The original of tliis book is in 
 tlie Cornell University Library. 
 
 There are no known copyright restrictions in 
 the United States on the use of the text. 
 
 http://www.archive.org/details/cu31924029039075 
 
On the Perseverative Tendency 
 
 BY 
 WILLIAM SILLIMAN^FOSTER 
 
 A THESIS 
 
 Presented to the Faculty of the Graduate School of Cornell 
 
 University in partial fulfilment of the req"uirements 
 
 for the degree of Doctor of Philosophy 
 
 Reprinted from the American JotmNAL op Pstchologt 
 July, 1914, Vol. XXV, pp. 393-426 
 
On the Perseverative Tendency 
 
 BY 
 WILLIAM SILLIMAN FOSTER 
 
 A THESIS 
 
 Presented to the Faculty of the Graduate School of Cornell 
 
 University in partial fulfilment of the requirements 
 
 for the degree of Doctor of Philosophy 
 
 Reprinted from the American Journal op Pstcholoqt 
 July, 1914, Vol. XXV, pp. 393-426 
 
ON THE PERSEVERATIVE TENDENCY^ 
 
 By William Siluman Foster 
 
 CONTENTS 
 
 I. Introductory 393 
 
 II. Historical / .- 394 
 
 III. Experimental 402 
 
 Preliminary Experiments 402 
 
 Experiments by the Method of Right Associates 407 
 
 Experiments by the Method of Retained Members 421 
 
 IV. Conclusion 425 
 
 I. Introductory 
 
 The spontaneous appearance of formations in conscious- 
 ness, that is to say, their appearance without peripheral stimu- 
 lation or associative connection, is a familiar phenomenon in 
 the psychology of common sense, and a fundamental concept 
 in the metaphysical psychology of Herbart. In the terms 
 perseveration and perseverative tendency the concept has 
 found continued and increasingly extended use in experi- 
 mental psychology, since the introduction of these terms in 
 the investigation of memory by Miiller and Pilzecker (1900). 
 In the last twenty volumes of the Zeitschrift fiir Psychologie 
 the article on Memory which does not involve one or both of 
 these terms is the exception rather than the rule, and much 
 use is made of them also in the treatment of memory in other 
 journals and in text-books. Not in the field of memory alone, 
 however, but also in such different fields as those of feeling, 
 rhythm, association, action and thought, do we find the ap- 
 peal to the freely rising or to the persisting mental formation, 
 and at least one writer has based a theory of introspection 
 upon the perseverative tendency. 
 
 When an attempt is made to bring together the knowledge 
 of perseveration which has been gained in such investigations, 
 however, great systematic difficulty arises. As the expression 
 " freely rising or persisting mental formations " briefly indi- 
 cates, and as we shall try to show in greater detail in the 
 historical section which -follows, no sinpe meaning has at- 
 tached to the terms. They have,, on the contrary, been most 
 
 1 From the Psychological Laboratory, Cornell University. 
 
394 FOSTER 
 
 frequently used without apology and with a great variety of 
 implications. Where criticisms have been offered, there have 
 resulted numerous different definitions, or in some cases there 
 has come an outright denial of the admissibility of the con- 
 cepts. 
 
 While the associative tendency and even the more recently 
 supposed determining tendency have been made the objects of 
 special investigation, and while methods have been developed 
 for establishing their conditions and measuring their strength, 
 we have as yet no direct investigation of the perseverative 
 tendency. For the most part, rather, the perseverative ten- 
 dency has been invoked to explain results incidentally obtained 
 in attacks on other problems, especially where such results 
 seemed difficult to explain in the better recognized ways. It 
 is in view of this disagreement as to its significance and status 
 that we have undertaken an historical review and direct ex- 
 perimental investigation of the perseverative tendency. 
 
 II. Historical^ 
 
 The term perseveration seems to have been first used by 
 the psychiatrist Neisser^ to indicate the fact of abnormally 
 persistent repetition or continuation of an activity, after the 
 activity had been once begun or recently completed, such as 
 is evidenced in the frequent speaking or writing of a given 
 word or words in unsuitable places. For Neisser the term is 
 descriptive, the name of a symptom to be recognized in the 
 clinic. Later psychiatrists -do not limit the term to this use, 
 however. Not merely the externally evident reactions, but 
 " perceptual and purely internal " reactions, ways of appre- 
 hending and perceiving things or situations, soon come also to 
 be classified as perseverations. Perseveration takes on an ex- 
 planatory aspect as well. For some it indicates that " inner- 
 vation processes which from some cause do not attain their 
 goal or express themselves in one way, do so in another " ; for 
 
 2 A review of the psychiatrical literature of perseveration is given 
 by K. Heilbronner, Ueber Haftenbleiden und Stertotypie, rMonatsch. f. 
 Psychiatrie _u. Neurologie, i8, 1905. Erg.-Hft., 293 flf. We are indebted 
 to this review chiefly for the details upon which the generalizations 
 of the first paragraph of our historical discussion are based. On the 
 psychological side are reviews of the literature on memory for certain 
 years by H. J. Watt. Arch. f. d. ges. Psych., 7, 1906, Literaturber., 
 16 ff.; ibid., 9, 1907, Literaturber., 22 ff., and a critical review by C. 
 Jesinghaus, Zur psychologischen Theorie des Gedachtnisses, Psychol. 
 Stud., 7, 1911, 365 ff. 
 
 3 65 Sitzung des Vereins deutscher Irrenartze, Zeitseh. f. Psychia- 
 trie, SI, 1894, 1016. 
 
ON THE PERSEVERATIVE TENDENCY 395 
 
 others it means " psychical after-images," or " weakness and 
 slowness of association," and so on. The further uses of the 
 term in psychiatry need hardly concern us. It is enough to 
 point out here that psychiatry offers no single standard con- 
 tent for the concept. 
 
 The term perseveration was not common in psychology 
 until after 1900 when, as we have previously indicated, it was 
 introduced by Miiller and Pilzecker* in connection with the 
 perseverative tendency. The latter term is defined by these 
 authors as follows : "After once having entered conscious- 
 ness, every idea possesses a perseverative tendency, that is, 
 a tendency to rise freely into consciousness. The tendency 
 generally dies rapidly away. The more intensely attention 
 was directed upon the idea, the stronger is the tendency, and 
 it increases if the given idea or series of ideas is very soon 
 repeated. With cumulative repetition it may easily happen 
 that, at times when the other iactors which are besieging 
 consciousness are not of especial strength and persistence 
 (Nachhaltigkeit) , the given idea or series of ideas will come 
 to consciousness solely in consequence of their perseverative 
 tendency." As proofs of the necessity of assuming the pres- 
 ence of a perseverative tendency, they give a list of observa- 
 tions and experimental results, (i) After one has been busy 
 for several hours in attentive observation of a phenomenon, 
 such for instance as the movements of a pointer over a scale, 
 the phenomenon may suddenly present itself again, in image, 
 with the distinctness of sensory experience, especially at times 
 when attention is not demanded in other directions (Fechner's 
 phenomena of sensory memory)." (2) Impressive experi- 
 ences sometimes reappear in the form of hallucinations, at 
 least at times when attention is not demanded strongly in 
 other directions." (3) A long continued motor activity sets 
 up in the lower centers a tendency automatically to continue 
 or to repeat the activity. After a strong disposition of this 
 sort has been established, the center affected may " reel off " 
 the activity for a certain length of time, without the necessity 
 of further incitement (motorische Einstellung) .'' (4) Cases 
 of perseveration, in the sense of the psychiatrists, as described 
 above. (5) Quite against our will, ideas and thoughts in 
 
 * G. E. Miiller u. A. Pilzecker, Experimentelle Beitrage zur Lehre 
 vom Gedachtniss. 'Zeitsch. f. Psychol, Ergbd. i, 1900, 58 ff- 
 
 5 G. T. Fechner, Elemente der Psychophysik, II, i860, 498 ff. 
 
 " Fechner, op. cit., Si3- 
 
 ^ G. E. Miiller u. F. Schumann, Arch. f. d. ges. Physiol, 45, 1889, 
 Z7 ff. Cf . L. Steffens : Zeitsch. f. Psychol, 23, 1900, 242 ff. 
 
396 FOSTER 
 
 which we have been deeply interested for some time — melo- 
 dies, chess games, and the like — may keep coming back into 
 our minds. (6) We often make mistakes in reading, hear- 
 ing, speaking or writing words, where words had in mind 
 previous to the mistakes seem, by their continuance (Nach- 
 klingen), responsible for the disturbance. (7) In experiments 
 on memory, ideas given as reactions in the first part of a 
 series may recur during the latter part of the series without 
 the volition of the observer, and without any evidence of 
 special reproductive tendencies. (8) In spite of instructions 
 not to think of the series, many observers cannot prevent cer- 
 tain syllables of a series which they have learned from com- 
 ing to mind in the interval between the learning and the test. 
 (9) In experiments by the method of right associates, not 
 only may a syllable which has caught the observer's attention 
 be given once or oftener as a response in the test of the series 
 of which it was a part, but it may also recur as a response 
 on a number of later days (habituelle Aushiilfesilben). In 
 cases where the correct associate is partially identical with 
 such a syllable, the chances of this sort of incorrect response 
 are increased. (10) Incorrect responses, in which the syllable 
 spoken is a syllable from the series, but is given in the wrong 
 place, are much more frequent in tests which follow shortly 
 after learning than in tests which follow at longer intervals, 
 even though the number of correct responses in the two cases 
 is approximately the same. (11) Series tested after a short 
 interval give shorter reaction-times than series tested after 
 longer intervals, even though the number of correct replies 
 is approximately equal in the two cases ; and mere rereading 
 of recently read material may be accomplished faster than 
 rereading of older material. (12) Mental activity during a 
 period shortly after learning not only disturbs the associative 
 connections (retroactive inhibition), but also brings it about 
 that the syllables are less likely to rise freely into conscious- 
 ness. (13) Observers differ greatly in respect to the degree 
 in which the above mentioned symptoms of perseverative ten- 
 dency appear. 
 
 It has often been pointed out that the definition of persevera- 
 tive tendency given above is not altogether unambiguous.* After 
 its entrance (nach ihrein Auftreten) into consciousness, and 
 while it is still in consciousness, an idea cannot well be said 
 to have a tendency to rise freely into consciousness, but only 
 
 ^W. Wundt: Grundsuge d. psysiol. Psychol., Ill, 191 1, 572. Jesing- 
 haus, op. cit., 365. W. Poppelreuter, Ueber die Ordnung des Vorstel- 
 lungsverlaufes, Arch. f. d. ges. Psychol, 25, 1912, 293. , 
 
ON THE PERSEVERATIVE TENDENCY 397 
 
 to have a tendency to continue in consciousness, or to run a 
 temporal course. Only after a formation has disappeared 
 from consciousness is it possible to speak of its tendency to 
 rise freely. It is difficult to say whether Mxiller and Pilzecker 
 mean that the tendency explains merely the free arousal of 
 the idea, or both this and its continuation. The analogous 
 motor disposition brings about a repetition or continuation of 
 the activity ; the lower center " reels off " the activity. The 
 syllables in the interval between learning and test may be 
 a continuation of the series as well as a recurrence. ° It may 
 be the continuation of the series which is damaged by imme- 
 diately subsequent mental work. The previous word seems 
 to linger (nachklingen) in the cases where we make mistakes 
 in speaking; and so on. 
 
