method of learning by reading alone is twice 
 as great for the lower grades, the average for grades three and 
 four being 35.99 per cent, as compared to 18.00 per cent, the 
 average for grades six and eight. 
 
 b. Introducing recitation earlier than the stage indicated in 
 (i) above, had a disadvantageous effect upon the learning of the 
 lower grades, but little or no ill effect upon the work of the upper 
 grades. 
 
 c. The upper grades, in comparison with the lower, learn more 
 effectively under the methods involving a relatively large amount 
 of reading. 
 
 RESULTS AS REGARDS RETENTION OF NONSENSE MATERIAL 
 
 Tests for retention of nonsense syllables were given from three to 
 four hours after the learning period, the exact intervals varying for 
 
46 
 
 RECITATION AS A FACTOR IN MEMORIZING 
 
 different classes but being always the same for all squads of any 
 one class. The pupils were simply asked to write down in proper 
 order all the syllables they could remember. It is impossible to 
 determine the unreliability of the retention results due to review, 
 intentional or otherwise, on the part of the pupils during the inter- 
 val between the learning period and the tests. That a few pupils 
 did review the material during the interim was obvious from the 
 fact that they obtained a higher score in the retention tests than in 
 the immediate test. Such results were, of course, discarded. It 
 was impossible to detect other cases in which the reviewing was 
 less extensive. With the exception of a few suspicious cases, the 
 results showed little or no indication of such procedure. An effort 
 in the way of appeal from teachers and the experimenter was made 
 to discourage such practices, and, on the whole, there are good 
 reasons for believing the results, aside from the exceptions men- 
 tioned, are quite reliable enough for broad interpretation. It would 
 be unwise, however, to give the data much weight for the inter- 
 pretation of fine differences, such as the differences between closely 
 related classes. 
 
 Table XXV shows the results in the form of averages with P. E.'s 
 computed in the manner previously described. Table XXVI like- 
 wise shows the results on a relative basis. 
 
 TABLE XXV 
 
 Showing the average scores obtained in the retention tests 
 
 Method 
 
 I 
 
 2 
 
 3 
 
 4 
 
 5 
 
 Grade eight 
 
 Average score 
 
 7.02 
 
 12.55 
 
 13.66 
 
 17-55 
 
 22.89 
 
 Interval four hours 
 
 P. E. 
 
 0.42 
 
 0.78 
 
 0.57 
 
 0.67 
 
 0.88 
 
 Grade six 
 
 Average score 
 
 523 
 
 7.12 
 
 9.91 
 
 12.58 
 
 20.38 
 
 Interval three hours 
 
 P. E. 
 
 0.45 
 
 0.54 
 
 0.72 
 
 0.82 
 
 1. 19 
 
 Grade four 
 
 Average score 
 
 3-49 
 
 5.89 
 
 8.35 
 
 10.58 
 
 14-25 
 
 Interval three hours 
 
 P. E. 
 
 0.38 
 
 0.56 
 
 0.53 
 
 0.47 
 
 0.53 
 
 Grade one 
 
 Average score 
 
 4-03 
 
 317 
 
 3-57 
 
 3-37 
 
 3-II 
 
 Interval three hours 
 
 P. E. 
 
 0.18 
 
 0.24 
 
 0.21 
 
 0.18 
 
 0.27 
 
 It is at once apparent that in a general way the results of the 
 retention tests are very similar to those of the immediate test. 
 Grade one stands by itself again for reasons that have been men- 
 tioned. Table XXVII shows the results of Grades four, six, and 
 eight combined. 
 
RECITATION AS A FACTOR IN MEMORIZING 
 TABLE XXVI 
 
 Showing the data of Table XX V on a relative basis 
 
 47 
 
 Method 
 
 I 
 
 2 
 
 3 
 
 4 
 
 5 
 
 Grade eight 
 
 Relative score 
 
 47-65 
 
 85.20 
 
 92-73 
 
 119. 14 
 
 155-46 
 
 
 P. E. 
 
 2.85 
 
 5-30 
 
 3-87 
 
 4 
 
 55 
 
 5-98 
 
 Grade six 
 
 Relative score 
 
 47-37 
 
 64.49 
 
 89.76 
 
 113 
 
 95 
 
 184.60 
 
 
 P. E. 
 
 4.07 
 
 4.88 
 
 6.51 
 
 7 
 
 42 
 
 10.76 
 
 Grade four 
 
 Relative score 
 
 41.01 
 
 69.21 
 
 98.12 
 
 124 
 
 32 
 
 167.45 
 
 
 P. E. 
 
 4-44 
 
 6.55 
 
 6.20 
 
 5 
 
 49 
 
 6.20 
 
 Grade one 
 
 Relative score 
 
 116. 81 
 
 91.88 
 
 103.48 
 
 97 
 
 68 
 
 90.14 
 
 
 P. E. 
 
 5.22 
 
 6.96 
 
 6.09 
 
 5 
 
 22 
 
 7.83 
 
 TABLE XXVII 
 
 Method 
 
 I 
 
 2 
 
 3 
 
 4 
 
 5 
 
 Relative score — average for Grades 
 four, six, and eight 
 P. E. 
 
 45-34 
 1. 19 
 
 72.96 
 3-32 
 
 93-53 
 1-34 
 
 119-13 
 1.92 
 
 169.17 
 
 4-47 
 
 Table XXVIII following shows the differences between the 
 various methods computed from Table XXVII. 
 
 TABLE XXVIII 
 
 Showing the differences between the various methods with P. E.'s of the differences 
 
 Differences of methods 
 
 Differences of methods 
 
 Differences of methods 
 
 2 — I =27.62 ±P. E. 3.51 
 3— i=48.i9=tP. E. 1.78 
 4— I =74.79 ± P. E. 2.25 
 5— I =80.75 =t P. E. 4.58 
 
 3— 2 =20.57 =t P. E. 3.60 
 4— 2=46.i7±P. E.3.87 
 5— 2=96.2i±P. E. 5.56 
 
 4— 3 =25.60 =t P. E. 2.34 
 5— 3 =75-64 =t P. E. 4.62 
 5— 4 = 50.04 ± P. E. 4.84 
 
 The steps from Method One to Method Five are all large and 
 reliable. Nearly four times as much is recalled when the learning 
 was predominantly recitation (Method Five) as when it was entirely 
 reading (Method One). As the amount of recitation increases the 
 amount recalled becomes greater. This increase in the amount 
 recalled is fairly uniform with the exception of the comparatively 
 
48 
 
 RECITATION AS A FACTOR IN MEMORIZING 
 
 great difference between Methods Four and Five. An explanation 
 for this was suggested earHer. 
 
 It will be recalled that in the immediate tests no differences were 
 found between the performances of the grades (except Grade one) 
 so far as the effects of the different methods of learning were con- 
 cerned. In the recall tests, there seems to be a slight difference 
 between Grades eight and five with respect to the superiority of 
 Method Five over Method One. From Table XXVI the differences 
 between Methods Five and One have been computed with results as 
 follows : 
 
 Differences for grade eight is 107.81 =*= P. E. 5.83 
 Differences for grade six is 137.23 =*= P. E. 11.70 
 Differences for grade five is 126.44 =>= P. E. 7.61 
 
 The superiority, in this respect, of Grade five over Grade eight is 
 18.63 per cent. ± P. E. 9.53. The P. E. of the average of Grade six 
 is so large as to make comparisons with that grade meaning- 
 less. Although Grades eight and five do differ by twice the 
 P. E., the exception in the case of Grade six and the possibility 
 
 TABLE XXIX 
 
 Showing the score obtained in the retention tests for sense material 
 
 Method 
 
 I 
 
 2 
 
 3 
 
 4 
 
 s 
 
 6 
 
 Grade eight 
 
 Relative score 
 
 9 59 
 
 II .60 
 
 15 29 
 
 15-77 
 
 15-51 
 
 14 53 
 
 Interval four hours 
 
 P. E. 
 
 0.37 
 
 0.50 
 
 0.46 
 
 0.54 
 
 0.47 
 
 0-59 
 
 Grade six 
 
 Relative score 
 
 8.13 
 
 8.61 
 
 12.36 
 
 13 -43 
 
 12.99 
 
 II. 13 
 
 Interval three hours 
 
 P. E. 
 
 0.43 
 
 0.40 
 
 0.60 
 
 0.60 
 
 0.62 
 
 0.57 
 
 Grade five 
 
 Relative score 
 
 717 
 
 8.20 
 
 10.51 
 
 12.28 
 
 10.79 
 
 II .62 
 
 Interval three hours 
 
 P. E. 
 
 0.27 
 
 0.29 
 
 0.39 
 
 0.44 
 
 0.36 
 
 0.39 
 
 Grade four 
 
 Relative score 
 
 7.66 
 
 9.14 
 
 9.67 
 
 11.23 
 
 10.36 
 
 9.90 
 
 Interval four hours 
 
 P. E. 
 
 0.49 
 
 0.44 
 
 0.47 
 
 0.61 
 
 0.47 
 
 0.47 
 
 Grade three 
 
 Relative score 
 
 4-75 
 
 5.83 
 
 8.16 
 
 9.40 
 
 8.89 
 
 8.70 
 
 Interval three hours 
 
 P. E. 
 
 0.36 
 
 0.39 
 
 0.41 
 
 0.43 
 
 0.44 
 
 0.34 
 
 of a more general unreliability of the data (see p. 46) for fine 
 distinctions, casts doubt upon this apparent difference between 
 grades. 
 
 In general, then, the results for the retention of nonsense syllables 
 are similar to those found in immediate tests, with the important 
 
RECITATION AS A FACTOR IN MEMORIZING 
 
 49 
 
 difference that the superiority of the methods involving recitation 
 is much greater. 
 
 RESULTS AS REGARDS RETENTION OF SENSE MATERIAL 
 
 Tests for retention of the sense material were given from three 
 to four hours after the learning tests, the time always being the 
 same for each class. The names of the individuals whose biogra- 
 phies had been studied were written on the board and the pupils 
 were asked to write all they could remember about each person. 
 Ample time was given. 
 
 TABLE XXX 
 
 Showing the data of Table XXIX on a relative basis 
 
 Method 
 
 I 
 
 2 
 
 3 
 
 4 
 
 5 
 
 6 
 
 Grade eight 
 
 Relative score 
 
 79 58 
 
 96.26 
 
 126.88 
 
 130.87 
 
 128.71 
 
 120.58 
 
 
 P. E. 
 
 3.06 
 
 4.14 
 
 3-81 
 
 4-47 
 
 3-89 
 
 4.89 
 
 Grade six 
 
 Relative score 
 
 74-31 
 
 78.70 
 
 112.97 
 
 122.76 
 
 118.73 
 
 101.73 
 
 
 P. E. 
 
 3 93 
 
 3-65 
 
 5-48 
 
 548 
 
 5-66 
 
 5.20 
 
 Grade five 
 
 Relative score 
 
 71 .06 
 
 81.26 
 
 104. 16 
 
 121 .70 
 
 106.93 
 
 115. 16 
 
 
 P. E. 
 
 2.71 
 
 2.90 
 
 3-89 
 
 4.42 
 
 3.60 
 
 3 90 
 
 Grade four 
 
 Relative score 
 
 79.29 
 
 94.61 
 
 100. 10 
 
 116.25 
 
 107.24 
 
 102.48 
 
 
 P. E. 
 
 5-04 
 
 4-53 
 
 4.84 
 
 6.28 
 
 4.84 
 
 4.84 
 
 Grade three 
 
 Relative score 
 
 62.33 
 
 76.50 
 
 107.08 
 
 123.35 
 
 116.66 
 
 114. 17 
 
 
 P. E. 
 
 4-71 
 
 5-10 
 
 5-37 
 
 5-63 
 
 5-76 
 
 4 45 
 
 Table XXIX shows the results for the various grades in the form 
 of averages with P. E.'s computed as before. Table XXX gives the 
 same data on a relative basis. Figure 3 gives the data of Tables 
 XXX and XXXI in graphic form. 
 
 TABLE XXXI 
 
 Showing the average of the results for all grades 
 
 Method 
 
 r 
 
 2 
 
 3 
 
 4 
 
 5 
 
 6 
 
 Average 
 P. E. 
 
 73-31 
 1-93 
 
 85 -44 
 2-45 
 
 110.23 
 
 2.83 
 
 122.98 
 1-43 
 
 115-65 
 2.46 
 
 110.82 
 2.27 
 
 Table XXXII following shows the differences between the vari- 
 ous methods with the P. E. of the differences, computed from the 
 data of Table XXXI. 
 
50 
 
 RECITATION AS A FACTOR IN MEMORIZING 
 
 TABLE XXXII 
 
 Showing the differences between the various methods 
 with the P. E. of the differences 
 
 Methods 
 
 Methods 
 
 Methods 
 
 2- 
 
 -I =I2.I3±P. 
 
 E. 3.12 
 
 3— 2 =24.79 ± P. 
 
 E. 
 
 3-74 
 
 4—3 = 
 
 i2.75=fcP. E. 3.16 
 
 3- 
 
 -I =36.92 ± P. 
 
 E. 342 
 
 4— 2 =37-54 =^ P- 
 
 E. 
 
 2.83 
 
 5—3 = 
 
 5.42 ±P.E. 3.74 
 
 4- 
 
 -I =49.67=1=?. 
 
 E. 2.40 
 
 5— 2 =30.21 ±P. 
 
 E. 
 
 346 
 
 6-3 = 
 
 o.59±P. E. 3.63 
 
 5- 
 
 -I =42.34 ± P. 
 
 E. 3.12 
 
 6— 2 =24.38 ± P. 
 
 E. 
 
 3-34 
 
 5—4 = 
 
 -7.33=*= P. E. 2.84 
 
 6- 
 
 -i=37.ii=tP. 
 
 E. 2.98 
 
 
 
 
 6-4 = - 
 6-5 = 
 
 -i2.i6=fcP. E. 2.68 
 -4.83 ±P.E. 3.34 
 
 From these tables it is clear that the general results for retention 
 are very similar to those obtained in the immediate test. Method 
 Four is the best, while Method One is the poorest. In the immediate 
 test (see Table XXIV) Method Four showed a superiority over 
 Method One of 26.64 per cent. ± P. E. 2.5 while in the retention 
 test (see Table XXXII) the difference is 49.67 per cent. ± P. E. 
 2.4 or nearly twice as great. The earlier finding that Methods Five 
 and Six are somewhat inferior to Method Four is borne out by the 
 similar result in the retention test. On the whole the results of the 
 immediate and the delayed tests are similar except that the differ- 
 ences between methods are more pronounced in the retention tests. 
 In the immediate test Method Four was superior to Method One by 
 26.64 per cent. =*= P. E. 2.5; while in the retention test the super- 
 iority is 49.67 per cent. ± P. E. 2.4. 
 
 Differences Among Grades 
 
 When the differences between grades are considered, the corre- 
 spondence of the results for immediate and delayed memory is not 
 so close. The finding in the immediate tests, that the difference 
 between the best method and the poorest method was much 
 greater for the lower grades, is not borne out by the results for 
 retention. 
 
 The average superiority of Method Four over Method One 
 
 For grade eight = 51.29 per cent. ± P. E. 5.38 
 
 For grade six = 48.45 per cent. =fc P. E. 6.70 
 
 For grade five = 50.64 per cent. =fc P. E. 5.19 
 
 For grade four = 36.96 per cent. =t P. E. 8.06 
 
 For grade three = 61.02 per cent. =*= P. E. 7.28 
 
 The magnitude of these differences shows no correspondence to 
 order of grades. 
 
Method I 
 
 RECITATION AS A FACTOR IN MEMORIZING 
 FIGURE 3 
 Based on the data of Tables XXX and XXXI 
 2.3 4 5 
 
 51 
 
52 RECITATION AS A FACTOR IN MEMORIZING 
 
 Another difference between grades previously found, namely, 
 that the methods introducing recitation very early worked a hard- 
 ship upon the lower grades but not on the upper ones, is not shown 
 by the results for retention. 
 
 The superiority of Method Four over Method Six, for example, is 
 
 For grade eight = 10.29 per cent. ^ P. E. 6.63 
 
 For grade six = 11.03 per cent. =t P. E. 7.55 
 
 For grade five = 6.54 per cent. =*= P. E. 5.83 
 
 For grade four = 13.77 per cent. =fe P. E. 7.93 
 
 For grade three = 9.18 per cent. =*= P. E. 7.14 
 
 The differences between grades in this respect are nil. In fact, all 
 of the differences between the two methods are so small in compar- 
 ison with the P. E. as to be of very doubtful significance. 
 
 A third difference found on the immediate tests — that for the 
 upper grades a method involving sixty per cent, reading (Method 
 Three) was quite as good as any other, while for the lower grades 
 this method was distinctly inferior to the methods giving more time 
 to recitation — is quite clearly indicated in the findings for 
 retention. 
 
 The superiority of Method Four over Method Three 
 
 For grade eight = 3.99 per cent. =t P. E. 5.83 
 
 For grade six = 9.79 per cent. =t P. E. 7.74 
 
 For grade five = 17.54 per cent. =t P. E. 5.83 
 
 For grade four = 16.15 per cent. ± P. E. 7.93 
 
 For grade three = 16.27 P^r cent. =t P. E. 7.68 
 
 For Grades six and eight. Method Four shows no real superiority 
 over Three, but in the case of Grades five, four, three real differences 
 appear. 
 
 Summary of Results for Retention of Sense Material 
 
 In general, the results are similar to those found for immediate 
 tests, the differences between the best and poorest methods being 
 somewhat greater. The superiority of Method Four over Method 
 One in the immediate test was 26.64 per cent. =t= P. E. 2.5 as 
 compared to 49.67 per cent. ± P. E. 2.4, or very nearly twice 
 as great. 
 
