of long standing that the room does move; will this man entering the rotating chamber with his knowledge grow giddy, or will he not? Practice in the room may no doubt modify the effect, just as does practice on the merry-go-round. But it is only by practice and not by knowledge or experience that the individual can succeed in maintaining a fixed level of any sort under these trying circumstances. With the aid of our level-concept we can also understand the so-called (physiological) relativity of movement. A person looking from a bridge into a rapid stream soon has the impression of being himself moved. Seated in a train which is standing in a station, we are often unable to decide whether it is our train or the one on a neighboring track which is beginning to move. One explanation of the former effect actually maintains that the movement of a small piece of the bridge which belongs to one's field of vision is more probable than the movement of so large a surface as that of the stream! But we should say that normally there is no choice as to where we shall place our anchorage; for, in most cases, even with the strongest impulse of our will, we cannot alter this anchorage (see Wertheimer [64]). Normally it is something quite independent of our will a compulsory perception founded in properties of the objective field which determine for us what parts are to appear as figures and what parts as ground; as v. Hornbostel (19) puts it: things are not holes in the world of experience. On the other hand ambiguous situations occur in which two or more anchorages are equally possible, though here, too, law reigns and not chance. The chief rule for these ambiguous cases is this: that the objects which form the (dynamic) center of our visual world are at the same time our points of anchorage. When I am playing cards in my compartment I see the train move on the next track even if it is in reality my own train which is moving, but when I am looking at the other train, searching perhaps for an acquaintance in the coach, then it is my own train which seems to be moving. Psychologically, i.e., phenomenally, there is no relativity of movement.19

But our level-concept has still a wider application. We have already referred to certain instances, such as fashion and style. Experimental psychology has also studied certain facts about the phenomenon of the level without recognizing them to be such. What I mean is best explained by referring to some of Hollingworth's experiments upon the indifference-point (I.P.) (17, 18). Many

investigators, testing a scale of magnitudes, have found the existence of an I.P.; that is, while most members of the scale were estimated with a constant error, positive or negative, small magnitudes being overestimated and large ones underestimated, there comes a point where no constant error occurs. Though the fact has been confirmed over and over again, and in very different fields, yet, strangely enough, there has been a wide divergence of opinion as to the absolute position of the. I.P. This startling fact suggested to Hollingworth the idea that there must be a mistake somewhere in the way the question is put. Is there, he asks, an absolute I.P. independent of the position and extent of the test-series, or is this I.P. a function of the total scale? He was able to demonstrate by a number of ingenious experiments that the latter is the correct assumption. Working with the reproduction of hand-and-arm movements, he arranged three series of experiments; A, including magnitudes of 10 to 70 mm. (with increments of 10 mm.); B, magnitudes of 30 to 150 mm. (with increments of 20 mm.); and C, magnitudes of 70 to 250 mm. (with increments of 30 mm.); each scale consisting of seven different magnitudes. Upon a given day only one of these series was used. The I. P. of series A fell at about 40 mm., of B, at about 75 mm., and of C at about 125 mm., that is, it was always found to be approximately at the center of the scale. Smaller magnitudes were overestimated and larger ones underestimated. There was no absolute I.P. The magnitude of 70 mm., being the upper limit of A, and the lower limit of C, and near the middle of B, was underestimated in A (minus 10.2), overestimated in C (plus 16.5), and reproduced fairly accurately in B (plus 1.7; p.e. 10.3). To check this result, four months later, the three magnitudes of 10 mm. (always overestimated), 250 mm. (always underestimated) and 70 mm. (variable with the series) were tested singly on occasions several days apart, but for none of these three did a constant error

occur.

In still another very clear experiment, the shifting of the I.P. itself was demonstrated. A set of standard magnitudes was prepared ranging from 10 mm. to 60 mm. (by increments of 10), from 60 mm. to 150 mm. (by increments of 15) and on to 250 mm. (by increments of 20). The standards of the 10 mm.-60 mm. were now given and reproduced in chance order, five trials being given for each magnitude. Then, without the knowledge of the O., the next magnitude was added and again five trials made of each standard. This was continued until the whole series of seventeen standards had been offered. The success of this experiment was remarkable. The I.P. rose with the introduction of each new standard magnitude; constant errors which were positive from the beginning, increased throughout the series, while constant errors which were negative in the beginning likewise underwent a continual change, decreasing to the zero point and emerging again as positive increments.

Hollingworth concludes, "that the phenomenon of the indifference point is of purely central origin" (17, p. 21), and this theory is as close to the one we propose as the general theoretical position of psychology at the time of his investigation would admit. According to his results, the I.P.-phenomenon belongs not to memory but to perception, and as an analogy he refers to type-concepts, such as race and class (18, p. 468). He also speaks of a "mental set", meaning by this "that we are adjusted for or tend to expect the average magnitude, and to assimilate all other magnitudes toward it, to accept them in place of it" (17, p. 39). But he insists on employing the term judgment"; the error to which this tendency leads, he says, "is distinctly an error of judgment, and is quite independent of sensory or physiological conditions" (18, p. 469).