 Whether or not these authors mean the perseverative ten- 
 dency to explain both continuance and free recurrence, how- 
 ever, it is certain that others who use the term later do so 
 understand it. Out of the many cases in which both mean- 
 ings appear in the work of a single author, we choose two in 
 which the twofold use is expressly stated. Ach^° says : " The 
 tendency of a psychical experience to rise freely into con- 
 sciousness or to remain (verharren) in consciousness as an 
 after-effect has already been established experimentally by 
 Miiller and Pilzecker, and called perseverative tendency " ; 
 and Ach refers to the memory after-image. Wertheimer^^ 
 
 " Miiller would probably deny the possibility of such a continua- 
 tion. In his review of Ebert and Meumann's work on practice in 
 memory (Zeitsch. f. Psychol., 39, 1905, 124) we find : " It is evident 
 at once that in immediate retention it is not a matter of after-images. 
 The after-images of a series of auditory or visual perceptions of 
 syllables or the like would give a simultaneous mixture of sounds or 
 colors, and not a successive appearance of distinct single ideas of 
 syllables, visual signs, and the like." Meumann believes, however, 
 that in immediate recall " reproduction takes place on the basis of 
 an immediate after-effect of the impressions and before these impres- 
 sions have yet faded out" (E. Ebert and E. Meumann: Ueber einige 
 Grundfragen der Psychologic der Uebungs-Phanomene im Bereiche 
 des Gedachtnisses, Arch. f. ges. Psychol.,^, 1905, 205). One of the 
 observers reports (p. 20) that reproduction followed " almost auto- 
 matically " on the basis of auditory after-images. Compare the experi- 
 ments on the range of consciousness, where forty impressions may be 
 present simultaneously according to Wundt {op. cit., 335), apparently 
 without any sign of "mixture." 
 
 1" N. Ach, Ueber die Willensthdtigkeit u. d. Denken, 1905, 10. 
 (Italics mine.) 
 
 " M. Wertheimer, Experimentelle Untersuchungen zur Tatbestands- 
 diagnostik, Arch. f. d. ges. Psychol, 6, 1906, 123. Cf. the nachkUn- 
 gende Bereitschaft of Pohlmann, Exp. Beitr. sur Lehre vom Ge- 
 ddchtnis, 1906, 28. 
 
398 FOSTER 
 
 " understands by perseveration (a) the tendency of a psychi- 
 cal content to remain in consciousness, to continue even m 
 the face of an opposing tendency, and (b) the tendency which 
 may be ascribed to a psychical content b to return into con- 
 sciousness, not only because ab, but because b (xb), was once 
 present." Some authors, on the other hand, use persevera- 
 tion to indicate not both, but only one of these tendencies. 
 Jung and Riklin" " use the name ' perseveration ' in the sense 
 of the investigations of Miiller and Pilzecker," and mean by it 
 " merely the continuance of the previous idea, in so far as 
 this continuance is manifest in the character of the following 
 reaction." Ebert and Meumann^' limit the conception in an- 
 other way. " Perseveration apparently played a very unim- 
 portant role with our observers. Four men . . . main- 
 tained very definitely that outside of the special experimental 
 hours a spontaneous appearance of the material which we had 
 presented to be learned never occurred." Even within a sin- 
 gle one of these meanings, we often find further limitations 
 and extensions. Jung and Riklin,^* for example, use the term 
 merely as classificatory and descriptive ; " it is not meant to 
 be explanatory." In their sense, " the concept has no relation 
 to the perseveration in organic brain-processes," although 
 Miiller and Pilzecker had quoted just these cases in support 
 of the perseverative tendency. Nor has it any relation to 
 " the hypothetical ' secondary function of the brain-cells,' 
 which for some authors explains the psychological after-effects 
 of the directive idea " ; although Muller and Pilzecker^" insist 
 that the perseverative tendency has a definite influence in hold- 
 ing the mind to one particular course of thought, or in bring- 
 ing it back to a previous course after an interruption. Jung 
 and Riklin do not count the recurrences of the same reply 
 at different times during an experiment as perseverations, but 
 have a separate category, repetitions, for these reactions. 
 They give, as instances of perseveration, such cases as melt 
 — hot followed by slow — Hre, where, as they say, the condi- 
 tioned reaction is an association to the previous stimulus - 
 word ; although Muller and Pilzecker attribute to persevera- 
 tion alone those cases solely in which association is not con- 
 cerned. They take account of perseveration only in case the 
 influence is apparent in the very next reaction after that in 
 
 12 C. G. Jung u. F. Riklin, Diag. Assoziationstudien, 1906, 31. 
 (Italics mine.) Cf. E. Stransky: Ueber Sprachvenmrrtheit, 1905, 28. 
 
 13 Op. cit., 154. 
 " Op. cit., 31 f . 
 
 1" Op. cit., 75. Cf. W. Schafer, Ueber die Nachwirkung der Vor- 
 stellungen, 1904. 
 
ON THE PERSEVERATIVE TENDENCY 399 
 
 which the influencing word is found. In general, we find a 
 large number of instances in which the concepts of persevera- 
 tive tendency and perseveration are changed from their origi- 
 nal sense, and made to fit in with the purposes and conven- 
 ience of the particular author using the terms. 
 
 As we have indicated above, the perseverative tendency is 
 sometimes equated to the " impressional tendency," or to 
 that " readiness " of the idea to be reproduced, which is de- 
 pendent upon the recency, frequency, etc., with which the 
 impression has been made, and which does not depend upon 
 the strength of its associative connections. With such a con- 
 ception, the perseverative tendency becomes effective for re- 
 production only when associative tendencies are also brought 
 into play. This view is held, for example, by Wreschner.^"* 
 " Rather must we say that associative tendencies to .reproduc- 
 tion are always added to the perseverative. Certainly the 
 former may be of such little strength or so highly complicated 
 that the single elements are unconscious or, if you prefer, only 
 dimly conscious, and so the appearance of freely aroused ideas 
 may be given. . . . But if I understand correctly the de- 
 velopments made by Miiller and Pilzecker, they only maintain 
 the fact that the associative factors do not alone determine 
 the course of ideas, but that along with these factors the per- 
 severative tendencies also have to be taken into account, since 
 they are often present in such a high degree that the associa- 
 tive factors, although never entirely lacking, lose their im- 
 portance." This also is the way in which Watt^' conceives 
 the perseverative tendency. " Ideas which have recently been 
 in consciousness have for a certain time the property of 
 coming to consciousness again very easily and quickly, or of 
 being reproduced. . . . This property is called the perseve- 
 rative tendency of the ideas, and we say that such ideas are 
 in a state of readiness." The question whether or not ideas 
 may rise freely into consciousness. Watt regards as in need 
 of experimental test. The close relation of the perseverative 
 tendency to the impressional tendency is pointed out by Titch- 
 ener.^' " Why, then, should we separate impressional and 
 perseverative tendency? Why should we not say, at once, 
 that the impressional tendency varies, oscillates, fluctuates ? " 
 And the answer given is that the status of the perseverative 
 
 1^ A. Wreschner, Die Reproduktion und Assoziation von Vorstel- 
 lungen, Zeitsch. f. Psychol, Ergbd. 3, 1907, 13 f . 
 
 !■' H. J. Watt, Experimentelle Beitrage zu einer Theorie des Den- 
 kens. Arch. f. d. ges. Psychol, 4, 1905, 341. 
 
 18 E. B. Titchener, Text-Book of Psychology, 1910, 400 f . 
 
400 FOSTER 
 
 tendency (as a sort of rhythm imposed on the impressional 
 tendency, making it now and again effective for reproduction) 
 is uncertain, and that the alleged cases of perseveration are 
 explainable in better recognized ways. 
 
 Perseveration, moreover, has not been limited to ideas and 
 other content-processes, but has been extended to explain the 
 course of consciousness at large. Witasek^' speaks of the 
 perseveration of feeling. " After it has just disappeared, an 
 act-feeling of definite character establishes a tendency to be 
 aroused again upon another provocation soon after, even 
 though this provocation under other circumstances would not 
 in and of itself be suitable to arouse just that feeling." Stein'^" 
 and other workers in the psychology of testimony speak of 
 the perseveration of feeling-tone. According to some writers 
 not only do we have a perseveration of mental processes, but 
 also of forms of consciousness and of mental acts. One of 
 the clearest and most definite statements of this position is 
 given in the following quotation from the conclusion of 
 Koffka's" article on rhythm. " We must ask what it is that 
 perseveres, the form of the groups, the activity, or both to- 
 gether. From our results it follows with certainty that the 
 forms of the groups as such have perseverative tendencies. 
 Such groups persisted, even when the rhythmic experience 
 had ceased, and mere regularity had taken its place. A per- 
 severation of the activity itself is to be recognized when, 
 after a rhythm is once established, a new rhythm is more 
 easily called forth than it would have been without the pre- 
 vious one." Maltzew^* finds that observers required to judge 
 the pitch of tones or the interval between tones often report 
 a pitch or interval just previously reported, and calls this a 
 perseveration of the pitch or of the interval. Watt"^ speaks 
 of the perseverative tendency of the Aufgabe; Ach^* discusses 
 the perseveration of a determination, as indicated by the con- 
 tinued correctness of reaction without repetition of directions 
 before every experiment, and by the occasional slip into an 
 old disposition when a new or difficult task is set, and the 
 
 18 S. Witasek, Grundlinien der Psychologie, 1908, 341 f. 
 
 2" P. Stein, Tatbestandsdiagnostische Versuche bei Untersuchungsge- 
 f angenen, Zeitsch. f. Psychol., 52, 1909, 162. Cf. Jung and Riklin : 
 op. cit., Wertheimer, loc. cit. 
 
 21 K. Koffka, Experimental-Untersuchungen zur Lehre von Rhyth- 
 mus. Zeitsch. f. Psychol., 52^ 1909, io8. 
 
 22 C. V. Maltzew, Das Erkennen sukzessiv gegebener musikalischer 
 Intervalle in den ausseren Tonregionen. Zeitsch. f. Psychol., 64, 1913, 
 224 ff. 
 
 23 Op. cit., 343 f . 
 
 2* Ueber den Willensakt und das Temperament, 1910, 55 f . 
 