 In some respects, the findings for the retention tests have not 
 borne out the earlier results concerning differences between the 
 various classes. But, as was explained before, certain possible 
 sources of error in the data from the retention tests render these 
 
RECITATION AS A FACTOR IN MEMORIZING 53 
 
 results of doubtful value for fine comparisons. The findings in the 
 immediate tests are probably better indications of real distinctions 
 between grades. 
 
 RESULTS FROM EXPERIMENTS UPON ADULTS 
 WITH NONSENSE MATERIAL 
 
 In addition to the experiments carried out with children as 
 subjects, tests were made upon adult students, using materials and 
 methods similar in most respects to those previously described. 
 The purpose of this extension of the work was threefold : first, to 
 furnish a basis of comparison of the present results with those of 
 earlier investigations; second, to permit a comparison of the work 
 of children with adults in similar tests; and third, to obtain more 
 detailed information concerning the nature of the particular func- 
 tions employed in the two methods of learning. 
 
 The subjects for the tests to be described were for the most part 
 graduate students, members of classes in experimental psychology. 
 The materials used were qualitatively the same as those employed 
 with the children. 
 
 Fifteen such students of psychology at Columbia were given, on 
 three different days, ten-minute tests with series of twenty nonsense 
 syllables. Each day one of the three different methods of study was 
 used; first, lo' reading; second, 5' L + 5' R; third, 2' L + 8' R. 
 
 The group was divided into three squads, and practice effects, 
 individual differences and differences in tests were neutralized in 
 the total by employing a method in all essentials the same as that 
 described on p. 26 ff. Therecordsof the individual students, however, 
 are fairly reliable as such for the reason that all of these subjects 
 had just completed a series of experiments on the learning process 
 and memory extending over three months. Each had learned 
 during this time several hundred nonsense syllables as well as much 
 other material and were thus fairly highly practised subjects. 
 Detailed introspective accounts of the factors involved in the 
 several methods of learning were requested. These will be consid- 
 ered later. Each person acted as subject for his or her regular 
 laboratory partner, who kept the time, and noted the number of 
 repetitions made. Later the two reversed positions, the former 
 experimenter now acting as subject. Each used a different series 
 of syllables, and six different texts were used altogether. 
 
 The data were scored by giving a grade for each correct letter 
 when there were two or more correct, and an additional credit when 
 the syllable was in correct position. The highest possible score for 
 the twenty syllables would thus be eighty.'' 
 
 'See Lyon, D. O., 'The Relation of Quickness of Learning to Retentiveness", Archives of 
 Psychology, 1916, No. 24, p. 27. 
 
54 
 
 RECITATION AS A FACTOR IN MEMORIZING 
 
 Table XXXIII shows the records of the various individuals with 
 the average, the A. D.'s, for the number of repetitions and the 
 P. E.'s for the obtained average. The P. E.'s were computed by 
 
 a- dis. 
 
 the formula P. E. = .6745 — 
 
 Vn 
 
 The differences between the methods as shown by the average 
 results are large and reliable, the P. E.'s being small. In general 
 Method Three results in more than twice as much material learned 
 
 TABLE XXXIII 
 
 Showing the scores obtained by adults in learning nonsense 
 syllables. Highest possible score eighty 
 
 
 Method I. Lio' 
 
 M 
 
 ;tIiod 2. Ls' 
 
 R5' 
 
 Metliod 3- L2' 
 
 R8' 
 
 Subject 
 
 Repeti- 
 
 Repetitions 
 
 
 Repetitions 
 
 
 
 tions Score 
 
 L 
 
 R 
 
 Score 
 
 L R 
 
 Score 
 
 
 Day one 
 
 
 Day two 
 
 
 Day three 
 
 
 Hm. 
 
 62 22 
 
 26 
 
 21 
 
 60 
 
 6 25 
 
 78 
 
 Dn. 
 
 54 32 
 
 22 
 
 18 
 
 49 
 
 5 21 
 
 80 
 
 Bm. 
 
 23 16 
 
 13 
 
 II 
 
 58 
 
 5 10 
 
 74 
 
 Bs. 
 
 22 32 
 
 15 
 
 10 
 
 60 
 
 4 9 
 
 80 
 
 Tr. 
 
 50 23 
 
 Day two 
 
 24 
 
 23 
 
 Day three 
 
 38 
 
 6 15 
 
 Day one 
 
 68 
 
 Py. 
 
 55 7 
 
 20 
 
 6 
 
 19 
 
 6 9 
 
 47 
 
 Gl. 
 
 68 9 
 
 17 
 
 5 
 
 40 
 
 7 13 
 
 49 
 
 Bn. 
 
 26 38 
 
 23 
 
 14 
 
 52 
 
 7 16 
 
 78 
 
 Jy. 
 
 26 28 
 
 9 
 
 7 
 
 32 
 
 6 14 
 
 42 
 
 Hr. 
 
 38 36 
 Day three 
 
 12 
 
 10 
 
 Day cue 
 
 56 
 
 5 16 
 
 Day two 
 
 50 
 
 Mk. 
 
 37 24 
 
 18 
 
 H 
 
 48 
 
 6 16 
 
 60 
 
 Sn. 
 
 50 29 
 
 24 
 
 20 
 
 58 
 
 9 29 
 
 74 
 
 At. 
 
 36 56 
 
 10 
 
 8 
 
 56 
 
 4 18 
 
 63 
 
 Ly. 
 
 23 12 
 
 25 
 
 15 
 
 43 
 
 6 II 
 
 62 
 
 Wp. 
 
 29 23 
 
 10 
 
 14 
 
 30 
 
 4 15 
 
 50 
 
 Average 
 
 40 25.8 
 
 18 
 
 13 
 
 46.6 
 
 5-7 16 
 
 63.7 
 
 
 A.D.14 P.E.I. 2 
 
 A.D.5 
 
 2 A.D.5.1 
 
 P.E.1.4 
 
 A.D.i.o A.D.4.0 
 
 P.E.1.4 
 
 as Method One, Method Two stands slightly above the average for 
 the other two. Minor diflferences among individuals will be appa- 
 rent on observation, but it will be noticed that in but one case is 
 Method Three inferior to Method Two and in no case is Method Two 
 inferior to Method One, although in one case Method Two is equal 
 to and in two other cases very slightly superior to Method One. The 
 
RECITATION AS A FACTOR IN MEMORIZING 55 
 
 superiority of Method Three over the others, is somewhat greater 
 than the data show in two cases. Subject Bs. had completely 
 learned the series in 8' 42" under Method Three and Subject Dn. 
 had completed the learning some time (exact amount not known) 
 before the end of the ten-minute period. 
 
 Great individual differences appear with regard to the number 
 of repetitions made during the ten-minute study period. The 
 average number of repetitions when the learning was entirely by 
 means of reading was forty, with a mean variation of fourteen. 
 The extreme rates were those of Subject Bs. with twenty- two 
 repetitions and Gl. with sixty-eight, or three times as many as Bs. 
 Method Two shows similar individual differences in the learning by 
 reciting as well as by reading. The average number of repetitions 
 for five-minute reading being eighteen with a M. V. of 5.2 and for 
 five-minute recitations the average number is thirteen with a M. V. 
 of 5.06. The average figures also show that the rates of repetition 
 were less for learning by reciting than for learning by reading, 
 although as far as this test is concerned, the difference may be 
 taken to mean merely that the repetitions in the last half of a 
 period of learning are longer than those of the first half. That the 
 former interpretation is more likely to be the correct one is indicated 
 by the fact that the sum of the repetitions for the all reading test 
 (Method One) is greater than for the half reading, half recitation 
 test (Method Two), i. e., forty as compared to thirty-one. This 
 greater speed of repetitions in the reading portion of Method Two 
 is shown by fourteen of the fifteen individuals. Method Three 
 shows the same situation, the total number of repetitions here 
 being twenty-two, with rather wide differences among individuals. 
 
 More Intensive Work with Nonsense Syllables 
 
 Somewhat more extensive work was done with two graduate 
 students, more skilled in introspective observation. Each of these 
 subjects was given several preliminary tests to insure an acquain- 
 tance with the procedure and to eliminate practice effects to some 
 extent, before the main experiment was begun. Series of twenty 
 nonsense syllables were studied for eight minutes according to six 
 different methods. Three tests were made by each method, and in 
 each case the number and duration of the repetition were noted by 
 the writer who kept the time with a stop watch. But one test was 
 made on a single day. A recall test was made after approximately 
 six hours for Subject 5 and after twenty-four hours for Subject T. 
 The following table gives the results in detail.^ The data were 
 
 ' The durations of the repetitions are not presented here, but will be mentioned in a later 
 section. 
 
56 
 
 RECITATION AS A FACTOR IN MEMORIZING 
 
 scored in the manner described on p. 53, eighty being the highest 
 score possible. The 'natural method' gave the subject liberty to 
 study in any way he might choose. 
 
 The results for Subjects S and T are very much the same as the 
 average results just found for the larger group. The differences 
 between Methods Four and Five for both subjects are too small to 
 
 TABLE XXXIV 
 
 Showing the average results for three trials of each subject 
 
 
 Metliod One 
 8'L 
 
 Method Two 
 6'L 2'R 
 
 Method Three 
 4'L 4'R 
 
 Combinations 
 
 Repeti- 
 tions 
 
 Score 
 
 Repetitions 
 
 Score 
 
 Repetitions 
 
 Score 
 
 
 L 
 
 L 
 
 R 
 
 L 
 
 R 
 
 
 Subject S, Average 
 Subject T, Average 
 Subject S, recall after 
 
 six hours. Average 
 Subject T, recall after 
 
 twenty-four hours. 
 
 Average 
 
 17 
 32 
 
 24 
 16 
 
 7.6 
 8.6 
 
 8 
 
 9 
 
 2.3 
 
 3-0 
 
 34 
 27 
 
 10.3 
 II-3 
 
 6 
 13 
 
 4-3 
 
 lO.O 
 
 48 
 34 
 
 17.6 
 13-3 
 
 
 Method Four 
 2'L 6'R 
 
 Method Five 
 I'L 7'R 
 
 Method Six 
 Natural Method 
 
 Combinations 
 
 Repetitions 
 
 Score 
 
 Repetitions 
 
 Score 
 
 Repetitions 
 
 Score 
 
 
 L 
 
 R 
 
 L 
 
 R 
 
 L 
 
 R 
 
 
 Subject S, Average 
 Subject T, Average 
 Subject S, recall after 
 
 six hours. Average 
 Subject T, recall after 
 
 twenty-four hours. 
 
 Average 
 
 3-5 
 4.0 
 
 5-6 
 18.0 
 
 49 
 49 
 
 27-3 
 22.6 
 
 2 
 3 
 
 8 
 22 
 
 52 
 
 26.3 
 
 24.0 
 
 2.6 
 30 
 
 9 
 20 
 
 51 
 50 
 
 29.0 
 21.3 
 
 be of significance and the 'natural method' produces results that 
 are quite as good as any other. This means that trained subjects 
 are capable of discovering and employing the best methods of 
 attack. Subject 5 began to attempt to recite in the 'natural 
 method' after two, four, and two repetitions respectively or at 
 about the same stage at which recitation was begun in Methods 
 Four or Five. The case is similar for Subject T. Subject 5 in Methods 
 Four, Five, or Six learned about twice as much as in Method One, 
 
RECITATION AS A FACTOR IN MEMORIZING 57 
 
 while Subject T learned about three times as much. The recall 
 tests after six or twenty-four hours bear out the findings for the 
 learning test, being somewhat more emphatic. For Subject 5 under 
 the optimum methods shows an amount recalled three times as 
 great as under Method One while for Subject T the ratio is nearly 
 four to one. 
 
 The speed of repetitions varies considerably for the two sub- 
 jects and for the same subject at different times, but in nearly 
 all cases reading seems to be done at a higher speed than recita- 
 tion, although as will be found later the duration of repetitions 
 during reading are very uniform while those during recitation are 
 very irregular. 
 
 In addition to the data here presented, a few additional experi- 
 ments, somewhat more specialized in nature, were performed and 
 are presented in a later section (pp. 71 and 72), in which four- 
 teen adult subjects participated in two five-minute periods of 
 studying sixteen nonsense syllables, according to two methods: 
 first, in which only reading was permitted ; and second, in which 
 recitation was permitted from the first. The average results show 
 a score 16.4 for the reading method and 32.85 for the recitation 
 method, or exactly twice as much. A similar test (p. 81) with eleven 
 subjects gave similar results, 5.54 syllables being correctly recalled 
 in the reading test as compared to 11. 4 in the recitation test. 
 
 ^Summary of Res ults for Adul ts with Nonsense Syllables 
 
 1. Several different experiments upon adult students in learning 
 nonsense syllables produce results similar to those found for children. 
 The advantage of methods affording an optimum amount of recita- 
 tion over the reading methods is very great, the two methods 
 showing in general a ratio of about two to one. 
 
 2. Although considerable individual differences were found, no 
 .subject was dis covered who did not o btain better results with 
 
 recitation than without it. 
 
 3. Great individual differences were found in the rate at which 
 the series were read or recited, but in general the duratio ns o f 
 
 Recita tions are longer than the duratio ns of read ings. 
 
 4. ^The ad vantage of the methods combining recitation with 
 reading in the learning period is more pronounce d in delayed than 
 in immediate recall. 
 
 EXPERIMENTS UPON ADULTS WITH SENSE MATERIAL 
 
 Non-connected Sense Material 
 Two graduate students, 5 and T, acted as subjects for a few tests, 
 in studying for eight-minute periods series of thirty words of four 
 
58 
 
 RECITATION AS A FACTOR IN MEMORIZING 
 
 letters each, according to several different methods. Two prelim- 
 inary trials were given in each case before the actual series were 
 started. Two series of tests were given, the order of methods being 
 reversed in the second series. No word was repeated in the series 
 of lists used. The data were scored by giving a credit of two for a 
 correct word and an additional credit if it were in the correct posi- 
 tion. Thus the highest possible score would be ninety. Table 
 XXXV gives the results: 
 
 TABLE XXXV 
 
 
 Method one 
 8' L 
 
 Metliod two 
 4'L 4'R 
 
 Method three 
 2'L 6'R 
 
 Method four 
 Natural method 
 
 Combina- 
 tions 
 
 Repeti- 
 tions 
 
 Score 
 
 Repetitions 
 
 Score 
 
 Repetitions 
 
 Score 
 
 Repetitions 
 
 Score 
 
 
 L 
 
 L 1 R 
 
 L 1 R 
 
 L R 
 
 
 
 
 
 
 Subject S 
 
 
 
 
 
 
 First 
 
 
 
 
 
 
 
 
 
 
 
 
 series 
 
 12 
 
 30 
 
 7 
 
 4 
 
 49 
 
 3 
 
 10 
 
 53 
 
 3 
 
 II 
 
 51 
 
 Second 
 
 
 
 
 
 
 
 
 
 
 
 
 series 
 
 15 
 
 35 
 
 6 
 
 6 
 
 58 
 
 2.5 
 
 II 
 
 60 
 
 2 
 
 II 
 
 54 
 
 Average 
 
 13-5 
 
 32.5 
 
 6.5 
 
 5 
 
 53-5 
 
 2.7 
 
 10.5 
 
 56.5 
 
 2.5 
 
 II 
 
 52.5 
 
 
 
 
 
 Subject 
 
 T 
 
 
 
 
 
 
 First 
 
 
 
 
 
 
 
 
 
 
 
 
 series 
 
 14 
 
 27 
 
 8 
 
 7 
 
 39 
 
 3 
 
 14 
 
 45 
 
 3 
 
 13 
 
 50 
 
 Second 
 
 
 
 
 
 
 
 
 
 
 
 
 series 
 
 16 
 
 34 
 
 9 
 
 7 
 
 42 
 
 3 
 
 13 
 
 47 
 
 3 
 
 15 
 
 47 
 
 Average 
 
 15 
 
 30.5 
 
 8-5 
 
 7 
 
 40.5 
 
 3 
 
 13-5 
 
 46 
 
 3 
 
 14 
 
 48.5 
 
 
 
 
 Subject S. 
 
 Recall 
 
 after six hours 
 
 
 
 
 First 
 
 
 
 
 
 
 
 
 
 
 
 
 series 
 
 
 15 
 
 
 
 30 
 
 
 
 
 
 
 20 
 
 Second 
 
 
 
 
 
 
 
 
 
 
 
 
 series 
 
 
 12 
 
 
 
 17 
 
 
 
 30 
 
 
 
 
 Average 
 
 
 13-5 
 
 
 
 23 -5 
 
 
 
 30 
 
 
 
 20 
 
 
 
 Subject 
 
 T. 
 
 Recall 
 
 after six hours 
 
 
 
 
 First 
 
 
 
 
 
 
 
 
 
 
 
 
 series 
 
 
 12 
 
 
 
 25 
 
 
 
 
 
 
 
 Second 
 
 
 
 
 
 
 
 
 
 
 
 
 series 
 
 
 14 
 
 
 
 29 
 
 
 
 30 
 
 
 
 
 Average 
 
 
 13 
 
 
 
 27 
 
 
 
 30 
 
 
 
 
RECITATION AS A FACTOR IN MEMORIZING 
 
 59 
 
 While this experiment is far from being extensive enough to be 
 decisive, it is suggestive. The subjects were well habituated to this 
 kind of learning, having previously learned nearly thirty series of 
 nonsense syllables. Both agree in showing that lists of words can 
 be more readily learned by a method which permits recitation, but 
 the difference between the methods is not so pronounced as was 
 found with nonsense syllables. For Subject S, in learning nonsense 
 syllables. Method Four was related to Method One as two to one; 
 for Subject T the ratio was nearly three to one ; while for lists of 
 meaningful words the ratios of the corresponding methods are for 
 Subject 5 about one and seven-tenths to one, for Subject 7" one and 
 five-tenths to one. The retention tests for series of words show a 
 similar ratio, although the data are too few for reliable results. 
 
