Open Access

JSON Export

{"created":"2022-01-31T16:19:18.845565+00:00","id":"lit23180","links":{},"metadata":{"alternative":"Studies from the Yale Psychological Laboratory","contributors":[{"name":"Seashore, C. E.","role":"author"}],"detailsRefDisplay":"Studies from the Yale Psychological Laboratory 3: 1-67","fulltext":[{"file":"p0001.txt","language":"en","ocr_en":"MEASUREMENTS OF ILLUSIONS AND HALLUCINATIONS IN NORMAL LIFE,\nBY\nC. E. Seashore, Ph.D.\nPART FIRST.\nIllusions op weight : influence of knowledge of size on\nJUDGMENT OF WEIGHT.\nWhen an object lifted is found heavier than was expected, it is overestimated, and when it is found lighter than was expected, it is underestimated. The strongest and most frequent illusion of this kind in normal life is perhaps that which is caused by our accustomed associations between the properties of size and weight of objects. The aim in Part First of this research is to investigate the nature and extent of the illusions of weight as caused by knowledge of the size of the body lifted. Illusions of weight from other sources are incidentally considered. The experiments were made between October, 1893, and May, 1895.\nThe problem is a development from a test made by Gilbert1 on the suggestive force of size on judgment of weight. Dresslar2 has made a statistical study of the same illusion, and it had previously been noticed and subjected to experiment by others.3 4 5 Since the present experiments were completed, a monograph by Griffing* has appeared, in which he touches upon the same illusion in so far as it is connected with the sense of impact and pressure ; Flournoy' also reports a very popular experiment on the same illusion.\n1 Gilbert, Researches on the menial and physical development oj school children. Stud. Tale Psych. Lab., 1894 II 43-45, 59-63.\n! Dresslar, Studies in the psychology of touch, Am. Jour. Psych., 1894 VI 313.\n8 Charpentier, Analyse de quelques \u00e9l\u00e9ments de la sensation de poids, Archives de Physiol., 1891 (5) III 126.\nM\u00fcller and Schumann, Ueber die psychologischen Grundlagen der Vergleichung gehobener Gewichte, Archiv, f. d. ges. Physiol. (Pfl\u00fcg.), 1889 XLV 37.\n4\tGriffing, On the sensations of pressure and impact, Psych. Rev., 1895 II Suppl. I.\n5\tFlournoy, De l'influence de la perception visuelle des corps sur leur poids apparent L\u2019Ann\u00e9e Psych., 1894 I 198.","page":1},{"file":"p0002.txt","language":"en","ocr_en":"2\nC. E. Seashore,\nFirst series of experiments : Influence of size on judgment of weight when size is estimated by direct sight.\nTwo sets of cylindrical blocks 31mra in length were made of brass tubing with hard-rubber ends. An additional hard-rubber disk on each end, 2mm thick and 15mm in diameter, served for the place of grasping. The entire length of each block was thus 35mm. In order that the appearance of the surface should not suggest any definite material, the blocks were all painted a dull, smooth black with optical varnish.\nEach set consisted of 17 blocks. Set A varied in size and had a uniform weight, while Set B varied in weight and had a uniform size. The blocks in Set A varied in diameter according to a geometric series in which the regular increment is one-tenth. Those in Set B were arranged in arithmetic series according to weight with a successive difference of 5\u00ae.\nIn the following account the blocks will be distinguished by the names A and B with their respective numbers in the series.\nThe blocks of Set A were of a constant weight, 80\u00ae, and of diameters in millimeters as follows, beginning with the smallest : 20.0,\n22.0,\t24.2, 26.6, 29.3, 32.2, 35.4, 39.0, 42.9, 47.2, 51.9, 57.1, 62.8, 69.1,\n76.0,\t83.6, 91.9.\nThe blocks of Set B were of a constant diameter, 42.9mm, and of weights in grams as follows, beginning with the lightest : 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100, 105, 110, 115, 120.\nIt is to be observed that the uniform weight for Set A is the same as the weight of B (9), the middle block in Set B ; and the uniform size in Set B is the size of A (9), the middle block in Set A. The limit in diameter of the extremes was determined by the size which will admit of uniform grasp.\nThe material of which the blocks were made was not appreciably subject to change from the ordinary variations of temperature and atmospheric condition. The weight was made accurate within a limit of 0.025\u00ae.\nThe two sets were placed on a tray that had a soft, black cloth bottom, Set A being arranged in order of size and Set B in order of weight.\nThe observer placed himself by the table on which the tray stood in such a position that by moving back and forth he could lift any block from its place in Set B and still retain approximately the same angle of the arm and hand. He was requested to select for each","page":2},{"file":"p0003.txt","language":"en","ocr_en":"Measurements of illusions and hallucinations in normal life. 3\nweight in Set A a corresponding one in Set B, by taking one at a time from A and placing it by the side of successive blocks in B with which he wished to compare it, lifting one at a time until he found the one in B which he thought had the same weight as the one from A. He was required to use the same hand, in the same position, and to clasp the disks of the blocks carefully between the thumb and middle finger, so that in every case the touch and the\nTable I.\nIllusion of weight when the blocks are seen directly.\nA\tB\tG\tB\tMV\n20.0\t110.2\t\u201422.9\t+30.2\t7.5\n22.0\t103.8\t\u201420.9\t+23.8\t7.0\n24.2\t98.2\t\u2014 18.7\t+18.2\t5.0\n26.6\t94.4\t\u2014 16.3\t+14.4\t6.5\n29.3\t94.0\t\u201413.6\t+14.0\t6.5\n32.2\t89.2\t\u201410.7\t+ 9.2\t8.0\n35.4\t86.3\t\u2014 7.5\t+ 6.3\t5.0\n39.0\t85.4\t\u2014 3.9\t+ 5.4\t4.0\n42.9\t83.8\t0\t+ 3.8\t6.0\n47.2\t80.4\t+ 4.3\t+ 0.4\t5.0\n51.9\t75.6\t+ 9.0\t\u2014 4.1\t4.0\n57.1\t71.6\t+14.2\t\u2014 8.9\t5.5\n62.8\t69.0\t+19.9\t\u201411.0\t6.5\n69.1\t65.8\t+26.2\t\u2014 14.2\t6.5\n76.0\t64.2\t+33.1\t\u201415.8\t6.5\n83.6\t61.2\t+40.7\t\u2014 18.8\t6.0\n91.9\t58.6\t+49.0\t\u201421.4\t6.5\nA,\tsize of the block in Set A (having a B, grams by which the estimated weight\nweight of 80s).\tof the block in Set A differed from its\nB,\tweight of the block in Set B (having true weight ; average of a total of 25 a diameter of 42.9ram) chosen as equal in experiments on 15 persons.\nweight to the block of Set A.\tMV, mean variation ; to obtain the\nC,\tnumber of millimeters by which the mean variation for the series each result diameter in Set A differed from that in j is to be divided by 5.\nSet B.\ngrip would be constant. He was at liberty to try each block a sufficient number of times to satisfy himself, but was warned against making so many repetitions that a disturbing fatigue would set in. When the experiment was repeated on the same day, sufficient rest was allowed between each set, and the observer was always cautioned to guard against any guess-work, theory, or memory of previous judgments. The observer had no names for the blocks, but,","page":3},{"file":"p0004.txt","language":"en","ocr_en":"4\n0. K Seashore,\nas he indicated his choices, the experimenter kept private record. He saw the difference of size in A and was told the difference and order of weight in B. Before beginning the experiment, he had the privilege of trying the blocks at random, and in the experiment no restrictions were made in regard to what order the A\u2019s should be taken in the set.\nInfluence of size on judgments\nof WEICHT\nFig. X.\nInfluence of size on judgments\nOF WEIGHT\nDifference in size,x=A-B in mm. Difference in weight,y=B-A in c.\nActual results\nIoEAL RESULTS\nThis method insures a direct measurement of the illusion caused by the influence of size in discrimination of weight. The results are given in Table I, and are graphically represented in figures 1 and 2.\nIn fig. 1 each successive increment of of the size is represented by one unit of the abscissa. If, according to Weber\u2019s law,","page":4},{"file":"p0005.txt","language":"en","ocr_en":"Measurements of illusions and hallucinations in normal life. 5\nproportional increments are perceived as equal increments, the successive differences are equal. The chart thus represents the differences in diameter as perceived.\nThe curve, fig. 2, shows that, with slight exception, proportional increments of size produce illusions of equal absolute differences in weight.\nThe uniformity of the mean variation is strong evidence of the regularity of the illusion. In these 25 experiments it varies between 4\u00ae and 8s. The 15 observers are in such consensus that even in the case of the greatest illusion the mean variation is only 6.5\u00ae, while for the ninth weight, where there is no illusion, the error of observation is 6\u00ae. This fact tends to show that the influence due to size in the determination of weight within a middle range is almost as potent as an absolute difference in weight.\nThis regularity is further emphasized by the fact that there is not, for all these observers, a single exception as to the direction of the illusion in the first three and last six blocks in the set.\nThe psychological laws expressing the influence of knowledge of size on judgment of weight, within a limited range, may be formulated as follows :\n1.\tBodies of similar material that have the same weight, but differ in size, appear to differ in weight when compared.\n2.\tThe larger is underestimated and the smaller is overestimated.\n3.\tThe intensity of this illusion varies directly with the perceived amount of difference in size between the bodies compared.\n/Second series of experiments : Persistence of the illusion of weight.\nA. Actual relations of weight unknown.\nHaving determined that the direction of the illusion was the same for all persons tried, and that it approximated the same amount, the next step was to test its persistence.\nIn this problem the first question to be settled was this : Does the illusion persist in spite of continued practice in attempting to gain accuracy in discrimination for weight, as long as the observer does not know the actual weight or the presence of the illusion ?\nFour careful observers were selected and subjected to the same tests independently, according to the same method of experimenting as in the preceding series.\nEach observer tried the same complete experiment twenty times, under similar circumstances, making two experiments each day. No","page":5},{"file":"p0006.txt","language":"en","ocr_en":"6\nG. JE. Seashore,\nsuggestion as to the degree of accuracy or success was given during the progress of the experiments. The instructions emphatically expressed were : \u201c The object of these experiments is to determine whether you can improve in the accuracy of this discrimination by practice. Do not allow yourself to be influenced in the least by memory of previous judgments or any theory of order or expected results.\u201d\nIn addition to the error of observation, expressed by the choice of corresponding B\u2019s for A (9), an extra test for the normal degree of\nTable II.\nPersistence of the illusion of weight when the fact of the illusion is not known.\nG\tDr\tMV\tDn\tMV\tDm\tMV\tDiv\tMV\tAD\n\u201422.9\t+22.8\t2.0\t+33.8\t1.8\t+31.8\t1.7\t+27.8\t1.9\t+28.9\n\u201420.9\t+18.5\t0.8\t+31.0\t1.4\t+28.3\t1.5\t+23.3\t2.8\t+25.3\n\u201418.1\t+15.5\t0.9\t+25.5\t2.1\t+23.8\t1 0\t+19.0\t2.1\t+20.9\n\u201416.3\t+13.5\t1.1\t+19.0\t3.0\t+20.1\t1.2\t+ 16.5\t2.0\t+17.3\n\u201413.6\t+11.8\t1.3\t+16.8\t2.8\t+15.8\t1.0\t+ 14.8\t2.2\t+14.4\n\u201410.7\t+ 8.5\t1.7\t+10.8\t3.1\t+ 12.5\t1.4\t+11.5\t2.4\t+ 10.8\n\u2014 7.5\t+ 5.8\t1.2\t+ 6.3\t4.8\t+ 8.3\t1.8\t+ 6.5\t2.1\t+ 6.7\n\u2014 3.9\t+ 3.5\t1.6\t+ 1.3\t3.0\t+ 3.8\t2.0\t+ 4.0\t1.9\t+ 3.1\n0\t+ 2.3\t1.6\t\u2014 2.3\t3.2\t\u2014 1.0\t1.5\t\u2014 0.5\t1.1\t\u2014 1.5\n+ 4.3\t\u2014 0.8\t1.2\t\u2014 7.5\t2.5\t\u2014 5.3\t0.2\t\u2014 3.0\t2.0\t\u2014 4.1\n+ 9.0\t\u2014 4.0\t0.9\t\u201411.0\t4.2\t\u201410.5\t1.7\t\u2014 7.0\t2.2\t\u2014 8.9\n+14.2\t\u2014 7.0\t1.5\t\u201415.3\t3.3\t\u201415.6\t0.5\t\u2014 12.5\t3.0\t\u201412.6\n+19.9\t\u2014 9.8\t1.4\t\u201419.3\t2.3\t\u201420.5\t1.0\t\u201418.0\t2.8\t\u201416.9\n+26.2\t\u2014 12.3\t1.5\t\u201421.8\t2.5\t\u201425.0\t1.1\t\u201421.3\t3.3\t\u201420.1\n+33.1\t\u201416.0\t2.1\t\u201423.8\t2.6\t\u201428.2'\t2.6\t\u201425.0\t1.8\t\u201423.0\n+40.7\t\u201416.0\t1.6\t\u201426.8\t4.0\t\u201432.6\t3.8\t\u201430.3\t2.1\t\u201426.4\n+49.0\t\u201422.8\t1.8\t\u201433.0\t2.8\t\u201437.4\t4.5\t\u201437.3\t4.2\t\u201432.6\n0, number of millimeters by which the diameter in Set A differed from that in SetB.