ON THE PERSEVERATIVE TENDENCY 401 
 
 instructions are not thoroughly ingrained in the observer. In 
 this general and more or less descriptive and classificatory 
 sense, we find a very large number of cases in which recourse 
 is had to the term perseveration. Almost as numerous as the 
 users of the term are the sorts of behavior that are regarded 
 as symptomatic of perseveration ; the observer in an associa- 
 tion-experiment reacts with a previously given stimulus-word 
 or reaction-word, reacts continuously in the same psychological, 
 logical, or grammatical form (rhyme, similarity, opposition, 
 etc.), reacts with a topical sequence of words, and so on.^' 
 
 The result of our historical survey, then, is as follows. 
 The terms perseveration and perseverative tendency are used 
 almost interchangeably. They may mean that there is a ten- 
 dency for ideas ^ once impressed to come to consciousness 
 rhythmically or otherwise, with or without the assistance of 
 the associative tendencies ; that ideas once in consciousness 
 tend to remain in consciousness for a certain length of time 
 or to run a temporal course ; that certain ideas, from the mere 
 fact that they have been more strongly impressed, are readier 
 to appear in consciousness than are ideas less strongly im- 
 pressed ; that feelings, acts, and dispositions may show the 
 same tendency as ideas in one or more of these senses ; and 
 that certain types of physical or ' mental ' behavior may be 
 expected under given circumstances. The theory of perse- 
 veration, we conclude, derives for the most part from inci- 
 dental psychiatrical observation and from the psychological 
 investigation of memory and association.^* The problem of 
 perseveration has not been attacked by itself or for its own 
 sake.^' 
 
 Many explanations have been offered of the facts alleged 
 to prove the necessity of the assumption of perseveration in 
 
 25 See Wreschner, op. cit., ii f., 15, 237 ff.; Jung and Riklin, op. cit; 
 E. Meumann, Intelligenzpriifungen an Kindern der Volkschule, Exp. 
 Padagogik, i, 1905, 96 f. ; R. R. Rusk, Mental Association in Children, 
 Brit. Jour. Psychol., 3, 1910, 365; T. Ziehen, Ideenassoziation des 
 Kindes, Samml. v. Abhand. aus d. Gebiete der pad. Psychol, u. 
 Physiol, I, Hft. 6, 1898, as f . 
 
 2» P. Ephrussi, Experimentelle Beitraage zur Lehre vom Gedachtnis. 
 Zeitsch. f. Psychol, ^7, 1904, 56 ff., 161 ff. ; W. Jacobs, Ueber das Lernen 
 mit ausserer Lokalisation, Zeitsch. f. Psychol, 45, 1907, 43 ff., 161 ff. 
 Many others might be added to the list of references. 
 
 ^'f Heilbronner, Stransky, Schafer, opp. citt., and Poppelreuter, 
 Nachweis der Unzweckmassigkeit der gebrauchlichen Assoziations- 
 experiraente, usw. Zeitsch. f. Psychol, 61, 1912, i ff., and Ueber die 
 Ordnung der Vorstallungablaufes, Arch. f. d. ges. Psychol, 25, 1912, 
 209 ff., touch the problem perhaps more directly than others. 
 
402 FOSTER 
 
 the sense of spontaneous appearance, and the question as to 
 whether or not the facts can be explained on the basis of the 
 selective or combined action of such factors as mere persist- 
 ence of peripheral excitation, unrecognized associations, re- 
 mote associations, place associations, impression, determina- 
 tion, and the like, seems an open one.^' We hope that the 
 following experiments may contribute towards an answer to 
 this question. 
 
 III. Experimental 
 
 Prelinuna/ry experiments. 
 
 In a small number of preliminary experiments an attempt 
 was made to eliminate or to neutralize the effect of associations 
 between nonsense-syllables, while conditions still remained 
 favorable for the induction of strong perseverative tendencies. 
 Under such conditions it was thought that syllables might be 
 spontaneously reproduced, or that they might show their in- 
 fluence in consciousness in some other way. These conditions 
 we tried to secure by the following method. 
 
 Five series of six syllables were made up. The syllables 
 were selected by chance from the list in Whipple's Manual of 
 Mental and Physical Tests (p. 247) ; but for the same series 
 no syllable was chosen which, combined with another syllable 
 of the series, would make a sensible word ; and no syllable 
 was taken which corresponded with a syllable already chosen 
 in respect to an initial or final consonant. The syllables actu- 
 ally employed were : dro, zig, luh, ild, tud, pum, aum, rin, orp, 
 ool, sef, nuc, loin, set, ruj, elt, smi, nen, gla, lei, gur, baw, 
 ime, acq, orm, tob, pru, gos, ith, tas. The syllables of a series 
 were then written down in the order : 
 
 123456 I 352641 I 546132 I 425316 I 214365 
 In such an order no syllable follows any other given syllable 
 more than once; and no syllable appears to have any special 
 advantage, due to position in the series, over any other. 
 
 Four observers took part in the experiments: Miss L. M. 
 Day (D), a graduate student in psychology with four years 
 of psychological training; Mr. A. S. Edwards (E), fellow in 
 psychology; Dr. L. R. Geissler (G), instructor in psychology; 
 and Dr. E. Jacobson (Ja), honorary fellow in psychology. 
 
 On a single day, one only of the above described series was 
 read aloud to the observer with an attempt not to introduce 
 rhythm. The writer served as experimenter in this and in all 
 the following experiments. The rate of reading was regulated 
 
 28 See Jesinghaus : loc. cit.; Titchener: loc. cit.; Wundt: loc. cit; 
 H. Ebbinghaus: Grundsuge der Psychologic, 1903, 657 f. 
 
ON THE PERSEVERATIVE TENDENCY 403 
 
 by a noiseless pendulum set to beat three-quarter seconds, 
 and a syllable was given at each beat with the skip of a single 
 beat between each group of six syllables. The observers were 
 instructed to give equal and best possible attention to all syl- 
 lables ; but they were told that they would not be required 
 to recall the syllables afterward. During the reading the 
 observer, with closed eyes, sat at the end of a couch upon 
 which, as soon as the reading was finished, he lay down. Dur- 
 ing the following interval of one minute he was instructed 
 not to think of the syllables, but to be as passive as possible 
 and to think of going to sleep, or to attend passively to sen- 
 sations from the skin and body, if such attention should seem 
 necessary in order to keep his attention from the syllables. 
 At the end of the interval the observer had the task of giving 
 slowly, at a suggested rate of approximately one per second, 
 a list of nonsense-syllables of any character, " just such as 
 came into his head." ^^ The experimenter recorded for thirty 
 seconds the syllables given, and took an introspective report 
 of the whole experience in so far as the observer was able to 
 give it. 
 
 It would seem, theoretically, as if such a method should by 
 the arrangement of the series eliminate definite associations, 
 to a great extent at least, through both associative and repro- 
 ductive inhibition; while by the factors of strong attention, 
 repetition, recency, small demand upon the attention of the 
 observer in other directions at the time when the list of sylla- 
 bles was being said, and perhaps also by the shortness of the 
 series, the perseverative tendency should still be favored.'"' 
 
 The following table shows (i) the number of syllables out 
 of the thirty given in the lists by the various observers, (2) 
 the number of times which such syllables appeared as repeti- 
 tions, (3) the number of syllables which corresponded to sylla- 
 bles in the series in respect to two letters only, (4) the sums 
 of these numbers (totals of "influenced" syllables), (5) the 
 number of other syllables, and (6) the total number of sylla- 
 bles in the lists. 
 
 A study of the Table shows that for all observers except 
 Ja the number of " uninfluenced " syllables is much larger 
 than that of " influenced " syllables. In spite of the fact that 
 the perseverative tendency has been favored by the factors 
 
 29 Compare the method used by W. Peters, Ueber Ahnlichkeitsasso- 
 ziation, Zeitsch. f. Psychol, 56, 1910, 167. 
 
 «" Muller and Pilzecker speai (p. s8) of the perseveration not only 
 of an idea but also of a series of ideas, and state (p. 64) that the 
 perseverative tendency of the syllables of a series is still of some 
 strength after five minutes. 
 
404 
 
 FOSTER 
 
 which we have mentioned, all syllables cannot be ' persevera- 
 tions.' Perhaps the instruction to give a list of syllables 
 prejudiced our observers against repetition of syllables. As 
 has been shown by other investigators, learning vvithout the 
 intent to remember, and reproduction without the instruction 
 to give right syllables only, usually result in a small number 
 of " right cases," " though we have little evidence to show 
 whether, under these conditions, associative, perseverative, or 
 " impressional " tendencies are weak. 
 
 TABLE I 
 
 
 D 
 
 E 
 
 G 
 
 Ja 
 
 (1) stimulus syllables given 
 
 (2) Repetitions of stimulus sylla- 
 
 bles 
 
 11 
 
 9 
 29 
 
 49 
 
 93 
 
 142 
 
 10 
 
 13 
 23 
 
 46 
 113 
 159 
 
 8 
 
 5 
 21 
 
 34 
 
 94 
 
 128 
 
 14 
 15 
 
 (3) Part repetitions 
 
 31 
 
 (4) Total " influenced " list-sylla- 
 bles 
 
 60 
 
 (5) " Uninfluenced " syllables 
 
 (6) Total syllables in list 
 
 58 
 118 
 
 
 
 A large number of rhymes appear in the lists, and it is due 
 to these in large measure that the numbers of part repetitions 
 [see Table I, line (3)] largely exceed the numbers of stimu- 
 lus-syllables and repetitions of stimulus-syllables [see lines 
 (I) and (2)]. 
 
 The observers report that it was not difficult to keep the 
 syllables from coming to mind during the interval. When 
 syllables did come, a probable basis for association was 
 usually given. 
 
 D says: "The sounds were absolutely dismissed from mind by 
 moving on the couch. They did not recur. During the rest-period I 
 thought of sleep and said to myself ' Go to sleep.' I tried to keep 
 my eyes fixed toward the window and not turn them towards the 
 experimenter. Once I did turn my eyes involuntarily toward the 
 experimenter and had an indistinct thought of the series, but no defi- 
 nite syllables." " When I lay down, a thought of the syllables came 
 up. (Eyes turned toward the experimenter, vague visual image of the 
 
 '1 See, e. g., Poppelreuter, Zeitsch. f. Psychol., 61, 1912, i fl. ; A. 
 Aall, Ein neues Gedachtnisgesetz ?, Zeitsch. f. Psychol., 66, 1913, i ff . 
 
ON THE PERSEVERATIVE TENDENCY 405 
 
 experimenter and table, and kinesthetic feel of the beat of the series.) 
 I looked around and moved eyes before any syllables came." 
 
 E says : " Felt pressure of eyelids on eyeball, (I often use this as a 
 means of getting to sleep), thought of time passing, and thought the 
 time seemed long. No clear ideas, and no syllable came up even 
 indefinitely. Attention on feel of hands on chest, feet hanging over 
 end of couch, etc." 
 
 G reports : " Not always a blank as required, because the period was 
 too long. At first attention on drowsy sensations, sort of swimming 
 in head, pressure sensations from face on pillow, noticed ticking of 
 clock, hiss of steam, then a few verbal ideas to which I attended, 
 told myself not to think of words. This brought back two or three 
 syllables in verbal form, sef and dag. Then another blank — forgotten 
 its contents." 
 
 Ja reports : " Persistently increased tonicity of the muscles of the 
 face, twitching of the muscles in the left arm, memory of conversa- 
 tion about a dream which I carried on with G a few minutes ago. 
 Turned my eyes toward experimenter, visual image of him, and then 
 the syllables zin, lorn. Then I noticed that I was introspecting, clock 
 and steam catch attention, then attention on kinaesthetic sensations in 
 arm and legs." 
 
 Introspections of the period during which the list is given 
 show that very frequently the observer voluntarily or involun- 
 tarily " thinks of the series," and that syllables belonging to the 
 series are thereby brought into consciousness. Rhyming seems 
 to be dependent, in part at least, upon a partial persistence of 
 the previous syllable. The greater number of " influenced " 
 syllables in the case of Ja may be explained by his report that 
 it is easiest to give a list by sometimes thinking of the series. 
 
