Classics in the History of Psychology

An internet resource developed by
Christopher D. Green
York University, Toronto, Ontario

(Return to Classics index)

James McKeen Cattell (1887)

Assistant in the Psychological Laboratory, University of Leipsic.

First published in Mind, 11, 524-538.

Part 4 of 4

IV. The Will-time.[2]

In the experiments described in the foregoing section the motion to be made was always the same, and took up the same or about the same time. In this section the nature of the motion depends on the nature of the impression. The experiments about to be described will throw further light on the Perception-time, but we shall find in addition a variable Will- (or Motor) time. The perception-process, further, is different from what we considered in the foregoing section: then the observer expected a certain impression and saw whether it was present or not; in the experiments now to be described the observer, not awaiting a given impression, had to identify the one occurring. We might perhaps expect the perception to be more difficult and consequently to last longer in the latter case; the experiments however show that there can be no great difference in the time.

Experiments have been made in this direction by Donders[3] and others, they letting the observer lift his right hand if (for example) the light is red, the left if it is blue. Under Wundt's direction Merkel[4] extended this method, the observer lifting a different finger for each of ten different impressions. My first experiments (carried out in the winter 1883-4) were made with aid of electric lights, as above described, and were similar to those of previous experimenters; they gave as the time for distinguishing the colour and choosing the motion 120s for B, 168 for C. Afterwards I used the gravity chronoscope, which enabled me to use daylight reflected from coloured surfaces. The current controlling the chronoscope was led through two keys (K and K' Fig. 8, MIND 42), the observer holding one closed with his right, the other with his left hand. Two colours, say red and blue, were used in the same [p. 525] series of experiments. If red appeared the observer lifted his right hand, if blue his left. The times are given in Table XXIX., the pairs of colours used being red and blue, and green and yellow. The reaction on red and on green was made with the right hand, on blue and on yellow with the left. Each number gives the average of six series (78 reactions in the uncorrected series).

If from the average time for the four colours we substract[sic] the simple reaction-time, we find that it took B 145, C 190s to distinguish the colour and find the proper motion. If these times are compared with those given in the preceding section (Table XVI.) we find that it took B 18, C 34s longer to send out the proper and corresponding motion, than the command sending out a motion already determined. As I have already remarked, the perception-process is slightly different in the two cases; it being necessary in the first to see whether the light is red or blue, in the second only to recognise the red light. The results of the experiments show that there can scarcely be a difference in the times taken up by the two processes.

Quite a similar method was used with letters, the observer lifting his right hand if A was present, his left hand if Z. The numbers given in the Table are taken from six series.

[p. 526] The perception-time was thus for B 38, for C 40s longer than for colours, and the will-time for B 11, for C 43s longer for the choice between two motions than for the choice between making a motion and not making it.

In most of my experiments the motion corresponding to the impression was made with the organs of speech. I consider the time of special interest, as we are constantly reading a word, naming a colour, &c. In the experiments first to be considered two impressions were used; the observer did not know which was to come, but named the one occurring as soon as possible after seeing it. The motion was registered by means of the sound-key. These experiments are an extension of those given in Tables XVIII., XXII., XXVII. and XXVIII. There the observer made a determined motion (i.e., named an expected object), here the motion had to be found after the impression had been distinguished. The relation between the processes is exactly the same as when the motion is made with the hand, the only difference being that we are constantly giving the name blue (for example) to a certain colour, whereas the association between a motion of the left hand and the colour blue must be made for the experiments. The impressions were taken in pairs as indicated in Table XXXI. 26 reactions were made as usual in a series, 13 on each of the two impressions.

These results in Table XXXI., when compared with those given in Tables XVIII., XXII., XXVII. and XXVIII., give the increased will-times shown in Table XXXII.