\nDi, Du, Dm, Dir, number of grams by which the estimated weight of the block in Set A differed from its true weight, for\nfour observers, I, II, III, IT ; averages of 20 experiments each.\nAD, average for the four observers.\nMV, mean variation; the mean variation for the series is obtained by dividing by 4.5\naccuracy of the observer was obtained by employing the method of suspending the blocks which will be explained in the next series of experiments. This test was made after the twenty regular experiments. The observer seated himself behind a screen and, keeping his hand in a comfortable position on the other side of the screen, lifted the blocks by the handles in such a manner that he had no knowledge or intimation of the size of the bodies lifted. The mean","page":6},{"file":"p0007.txt","language":"en","ocr_en":"Measurements of illusions and hallucinations in normal life. 7\nvariation thus found for the whole set of blocks is slightly less than the mean variation recorded in the table for twenty trials on any single block.\nThe results of these experiments, showing the persistence of the illusion when the actual weight is not known, are given in Table II.\nThis table proves that the illusion is regular and persistent. It cannot be eliminated by practice as long as the actual weight is not known.\nPersistence of the illusion of weight\nA&B.fact of illusion unknown C.FACT OF ILLUSION known\nThis persistence of the illusion justifies the procedure of making successive tests of the principle on the same observer under varied conditions for the purpose of comparison.\nB. Actual relations of weight known.\nDoes the illusion persist when the observer is familiar with the apparatus and knows the nature of the illusion and the actual relations of weight? To test this, I used the same apparatus as before and the same method, except that the observer was made acquainted with every essential of the apparatus before beginning the experiment.\nFor observers I secured ten professors and graduate students who had all done special work in psychology and could be depended upon for reliable introspective analysis and critical judgment. Each","page":7},{"file":"p0008.txt","language":"en","ocr_en":"8\nC. JS. Seashore,\none knew the general trend of the illusion as found in previous measurements, but no one knew what the extent of the illusion was for other persons under the present conditions. They knew the actual weight and were to determine the apparent weight. This was very difficult because there was a conflict between knowledge and feeling, and the former was to be governed by a conscious effort. In previous experiments the observer was supposed to exercise all possible skill in matching the blocks correctly by lifting them, but\nTable III.\nPersistence of the illusion of weight when the fact of the illusion is known.\nG\tAD\tMV\n\u201422.9\t+16.5\t4.0\n\u201420.9\t+16.5\t4.2\n\u201418.7\t+ 13.5\t4.1\n\u201416.3\t+11.0\t4.5\n\u201413.6\t+11.0\t4.9\n\u201410.7\t+ 7.8\t5.1\n\u2014 7.6\t+ 6.5\t3.4\n\u2014 3.9\t+ 5.0\t4.0\n0\t+ 2.3\t4.4\n+ 4.3\t\u2014 1.0\t3.2\n+ 9.0\t\u2014 6.0\t3.0\n+14.2\t\u2014 6.5\t3.7\n+19.9\t\u201411.5\t3.9\n+26.2\t\u201411.5\t2.2\n+33.1\t\u201414.5\t3.5\n+40.7\t\u201417.0\t3.5\n+49.0\t\u201419.5\t5.1\n0, number of millimeters by which the diameter in Set A differed from that in Set B.\nAD, number of grams by which the estimated weight of the block in Set A\ndiffered from its true weight; average for ten observers.\nMV, mean variation; the mean variation for the series is obtained by dividing by 3.2.\nhere it was required to discriminate between known weight and felt weight and to estimate the amount of difference by matching the blocks as before.\nThe full record of the experiments upon these ten observers is contained in Table III.\nThe curves in fig. 3 give a diagrammatic representation of the intensity of the illusion (A,B) when the fact of the illusion is unknown, and (C) when it is known. A is the curve from Table I, fig. 2, placed","page":8},{"file":"p0009.txt","language":"en","ocr_en":"Measurements of illusions and hallucinations in normal life. 9\nhere for comparison ; B, dotted line, is the average for the four observers, Table II ; and C is the average for the ten observers, Table III.\nThe coincidence of the judgments of the observers is an extraordinary phenomenon, considering that the discrimination is very difficult under these circumstances, and that no one could he influenced by the records of others. The conclusions are plainly warranted that :\n1.\tThe illusion of weight persists even after the observer knows its nature, extent, and cause ;\n2.\tIts intensity is somewhat less than when the actual conditions are unknown to the observer.\nTo carry the test further and require the observer to repeat it a great number of times, would be to find out how long it would take him to learn to make the proper allowance for the illusory feeling. It is difficult to say whether that feeling can be wholly educated away, but we must suppose that a person can finally educate himself to make proper allowance for it or neglect it.\nThird series of experiments : Dependence of the illusion of weight upon the directness of sight.\nThus far only the diameter in Set A had been varied. In the present series another quantity, the directness of sight, was varied. Each experiment consisted of four sets of tests. The aim was to secure a judgment of weight when the size of the bodies compared was made prominent in different degrees as follows : (1) size estimated by direct sight, (2) size estimated by indirect sight, (3) visual memory of size, and (4) no knowledge of size.\nThese conditions were obtained through the following four sets of variations, corresponding respectively to the above requirements :\n1.\tThe conditions of the first set were fulfilled by the method pursued in the previous experiments. The observer looked directly at the block as he grasped it between the thumb and second finger, and its proportions were emphasized by the comparative amount of space it occupied in the limited opening between the grasping fingers. The effort to grasp the disks accurately attracted the attention of his eyes in that direction.\n2.\tIn this series of experiments, additional apparatus was required for suspending the blocks. It consisted of handles, made of exactly the same length as the blocks and supplied with disks on the ends,","page":9},{"file":"p0010.txt","language":"en","ocr_en":"10\nG. E. Seashore,\nmade of the same material and in the same shape as the disks on the blocks. To secure lightness they were made in the shape of a slender spool. These handles rested on a support in the same position as the tray, and from them the blocks were suspended 100mm below, by means of silk cords. Each block was supplied with a loop of silk cord, which could readily be slipped off and on a hook at the lower extremity of the suspended cord. The combined weight\nTable IT.\nIllusion of weight under various circumstances.\na\tD\tMV\tE\tMV\tF\tMV\tG\tMV\n\u201422.9\t4-24.5\t8.9\t4-12.0\t6.0\t4- 8.5\t4.4\t\u20140.5\t2.4\n\u201420.9\t4-23.0\t8.0\t4-10.0\t5.0\t+ 7.5\t3.8\t4-1.5\t3.7\n\u201418.7\t4-19.5\t9.0\t+ 9.0\t5.0\t4- 6.5\t3.9\t\u20140.5\t3.0\n\u201416.3\t4-19.5\t8.4\t4- 6.5\t45\t4- 2.5\t3.6\t0\t2.0\n\u201413.6\t4-17.0\t7.8\t4- 4.0\t5.0\t4- 1.5\t4.8\t4-0.5\t1.5\n\u201410.7\t4-12.0\t9.3\t4- 3.5\t3.6\t4- 2.5\t4.6\t\u20140.5\t1.5\n\u2014 7.5\t4- 9.0\t6.0\t4- 2.5\t3.9\t4- 3.5\t5.0\t\u20140.6\t2.5\n\u2014 3.9\t4- 4.5\t2.5\t4- 2.0\t3.8\t4- 1.0\t3.6\t\u20141.5\t2.5\n0\t4- 1.5\t2.8\t\u2014 1.0\t4.0\t4- 0.5\t4.3\t4-0.5\t2.5\n+ 4.3\t\u2014 1.6\t2.7\t\u2014 2.5\t3.7\t\u2014 0.5\t5.1\t4-1.0\t1.0\n+ 9.0\t\u2014 6.0\t2.8\t\u2014 4.5\t5.5\t\u2014 2.5\t3.8\t\u20141.0\t1.6\n4-14.2\t\u2014 7.5\t2.6\t\u2014 6.0\t4.6\t\u2014 3.5\t4.0\t\u20141.0\t1.7\n4-19.9\t\u201414.0\t3.2\t\u2014 7.0\t4.9\t\u2014 4.5\t3.8\t\u20141.0\t5.8\n4-26.2\t\u201419.0\t4.2\t\u2014 7.0\t4.6\t\u2014 4.5\t4.5\t\u20141.5\t4.6\n4-33.1\t\u201418.5\t3.5\t\u2014 9.5\t54\t\u2014 5.5\t1.5'\t\u20141.0\t4.0\n4-40.7\t\u201424.5\t4.8\t\u201410.5\t7.0\t\u2014 8.0\t2.6\t4-0.5\t3.5\n4-49.0\t\u201427.5\t5.2\t\u2014 12.5\t8.2\t\u201412.0\t4.9\t\u20142.0\t3.4\nG, number of millimeters\t\t\tby which the\tD,\twith direct vision of size.\t\t\t\ndiameter\tin Set A differed\t\tfrom that in\tE,\twith indirect vision of\t\tsize.\t\nSet B.\nI), E, F, O, number of grams by which the estimated weight of the block in Set A differed from its true weight; averages for ten observers in each series.\nF,\twith visual memory of size.\nG,\twith no knowledge of size.\nMV, mean variation; the mean variation for the series is obtained by dividing by 3.2.\nof a handle with hook and cord was 2\u00ae. Employing this method of suspending the blocks, the observer naturally directed his attention to the handle by which he lifted. Since the blocks hung 100mm below the handle, they were out of the line of direct vision and not in contact with any body whose contiguity would limit the space occupied or contrast its size.","page":10},{"file":"p0011.txt","language":"en","ocr_en":"Measurements of illusions and hallucinations in normal life. 11\n3.\tThe weights were again suspended. The observer looked at each block as it was hooked on but shut his eyes while lifting it.\n4.\tThe observer was blindfolded for the final set. He was then required to lift the suspended blocks by the handles without knowing whether he was lifting a big one or a little one. He remembered the range of the sizes but did not know which block in the series was presented.\nThe purpose of the variations was not told to the observer. He was simply asked to subject himself to four tests necessarily to be taken successively on the same observer. Definite instructions were\nInfluence of size on judgments OF WEICHT\nD, OINECT VISION\nE INDIRECT VISION\nF\tOF SIZE\n. VISUAL MEMORT\nG NO KNOWLEDGE\nFig. 4.\ngiven at each step so far as this could be done without making any suggestions pertaining to the illusion. The whole experiment on one observer occupied from 80 to 100 minutes, sufficient rest being allowed between the different sets. The observers were especially cautioned not to let the results of one test influence their judgment in another.\nThe following statements are based upon the results contained in Table IV (fig. 4).\n1.\tVisual knowledge of size causes the greatest illusion, in determination of weight, when the body lifted is directly looked at and its size is brought into prominence by the comparative amount of space that it occupies in a limited area.\n2.\tThe influence of size on judgment of weight is lessened when the object is placed out of the line of direct vision and isolated from any object whose contiguity would favor a comparison of size.","page":11},{"file":"p0012.txt","language":"en","ocr_en":"12\nG. JE. Seashore,\n3.\tVisual memory of size causes a less intense illusion of weight than that which is produced by looking at the blocks while lifting*\n4.\tSize has no influence on the perception of weight when the observer has no knowledge of it.\nfourth series of experiments : Dependence of the illusion of weight upon the senses by which knowledge of size is acquired.