 D reports : " Consciousness of the Aufgdbe, setting of my throat, 
 eyes turned toward the experimenter, vague, visual recollection of 
 the series (table and experimenter) ; looked around and moved eyes, 
 scowled, kinaesthetic verbal image of /, started to say it and the rest 
 of the word came of itself; auditory-kinaesthetic images of two or 
 three syllables ; said to myself, ' you are making them all from one 
 letter ' ; remembered that I could do so, but must not do so consciously. 
 Let throat take care of itself unless I felt stuck, then visual image 
 of experimenter and table, j came up, and rest of syllable came up 
 of itself ; other syllables seemed to follow automatically." " Volun- 
 tarily set my throat and it keeps setting itself the same way." 
 
 E says : " While I was giving one syllable, the next one often came 
 up auditorily; wondered whether I should think of the syllables read 
 or make an effort not to. They did not seem to come up of them- 
 selves, except one — lorn." " Kinaesthetic set which meant the problem, 
 i. e., to give nonsense-syllables ; kinaesthesis in throat when searching 
 for consonants; consonant I came up auditorily, and started to speak 
 at the same time; a second syllable came up auditorily as I began to 
 say the first. Sometimes several consonants seemed to be there at 
 once, and one gains its place by the vocal organs following the line 
 of least resistance. The vowels come simply as an automatic filling- 
 in ; very often do not get the vowel auditorily until after I have 
 said it. Noticed when I gave bil that the / was clear in consciousness 
 and began the next syllable." 
 
406 FOSTER 
 
 G says : " Some syllables seem to come of their own accord after 
 a consciousness in which attention is on the sensations from the 
 articulatory organs. Others are suggested by the sound of the pre- 
 vious ones, that is, certain vowels or consonants — chiefly consonants — 
 seem to repeat themselves because the organs are still set for their 
 reproduction. Search for a syllable generally means a kinaesthetic 
 memory-image of the previous vowel, and anticipatory kinaesthetic 
 image of one or two other vowels. The initiatory consonant seems 
 to take care of itself. Sometimes an image of the last consonant of 
 the previous syllable." 
 
 Ja. " As soon as E said, ' Say the list,' verbal motor setting and 
 start to speak. Syllables seem to say themselves. I said blu and 
 then plu, which suggested that I was voluntarily trying to give the 
 series read over again. Then I attempted to let the syllables come 
 of themselves in a greater degree. After a little no syllables formed 
 themselves naturally, and I came back to a thought of the series. 
 This seems to be the easiest way to give a list." 
 
 A close scrutiny of the " uninfluenced " syllables shows 
 that a great many of them are very similar. Certain letters 
 are repeated over and over again. Table II will serve to 
 indicate the number of times in which the more frequent 
 initial and final consonants appear in the "uninfluenced" sylla- 
 bles with. the different observers. In the case of Ja vowels 
 are given five times as initial and six times as final letters. 
 So great a frequency of particular letters and so great an 
 agreement among observers as to the most frequent initial 
 and final letters can not be the result of chance. Neither is 
 explanation possible on the supposition that certain letters in 
 the series which was read were more frequent or attracted 
 attention more definitely; for the most frequent initial and 
 final letters given by the observers in the " uninfluenced " 
 syllables are not especially frequent as initial and final letters 
 in the series read; and as matter which will be presented 
 later shows, greater attention-compelling power of certain 
 letters is not such as to explain the preponderance. 
 
 Certain letters, especially I, m. and n, are likely to appear 
 as initial letters; and others, such as /, m, n, r, and j are 
 likely to appear as final letters, when one decides to say some- 
 thing without knowing what one is going to say. Further 
 account of this phenomenon will be given in our discussion of 
 the results obtained in later experiments (p. 420). 
 
 The main conclusion to be drawn from these preliminary 
 experiments is that, under conditions such as ours, where the 
 perseverative tendency is intentionally favored, and the asso- 
 ciative tendencies are neutralized, there appears little or no 
 evidence of spontaneous reproduction. The apparent influence 
 of the syllables read may be explained in terms of association, 
 impression determination, and continuance. 
 
ON THE PERSEVERATIVE TENDENCY 
 TABLE II 
 
 407 
 
 Obs. 
 
 Total Cases 
 
 Initial Letters 
 
 D 
 E 
 G 
 Ja 
 
 D 
 E 
 G 
 Ja 
 
 93 
 
 113 
 
 94 
 
 58 
 
 93 
 
 113 
 
 94 
 
 58 
 
 1 
 
 24 
 
 1 
 20 
 
 1 
 
 14 
 
 n 
 9 
 
 1 
 26 
 
 k 
 25 
 
 n 
 28 
 
 m 
 
 P 
 
 s 
 
 15 
 
 12 
 
 8 
 
 z 
 
 d 
 
 r 
 
 15 
 
 14 
 
 10 
 
 m 
 
 g 
 
 f 
 
 13 
 
 12 
 
 10 
 
 1 
 
 t 
 
 p, z, w 
 
 8 
 
 5 
 
 4 
 
 b, k, m, p 
 5 
 
 Final Letters 
 
 m 
 
 P 
 
 n 
 
 25 
 
 18 
 
 7 
 
 g 
 
 m 
 
 b 
 
 16 
 
 15 
 
 9 
 
 r 
 
 m 
 
 f 
 
 17 
 
 12 
 
 9 
 
 k 
 7 
 
 s 
 10 
 
 d 
 
 7 
 
 Variations of the experiments, in which parallel series in 
 the same (not in a mixed) order were learned along with such 
 series for comparison, will be described later (p. 421). 
 
 Experiments by the Method of Right Associates. 
 
 According to Muller and Pilzecker, the influence of the per- 
 severative tendency is especially evident in the results of the 
 method of right associates.*'' We determined therefore to per- 
 form certain experiments by this method. We began with two 
 observers, the D and E of the preliminary experiments, neither 
 of whom had previously worked with this method or was 
 especially well acquainted with the literature in this field. 
 Experiments were made on three alternate days per week, 
 and always at the same hour of the morning. 
 
 The nonsense-syllables used were adapted from the list r-60 
 of Muller and Schumann's twelve-syllable normal series, pub- 
 
 82 Op. cit., 281. 
 
408 FOSTER 
 
 lished by Rupp in 1909. Certain changes were made in these 
 syllables, in order that they should be easier for our observers 
 to pronounce ; sch was changed to sh, u to au, and so on, and 
 syllables which had meaning in English were modified. Subse- 
 quent series were made up by the experimenter in accordance 
 with Miiller and Schumann's rules. The syllables were type- 
 written and presented step-wise by means of the Spindler and 
 Hoyer memory apparatus.^' As usual, two blank spaces were 
 left at the end of the series. After preliminary practice, the 
 speed of rotation of the drum was held constant at 8.2 sees. 
 The accented syllables were exposed in the test in the space 
 to the right of the place of presentation. Reaction-times were 
 measured by the Hipp chronoscope, connected in make-to- 
 break series with the exposure apparatus and a Cattell lip- 
 key. The chronoscope was situated in a distant room, and 
 the readings were taken by a second experimenter. Before 
 each experimental hour the chronoscope was tested by means 
 of the large Wundt control-hammer, set at 1500-, and adjust- 
 ments were made to reduce the mean variation of the reading 
 to 1.5 aoT less. The absolute time of the chronoscope reading 
 was found by the use of Wundt's chronograph.'*- That the 
 chi onoscope-times may be regarded for practical purposes as 
 absolute is shown by the fact that ten trials with the Wundt 
 hammer gave 150 ± 1.3 a as measured by the chronoscope, 
 and 149 ± 1.2 0" as measured by the chronograph; with a dif- 
 ferent setting of the control hammer, 180.2 ± 1.3 o- with the 
 chronoscope, and 179 ± 0.50- with the chronograph. Sixteen 
 tests with a large control-pendulum gave with the chronoscope 
 1758.4 ±1.50-, and with the chronograph 1757 ± i.6o-. These 
 figures show that the chronoscope, adjusted for constancy at 
 1500-, was also reliable for the much longer times which fre- 
 quently occurred in our experiments. 
 
 In the first experiments, much as in the earlier experiments 
 of Miiller and Pilzecker,''' the observers were instructed to 
 read the series in trochaic rhythm, giving equal and best pos- 
 sible attention to all syllables, not to think of the syllables 
 during the interval between learning and test, to react in the 
 test as quickly as possible, trying to give the unaccented syl- 
 lable immediately following the stimulus- syllable in the learn- 
 ing series, but not waiting for absolute conviction of the cor- 
 rectness of a syllable which came to mind without such 
 
 '3 Catalogue 21, 1908, No. 215. 
 " Op cit., Ill, 383 ff. 
 »= Op. cit. 8 ff. 
 
ON THE PERSEVERATIVE TENDENCY 
 
 409 
 
 conviction, and to respond with the word " Nothing " if 
 no syllable came. After the test, which on the average 
 took less than a minute, the observer was required to write 
 as complete an introspection of the course of the experiment 
 as he could. With this requirement it was found possible to 
 complete in an hour two experiments, with a five-minute rest 
 between them. 
 
 Table III shows the quantitative results obtained from 12 
 series by each of the two observers. The interval between 
 learning and test was three minutes and the number of repeti- 
 tions was eight. The values were computed as by Miiller and 
 Pilzecker (pp. 26 f). 
 
 TABLE III 
 
 Obs. 
 
 *r 
 
 Tr 
 
 f 
 
 Tf 
 
 V 
 
 Tv 
 
 D 
 
 .39 
 
 3,290 
 
 .42 
 
 4,790 
 
 .10 
 
 16,970 
 
 
 
 E 
 
 .09 
 
 1,480 
 
 .83 
 
 1,190 
 
 .07 
 
 1,490 
 
 
 
 * r =: percentage of correct replies ; f = percentage of incorrect re- 
 plies ; V = percentage of cases in which no reply was given. Tr, Tf , 
 Tv, are the rounded averages of the reaction times of the correct, 
 incorrect, and no replies, respectively. The percentage of cases in 
 which the reply corresponded to the correct reply in respect to at 
 least two parts out of the three (beginning consonants, vowel or 
 diphthong, end-consonants) may be obtained by i.oo-r-f-v. 
 
 The striking difference between' the results of the two ob- 
 servers is at once apparent. E gives many less correct replies 
 and many more incorrect replies than D, and his times 
 throughout are very much shorter. Although, as we shall see 
 from the results of later experiments, E's command of a series 
 in learning and his retention and reproduction of it are prob- 
 ably in general considerably less than D's, this fact alone is 
 not sufficient to explain the differences in the above results. 
 The difference is rather due, primarily, to the fact that E 
 evidently laid the greatest stress on that part of the instruc- 
 tions^^ which requires a reaction as quickly as possible, and 
 permits reactions of whose correctness the observer is not 
 certain. While both the percentages of cases and the times 
 of the right and incorrect responses in the results of D cor- 
 respond very closely with the values given by Miiller and Pil- 
 zecker's observer Jahn under the same conditions (.39 to .39, 
 
 3* Before the work of every hour the instructions were always read 
 to the observer, or read by him from a slip pasted in the note book 
 in which introspections were written. 
 
