Chapter XXIII: Section IV: , 1. The experiments of Table IX, C, repeat those of A with (3)

1 - 10 72 20 88 133 5 6 5 33 2 0 1 6 0 1 0 2 0
11 - 20 114 119 109 94 33 29 40 83 7 5 15 25 2 2 7 10 0
21 - 30 50 87 54 29 70 57 73 74 16 13 25 52 4 1 3 27 0
31 - 40 17 17 10 6 51 87 53 30 24 21 24 30 2 1 11 14 0
41 - 50 10 14 7 4 40 35 39 20 35 18 37 32 14 6 17 34 2
51 - 60 2 10 3 1 43 30 40 14 48 44 47 53 30 18 35 36 5
61 - 70 1 3 0 3 11 7 5 2 28 50 20 21 8 11 23 22 18
71 - 80 3 2 1 2 5 8 2 10 29 32 23 14 14 21 20 14 57
81 - 90 0 0 1 1 9 3 6 2 27 29 28 14 41 35 35 19 67
91 -100 0 2 0 1 4 2 1 16 11 15 5 25 35 17 11 50
101 -110 0 0 0 1 1 1 4 7 6 0 9 24 7 6 25
111 -120 0 0 2 3 6 3 17 20 10 8 52 38 33 16 19
121 -130 2 0 0 1 0 3 4 3 3 4 15 6 8 13
131 -140 0 1 1 0 1 2 1 4 6 9 6 6 5
141 -150 0 1 0 1 5 4 4 2 15 13 16 7 1
151 -160 1 1 0 0 0 4 2 0 3 7 2 2 2
161 -170 0 0 0 1 1 0 0 3 2 0 1
171 -180 1 0 0 7 4 7 4 22 14 16 22
181 -190 0 0 1 0 1 0 2 2 3 0
191 -200 0 0 1 1 3 0 3 5 6 6
201 -210 1 0 0 0 0 0 1 0 1 0
211 -220 0 0 1 0 0 1 2 0 0
221 -230 0 0 1 0 0 1 0 6 2
231 -240 0 2 1 0 1 3 3 2 4
241 -250 0 1 0 0 3 0 0
251 -260 0 0 2 0 0 1
261 -270 0 0 1 0 0 0
271 -280 0 0 0 0 0 0
281 -290 0 0 0 0 0 0
291 -300 1 2 4 2 1 1
301 -310 0 0 0 0
311 -320 1 0 0 0
321 -330 0 0 0 0
331 -340 1 0 0 1
341 -350 0 0
351 -360 1 1
361 -370 0
371 -380 0
381 -390 0
391 -400 2

Totals
271 274 274 273 274 273 274 273 274 274 274 274 274 273 274 272 264

Tables 7 and 8 show that the range of the judgments increases with the length of the interval judged, and that the modal class is always much nearer the lower than the upper limit. Asymmetry is characteristic of the distribution of organic data, and in certain instances, as for example writing 18 seconds, males, the choice of a 10-second class interval results in extreme asymmetry, and one is reminded of the tables which Fechner gave as examples of his logarithmic method in statistics.[130]

It is not to be expected that a method of grouping should be found which will give regularity of distribution throughout, but it is important that there should be regularity about the mode. In the table of distribution for the males (Table 7) all the intervals from idleness 18 seconds to reading 72 seconds are regular.[131] The remaining intervals, with the exception of estimating 108 seconds, are irregular.

Trial proves that for these intervals increase of the size of the class to 30 seconds is sufficient to give regular distributions, as is obvious from Table 9. Grouping by 30-second classes gives regularity for most of the female judgments, but for idleness 108 seconds and writing 108 seconds there are still slight irregularities, as Table 10 indicates.

Tables 7 and 8 show that the distribution is far less regular for the females than for the males. The fact that it becomes regular when the class interval is increased to 30 seconds suggests that the irregularities of distribution which appear in the tables are due to those influences which favor simple fractions of a minute and not to the small number of judgments.

Having now noted certain important characteristics of the time-estimation judgments and the nature of their distribution, we may examine the arithmetical means and other statistical quantities which have been determined for our data. Those quantities which have been determined for the several ages, intervals, and fillings as well as for the sexes are: (1) The Mean (_M._ in tables), (2) the average variability (_M. V._), (3) the positive variability (+ _V._), (4) the number of judgments with positive variation (No. + _V._), (5) the negative variability (- _V._), (6) the number of judgments with negative variation (No. - _V._), (7) the relative variability (_R. V._) = _M_ + _V_/_M_ × 100.

Since the sums of the positive and the negative variations are equal, it is possible to make certain of the accuracy of the means and average variabilities by comparison of the + _V._ and - _V._ As this was done in all cases we feel confident of the reliability of the statistical quantities presented in the tables.

In Table 11 have been arranged the various quantities as determined for the males and females for each interval. The values given in this table are averages of the determinations made for the several ages separately.

TABLE 9

DISTRIBUTION OF CERTAIN MALE JUDGMENTS IN 30̋ CLASSES

_Classes._ W. 72̋ I. 108̋ R. 108̋ W. 108̋

1- 30 92 6 7 30
31- 60 115 61 52 91
61- 90 35 90 100 68
91-120 6 62 59 39
121-150 1 16 20 12
151-180 1 9 7 6
181-210 3 5 3
211-240 3 1 0
241-270 1 1
271-300 1

250 251 251 251

TABLE 10

DISTRIBUTION OF FEMALE JUDGMENTS IN 30̋ CLASSES

18̋ 36̋ 72̋ 108̋
_Classes_ I E R W I E R W I E R W I E R W
1- 30 236 226 251 256 108 92 118 190 25 18 41 83 6 4 10 39
31- 60 29 41 20 11 134 152 132 64 107 83 108 115 46 25 63 84
61- 90 4 5 2 6 25 18 13 14 84 111 71 49 63 67 78 55
91-120 0 2 0 3 8 9 5 37 38 31 13 86 97 57 33
121-150 2 0 2 2 1 9 10 8 9 25 37 28 21
151-180 1 2 0 0 8 9 9 4 28 23 18 25
181-210 1 0 2 1 4 0 6 7 10 6
211-240 0 2 3 0 1 5 5 8 6
241-270 0 1 0 3 3 0 1
271-300 1 2 4 2 1 1
301-330 1 0 0 0
331-360 1 1 1 1
361-390 0
391-420 2