\nIn all the foregoing experiments the perception of difference in size was purely visual. The next problem was to settle the question : Does the illusion of weight vary with the different senses by which the image of size is produced ?\nIt is impossible to draw a strict line of demarcation between the different sensory elements which, besides sight, unite in building up a percept of size, because they generally cooperate, and their effects fuse. For the purpose of comparison, I aimed to get a measure of the illusion of weight due to size when knowledge of size was acquired through each of the following channels predominatingly : (1) muscle sense, (2) touch, (3) sight, (4) muscle sense, touch and sight.\nThese conditions were approximately attained by the following respective methods of procedure.\n1.\tWithout having seen the blocks, the observer seated himself on a high stool behind a screen, in such a position as was occupied when he stood by the tray and looked at the apparatus, so that his arm and hand would be comfortably adjusted over, the tray, on the other side of the screen. The blocks were then handed him in the same order as before, but placed on end on the tray, so that by dropping his thumb and fingers around a block the observer included it in his grasp, holding it by the circumference ; he thus acquired a knowledge of its size mainly by the muscle sense in the fingers.\n2.\tSeated in the same position, after a brief rest, the observer held out his hand, not resting it on anything, and the blocks were placed on his flat palm, one at a time. An image of the height (length) of the block was inevitably transmitted from the experience in the first set, but the proportional size was here estimated chiefly by the area of touch or pressure.\n3.\tThe method of estimating comparative size by direct sight, as pursued in the previous experiments, eliminates all other sensory elements than sight.","page":12},{"file":"p0013.txt","language":"en","ocr_en":"Measurements of illusions and hallucinations in normal life. 13\n4. In the final set the blocks were again placed on end, and the observer occupying the same position as in the third set, grasped the blocks by the circumference as in the first set of this series and, in addition, looked directly at them. Hence, there was the combined effect of muscle sense, touch and sight.\nUpon the basis of the results of the second series of experiments,\nI considered it justifiable to make these four successive tests on the same person. Furthermore the advantage of the order will be noticed. In the first and second sets he had not seen the blocks. The estimate of the length was a proper influence to carry over to the second set ; it was, namely, a memory of the fact that the length was a constant quantity. The third set could not be influenced by any previous judgments, because the record was kept secret, and the blocks had not been recognized. In the fourth set a warning against influence from previous judgments was really superfluous, because the observer was after truth and not consistency.\nNo test on the error of observation, or mean variation when there was no influence of sight, was here included, because that would practically be the same as found in the fourth set of the third series, and the record of A (9) in each set.\nThe sensations of touch and pressure cannot be distinctly separated. Indeed, since the blocks cannot be placed without some velocity, there is really a sense of impact in the second set. For brevity, the space-giving characters of these sensations is spoken of as \u201c touch.\u201d\nThat muscle sense, touch, sight and other sensations cannot be isolated does not detract any from the value of this series of experiments. In the first set, skin and joint sensations assist ; and in the second set, joint and muscle senses are slightly involved.\nIn the third set, all of these senses join in determining the size of a body, but since all but one are constant, it remains for that sense, sight, to discriminate for differences in size.\nIn the first set, the observer had not seen the blocks. The experience that they differed in size puzzled him. Hence every time he picked one up his first concern was to fix an image of its size by estimating the relative extent of the grasp required in lifting it. In the second set, attention was again called to the varying size in a similar manner, but here the estimate of the dimensions was more indefinite and uncertain. In the third set, he determined the dimension at a glance, without any effort, and hence did not concern himself particularly about the estimation of size. But, in the final set, his attention was again called to the appreciation of size, because","page":13},{"file":"p0014.txt","language":"en","ocr_en":"14\nG. JS. Seashore,\nhere he noticed that he combined several sensory elements, in that way acquiring a very accurate knowledge of size.\nIn brief, by the above four variations, different degrees of attention were given to discrimination of size, while to avoid the illusion all attention should have been given to the perception of weight.\nAccording to this method a constant error, due to the fact that A was lifted first, will be implied in the first two sets. I could not\nTable V.\nIllusion of weight for different senses.\nG\tH\tMV\t1\tMV\tJ\tMV\tD\tMV\n\u201422.9\t+27.5\t10.0\t+12.5\t16.5\t+21.5\t9.0\t+20.5\t9.1\n\u201420.9\t+20.0\t12.2\t+ 15.0\t13.0\t+ 16.0\t8.8\t+20.5\t8.8\n\u201418.7\t+21.5\t13.0\t+10.0\t12.5\t+17.0\t8.6\t+ 19.5\t8.5\n\u201416.3\t+ 12.5\t9.4\t+ 6.0\t11.7\t+12.5\t9.5\t+17.5\t5.4\n\u201413.6\t+ 9.0\t6.6\t+ 1.0\t12.5\t+ 7.0\t7.6\t+18.5\t7.4\n\u201410.7\t+ 4.0\t9.0\t+ 5.5\t9.4\t+ 8.0\t8.0\t+10.0\t6.0\n\u2014 7.5\t+ 2.0\t6.2\t+ 0.5\t8.8\t+ 3.5\t7.4\t+11.0\t7.2\n\u2014 3.9\t0\t3.0\t- 1.0\t7.1\t+ 5.0\t6.5\t+ 4.5\t6.5\n0\t- 7.0\t6.2\t- 5.0\t5.5\t- 1.0\t8.1\t- 1.0\t6.8\n+ 4.3\t-12.0\t5.5\t- 9.5\t6.3\t\u2014 2.5\t9.8\t- 1.5\t7.2\n+ 9.0\t-18.5\t7.0\t\u2014 14.0\t7.8\t- 5.5\t7.5-\t- 4.0\t9.3\n+14.2\t-22.0\t8.4\t-18.0\t6.0\t\u2014 7.5\t5.8\t- 9.5\t6.5\n+19.9\t-26.5\t7.5\t-19.5\t7.5\t\u2014 9.0\t6.8\t-14.5\t4.9\n+26.2\t-28.5\t6.8\t-24.5\t8.2\t-15.0\t9.0\t-15.5\t5.2.\n+33.1\t-35.0\t5.0\t-33.5\t7.2\t-13.0\t7.2\t-22.5\t6.6\n+40.7\t\u201434.0\t4.4\t\u2014 31.0\t8.7\t-18.0\t8.1\t-25.5\t7.5\n+49.0\t-38.0\t4.9\t-33.0\t11.0\t-20.0\t8.0\t-28.0\t8.0\nG,\tnumber of millimeters by which the diameter in Set A differed from that in SetB.\nH,\tI, J, D, number of grams by which the estimated weight of the block in Set A differed from its true weight ; averages for ten persons.\nS, size known by muscle sense.\n1j size known by touch.\nJ' size known by direct sight.\nA, size known by muscle sense, touch and sight.\nMV,\\ mean variation; the mean variation for the series is obtained by dividing by 3.2.\neliminate it by repeating the test in the reverse order, because it would be too fatiguing to take each set twice. It was more important to take the four sets in regular succession than to divide up the experiment in order to eliminate the error. Hence, the fact of this constant error must be borne in mind in interpretation of the results. An approximate correction of it may be made by noticing its extent in the middle block of Set A.","page":14},{"file":"p0015.txt","language":"en","ocr_en":"Measurements of illusions and hallucinations in normal life. 15\nThe results are given in Table V and are expressed in the curves of fig. 5.\nThe variation of the intensity of the illusion of weight, due to the acquisition of knowledge of size by different senses, may be formulated on a comparison of the above four sets of results as follows :\n1.\tThe illusion of weight dependent on size is greatest when size is estimated mainly by muscle sense, and the weights have not previously been seen.\n2.\tThe illusion is more fluctuating and on the whole not quite so strong when size is estimated by the area of pressure in the flat palm, including a memory of the third dimension.\nIllusion OF WEICHT FROM OIFFEREHT SENSES\nSize ESTIMATED BY H. MUSCLE SENSE I .TOUCH i. SIGHT\nD. MUSCLE SENSE TOUCH & SIGHT\nFig. 5.\n3.\tIn these variations the illusion is weakest when size is'estimated by direct sight.\n4.\tWhen size is estimated by the combined effect of all the spatial senses, the illusion is weaker than when depending on muscle sense or touch and stronger than when dependent on sight alone.\nFifth series of experiments : Illusion of weight due to the knowledge of the material of which the weights are made.\nThis series of experiments pertains to a corollary of the foregoing conclusions. Is there an illusion of weight dependent upon disappointed expectation caused by a supposed knowledge of the material","page":15},{"file":"p0016.txt","language":"en","ocr_en":"16\nG. E. Seashore,\nlifted ? Theoretically, if size has a suggestive influence upon judgment of weight by leading the observer to expect a body to be heavy or light according as it is large or small, a similar association between judgment of weight and knowledge of the specific gravity of the material lifted ought to caifse an illusion of weight, under favorable circumstances. If a body, seen to be made of a heavy material but containing a light deceptive core, be lifted, it will rise too easily because lifted with an effort calculated to raise a heavier weight, and it ought to be underestimated for the same reason that a large block in Set A is underestimated. In the same manner, if a body apparently made of a solid piece of light material but containing a heavy deceptive core be lifted with a motor adjustment adapted to the apparent weight, it will require more effort than was expected to raise it, and it ought to be overestimated for the same reason that the small block in Set A is overestimated.\nThis problem suggests a wide range of experiments, but there are very great difficulties in the way of an accurate measurement. The methods I have employed in this series are only partially successful. An attempt was made by three sets of experiments to bring out the following variations in the conditions of the bodies to be lifted :\n1.\tSame size and weight but different material ;\n2.\tSame size but different material and weight ;\n3.\tDifferent size, weight and material.\nFirst set.\nFour blocks were made of the same shape and size as the B\u2019s, i. e. 42.9mm in diameter, and the same weight as the A\u2019s, i. e. 80s. One was made of cork, one of wood, one of copper and one of lead. The first two were filled and the last two were cored out until each one weighed 80s. This was done so carefully that the surfaces in their natural color suggested that the weights were turned out of solid pieces of the respective materials and were uniform throughout. It was thought that the observers would possess a fairly definite .knowledge of the specific gravity of these four kinds of material.\nThe method was to require the observer to take one of these at a time and match it in the original Set B. Ten observers tried the experiment, but the amount of error which may be due to the illusory influence is not more than what must be allowed for the mean variation. This set was a failure because the lead was too light to allow any one to suppose that it was solid, and the cork was so heavy that","page":16},{"file":"p0017.txt","language":"en","ocr_en":"Measurements of illusions and hallucinations in normal life. 17\nthe observer at once concluded that it was filled. The illusion was quite effective in the wood and the copper blocks, but the judgments of these were influenced by the deceptive appearance of the other two, so that the mean variation was large.\nSecond set.