410 FOSTER 
 
 3290 to 2230, .42 to .38, 4790 to 3200, .10 to .18), E's times 
 fall far below the times both of D and of Jahn. In the early 
 experiments, when judging the correctness of his responses, 
 E would sometimes report : " The syllable is all right, but I 
 and doubtful about its position," meaning that he recognized 
 the syllable as a member of the learned series, but that he 
 could not tell whether or not it was the syllable which had im- 
 mediately followed the stimulus-syllable. E's attention was 
 specifically called to the fact that no such syllable could prop- 
 erly be called correct or " all right," but no appreciable effect 
 on the number of subsequent right cases appeared. E merely 
 changed the form of his report on correctness. E rarely re- 
 ported that a syllable came to mind and was rejected; he 
 usually thought about the correctness of a response only after 
 the response had been given. 
 
 In learning, E's attention was chiefly on the pronunciation of the 
 syllables, and he paid " much less attention to the way the syllables 
 looked or sounded." He had little difficulty in not thinking of the 
 series during the interval after learning until the experimenter said 
 " Get ready." Then, after E had adjusted the lip-key, he began to 
 ■'get ready to reply." The following is a characteristic account of 
 this period. " Strong strain sensations. Fixation of the screen. Vague 
 and fleeting visual images, — indistinguishable dark forms on a white 
 ground — a tendency to form the vocal organs." Sometimes during 
 this period images of actual syllables appeared, usually in 
 auditory or auditory-kinaesthetic form. With the experimenter's 
 " Ready," strains increase. The stimulus-syllable is nearly always re- 
 ported as having been read visually without internal speech. After 
 the perception of the stimulus-syllables, various processes are re- 
 ported. In a very few cases recognition of the stimulus-syllable oc- 
 curs, or some other reference is made to the place of the syllable in 
 the learning series. Sometimes a " blank " or a " mass of strain 
 sensations in the head " is reported. Most frequently, however, the 
 perception of the stimulus is at once followed by a " taggy " image 
 of the .reply-syllable or of a part of it. These taggy images are about 
 equally often visual and auditory-kinaesthetic. When these images 
 come, E at once starts to speak, developing the syllable " vocally " if 
 only part of it is given in image. Almost as frequently E reports 
 that a syllable or part of a syllable " forms itself mechanically upon 
 the tongue " and " says itself," or the syllable " shoots oflf^Epmati- 
 cally." Very often this mechanical reproduction is followed byfeE.^iai 
 or other imagery of the spoken syllable, which helps the obse^^ to 
 spell what he has spoken and to decide whether the syllable is a true 
 member of the series or not. Sometimes, in the period between suc- 
 cessive stimulus-syllables in the same test, E gets ready for the next 
 reaction as he does for the first reaction, and reports images of the 
 learned series or of parts of it; but usually the getting ready consists 
 merely in fixation of the screen and increased strain in the eye- 
 muscles and in the throat and chest. 
 
 D very frequently finds the nonsense syllables meaningful, especially 
 in the first few series. Certain syllables were felt to be like ' some- 
 German word," heil-gaam suggested a proper name, zusshaun 
 
ON THE PERSEVERATIVE TENDENCY 411 
 
 seemed like suss-schon, and so on. Certain syllables were more pleas- 
 ant than others. 
 
 D, like E, finds that most attention needs to be given to the pro- 
 nunciation of the syllables, but unlike E she finds that, after two or 
 three repetitions, she can pay attention to the " looks " of the sylla- 
 bles, can " run over " them visually, sometimes fixating certain letters 
 in turn, and can very frequently anticipate the coming syllable in 
 visual or kinaesthetic imagery. D has considerable difficulty in pre- 
 venting herself from thinking of the series during the interval, and 
 says she must " keep talking or else think about the experiment." The 
 thought of the series, when it does arise, is reported sometimes as 
 " a visual image of a white strip over there," sometimes as " a white 
 strip on the drum," sometimes as merely " a tendency to turn the 
 eyes towards the exposure-slit." Sometimes in these cases specific 
 syllables appear, in auditory-kinaesthetic or in kinaesthetic imagery. 
 When getting ready for the test, D reports a consciousness of the 
 Aufgabe, represented in a number of ways, at diflferent times. 
 Most frequently at first appear verbal images, such as " quickly " 
 and " get that syllable," often with " taggy visual images " of the 
 strip or of a part of the strip on which the syllables are written. 
 Sometimes these taggy images are all that is reported. Sometimes 
 no visual images appear, but D reports " kinaesthesis of eye-move- 
 ment to the left," or a " kinaesthetic attitude, which means ' get the 
 syllables.' " Quite often, with these representations of the Aufgabe, 
 D has auditory or auditory-kinaesthetic (rarely visual) images of 
 specific syllables. Strain sensations and fixation of the screen seem 
 to be less prominent with D than with E. 
 
 D generally reads the stimulus-word in internal speech, and some- 
 times repeats the word once or oftener. Immediately after this D 
 practically always realizes or attempts to realize the place which the 
 stimulus had in the learning series. This realization is characteris- 
 tically in terms of actual or imagined eye-movement, with or without 
 vague visual images of syllables or dark spots on a white background 
 to the left. " Impulse to turn eyes to the left and down," " hurried 
 eye-movement," " a quick dart of the eyes toward the upper part of 
 the series " are introspections of this period. After this kinaesthetic 
 or kinaesthetic and visual localization, in a large number of cases 
 " taggy " images of the reply-syllable appear in visual or auditory- 
 kinaesthetic form, and the reaction is immediately made. Sometimes, 
 however, as in the case of E, D reports that " the reaction-word said 
 itself as a muscular continuation of reading the stimulus," and some- 
 times that " there was effort to get something to say (throat-move- 
 ments)." Sometimes, when no imagery or imagery of only a part 
 of the reply-syllable appears, D replies " by starting to say something 
 and trusting the throat to finish it." 
 
 E's reactions, then, if we may be permitted to make the 
 comparison, seem to approach the abbreviated or muscular 
 type; D's the natural or central type. In order to get a more 
 complete and accurate description of all periods of the learn- 
 ing and of the reproduction consciousness, and especially of 
 the processes concerned in the reproduction of the incorrect 
 replies, we determined to favor quick reproduction by definite 
 instructions ^.r* increase the number of our observers, and to 
 
412 
 
 FOSTER 
 
 carry out a number of purely qualitative experiments before 
 and after the experiments of the main quantitative group. 
 The conditions in this series of experiments were identical 
 with those just reported, save that the four new observers 
 were tested after an interval of one minute instead of after 
 three niinutes, and save that here all observers were told to 
 place the chief emphasis upon reacting as quickly as possible. 
 They were still to try to give the right associate, but were 
 told not to discard any syllable which came to mind, unless 
 that syllable were felt to be incorrect before a reply could be 
 made. Besides the two original observers, D and E, there 
 took part in this work Dr. H. P. Weld (W), assistant profes- 
 sor of psychology; Mr. C. A. Ruckmich (R), instructor in 
 psychology; Mr. R. C. Holl (H.), and Mr. J. S. Johnston 
 (J), graduate students in psychology. W and J had had 
 previous experience in experiments on memory. Twelve series 
 were given to each observer. The results of the experiments 
 are shown in Table IV. 
 
 TABLE IV 
 3 Minute Interval 
 
 Obs. 
 
 r 
 
 Tf 
 
 f 
 
 Tf 
 
 V 
 
 Tv 
 
 D 
 
 ,22 
 
 1,360 
 
 .58 
 
 1,630 
 
 .11 
 
 2,430 
 
 
 E 
 
 .07 
 
 1,390 
 
 .70 
 
 700 
 
 .11 
 
 1,490 
 
 1 Minute Interval 
 
 w 
 
 .21 
 
 810 
 
 .52 
 
 1,430 
 
 .21 
 
 2,440 
 
 
 R 
 
 .29 
 
 750 
 
 .56 
 
 1,050 
 
 .03 
 
 2,570 
 
 
 H 
 
 .25 
 
 680 
 
 .39 
 
 1,020 
 
 .27 
 
 3,470 
 
 
 J 
 
 .06 
 
 590 
 
 .64 
 
 780 
 
 .29 
 
 3,150 
 
 
 After this group of experiments, the instructions were again 
 changed. The other conditions remained the same, for both 
 groups of observers, as in the previous experiments. The 
 observers were, however, now instructed to take as much 
 time for finding the correct unaccented syllable as seemed to 
 them likely to make for correctness, and not to reply with the 
 first syllable which came to mind unless they felt that it might 
 at least be correct. Table V gives the results of twelve series 
 with each of the observers. 
 
ON THE PERSEVERATIVE TENDENCY 
 
 413 
 
 TABLE V 
 3 Minute Interval 
 
 Obs. 
 
 r 
 
 Tr 
 
 f 
 
 Tf 
 
 V 
 
 Tv 
 
 D 
 
 .43 
 
 4,610 
 
 .42 
 
 6,600 
 
 .01 
 
 9,052 
 
 
 
 E 
 
 .25 
 .29 
 
 2,970 
 
 1 MiNUT 
 
 3,200 
 
 .58 
 
 E Interv 
 .41 
 
 11,410 
 
 AL 
 6,840 
 
 .08 
 .12 
 
 19,000 
 
 W 
 
 14,780 
 
 
 
 R 
 
 .24 
 
 1,130 
 
 .50 
 
 1,540 
 
 .03 
 
 4,560 
 
 
 
 H 
 
 .47 
 
 3,330 
 
 .39 
 
 4,770 
 
 .04 
 
 8,470 
 
 
 
 J 
 
 .19 
 
 5,610 
 
 .60 
 
 7,860 
 
 .14 
 
 18,880 
 
 
 
 As might perhaps be expected from the fact that he reacted 
 very rapidly in the experiments whose results are given in 
 Table III, E's results in Table IV do not differ greatly from 
 his results in that table. The number of right associates re- 
 mains approximately the same. There is a slight decrease 
 in the number of incorrect replies, with a corresponding in- 
 crease in the number of partly right syllables, due perhaps 
 partly to practice in learning. AH the reaction-times are 
 somewhat shortened, but Tf still remains less than Tr. With 
 D, however, the change is considerable. Her correct replies 
 are fewer, her incorrect replies are more, and her reaction- 
 times are much shortened. W, R, and H (Table IV) show 
 great similarity of results in the experiments under the in- 
 struction to put the emphasis on rapidity of reaction, both 
 with regard to percentages of right and wrong cases, and in 
 the length of reaction-times. These reaction-times are much 
 shorter than those of D, who gives approximately the same 
 r-percentage under the same instruction, but who is tested 
 after a longer interval.'' J, who had considerable difficulty in 
 reading the syllables at the rate of speed at which they were 
 presented, falls far below the other observers of his group in 
 respect to the number of correct replies. 
 
 With the instruction to put the emphasis on correctness 
 rather than on speed all of the observers except R, for whom 
 there is a small decrease (more than compensated by the 
 greater number of partly right responses), give more correct 
 replies than with the second instruction, and all save H, who 
 
 3' Cf. Muller and Pilzecker, op. cit., 47 ff. 
 
414 FOSTER 
 
 gives an equal number, give fewer incorrect replies. The re- 
 action times without exception show an increase in length. 
 
 Before and after the groups of experiments whose results 
 are shown in Tables IV and V, we carried out a number of 
 experiments identical with those of the main groups except 
 for the fact that the experiment was interrupted at a point 
 in its course known beforehand to the observer, e. g., after 
 the learning, or after the first or a following reaction; and 
 an introspection of the period immediately before the inter- 
 ruption was required. On the basis of these introspections, 
 as checked by the introspections given in the main groups, 
 the following attempt is made to characterize the conscious 
 processes of our observers during these periods. 
 