Totals 271 274 274 273 274 273 274 273 274 274 274 274 274 273 274 272

TABLE 11

MEANS, ETC., FOR EACH SEX, INTERVAL AND FILLING

KEY: Ma. = _Males_
Fe. = _Females_

M. M. V. +V. No. +V.
Ma. Fe. Ma. Fe. Ma. Fe. Ma. Fe.

{I 17.7 20.74 5.4 10.35 5.8 15.48 15.9 22.75
18" {E 19.5 25.55 4.9 9.84 6.0 13.98 16.9 23.75
{R 15.5 18.46 4.9 9.14 6.3 13.03 14.6 24.25
{W 11.5 15.58 3.7 8.56 4.6 11.83 14.3 25.75

{I 33.3 42.81 9.1 16.55 12.6 20.63 14.2 28.00
36" {E 33.1 41.54 8.4 15.23 10.7 23.34 14.2 23.00
{R 32.1 41.71 8.4 16.42 9.7 20.78 16.3 27.75
{W 24.7 30.10 9.0 14.71 10.9 22.08 14.4 23.50

{I 63.3 73.00 17.2 27.20 20.9 33.72 15.7 27.75
72" {E 63.1 77.13 16.0 26.56 18.8 37.27 16.9 24.75
{R 57.9 70.78 17.3 30.30 20.9 39.28 15.2 26.75
{W 51.2 54.93 19.8 24.21 23.7 29.47 15.0 38.25

{I 92.7 113.37 29.8 40.13 35.3 44.08 15.2 32.00
108"{E 99.8 114.88 26.3 36.38 29.3 49.26 14.9 26.25
{R 90.1 100.47 28.3 40.18 33.9 47.90 15.3 34.25
{W 75.5 87.45 32.4 45.33 40.8 61.12 14.9 34.25

-V. No. -V. R. V.
Ma. Fe. Ma. Fe. Ma. Fe.

{I 5.7 7.78 20.0 45.00 30 49.83
18" {E 4.4 7.69 19.0 44.75 25 39.10
{R 4.4 7.07 21.2 44.25 31 49.26
{W 3.2 7.14 21.6 42.50 33 54.55

{I 8.1 14.01 21.7 40.50 27 38.32
36" {E 7.2 11.57 21.7 45.25 25 36.66
{R 7.5 13.84 19.6 40.75 26 38.84
{W 8.7 11.24 21.5 44.75 36 48.64

{I 15.5 22.83 20.2 40.75 27 36.87
72" {E 14.7 20.98 19.0 43.75 27 34.43
{R 15.7 24.91 20.7 41.75 30 42.77
{W 16.0 21.26 20.9 36.75 37 43.84

{I 25.9 37.87 20.7 36.50 32 35.34
108"{E 24.4 29.51 19.9 42.00 26 31.57
{R 24.8 34.60 20.6 39.75 31 39.95
{W 27.5 35.86 19.9 42.75 42 51.67

The following facts are revealed by comparison of the results for the two sexes.[132] Without exception the means for the females are larger than those for the males. All but one of the sixteen intervals (E 18̋) are underestimated by the males, whereas all but six are overestimated by the females. The amount of over- and under-estimation is given in Table 12. In every instance the females overestimate in comparison with the males. The mean variability is very much greater for the females, as is also the relative variability. If variability be taken as a measure of reliability of judgment the males are far superior to the females.

As is obvious from Table 12, both under- and over-estimation increase with increase in the length of the interval. For 18̋ intervals they are least, for 108" intervals greatest.

The influence of the fillings is marked. The writing intervals without exception are judged as shortest; reading gives the next shortest intervals, while sometimes idleness, sometimes estimating, comes third. In order of increasing length of average estimates of the intervals the fillings stand: writing, reading, idleness, estimating. As a rule the averages for the idleness and the estimating intervals are nearly the same, but it is worthy of note that the females always overestimate to a greater extent when estimating than when idle. This is another indication of the discrepancy between the female time-unit and the objective unit.

About ninety per cent of the subjects estimated by some counting method. The methods most frequently used were "counting seconds," counting "1 and 2 and 3 and 4, etc.," counting the swings of a pendulum, tapping, and counting imaginary watch-ticks.

The above statements might suggest that the overestimation characteristic of the female judgments is due to a more rapid counting rhythm. This, however, is not true, for the letter-counting tests indicate a slightly more rapid rhythm for the males, 93.42 as opposed to 91.89.

TABLE 12

AMOUNT OF OVER- AND UNDER-ESTIMATION OF INTERVALS

_Males._ _Females._ _Males._
{I - 0.3 + 2.74 Age 17 18 19 20 21 22 23
{E + 1.5 + 7.55 Over- or under- -8.55
18̋ {R - 2.5 + 0.46 estimation -3.60 -7.50 -13.02 -13.49 -8.71 -13.20
{W - 6.5 - 2.42 No. + 6 2 2 0 2 2 2

{I - 2.7 + 6.81 No. - 10 14 14 16 14 14 14
{E - 2.9 + 5.54
36̋ {R - 3.9 + 5.71
{W -11.3 - 5.90 _Females._

{I - 8.7 + 1.00
{E - 8.9 + 5.13 Age 17 18 19 20
72̋ {R -18.1 - 1.22 Over- or under-
{W -24.8 -17.07 estimation + 2.38 - 4.16 - 2.57 + 2.55

{I -15.3 + 5.37 No. + 11 8 8 12
{E - 8.2 + 6.88 No. - 5 8 8 4
108̋{R -17.9 - 7.53 + = overestimation
{W -32.5 -20.55 - = underestimation

In Table 13 are the means, variabilities, errors, etc., for the letter-counting tests. In this table we have presented the values of the various statistical quantities for the several ages of both males and females, for there are certain interesting differences which should be noted. Similar age-tables have been prepared for all of the other results, but in no case have noteworthy differences appeared. From Table 13 it will be observed that the males on the average count more letters in thirty seconds than do the females, and at the same time make more errors. The mean and relative variabilities for the sex-groups are almost the same. Curiously enough the number of letters counted as well as the accuracy of counting decrease for the females with age (17 to 20 years being the age-limits of the group under consideration). The males within the same age-limits increase in rapidity of counting, but decrease in accuracy. In the examination of Table 13 it is to be remembered that the 17 and 23 year groups of males contain only 16 individuals each, and therefore cannot be compared to advantage with the other groups. Both mean and relative variabilities decrease from 17 to 20 years for the females, whereas for the males there is no constancy in the direction of the change.