\nThree blocks were made the same size as the B\u2019s, i. e. 42.9mm in diameter, but of the following materials with their respective weights : (1) cork, natural weight 40g, loaded to 55g ; (2) wood, natural weight 65g, loaded to 80s ; (3) lead, natural weight not determined, cored out so as to weigh 120s. Their appearance suggested that they were uniform throughout. The aim was to make them feel different from what the observer expected them to feel ; but not so much that he would suspect that they had a false appearance. Nothing was said by explanation or question as to the character of the material before beginning the experiments. After these experiments about 60$ of the observers stated that they had taken the blocks to be made solid of the materials seen on the surface. About 20$ were uncertain, and the rest thought that they had a deceptive appearance.\nFourteen observers gave this result in selecting corresponding B\u2019s for the lead block : one overestimated, six right, and seven underestimated. The average amount of the underestimation was 8s.\nIn the same manner fourteen observers gave the following results in matching the cork block : one underestimated, four right, and nine overestimated. The average amount of the overestimation was 7g, which may be considered as due to the 15g surreptitious weight.\nThe twenty-nine observers who matched the wooden block gave these results : three underestimated, seven right, and nineteen overestimated. The average amount of overestimation was 10g.\nThe apparatus cannot be counted a real success until it contains an appreciable illusion for all observers and still has no suggestion of deceptive appearance. Nevertheless this fragmentary set of experiments shows that :\n1.\tWhen a body containing a surreptitious light core is lifted with a motor adjustment adapted to the expected weight of a certain material and it feels lighter than was expected, that body will be underestimated in weight ; and\n2.\tWhen the body lifted contains a deceptive heavy core, the observer will overestimate the weight of that body, especially if he does not suspect the presence of the core.\n3","page":17},{"file":"p0018.txt","language":"en","ocr_en":"18\nC. JS. Seashore,\nThird set.\nThe largest and the smallest of the blocks of Set A were used and in addition two extra blocks were made, here called respectively P and L for brevity. P was a solid piece of pine wood turned to exactly the size of the largest A, which made it weigh 75\u00ae. L was a solid piece of lead turned in the shape of the smallest A until it weighed 75s ; the length was kept the same as that of A, but the diameter had to be reduced to 16mm in order to secure the proper weight. This made it a little more than two steps smaller than the smallest A, according to the ratio of the diameter of the A\u2019s.\nThe aim was to determine whether the illusion would be greater between the largest and the smallest blocks in A, which appeared to be made of the same material, than between P and L, whose apparent material suggested the proper weight. By the method followed each of these four blocks was to be matched twice in Set B, taking them in reverse order so as to avoid errors due to the order of sequence. Eighteen observers had two trials each. The averages of the B\u2019s selected were : For largest A, 4.6 ; for smallest A, 13.4 ; for L, 12.8 ; and for P, 5.3 ; with the corresponding average mean variations : For largest A, 6s ; for smallest A, SF ; for L, 10g ; and for P, 7s. Hence, although P weighed 5* less than largest A, it was estimated to be on the average 3.5\u00ae heavier, making the illusion in P 8\u00ae less than the illusion in largest A. And, although L was over two steps smaller than smallest A, and weighed 5\u00ae less, it was judged to weigh only three grams less. Taking the average of the results previously obtained for the A\u2019s, the illusion for L is found to be about 7s less than for smallest A.\nIn the thirty-six trials the illusion was stronger for largest A than for P twenty-nine times, equal six times and weaker once. Similarly it was stronger for smallest A than for L thirteen \u2022 times, equal nine times, and weaker only fourteen times, although L was over two steps smaller than smallest A.\nI place no great confidence in the quantitative determinations in this series of experiments, because not all the conditions were sufficiently controlled, but the general qualitative results are positive. It has been proved in two different ways that knowledge of the material lifted influences the judgment of weight, namely :\n1. When there is no illusion on account of size, there is under certain circumstances an illusion caused by knowledge of the material lifted ; and,","page":18},{"file":"p0019.txt","language":"en","ocr_en":"Measurements of illusions and hallucinations in normal life. 19\n2. When there is an illusion due to size, it is not so strong when the difference in size is proportional to the specific gravity of the material lifted as when there is no definite knowledge of the material.\nGeneral remarks on the experiments in Part First.\nOwing to the nature of the subject the method employed is necessarily a combination of the experimental and the statistical. It is experimental because, whenever practicable, I have varied the conditions and repeated the tests on the same individual. On the other hand, a sufficient number of persons from the various vocations have been tried to satisfy the statistical demands and justify the conclusion that the main laws derived are fairly universal.\nApparatus.\nThe difference between the B\u2019s was made 5g, because in an arithmetical series that is about the average limit of discrimination for weight within the range here investigated. The possible variable error due to this cause is \u00b12.5*. This source of error is evenly distributed throughout the series ; hence, it does not alter the expression for the general intensity of the illusion, but merely causes minute fluctuations in the curves of the results.\nIf a wider range had been investigated, the B\u2019s should have been made to vary in weight by an increment of about according to Weber\u2019s law, in order that the perceptions of difference should have been equally appreciable in discriminative consciousness. Here a constant difference was used merely for convenience, but that arrangement ought not to modify the results appreciably.\nThe additional 2s of the suspending apparatus is the same for all the blocks, and this increase in the total range of the weight has so small effect on the results that it may practically be neglected.\nIn a series of cylinders of equal length but varying diameter, there is a constant illusion in regard to length. The apparent length increases with the decrease of the diameter. In order that these blocks should have appeared to be of the same length they should actually have varied in geometric ratio with a very small increment. This could not be done in the present test because one of the required conditions was that the grip should be uniform.\nThe temperature of the blocks was kept practically constant.","page":19},{"file":"p0020.txt","language":"en","ocr_en":"20\nG. E. Seashore,\nObservers.\nThe following essential requisites were kept in view in selecting observers :\n1.\tThey shall be careful and competent observers ;\n2.\tThey shall have no previous knowledge of the special principles involved or of the direct purpose of the experiment ;\n3.\tThey shall not know how or of what material the blocks are constructed ;\n4.\tThey shall be in a comfortable condition and free from emotional disturbances ;\n5.\tThey shall be trusted not to work according to any prejudicial scheme or preconceived theory.\nThese qualifications could only be secured in a relative manner, but the aim was scrupulously adhered to so far as resources and temporary conditions permitted. Discrepancy in the attainment of these requirements accounts for many of the partial variations from the general law of the illusions. The requirements were altered for the fifth and the latter part of the second series so as to suit the special demands of those cases. On the whole, I was fortunate in securing the assistance of reliable observers, who worked with a scientific interest.\nElimination of errors in method.\nThere is a law that when two bodies of approximately the same, weight and size are lifted in succession; the weight of the latter is overestimated. The extent of this error varies considerably for different persons and for the same person at different times. In the present experiments it was, however, practically eliminated by requiring the observer to lift two blocks, to be compared alternately several times before forming his decision.\nIn certain cases there may have been a slight error due to the order in which the blocks were taken up to be matched. This applies to both the A\u2019s and the B\u2019s. Thus in the second series it was found that the curve of the illusion is a little more regular when the A\u2019s are taken in order than when they are taken at random. In the same manner the illusion might depend slightly on how many B\u2019s were picked up and tried before a decision was reached. These influences may have affected some single choices, but they do not appreciably change the expression for the general intensity of the illusion. This was proved by actual measurements.","page":20},{"file":"p0021.txt","language":"en","ocr_en":"Measurements of illusions and hallucinations in normal life. 21\nThe memory image for weight sinks very rapidly during the first ten seconds. The blocks were therefore placed so that the observer could change from one to the other by the least possible movement and change of position. He could deliberately lay down one and pick up another in three seconds. The time required for this transition was the same for all the blocks.\nThere is an illusion of weight caused by fatigue in lifting objects. This illusion did not affect the present experiments because the bodies compared were lifted alternately several times. Furthermore the number of trials was always limited to such an extent that there would be no necessary cause for any disturbing fatigue from continuous exertion.\nThe known error in judgment of weight caused by lifting a light body after a heavy one, was here avoided by the fact that the final discrimination was not made until the observer had reached two blocks that appeared to weigh the same. The error was frequently involved in the first trials of any experiment where there was a decided difference between the blocks, but as the observer found his way up to those that appeared more and more alike, it was gradually eliminated before he reached a final decision.\nThe weight of a body depends upon its mass and the velocity with which it is lifted. Therefore special precaution was always taken to get the observer to lift the block to the same height and with the same velocity as nearly as possible.\nIn the case of direct sight the observer would often for a moment try to avoid the distraction of sight by not looking at the block. Such judgment would then inevitably be exceptional and be governed by the conditions of indirect sight or visual memory.\nSome partial variations from the general trend of the illusion in direct sight are due to this cause. When the observer took such liberty only once or twice, I did not interfere, but if he proposed to continue I was obliged to inform him, in some guarded statement, about the conditions required.\nSuggestions for further experiments.\nI chose the diameter of middle A as the diameter for all the B\u2019s, merely because it was the middle. The B\u2019s might have been made any size within the range of the A\u2019s. As a development and further proof of the present test, the following variations might be tried.","page":21},{"file":"p0022.txt","language":"en","ocr_en":"22\nG. E. Seashore,\n\u00ef. Construct the B\u2019s all the same size as the smallest A, making one 90g and the rest decreasing by steps of 5g each, down to 25\u00bb. That would allow the same range of the illusion as the present apparatus, since the smallest A might be judged fully three times as heavy as the largest A.\n2.\tConstruct the B\u2019s all the same size as the largest A, making the lightest 70\u00ae and the rest increasing by 10s each step up to about 240\u00ae.