 Experiments with Connected Sense Material 
 
 Subject T endeavored in six different tests of ten minutes each 
 to learn twenty-line stanzas of poetry from Goldsmith's 'Deserted 
 Village', according to three different methods. Recall of the 
 material was attempted after six hours. The results show the 
 number of words learned or remembered. 
 
 TABLE XXXVI 
 
 
 Method One 
 lo' L 
 
 Method Three 
 5'L s'R 
 
 Method Four 
 
 Combinations 
 
 Repeti- 
 tions 
 
 Score 
 
 Repetitions 
 
 Score 
 
 Repetitions 
 
 Score 
 
 
 L 
 
 L 
 
 R 
 
 L 
 
 R 
 
 
 First series 
 Second series 
 
 Average 
 
 II 
 
 12.5 
 
 78 
 86 
 
 82 
 
 6 
 
 5 
 
 4 
 5 
 
 93 
 io6 
 
 99-5 
 
 3 
 3 
 
 7 
 8 
 
 84 
 108 
 
 96 
 
 Recall after six hours 
 
 First series 
 
 
 44 
 
 
 
 66 
 
 
 
 
 Second series 
 
 
 53 
 
 
 
 58 
 
 
 
 58 
 
 Average 
 
 
 49-5 
 
 
 
 62 
 
 
 
 58 
 
 The advantage of the methods including recitation over the 
 reading method is apparent although not very large in both the 
 immediate and the delayed memory test. The method employing 
 fifty per cent, reading and fifty per cent, recitation seems to be 
 quite as good as the method permitting seventy-five per cent, 
 recitation. The results, of course, are too few to be more than 
 
6o 
 
 RECITATION AS A FACTOR IN MEMORIZING 
 
 suggestive, although they do seem to be quite in harmony with the 
 findings for children. 
 
 The biographical sense material used with the school-children 
 (see p. 26) was studied by fifteen graduate students under three 
 different methods, as shown in Table XXXVII. The fifteen sub- 
 
 TABLE XXXVII 
 
 Showing the number of details or facts recalled 
 
 
 
 
 
 * 
 
 
 Combinations 
 
 Method One 
 
 Method Two 
 
 Method Three 
 
 
 
 8'L 
 
 4'L 4'R 
 
 2'L 6'R 
 
 
 Subjects 
 
 Day one 
 
 Day two 
 
 Day three 
 
 
 Bm 
 
 41 
 
 48 
 
 52 
 
 
 Ws 
 
 14 
 
 28 
 
 32 
 
 
 Ky 
 
 36 
 
 39 
 
 45 
 
 
 Tr 
 
 9 
 
 20 
 
 22 
 
 
 Sn 
 
 13 
 
 18 
 
 19 
 
 
 
 Day two 
 
 Day three 
 
 Day one 
 
 
 At 
 
 39 
 
 47 
 
 49 
 
 
 Rs 
 
 40 
 
 47 
 
 39 
 
 
 Gl 
 
 10 
 
 20 
 
 19 
 
 
 Py 
 
 8 
 
 22 
 
 20 
 
 
 Sa 
 
 8 
 
 18 
 
 21 
 
 
 
 Day three 
 
 Day one 
 
 Day two 
 
 
 Mn 
 
 14 
 
 21 
 
 26 
 
 
 Ce 
 
 23 
 
 27 
 
 31 
 
 
 An 
 
 18 
 
 24 
 
 27 
 
 
 Sn 
 
 19 
 
 18 
 
 21 
 
 
 Ms 
 
 18 
 
 30 
 
 26 
 
 A- 
 
 ^erage 
 
 20.6 
 
 28.4 
 
 29.9 
 
 P. 
 
 E. 
 
 2.1 
 
 19 
 
 19 
 
 jects were divided into three groups, each employing a different 
 method on the different days. Other details were the same as 
 those described on page 53. 
 
 The average results for Methods Two and Three are distinctly 
 superior to those for Method One, and this is true practically without 
 exception for all of the individual cases. The difference, in the 
 average results, between Methods Two and Three is small and being 
 no larger than the P. E. is unreliable. The general result of this 
 test upon adults is the same as that obtained from the older children 
 with the same sort of material. The value of recitation as compared 
 to reading is not so great as it is when nonsense material is used, 
 and no particular advantage is obtained by introducing the recita- 
 tion very early in the learning. 
 
RECITATION AS A FACTOR IN MEMORIZING 6 1 
 
 GENERAL SUMMARY 
 
 Nonsense Material 
 
 1. \n general, recitation, after a few ini tial readings, is of much 
 more value i n learning than m ore readin g. 
 
 a. Under the conditions of the present experiment a method 
 devot ing the first twenty per^nt. of the time to r eading followed 
 by eighty per cent, recitation will result in learnin g for immediate 
 
 jreproduction twice as much material as will a method of reading 
 _ only. 
 
 b. As measured by recall three to four hours later, the difference 
 between the two methods is about twice as great; four times as 
 much being recalled under the recitation method as under the 
 reading method. 
 
 2. After a certain amount of initial reading (one minute and 
 forty-eight seconds or twenty per cent, of the total time in this 
 experiment) the more quickly the recitation is introduced th e 
 better the results as measured by either immediate or delayed 
 recall. 
 
 3. No conspicuous differences appear between the results for 
 adult subjects and children or between the various grades with the 
 exception that the findings for the first grade differ from all others. 
 
 Sen se Material 
 
 1. In general the best results are obtained from a method devot- 
 ing about forty per cent, of the time to reading followed by an 
 equal amount of recitation . 
 
 2. In general, the optimum combination of reading and recita- 
 tion produces in immediate tests results superior by about twenty- 
 seven per cent, to those obtained from reading only. 
 
 a. The di fference shown by recall three or four hours later is 
 nearly twice as great as t hat shown in the immediate test . 
 
 3. In most respects the results for adults and for the various 
 grades are very similar. 
 
 4. In certain respects differences between the grades were found 
 on the basis of the results of immediate tests. 
 
 a. The advantage of the best methods over the poorest is 
 much greater in the lower grades than in the upper, e. g., the 
 average advantage for grades three and four of the best method 
 over the poorest is 35.99 per cent." as compared to 18 per cent., 
 the average for Grades six and eight. 
 
62 RECITATION AS A FACTOR IN MEMORIZING 
 
 h. Introducing recitation very early in the study period has a 
 disadvantageous effect upon the learning of the lower grades, but 
 has little or no ill effect upon the work of the upper grades. 
 
 c. The upper grades, in comparison with the lower, learn more 
 effectively under the methods involving a relatively large amount 
 of reading. 
 
 5, With the exception of (c) above, none of the differences 
 between grades were evident in the results of the retention tests. 
 a. This was believed to be due, in the main, to unavoidable 
 errors which crept into the retention tests (see p. 46). 
 
 Results from Tests on Adults 
 
 1 . The advantage of recitations over reading is greater the more 
 senseless and unconnected the material. 
 
 a. Advantage is greatest for nonsense syllables, less great 
 for lists of words, and still less great for connected prose or 
 poetry. 
 
 2. Great individual differences appear in the tempo of studying 
 by reading or recitation, some individuals completing a perusal on 
 the average in one-third of the time taken by others. 
 
 3. As a rule, the tempo is considerably quicker in reading than 
 in recitation, for most individuals. 
 
 4. Usually, a given individual during a single sitting, reads and 
 rereads at a very uniform speed, while -the rates for consecutive 
 recitations are very variable. 
 
 COMPARISON WITH RESULTS OF OTHER INVESTIGATIONS 
 
 The general findings in the present experiment upon children as 
 well as upon adults are in harmony with the results of most of the 
 earlier investigations, which were presented in Chapter II. It will 
 be necessary here to recall but briefly the conclusions obtained in 
 some of the more important of the earlier works. 
 
 Katzaroff found, by combining the results for three subjects, 
 four tests each, that fifteen readings of nonsense syllables — the test 
 being made seventy-two hours later — produced a score of six as 
 compared to twenty obtained from eight readings and seven recita- 
 tions. Other individuals in similar tests, showed even greater 
 differences. Witasek, Knors, and Abbott also verified the greater 
 effectiveness of recitation in learning nonsense syllables under 
 various conditions, although the quantitative determination of the 
 superiority of recitation has differed considerably. 
 
 The work of Kiihn, being more akin to the present experiments, 
 is of more value for comparative purposes. In immediate tests. 
 
 ^ 
 
RECITATION AS A FACTOR IN MEMORIZING 63 
 
 the superiority of recitation over reading found by Kiihn is very 
 similar to that found in the present work, for each of the several 
 materials used. Kiihn's conclusion (p. 422), "By the majority of 
 people [adults] recitation is much better than readings, and the 
 relative advantage is greater, the more senseless the material," is 
 verified by the present results with children as well as adult 
 subjects. 
 
 With regard to the present finding that the superiority of recita- 
 tion over reading is greater when measured by delayed than by 
 immediate recall, but little evidence has been produced by the 
 earlier studies. But the results that are available seem to be in 
 harmony with the present findings. For example Kuhn found 
 (see p. 8) that a lesson, although learned in very much less time 
 by means of recitation than by reading alone, was retained much 
 better and that the superiority of recitation in this respect became 
 greater the longer the retention test was delayed. 
 
 The matter of individual differences deserves consideration. 
 Abbott in experiments upon five subjects found one among these 
 for whom reading was a better method of learning than recitation 
 and Kiihn found the same in the case of three out of thirteen 
 subjects. Both investigators found that such learners employed a 
 'purely mechanical' form of learning or were of very strong visual 
 type — such that best results were obtained when the subject simply 
 'looks at a word and lets it soak in'. Abbott concludes, "We must 
 go back to the type of the individual to explain the processes and 
 relative efificiency in recall." This matter of learning types will be 
 taken up in more detail in the next section. For the present, while 
 there is no intention of contending that such extreme types as 
 those found by Abbott and Kiihn do not exist, the present work 
 indicates that they are in no wise numerically so prominent as their 
 findings would suggest. While Kiihn found three among thirteen 
 subjects, and Abbott one among five, in the present work, tests 
 upon fifty or more adults made under less artificial conditions have 
 not produced a single case of such 'mechanical' or 'strongly 
 visual' types. In no case has the method of learning by reading 
 given better results than a method in which recitation was also a 
 factor. Unfortunately, the data of the children cannot be employed 
 on this point with assurance, for the reason that the effects of a 
 particular method in the case of any individual may be marked by 
 practice effects, differences in texts, and the like. However, in spite 
 of all these differences, an examination of the individual data shows 
 that exceptions to the general rule that recitation is more effective 
 than reading are very, very rare. This fact has a very important 
 pedagogical significance, since it gives assurance that such appli- 
 
64 RECITATION AS A FACTOR IN MEMORIZING 
 
 cations as follow from a study such as the present one, may be 
 made by the teacher to her pupils as a whole without working a 
 hardship on more than a very few if any individuals. 
 
 Further considerations of interest to pedagogy, such as the opti- 
 mum point of introducing the recitation in the case of various 
 materials, and the efificacy of various minor functions employed in 
 learning, will be treated in more detail in the next two sections. 
 
V 
 
 AN ANALYSIS OF READING AND RECITATION 
 AS FACTORS IN LEARNING 
 
 The previous section, from an objective point of view, gave us 
 certain facts concerning two very broad and complex functions, 
 e. g., learning by reading and learning by recitation. It was found 
 that the results, measured in terms of the amount of material learned 
 in a given time and the amount retained after a given time, differed 
 considerably according to the proportion of time allotted to one or 
 the other of these two functions. One is interested to discover, if 
 possible, in just what manner these two broad functions differ, 
 since the result of their exercise is so markedly different. It is 
 likely that the best method of discovering these differences is to 
 analyse each of the complex functions, as far as possible, into their 
 elements, finding just what minor functions are operative and in 
 what manner they combine to make up the gross functions of learn- 
 ing in each case. If such an analysis can be successfully accom- 
 plished, the result should be a much better understanding of the 
 two functions as a whole and the production of valuable suggestions 
 with regard to the selection and combination of constituent func- 
 tions for the most economical methods of study. 
 
 But such a reduction of the complex functions into constituent 
 processes that shall be typical is by no means an easy or a certain 
 matter. Some of the elementary functions can be observed from 
 the outside and can be verified by objective tests, but most of the 
 facts can be observed only by the learner and we are forced to 
 limit ourselves to his reports upon them. Indeed, most of our 
 analysis is of the introspective, or more accurately retrospective 
 sort, subject to the limitations of this form of evidence. 
 
 In the present work an effort has been made to get reports as full 
 as possible, and as free from suggestion as possible, from subjects 
 believed to be reliable and capable. About forty subjects in all 
 were used, and they were subjects whose experience seems to have 
 fitted them for the retrospective work. Nearly all had had several 
 months' practice in introspective reporting, each having learned, 
 previous to the experiments, a large number of series of nonsense 
 syllables and other kinds of material and having had considerable 
 practice in describing their mental imagery in various sorts of 
 mental tasks. After each test in the present experiments the sub- 
 
66 RECITATION AS A FACTOR IN MEMORIZING 
 
 jects wrote a full account of the functions employed in the learning, 
 such as the kinds of imagery employed, the kinds of 'aids' used and 
 how, their attitude toward the work, the satisfyingness and annoy- 
 ingness of different methods, the fatigability of different methods, 
 and the like. Reports from the children were secured on many of 
 these points also. Wherever practicable the introspective accounts 
 were checked up or tested by manipulation of the data already at 
 hand or by new experiments devised to fit the case, and the results 
 of other studies have been freely drawn upon. 
 
 Before proceeding to the results, a few cautions should be indi- 
 cated. In the first place, individual differences play a large role. 
 No single individual at any time is likely to make use of all the 
 minor functions that will be described. Some subjects place more 
 emphasis upon certain functions, some upon others, and the same 
 individual usually changes his method to some extent according to 
 the nature of the material and the like. More constant differences 
 among individuals due to earlier training in learning methods or to 
 memory types will be mentioned. But just as we found in the pre- 
 ceding section no very sharp differences in mental type, and no 
 definite cases in which reading proved to be superior to recitation, 
 so we shall find that typical methods of learning contain the main 
 functions employed by nearly all learners. 
 
 A second caution is that wholesale conclusions from results 
 obtained mainly from adults should not be made to apply to the 
 learning of children. Necessarily the introspective accounts are 
 largely those of adults, but the reports of children have also been 
 considered to some extent, and where possible, introspective accounts 
 have been verified by objective data obtained from children. That 
 the minor functions employed by children should correspond closely 
 to those of adults has already been indicated by the fact that the 
 results of the exercise of the two general functions have been very 
 similar for both classes of subjects. 
 
 With these precautions in mind, a consideration of the various 
 activities, aids, and attitudes involved in learning and recalling any 
 material will now be taken up, special attention being given to the 
 differences that appear according to whether the method of learning 
 is reading or recitation. 
 
 Nearly all reports, in the first place, agree in emphasizing the 
 fact that learning even a series of sixteen or twenty nonsense 
 syllables is far from a simple mechanical task. The number and 
 variety of associative aids is remarkable. Where adults go to their 
 wit's end for such associations it can hardly be doubted that they 
 assist learning. A consideration of such aids is perhaps a good 
 place to begin. 
 
RECITATION AS A FACTOR IN MEMORIZING 67 
 
 In general it may be said that such aids to learning may be of 
 two sorts: one which is found in the material itself, needing only 
 to be noted and employed, and another sort which is worked into 
 the material by the learner. Of either sort some may be marked 
 off as motor in character and others as perceptual. 
 
 ARTICULATION 
 
 Although the tests for learning were always written, the majority 
 of adult subjects reported that practice in accurate pronunciation 
 of the material was an aid in learning. This was found to be par- 
 ticularly true in the case of nonsense material which was difficult 
 to articulate. Subjects report that the motor and auditory elements 
 of the words were secured better from reciting, especially when the 
 material offers great difficulty in pronunciation. The learner is 
 likely to begin by carefully articulating the material to himself 
 while reading, but if the reading is prolonged too long, these func- 
 tions are likely to be neglected. In many cases the explanation 
 given for this is that they were able to move down the series of 
 syllables more easily without articulating, depending more upon a 
 visual imprinting of the data. In recitation this is rarely the case. 
 When they attempt to recite the material, the articulation is a 
 most natural and in most cases an essential act. The reproduction 
 and practice of the motor act is an aid to learning. The school 
 children found considerable difficulty in pronouncing the syllables, 
 and for them actual articulation was more essential. The members 
 of the sixth and eighth grades in answer to the question: "Why are 
 the syllables so hard to learn?" wrote, many of them, "Because 
 they are hard to say." They also reported that they liked the 
 recitation because it gave them a better chance "to see if they 
 could say them." Movements of the lips, sometimes without, 
 although generally with whispering, especially in the lower grades, 
 were very marked in the recitation part of the learning period. 
 
 ACCENTS AND RHYTHMS 
 
 Articulation is usually accompanied by accenting or stressing 
 certain syllables or words, according to the report of nearly all 
 subjects. The following serve as samples. Subject Rs in one test 
 reported accenting syllables one, five, nine, thirteen, and seventeen 
 in the series, syllables five and nine being more strongly accented 
 than the others. Subject Py accented every third syllable. Sub- 
 ject Sn accented every fourth syllable strongly and every second 
 syllable less strongly. Subject Bn reported an increasing accent 
 within groups of four syllables, the last being most strongly accented, 
 followed by a drop to the minimum on the fifth. Sometimes these 
 
68 RECITATION AS A FACTOR IN MEMORIZING 
 
 accents are obvious to an observer who may notice the accompany- 
 ing motor activities such as nodding the head, tapping the finger, 
 or thumping the foot. That the children employ such accents was 
 usually evident from such signs and was usually indicated by the 
 whispering which accompanied the learning. 
 