TABLE 13

MEANS, ETC., FOR LETTER-COUNTING

_Males._

_Age_ 17 18 19 20 21 22 23 _Average_
M. 92.81 92.48 94.00 95.87 92.66 97.06 89.06 93.42
M. V. 10.29 15.31 16.50 12.90 7.81 13.04 19.20 13.58
+V. 11.76 22.96 16.50 11.70 8.13 16.30 21.94 15.61
No. +V. 7 9 20 37 24 14 7 16.96
-V. 9.15 11.48 16.50 14.40 7.51 10.87 17.09 12.57
No. -V. 9 18 20 30 26 21 9 19.00
R. V. 11.1 16.5 17.6 13.5 8.4 13.4 21.6 14.59
Errors 0.88 1.04 2.10 3.03 1.70 0.89 2.75 1.77

_Females._

_Age_ 17 18 19 20 _Average_
M. 97.74 95.51 89.21 85.09 91.89
M. V. 17.06 14.56 11.30 11.28 13.55
+V 18.67 18.49 12.06 14.56 15.94
No. +V. 32 21 30 31 28.50
-V. 15.71 12.14 11.66 9.21 12.18
No. -V. 38 32 31 49 37.50
R. V. 17.5 15.2 12.7 13.3 14.70
Errors 1.35 1.46 1.49 1.65 1.47

SUMMARY

(1) The length of a second is slightly overestimated by men, greatly underestimated by women.

(2) Intervals of from 18 to 108 seconds are usually slightly underestimated by men (ages 17 to 23 years), and greatly overestimated by women (ages 17 to 20 years).

(3) The time-estimates of women are far more variable than those of men, and on the whole markedly less accurate.

(4) Both men and women favor estimates which end in 0 or 5, as well as simple fractions of a minute, but the tendency is stronger in women than in men. Over one third of the estimates reported in this paper were 15 seconds or simple multiples thereof.

(5) In letter-counting the groups of subjects studied (251 men and 274 women) exhibited differences just the opposite of those in time-estimation, for the men counted more rapidly and less accurately than did the women.

(6) Of the four fillings for the intervals used in the experiments, "writing" gave the smallest estimates of the intervals, listening to "reading" next, while "idleness" and "estimating" were conditions in which the intervals seemed much longer to both men and women.

(7) This preliminary study of sex-differences in time-estimation, by which it has been learned that women overestimate and are notably inaccurate in comparison with men, is to be followed and supplemented by the results of an investigation now in progress concerning the relations of sex-differences in time-estimation to age and physiological rhythms.

FOOTNOTES:

[Footnote 125: Science, N.S., vol. 19, pp. 708-709, 1904.]

[Footnote 126: E. Mach (Analyse der Empfindungen, 1900, p. 161) thinks that his physiological time-unit has become larger with age. And he has also noted that the time-sense differs in animals of the same species which differ in size.]

[Footnote 127: The unbracketed number is that for the males, the bracketed that for the females.]

[Footnote 128: It may be asked in connection with the above statistics, what is the probability that even and uneven digits have unequal chances of occurrence? The analytic treatment of this problem indicates that the probability is ½[1+φ(4.25)] = 0.99999999891 for the males and ½[1+φ(15.93)] for the females. It is therefore almost certain, theoretically, that even and uneven numbers have not the same frequency. See Czuber, Wahrscheinlichkeitsrechnung, 1903, pp. 158-161. The constants of this problem are in Czuber's denotation, _a_ = 52.308 and _b_ = 222.]

[Footnote 129: The judgments might be grouped in 10 second classes beginning with the lowest number in the experiments, but this would have the disadvantage of rendering the distribution tables or curves for different groups of judgments incomparable.]

[Footnote 130: G. Th. Fechner, Collectivmaasslehre, pp. 338-346, 1897. Compare the asymmetry of writing 18 seconds for both males and females with Fechner's Table III, p. 442.]

[Footnote 131: The distribution may be called regular if the classes increase to a maximum, the mode, and then decrease without interruption. (See Fechner, Collectivmaasslehre, pp. 121-122.)]

[Footnote 132: The values of the means and variabilities was determined to the second decimal for the females, but only to the first for the males; consequently the relative variability could be computed exactly to the first decimal for the females and to the unit only for the males.]

ASSOCIATIONS UNDER THE INFLUENCE OF DIFFERENT IDEAS

BY BIRD T. BALDWIN

The purpose of the following investigation was to study the influence of two or more starting-points on the train of associated ideas. It was begun in October, 1902, and concluded in January, 1905. Nineteen graduate students acted as subjects, and the experiments were conducted with each individually for one hour per week, except in a few instances where two subjects were present together. Occasionally the experimenter acted as subject in order to get a clearer insight into introspective data. There are recorded here thirteen groups of one hundred and eight sections, including eight hundred and fifty-five separate experiments, with a sum total of eleven thousand, one hundred and thirty-five named associations.