 In the learning period, with the instruction to react as quickly as 
 possible, D reports less attention to the visual aspect of the syllables 
 than with the first instruction, and more attention to pronouncing 
 them correctly and emphatically. " Tried to keep my throat set for 
 the first syllable of a pair until the second was in the field, making 
 the accompanying kinaesthesis for the pair continuous. There was, 
 too, something in the eye-kinaesthesis that made the two syllables of 
 a pair group together and separate more definitely from the next pair." 
 " Tried deliberately at first to keep attention on each syllable while 
 it remained in the field, but found that after a few repetitions my 
 throat tended to continue with the second syllable of the pair while 
 the first was still in the field. There was also incipient eye-movement 
 upward toward the coming syllable, and occasionally a hazy antici- 
 patory visual image. . . . Before the last presentation the second 
 syllables of all the pairs but one were correctly anticipated." The syl- 
 lables less frequently have meaning than before. E's introspections 
 correspond very closely with his introspections for the first group of 
 experiments. 
 
 For W learning was " almost entirely a vocal-motor affair." The 
 syllables were " read off " with clear enunciation and strong accentua- 
 tion. W, like D, also " goes from the final consonant of the first to 
 the beginning of the next syllable of a pair without a break in speech " 
 and tends to " make the syllables (of a pair) Hke one word." Mean- 
 ings for the syllables are infrequent and W tries to " inhibit " them. 
 
 R pays little attention to the visual side of the experience, but the 
 kinaesthetic and auditory sides are very clear. He reads the series 
 as a whole into one rhythm, accenting most strongly the first and 
 ninth inembers. His voice drops slowly till the sixth syllable, and then 
 rises till the eleventh, with a drop at the end. The whole series forms 
 " a unitary kinaesthetic-auditory complex." He anticipates the second 
 syllables of the pairs after a few repetitions kinaesthetically, later 
 auditorily also. The series differ greatly for R in their pleasantness 
 or unpleasantness, depending, he maintains, upon all three factors 
 of ease of pronunciation, sounds of the individual syllables and of 
 combinations of syllables, and visual appearance of certain syllables. 
 
 H pays attention chiefly to the pronunciation and to the visual ap- 
 pearance of the syllables; after the first few repetitions, more espe- 
 cially to the latter. He anticipates the syllables, for the most part 
 
ON THE PERSEVERATIVE TENDENCY ,415 
 
 in visual, but sometimes in other terms. The syllables very frequently 
 have sensible meaning for him. 
 
 J, as we have before noted, has trouble in reading the series as 
 fast as it is presented, and in the first or second presentations occa- 
 sionally falls behind and has to skip a syllable or a foot. After the 
 third repetition, however, he begins to pay attention to the looks of 
 the syllables. 
 
 Learning under the third instruction differs little from learning 
 under the second, save in the fact that attention for most observers, 
 but especially for D, E, W, and H, is now more than before upon the 
 visual side of the experience. 
 
 In the interval between learning and test, D and E report slightly 
 fewer cases in which a thought of the series occurs, than in the first 
 series of experiments. Of all our observers W has the greatest 
 trouble in keeping his mind off the syllables. He made a practice 
 of rising from his chair and walking about the room singing or whis- 
 tling until called back for the test. R had little trouble after the first 
 few experiments, and J never reported the appearance of a syllable 
 during the interval in any experiment. H reported syllables in about 
 as many cases as did D. 
 
 The preparation of D and E for the test in the later experiments 
 is very like that in the first group of experiments. D reports more 
 and stronger strain sensations for the second instruction than for the 
 third, and notes especially the strains in chest and throat, and stronger 
 fixation of the screen. Reference to the learning series often ap- 
 pears, both in the visual and kinaesthetic forms before described and 
 as a general kinaesthetic set, and with this reference syllables some- 
 times come up in imagery, especially in auditory-kinaesthetic terms. 
 Sometimes the getting ready is reported by E merely as a total bodily 
 set, strain sensations especially in face, chest and throat, strong fixa- 
 tion of the screen. 
 
 Visual images of the white strip of paper, and auditory-kinaesthetic 
 images of specific syllables are very frequent with W during the 
 period of getting ready for the test. 
 
 R frequently has auditory-kinaesthetic imagery of the series in this 
 period. Thus he reports " a hazy group of auditory imagery which 
 seemed to be a jumble of a number of the nonsense-syllables with 
 no definite order or rhythm. Then some of the syllables were imaged 
 (auditory-motor) together in a series, but I can't now tell which 
 ones." " A number of syllable combinations started to come and 
 seemed to crowd together in the focus of attention." 
 
 H usually waits passively for the stimulus to appear in the experi- 
 ments both of the second and of the third instruction. Sometimes 
 he reports " a visual field with syllables in it as vague, darker lines," 
 but usually he reports that " there is just an empty visual field." 
 
 J first "gets set muscularly for a quick reply." Then before him 
 there is a sort of " dark, hazy, slightly transparent background " in 
 which at times certain syllables of the series appear as darker forms 
 to the right of the slot in the apparatus. 
 
 E finds no difference in conscious content between reproduction 
 under the second and under the first instruction. D less frequently 
 reports repetition of the stimulus-syllable in internal speech ; but prac- 
 tically always, when the reply does not follow as a " muscular con- 
 tinuation of the stimulus," D reports actual or imagined eye-move- 
 ments to localize the stimulus in the visual scheme which represents 
 
416 FOSTER 
 
 the learning series. Very frequently there is a visual or auditory 
 image of the reply-syllable preceding or accompanying utterance. 
 
 W most frequently reports that the reply was " purely vocal-motor, 
 without anticipatory imagery." " The reply comes immediately to 
 my throat, or tongue, and is followed by visual imagery of the sylla- 
 ble said." R either reacts in vocal-motor fashion after visual, or visual 
 and auditory-kinaesthetic perception of the stimulus, or the reaction is 
 preceded by an image — usually kinaesthetic or auditory-kinaesthetic — 
 of the syllable or of a part of it. H also responds " automatically " 
 sometimes, but for the most part, he reports that the syllable is pres- 
 ent in hazy visual imagery before the reply is given. J always reports 
 that the reply-syllable, as a darker form in the visual background, 
 appears before a reply can be made. Under the third instruction the 
 reproduction-consciousness seems much the same as with the pre- 
 vious instruction in the cases of W, R, and J, save that the strain 
 sensations of the preparatory period are less intense and that the 
 processes in the reproduction-period run their course more slowly. 
 For D and H the visual imagery is clearer and more definite, and the 
 repjy-syllables are hardly ever given " automatically." Frequently the 
 first syllable which comes is discarded, and a search (either imagined 
 eye-movement or visual imagery) takes place. 
 
 When, as in our experiments, a recently learned series is 
 tested by the method of right associates, a large number of 
 the incorrect replies are found to be identical with the sylla- 
 bles of the learning series. Other replies correspond with 
 syllables of the learning series in respect to two letters. Table 
 VI shows the percentages of syllables of the first kind under 
 the rubric frr (false row-right) ; the percentages of syllables 
 of the second kind under the rubric frrp (p=part). The 
 totals of frr-j-frrp are shown in the fourth column, and the 
 percentage of the false cases which are either frr or frrp is 
 shown in the fifth column. 
 
 By far the greater part of the incorrect replies — on the 
 average about three-fourths of them — ^are actual or partial re- 
 productions of syllables of the learning series. The small 
 percentage in the case of R under the second instruction is 
 due to the fact that sometimes his reply in this group of 
 experiments was a syllable which, added to the stimulus sylla- 
 ble, made a sensible word. 
 
 Especially with the purpose of investigating the relations 
 of the incorrect responses, in their dependence upon the in- 
 terval between learning and reproduction, a group of experi- 
 ments was undertaken with two observers, W, and Miss 
 J. N. Curtis (C), scholar in psychology, who had had a good 
 deal of previous practice in experiments on memory. Two 
 sets of- experiments of twelve series each, the one with four 
 repetitions of the series and a test after one minute, and the 
 other with twenty-four repetitions and a test after twenty- 
 four hours, were made with C as observer. A set of series 
 
ON THE PERSEVERATIVE TENDENCY 
 TABLE VI 
 
 417 
 
 
 
 
 
 
 frr + frrp 
 
 Instr. 
 
 Obs. 
 
 frr 
 
 frrp 
 
 Total 
 
 f 
 
 Group I / D 
 
 .29 
 .20 
 
 .10 
 .30 
 
 .39 
 • .50 
 
 .93 
 .60 
 
 
 \ D 
 
 .22 
 
 .35 
 
 .57 
 
 .98 
 
 
 E 
 
 .21 
 
 .19 
 
 .40 
 
 .57 
 
 Group II 
 
 W 
 R 
 
 .33 
 .03 
 
 .08 
 .10 
 
 .41 
 .13 
 
 .79 
 .23 
 
 
 H 
 
 .31 
 
 .04 
 
 .35 
 
 .89 
 
 
 . J 
 
 .39 
 
 .13 
 
 .52 
 
 .81 
 
 
 D 
 
 .19 
 
 .19 
 
 ,38 
 
 .90 
 
 
 E 
 
 .29 
 
 .25 
 
 .54 
 
 .93 
 
 Group III... 
 
 W 
 R 
 
 .10 
 .09 
 
 .22 
 
 .24 
 
 .32 
 .33 
 
 .78 
 .66 
 
 
 H 
 
 .32 
 
 .04 
 
 .36 
 
 .92 
 
 
 J 
 
 .24 
 
 .14 
 
 .38 
 
 .63 
 
 identical with the latter was made with W, who was given 
 one test and afterwards learned one series on each experimen- 
 tal day. C did the same, with the addition of learning and re- 
 producing a series with the shorter intervals between the test 
 of one twenty-four-hour series and the learning of the next. 
 The instructions and the other conditions of learning were 
 the same as in our second group of experiments by the method 
 of right associates, where the emphasis was placed upon re- 
 acting as quickly as possible. The reaction-times were re- 
 corded in twenty-fifths of seconds, written by a time-marker 
 connected with the Kronecker interrupter, and are expressed 
 in Table VII in tenths of a second. 
 
 TABLE VII 
 
 
 r 
 
 Tr 
 
 f 
 
 Tf 
 
 V 
 
 Tv 
 
 w 
 
 24 hours 
 
 .02 
 
 9.0 
 
 .73 
 
 14.3 
 
 .16 
 
 25.8 
 
 C 
 
 24 hours 
 
 .23 
 
 5.7 
 
 .30 
 
 8.4 
 
 .26 
 
 11.2 
 
 C 
 
 1 minute 
 
 .49 
 
 1.6 
 
 .22 
 
 5.2 
 
 .17 
 
 7.9 
 
418 FOSTER 
 
 C explained that, on account of past experience in memory experi- 
 ments, where " ingenious " learning was encouraged, she found it 
 almost impossible to learn nonsense-syllables without attaching sensi- 
 ble meanings to them. She found the attempt to repress such means 
 of memorizing so disturbing that it did not seem advisable to instruct 
 her to do so. To her mode of learning is to be attributed the high 
 percentage of right cases, shown in her results both for the shorter 
 and longer intervals. She reported that most of the syllables of any 
 series had meaning for her in learning, and that these meanings often 
 linked sensibly the accented to the unaccented syllables of a pair. 
 In tests after the one-minute interval, however, these sensible mean- 
 ings often played no conscious part for C. In general, she reports, 
 first, " an auditory-kinaesthetic image of the stimulus syllable as it 
 sounded in my own voice." Then, if a kinaesthetic or auditory- 
 kinaesthetic image of the reply-syllable does not appear quickly and 
 easily, C " tries to think of the meaning." If still no image of the 
 second syllable appears, the stimulus-syllable is repeated in internal 
 speech a number of times. 
 