 The value of such accents lies in the fact that a syllable comes to 
 be associated with its accent and the act of accenting tends to call 
 up the syllable. Although individuals show great differences in 
 their choice of accents and the same individual may often employ 
 different accents according to the material being studied, in any 
 one lesson the accentuation is usually constant and assists learning 
 through this tendency to repeat the same motor activity which 
 acts as a frame-work upon which the syllables may be affixed. 
 
 Such accentuation should, however, be considered in connection 
 with the almost universal employment of rhythm in the learning of 
 a series of syllables. Miiller and Schumann,^ Meumann^, and others 
 have shown the value obtained from the employment of rhythm in 
 learning. The kind of rhythm, like the kind of accentuation, varies 
 with individuals and materials. In learning a series of twenty 
 nonsense syllables, subject At divides the material into feet of three 
 syllables, the first being long and accented, the two following unac- 
 cented and short / yj w | '' u kj \ with a pause between 
 
 groups. Subject Py uses an identical rhythm. Subject Sii em- 
 ploys a trochaic measure with two pairs combined into a measure 
 of four by placing greater accentuation on the third and seventh 
 
 than on the first and fifth ^ . / kj \ kj / kj | . Subject 
 
 Tr employs a measure of four feet, a long accented syllable fol- 
 lowed by three short unaccented syllables with a pause between 
 measures, / w w v^ | '' v^ u l; 1 . 
 
 According to the reports of most individuals, the employment of 
 such rhythms is the most natural thing in learning by recitation, but 
 in reading they are not so frequently or easily used. Some report, 
 in the case of learning by reading, that they begin by arranging 
 the material for rhythmical perusal with accents and pauses but 
 abandon the method before the lesson is over because it seems to 
 be of no avail. It seemed that a method employing more visual 
 factors and less motor would work better; their efforts were directed 
 to 'looking hard' at the syllables to assist them to 'soak in'. 
 Several subjects, however, reported that they did use a rhythmical 
 division of the material throughout the reading, and their opinion 
 that it did not prove to be of great value was usually borne out by 
 the meagre results of the final tests. "With my eyes on the paper," 
 
 1 'Experimentelle BeitrSge zur Untersuchung des GedSchtnisses', Zeitschrift ftir Psychohgie, 
 1894, 6, pp. 81-191. 
 
 2 The Psychology of Learning, translated by Baird, 1913. 
 
RECITATION AS A FACTOR IN MEMORIZING 69 
 
 says subject S, "it is hard to do more than just read hard and think 
 about the individual syllables. I knew the rhythms and other aids 
 would be of more value if I could only look away from the list." 
 
 Auditory and Visual Types of Learners 
 
 It appears that there are certain differences in method according 
 to whether the learner rehes more upon auditory-motor elements 
 or upon visual elements in learning. In some cases in which the 
 learning is predominantly of the auditory-motor type, imprinting 
 consists in forming a series of auditory or vocal images of the 
 whispered words or a series of successive innervations of the vocal 
 muscles, which are often accompanied by sensations or images of 
 movements. The subject learns the sounds, muscular feelings, and 
 rhythmic sequences of the syllables which he memorizes. Repro- 
 duction may be a sort of melody in which the various syllables 
 assume their proper rhythmical positions. Usually in reproduction 
 the subject cannot get the whole series in consciousness at once. 
 He must start the series off and let it run its course. Now many of 
 these subjects report that reading is of value to a certain point, but 
 if no opportunity for recitation is afforded, the latter part of the 
 process of learning is very much hampered and complete learning 
 seems impossible. The presence of the words to the eye precludes 
 the subjective innervations which are essential for learning. A 
 different process seems to be involved when the visual stimuli are 
 absent. 
 
 Some subjects reported that they made use of visual imagery to 
 a much greater extent. They were not so greatly hampered by 
 lack of recitation. But no one was found who relied entirely upon 
 visual imprinting, auditory and motor elements being always em- 
 ployed as well. Of those who relied to the maximum upon visual 
 imagery, most employed a rhythmical division of the material to 
 some extent. Such subjects divided up the material into measures, 
 with a motor stressing of certain syllables coupled with a visualiza- 
 tion of all of the syllables, especially those that were accented. They 
 differed from the auditory-motor learners, apparently, only by 
 relying somewhat more upon visualization and less upon the audi- 
 tory and motor factors. None used visual imprinting alone. In 
 the learning by reading these subjects employed the visual factors 
 to the utmost, with the corresponding neglect of the auditory and 
 motor elements. While their results, as a rule, differed less for the 
 two methods than did those of the auditory-motor learners, in no 
 case were they so efficient in tests permitting no recitation as in 
 the tests in which recall was a factor. 
 
70 RECITATION AS A FACTOR IN MEMORIZING 
 
 Just as there was, among about forty different adult subjects, 
 none that could be called a purely visual learner, so there was none 
 that seemed to rely entirely upon auditory-motor factors. Visual- 
 ization to some extent usually entered into the learning of the latter. 
 The differences were merely those of emphasis upon one or another 
 factor, and, indeed, among the subjects were many who seemed to 
 be able to employ now some factors, now others, according to the 
 situation to be met. In general, learning by reading seemed to 
 throw the emphasis upon the visual method. 
 
 LOCALIZATION AND NOTING OF POSITIONS OF ITEMS 
 
 A number of aids to memorizing which are more of a 
 perceptual than a motor sort are usually employed. They 
 are closely connected with the motor aids of articulation, rhythm, 
 etc. Some of these depend upon peculiarities or divisions found 
 in the material itself, while others are worked into it by the 
 subject. 
 
 One important matter is the noting of the positions of certain 
 syllables. Such localizations seem always to be an aid to memory. 
 Sometimes localization is greatly aided by peculiarities within the 
 text, but often more arbitrary methods of obtaining a localization 
 schema are employed. Some report that they simply localize a 
 certain few 'head-liners' in the series by noting their positions in 
 visual space. Although they are not able to visualize all of the 
 items, a few are made to stand out plainly, serving as landmarks to 
 which others are attached. Other subjects divide the list into a 
 certain number of parts, a few syllables thus being denoted by their 
 numerical positions. A few report these localizations to be deter- 
 mined by modulation of the voice or dependent upon the rhythm 
 that is employed. But all report that these localizations are an aid in 
 memorizing and that they were more easily employed in recitation than 
 in reading. On the introspective side such reports as these are 
 found: (subject T) "In reading it was so easy to glide through the 
 series that I did not take the trouble to note any special points of 
 interest. It seemed that I could do more if I just looked hard at 
 the syllables, covering up my ears so that I could do nothing but 
 look. But when I began to recite I found that I had to note certain 
 syllables specially, which I afterwards used as starting and stopping 
 places." Evidently, recitation tests the value of the difTerent aids 
 and generally leads the learner to recognize the value of those which 
 serve the purpose desired. 
 
 In order to obtain some objective data on the matter of localiza- 
 tion, a test was given for that purpose. Fourteen graduate students 
 whose status and introspective training have been described, acted 
 
RECITATION AS A FACTOR IN MEMORIZING 
 
 71 
 
 as subjects. Lists of sixteen nonsense syllables were used as mater- 
 ial. All the subjects studied at the same time, half of them by the 
 reading method first and half by the recitation method first. Later 
 another experiment was given in which each used the other method. 
 Five minutes were devoted to the study and the syllables were 
 written down immediately afterwards. The subjects were then 
 asked to indicate those whose positions they felt certain were 
 correct, those which were doubtful, and those which they were 
 sure were incorrect in position. They were then asked to describe 
 the means or cues by which they made their judgments. The 
 results are shown in Table XXXVI I L 
 
 It should first of all be noted that almost exactly twice as high a 
 score was obtained by the recitation method, and this introduces a 
 factor which tends to produce a better showing in the matter of 
 accurate localization for the reading method. It will be noticed, 
 for example, that many subjects in the reading series were certain 
 of the positions of only two or three syllables, which were in nearly 
 every case the first, or the first and second, and the last. It is well 
 known that the first and last syllables are the easiest to learn and 
 to localize. In the reading series these two or three syllables form 
 
 TABLE XXXVIII 
 
 Results given in the absolute number of syllables 
 
 
 Reading 
 
 Subject 
 
 Judged 
 
 Really 
 
 
 Correct 
 
 Doubtful 
 
 Wrong 
 
 Correct 
 
 Wrong 
 
 Score 
 
 Ln. 
 
 2 
 
 2 
 
 
 
 I 
 
 3 
 
 7 
 
 Sa. 
 
 3 
 
 2 
 
 
 
 2 
 
 3 
 
 13 
 
 Sn. 
 
 3 
 
 4 
 
 I 
 
 4 
 
 4 
 
 17 
 
 Ms. 
 
 2 
 
 I 
 
 2 
 
 2 
 
 3 
 
 13 
 
 J. M. 
 
 7 
 
 2 
 
 
 
 6 
 
 3 
 
 19 
 
 Tr. 
 
 3 
 
 2 
 
 I 
 
 2 
 
 4 
 
 II 
 
 Wr. 
 
 2 
 
 2 
 
 2 
 
 3 
 
 3 
 
 10 
 
 Ce. 
 
 4 
 
 2 
 
 I 
 
 5 
 
 2 
 
 16 
 
 Py. 
 
 9 
 
 
 
 
 
 5 
 
 4 
 
 21 
 
 Gl. 
 
 4 
 
 4 
 
 
 
 3 
 
 5 
 
 14 
 
 Mn. 
 
 6 
 
 4 
 
 
 
 7 
 
 3 
 
 23 
 
 An. 
 
 6 
 
 3 
 
 
 
 4 
 
 5 
 
 18 
 
 At. 
 
 II 
 
 2 
 
 
 
 9 
 
 4 
 
 32 
 
 Rs. 
 
 6 
 
 2 
 
 
 
 5 
 
 3 
 
 16 
 
 Average 
 
 485 
 
 2.79 
 
 0.50 
 
 4.14 
 
 3-5 
 
 16.4 
 
 Per cent. 
 
 59-6 
 
 34-3 
 
 6.3 
 
 54-2 
 
 45-8 
 
 
72 
 
 RECITATION AS A FACTOR IN MEMORIZING 
 TABLE XXXVIU— Continued 
 
 
 Recitation 
 
 Subject 
 
 Judged 
 
 Really 
 
 
 Correct 
 
 Doubtful 
 
 Wrong 
 
 Correct 
 
 Wrong 
 
 Score 
 
 Ln. 
 
 6 
 
 I 
 
 3 
 
 4 
 
 
 
 26 
 
 Sa. 
 
 7 
 
 I 
 
 
 
 8 
 
 
 
 24 
 
 Sn. 
 
 i6 
 
 
 
 
 
 16 
 
 
 
 48 
 
 Ms. 
 
 7 
 
 3 
 
 
 
 10 
 
 
 
 23 
 
 J. M. 
 
 8 
 
 I 
 
 2 
 
 8 
 
 3 
 
 30 
 
 Tr. 
 
 13 
 
 2 
 
 
 
 14 
 
 I 
 
 46 
 
 Wr. 
 
 4 
 
 5 
 
 
 
 4 
 
 5 
 
 24 
 
 Ce. 
 
 12 
 
 2 
 
 
 
 13 
 
 I 
 
 42 
 
 Py. 
 
 8 
 
 
 
 
 
 8 
 
 
 
 24 
 
 Gl. 
 
 10 
 
 
 
 
 
 10 
 
 
 
 30 
 
 Mn. 
 
 8 
 
 2 
 
 
 
 8 
 
 2 
 
 26 
 
 An. 
 
 7 
 
 2 
 
 
 
 7 
 
 2 
 
 25 
 
 At. 
 
 15 
 
 
 
 
 
 15 
 
 
 
 44 
 
 Rs. 
 
 i6 
 
 
 
 
 
 16 
 
 
 
 48 
 
 Average 
 
 9.64 
 
 1-34 
 
 0.34 
 
 10.07 
 
 1 .0 
 
 32.85 
 
 Per cent. 
 
 85.1 
 
 II. 8 
 
 31 
 
 90.9 
 
 91 
 
 
 a larger portion of the whole number written down than in the 
 recitation series. 
 
 In spite of this advantage, the subjects, after learning by reading, 
 felt certain of the positions of but fifty-nine and six-tenths per cent, 
 of the syllables written down, as compared to eighty-five and one- 
 tenth per cent, in the recitation series. The reading series is con- 
 spicuous with respect to the number of 'doubtful' cases, which 
 amount to thirty-four and three-tenths per cent, as compared to 
 eleven and eight-tenths per cent, for the recitation series, or three to 
 one. So it is quite clear that the subjects are more confident of 
 their opinions in the recitation series. The data also show that a 
 much larger number of the syllables learned by recitation are in 
 correct position (ninety-one per cent.) than in the reading series 
 (fifty-four per cent.). Another important fact appears, namely, 
 that the judgments after learning by recitation are not only more 
 accurate but also more conservative than after learning by reading. 
 As to the first point, while fifty-nine and six-tenths per cent, of the 
 syllables written down after learning b}^ reading were judged to be 
 in correct position, only fifty-four and two-tenths per cent, actually 
 were, while in the recitation series, of those written a larger per- 
 centage (eighty-five and one-tenth) was judged to be in correct 
 
RECITATION AS A FACTOR IN MEMORIZING 73 
 
 position, and a still larger percentage (ninety and nine-tenths) 
 actually was. An examination of the original data showed that in 
 the reading series, of those judged 'doubtful' nearly all were really 
 in an incorrect position, as were also nine of the sixty-eight certified 
 as 'correct' in position, while in the recitation series some of those 
 judged 'doubtful' were really in a correct position, while only two 
 cases out of the total of 127 judgments of 'correct in position' were 
 wrong. It thus appears that after learning by recitation, the sub- 
 jects are both more accurate and more conservative in their judg- 
 ments. 
 
 NOTING UNUSUAL CHARACTERISTICS OF THE MATERIAL 
 
 The remark was just made that the noting of unusual words and 
 characteristics in the material was often an aid in localization. This 
 function is of value because it serves to break up the material into 
 units that can be more easily handled. A peculiar word or syllable 
 becomes a center around which other syllables are grouped, or it 
 may serve as a starting and stopping place within the series. 
 
 The kinds of peculiarities noted are myriad. Sometimes it is the 
 sound — the children especially are attracted by 'funny sounding' 
 syllables. Sometimes a syllable stands apart by having the con- 
 sonants each standing above or below the line, e. g., gop, lib. Some- 
 times the fact that one letter was printed light, or that the whole 
 was blotched or blurred, or that a mark appeared on the page 
 opposite it, is noted. More often the associations are meaningful, 
 and these will be considered more fully in the next section. 
 
 Subjects report that all such peculiarities are brought out more 
 clearly by reciting the material. They are not so effectively brought 
 into play when one is reading because the words before the eyes 
 render such aid unnecessary. The thing to do is simply to 'look 
 hard and try to avoid distractions'. Subject Fx reports: "After 
 the reading period was over [four minutes out of eight], I could 
 remember only three syllables. I had a hazy idea of some of the 
 others but I couldn't quite get them. But by picking out tw^o queer 
 looking syllables, the sixth and the tenth, I was soon able to fill in 
 those between." 
 
 MEANINGS OF TERMS AND RELATIONS OF PARTS 
 
 Subjects report that the nonsense syllables take on more meaning 
 during recitation. Some feel that in merely reading they take the 
 syllable as it stands; they may notice its form and position but 
 they do not try so hard to make it mean something. The meanings 
 come out more clearly when they are forced to reconstruct it in 
 recall. The kinds of meanings are various. Sometimes it is a far- 
 
74 RECITATION AS A FACTOR IN MEMORIZING 
 
 fetched resemblance to some familiar word, such as toq = toque, 
 soy = say, etc. Sometimes two words are combined to form a 
 single word, such as sor-dit = sordid, jor-kih = jerky. Often a 
 resemblance to a familiar foreign word is seized upon, qos = Latin 
 qiws, or a word is associated with a foreign equivalent, dit = French 
 'he says'. Again a syllable is employed as part of some familiar 
 word, as gov in governor, and still more common were associations 
 between the syllables and the 'nicknames' of known persons. Some- 
 times the recurrence of words having a similar look or sound is 
 noted, such as toq and doc, and occasionally the first letters of 
 successive syllables are combined to form a new word. Sometimes 
 the associations are less definite; the syllable merely feels big, or 
 dull, or bright, or buzzy, e. g., viz feels 'buzzy', likewise zop; dit is 
 short and snappy, qos seems to be 'such a mouth full'. 
 
 In the case of sense material, recitation leads to a more thorough 
 understanding, both of the minor details and of the meaning of the 
 thing as a whole. They size up the men described more definitely. 
 One subject reports, "In reading I was dealing more with a lot of 
 details, which I handled mostly in a verbal way. There was no 
 flesh and blood about the men. But during recitation, I could 
 really picture them as men of [such and such age, size, etc.]." It 
 appears that this better grasp of the meaning of the material 
 is an aid to memory. In this connection Meumann writes i^ "In 
 the case of coherent and meaningful material the chief memorial 
 support consists in the apprehension of the meaning and the 
 logical context." 
 
 CHANGING METHODS DURING A STUDY PERIOD 
 
 A few subjects reported that they believed one advantage of 
 recitation was to be found in the fact that they could shift from one 
 kind of imagery to another more readily. In reading they were 
 more likely to depend on visual imagery, or, as they reported, to 
 use no imagery at all, but simply look at the syllables, while in 
 recitation they would employ now one sort of imagery, now another, 
 or more accurately emphasize difi^erent sorts of imagery at different 
 times. One subject reports: "Sometimes I tried to recall by seeing 
 the words in my mind's eye, and sometimes by trying to remember 
 how it sounded, and again by trying to say several words quickly 
 without imagery. I think this helped since it made the work more 
 interesting and allowed me to resort to different methods when I 
 got stuck." This shifting from one method to another may have 
 made the work more absorbing, but its general value as an aid in 
 
 ' The Psychology of Learning, p. 297. ' 
 
RECITATION AS A FACTOR IN MEMORIZING 75 
 
 learning may well be doubted. At all events, the greater freedom to 
 employ any method that seems desirable is a notable characteristic 
 of learning by recitation. 
 