So far the field has not been studied experimentally, although Cordes,[133] while attempting to observe the effect of an unconscious intermediating factor in 'mediate association,' noticed that the accompanying factor (particularly with a tone) sometimes combined with the starting-point in determining the associated series. The fact is simply mentioned, and Scripture's[134] conclusions from the same are inadequate when he states, "In general we may say that two simultaneous ideas have an effect that depends on their relative masses; if one of the ideas is over-poweringly weighty the next idea will be chiefly influenced by it, but if the two are nearly balanced the next idea will be the result of the two." Miss Calkins[135] makes a near approach to the problem in a study of "Mental Combination," where auditory words were given as nearly simultaneously as possible and the subject was required to remain for four seconds in silence and then to write the train of imagery which passed in six following seconds; the words as auditory starting-points were finally excluded from the experiments "because the first word pronounced tends often to establish itself so firmly that its association-images are proof against the intrusion of the second word, which has therefore no chance to be grasped with the first." These results differ from the ones here recorded which were obtained independently and under different conditions. In her valuable monograph on Association[136] there is in one instance a suggestion of the problem, but no results are given.

The two or more starting-points were nonsense syllables, concrete or abstract words, pronounced by the investigator or shown on cards, or finally, pictures. The associations continued for a period of fifty seconds unless otherwise indicated. The same letters denote the same subjects throughout all the experiments, K. and R. being women. In the examples given, Roman numbers have been used, in place of the letters, as it was found impossible to get representative series which were entirely devoid of personal references.

After the associations had been "jotted down," the subjects in each case copied them, noting also the suggesting ideas, and giving, in many cases, a number of introspective notes. It is important to mention that in all of the work these records of suggesting ideas were made by the subjects without any questions on the part of the investigator.

GROUP I. TWO NONSENSE SYLLABLES

In order that a standard might be obtained that would serve as a basis for subsequent experiments, two nonsense syllables were pronounced with equal emphasis, as starting-points, and the subject was asked merely to wait until both had entered consciousness and not to favor nor inhibit either. If the subject stated that associations arose between the pronunciation of the starting-points the results were discarded.

Turning to the notes we find that K. reported, "The first just hovered around," and the associations followed the second, or as V. states it, "Sof was carried subconsciously for a while without influence," while By. frequently noticed that "The first may be in the background on the point of breaking in but be inhibited involuntarily," or in another, "It was frequently present but exerting no influence." A. "forgot what the other word was, but felt it trying to get a hand in." Ro. goes further and states in terms similar to A., "The words which might have been suggested by 'taf' were wholly inhibited, though I am sure 'taf' was present to me throughout the series, but was not efficacious as against the other syllable." It is unnecessary to multiply instances to show that the wraith, so to speak, of one may linger for a short time, or occasionally for the entire series, but that the associations are determined by the other. Or the lingering wraith of a starting-point may hover ineffectively in consciousness and exert no determining influence on the series until toward its close; then, however, although both starting-points still persist, they exchange places so far as effectiveness is concerned, and the former wraith becomes the primary factor in effecting the series of associations. For example, "The two syllables were present all the way through, but the former exhausted itself as a word-suggester after a few words." Again Ro. gives examples where "One alone may be in consciousness and determine the associations when the other may enter without effect upon the series, or may determine the series, the first persisting but losing its effect"; and a case of both ideas persisting and each alternately exerting an influence was noted by M. These last two conditions are exceptional.

Studying carefully the notes and graphic representations as given in the tables, it is apparent that one starting-point may be followed independently and exclusively; we find "'fef' suggesting 'theft,' and the subsequent associations monopolizing the field of consciousness and 'tuz' not again appearing." (V.) The results show that in the one hundred and seven experiments one was followed exclusively only twenty times. The following would be an example:

III. Naf-Tam

--tambour tam
--experiment tambour
--psychology experiment
--laboratory experiment
--space psychology
--time space
--ego space and time
--Ebbinghaus ego
--discussion psychology
--posted notice psychology
--door posted notice
--gray door
--stairs door
--fishes laboratory
--water fishes
--decay fishes
--animal fishes
--meat fishes
--odor meat

In other cases one starting-point may be followed and then the other for a few or many associations, each "occupying the mind to the exclusion of the other for a number of words, when suddenly the other appears and suggests a new series," which V. calls "a curious zigzagging in consciousness." This may take place at any point along the series. As Ro. indicates, "There was rivalry here; 'yud' wholly inhibited all the 'zid' associations, even 'zid' itself for the time, then, 'yud' apparently being exhausted, 'zid' entered again into the focus."

In contrast to the above forms, the tables indicate and the notes verify that while the associations primarily follow one there can be traced the modifying effect of the other; "The nature of the words was influenced by a fringe of consciousness which contains the other starting-point," M.'s comment runs, while J. adds, "The other helped in the mental picture but had no special significance." There are no experiments in the first group where this partial fusion took place throughout the series, and there are only six examples where it took place before the sixth word and continued to the end. While there are no cases here of alternating partial fusion, yet it sometimes happens, that is, the predominating influence is sometimes transferred from one to the other. The modifying influence may reach such a degree at any point in the series that the identity of each point of departure of the association is lost and there is total fusion, resulting in one idea or one train of thought. For example, 'rel' and 'mem' gave 'realm' for V.; for Ht. 'fef' and 'tuz' gave 'fezz'; for M. they gave 'fez'; for Ro. 'fuz.'

Of the twenty-one experiments where one starting-point was followed exclusively, there were two cases for the first and nineteen for the second. There were but two cases of alternating groups, a third one being unique in that after the first three words there was a persistent alternation between two apparently independent series of inter-related words and inter-related visual images. There were three cases of complete fusion throughout the series and nine where total fusion started before the fifth word and continued.

In a number of cases the initial starting-points were forgotten before the end of the series and the subjects were sometimes unable to recall them.

As the subjects were kept ignorant of the characteristic persistence of either starting-point in forming separate associations, and of the tendency towards fusion, as well as of the preponderance of influence of one over the other, it is to the tabulated results that we must look for quantitative measurement of these. 76.2% of the associations are due to the independent influence of one or the other of the starting-points; 23.9% are due to the combined influence; for the first starting-point we have 18.7%, for the second 57.5%. In 15.3% of all the associations the combined influence was the result of total fusion; in 4.1%, the result of a union where the first starting-point predominated in influence; in 4.4%, the second predominated. The following graphic representation indicates the order of influences for the first pair of syllables. The table gives the results for the ten pairs.