We have already seen that with the hand B needed less additional time than C to make the choice between two motions: the difference between the two observers is still more marked when the motions are made with the speech organs. Table XXXII. is further interesting in showing a difference between letters and words on the one hand, and colours and pictures on the other. The association between a printed letter or word and its name requires less time, and is consequently closer than between a colour or picture and its name. We can understand this, as the former association is being more continually practised; still we could not have foreseen it, as the association between a colour or object and its name is formed long before we learn how to read.

In the experiments now to be described there were not two objects and two corresponding motions, but a large number of objects; the one occurring to be named by the observer. In this case we determine the time it takes to see and name an impression, as a word or a colour. We have in the preceding section determined approximately the time taken to see an object: the difference between the two times gives us the time it takes to name the object. We shall first consider the time needed to see and name a letter. All the letters of the alphabet (capital letters of the largest size in the text of MIND) were used, each occurring once in the course of the series. After thirteen series had been [p. 527] made, the times for the separate letters were averaged together, so that we get the average of thirteen determinations on each letter; these series were corrected in the usual way, the three reactions varying most from the corrected average being dropped.

[p. 528] As the determinations for the same letter were made at different times we find the mean variation larger than usual. Table XXXIII. gives besides the results obtained with aid of the sound-key, series made with aid of a second observer. The first observer simply named the letter as quickly as possible, and the second observer made the reaction on the sound in the manner above described. Mr. Wolfe acted as second observer; in the Table I have subtracted his reaction-time on sound (150s) and his mean variation (10s).

We thus see that it takes the observers about four-tenths of a second to see and name (i.e., read) a letter. In this connexion [p. 529] results I have already published[5] should be considered. I there determined by two distinct methods the time it takes to see and name letters. In most of the experiments, however, the observer while seeing and naming one letter could begin to see and name the one or ones following, so that the processes overlapped and the times became much shorter, namely 279s for B, 224 for C. The times were still further shortened (becoming 96s for B, 89 for C) when the letters made words. Why B's times should be longer than C's under these circumstances and shorter for a single letter I do not know. We found in the preceding section that it took B 119, C 116s to perceive a letter. Supposing the perception-time to be the same in both cases, B needed 143s, C 176 to find the name belonging to a letter. It should be added that in later series of experiments B's time became shorter. This method of determining the relative legibility of the several letters has an advantage over that in the previous section in so far as all the letters occur in the same series; but it is greatly complicated by the fact that the time of pronouncing the several letters may be different, as also the motion registered by the sound-key or second observer.

Series were made on the German capital letters with the results given in

Numbers of one, two and three places were further used, and the time it takes to see and name them was determined. I did not take numbers of more than three places, fearing that they might not be seen and read as wholes. The results are given in Table XXXV.; from which it will be seen that it took B 33, C 38s longer to see and name a number of two places than of one, and B 57, C 47s longer for a number of three than of two places.[6]

[p. 530]

The time it takes to see and name a word was determined in the same way. Experiments were made on long and short English and German words, 26 of each sort being taken. In the case of the short words I made thirteen series, and found the time for the separate words as with the letters (Table XXXVII.). On the long words only five series (130 determinations) were made, and the times for the separate words were not calculated (Table XXXVI.).

[p. 531]

An examination of the Tables shows that it took longer (in English B 52, C 46s, in German B 39, C 46s) to see and name a long than a short word. In both cases, of course, the beginning of the motion was registered; so the time occupied in pronouncing the word does not come into consideration. We further learn that it takes longer (for short words B 17, C 35s, for long words B 30, C 38s) to see and name a word in a foreign than in one's native language.[7] Comparing the results here reached with those given in the foregoing section, we find that to name a short word in his native language B needed 104, C 114s. We find further that B named a word in 39, C in 62s less time than a letter. [p. 532] This is not surprising; we are constantly reading and using words, much more than letters; so the association between the concept and the name has become closer and takes place in less time.

The same method was used to determine the time it takes to see and name a colour. The ten colours taken occurred two to three times in a series, and the times for the separate colours (13 determinations) were afterwards averaged together. Table XXXVIII. gives the times as measured with aid of a second observer, and as directly registered. The order of the colours is that of the average time needed to name them, beginning with the colour most quickly named.