 PATTERNS AND GROUPINGS 
 
 Closely connected with the previous finding that recitation leads 
 to better articulation, accentuation, pauses, vocal inflections, use of 
 melody and rhythms, as well as to better localization, noting of 
 peculiarities and meanings in the material, is the finding that 
 recitation tends more toward a division and grouping of the mater- 
 ial. In reading, the syllables are handled more as isolated terms; 
 the learner tries to imprint each by itself. In recitation more of an 
 attempt is made to make the material over into some sort of pattern, 
 a more or less highly organized structure. The patterns differ 
 greatly among individuals and vary according to the list of syllables 
 used. Very often the structure is decidedly of a rhythmic character, 
 associations being formed between accented terms, their positions 
 and pauses, as we have seen. In these cases the associations between 
 members of a given foot are particularly strong, and the feet, al- 
 though they are in the beginning relatively independent, are bound 
 together in various ways. Sometimes the groups are of unequal 
 length, being determined by the location of syllables which for 
 various reasons stand out prominently. More often, of course, the 
 groups are of equal size, including from two to six syllables, usually 
 three or four. 
 
 Subjects report that this active process of dividing up the material 
 and making it over into groups is more easily done in recitation. It 
 is, however, very often done in reading also, but it is then more 
 difficult to do; the divisions cannot be made so sharply, and the 
 ease of reading down the series defeats their purpose. For example, 
 one subject {Bn) whose results were very poor in the reading tests, 
 said: "A certain amount of reading is valuable to get acquainted 
 with the material and to frame up a method of attack, but there- 
 after it seems to do me no good. I simply can't learn by more 
 reading, except by taking a small bit of the series, giving it special 
 attention at one time and later going through it very hurriedly. 
 The desire to look away from the paper to see if I can recite the 
 material is well nigh irresistible." This 'going through it very 
 ^JAirriedly', which the subject speaks of, is probable a very close 
 approach to recitation. 
 
 It thus appears that in the reading series the material is handled 
 more by separate items than by groups. Less effort is used to build 
 up a structural whole — there is less organization of the material. 
 Subjects 5 and T show in another way an advantage of recitation 
 
76 RECITATION AS A FACTOR IN MEMORIZING 
 
 which is dependent upon better organization of the material. The 
 following figures give the number of seconds for each of a number 
 of repetitions in several tests. 
 
 Subject T readhig 
 
 15, 12, II, 12, 12, 12, 14, 13, II, 15, 16, 13, II, 9, 15, etc. 
 14, 13, 10, 12, II, 15, 14, 17, 13, 13, 17, 9, 14, 13, II, etc. 
 
 16, 10, 12, 13, 14, 12, II, 16, 14, 9, II, 12, 16, 12, II, etc. 
 
 Recitation 
 37, 45, 62, 20, 45, 12, 45, 36, 6, 50, 35, 4 
 27. 5. 47, 52, 46, 8, 31, 45, 33, 12, 6 
 47. 27, 53> 12, 34, 5, 34, 2, 26, 53, 35 
 
 Subject S reading 
 24, 24, 18, 24, 32, 22, 25, 34, 26, 30 
 18, 14, 16, 18, 14, 20, 25, 23, 24, 19 
 
 Recitation 
 82, 90, 42, 72, 12, 87, 36, 12, etc. 
 72, 80, 36, 8, 46, 90, 42, 6, 45 
 
 In the first place it will be noted that the rates for readings are 
 very uniform. The subject reads and rereads in much the same 
 way, giving as we have seen about equal attention to all syllables. 
 But in the case of recitations, the rates of the repetitions are varied, 
 the average rate being slower with a much higher mean variation. 
 The subjects were able to account for this, in part at least. Usually 
 the material was divided into groups, different ones being featured 
 at different times. To begin with, the first group was hit hard, 
 perhaps also the last group, with the result that these two groups 
 were earliest learned. When these were fairly well under control, 
 attention was given to the second group, and so the learning pro- 
 gressed. The variations in the total time for repetitions are due to 
 the varied treatment of some of the groups. Usually a group was 
 perused very slowly when it first became an object of attack and 
 once having been fairly well mastered was passed over very rapidly, 
 except that now and then a more lengthy and more thorough 
 review might be given. 
 
 Of special interest and importance are the very short repetitions 
 of four, five, six, eight, etc. seconds which occur at various intervals, 
 being more numerous near the end of the study period. The sub- 
 jects reported that these amounted to very hasty reviews of the 
 whole series. In the beginning they served the purpose of providing 
 a better acquaintance with the material as a whole, while later on 
 they usually amounted to very hasty surveys of the material already 
 learned, either with or without much attention to the unlearned sylla- 
 bles. They served a two-fold purpose of economizing time and of work- 
 ing over the lesson as a whole. In the latter capacity they served the 
 
RECITATION AS A FACTOR IN MEMORIZING 77 
 
 purpose of building up associations between the various groups of 
 items and perfecting the organization of the whole structure. 
 
 G. E. Miiller,* who has made an extensive study of learning 
 methods, describes in the course of memorizing series of digits, non- 
 sense syllables, etc., several stages in the organization and grouping 
 of the material. With simultaneous presentation, the first stage is 
 a 'collective apprehension' of the row of items. This stage affords 
 opportunity to secure an acquaintance with the material generally 
 and to observe such near-lying cues as there may be that can be 
 employed in dividing up the material for further learning. A 
 second stage is called 'collective successive apprehension', which 
 consists of 'a speedy perusal of the individual members of the com- 
 plexes with attention'.^ The result is that 'the two successive 
 members of one and the same group are bound together by associa- 
 tions stronger than the associations between successive members of 
 different groups'. This is followed by a third stage, which consists 
 of an 'inner reconstruction' of the earlier apprehended groups. 
 Usually recitation is the chief constituent of the third stage. The 
 subject endeavors to reproduce the material without looking at it, 
 and this leads to the employment of the various kinds of aids that 
 have been previously mentioned. The learner must select the 
 bonds that are requisite to reproduction and exercise them until, 
 once set into operation, they will run their course without external 
 assistance. Of course, during the recitation, references may be 
 made to the text for purposes of prompting as well as for review of 
 material already partly learned. But the 'inner reconstruction' of 
 the material is the important function. Kiihn observed as the most 
 serious deficiency of learning by reading the almost unavoidable 
 tendency to neglect many of the functions which are essential to 
 recall, functions which as a rule can operate only in voluntary 
 recall. He writes:^ "Therefore we come to the conclusion that 
 recitation is better because it leads to a more fundamental, many- 
 sided working-over ('Verarbeiten') of the material." 
 
 The typical learner, we have seen, breaks up the material into 
 smaller groups which are dealt with as units. Similar to the present 
 findings, Kiihn noted that such manipulation of the material was 
 more characteristic of recitation. He states: "By learning with 
 recitation the construction of groups can be carried on more readily 
 than through reading. Many persons say, in fact, that in really 
 pure reading such a construction of groups is impossible. "'' 
 
 1 'Zur Analyse der Gedachtnistatigkeit und des Vorstellungsverlaufes', Zeitschrift filr Psy- 
 chologic, 191 1, Supplementary vol. 5, pp. 253-403. 
 
 6 Ibid., p. 254. 
 
 « 'IJber Einpragung durch Lesen und durch Rezitieren', Zeitschrift fiir Psychologic, 1914, 68, 
 p. 443. ' Ibid., p. 440. 
 
78 RECITATION AS A FACTOR IN MEMORIZING 
 
 The manner in which these groups are built up, the determina- 
 tion of their number, size, and distribution, has already been de- 
 scribed. In general, the nature of the grouping depends upon the 
 kind and length of the material and upon a host of peculiarities 
 which may be found within it. Great differences are also found 
 among different individuals and for the same individual at different 
 times. 
 
 The value of such groupings of the material as an aid in learning 
 has been pointed out by Miiller. They are in brief: 
 
 1. Although it is impossible to grasp in one span of attention a whole list of 
 items, the smaller groups can be utilized as units for attention, thus leading to 
 economy of time and energy in apprehending the whole group. 
 
 2. The factor of localization comes more eflfectively into play. One cannot 
 remember the positions of each member of a series of twenty nonsense syllables, 
 but he can remember the position of four or five groups, each being treated as 
 a unit. 
 
 3. Each group comes to have its own individuality and thus serves as a center 
 of attack. 
 
 4. Groupings assist rhythmical and melodic perusal. 
 
 5. Groups as such are more interesting than a series of single items which the 
 learner soon becomes familiar with, as such, and then permits attention to flag. 
 The groups, as interesting problems to be mastered, arouse and direct attention. 
 
 When the series is quite long, it is not enough that the individual 
 groups should be mastered, but the series of groups must be bound 
 together by additional associations. Sometimes the localization of 
 the groups in visual space or numerically is sufficient, but very 
 often other associative or mnemonic aids are employed. 
 
 Our previous analysis of the learning process would fit very 
 nicely into Miiller's scheme of three stages. That the reading 
 method should be employed to some extent in the beginning has 
 been pointed out by Miiller — in fact, the first two stages are entirely 
 dependent upon reading. The third stage of 'inner reconstruction' 
 is, as its name implies, primarily a stage of attempted recitation. 
 
 To limit the learner entirely to the reading method precludes 
 the possibilities of the active stage of 'inner reconstruction' and 
 thus greatly hampers the learning. The natural tendency of the 
 learner to resort to this latter method of study is shown in the oft 
 repeated statement that the desire to do so was 'well nigh irresist- 
 able' and the like. Most subjects can, to varied degrees, continue 
 to learn by reading, but there are some, perhaps, who can advance 
 only to a limited extent. Kiihn found, ^ in fact, that after a certain 
 number, additional readings may prove not only to be of no value 
 for imprinting, but may be positively harmful. For example, one 
 subject (Got.) required after 
 
 »0p. cit., p. 477. 
 
RECITATION AS A FACTOR IN MEMORIZING 79 
 
 40 readings, 17 additional recitations to learn 
 
 25 readings, 9 additional recitations to learn 
 
 12 readings, 6 additional recitations to learn 
 
 2 readings, 5 additional recitations to learn 
 
 Similar results were found for three other subjects. Such tests, 
 however, have been tried with several subjects in the present study, 
 but in no case were such negative results found, although two sub- 
 jects were found who were unable by reading alone to completely 
 master a long series of nonsense syllables. 
 
 HELPS MORE CONS TA NT AND MORE NATURAL IN RECITATION 
 
 It was pointed out earlier that recitation leads more successfully 
 to the empl o yment of various sorts of aids, such as modulations of 
 the voic e, rhythms, pauses, meaningful associ ations, and the like. 
 An additional point very often reported is that such aids not only 
 ^ome into play more rea dily in recitation but that they are more 
 consta nt- During reading, some report that they emphasize now 
 one syllable, now another; they now use one rhythm, later another; 
 the sight of the word suggests now one association, later another. 
 In recitation, when once adopted, the aids are more constant. This 
 is partly due to the fact that most learners do not like to refer to 
 the text unless it is absolutely necessary, and since recall is entirely 
 dependent upon the use of some association, a connection once 
 initiated is likely to be invariably employed. During reading, 
 since the syllable in each case is present to the eye, the previously 
 observed association, being less essential, is not so deeply impressed; 
 other connections, depending upon the attitude of the subject at 
 the moment, are likely to overrule it with the result that a new 
 association is substituted. This, in essence, is what many report: 
 "It is hard to keep my mind on the work in reading. Different 
 influences seem to come in continually that give the material a new 
 look. First a syllable means one thing and later I associated it 
 with something else." Subject Tr says: "I first thought oi fab as 
 part of fable, wab as Weber, etc., but it was often difficult to remem- 
 ber some of them because I didn't have to depend upon them." A 
 similar situation was found in the case of many subjects by Kiihn, 
 who concluded:^ "The helps in recitation seem to be more natural, 
 while in reading they appear manifold and artificial." 
 
 TESTING THE LEARNING 
 
 In an earlier section , evidence was fou nd that there was a greate r 
 
 , certaint y as t o wha t was known when recitation was emploved in 
 
 the learning. This, of course, is not only true at the completion of 
 
 ' op. cit., p. 440. 
 
8o RECITATION AS A FACTOR IN MEMORIZING 
 
 the learning but during the various stages. In addition to the ob- 
 jective evidence already presented (p. 71 f.) are the reports of manj'^ 
 subjects that when they read only, they are not at all certain how 
 much of the material is known or how well it is known. They may 
 have a general feeling that they can recite a certain part of the 
 material, but they cannot be sure until they have tried. The 
 recitation, of course, constitutes the test. 
 
 One of the values of recitation is that it gives exact knowledge of 
 the results that are being produced and serves to throw into relief 
 the efficacy of the different aids that are being employed as a means 
 to learning the lesson. Recitation leads more surely to the selection 
 and repetition of the desirable bonds and to the elimination of the 
 unfit. In other types of learning, Judd has shown that knowledge 
 of results of practice is essential to improvement. ^° He found that 
 practice in locating the continuation of sloped lines, part of which 
 was concealed from the subject, produced no improvement when 
 the results of the practice were not disclosed, but improvement 
 immediately resulted when the subject was permitted to view 
 briefly the results of his efforts. 
 
 In an earlier section (see p. 71 f.) it was found that recitation 
 leads not only to better localization of the syllables but it also 
 leads to a more accurate knowledge of the correctness of the posi- 
 tion of syllables. In learning by recitation, out of 127 judgments 
 of 'correct in position' but two were wrong, while in the reading 
 series nine out of sixty-eight such judgments were wrong. 
 
 In order to find if there is a greater certainty with respect to the 
 form of the syllables without regard to their position, another 
 similar experiment was made. Eleven graduate students acted as 
 subjects in two tests of five minutes each, one by the reading method 
 and one by the recitation method. Half of the subjects took the 
 former and half the latter test first, the order being reversed for 
 the second test. Table XXXIX gives the results. 
 
 In the first place, a greater number of syllables are written down 
 after the recitation test than after the reading test (twelve and two- 
 tenths as compared to eight). The absolute number judged correct 
 in the recitation series is about twice the number so judged in the 
 reading series, eleven and three-tenths as compared to six syllables. 
 Likewise, the number of syllables that were actually correct was 
 about twice as great for the method including recitation, eleven and 
 four-tenths as compared to five and fifty-four one-hundredths. Of 
 the total number of syllables written down in the reading series 
 seventy-five per cent, were judged to be correct, while in the recita- 
 
 •" 'Practice without Knowledge of Results', Psychology Review Monographs, 1905, 7, pp. 
 185-198. 
 
RECITATION AS A FACTOR IN MEMORIZING 
 
 8i 
 
 tion series ninety-two and six- tenths per cent, were judged to be 
 correct. That is to say, there was a greater assurance of correctness 
 when the learning involved recitation. Moreover, in the recitation 
 tests, of those written down ninety-three and four-tenths per cent, 
 were actually correct as compared to sixty-nine and two-tenths per 
 cent, for the reading series, indicating again that there is less 
 certainty about the knowledge of results during reading. It should 
 be noted that in the reading series there is a considerable discrep- 
 
 TABLE XXXIX 
 
 Results given in number of syllables correct in form without regard to position 
 
 
 After 
 
 five minutes reading 
 
 After five minutes recitation 
 
 Subject 
 
 Number 
 written 
 
 Number 
 judged 
 
 Number 
 
 actually 
 
 Number 
 written 
 
 Number 
 judged 
 
 Number 
 actually 
 
 
 
 correct 
 
 corre.ct 
 
 
 correct 
 
 correct 
 
 At. 
 
 13 
 
 13 
 
 II 
 
 16 
 
 15 
 
 145 
 
 Rs. 
 
 8 
 
 8 
 
 6 
 
 16 
 
 16 
 
 16 
 
 Py. 
 
 9 
 
 7 
 
 7 
 
 10 
 
 9 
 
 9-5 
 
 Tr. 
 
 7 
 
 5 
 
 5 
 
 15 
 
 14 
 
 14 
 
 Gl. 
 
 8 
 
 4 
 
 5 
 
 10 
 
 10 
 
 10 
 
 Mn. 
 
 II 
 
 6 
 
 7 
 
 12 
 
 10 
 
 10 
 
 An. 
 
 8 
 
 4 
 
 5 
 
 12 
 
 II 
 
 II 
 
 E. M. 
 
 5 
 
 4 
 
 4 
 
 10 
 
 9 
 
 9 
 
 J. M. 
 
 7 
 
 6 
 
 4 
 
 II 
 
 9 
 
 10 
 
 Sn. 
 
 7 
 
 5 
 
 4 
 
 16 
 
 16 
 
 16 
 
 Sa. 
 