\n3.\tSimilarly construct the B\u2019s in separate sets respectively of the same size as each of the intermediate A\u2019s.\nFrom tentative tests made it appears that the illusion would have, approximately the same intensity under these conditions as was determined by the apparatus and method employed. The difference between the principle of the first two classes of apparatus suggested and the one employed is that, according to the former, a quantity is measured by a direct comparison which, according to the latter, is measured by an intermediate step. This would be true when two A\u2019s are compared by means of the B\u2019s. When an A is compared with a B, the process is the same in both cases. It will be observed that in all these suggested experiments it is the same objects that are measured, but the measuring is done by different means in each set of experiments.\nHow does the illusion vary with a greater variation in diameter of the cylinders, with different ranges of weight, and with variation in the shape of the weights? These and similar questions demand solution by experiment, but the measurements can not be made as definite and as accurate as those in the present research.\nDr. Scripture has a set of weights made large enough to use in demonstrations before large audiences. It consists of cubical wooden boxes loaded with lead to the desired weights. There are only two corresponding to my A\u2019s, the smaller 8.2om cube and the larger 60cm cube, each weighing 8f lbs. Those that correspond to my B\u2019s are all 15cm cube and range by f lb., each step from 2 lbs. up to 17 lbs.\nI tried the experiment with this apparatus on four men with the following result : The large box was matched with boxes of 5.75 lbs., 4.25 lbs., 4.25 lbs., and 2 lbs., by the respective observers\u2014 average 4 lbs.\u2014that is there was an underestimation of 3 lbs., 44-lbs., 4.25 lbs., and 6.75 lbs. respectively\u2014average 4.4 lbs. The small box was matched with 8.5 lbs., 14.75 lbs., 13.25 lbs., and 10.25 lbs. respectively\u2014average 11.9 lbs.\u2014that is, there was a corresponding overestimation of 0.5 lb., 6 lbs., 4.25 lbs., and 1.5 lbs.\u2014average 2.9","page":22},{"file":"p0023.txt","language":"en","ocr_en":"Measurements of illusions and hallucinations in normal life. 23\nlbs. That means that mainly on account of this difference in size the observers made a difference of 2.75 lbs., 10.5 lbs., 9 lbs., and 8.25 lbs. respectively\u2014average 7.6 lbs.\u2014between the two boxes, which both had the same weight, 8.75 lbs. That is, the smaller is judged to weigh 2.9 times as much as the larger. This, it will be observed, slightly exceeds the average amount of the illusion between the extremes in my set.\nThis illusion is based upon the difference in volume, but a comparison between Dr. Scripture\u2019s set and mine shows that it depends very much upon the shape of the weights and whether two or three dimensions are varied. The diameters of the smallest and the largest A are in the ratio of 20: 91.9; while the one dimension of the boxes stands in the ratio 8.2 : 60, and still the illusion is not much stronger in the latter case. It is therefore evident that it does not vary directly with the volume. The range of the weight is the great factor which must here be taken into consideration. It would be very desirable to try the folio wing variations :\n1.\tExtend Set A with additional blocks on both ends, making the smallest as small as possible when a shell is filled with mercury to 80\u00ae and the largest as large as possible without exceeding the given weight. Handle them in the most convenient way and match them in an extended series of the B\u2019s to determine whether the illusion varies with the variation in diameter according to Weber\u2019s law.\n2.\tWith the same scale of variations in size make different sets of different weight. Sets of 40\u00ae, 160\u00ae, and 320\u00ae would perhaps be the best weights to try in addition to the original set of 80\u00ae.\n3.\tSince the illusion depends very much upon the shape of the body lifted, it would be very interesting to determine the comparative amount of the illusion when all three dimensions are varied as in regular cubes or spheres.\nDresslar' found that of bodies (lead sheets) which have the same weight and the same size but different shape, the one which appears the most compact, the circle, will be judged to be the heaviest. The same principle ought to apply to the cube and the sphere ; and it actually appears that this is confirmed by Dr. Scripture\u2019s cubical boxes as well as by Charpentier\u2019s1 2 brass balls.\nSince on impact the weight of a body varies with its mass and the square of the distance through which it falls, the normal association\n1 Dresslar. Studies in the psychology of touch, Am. Jour. Psych., 1894 IV 49.\n\u201cCharpentier, Analyse de quelques \u00e9l\u00e9ments de la sensation de poids, Archives de Physiol., 1891 (5) III 126.","page":23},{"file":"p0024.txt","language":"en","ocr_en":"24\nG. E. Seashore,\nbetween weight and velocity of impact is quite strong and definite. Hence on the basis of what has been determined the following experiment may be suggested : Construct an apparatus such that a ball may appear to fall through a given distance and impinge upon the hand ; but, by a secret device, it falls only a part of' that distance or conversely, through a greater distance. The disappointed expectation in regal\u2019d to the velocity of impact ought to cause an illusion which might readily be measured.\nThe difficulties encountered in the fifth series of experiments urge the investigation of another problem, the effect of color on discrimination of weight. Several observers stated that it had a decided influence on their judgments. The experiment might be performed somewhat as follows : Make a standard set of twenty blocks corresponding to the B\u2019s above, of an \u201c indifferent \u201d color. Let these vary by steps of 3\u00ae each, making the ninth weigh 80s. Make another set of eighteen blocks, nine of the \u201c indifferent \u201d color and the remaining nine of each of the respective spectral colors, together with black and white. Let these last nine be weighted to 80s ; four of the \u201cindifferent\u201d to 77K, 74s, 71s, and 68s each ; and the remaining five \u201cindifferent\u201d to 83s, 86s, 89s, 92s, and 95s each. Require the observers to match each block from the second set with one in the first set, which will serve as a measure.\nDresslar1 seems to make an unjustifiable inference in regard to the variation of this illusion with the degree of intelligence of the observer. He used eight cylinders having a constant weight 132s, a constant diameter 1 inch, .and length varying by steps of one-half inch each. Such a difference causes quite a decided illusion for everybody. He concludes that those who arranged them in regular order according to the illusion were the most suggestible. That is not necessarily so. All that he can conclude is that those who arrange them in regular order have the finest power of discrimination ; for since the illusion exists for all persons, an error in the arrangement is the same as an error of observation when there is an absolute difference in weight. One who has a very fine power of discrimination may arrange the weights in perfect order according to the illusion even if the illusion is not half as strong with him as with another person who is a poor observer and cannot judge weights well enough to arrange these in order. Deesslar\u2019b apparatus does not afford him any definite measure for the intensity of the illusion.\n1 Dresslar, Studies in the psychology of touch, Am. Jour. Psych., 1894 IV 41.","page":24},{"file":"p0025.txt","language":"en","ocr_en":"Measurements of illusions and hallucinations in normal life. 25\nTo justify his conclusion he should have measured the extent of the illusion for the different classes of observers. This Gilbert1 did in his tests on school children, and he found that the intensity of the illusion decreases quite regularly in children from the age of 9 to IV. From 6 to 9 the suggestion increased, but from 9 to 17 inclusive it decreased, the amount of the illusion for his apparatus being as follows for the respective ages beginning with 9 : 50.0s, 43.5s, 40.0s, 40.5s, 38.0s, 34.5s, 35.0s, 34.5s, and 27.0s. If we admit that children at 17 have a higher degree of intelligence than children at 9, this proves just the reverse of Dresslar\u2019s conclusion. To experiment further upon this very important problem, a very satisfactory arrangement would be had by taking Dresslar\u2019s method of classifying pupils into three classes according to general intelligence and testing them by the apparatus used in this research.\nPsychological analysis.\nPerception of weight is an interpretation by discriminative consciousness of a very complex series of peripheral sensory elements with a conative feeling of effort in the light of an intricate series of associated ideas. An illusion of weight is caused when any of these factors stand out in an unnatural relation. Thus if the skin be inflamed, the haptic sensations will be abnormal ; if the muscles be fatigued or the joints disordered, the muscle sense and the joint sense will give a wrong impression ; if the central associations through memory, imagination, and comparison be mistaken, the interpretation will be false ; and if the exerted effort be disproportionate to the object, the expended energy will be incorrectly estimated.\nIn order to see the psychological process in the present illusion, let us take an example, magnify that instantaneous perceptive process, and analyze it into its constituent elements in a schematic way.\nThe weight of an object is to be determined by lifting it in view. The first step consists in the presentation of a sensation complex, which is the sensory report of the eye and hand upon seeing and grasping the object to determine its spatial relations. Though knowledge of size may not be necessary for the percep-\n1 Gilbert, Researches on the mental and physical development of school children, Stud. Yale Psych. Lab., 1894 II 60.","page":25},{"file":"p0026.txt","language":"en","ocr_en":"26\nC. JS. Seashore,\ntion of weight, it always enters as a preliminary determination. I cannot conceive what kind of a sense of weight a person would have without some estimate of the volume through which it is distributed. The presentations of the spatial sensations first fix in discriminative consciousness an image of the size of the object, and this is compared with the memory images of experiences with similar objects. In a comparison of the surface appearances in color, shape, nature of material, smoothness, hardness, etc., it is the function of attention to bring all these considerations into service to determine the probable weight on the basis of previous experiences.\nAfter an elaborate and infinitely complex process of discrimination and comparison, the result is that a probable weight, w, is assigned to the object. This whole mental process was performed before the object was raised from its base. The motor apparatus had waited for a command from discriminative consciousness, directing what amount of effort should be put forth in order that the motor adjustment should correspond to the mass of the object. In consideration of the estimated probable weight, to, the gauge of the motor impulse is set to a corresponding amount of motor effort, to. The judgment of weight is an interpretation and estimation of the amount of this effort, w, expended.\nIf the effort put forth just meets the demand, the weight will be judged accurately, other things being equal. But if some greater effort, w + d, should be required in succeeding to lift the object, the extra effort, d, is brought into exceptional relation to to, and this causes d to become prominent and overestimated. Since d is an inci'ement of w, w + d will be overestimated, i. e. the weight of the object will be overestimated. And, if w-d should be required, d will again be overestimated for the same reason. Since d must be subtracted from w, to-d will be underestimated, i. e. the weight of the object will be underestimated.