 In tests after twenty-four hours, C usually reports a perception of 
 the stimulus-syllable as a verbal meaning; then follow various sorts 
 of imagery; usually, however, predominantly visual. An auditory- 
 kinaesthetic image of a reply-syllable usually appears, but is frequently 
 modified in expression, though not in meaning, before it is given. 
 
 W learns " in vocal-motor fashion." In the test he repeats the 
 stimulus-syllable over and over, sometimes a reply-syllable is spoken 
 " as if a continuation of speaking the stimulus-syllable," but usually 
 after considerable repetition W decides " to say something " without 
 knowing what he is about to say. He rarely reports during the test 
 a reference to the learning series. 
 
 Both observers report cases in which syllables appear outside the 
 experimental hour; but both maintain that in all cases the syllable 
 appears in connection with a thought of the experiment or of the 
 laboratory, usually represented visually. 
 
 In our second group of experiments by the method of right 
 associates, when the test took place after a one-minute inter- 
 val, and where eight repetitions of the series were given, 63% 
 of the incorrect replies given by W were nevertheless syllables 
 of the series. In the tests after twenty-four hours, with 
 twenty-four repetitions, only 4% of his incorrect replies were 
 members of the series. In tests after one minute with four 
 repetitions, 28% of C's incorrect reactions were syllables be- 
 longing to the series, while in tests after twenty-four hours 
 with twenty-four repetitions, 20% of these reactions were 
 members of the series. 
 
 As we have seen in our historical discussion (p. 396) Miil- 
 ler and Pilzecker consider the fact that in general, in tests 
 of recently learned series, a greater number of the incorrect 
 syllables are identical with syllables of the learning series than 
 in tests of series after longer intervals, as an indication of 
 the greater strength of the perseverative tendencies in the first 
 case. If, instead of perseverative tendencies, we read readi- 
 
ON THE PERSEVERATIVE TENDENCY 419 
 
 ness of the syllables for reproduction, or strength of their 
 impression, the explanation seems as adequate. Certainly 
 when strong secondary associations are present, as in tests 
 which follow soon after learning, the free appearance of sylla- 
 bles cannot be supposed. And if, as is usually maintained,''^ 
 the secondary associations lose in strength relatively more 
 rapidly with time, then it is to be expected that the direct asso- 
 ciations will be those more effective after some time has 
 elapsed, and that frr cases will be the exception. 
 
 We have found numerous cases in which a particular syl- 
 lable is given as a reply several times in the same test 
 and in the tests on different days ; but the explanation can, 
 at least in many cases, be made without supposition of 
 free arousal. D, for instance, gave maan as an incorrect 
 reply in experiment 3 with the report : " Visual perception of 
 the stimulus ; eye-movement to place it in the series, but no 
 syllable came; eye-movements continue over place of series, 
 then suddenly visual image of mamt (the spelling of maan 
 had made a distinct impression on me in learning). Said it 
 at once." This is an instance of an frr syllable aroused through 
 a place association. In experiment 9 D gives the correct re- 
 ply maun with the report : " The stimulus-word hos made a 
 strong impression on me (in learning) as the first syllable of 
 a person's name, and the second syllable as meaning man was 
 connected with it; pleasant organic sensations. Now with 
 perception of hos and recognition of its place in the series, 
 in terms of eye-movement and visual image of the exposure 
 field, the same organics as before. Set my throat, a lot of 
 eye-movement; then maun came out in a great hurry." In 
 the experiments under instruction to react quickly, 13-24, 
 maan appears twice as an incorrect reply, mual three times, 
 m^n once and mmm . ., an uncompleted reply, once. In every 
 case D reports that after the stimulus-syllable nothing would 
 come; then occurred a hurried eye-movement over the place 
 of the series and a setting of the throat, which finally is re- 
 leased by utterance. In one case a visual and in one case 
 an auditory-kinaesthetic image of the syllable is reported as 
 preceding utterance. Not until experiment 24 did D realize 
 that a similar syllable had previously been given as a reply. 
 D, then, especially when no syllable will come visually, sets 
 her throat to say something. With this setting of the throat 
 the saying of a syllable beginning with m has become strongly 
 associated, and is now realized in the absence of other asso- 
 ciations. W gives as incorrect replies, under the instruction 
 
 38 Ebbinghaus, op. cit., p. 647. 
 
420 FOSTER 
 
 to reply quickly, neuf twice, neb twice, nes once, nef once, 
 nev once, nif once, and other syllables beginning with n and 
 m once each. Under the third instruction he gives mal three 
 times, ner twice, nur twice, neuf twice, and other syllables 
 such as nar, nal, once each. In practically all such cases W 
 reports strain in the throat, hesitation, or a blank, " I am 
 stuck " feeling. In the incorrect replies of the other observers 
 similar syllables are given when the observers are at a loss 
 for the reply and " set their throats to say something." 
 
 It is noteworthy that all the five habituelle Aushulfe- 
 silben nof, laf, noch, mm and meun, listed by Miiller and 
 Pilzecker (p. 62), begin with /, m, or n, the letters which we 
 found to occur so frequently as initial letters in our experi- 
 ments by the method of right associates, and in the " unin- 
 fluenced " reactions in the preliminary experiment. With the 
 method of right associates in 207 reactions, in which the in- 
 correct response did not correspond even in respect to two 
 letters with the syllable of the series, the frequency of the 
 initial letter was: n, 48; m, 35; 1, 22; t, 19; r, 13. The fre- 
 quency of the final letters was : m, 28 ; 1, 26 ; r, 26 ; f , 18 ; s, 14. 
 By chance, each letter should occur about ten times. If we 
 arrange the total number of letters, substituted for the correct 
 letters in the reactions where a syllable is given which corre- 
 sponds with a syllable in the series in respect to two letters, 
 we find the following relations of frequency. For 183 initial 
 letters : n, 32 ; t, 28 ; m, 25 ; 1, 20 ; d, 20. For 171 final letters : 
 r, 31 ; 1, 28 ; s, 27 ; m, 22 ; f , 20. By chance, each letter would occur 
 here about nine times. Observers seem to differ very little 
 from one another in respect to the letter most frequently given 
 in these cases, but the smallness of the number of cases pre- 
 vents us from drawing more definite conclusions. Certain ex- 
 periments without the lip-key and other experiments, in which 
 the Rousselot microphone was used, show that the form of 
 key or the absence of any key makes no appreciable difference 
 in the character of the response, so far as the predominance 
 of these letters is concerned. These same initial and final let- 
 ters appear frequently in the totally wrong cases in tests of 
 series of syllables in which the letters themselves do not ap- 
 pear, and even with observers unaccustomed to work with 
 nonsense syllables. It does not seem, then, as if these habi- 
 tuelle Aushulfesilben were necessarily recurrences of syllables 
 recently attended to.°° 
 
 '" The letters, which, from their frequency in the right cases, ap- 
 pear to have an especial attraction for attention, are not by any means 
 in our experiments the same letters as those appearing most frequently 
 in the Aushulfesilben. Cf. Muller and Pilzecker, op. cit., pp. 254 ff. 
 
ON THE PERSEVERATIVE TENDENCY 421 
 
 Experiments by the Method of Retained Members. 
 
 A series of experiments by the method of retained members 
 in which the method somewhat resembled that of our pre- 
 liminary experiments, was also performed. 
 
 ^yith the exceptions mentioned below all conditions of ex- 
 perimentation were the same as in the experiments by the 
 method of right associates. On each day a twelve-syllable 
 series of nonsense-syllables was presented for eight repetitions 
 and read by the observer precisely as in the method of right 
 associates (repeated series). After an interval of one minute 
 the observer gave as many of the syllables as he could recall 
 in 45 seconds, reacting with the lip-key. Alternately before 
 and after this series a second twelve-syllable series was pre- 
 sented and tested. In the successive presentations of this 
 second series (mixed series), as in our preliminary experi- 
 ments, the order of the syllables was so changed that no 
 single syllable was ever in the eight presentations followed by 
 the same syllable as in any previous presentation. The sylla- 
 bles were so arranged also that no syllable had any special 
 advantage over any other by reason of its position (first or 
 last in the presentation) or by reason of accentuation (first 
 or last in a pair), and two different arrangements were used 
 still further to decrease the chances of advantage of syllables 
 by place in the series. By this arrangement strong direct as- 
 sociations should be avoided. The syllables of this second 
 form were typewritten on a long strip of kymograph paper, 
 the eight arrangements following one another with the cus- 
 tomary two blank spaces between presentations. The actual 
 exposure was made by the Spindler and Hoyer apparatus, the 
 strip being rolled as a band over a second light drum at the 
 back of the apparatus. The times of the reactions were re- 
 corded graphically on a Ludwig-Baltzar kymograph, and the 
 actual times recorded alongside of these in twenty-fifths of a 
 second by a marker connected with the Kronecker inter- 
 rupter. After preliminary practice the observers found little 
 difficulty in using the lip-key for successive reactions, though 
 in a few cases the observer believed that the second reaction 
 may have been slightly delayed by the necessity of closing the 
 key after the first reaction. After a short preliminary prac- 
 tice, 12 experiments with each sort of syllable order were 
 made with all of the four observers W, R, H, and J. The 
 observers were told to give not only such syllables as they 
 were sure belonged to the learning series, but also other sylla- 
 bles which occurred to them, if they believed that these other 
 syllables might possibly be correct. 
 
422 
 
 FOSTER 
 TABLE VIII 
 
 Obs. 
 
 Total Reactions 
 
 Right Reactions 
 
 Incorrect Reactions 
 
 Rep.* 
 
 Mix.t 
 
 Rep.* 
 
 Mix.t 
 
 Rep.* 
 
 Mix.t 
 
 W 
 
 79 
 
 74 
 
 .51 
 
 .55 
 
 .18 
 
 .27 
 
 R 
 
 103 
 
 95 
 
 .46 
 
 .60 
 
 .32 
 
 .21 
 
 H 
 
 88 
 
 81 
 
 .83 
 
 .77 
 
 .05 
 
 .09 
 
 J 
 
 66 
 
 66 
 
 .59 
 
 .52 
 
 .06 
 
 .17 
 
 '■ Rep.= Repeated order. 
 
 tMlix.= Mixed order. 
 