TABLE I. TWO NONSENSE SYLLABLES SPOKEN

Time--50 seconds.

Characters: | represents the influence of the first, - the second, +
complete fusion, α fusion with the first predominating, β fusion with
the second predominating.

1) Taf--Coz

M. | - - - - | | | | | - - - -
F. - - | | + + + + | | | + + β β β β β β
H. - | | | | | | | | | | | | | | |
R. | - - - - - - - - - - - - - - -
K. - - - - - - -
V. | | - - - - | - - | + + + +
Ro. | | | - - - - - - - - - - - - - - - -
Bl. - - - - - - - - - - - - - - -
By. | - | | | - - - - - - - - - - - - - - - - - - -
Bs. - - - - - - - - - - - - - - -
Ht. - - | | | | | | | | | | | | |

Starting-points:

(1) Taf--Coz.
(2) Cim--Bef.
(3) Yud--Zid.
(4) Sof--Deb.
(5) Naf--Tam.
(6) Fef--Tuz.
(7) Sar--Nef.
(8) Sek--Lub.
(9) Hov--Bes.
(10) Rel--Mem.

| - + α β

M. { 20 69 26 4 3
{ 16.5% 56.6% 21.2% 3.3% 2.4%

F. { 28 22 37 0 39
{ 22.2% 17.5% 29.4% 30.9%

H. { 39 98 0 11 0
{ 26.3% 66.2% 7.5%

R. { 9 102 32 0 0
{ 6.3% 71.4% 22.3%

K. { 47 58 1 14 0
{ 39.1% 48.3% .7% 11.9%

V. { 30 79 54 0 0
{ 18.4% 48.4% 33.2%

Ro. { 32 111 0 5 0
{ 21.4% 75.1% 3.5%

Bl. { 5 131 15 15 0
{ 3.1% 78.9% 9.0% 9.0%

By. { 43 102 23 15 22
{ 16.8% 52.6% 11.7% 7.7% 11.2%

Bs. { 9 72 31 0 6
{ 11.4% 58.5% 25.3% 4.8%

Ht. { 32 58 20 0 0
{ 29.0% 52.8% 18.2%

Totals{294 902 239 64 70
{ 18.7% 57.5% 15.3% 4.1% 4.4%

Number of subjects, 11; number of sections, 10; number of
experiments, 110; number of associations, 1569.

GROUP II

Two concrete nouns with apparently equivalent connotation were pronounced. In order that the subject get no clue as to the preponderance of one starting-point over the other, the nine sections were given at irregular intervals in connection with other experiments.

The tables show few cases of fusion, there being but one case in the eighty experiments where all the associations were the result of the combined influence of both starting-points. There was no other instance where complete fusion took place before the sixth association and continued throughout, and only six cases where any form of fusion took place in the eighty experiments within the first five words (7.6%), while in the one hundred and seven experiments of Group I there were twenty-five (23.9%).

The tables show a tendency which was noted when two syllables were used and which is emphatically brought out here, namely, it is the position or sequence which determines which of the two equivalent starting-points shall produce the greater influence. When the two starting-points are given in immediate succession, it is the second which predominates in influence. This preponderance may be clearly demonstrated by the tables alone, though there are many notes which show "the greater influence of one" (the second). A. seems to have realized this when he wrote, "While the first was uppermost the second hovered in subconsciousness, resting content, knowing that it would have its turn soon." In corroboration the tables give--abstracting from all cases of fusion--268 words (23.2%) for the first, 721 words (62.3%) for the second. In fifty-five of the eighty experiments the last had full control at the end of the series. Taking the cases where one was followed exclusively throughout the series, starting between the first and the sixth named association, we find eight for the first and twenty-nine for the second. This greater influence is again to be found in the cases of partial fusion; the first predominating in 2.6%, and the second in 4.3%.

The results of both groups show that the starting-points tend in a high degree toward independent influence, and also that such a method of presentation is one of sequence rather than simultaneity, as the two words show unequal influence. The former must in a manner be reproduced to become a point of departure for associations, while the latter acts directly. Is it then a condition of mind that when similar impressions are presented the one in the presence of which consciousness is reacting directly has a greater influence in arousing associations than one which is just past? This we are forced to conclude is the case, but the proof of the conclusion will be supplemented by later results.

TABLE II. TWO WORDS SPOKEN

Time--50 seconds.
Characters--same as Table I.

(1) Library--River

M. | + + β - - - - - - - - - - - -
F. | - - - - - - - - - | | | |
H. - - - - - - - - - - - - - - - -
Ro. | | | | - - - - - - - - -
J. | - | | | | - - - - - - - - - - - - - - -
Bl. | | - - - - | | | | | | | | | - - /
By. - - - - - - - - - - - - - - - - - - - - - -
Bos. | - - - - - - -
Bur. - + - - - - - - - - - - -

(1) Library--River.
(2) Hat--Road.
(3) Newspaper--Medicine.
(4) Bicycle--Drum.
(5) Theatre--Magazine.
(6) Store--Church.
(7) Soldier--House.
(8) Ship--Boy.
(9) Furniture--Tree.

TABLE II.--_continued_

| - + α β

M. {32 66 26 0 13
{23.4% 48.2% 18.9% 9.5%

F. {32 67 13 5 1
{27.2% 56.7% 11.0% 4.3% .8%

H. {59 79 5 0 0
{41.3% 55.2% 3.5%

Ro. {44 85 0 0 0
{34.1% 65.9%

J. {21 121 8 5 0
{13.5% 78. % 5.3% 3.2%

Bl. {22 83 6 0 1
{19.7% 74.1% 5.4% .8%

By. {16 97 18 14 23
{ 9.5% 57.8% 10.7% 8.3% 13.7%

Bs. {26 64 3 7 1
{25.8% 63.5% 2.9% 6.9% .9%

Bur. {16 59 8 0 10
{17.2% 63.4% 8.6% 0 10.8%

Total {268 721 87 31 49
{ 23.2% 62.3% 7.6% 2.6% 4.3%

Number of subjects, 9; number of sections, 19; number of
experiments, 80; number of associations, 1156.