 5 
 
 4 
 
 3 
 
 7 
 
 6 
 
 6 
 
 Average 
 
 S.o 
 
 6.0 
 
 5-54 
 
 12.2 
 
 II-3 
 
 II. 4 
 
 P. E. 
 
 1-4 
 
 1-5 
 
 1-3 
 
 2.0 
 
 2.3 
 
 2.2 
 
 Per cent, of number 
 
 
 
 
 
 
 
 written 
 
 
 75-0 
 
 69.2 
 
 
 92.6 
 
 93-4 
 
 Per cent, of number 
 
 
 
 
 
 
 
 judged correct 
 
 
 
 92.3 
 
 
 
 100. 
 
 ancy between the number of syllables 'judged correct' and the 
 number 'actually correct'; while for the recitation method these 
 two figures are almost identical. This means that after you have 
 studied a lesson by the recitation method you are practically certain 
 how well you know it, but after you have studied by reading you 
 are not only uncertain about your knowledge but your honest 
 opinion is likely to be an overestimation of your attainment. A 
 closer examination of the table, however, will reveal the fact that 
 individuals differ in this respect. Under the reading method, three 
 people correctly estimate their knowledge {i. e., the number of 
 syllables 'judged correct' equals the number 'actually correct') ; 
 
82 
 
 RECITATION AS A FACTOR IN MEMORIZING 
 
 five people overestimate their knowledge; while three underestimate 
 their knowledge. Confronted by this general uncertainty of results, 
 some subjects are likely to be very conservative in their judgments 
 and others much less so, the general result being an overestimation 
 of attainment. 
 
 In the case of the recitation series, eight subjects correctly esti- 
 mate their knowledge, while one overestimates and two under- 
 estimate their knowledge. The sum of the differences between the 
 number 'actually correct' and the number 'judged correct' is two 
 syllables for the recitation series and eleven for the reading series. 
 
 On the whole then, learning by reading makes it very difficult 
 to estimate one's attainment, while learning through recitation 
 leads to very accurate knowledge of results. This should be thought 
 of in connection with the fact that in our tests the amount learned 
 by recitation is about twice as great, a fact which can only emphasize 
 the greater accuracy in that case. Other things being equal , we should 
 expect twice as many errors of judgment in the recitation results. 
 
 Some evidence can be obtained from the children's data to 
 indicate a similar result. From the data of several classes was 
 computed the total number of syllables written down, and the 
 total number of syllables that were correct in form. From the 
 various methods of study including recitation, certain ones were 
 chosen in order to make practice effects, etc., balance up with the 
 reading series. The following is a sample result, based on forty 
 pupils of the sixth grade. 
 
 Number of syllables 
 Relative number 
 
 Reading 
 Written down Correct 
 
 7. 12 
 1 00.0 
 
 4-5 
 
 63.2 
 
 Recitation 
 Written down Correct 
 
 II . 12 
 
 100. 
 
 9.06 
 51.4 
 
 The results show a clear superiority in favor of the recitation 
 method of learning. 
 
 A similar result was found with sense material, a sample of 
 which follows, showing in the case of thirty-nine eighth-grade 
 pupils the number of details of facts written and the number 
 correct. 
 
 
 Reading 
 Written down Correct 
 
 Recitation 
 Written down Correct 
 
 Number of facts 
 Relative number 
 
 27.2 
 100. 
 
 22.6 
 83.1 
 
 28.7 
 100. 
 
 26.1 
 90.9 
 
RECITATION AS A FACTOR IN MEMORIZING 83 
 
 Objective data, thus, support the introspective opinion, previ- 
 ously given, that one has better knowledge of results in learning by 
 recitation and that this is an aid in learning. Some of the concrete 
 ways in which this knowledge may be of assistance may be briefly 
 considered. 
 
 First: There is a feeling of satisfyingness in the certainty of 
 progress, in knowing that headway is actually being made. Con- 
 versely, it is annoying to be uncertain whether the study is bring- 
 ing returns. The satisfyingness results in better attention and 
 better application to the work, while annoyingness is distracting 
 and hampers learning. Subject An gives a typical report: "It 
 [reading] was discouraging because I did not feel that I was making 
 much progress during the last part of it. There was no way to tell." 
 
 Second: A certain saving of energy may result from knowing what 
 parts of the material are known and what are not known, a. Over- 
 learning of certain portions may be prevented. Usually the first 
 and last syllables are first learned and when the subject knows that 
 these are mastered, they can be passed over hastily in subsequent 
 perusals, a very slight amount of review being sufficient to keep 
 them intact. Subject Rs says: "I saved time during recitation by 
 skipping hurriedly over the words I already knew." b. An oppor- 
 tunity is afTorded to direct special attention to those portions that 
 are still unlearned. Subjects report that certain syllables offer 
 special difficulty which is often not suspected until they endeavor 
 to recite, c. The two factors together, easing down on familiar 
 or learned portions and attending more intensely to unfamiliar or 
 especially difficult portions, result in a saving of energy in the long 
 run. It makes the work more absorbing, and also makes possible 
 short periods of relaxation of attention or breathing spells, which 
 may result in a rebound of energy for learning the more obstinate 
 portions. 
 
 ERRONEOUS RECALL 
 
 It is obvious that an attempted recitation may result not only 
 in a failure to recall a certain syllable, but it may also result in 
 erroneous recall, neither of which could occur during reading in the 
 strict sense. If the errors are too numerous or if they are not dis- 
 covered in the case of recall, they become a harmful rather than a 
 beneficial factor in learning. Failures to recall are very frequent in 
 some cases in which the recitation is introduced too early, with the 
 result that time is frequently lost in unfruitful endeavor to recall 
 items that are not as yet sufficiently fixed in mind. Erroneous 
 recalls, under the same conditions, are also frequent, but they sel- 
 dom occur without some feeling or indication of incorrectness. 
 
84 RECITATION AS A FACTOR IN MEMORIZING 
 
 Many subjects report that they profit much by these mistakes. 
 Noting and correcting an error helps to fix the proper item in 
 mind; it receives better attention at that moment and will receive 
 special attention on the next repetition. Subject Rs said as a 
 sample concerning a test in which recitation was begun after five 
 minutes' reading: "Twice I failed on piw, each time saying poy 
 [evidently confused with soy which followed]. But after twice cor- 
 recting it, I had it so well in mind that I will probably remember it 
 longer than any other in the series." 
 
 In connection with the matter of unsuccessful and erroneous 
 recall, Katzaroff pointed out, as an advantage of recitation, a kind 
 of growing satisfyingness in the task. Successful recall is satisfying 
 and failure is annoying. As we proceed, the proportion of satisfy- 
 ingness becomes steadily greater, toning up the learner and enabling 
 him to keep up interest and application in spite of growing fatigue. 
 He states :^^ "The learner is active, he has to seek, he rejoices when 
 he has found and is irritated at the syllables which evade his call. 
 Here crowd sentiments of affection for certain syllables, of antipathy 
 for others, which contribute to enrich the associative bonds and 
 favor conservation and recall." One of the workers in the present 
 study similarly said: "In reading, it is the last part of the test that 
 is most wearisome, but in recitation, it becomes almost a pleasure 
 as I approach a mastery of the whole bunch of words." 
 
 UNINTENTIONAL RECITATION DURING THE READING TESTS 
 
 A great many of the subjects found it difficult to resist their 
 natural tendency to recite, during the reading series; in fact, 
 the reading was nearly always combined with more or less recall 
 of an unintentional, practically unavoidable sort. The effort to 
 avoid reciting acted as a positive disturbance and source of annoy- 
 ance, thus distracting attention and consuming energy to no pur- 
 pose. Subject Md speaking of the reading method said: "Very 
 difficult and disagreeable, because I was constantly inhibiting the 
 tendency to test what I had been trying to learn." 
 
 SATISFYINGNESS AND ANNOYINGNESS IN READING AND RECITATION 
 
 According to the introspections of many adult subjects and the 
 reports of many school children, one conspicuous difference between 
 reading and recitation lies in the greater satisfyingness of the latter. 
 That the matter of satisfyingness and annoyingness of mental 
 work is important has been emphasized by Meumann.^^ "Xhe emo- 
 tional condition in which we find ourselves during the performance 
 
 » op. cit., p. 257. 
 
 '2 The Psychology of Learning, p. 281. 
 
RECITATION AS A FACTOR IN MEMORIZING 
 
 85 
 
 of a mental task is of profound importance for the accomplishment 
 of the task. In general, it may be said that an emotion of pleasant- 
 ness facilitates the function of memory, and that unpleasantness 
 has a very detrimental effect upon memory." Thorndike is more 
 cautious :i^ "No one probably doubts that interest in the exercise 
 of a function favors improvement at it," and "such statements 
 appeal to our common sense as probably true, though they have 
 not been fully verified." 
 
 It shall be our purpose, first, to inquire as to what differences 
 appear between recitation and reading as producers of satisfying- 
 ness and annoyingness, and then to consider briefly in what way or 
 by means of what minor functions these effects are brought about. 
 
 That there is greater satisfyingness in studying by the recitation 
 method is indicated by the witness of nearly every subject, child or 
 adult. At the close of the experiments with the school children 
 they were asked to state what method of learning they liked best. 
 For ease of selection the cases considered were three: one in which 
 they read all the time, one in which they read about half of the 
 time, and another in which they recited nearly all the time. The 
 following table gives the distribution of opinion. 
 
 With nonsense material 
 
 
 All reading 
 
 Half and half 
 
 Mostly recitation 
 
 Grade eight 
 Grade six 
 
 3 
 2 
 
 2 
 10 
 
 29 
 
 27 
 
 Sense material 
 
 Grade eight 
 Grade six 
 
 20 
 
 28 
 
 It is clear that the children strongly preferred the methods in 
 which recitation was included. 
 
 The reasons for their preferences are varied and not very specific. 
 Such statements as, "It isn't such hard work," "I learn better that 
 way," were common. Some explained their preferences as follows: 
 "I knew I was learning them when I recited"; "I get so tired when 
 I read"; "When I recite, it's fun to see if I can say more every time 
 than I ever did before." 
 
 The introspective accounts of adults are even more emphatic. 
 Among the subjects listed in Table XXXIII, fourteen reported that 
 
 ^> Educational Psychology, vol. II, p. 219. 
 
86 RECITATION AS A FACTOR IN MEMORIZING 
 
 Method Three (two minutes reading and eight minutes recitation) 
 was most satisfying, one that Method Two (half and half) was 
 most natural and satisfying, and all reported that Method One (all 
 reading) was least so. In fact, most of them report that the last 
 four or five minutes in the reading test were positively 'annoying', 
 'monotonous', 'tiresome', 'very fatiguing', etc. Subject Hn 
 declared: "Without a doubt, trying to learn a series of nonsense 
 syllables in this way is the most monotonous work I have ever done. 
 The syllables came to have absolutely no connection or association, 
 and the typewritten letters became, after four or five minutes, so 
 many stupid hieroglyphics." 
 
 Many of the actual ways through which reading becomes annoy- 
 ing and recitation satisfying have already been indicated, and they 
 will receive but brief mention here. a. There is satisfaction in 
 the realization that progress is actually being made. We have seen 
 earlier that this is the case during recitation. Conversely, it is 
 annoying to be uncertain of one's progress in the learning, b. Reci- 
 tation is satisfying because it offers the learner more freedom 
 to employ such aids, and work with such methods, as he may 
 desire. Reading becomes annoying because it hinders or prohibits 
 the exercise of many of the desired functions, c. The facts of a 
 and b taken together explain other sources of satisfyingness in reci- 
 tation. For example, it is satisfying, as Katzaroff pointed out, to 
 attack portions of the lesson that offer special difficulties — difficul- 
 ties that are often not realized until one begins to recite. Again, 
 the opportunity that recitation affords the learner to ease off on 
 familiar portions, and strike hard at difficult portions, seems to be 
 a good remedy for boredom and fatigue, d. Annoyingness attends 
 the constant effort exerted by many in resisting the natural 
 tendency to recite during the reading series. 
 
 IS THERE GREATER ACTIVITY IN RECITATION THAN IN READING? 
 
 The early investigators on this subject gave great emphasis to 
 the conclusion that recitation, as compared to learning by reading, 
 produced a greater activity on part of the learner, and to this greater 
 expenditure of energy was attributed in large measure the better 
 results obtained. For example Katzaroff says:i* "In the readings, 
 the subject is passive, calm, indifferent; in recitation he is active." 
 
 The introspections and observations from the present work do 
 not lead to exactly this conclusion. The distinction seems to be 
 one of kind rather than one of quantity. It appears that recitation 
 does not always, in fact, does not generally result in greater activity, 
 effort, or expenditure of energy on part of the learner, but the indica- 
 
 " op. cU., p. 257. 
 
RECITATION AS A FACTOR IN MEMORIZING 87 
 
 tions are that the energy is expended in a different way. Certainly 
 the conscientious learner by the reading method is not 'calm and 
 indifferent'. An apparent, but certainly not a real lack of activity 
 is indicated by the already mentioned fact that many of the motor 
 functions such as articulation, accentuation, the use of rhythm, etc., 
 are much less prominent in reading. In the next section, however, 
 it will be seen that the subjects declare that every internal symp- 
 tom indicates that reading is more consuming of energy than reci- 
 tation. 
 
 FATIGUE EFFECTS OF READING AND RECITATION 
 
 Other things being equal, we should expect, were it true that 
 recitation results in greater activity and expenditure of energy than 
 reading, that it would also be more fatiguing. Unfortunately, in 
 this study, we have no indisputable measure of fatigue, but it never- 
 theless appears, in so far as one is able by a subjective judgment to 
 estimate fatigue, that recitation is much less fatiguing. The findings 
 reported in the section on the satisfyingness and annoyingness 
 of the two methods bear strongly on this point. In so far as feelings 
 of fatigue, boredom, monotony, and the like are indices of real 
 fatigue, there can be no doubt that recitation is less fatiguing than 
 learning by reading. Whether or not they are measures of real 
 fatigue, they are at least very important from the point of view of 
 work in the school-room. In the face of such statements as those 
 following, there can be no doubt that recitation is to be preferred 
 to learning by reading in this respect. Subject Hn: "Reading is 
 the most monotonous work I have ever done." Subject Dn: 
 "Reading is most fatiguing because there is no variation." Subject 
 Fx: "Reading most fatiguing — monotonous — took all my energy 
 to keep up interest." Subject Sn: "This method very tiresome — 
 effort seemed to be fruitless." Subject Mk: "Very tiresome and 
 disagreeable." On the other hand, learning by recitation may, as 
 Subject Py said, "be almost a pleasure," or, as Subject Mk states, 
 "much more satisfying," or, as Subject Rs says, "not so bad as 
 reading, that's certain." 
 
 Subjects also report that the after-effects of learning by reading 
 are greater than learning by recitation. Subject Bn reports: "I 
 couldn't apply myself to work for an hour after the experiment.' 
 Subject T: "I felt tired all the rest of the afternoon." 
 
 From a practical point of view, it should also be considered that 
 the fatigue, based on the amount learned, rather than the time 
 spent, would be relatively very much greater in the case of learning 
 by reading. The subjects report that they are very much less 
 fatigued by ten minutes of study by recitation than by ten minutes 
 
88 RECITATION AS A FACTOR IN MEMORIZING 
 
 study by reading, yet they have learned twice as much. If the study 
 by reading were continued until the amount learned was equal to 
 that learned by the recitation method, the fatiguing effects of the 
 former would doubtless be still more marked. 
 
 SUMMARY AND CONCLUSIONS AS TO THE NATURE OF READING 
 AND RECITATION AS FUNCTIONS IN THE LEARNING PROCESS 
 
 Our analysis of learning has shown the memorization of any 
 m.aterial, especially of nonsense material, to be a complex process 
 involving the formation of a host of bonds. It has appeared, more- 
 over, that many, in fact, most of these bonds can be properly formed 
 only by means of recitation. Our subjects have reported that it 
 was 'dif^cult', 'unnatural', 'annoying', 'fatiguing', or 'impos- 
 sible' to establish most of the essential bonds during reading. 
 Efforts to learn the material by rote, to memorize it mechanically 
 or by means of 'visual imprinting' during a series of readings 
 proved to be futile. Memorization was possible only by means of 
 establishing bonds between items and their pronunciation, sound, 
 or look : between items and accents, pauses, or elements of a rhythm ; 
 between items and their position in a series ; between an item and 
 other items which it may be considered a part of, similar to, or 
 somehow related to and the like, as well as additional bonds between 
 characteristics of successive groups of items. We have found that 
 it is to the formation of just these bonds that recitation leads, and 
 that it is just these functions that it is difficult or impossible to 
 exercise adequately during reading in its pure form. Consequently 
 it seems to be a justifiable conclusion that complete learning is 
 possible only by means of some form of recitation. Pure reading 
 alone will scarcely enable one to completely learn a lesson which 
 exceeds the memory span by any considerable length, yet it serves 
 an important function in the learning process as we have seen. 
 
 The considerations of the present chapter have shown that read- 
 ing and recitation are very broad functions made up of many minor 
 ones. Economical learning consists not only in selecting and exer- 
 cising those more minute functions which are essential and elimi- 
 nating those that are valueless, but also in exercising them in 
 proper sequence and each for an optimum time. It will be neces- 
 sary here to review but briefly some of the essential functions, 
 indicating to which of the two broader functions they belong and 
 the order in which they are customarily exercised. 
 
 The first stage of the learning consists, as we have seen, in looking 
 over the whole material with the purpose of obtaining an idea of 
 its general make-up, noting the individual items in the group, 
 getting the pronunciation or look or sound of the terms to some 
 
RECITATION AS A FACTOR IN MEMORIZING 89 
 
 degree, and noting outstanding 'aids' which may be employed in 
 breaking up the material to further the learning. Much may be 
 done in this stage to determine upon a method of attack. The 
 length of the material, its apparent difficulty, its peculiarities, the 
 possibilities for rhythmical division, and its ready-made associations 
 are considered. The material may be thus perused for several times 
 until the reader feels 'familiar' with it and a method of procedure 
 is tentatively adopted. This stage is the reading stage, includ- 
 ing what Miiller has termed the stages of 'collective' and 'suc- 
 cessive apprehension'. That the functions of reading are 
 essential and satisfying here has been indicated by all intro- 
 spective data. 
 
 The optimum duration of this stage depends upon many factors, 
 such as the length and difficulty of the material, the age, training 
 and capacity of the learner, and the like. These considerations will 
 be taken up on a later page. 
 