 As is evident at once from the above Table the percentage 
 of correct repHes after an interval of one minute is not directly 
 dependent upon the fact that the direct associations are re- 
 peated and strengthened in learning, for the percentage of 
 right reactions is nearly as great (H, J), or even greater 
 (W, R.), when no such repetition occurs, as is the case in the 
 mixed order. The rather surprising fact that W and R give 
 slightly more correct syllables and that H and J give almost as 
 many correct syllables in the series with the mixed order, as in 
 the other series where the direct associations are repeated eight 
 times, is undoubtedly to be explained by strong indirect asso- 
 ciations and by strong impression. Three of the observers re- 
 port that the individual syllables in the mixed order stand out 
 more clearly in attention. In the mixed order the observers do 
 not know, when one syllable is in the field, what the next sylla- 
 ble will be, as they very often do in the repeated order. The 
 consequence of this unpreparedness is that the observer's atten- 
 tion to the presented syllable is always demanded in a high de- 
 gree. All observers report that more attention has to be paid to 
 the visual appearance of the syllables in the mixed order. That 
 direct associations are not of primary importance for recall 
 of a very recently learned series is also shown by the fact that 
 the usual order of recall of syllables with the repeated order 
 of presentation is not the order of the syllables in the learning 
 series. The first syllable to be recalled is two-thirds as often 
 the eleventh of the learning series as the first, and the other 
 syllables follow in very irregular order. Still another fact 
 that tends to show the same thing is that out of 48 cases ir 
 which a correct syllable is followed by a correct syllable in the 
 reproduction of the series with repeated order, R gives the 
 directly associated syllable only 11 times; W gives it in only 
 
ON THE PERSEVERATIVE TENDENCY 
 
 423 
 
 lo cases out of 49; H in only 21 cases out of 70; and J in 
 9 cases out of 48. The number of times in which the follow- 
 ing syllable in these cases is the syllable immediately preceding 
 the given syllable in learning for R is 9, for W is 4, for H 
 is II, and for J is 6. 
 
 The average length of time elapsing after the experimenter's 
 " Now " and before the observer's first, second, third reac- 
 tion and so on, is given in the following Table. No averages 
 are calculated for those places in which less than six reactions 
 appeared, and the times which were measured in twenty-fifths 
 are here expressed by the nearest tenth of a second. 
 
 TABLE IX 
 Observer W 
 
 Place 
 
 1 
 
 2 
 
 3 
 
 4 
 
 5 
 
 6 
 
 7 
 
 8 
 
 9 
 
 Time: rep 
 
 3.9 
 
 6.8 
 
 8.0 
 
 9.9 
 
 15.9 
 
 22.0 
 
 31.3 
 
 35.4 
 
 
 
 
 Time: mix 
 
 2.4 
 
 5.4 
 
 7.9 
 
 11.9 
 
 22,2 
 
 28.9 
 
 34.3 
 
 
 
 
 
 
 Observer R 
 
 Place 
 
 1 
 
 2 
 
 3 
 
 4 
 
 5 
 
 6 
 
 7 
 
 8 
 
 9 
 
 10 
 
 Time: rep 
 
 1.3 
 
 2.2 
 
 3.7 
 
 6.3 
 
 9.7 
 
 17.1 
 
 23.3 
 
 24.3 
 
 32.6 
 
 S'l A 
 
 
 
 
 1.2 
 
 2.6 
 
 4.6 
 
 6.9 
 
 12.1 
 
 17.4 
 
 26.1 
 
 34.7 
 
 39.8 
 
 
 
 
 Observer H 
 
 Place 
 
 1 
 
 2 
 
 3 
 
 4 
 
 5 
 
 6 
 
 7 
 
 8 
 
 9 
 
 
 1.9 
 
 4.1 
 
 7.4 
 
 11.2 
 
 17.1 
 
 18.2 
 
 20.1 
 
 26.7 
 
 W 6 
 
 
 
 
 1.7 
 
 3.9 
 
 5.8 
 
 9.4 
 
 18.1 
 
 23.9 
 
 27 4 
 
 28.6 
 
 
 
 
 
 Observer J 
 
 
 
 
 
 
 
 Place 
 
 1 
 
 2 
 
 3 
 
 4 
 
 5 
 
 6 
 
 7 
 
 8 
 
 9 
 
 
 2.9 
 
 5.3 
 
 7.9 
 
 9.2 
 
 10.6 
 
 14.2 
 
 26.5 
 
 
 
 
 
 
 
 2.1 
 
 3.5 
 
 5.9 
 
 9.5 
 
 12.8 
 
 20.3 
 
 31.3 
 
 
 
 
 
 
 In the recall of the series with repeated order, the observers 
 give the first or first few reactions later on the average, and 
 
424 FOSTER 
 
 the following reactions sooner on the average, than in the 
 series with mixed order. The difference is especially great 
 in the case of W, and can at least in his case be explained 
 on the basis of his introspections. " With the short [i. e., the 
 rep. order] series it takes time to readjust, to put myself into 
 the set of the learning series. I start to visualize the series." 
 "At 'now' very brief time in which I readjusted. Blank; 
 visual field dark, visual image of syllables." " This reproduc- 
 tion is very different from the reproduction in the long {i. e., 
 mixed order] series. There I was waiting for a release ; here 
 there is a real recall." In general, then, in the repeated series 
 W tries to get back into the learning situation and to visualize 
 the syllables. In the mixed series, however, the first syllables 
 follow in vocal-motor fashion a set of the throat. After the 
 first three or four syllables, usually comes a time during which 
 no syllables are given. W reports : " Relief, passivity, atten- 
 tion on nothing in particular." Then W tries to get the syl- 
 lables visually or repeats over in internal speech the syllables 
 already spoken. A similar state of affairs is also reported 
 by H. R reports that reproduction follows the " now " in 
 autditory-kinaesthetic terms in both cases; reproduction for J 
 is usually in visual terms. The shorter time of the first replies 
 in the mixed series may depend upon vocal-motor readiness, 
 due to higher attention. The increased time in the later re- 
 plies is explained by the fact that the direct and the place 
 associations in this form of series are weak. In the repeated 
 series place and direct associations continuously aid the ob- 
 servers. 
 
 The incorrect replies of this method are distributed almost 
 entirely in the last part of the recall-period. A large number 
 of them are very similar, and resemble the hdbituelle Aushiilfe- 
 silben of the method of right associates. With W, for in- 
 stance, nev is given as an incorrect reply five times before 
 the series of experiments is two-thirds finished, and not until 
 that time does he realize that it has been given before. When 
 the fact is noticed, he remarks : " I think this syllable has come 
 up a number of times before. I don't think it is in to-day's 
 series at all. It came when I was balked ; I felt helpless, my 
 throat was tense, I was waiting for something to come, and 
 my face and eyes were screwed up. Feeling of emptiness. 
 Determined to say something and nev was uttered. Then I 
 saw it in type, not black, but reddish." After this report the 
 syllable did not appear again in the series. R and J also 
 report that such syllables are uttered after the throat is set in 
 an attempt to say something. 
 
ON THE PERSEVERATIVE TENDENCY 425 
 
 IV. Conclusion 
 
 In the historical section of this paper, we pointed out that 
 the concept of perseveration has acquired a wide variety of 
 meanings. Perseveration may be the designation of at least 
 four psychologically distinguishable states of affairs, (i) It 
 may mean that a content-precess has a tendency to run a 
 temporal course, to have a continued existence. Since this 
 meaning is implied in the very term " process," it seems use- 
 less to adopt a special term for designating it. (2) Persevera- 
 tion may mean that a content-process, having once appeared 
 in consciousness, does not at once disappear from conscious- 
 ness, but stays or remains in consciousness for a certain time 
 without reference to a change in its durative character.*" 
 (3) Perseveration may mean that a given mental formation 
 which has once been in consciousness is likely to appear in 
 consciousness again. (4) Perseveration may mean that, if a 
 given mental formation has once been in consciousness, its 
 recurrence need not be attributed to any condition save this 
 of having once been in consciousness. 
 
 It is the factual justification for this latter concept that our 
 experimental work has called into question. In our experi- 
 ments we have met no facts which seem to make necessary 
 the supposition of a tendency for an idea to arise without the 
 probable cooperation of associative tendencies. When a syl- 
 lable-idea came to the mind of our observers, in the interval 
 between the learning and testing of series, or in the intervals 
 outside of the experimental hour, the observers reported that 
 such an idea followed or accompanied an idea of a larger 
 situation of which it formed a part. In the " natural " rhym- 
 ing of syllables, which occurs when the problem of relatively 
 unrestricted speaking of a list of nonsense-syllable is given, our 
 observers find cases in which the previous sound is still in 
 consciousness when the new syllable arises. We have main- 
 tained that the syllables which are habitually given as reac- 
 tions by observers, in the method of right associates, when 
 the observer is at a loss for vocal-motor expression, are in 
 many cases not syllables which have been previously learned, 
 but are rather natural reactions, due merely to the intention 
 to utter some nonsense syllable ; and we believe that the re- 
 currence of such replies may be sometimes partly explained by 
 an association of the given motor impulses to the conscious 
 processes characteristic of the state. The greater number of 
 false cases in the tests of recently learned syllables is explica- 
 
 40 Jesinghaus, op. cit., 366. 
 
426 FOSTER 
 
 ble, we believe, by the assumption that indirect or remote as- 
 sociations lose relatively much more rapidly in strength than 
 do direct assiciations, and that, as more and more repeti- 
 tions are given to a series, the repetitions become more and 
 more effective for the direct and less and less effective for the 
 indirect associations within the series, through the partial 
 inhibition of the latter by the former. In a series, there- 
 fore, in which direct associations are equal to those in another 
 more recently learned series through the advantage of more 
 frequent repetition, the remote associations will be much 
 weaker, and in it the number of false cases which are syllables 
 of the series will be less. 
 
 The shorter reaction-times of right syllables in recently 
 learned series, whose percentage of right cases is equal to 
 that of another less recently learned series, may be explained 
 as readily on the basis of strong impressional tendencies or 
 greater readiness of the syllables, as by postulating persevera- 
 tive tendencies. 
 
 The non-experimental observations which have been urged 
 in support of the theory of a free arousal of ideas or feelings 
 and the continuance or recurrence of a given form of mental 
 activity, may be explained, on the basis of some or all of the 
 following factors ; unnoticed motives to association, such, for 
 instance, as kinaesthetic or organic sensations; strong impres- 
 sion or determination ; remote associations ; place associations, 
 and the like. Many alleged instances of perseveration may 
 be explained as the mere continuance of the given process 
 instead of its re-arousal. 
 
 Finally, conditions which are supposed to favor persevera- 
 tion do not always bring about a reproduction. We may have 
 been recently awakened from a vivid dream, and we may try 
 in vain to remember the ideas of the state, whereas an idea 
 associated to a dream idea will often accidentally bring up 
 that idea without our endeavor to remember. The stimulus 
 associated to an action in the patient's mind by the operator in 
 hypnosis may touch oflF the suggested but forgotten action at 
 a later time, even though the action be foolish. 
 
 Not only may the necessity of supposing a spontaneous 
 recurrence of mental contents be questioned, but a spontane- 
 ous activity in the nervous system, which seems thereby to 
 be implied, is also made improbable by physiological facts so 
 far known .*^ 
 
 ■'I See Wundt, op. cit., 572 i.; Jesinghaus, loc. cit; Poppelreuter, 
 Arch. f. d. ges. Psych., 25, 1912, 293 ff. ; Titchener, loc. cit. 
 
Cornell University Library 
 BF390 .F75 1914 
 
 On the perseverative tendency, by Willia 
 
 olin 
 
 3 1924 029 039 075