GROUP III

For two syllables and like words the second strongly predominates in awakening associations; will the same be true when two simple outline pictures are shown in the same order? The following results show in attestation of the above conclusion percentages remarkably similar to those of the words. There are 23.7% for the first and 61% for the second; 84.7% of the associations show no sign of fusion and only 15.3% for fusion and the different forms of partial fusion, the second still holding the ascendency. When the two pictures did fuse, R. tells us, "All the associations were more elaborate pictures than when mere words were given."

A comparative study, inadequate though it is, offers a partial parallelism between the predominating memory type and stimulation. The subjects who are preëminently of the visual, K., visual motor, M., H., V., and visual lingual motor, Ht., find the pictures more suggestive than the syllables, while those of the pure motor, F. and By., reverse the order and find that the syllables offer more numerous and "more vivid" (F.) associations.

Common experience, however, immediately shows the limitations of reaching conclusions, when considering prolificacy as criteria of suggestiveness, inasmuch as starting-points which are abundantly rich in associations tend to produce so many points of departure that they tend to inhibit one another. Again there is to be considered the kind of associations, those of the syllables being of a very elementary character and in serial form.

A few experiments were given in which colored slips of paper were used as starting-points. These proved very suggestive for the subjects. Also a few tones were given, but these were soon discontinued, as other tones, which could not be recorded, were frequently suggested. The sentence which forms a very satisfactory starting-point where one is used could not be used to an advantage where several were given. As we are here interested in the mutual influence of the starting-points, our remaining study will be confined to a quantitative and qualitative variation of the forms used in the previous groups.

TABLE III. TWO PICTURES SHOWN

Time--50 seconds.
Characters--same as Table I.

(1) Boy Rolling Hoop--Blacksmith

M. | | | - - - - - - | | | | |
F. | | - - | + +
H. - - - - - - - - - - - -
R. - - - - - - - - - - - - - - - -
K. - - - - - - - - - - - - - - - - -
V. - - | | + + + + + + + + + + + + + +
S. - - - - - - - - - | | | | | | |
By. | - - - - - - - - - - - - - - - - - - - - -
Bs. - - - - - - - - -
Ht. | | - - - - - - - -

(1) Boy Rolling Hoop--Blacksmith.
(2) Old Man With Umbrella--Bird House.
(3) Carpenter--Mower.
(4) Children Playing--Boy with Basket.
(5) Horse--Dog House.
(6) Shoemaker--Fisherman.
(7) Little Girl--A Chicken.
(8) Boy--A Sheep.

| - + α β

M. {41 65 2 0 1
{37.6% 59.7% 1.8% .9%

F. {11 42 21 0 5
{13.9% 53.2% 26.6% 6.3%

H. {37 52 11 26 9
{27.4% 38.5% 8.1% 19.3% 6.7%

R. {41 58 8 0 15
{33.4% 48% 6.5% 12.1%

K. {32 82 0 0 0
{27.1% 72.9%

V. {12 95 21 3 5
{ 8.8% 70% 15.4% 2.1% 3.7%

S. {36 44 0 0 0
{45% 55%

By. { 8 111 12 0 13
{ 5.4% 77.2% 8.3% 9.1%

Bs. {13 40 7 0 0
{21.7% 66.7% 11.6%

Ht. {22 72 7 0 4
{20.9% 68.6% 6.7% 3.8%

Totals {253 661 89 29 52
{ 23.7% 61% 8.2% 2.7% 4.4%

Number of subjects, 10; number of sections, 8; number of experiments,
76; number of associations, 1084.

GROUP IV

The previous experiments suggest the conclusion that that starting-point in the presence of which consciousness is reacting exerts a greater influence than one just past. The proof of this anticipated result is supplemented by subsequent consectary data.

It is necessary to test the outcome when the nature of the impressions is varied and the starting-points are given simultaneously. An outline picture and a word may be so given. These experiments verify from a different standpoint the statement that the picture establishes itself more permanently and is more influential, there being 47.8% of the associations produced by the picture alone, 14.8% by the words alone, 25% of a fusion of the two, and 12.4% of a uniting influence with the first predominating. The subject R. is an exception in that she favors the word when not favoring a combination of the two, a fact which I am unable to explain except to add that the subject stated that the æsthetic pleasure connected with the picture was sufficient to inhibit the associations. K. has 100% for the pictures, which is partially explained by the fact that she is a remarkable visualizer who reproduces all situations in visual terms.

TABLE IV. PICTURE SHOWN AND WORD SPOKEN

Time--50 seconds.
Characters same as Table I.

(1) (a rabbit)--Table

M. | - - | | | | | | | + - - - -
R. + + + + + + + + + + + + + + + + + +
K. | | | | | | | | | | | | | | |
S. | | | | | - | | | | |
J. | - - | | | | | | | | | - - - | | | | | | | | |
Ht. | - | | | | | | | | | | | | - - - -

(1) (a rabbit)--Table.
(2) (horse)--Book.
(3) (boy)--Dish.
(4) (duck)--Iron.
(5) (dog)--Paper.

| - + α β
M. { 38 9 16 6 0
{ 55.0% 13.1% 23.3% 8.6%

R. { 1 20 55 0 0
{ 1.3% 26.3% 72.4%

K. { 65 0 0 0 0
{100.0%

S. { 29 15 3 10 0
{ 50.8% 26.3% 5.3% 17.6%

J. { 75 19 14 9 0
{ 64.2% 16.3% 11.9% 7.6%

Ht. { 17 7 30 33 0
{ 19.5% 8.1% 34.5% 37.9%

Totals {225 70 118 58 0
{ 47.8% 14.8% 25.0% 12.4%

Number of subjects, 6; number of sections, 5; number of experiments,
30; number of associations, 471.