 Following this stage, new functions may be introduced and the 
 original functions may be employed in a somewhat different manner. 
 This is the stage of recitation. It consists essentially in the final 
 selection of the bonds requisite to recall and the exercise of these 
 bonds until they are firmly established. What these bonds are, it 
 was the purpose of the preceding sections of this chapter to point 
 out. Thus it appears that memorizing is in no essential way differ- 
 ent from any other form of learning. The bonds selected are 
 exercised, those found to be unfit are eliminated, and new bonds 
 are added as the case demands, the period of practice being con- 
 tinued until, once initiated, the series of desired responses runs ofif 
 in the proper order. 
 
 Like other processes of learning memorizing may be explained in 
 physiological terms. An adequate explanation of this sort would 
 make the difference between the functions of reading and recitation 
 more intelligible. The learning of a series of nonsense syllables, 
 like the formation of any habit, involves two things: a sensori- 
 motor response or the formation of a bond between a situation and 
 a particular response, and a sequential connection between the 
 various situation-response bonds in serial order.^^ The following 
 diagrams illustrate in a very rough way, what physiological actioHS 
 and changes are involved in the learning of a series of nonsense 
 syllables or any other material. 
 
 '5 A standard treatise in English upon the physiological aspects of learning is Ladd and 
 Woodworth, Physiological Psychology, New York, 1911. For an abbreviated but excellent account, 
 see also Thorndike, E. L., Educational Psychology, New York 1913, vol. i, chapter XIV. The 
 illustrations used in the present article are similar in some respects to those employed by Bair to 
 explain the development in skill in typewriting. Compare 'The Practice Curve', Psychological 
 Revie'iV Monographs 1902, No. 19, pp. 1-70. 
 
90 
 
 RECITATION AS A FACTOR IN MEMORIZING 
 
 In Figures 4, 5, and 6, S, S \ S -, — represent the stimuli, that is, 
 the sight of the syllables, occurring in serial order as they would in 
 reading. These stimuli are conducted to the sensory centers A, A^ 
 A^,— , which discharge respectively into the higher centers B, B \ 
 B 2 — , and these cells in turn discharge into M, M ^ M ^ — , the 
 effectors which produce the motor responses of writing or speaking 
 the syllables. 
 
 Let us consider a case of pure reading; p^^re in the sense of being 
 entirely devoid of all elements of recall, waiving for the moment 
 
 /V' ./f^ j.f»^ 
 
 5' Q- i^ 
 
 Fig. 4 — 'Pure' reading. 
 
 Fig. 5 — Reading with formation 
 of associations. 
 
 [,/ir-^ 1 n^ 
 
 Fig. 6 — Recitation. 
 
 the question of whether such reading actually exists. Pttre reading 
 would consist in the exercise of bonds S A B M, S ^A^B ^M \ etc., 
 as distinct units. The more often these bonds are exercised, the 
 more definite becomes the connections and the more automatic 
 the response. But it is obvious that however firmly these bonds 
 become fixed, they cannot of themselves make possible voluntary 
 recall, since S, S \ S ^, — , (the sight of the syllables being learned) 
 is an essential link in the process. 
 
 Figure 6 illustrates roughly the requirements for voluntary re- 
 call. In this case the expression of the syllables (designated as 
 M, M', M-, — ,) are produced in the absence of the stimuli, S, S^ 
 S^, — , of the visible words. What is required here is that bonds 
 should have been formed between the various higher units. Con- 
 nections between B and B\ M and B\ or both, must be established. 
 The result is that once the series is started, the physiological pro- 
 
RECITATION AS A FACTOR IN MEMORIZING 9 1 
 
 cesses which produce the first syllable act as the stimuli for the pro- 
 duction of the processes which bring about the response of the 
 second syllable and so on. 
 
 Perhaps few would doubt that this illustration would account, in 
 a very rough way, for the process of recall, but many might be 
 unwilling to admit that Figure 4 is a correct representation of the 
 processes involved in reading. The doctrine of association by con- 
 tiguity might insist that the mere repetition of the syllables one 
 after another would result in the establishment of bonds between 
 them. Bair,^^ consequent to his study of the development of skill 
 in typewriting, concluded, although he really gave very little weight 
 to it, that "connections are formed between cells that for a number 
 of times have been stimulated or discharged — in succession," 
 Woodworth has pointed out the inadequacy of this doctrine. To 
 quote :^^ "contiguity is a necessary condition of association. But 
 is it a sufficient condition? There is little in the experimental work 
 on memory to indicate that it is sufficient, and much to indicate 
 that it is not usually depended on to accomplish results. The 
 things to be connected must be together, in order to arouse the 
 reaction connecting them; but, unless they arouse some such 
 reaction, they do not become connected, except it be very weakly." 
 Professor Woodworth has shown some convincing experimental 
 evidence ^^ in support of his view and doubtless much more could 
 be discovered by search through studies already in print, ^^ but 
 space will not permit us to go into the matter here. 
 
 In attempting to learn by reading, the subject does not rely 
 entirely upon mere repetition of the syllables — upon the alleged 
 efficacy of contiguity — alone, but in most cases, tries to form the 
 serial associations upon which he must rely to recall the series 
 when the time comes. Why are these bonds not definitely formed? 
 The reason is that the presence of the printed words (5, S\ 5-, etc.) 
 makes it so unessential, during reading, to connect B with B^, or 
 M with B^, that the learner's purpose to strengthen these bonds is 
 defeated. Since, according to prescription, the learner must, 
 
 i^Op. cit., p. 51- 
 
 1' 'A Revision of Imageless Tliought', Psychological Review, 1915, 22, pp. 1-27, especially 
 pp. 16-22. 
 
 '8 It may be well to give the following sample test, in the words of the author. "I read a list 
 of twenty pairs of unrelated words to a group of sixteen subjects, instructing them beforehand to 
 learn the pairs so as to be able to respond with the second of each pair when the first should be given 
 as a stimulus. But, after reading the list three times, I told them that they should, if possible, 
 give also the first word of the following pair on getting the second word of the preceding pair as a 
 stimulus. — The results were most definite: the second members of the pairs were correctly recalled 
 in seventy-four per cent, of all cases, but the first members were recalled in only seven per cent, 
 of the cases." 'A Revision of Imageless Thought'. Ibid., p. 18. 
 
 19 For example, see Hollingworth, H. L. 'Characteristic Differences between Recall and Recog- 
 nition', American Journal Psychology, 1913. 24, pp. 532-544- Also 'The Influence of Caffein on 
 Efficiency', Archives of Psychology, 1912, No. 22, p. 17. 
 
92 RECITATION AS A FACTOR IN MEMORIZING 
 
 on completing the series S A B M, then read S\ the connections 
 S^ A^ B^ M^ being by previous practice better estabhshed, are 
 thrown into action before the incipient bonds B-B^, M-B^, are 
 awakened. The response follows directly upon the stimulus pro- 
 voked by seeing the word. 
 
 How is it possible, then, as shown by our quantitative results 
 given earlier, that some memorization does result from reading? 
 In all probability pure reading is a fiction; recall, to some degree, 
 being always present. Nearly all subjects were able, introspectively, 
 to discern this fact. Figure 5 shows, roughly, the physiology of 
 this situation. The dotted lines S-A, S^-A^, etc., indicate that 
 these bonds between the sight of the word and its expression are 
 more feebly exercised; are less depended upon than is the case in 
 Figure 4, which illustrates the hypothetical pure reading. The 
 manner in which the items (S^, etc.) are required to play a minor 
 role are various. Sometimes the subject pauses between the 
 series, S A B M and S^ A^ B^ M\ etc., thus permitting the bonds 
 B-B\ M-B^, etc., to be thrown into action before S^ is observed. 
 That is to say, the subject anticipates the next word in the series, 
 more or less, before he reads it. Sometimes the syllables are read 
 in a hazy, inattentive way, in which case the subject relies partly 
 upon the exercise of the serial bonds as well as upon the objective 
 stimulus of the printed word. In these and other ways, actual 
 reading departs from pure reading and in consequence leads more 
 effectively to memorization. In short, the actual reading which 
 the subject practices is a sort of hybrid between the hypothetical 
 pure reading and recitation. 
 
 A more accurate picture of the anatomical substrata of memoriz- 
 ing would undoubtedly be much more complex than our simple dia- 
 grams. As we have seen, consciousness of the meaning and form 
 of the material is a prominent factor in learning. Consequently, 
 the diagram should contain at least a symbolic representation of 
 the centers upon which, presumably, such consciousness depends. 
 Thus, in Figure 7, P, a 'psychic' center may be added, in which 
 elaborations of the sensory data take place.-" Probably in learning 
 a passage, as well as during the recall for some time, P is called into 
 activity, discharging into B. As practice continues, in all likeli- 
 hood, pathway S A B M becomes relatively more and more per- 
 meable, until finally conduction through A P B ceases almost 
 entirely — the process becomes practically unattended by conscious- 
 ness of meaning. To illustrate this condition for reading (illus- 
 
 2" For an account of such distribution of functions, see Ladd and Woodworth, Physiological 
 Psychology, chapters IX and X. The 'association' or 'psychic' areas are given special treatment 
 on pp. 251-263. 
 
RECITATION AS A FACTOR IN MEMORIZING 93 
 
 trated by Figure 7) we may repeat the statement of one of our 
 subjects: "The typewritten letters became, after four or five min- 
 utes, so many stupid hieroglyphics." 
 
 In recitation also, the 'psychic' center P is involved. Figure 8 
 is micrely a suggestion of the possible connections of this center 
 with others. P may have connections with B^, as well as with P^, 
 which in turn, is connected with B^ through which the discharge 
 into the organ of expression takes place. If now we consider that 
 each of these possible connections indicated by a straight line in 
 the figure is a representation of hundreds, perhaps thousands of 
 different neurones that may be employed, the complexity of the 
 
 ^c^-:__^^^._.^^-._.>3 
 
 Fig. 7 — Reading with awareness Fig. 8 — Recitation with awareness 
 
 of meaning. of meaning. 
 
 neural substrate involved in learning is suggested. But recitation 
 of the series of syllables may become short-circuited to a nearly 
 mechanical activity, such that once initiated, the series of responses 
 occurs automatically while attention is occupied with other matters. 
 This may be typified by considering that the connections through 
 P and P^ etc., and perhaps even the connections B-B^, etc., drop 
 out, so that the connection of M with B^, which leads directly to 
 M^, is so close that once the series of responses is started, each 
 follows its predecessor with mechanical precision. 
 
 Recitation, in brief, differs from reading physiologically by the 
 fact that it selects and exercises the bonds upon which the estab- 
 lished habit depends, while reading calls into action some bonds 
 that are not strictly needed for recall, omits some that are requisite, 
 and does not so well exercise the remaining few, needed for recall. 
 Recitation is for memorizing what practice is for other habits.^' 
 The physiological basis is the same. 
 
 21 Such a physiological explanation, for example, has been worked out in detail by J. H. Bair 
 for typewriting. See 'The Practice Curve', Psychological Review Monographs, 1902, No. 19, 
 pp. 1-70. 
 
94 RECITATION AS A FACTOR IN MEMORIZING 
 
 Some of the differences between reading and recitation, which 
 appeared from the introspective analysis of the two functions, may 
 appear with more clarity when considered from the physiological 
 side. For example, many subjects reported that the associative 
 aids adopted during recitation were more constantly employed than 
 those adopted during reading. Recitation results in the continued 
 exercise of particular bonds, as we have seen, and of course each 
 repetition strengthens those bonds, with the result that the nervous 
 impulse once initiated flows along the most frequently traversed 
 pathway. In reading, none of the serial bonds receives adequate 
 exercise, with the result that none has a great advantage over any 
 other, and now one, now another pathway may be traversed. 
 
 Annoyingness and fatigue in the case of prolonged reading may 
 be considered as largely due to a check placed in the way of the 
 exercise of the bonds desired. "When any conduction unit is in 
 readiness to conduct, for it to do so is satisfying. When any con- 
 duction unit is not ready to conduct, for it to do so is annoying. 
 When any conduction unit is in readiness to conduct, for it not 
 to do so is annoying. "22 After the preliminary exercise of the 
 conduction units SAB M, further exercise of that bond becomes 
 annoying; the serial bonds are then ready to conduct. For them 
 to do so, under the conditions specified in reading, is practically 
 impossible, because the stimuli S, S^, S'\ etc., by virtue of their 
 firmer establishment, cause the conduction to take the habitual 
 course, S A B M, etc. 
 
 Space will not permit further illustrations of this sort. By way 
 of summary of this section, it is only necessary to repeat that read- 
 ing and recitation are relatively distinct yet essential functions of 
 the learning process. Each has its proper place, and as we have seen, 
 introducing recitation too early or withholding it too long retards 
 learning. The important matter is to determine the optimum point 
 at which to introduce it, a matter which will receive consideration 
 in the following section. 
 
 THE OPTIMUM TIME AT WHICH TO INTRODUCE RECITATION 
 
 The quantitative results presented in Chapter IV indicated that 
 the optimum time at which to introduce recitation varied somewhat 
 according to the age and training of the different groups of school 
 children, and the data obtained from adults pointed to differences 
 among individuals of approximately the same age and training. It 
 is obvious that the determination of the optimum time at which to 
 introduce recitation is a most important matter for purposes of 
 economizing time and energy in learning. That the quantitative 
 
 '2 Thorndike, E. L., Educational Psychology, vol. II, pp. 1-2. 
 
RECITATION AS A FACTOR IN MEMORIZING 95 
 
 determination of the best combinations of reading and recitation 
 made in the present study apply only to the particular conditions 
 here employed, has already been pointed out. The optimum time 
 for the beginning of recitation will doubtless vary not only according 
 to the age, training, and capacity of the learner but also according 
 to the kind of material, the length of the lesson, and the purpose in 
 view, i. e., whether the lesson is to be learned verbatim, whether the 
 substance without the exact form is to be reproduced, or whether a 
 less definite mastery is all that is desired. Consequently, a quanti- 
 tatively precise rule cannot be made. 
 
 The present study, however, has produced some results that are 
 suggestive. In general, all the evidence, quantitative, introspective, 
 and interpretative seems to imply that recitation should be intro- 
 duced early. Only a very small percentage of the total time required 
 to learn should be devoted to reading. However, it seems to be a 
 natural tendency of many adult subjects to make too early an 
 attempt at recitation. Some of the reasons for the disadvantageous 
 effect of introducing recitation may be pointed out. First, The 
 bonds between the words and syllables and their correct pronuncia- 
 tion are not sufficiently well formed to permit successful recitation. 
 Second, The advantageous effect of a preliminary determination of 
 a line of attack is foregone in whole or part. Third, The amount 
 of data that can be recalled at so early a moment is insufficient. 
 The learner is likely to waste time in fruitless endeavor to recall 
 syllables that are simply not as yet forthcoming. Fourth, Too much 
 time is wasted looking on and off the text, 'finding the place' and 
 the like. Fifth, Too frequent failures in attempted recalls break 
 attention and may develop an unpleasant attitude on part of the 
 subject. Sixth, Too many erroneous recalls may be made. When 
 the learner has such slight acquaintance with the material as a 
 whole, errors once made are likely to be repeated. Later these 
 undesirable bonds must be broken down before the correct bonds 
 can be formed. 
 
 Just as introducing the recitation too early has a deleterious 
 effect, so does introducing it too late retard learning. The abundant 
 quantitative evidence for this has been presented in Chapter IV. 
 
 The optimum combination of the two functions can be best 
 expressed in this way. Reading should be continued until the 
 learner is fairly familiar with the material as a whole and with the 
 items of which it is composed. The learner should have decided 
 meanwhile upon his general method of attack. Enough of the 
 material should be clearly in mind so that the learner's first attempts 
 at recall will meet with some success. Just how much is enough 
 will depend largely upon the learner. As a guiding principle one 
 
96 RECITATION AS A FACTOR IN MEMORIZING 
 
 may consider that the first few recitations should not result in too 
 great a distortion of the material, nor should it cause a waste of 
 time in fruitless endeavor to recall. The capacity of the learner to 
 quickly judge the status of his knowledge is of prime importance; 
 he should be able to know at once whether continued effort to recall 
 this particular syllable will end in success or not, and in the latter 
 case time should not be wasted before reference to the text is made. 
 The early stages of learning will thus employ both reading and reci- 
 tation, the relative amount of the former decreasing as the learning 
 progresses. Economical learning would consist, in part, in employ- 
 ing recitation, after it is once introduced, to the full, coupled with 
 the capacity to speedily resort to reading where it is essential. 
 
 THE EFFECTIVENESS OF RECITATION IN LEARNING NONSENSE 
 MATERIAL AS COMPARED TO LEARNING SENSE MATERIAL 
 
 The quantitative results of Chapter IV seemed to indicate two 
 things: First, that the optimum time for introducing recitation was 
 considerably earlier for nonsense than for sense material; and 
 Second, that recitation seemed to be a more fruitful method of 
 study in the case of nonsense material than in the case of sense 
 material. 
 
 The first result is apparent rather than real. While it is true that 
 recitation introduced very early produced richer returns for non- 
 sense than for sense material, this should be considered in connec- 
 tion with the fact that the amount of material forming the lesson in 
 the former case is but a small fraction of that used in the latter. Yet 
 the amount of material should be considered only in connection 
 with the difficulty of the material. While the nonsense material 
 was much less in amount, it was very much more difficult to learn. 
 A further consideration of this point is unnecessary since the factors 
 which influence the introduction of recitation, just considered, are 
 the same in either case. 
 
 The point with regard to the value of recitation as dependent 
 upon the kind of material is important and demands further 
 consideration. 
 