\nHere we have taken for granted that w was determined by our general past associations between weight and size. But, suppose in a particular case, largest A and smallest A are to be lifted in succession. Since they have the same appearance except in size, there is a special and very definite association formed between size and weight, which raises the value of d by contrast and makes it more definite. The greater the value of d, the greater will be the consequent illusion. Hence, this illusion of weight","page":26},{"file":"p0027.txt","language":"en","ocr_en":"Measurements of illusions and hallucinations in normal life. 27\ndepends upon our more or less fixed associations between size and weight.\nSight first scans the object, estimating its size ; then come the skin, joint, and muscle sensations and emphasize its reality and solidity. When this is done, active motor consciousness proceeds to estimate its weight. Or, to emphasize the different steps again, the size of the object is first estimated ; then, on the basis of that and the appearance of the material, a mental estimate of its probable weight is formed ; and, with the muscles accommodated to this, the actual weight is judged by estimating the amount of expended effort.\nThe above analysis corresponds to the actual process of perception in the case of the illusion under consideration. Let us trace its process more particularly in different stages.\nIn a preliminary way two objects, largest A and smallest A in the above apparatus, are to be compared in weight. The observer looks at both and then lifts largest A, noticing its size particularly. It is found lighter than was expected. Consequently a certain amount of effort, d, must be subtracted from w. Therefore there is an illusion : largest A is underestimated. Then he lifts smallest A in the same manner. Since the two objects appear to be made of the same uniform material, and the weight of largest A is known, the probable weight of smallest A is estimated to be in proportion to the size of the objects ; hence the intensity of w is based upon a comparison of the diameters of the two objects. Consequently d is raised to a very high value by the disappointed expectation due to contrast, and smallest A is very much overestimated. In both the overestimation and the underestimation there is evidently a surprise or disappointed expectation.\nThis explanation is applicable in the first trial, where there is a surprise. But how explain the persistence of the illusion in repeated trials with the same block as well as with a series of blocks ?\nSuppose the observer has matched the blocks 19 times and come to a fairly positive conclusion that the smallest one is twice as heavy as the largest one. Then he knows just what to expect the 20tli trial, and there ought not to be any surprise. But the illusion is there just as before. Could we neglect or break up a series of fixed associations that have been forming in all our past life, then there ought theoretically not to be any extensive persistence of the illusion. But we cannot. The judging activity is not in such perfect control that all the information gained by the first trial can","page":27},{"file":"p0028.txt","language":"en","ocr_en":"28\nG. K Seashore,\nbe utilized and all previous associations set aside. Instead of taking a new course in the second trial and estimating the amount of effort only by the experience with these particular blocks, interpretative consciousness ignores that and follows the formerly set paths of habit. It receives the new i'eports from the spatial senses for each trial, compares, discriminates, and commands as before, often irrespectively of the immediately preceding experience. This is liable to continue as long as the observer is not informed of the illusion.\nBut how explain the persistence of the illusion when the observer knows all about it ? Suppose he has been told that two objects have the same weight, and he has empirically discovered that the weight seems to vary inversely with the size. He lifts them and the original illusion persists. The same explanation must here be extended. He does not possess will power enough to disregard the accustomed manner of association and put into practice what he ought to know. Size has ever before been influential in determining weight, therefore, relatively, it can not be suppressed. This is not a sign of weakness in discrimination or judgment ; it is the working principle for those whom we consider most intelligent. That feeling of interest which sight commands is persistent and insists on distributing the felt weight throughout the perceived volume ; and in the ordinary flow of conscious activity, it is almost impossible to muster force enough to dam it up.\nThe same principle is confirmed by the variation of the illusion with the directness of sight. It is illustrated by the four steps in the third series of experiments. If we may represent attention as a force, attention to spatial sensations and attention to direct sensations of weight are two forces which have generally worked in the same direction; but, under these illusory circumstances, the former has turned against and opposes the latter. As one decreases the other increases. In direct sight attention to size has a strong opposing force to attention to weight ; in direct sight the former decreases and the latter increases ; in visual memory the corresponding increase and decrease is still greater ; and, when there is no knowledge of comparative size, attention to size loses all force and there is no illusion.\nThe same principle is again exemplified in the fourth series of experiments, though somewhat differently. Previous to this the knowledge of comparative size had been acquired by sight. Here sight was excluded in the first two steps. Size was estimated by the other spatial senses. It is more of an effort to determine size by","page":28},{"file":"p0029.txt","language":"en","ocr_en":"Measurements of illusions and hallucinations in normal life. 29\ntouch or muscle sense than by sight. Hence it requires more attention, which would otherwise be accorded to the sense of weight.\nAgain it is evident that the same principle suffices to explain the illusion due to the suggestive force of the material.\nIn a word, then, the illusory influence of size upon judgment of weight consists in disappointed expectant attention. We continue to be disappointed because we will not take the temporary circumstances into sufficient account.\nPART SECOND.\nExperimental applications of the principle of suggestion in\nNORMAL PRESENTATIONS OF SENSE.\nThere is no dearth of observations on \u201c errors of sense,\u201d nor need we review much of our normal experience in order to discover the presence of the force of suggestion. I have not undertaken to search particularly for new classes of illusions and hallucinations. The purpose in the following experiments has rather been to take up some of the simplest forms and principles of illusions and hallucinations and submit them to examination by applying the laboratory methods of psychological experiment to normal waking life.\nThe chaotic and superstitious view of mind represents it as lawless and unknowable, while on the other hand we have the view that by scientific method and philosophical introspection the laws of mind-mind, normal and abnormal\u2014may. be studied and classified just as well as the laws of matter. Again, on the one hand we have the phenomena of hypnotism, mind reading, crystal vision, apparitions, telepathy, etc.\u2014considered as occult practices, while on the other hand there is an attempt to bring them in conformity with principles at work in normal every-day life. And yet again we have the claim that the world of things is but an illusory creation of imagination, while on the other hand there are those who regard the states and actions of mind as fleeting phantasms and place implicit faith in the presentations of sense.\nAny attempt to discover the nature of sensory illusions and hallucinations deals w'ith the core about which these contentions center. It is surprising that so little real scientific work should have been done in a field so important. The whole subject of \u201c Trugwahrnehmungen \u201d readily presents itself for experiment and it may","page":29},{"file":"p0030.txt","language":"en","ocr_en":"30\nG. JS. Seashore,\nwell profit by the recent developments in laboratory methods and facilities for experimental psychology. Working in this direction, I have the disadvantage of being one of the first to enter this domain, and consequently the method is tentative and cursory.\nHallucinations of warmth.\nApparatus.\nThe apparatus1 for this experiment was constructed on the principle that a wire will be heated by passing an electric current through it. Two binding posts were placed 8cm apart and connected by a German silver wire. They were then connected in circuit with a bichromate dip cell and a secret open-circuit key. The battery and the stand supporting the German silver wire were placed together on a table in front of the observer, while the open-circuit key was fastened under the table, where it could be readily opened and closed with the knee without being detected.\nMethod of experimenting and results.\nThe experimenter and the person experimented upon were seated, one on each side of the apparatus, facing each other. The preparatory explanation had to vary with the observer\u2019s knowledge of the principles involved. His .attention was called to the fact that an electric current will heat a wire and that it will take an appreciable time for a person touching the wire to perceive this heat. The instructions were : \u201c At a signal which I will give, after the dip is lowered, grasp that wire between your thumb and index finger ; concentrate your attention on what you are doing, and when you feel the wire become heated, say hot.\u201d\nAll being ready, I lowered the dip as mechanically as possible and a second later gave the signal to grasp the wire. ' The time between the signal and the reaction was recorded in seconds. In the first two to five trials I closed the circuit with the knee key, the wire was heated, and we found roughly the time that it took to perceive the heat. The observer thus received a distinct perception of how the wire became heated, associated this with the lowering of the dip, and formed a distinct estimation of the required time.\nThen I suggested that we repeat the experiment a sufficient number of times to secure a fair average. We proceeded exactly as\n1Soeipture, Tests on school-children, Educ. Rev., 1893 V 61.","page":30},{"file":"p0031.txt","language":"en","ocr_en":"Measurements of illusions and hallucinations in normal life. 31\nbefore, except that I left the current broken ; thus, although I lowered the dip, no current went through the wire.\nFive different signals were used, namely : (1) a wire coil whirled so as to rattle and expose a red card, having an effect on eye and ear (the starting of this whirl was the signal to grasp the wire); (2) tapping the right hand with a pencil (same hand as held the heated wire) ; (3) tapping the left hand with a pencil ; (4) tapping on the table ; (5) saying \u201cNow;\u201d (6) tapping the wrist.\nThe experiments were made in sets of ten trials, using a different signal for each set. The signals were taken in the order given above. When no heat was perceived in one minute, that experiment was recorded as a failure. The results are contained in Table VI.\nTable VI.\nHallucination of warmth.\n0 i\tSignal l.\\ T MV F\t\t\tSignal 2. T MV F\t\t\tSignal 8. T MV F\t\tSignal U.