These results are interesting and were not foreseen. We found in the preceding section that it takes less time to perceive a colour than a letter or word; we now find that it takes comparatively a very long time (B 286, C 400s) to find the name of the colour. This was especially the case at first and for certain colours. The colour was recognised with ease, but the name could only be found (more especially by C) with great difficulty. The colours most frequently seen and named in our daily life, red, yellow, green, blue and black, were named with greater ease and in decidedly less time (B 61, C 150s) than the other five colours, pink, violet, orange, gray and brown. In the case of these latter colours the time was considerably shortened by practice.

[p. 533] The twenty-six pictures already described were in like manner seen and named (by B in German, by C in English), the times for the several pictures (13 determinations) being given in the following Table.

We found in the foregoing section that pictures (and, we may assume, the objects themselves) can be seen in about the same time as colours; we now find that they can also be named in about the same time (by B in 251, by C in 278s) as the colours most frequently used. The difference in time for the several pictures is interesting; both B and C named the picture of a hat in the shortest time; B required the longest time to name 'bird' and 'teapot,' C to name 'teapot' and 'moon'. It is an interesting fact that the picture of a chair can be recognised in less time than the word 'chair,' but that it takes over a tenth of a second longer to name it.

It will be useful to collect together certain of the results of these experiments. In the following Table I do not give to the thousandth of a second the averages of the determinations made, but what from the experiments seems to be the time it takes B and C to perceive and find the name of the objects we have been considering.

[p. 534]

We have thus found the time it takes us to see and name the objects which we spend a great part of our life in seeing and naming. We have not been dealing with artificial processes or things outside the circle of our natural interests. If in the course of evolution, as is probable, the molecular arrangement of the nervous system becomes more sensitive and delicately balanced, we may suppose that the times taken up by our mental processes become shorter, and we live so much the longer in the same number of years. It will therefore be of great interest to make experiments such as these on the lower races, as well as on persons of different age, sex, occupation, &c.

V. The Influence of Attention, Fatigue and Practice on the Duration of Cerebral Operations.

We have seen that while the time of a reaction is somewhat lengthened when the brain cannot so well prepare itself, it does not vary greatly with different degrees of Attention. I have made similar determinations for cerebral operations in which complications have been added to the simple reaction-time. I chose as typical cases the time it takes to see a white surface and show this by a motion of the hand, and the time it takes to see and name a letter. On the one hand the observer tried by great concentration of the attention and effort to react as quickly as possible; on the other hand the impression was produced at irregular intervals (three-fourths to fifteen seconds), so that the brain could not be held in a maximum state of readiness.

We find, from Table XLI., under the two degrees of attention or preparation a difference in the times of seeing and reacting on the white surface of 75s for B, 15 for C; in the time of seeing and naming a letter 29s for B, 25 for C.

[p. 535]

As I have given throughout this paper the dates on which the series were made and have not omitted any series, the results of continued Practice can be studied to advantage. B and C had previously had considerable practice in making simple reactions, but none in the other processes here considered. In the twenty series of reactions on light (Table I.) made during a period of six months, no reduction in the time is to be noticed. B's reaction-time was however shorter in 1884-5 than in 1883-4, as can be seen from Table VI., where his times, especially for light, are considerably longer than C's. I repeated at the close of the investigation the determinations made at the beginning in which the observer reacted on one of a number of colours, letters or words (the results are given in Tables XIX., XXIV., XXVII.), and found that the times had become shorter. I give the decrease in time.

                                                                              B                               C
                        Colours . . . . . . . . . . . . . . . . . . . . 28                              20
                        Letters . . . . . . . . . . . . . . . . . . . . . 50                              25
                        Words . . . . . . . . . . . . . . . . . . . . . 54                              35

[p. 536] As I have already mentioned, the time of naming the colours and pictures became shorter through practice. In some cases where the attention was distracted the brain accommodated itself to the changed conditions. It can be stated as a law that the times of cerebral operations become shorter as they become more automatic, but that a limit is reached beyond which further practice has little or no effect.