 The results have shown clearly that equal amounts of recitation 
 produce richer returns in the case of senseless non-connected material 
 than when connected senseful material is used. The reasons why 
 this should be the case have been given in the previous sections of 
 this chapter, and it is only necessary here to summarize the factors 
 upon which this difference depends. 
 
 In the first place, it was found that recitation was of great service 
 in assisting the subject to organize the material into some sort of 
 compact and connected whole, such an organization being essential 
 
RECITATION AS A FACTOR IN MEMORIZING 97 
 
 to a thorough mastery of it. The particular means of accomplishing 
 this organization were: the formation of bonds between the items 
 and accents, modulations of the voice, pauses, and elements of a 
 rhythm; the formation of bonds between items and their meaning, 
 immediate or distant; the noticing of peculiarities in the text and 
 the formation of bonds between items and their position for the 
 purpose of breaking up the material into groups; the noticing of 
 bonds between items and characteristics of the groups and so on. 
 In short, recitation rendered great service in creating usable asso- 
 ciations within the material where there was none, or in more 
 adequately noticing and exercising those that were already present. 
 In nonsense material these bonds between items are absent, and this 
 process of organization and creation of associations is difificult and 
 essential; learning of such material consists in accomplishing just 
 this organization. In the connected sense material such as that used 
 in the present experiment, most of these associations are already 
 present; the material is already organized, the items are connected 
 by serial connections of meaning, rhythms, and the like, by means 
 of which the various elements are firmly knit together. The func- 
 tion of recitation for the formation of these bonds is not required. 
 What is needed is that the ready-formed associations be noticed 
 and exercised, although, in most cases, bonds in addition to those 
 found in the material will be required. 
 
 A second reason for the better results obtained by reading in the 
 case of sense material is closely related to the first and lies in the 
 fact that reading is less 'pure' in studying sense material. As 
 was remarked earlier, after a certain number of perusals the read- 
 ing of either kind of material is probably not pure and becomes 
 less and less so as the subject becomes more familiar with it. The 
 more easily the material can be grasped, the less pure the reading 
 becomes, as a rule. Nonsense material is always rather hard to 
 articulate and hard to work with generally, and as a consequence, 
 there is less of a tendency to depart from reading when it is pre- 
 scribed. But in the case of connected sense material, the reader is 
 usually already familiar with the words and phrases as such; only 
 the combinations are new and doubtless not all of them. The greater 
 fluency and greater familiarity of the material results in combining 
 recitation with reading; only certain key words need be noticed, 
 the gaps being filled in by recall. The learner can glance along the 
 lines, scarcely seeing more than an occasional word which sug- 
 gests the context. 
 
 The physiological explanation that was applied to reading and 
 recitation in general can be equally well utilized to illustrate these 
 points. In recitation the connections between the items (repre- 
 
98 RECITATION AS A FACTOR IN MEMORIZING 
 
 sented in Figures 7 and 8 by P-P^, P-B^, B-B^, etc.) are, in con- 
 siderable degree, already given in the material. In fact, the serial 
 associations between the words of familiar phrases are already 
 fixed in one's nervous system through earlier practice. Recitation, 
 as a factor making possible the formation of many connections, is 
 consequently not needed. In other cases where the connections 
 are less definitely formed, only a small amount of practice is re- 
 quired to stamp them in. The result is that in so far as the con- 
 nections are ready-formed, reading amounts in all essentials to 
 recitation. The eye neglects many of the words as such, fixating 
 only occasional points. Reading thus becomes far from pure and 
 approaches recitation, in all likelihood, more and more closely as 
 the learning advances. 
 
VI 
 
 CONCLUSIONS AND PEDAGOGICAL IMPLICATIONS 
 
 A detailed summary of results will not be attempted at this 
 point. Only a few of the results which are of practical importance 
 for the work of the school-room will be repeated. The reader who 
 wishes a more detailed account of the findings may refer to the 
 summaries that are to be found at the close of the previous chapters. 
 
 The compilation of quantitative and introspective evidence has 
 shown that reading and recitation are relatively distinct functions 
 jn the process of learning. Each has its proper office to perform, 
 and to restrict the lea rning entirely to one or t he oth er results in 
 
 loss of time and energ y. Reading, as the introductory function, 
 
 should be employed until the learner is fai rl y well acquainted with 
 the mate rial as a whole ; until a method for further attack has been 
 tentatively adopted; and until the first attempt at recall will 
 meet with some success without too great a distortion of the ma- 
 terial. The optimum point for introducing it, thus, occurs early in 
 the process, but to introduce it too early, as well as to introduce it 
 too late, will have a detrimental effect. In determining the exact 
 moment at which recitation can best be introduced, one must take 
 into account the length of the lesson, the difficulty of the material, 
 the kind of learning that is desired, the age, training, and general 
 capacity of the learner. 
 
 The function of recitation, as we have seen, is similar to that of 
 practice in any form of sensori-motor learning. Memorization 
 consists in selecting certain essential bonds, eliminating the unfit, 
 and exercising the former until the connections are so well formed 
 that once initiated, the series of responses will occur in proper 
 sequence. The laws of use and disuse apply here as in other forms 
 of learning; the physiological basis is the same. 
 
 Since recitation is equivalent to practice in other forms of learn- 
 ing, we should expect as a matter of course that any restriction 
 upon its employment during the process of memorization should 
 result in retarding improvement. Our experiments upon this point 
 have shown that this is the case. This was true for all subjects, 
 except children so young as to be unable to meet the requirements 
 of the test, and for all materials employed, although, as might 
 reasonably be expected, minor differences are to be found. In 
 general, a method in which recitation is introduced at the optimum 
 
100 RECITATION AS A FACTOR IN MEMORIZING 
 
 time, in comparison with a method in which the learner is entirely 
 restricted to reading, enables the learner to reproduce immediately 
 after a short period of study approximately twice as much material. 
 The advantage of recitation as one should expect, is much more 
 pronounced in delayed recall. After an interval of three or four 
 hours, r ecitation makes possible the recall of four times as much 
 .materia l as does reading. This is to be expected, since recitation is 
 understood as a process of adequate practice, while reading, whose 
 function is introductory, restricts or inhibits the exercise of the 
 bonds upon whose strength recall depends. In reading, while 
 many of the bonds may be well enough established for immediate 
 use, the neural connections rapidly disappear with disuse. 
 
 As the nature of reading and recitation now appears, the ques- 
 tion is not so much — How is it that reading produces such poor 
 results? — but rather — How is it that reading permits of any memo- 
 rization at all? The evidence that has been gathered makes it 
 doubtful whether pure reading would result in memorization. But 
 there is little doubt that pure reading is a fiction ; more or less recita- 
 tion is always present in any prolonged effort to learn. 
 
 The fact that reading is seldom if ever pure can be most clearly 
 illustrated in the case of learning sense material, and this fact helps 
 us at the same time to understand why reading as a method of 
 learning is more fruitful when applied to such material than when 
 employed with non-connected senseless material. Nearly all of 
 the subjects admitted that their learning, especially of sense mate- 
 rial, was not limited to pure reading. The eye moved along the 
 line actually seeing only occasional words. Other words, in fact 
 whole phrases, were filled in by recall. The text served only to 
 suggest groups of words or ideas which were for the most part filled 
 in by the learner. In so far as this subjective reproduction of the 
 material was carried on, to just that extent the learner was reciting 
 rather than reading, and without doubt this sort of recall was at 
 all times considerable, becoming more and more so as the learning 
 progressed. Consequently, it appears that the memorization of 
 the material, technically speaking, must, after all, be attributed to 
 recitation. 
 
 The findings of Chapter IV were to the effect that reading 
 was much more productive when the material was senseful and con- 
 nected than when senseless and non-connected. The previous para- 
 graph explains in part why this should be so. Reading of senseful 
 connected material is far from pure, while with senseless material, 
 on account of its less fluency and lack of senseful serial associations, 
 the learner finds it less unnatural to actually see and read each item. 
 No associations are present in the material which enable the learner 
 
RECITATION AS A FACTOR IN MEMORIZING lOI 
 
 to fill in the gaps when only occasional syllables are read. The 
 bonds between items must be built up by the learner himself, and 
 it is in this process that recitation is of the greatest value. These 
 two factors together, namely, that the bonds between items in 
 nonsense material must be worked in by the learner and that 
 reading is much more pure with this material, explain the relatively 
 greater advantage which recitation brings about with nonsense as 
 compared to senseful material. 
 
 In addition to the fact that recitation as compared with reading 
 enables the learner to form the requisite bonds more quickly and 
 more permanently, the results of Chapter V have indicated other 
 advantages of recitation as a form of learning. It was found that 
 recitation leads to greater certainty of one's knowledge. It enables 
 the learner not only to know but to be aware of how well he knows. 
 Fewer blunders and erroneous recalls are made. The material is 
 better organized; it is in more usable form. The meaning of the 
 material is better obtained, and the relations among parts become 
 more clear. In addition to this, as Katzaroff found (see page 5), 
 material learned by means of recitation can be more promptly 
 recalled; the recitation time is less. 
 
 From every point of view the superiority of recitation over 
 reading, beyond the few perusals required to furnish the initial 
 grasp of the material, is very clear. It holds for all materials and 
 for practically all subjects. Consequently, the applications of the 
 results to pedagogy are direct and manifestly important. 
 
 For the improvement of methods of study among school chil- 
 dren, it is first of all necessary that the teacher should be aware of 
 the value of recall in learning and that she should endeavor to 
 impart this information in a practicable way to the pupils. That 
 the pupils cannot be depended upon to discover economical methods 
 of studying by themselves has often been discovered by inquiry. 
 Miss M. J. Baldwin,^ for example, found for grammar and high 
 school pupils "that eighty-two per cent, studied words rather than 
 thoughts, that they study in a mechanical sort of way which enables 
 them to say that they have studied the lesson and spent the re- 
 quired time. They read the words over and over and doubtless 
 get more confused the more they read." 
 
 It is perhaps not sufficient, however, that the pupils should be 
 merely aware of the fact that attempted recitation is an essential 
 process in learning. The teacher must devise means by which the 
 pupils may be induced to study by trying to recall the material 
 rather than by merely continuing slavishly to read and reread the 
 words. The determination of these means, of course, does not lie 
 
 1 'How Children Study', Archives of Psychology, 1909, No. 12, p. 70. 
 
102 RECITATION AS A FACTOR IN MEMORIZING 
 
 within the scope of this study. It has been the purpose here merely 
 to show that the recitation method can be employed, and employed 
 very effectively, by pupils from the third grade up. 
 
 A few things which may induce the pupil to rely more upon recall 
 will occur to anyone. Any method which requ ires the summarizar 
 tion of t he facts of the lesson brings recall into play, since i n such a 
 ^proc ess the pupil must think over the whol e material, cul l out the 
 ess entials, and state in his own word s the main points . The teacher 
 should encourage the pupil to react to the lesson in this way and 
 reward him for successful attempts. Cond ensations. of the ideas in 
 written form, or even better, if possible, summarizing the content 
 jjf the lesson mentally, is almost certain to bring rich returns . It 
 brings into play the beneficial factors involved in recitation, devel- 
 ops power to distinguish the essentials from the unessentials, and 
 may develop confidence and satisfaction in the pupil, since it 
 enables him to be more certain of his mastery of the material. The 
 pupil who has reasonable assurance that he has the lesson in hand 
 can approach the recitation before the teacher in a more effective 
 frame of mind. 
 
 The fact that recall is of such great importance in learning has 
 a significant bearing on the nature of the recitation period in the 
 school-room. As Colvin has pointed out:^ "The fact that the reci- 
 tation, as such, is largely ignored in higher grades of instruction is 
 doubtless a serious pedagogical defect, which can be remedied only 
 by accustoming the student to practise on his own initiative recall 
 in his learning." The recitation should be regarded not merely as 
 an opportunity afforded the teacher to find out what the pupils 
 know, how hard they have studied, and what grade they should be 
 given. Instead of an inquisition it should become a period of in- 
 struction. It should offer the pupil an opportunity to recite material 
 he has previously more or less completely mastered. If the silent 
 pupils could be induced to recall the material as well as the pupil 
 who is orally reciting to the teacher, the period could become a 
 valuable opportunity for review. Its most admirable function 
 would consist in affording the pupils an opportunity to discover 
 where their knowledge is hazy, inexact, and uncertain. 
 
 In addition, the teacher should make of the recitation a means 
 of discovering the methods of studying employed by the pupils 
 and of suggesting improvements in that respect. The unprepared 
 student should not simply be met by the remarks: "How many 
 times did you read your lesson?" and "Go read it again!" but more 
 detailed inquiry into the cause of failure, followed by more valuable 
 suggestions with regard to methods of study, should be the pro- 
 
 2 The Learning Process, New York, 1913, p. 165. 
 
RECITATION AS A FACTOR IN MEMORIZING IO3 
 
 cedure. In a word: "It should be remembered that instruction 
 in the technique of learning is perhaps as important as instruction 
 in the content of the subjects of the school curriculum."^ 
 
 More advance d _^^s tudigi tsjmay prqfi t^ b^ the knowledge of the 
 indispensible value of recitation. The college student is confronted 
 by a situation in which the 'absorbing' of knowledge seems para- 
 mount, and where reaction is too little required. Listening to 
 lectures and reading the texts require most of his time; recitations 
 are few and far between. That they 'read lots but learn little' is 
 a stock criticism, and it is indeed not seldom true that the college 
 student is quite as ignorant of economical methods of study as the 
 grammar school pupil. Recently the writer heard the case of a 
 college student who came to a professor of psychology for an exam- 
 ination of what he believed to be a very poor memory. The student 
 asserted that he could read a lesson over a dozen times and still not 
 know it. A brief examination showed his memory not to be below 
 par, but all the evidence indicated entirely inadequate methods of 
 study. The stu dent relied upon impre ssion with little or no effort 
 at expression ; recall of the main points of his lesson was seldom 
 tried. Yet for the college student who is so seldom called to account 
 for his acquirements, recitation is more than usually essential. Fre- 
 quen t reviews,^ thinking the matter oyer by one's self, writing briefs 
 of the main points, conversation with other st udents, and the like, 
 are valuable because they th row into relief the portions that are 
 hazy, ine xact, and confused as well a s because they fix more clearly 
 in mind the m aterial that is rehearsed . 
 
 Various opinions have been expressed with regard to methods 
 of taking notes during lectures."* Doubtless the method must be 
 varied somewhat to suit the material that is presented, but the 
 findings in the present study suggest a method which, although 
 seldom employed, should bring good results. Instead of making of 
 one's self a mechanism for transferring spoken words to paper with 
 but little heed to the meaning, the student devotes his attention to 
 a thorough understanding the material presented, selecting the 
 important points, organizing them into a systematic whole as the 
 lecture progresses, and for the most part, delaying to a later hour 
 the writing of the notes. Later in the day or evening, the lecture 
 is rehearsed an dean outline written down for future reference . While 
 some disadvantages, or more likely, inconveniences, of such a 
 method may appear, certain advantages of an important nature 
 are obvious. First of all, the student may develop better habits 
 
 3 Colvin, op. cit., p. 178. 
 
 * This subject will be found discussed at length in two recent books: G. V. N. Dearborn, How 
 to Learn Easily, Boston, 1916, Chapter II, and Harry D. Kitson, How to Use Your Mind, Phila- 
 delphia, 1916, Chapter II. 
 
104 RECITATION AS A FACTOR IN MEMORIZING 
 
 of attention during the lecture. JHje fo rces himself to pick out the 
 , essen tials, to grasp the relations of ideas and to unify and organize 
 the materia l presented. The will to remernber,^, which Meumann so 
 strongly emphasizes , com es into play. _^ The student must actively 
 grasp the meaning of the lecture in order to be able to reproduce it 
 later. Secondly, the writing of a brief of the lecture at a later hour 
 combines the advantage of a recitation, which the copious note- 
 taker too seldom practises, with the well known benefits to be 
 derived from the distribution of learning periods.^ A few students 
 who have tried this method speak enthusiastically of its effectiveness. 
 Finally, a word with regard to a more technical application of the 
 results of this study. Individuals, when permitted to study by 
 their 'natural method,' were found to employ various methods, not 
 only for different materials, but for lessons of the same material 
 and of the same length, at different times. ^ The quantitative 
 results consequently vary considerably, according to whether the 
 subject does or does not happen to employ an optimum combination 
 of reading and recitation. In experimental work on memory and 
 learning in which successive tests under constant conditions are 
 required, it w ould seem tojbe^an important precautio n to specify 
 the time at whic h the learne r shou l d change from reading to at- 
 tempted recall, with instructions to employ thereafter the recitatioy 
 method until learni ng is complet e . 
 
 5 See Jost. A., 'Die Assoziationfestigkeit in ihrer Abhangigkeit von der Verteilung der Wieder- 
 holungen', Zeitschrift fiir Psychologic, 1897, 14, pp. 436-472, or Ebbinghaus, H., Memory, trans- 
 lated by H. Ruger and C. Bussenius, New York, I9I3. 
 
 6 But little of the actual data bearing on this point has been presented in this paper. For the 
 most part, such data were obtained from the practice tests conducted preliminary to those here 
 presented. 
 
VITA 
 
 Born in Red Wing, Minn., September 22, 1890. Graduated 
 from primary school of Fortuna, California, in 1905, and from 
 Fortuna High School in 1909. Received degree of B.L. in 1914, 
 and M.A. in 191 5, from the University of California. Attended 
 Columbia University during the academic year 1916-1917. Assistant 
 in Psychology under Professors G. M. Stratton and Warner Brown 
 in the University of California during the academic years 1914-1915, 
 and 1915-1916. Assistant in Psychology in Columbia University, 
 1916-1917. Major work at Columbia taken with Professors J. 
 McKeen Cattell, R. S. Woodworth, E. L. Thorndike, and Dr. 
 A. T. Pofifenberger, Jr. 
 
019 820 136 4