\\ T MV F\t\t\tSignal 5. T MV F\t\t\tSignal 6. |1\tSummary T MV F\tM\tN\tE\t\t\t\tA\nI\t10\t2\t0\t13\t6\t0\t17 10\t0\t9\t3\t2\t\t\t\t\t40\t38\t2\t12\n11 !\t14\t5\t0\t16\t4\t0\t21 3\t0\t19\t3\t0\t;9\t4\t0\t..\t....\t50\t50\t0\t18\nIll\t27\t7\t2\t26\t4\t0\t19 6\t0\t23\t6\t0\t21\t4\t0\t\t\t, 50\tCO\t2\t23\nIV\t3\t1\t1\t4\t2\t0\t2 5\t0\t10\t1\t1\t7\tQ\t1\t14\t6 0 60\t60\t0\t8\nV\t3\t1\t0\t3\t1\t0\t3 0\t0\t3\t0\t0\t2\t0\t0\t4\t1 0 60\t60\t0\t3\nVI\t13\t6\t0\t17\t5\t0\t12 6\t0\t8\t3\t1\t5\t2\t0\t\t50\t49\t1\t11\nVII\t6\t2\t0\t8\t2\t0\t8 2\t0\t10\t2\to!\t17\t3\t0\t\t....\t50\t50\t0\t10\nVIII\t3\t1\t0\t3\t0\t0\t3 1\t0\tI3\t1\t0\t3\t1\t0\t7\t2 0 60\t60\t0\t4\n0, observer.\nT, average number of seconds required for the hallucination to arise.\nMV, mean variation.\nF, number of experiments in which no hallucination was produced within one minute.\nM,\ttotal number of experiments.\nN,\ttotal number of hallucinations out of M.\nE, total number of no hallucinations within one minute.\nAI average number of seconds for all T.\nThe object of using different signals was to find if the hallucination would be influenced by the character of the signal, but the conditions could not be satisfactorily controlled for that purpose, and the difference in time depending on this is not great enough to warrant any generalization from the figures on that point. It is, however, quite probable that this division and use of different signals tended to increase the force of the original suggestion by intensifying the expectant attention.","page":31},{"file":"p0032.txt","language":"en","ocr_en":"32\nG. E. Seashore,\nThe leading mental process in these hallucinations' was an inference. The fact that the experimenter performed apparently the same manipulations that in the preparatory trials had produced a distinct sensation, formed the definite suggestion that, since the conditions were in toto repeated, the resultant sensation would recur in the same time and manner as before. By force of a firm expectant attention, caused by this inference, the image of the sensation realized itself in a sensation projected into the peripheral organs. And in the positive instances, the observer felt it just as he expected to feel it, although there was no physical stirdulus.\nRemarks like these were frequent : \u201c faint,\u201d \u201c distinct,\u201d \u201c quite warm,\u201d \u201cvery distinct,\u201d \u201c suddenly hot,\u201d showing that the hallucination existed in different degrees of intensity. It was difficult to convince those, to whom the experiment was afterwards explained, that the actual stimulus had really been withheld. The eighth observer had heard a lecturer refer to the test. When questioned after the test in regard to his experience, he said he felt the heat distinctly every time, although at first he thought there was some trick about it. \"From these experiments I conclude that a mental image of a definite liminal sensation of heat can be realized in a peripheral sensation in the absence of any physical stimulus if there is no incongruity in the phenomena which serve as suggestions.\nIllusions of'photometric changes in gray.\nApparatus.\nThis apparatus was constructed on the principle that the intensity of the shade of a white surface varies with the angle at which the rays of light strike it. The aim in its construction was to present two symmetrical white surfaces so adjusted that they could be turned independently at a graduated rate and distance, changing the angle of incidence of the rays of light that fell on both surfaces from one artificial light.\nA front opening in a wooden frame, fig. 6, was covered by black card-board, which had two symmetrical openings, 2cm apart. The\n1 X have observed the distinction between illusion and hallucination first made by Arnold (1806) and now accepted in a general way, namely, that sensory illusion is a false interpretation of an object, while sensory hallucination is a subjective sense image objectively realized by external projection ; that is, in illusion there is a physical cause, while hallucination is purely subjective. It is, however, probable that there is a \u201cpoint de rep\u00e8re\u201d in many of the phenomena I have classified as hallucinations.","page":32},{"file":"p0033.txt","language":"en","ocr_en":"Measurements of illusions and hallucinations in normal life. 33\nopenings were 6cm wide and 8cm high. Back of them were two leaves swinging on hinges at the middle of the partition between the openings. The fronts of these leaves were covered with fine white Bristol board. To the outer ends of these leaves, cords were attached, which wound around an axle in the rear of the apparatus in such a manner that the motion of the leaves could be controlled\nFig. 6.\nby a crank on the axle. One revolution of the axles constituted th\u00e9 standard for measurement of change in the angles, and was marked off in corresponding degrees on the arcs described by the outer ends of the swinging leaves. Thus the term degree is an arbitrary unit and not of a circle. For convenience, I will speak arbitrarily of the amount of change in the intensity of shade produced by the swinging of a leaf one degree in one second as a change of one 3","page":33},{"file":"p0034.txt","language":"en","ocr_en":"34\nG. E. Seashore,\ndegree-second. The figures in all the tables of this section record the results in terms of such degree-seconds.\nThe apparatus was placed in the dark-room. A standard candle, one meter in front of it, lighted up both surfaces equally. The observer\u2019s chair was placed at a distance of 4m directly in front of the apparatus. The range of vision was limited by two slits in cardboards, the one placed near the observer\u2019s eye, and the other lm away. By this means, distraction of the eyes was avoided and the tops and bottoms of the surfaces to be observed, where shadows were necessarily cast, were cut out from view. The surfaces were spoken of as right, R, and left, L, from the observer\u2019s point of' view.\nAt the beginning of an experiment the white surfaces lay in the same plane and the incident rays of light struck them practically at right angles ; i. e. the inner edges of the surface were 20m apart and the candle was placed directly in front of that partition. The angle of incidence decreased as a leaf was turned back on its hinge. The surfaces at no time appeared to be white but rather a light gray. The surfaces increased in intensity of gray inversely with the angle of incidence of the light rays. There was only one light in the room, and by the above arrangement all shadows were excluded from the surfaces in view. The crank was turned one revolution, i. e. 1 degree per second.\nThe apparatus was not constructed for very accurate photometric measurements, but for comparative measurements it serves very well.\nMethod of experimenting and results.\nIn the first series of experiments the rate of change was determined by the beating of a metronome. After a warning, R was turned 1 degree per second until the observer found it darker than L. Then L was turned until he thought L was equal to R. When this had been repeated a few times, I interspersed chances for illusions. That was done by simply allowing time after the warning without any movement of the leaves. The warning consisted in saying, \u201c Ready,\u201d and the record was kept in seconds marked by the metronome. Nothing was said to the observer about the time of the discrimination. He was simply asked to state when he saw one surface lighter or darker or both alike. He could not see any manipulations of the apparatus. His attention was concentrated on watching the change of the shades. The fact of darkness and quiet","page":34},{"file":"p0035.txt","language":"en","ocr_en":"Measurements of illusions and hallucinations in normal life. 35\nin the room heightened his expectant attention and threw him into a suggestible state of mind.\nSeveral persons were experimented upon by this method with the result that the required change appeared to them universally from 6 to 13 sec. after the signal had been given, whether there was any real physical change or not.\nThough the metronome which was used in the first experiments increased the force of the suggestion of change, it afforded an element of distraction, which was avoided in the later experiments by not using it. The reason for this was that the observer could not escape paying attention to the time, and this distraction might be so strong that after the first trial he would be in danger of interpreting a certain number of seconds as so much change of shade.\nBut even when the metronome was not used, the suggestion was so effective that the illusion could be produced alternately with the normal sensation.\nIn the following records S will denote the presence of a physical stimulus and I its absence. The column headed S in the table shows the number of degree-seconds of the change, i. e. the number of equal steps described on the graduated arc in an equal number of seconds. The column headed I indicates the illusions by giving the average time that it took them to arise. This time may be compared with the time given in column S.\nTable VII.\nIllusions of photometric changes in gray : Increasing and decreasing deepness of shade.\n0\tS(o)\tMV\t\u00a3(25)\tMV\t1(0)\tMV\t-1(25)\tMV\nj\tI\t11.7\t1.0\t12.6\t4.7\t18.1\t4.6\t23.0\t3.5\nII\t13.8\t5.0\t17.6\t4.9\t19.7\t6.1\t19.4\t7.9\nIII\t13.1\t46\t12.0\t2.0\t18.3\t6.2\t17.0\t2.9\n0, observer.\nS(o), number of degree-seconds the surface was darkened.\nMV, mean variation.\nS (25), number of degree-seconds it was made lighter.\nI(o) and 1(25), corresponding illusions.\nCan the illusion be produced in decreasing as well as increasing intensity of shade ? The question was answered in the affirmative by the following test of which the final averages of the results are","page":35},{"file":"p0036.txt","language":"en","ocr_en":"36\nG. JE. Seashore,\ngiven for three observers in table VIL L was left constant at 10 and R was moved toward 10 alternately from the 0 and 25 points, the number of degree-seconds indicated in the table and the observer was to tell when he judged the two surfaces to be alike. The I trials were alternated with the S trials.\nIn order to keep up the suggestion with some observers, I found it expedient to start moving the surface at the end of 20 seconds, if the suggestion had not then been effective. Hence the records higher than 20 contain only partial illusions, and, in rare cases, failures. Thus, in table VII, 23 shows that there was 3 degree-seconds of physical stimulus present.\nTable VIII.\nIllusions of photometric changes in gray Dependence upon the number of repetitions.\n\t\t\t\t\t\t\t\t\n0\ts\t\tMV\t/\t1* S'\tMV\tI'\tMV\n\tN=9\t\t\trH II 5zi\tN=5\t\tN=5\t\nI\t11\t\t3\t16\t15\t5\t22\t9\n\t13\t\t3\t30\t12\t3\t23\t1\n\t4\t\t1\t1\t6\t1\t8\t2\nII\t6\t\t1\t9\t9\t2\t12\t3\nIII\t9\t\t3\t7\t8\t2\t.13\t5\n\t11\t\t1\t10\t13\t2\t15\t2\nIV\t12\t\t1\t19\t16\t2\t25\t2\n\t15\t\t1\t24\t12\t1\t27\t2\nAye.\t10.1\t1.7\t\t15.2\t11.4\t2.2\t18.1\t4.0\nN, number ol trials on each point. | Other notation same as before.\nIn accordance with the principle on which the apparatus was constructed, the deepness of the shade changed proportionally faster in the dark field. Thus, starting the leaves successively from 0, 10 and 20 and moving them at the same rate the change in shade would appear sooner after 20 than after 10 and sooner after 10 than after 0. Experiments on this point show that the time for the illusion to arise is proportional to the rate of actual change of shade, although the observer does not knqw that there is any difference in the rate of change.","page":36},{"file":"p0037.txt","language":"en","ocr_en":"Measurements of illusions and hallucinations in normal life. 37\nThe following method was used in determining the dependence of the illusion upon the frequency of the repetition of the suggestion in form of the real stimulus.