GROUP V

The character of the starting-points determines which shall be followed in a simultaneous presentation; can the character of two immediately successive starting-points be so varied as to overcome the factor of temporal difference? The few experiments in this group show an attempt to shift and measure this influence, selecting as variants words of general and specific character, the first being a term of everyday experience with the subjects, of wide connotation, and therefore presumably rich in associations, and the second, which had the advantage of position, one of more limited and specific connotation, thus in a measure making it possible to test the stability of the second in establishing itself as a word-suggester. For the words "Gymnasium--Stamens" the second still strongly predominates. There is a slight increase in the amount of fusion (24.5% in all), the explanation of which would probably lie in the fact that the first is more assertive than before and offers more elements common to the two. There was one case of total fusion, one case where fusion started before the sixth word and continued, no cases of alternating fusion, four cases of partial fusion, five cases where the second was followed throughout, no case where the first was followed throughout, five cases of alternation between the first and the second. In 75.5%, the influence was exerted without fusion. Eight of the nine subjects favor the specific term.

TABLE V. GENERAL AND FAMILIAR WORD--SPECIFIC WORD SPOKEN

Time--50 seconds.
Characters--same as Table I.

(1) Laboratory--Rake

M. | | | | | - - - - - - - - - - - - - - - -
F. | - - - - - - - - -
H. + + + + + + + + + + +
V. | - - - - - - - - - - - - - | | | | |
S. - - - - - - - - - - | | | |
Bl. - - - - - - - - - - - - - - - - -
By. | - - | | | | | | | | |
Bs.
Ht. - - - - - - - -
J. | | β β β β β β β β β β β β | | | | |

(1) Laboratory--Rake.
(2) Experiment--Butterfly.
(3) Gymnasium--Stamens.

TABLE V.--_continued_

{ 22 25 4 6 0
M. { 38.6% 43.9% 7.0% 10.5%

{ 4 19 0 0 0
F. { 17.4% 82.6%

{ 4 25 13 0 0
H. { 9.5% 59.6% 30.9%

{ 7 38 2 0 7
V. { 12.8% 70.7% 3.7% 12.8%

{ 11 34 0 0 0
S. { 24.4% 75.6%

{ 0 36 0 0 0
Bl. { 100.0%

{ 11 16 7 0 9
By. { 25.5% 37.1% 16.5% 20.9%

{ 0 9 12 2
Bs. { 39.1% 52.2% 8.7%

{ 10 11 10
Ht. { 32.3% 35.4% 32.3%

{ 17 8 15 0 13
J. { 32.1% 15.1% 28.3% 24.5%

{ 86 221 63 8 29
Totals { 21.2% 54.3% 15.4% 1.9% 7.2%

Number of subjects, 10; number of sections, 3; number of experiments,
27; number of associations, 407.

GROUP VI

What modification of influence takes place when an auditory impression in the form of a word of abstract nature (the pure abstract words are taken later) is given in comparison with a concrete noun? The tables indicate that we are now able partly to overcome the disadvantage of first position by the advantage of concrete content. 24% of the associations are the result of total fusion; there are eight cases of total fusion throughout the series, and seven where total fusion took place before the third word and continued. There are two cases where the first was followed throughout, and but one where the second was followed exclusively. The tables indicate that there were very few words. Some of the subjects claim, "These words did not seem equally rich in associations. I was not at all conscious of the one while the other was in consciousness." (H.) The most characteristic feature here was, as Br. also indicates, "Great amount of rivalry at the beginning of the series"; while Bs. states, "There was a lot of confusion and a feeling of groping for words." Br. adds later, "For some seconds association seemed obstructed. Then by an effort the process was started which followed an involuntary course. A kind of confused presence of both words." Another subject adds, "There was a long blank after the words were said in which both words were balancing off in the fringe of consciousness and the mind expectant, passively waiting for an association to turn up. The hesitant period seemed marked by an attempt at a synthesis of these two words in some way."

SUBJECT XIX

Lamp--Justice.

| chimney lamp.
| white lamp chimney.
| yellow white chimney.
| flame yellow.
| nickel lamp plus the other words.
- scales justice.
- purple robe justice.

As the tables indicate, there was frequently a strong tendency here for the abstract terms to fuse. As H. noticed, "It does not tend to call up associations of its own stripe, but in some way becomes concrete."

TABLE VI. TWO WORDS: CONCRETE--ABSTRACT SPOKEN

Time--15 seconds.
Characters--same as Table I.

(I) Desert--Hate

A. | | | + + +
M. | | | | - -
F. + + +
H. | - | | | | | | |
Ro. | | - - | | | |
Bl. + + + + + + +
By. | | | + + + + + +
Bs. - | | | | |
Br. | | | + +

(1) Desert--Hate.
(2) Lamp--Justice.
(3) Pen--Fatigue,
(4) Gate--Fear.
TABLE VI--_continued_

| - + α β
A. { 10 2 6 0 0
{ 55.5% 11.2% 33.3%

M. { 7 8 10 0 0
{ 28.0% 32.0% 40.0%

F. { 3 8 5 0 0
{ 18.7% 50.1% 31.2%

H. { 17 16 1 0 0
{ 50.0% 47.1% 2.9%

Ro. { 15 11 0 0 0
{ 57.7% 42.3%

Bl. { 4 15 7 2 0
{ 14.2% 53.5% 25.1% 7.2%

By. { 4 21 8 0 1
{ 11.8% 61.8% 23.5% 2.9%

Bs. { 5 1 15 0 2
{ 21.8% 4.3% 65.2% 8.7%

Br. { 11 8 2 0 0
{ 52.3% 38.1% 9.6%

Totals { 76 90 54 2 3
{ 33.8% 40.0% 24.0% .8% 1.4%

Number of subjects, 9; number of sections, 4; number of experiments,
34; number of associations, 225.