The investigation was concluded in April; in July, after an interval of three months during which no reactions were made, the times of the more important processes were again measured. The Table gives the results of five series of simple reactions on light and sound, of five series in which the observer showed by a motion of the hand that he had perceived a white surface, a letter and a colour, and of three series in which he perceived and named a letter, a word and a colour. The increase or decrease of the time is given in the column headed Df.

We now come to the effects of Fatigue. These, like the effects of attention, have been greatly overestimated, experimenters having made but few reactions in a series or at a sitting, fearing lest the observer should become fatigued and the times unduly long. In order to determine the influence of fatigue in successive reactions, I took thirty series of simple reactions on light and averaged all the first reactions (as also the mean variation of these reactions) together, all the second reactions, and so on through the twenty-six reactions of which the series was made up. In the same way I took two hundred series where the subject had to react (with the hand) after distinguishing an impression, and averaged all the first, second, &c. reactions together. The impressions were different in the different series, but of course the same throughout each series. In these series only thirteen determinations were made, but twenty-six mental processes took place, it being as fatiguing to see that the object was not there and keep from reacting, as to distinguish the object and react.

[p. 537]

It will be seen that, though the difference is not great, the first reactions of a series are the shortest. It seems that in the first experiments the observer involuntarily strains his attention more, and so gives shorter times. This is the more marked the less automatic the process is; that is, with the white light than in the simple reaction, and in the case of B than in the case of C. The further course of the series shows no lengthening in the times or increase in the mean variation; so the brain is not considerably fatigued by making (or refraining from making) twenty-six reactions in succession.

In order further to investigate the effects of fatigue, I made extended series of experiments in which 1950 reactions were made in succession, the observer reacting almost continuously from early in the morning until late into the night. Three series (78 reactions) were made with light, then three series (39 determinations, but 78 mental processes) in which white light was distinguished and reacted on, then three series in which letters were seen and named, then two series in which associations were made, lastly three series of reactions on sound. This entire combination of series was repeated six times. The experiments were begun both days at 7:30 a.m. and were concluded in the case of C at 1:30 a.m., in the case of B at 11 p.m., short pauses being made for meals. One series of each variety was made the following morning and again in the evening; in the case of C a further set of series the day after. In the Table I give the average time and mean variation of the first set of series, afterwards the increase or decrease as compared with these. I do not take up space to give the corrected series, as they scarcely differ from the others.

[p. 538]

The first result to be noted from the Table is the very slight effects of fatigue; in no case is the time lengthened more than a couple of hundredths of a second, and the mean variation is but little increased. We reach the unexpected result that the processes which are the most automatic (naming the letters, and C's simple reaction-time) are the most affected by fatigue. The determinations made on the following day show that B had recovered from all fatigue; in the case of C, however, the brain substance concerned in the simple reaction seems to have been so far exhausted that his reaction-time remained abnormally long for two days.

I think these experiments show that it is possible to apply scientific methods to the investigation of mind. We have determined the times required for those processes which make up a great part of our mental life, and found these times to be constant; they are no more arbitrary, no less dependent on fixed laws than, for example, the velocity of light. I shall soon print an account of experiments going a step farther and determining the times of mental processes more removed from psycho-physical operations having to do with sensation and motion.


[1] Concluded from MIND 42 and 43.

[2] I use the term 'Will-time' for lack of a better; in Germany 'Wahlzeit' is used. The motion is in most cases simply the result of the perception, and 'Association-time' might be used, were it not already taken up. 'Motor time' would perhaps best explain the process, but might cause confusion.

[3] Arch. für Anat. u. Physiol., 1868.

[4] Phil. Studien, ii, 1.

[5] Phil. Studien, ii. 4; MIND 41.

[6] See Friedrich, Phil. Studien, i. 1.

[7] See Cattell, Phil. Studien, ii. 4; MIND 41.