Again the distinction drawn between sensory components and judgments of peripheral and central factors vitiates his theory. Leaving these out of account, and referring the reader to the next article for a discussion of the difference between peripheral and central factors, which takes on a very different aspect in our theory, we may here draw the following conclusions. In reacting to a definite scale of stimuli we establish a general level which, in the case described, as in many others, is both motor and sensory. The effect of each single stimulus is dependent upon this level, much as the figure is dependent upon its ground. And secondly, the general level holds together the whole group of phenomena corresponding to the scale of stimuli. Although they may rise or fall from this level, the phenomena never lose their existential connection with it, and being attracted by the level, the result is often a wrong judgment or a false reproduction. This attracting or assimilating effect of the level is a special case of our general law of levelling (discussed above). We see further that this level adapts itself automatically to the scale and this process of adaptation must therefore be explicable in terms of our general physiological theory.

Hollingworth rightly gives a wide application to his results, comparing the I.P.'s of different investigators with the range of their scales, and he has himself confirmed his "law of central tendency in a purely sensory field by experiments upon the size of gray squares. I may also add in this connection that what G. E. Müller calls the "absolute impression" (Absoluter Eindruck) is just such a rise or fall from a general level. Whenever an O. makes a judgment that is not based upon a comparison between two stimuli, he is reacting not to a stepwise phenomenon, but to an emergence from the general level. With this I must bring my first article to a close.

13. V. HELMHOLTZ, H. Handbuch der physiologischen Optik. 3 Aufl. Erg. u. her. in Gemeinschaft mit A. Gullstrand und J. v. Kries von W. Nagel.t 3 Bde. Hamburg & Leipzig, 1909-1911 (1. Aufl. 1856-1866).

14. HENNING, H. Experimentelle Untersuchungen zur Denkpsychologie. I. Die assoziative Mischwirkung, das Vorstellen von noch nie Wahrgenommenem und deren Grenzen. Zeits. f. Psychol., 1919, 81, 1-96.

15. HERING, E. Grundzüge der Lehre vom Lichtsinn. Berlin: Springer, 1905-1920, pp. v+294. Sonderabdruck a. d. Handbuch d. Augenheilkunde (Graefe-Sämisch) 1. Teil, XII Kop.

16. HÖFLER, A. Gestalt und Beziehung-Gestalt und Auschanung. Zeits. f. Psychol., 1912, 60, 161-228.

17. HOLLINGWORTH, H. L. The Inaccuracy of Movement With Special Reference to Constant Errors. Arch. of Psychol., 1909, No. 13, Columbia Contrib. to Philos. & Psychol., 17, pp. 87.

18. HOLLINGWORTH, H. L. The Central Tendency of Judgment. J. of Philos., Psychol. and Sci. Meth., 1910, 7, 461-469.

19. V. HORNBOSTEL, E. M. Über optische Inversion. Psychol. Forsch., 1922, 1, 130-156.

20. V. HORNBOSTEL, E. M. UND WERTHEIMER, M. Über die Wahrnehmung der Schallrichtung. Sitzungsber. d. Preuss. Akad. d. Wissensch., 1920, 20, 388-396.

21. JAENSCH, E. R. Einige allgemeinere Fragen der Psychologie und Biologie des Denkens, erläutert an der Lehre vom Vergleich. (Mit Bemerkungen über die Krisis in der Philosophie der Gegenwart.) Arb. z. Psychol. und Philos. Leipzig: Barth, 1920, 1, pp. 31.

22. JAENSCH, E. R. UND MÜLLER, E. A. Über die Wahrnehmung farbloser Helligkeiten und den Helligkeitskontrast. Über Grundfragen der Farbenpsychologie. Zugleich ein Beitra zur Theorie der Erfahrung. Hn. v. E. R. Jaensch I. Zeits. f. Psychol., 1920, 83, 266-341. (Auch separat bei Barth.)

23. JAENSCH, E. R. Über den Farbenkontrast und die so genannte Berücksichtigung der farbigen Beleuchtung. Uber Grundfragen u.s.w. III. Zeits. f. Sinnesphysiol., 1921, 52, 165-180.

24. KATZ, D. Die Erscheinungsweisen der Farben und ihre Beimflufsung durch die individuelle Erfahrung. Erg. Bd. d. Zeits. f. Psychol. Leipzig: Barth, 1911, pp. xviii+425.

25. KOFFKA, K. Beiträge zur Psychologie der Gestalt. I. Untersuchungen über den Zusammenhang zwischen Erscheinungsgrösse und Erscheinungsbewegung bei einigen sagenannten optischen Täuschungen, von F. Kenkel. Zeits. f. Psychol., 1913, 67, 358-449.

26. KOFFKA, K. The same. III. Zur Grundlegung der Wahrnehmungspsychologie. Eine Auseinandersetzung mit V. Benussi, von K. Koffka. Zeits. f. Psychol., 1915, 73, 11-90.

27. KOFFKA, K. The same. IV. Zur Theorie einfachster gesehener Bewegungen. Ein physiologisch-mathamatischer Versuch von K. Koffka. Zeits. f. Psychol., 1919, 82, 257-292.

Beiträge I-IV also separate, as Beiträge, etc. Vol. I, Leipzig: Barth, 1919, pp. v+323.