Reversing the order by placing the concrete noun second, it gains in influence. We are told by the subjects at this point, "The choice seems to be determined by the concreteness of the word." (H.) "The abstract soon exhausted itself as a word-suggester." (Ro.) There was fusion in 24% of the associations of the first group, and 22.2% in the second. There are six cases where the second alone prevails, two for the first starting-point and four for the second. There are eleven cases of fusion beginning before the sixth word and continuing throughout the series. There is much partial fusion, with the second predominating in influence.

The results again emphasize the fact that the influence is transferable, also that normally the second has the advantage; furthermore, they illustrate the preponderance of the concrete word as a starter of associations, and that the abstract term when it exerts an influence tends to fuse rather than persist in having separate associations; all of which shows that concrete terms produce more vivid impressions than abstract ones, and would, when it is possible to use them, be of direct aid to the learner.

TABLE VII. TWO WORDS: ABSTRACT--CONCRETE SPOKEN

Time--15 seconds.
Characters--same as Table I.

(1) Honesty--Tide

A. - | - - - -
M. - - - - | | | |
F. | - - - - -
H. | - - -
Ro. | | - - - -
Bl. | - - - - - -
By. - - + +
Bs. - - - - - - -
J. - + + + + + +

(1) Honesty--Tide.
(2) Skill--Coal.
(3) Terror--Sky.
(4) Refined--Flag.

| - + α β
A. { 2 7 11 0 4
{ 8.4% 29.2% 45.8% 16.6%

M. { 4 11 10 1 5
{12.9% 35.5% 32.3% 3.2% 16.1%

F. { 2 8 9 0 0
{10.5% 42.2% 47.3%

H. { 8 11 1 3 0
{39.2% 43.5% 4.1% 13.2%

Ro. {10 13 3 0 0
{38.4% 50.0% 11.6%

Bl. {15 10 0 0 0
{60.0% 40.0%

By. { 2 20 3 0 0
{ 8.0% 80.0% 12.0%

Bs. { 1 11 6 0 2
{ 5.0% 55.0% 30.0% 10.0%

J. { 6 16 6 0 0
{21.4% 57.2% 21.4%

Totals {50 107 49 4 11
{22.6% 48.4% 22.2% 1.9% 4.9%

Number of subjects, 9; number of sections, 4; number of experiments,
36; number of associations, 221.

Increasing the disparateness by making the one a proper name and the other a pure abstract noun, we find the name dominates consciousness, almost to the exclusion of the abstract term. The tables confirm the conclusions that the abstract term, even when given the advantage of position, exerts little influence, for in the first group of eighteen experiments of two hundred and eighty-four associations there are 13.4 times as many associations for the first as for the second, or two hundred and fifteen words for the first (75.8%) and sixteen (5.5%) for the second.

Reversing the order, the burden of influence swings back to thirty-eight (12.8%) for the first and one hundred and fifty-six (52.7%) for the second with an amount of fusion increased to ninety-nine (33.4%).

TABLE VIII. TWO WORDS: PROPER NOUN--ABSTRACT NOUN SPOKEN

Time--50 seconds.
Characters--same as Table I.

(1) Lowell--Liberty

M. | | | | α α α | | | | | |
R. | α α | | | | | | | | | | | | |
K. | | | | | | | | | | | | | |
J. | | | + + + α α α α α α α α α α α α α
S. | | | | | - | | | | | | |
Ht. - - | | | | | | | | | | | | | |

(1) Lowell--Liberty.
(2) Roosevelt--Fidelity.
(3) Eliot--Integrity.

| - + α β

M. { 29 3 2 7 0
{ 70.7% 7.4% 4.8% 17.1%

R. { 30 0 17 2 0
{ 61.2% 34.7% 4.1%

K. { 43 2 0 0 0
{ 95.5% 4.5%

J. { 42 0 8 13 0
{ 66.6% 12.7% 20.7%

S. { 40 2 0 0 0
{ 95.2% 4.8%

Ht. { 31 9 3 1 0
{ 70.4% 20.4% 6.9% 2.3%

Totals {215 16 30 23 0
{ 75.8% 5.5% 10.7% 8.0%

Number of subjects, 6; number of sections, 3; number of experiments,
18; number of associations, 284.

TABLE IX. TWO WORDS: ABSTRACT NOUN--PROPER NOUN SPOKEN

Time--50 seconds.
Characters--same as Table I.

(1) Individuality--Lincoln

M. | + | + + + + + + + + + + + +
R. - - - - - - - - - - - - - - - -
K. - - - - - - - - - - - -
J. | - - - - - - - - - - - | | | | | | | | | |
S. - - - - - - - - - - - - - - -
Ht. - - - - - - - - - - - - - - - - - -

(1) Individuality--Lincoln.
(2) Brevity--Webster.
(3) Justice--Hanus.

| - + α β

M. { 4 19 25 0 0
{ 8.3% 39.6% 52.1%

R. { 5 31 12 0 0
{10.4% 64.5% 25.1%

K. {14 31 0 0 0
{31.2% 68.8%

J. {13 11 44 0 0
{19.1% 16.1% 64.8%

S. { 2 40 0 0 0
{ 4.8% 95.2%

Ht. { 0 24 18 0 3
{ 53.4% 40.0% 6.6%

Totals {38 156 99 0 3
{12.8% 52.7% 33.4% 1.1%

Number of subjects, 6; number of sections, 3; number of experiments,
18; number of associations, 296.

GROUP VII

Occasional experiments of the previous groups indicated an abnormal influence of those ideas which were accentuated by the special interests of the subjects. We ask therefore: What is the relation of the newly aroused associations to the present content of consciousness? May the present content of consciousness be so varied as to reënforce or inhibit the characteristic influence of the individual impressions successively presented? In order to test this, words or a sentence were pronounced just previous to the presentation of the two starting-points related to them as follows; in the first divisions the preparatory words lead to the second, in the second divisions to the first, in the third divisions to neither, in the fourth divisions to both, and in the last we have a slight change in the marginal setting and a sentence leads to the second.

The following are examples of the first and second divisions:

Quartz--Granite--Shale.
HORN--SLATE.
Dog--Sheep--Horse.
SQUIRREL--TELEGRAM.

SUBJECT XVIII