\nFour sets of ten trials each constituted an experiment. In the alternate sets the stimulus was given nine times successively and followed by an illusion, while in the other sets the stimulus was presented for the alternate trials and withdrawn from the rest.\nL was kept constant at 0 and R was started at 0 and moved the number of degree-seconds indicated in table VIII, which gives the averages.\nThis shows that :\n1.\tThe required time and amount of change in intensity for discrimination is greater when the stimulus is given in alternate trials than when it is given 9 times in succession. Ratio, 11.4 degree-seconds to 10.1 degree-seconds.\n2.\tThe time for the illusion to arise is greater when the illusion is produced in alternate trials than when produced only once after 9 repetitions of the stimulus. Ratio, 18.1 seconds to 15.2 seconds.\nI would not ascribe much value to these averages were it not that they express the relation which I find obtains, on the whole, in experiments on each of the other senses.\nTime, it is evident, is the leading element in this kind of suggestion. By the following method I have secured a relative measure of its influence. The illusory element in the method of the next two experiments was the fact that, after giving a regular warning, I did not begin the movement of the leaf until after the number of seconds indicated at the head of each column.\nThe first observer had 100 trials and the second 40, but only the averages of the respective divisions are recorded in the table. The trials were so distributed with alternate S and I and a proper succession in regard to time as to eliminate systematic and progressive errors.\nSince possibility of direct suggestion was precluded by the present method, the efficient cause in producing these illusions was simply the fact that the observer falsely inferred that the same change which took place after the first signals would recur after each succeeding signal. The warning or signal was given in such a mechanical way as to have no direct influence upon the discrimination.\nThe main results of these experiments on illusions of photometric changes in intensity of shade may be summarized as follows :","page":37},{"file":"p0038.txt","language":"en","ocr_en":"38\nC. JE. Seashore,\n1.\tVisual illusions of deepness in shade can be experimentally produced by securing a firm expectant attention to a known sense percept ;\n2.\tThough time is one of the main elements in a suggestion, the illusion will occur in the absence of any external suggestion or special emphasis of passing time (i. e. the observer does not know that time is recorded) ;\n3.\tA suggestion of change from dark to light shade is just as effective as the converse ;\n4.\tThe time for the illusion to arise varies directly with the time for ordinary discrimination ;\n5.\tThe discrimination-time and the illusion-time are both longer when the illusion is given in the alternate trials than when it is only given once after a series of ordinary discriminations.\nTable IX.\nIllusions of photometric changes in gray : Dependence upon the duration of the stimulus.\n0 I II\tS MV\t\t5 Sec. A MV D I\t\t\t\t10 Sec. A MV D 1\t\t\t\t15 Sec. A MV D I\t\t\t\t20 Sec. A MV D 1\t\t\t\n\t16.3 10.5\t2.7 1.6\t17.2 12.0\t2.4 1.5\t0.9 1.5\t4.1 3.5\t19.0 15.0\t2.6 3.5\t2.7 4.5\t7.3 5.5\t22.3 17.3\t2.5 5.2\t6.0 6.7\t9.0 8.3\t24.9 15.5\t1.6 5.2\t8.6 5.0\t11.4 15.0\nS, regular stimulus present.\nMV, mean variation.\nA, result when the stimulus is delayed with the number of seconds indicated.\nD, difference between and A.\nI, amount of the illusion expressed relatively in degree-seconds of change.\n0, observer.\n6.\tThe amount of the illusion varies directly with the amount of variation in the time of producing the physical stimulus.\nIllusions of light.\nApparatus.\nThe main part of this apparatus was constructed for the purpose of measuring the influence of the rate of change in determining the least perceptible difference in intensity of light, but has not before been described. It consisted of (1) a kymograph with attachment for rotating a wheel at a slow, measured rate ; (2) a light with lenses, diaphragms, shutter, and reflectors ; (3) a stop-watch, namely, the Runne chronometer with electric start and stop ; (4) sounder and telegraph key in the dark-room.","page":38},{"file":"p0039.txt","language":"en","ocr_en":"Measurements of illusions and hallucinations in normal life. 39\nAn attachment was put on the Zimmerman kymograph for rotating a large card-board disk at adjustable, slow rates. On a radius of this disk, and in metallic contact with the frame of the kymograph, an arm projected with a platinum point, which made contact with another point insulated from the frame, at any position in the revolution. The edge of the disk passed between the two glasses of a double photographic lens. In this passage a diaphragm was inserted. It was 20mm long on the inner edge, and the sides extended 25mm on the radii of the disk. The rotating disk had an opening near the edge, measuring 80mm on the inner side, with dimensions of the radii symmetrical with those of the opening in the diaphragm. The sides of the two openings coincided. The platinum point could be so adjusted that it made contact when the upper edge of the disk-opening was at any desired position in the diaphragm. There was another platinum point which could be so adjusted as to close contact when the lower edge passed any desired point in the opening of the diaphragm.\nThe light was cut out from the lens by a shutter which was supported by magnets, charged from a dip cell. Closing the platinum contact shunted this current through the apparatus, released the magnets, and dropped the shutter. The falling of the shutter opened the course for direct rays of light through the lens and at the same time closed a mercury contact, and thereby started the stop-watch.\nAn Argand burner was placed 50cm in front of the lens ; 5cm from this, between the burner and the lens, there was a white tissue paper, which diffused the rays of light. Two reflecting plates, having facing sides coated with magnesium oxide, were placed on the other side of the lens, parallel to each other, at an angle of 45 degrees to the rays of light. The light thus passed through the tissue paper, the lens, and the diaphragm, and was reflected by the coated surfaces. It was then seen in the dark-room through a circular opening 10mm in diameter in a black diaphragm. There it appeared as a disk in the color of the gas-flame. Though it was very clear it did not have any excessive brightness.\nTelegraph sounders were used for signals. The telegraph key in the dark-room was so connected with magnets on the stop-watch that the reaction stopped the watch. Hence the time between the dropping of the shutter and the movement of reaction could be read off on the stop-watch in fifths of a second.","page":39},{"file":"p0040.txt","language":"en","ocr_en":"40\nG. E. Seashore,\nThe kymograph revolved the disk so that the opening in the diaphragm, between the lenses, was diminished or increased according as the lower or upper edge of the disk opening passed through it. The photometric principle involved was that the intensity of the stimulus changes with the variation in the size of the opening.\nMethods and results.\nWhile assisting another experimenter in tests on the threshold of change as dependent on the rate, it became apparent to me that suggestion might play an important part in the experiment. To determine this I repeated the experiment with the apparatus just described and by the same methods as were being used in the investigation referred to. I tried to keep all conditions similar, but put myself on guard to discover if any illusion was involved in the process.\nThe original problem was this: with the apparatus just as described, to find the dependence of the time and accuracy of discrimination upon the rate of change in the physical stimulus. The observer was to react as soon as he perceived any change in intensity of brightness of the disk. He knew that it was to grow darker at a definite rate.1\nThe running kymograph closed the lower platinum contact and thereby dropped the shutter the instant the lower edge of the opening coincided with the lower edge in the opening in the diaphragm. This act exposed the disk to view at its greatest intensity of brightness. But the continuous motion of the kymograph kept the cardboard disk revolving at a uniform rate. This regularly diminished the opening in the diaphragm, which caused a corresponding lessening of the intensity in the red disk, making it grow darker.\nThe observer was seated in the dark room, and received warning by a sounder 5 seconds before the shutter dropped. His instructions were : \u201cReact as soon as you see that the disk has become darker than it was at its first appearance.\u201d\nIt was not stated that the change would always be the same, but the observer generally inferred this when he saw the apparatus.\nIn sets of 20 observations each, I interspersed chances for illusion, at irregular intervals, by simply .stopping the kymograph just as the\n1 Scripture, On the method of regular variation, Am. Jour. Psych., 1891 IV 571. Scripture, Ueher die Aenderungsempfindlichkeit, Zt. Psych. Phys. Sinn, 1894 VI 472.","page":40},{"file":"p0041.txt","language":"en","ocr_en":"Measurements of illusions and hallucinations in normal life. 41\ndisk was exposed. It was thus kept constant at full intensity of brightness. These furtive interspersions were made from 2 to 6 times in the 20 trials, and the observers almost invariably reacted to these corresponding expected changes, perfectly confident that they had perceived real changes taking place.\nAfter the fact of this illusion was known, two persons continued the experiments on the threshold, for which this apparatus was originally intended. To test the accuracy of the discrimination the experimenter interspersed chances for illusions as in the above method. These illusions were never detected by the observer. For the purpose of the original problem, the experimenter had to agree not to indulge in any such tests because, after having been told of it, the observer became suspicious that it would be repeated, and was made over-cautious. This over-cautiousness led him to wait and verify his first judgment, and that made the discrimination-time worthless.\nTable X.\nIllusions of light.\n\t0\tLS\tLI\tDS\tDI\n\tI\t3\t3\t5\t3\n\tII\t8\t8\t5\t9\n\tIII\t5\t6\t5\t4\n\tIV\t4\t6\t3\t5\n0, observers.\t\t\tI\tDS, disk grew darker (actual physical\t\nLS, disk grew lighter (actual physical change), change).\tD1, disk grew darker (no physical\nLI, disk grew lighter (no physical change), change),\nThe record is kept in seconds. Each figure is the average of five observations.\nTo find whether the illusion could be obviated in an experiment of this kind, I instructed the observer to react as soon as he could tell definitely whether the disk was growing lighter or darker.\nThe two platinum contacts were so adjusted that, by closing the circuit with either one, the disk would be exposed at half its possible intensity. When the lower contact was used, the disk would continue to grow darker, and when the upper contact was used, it would continue to grow lighter. The results are shown in Table X. In this and the following tables of this section the unit of measurement is the amount of change in a second at the given rate. We may again call this a degree-second of change.","page":41},{"file":"p0042.txt","language":"en","ocr_en":"42\nC. JE. Seashore,\nAs these figures indicate, the illusion came out surprisingly strong even when the precaution was taken, to require the observer to discover the direction of the change. Though there was no suggestion as to whether the disk should grow darker or lighter, the firm expectation that a change would occur one way or the other proved a sufficient cause to call forth the illusion.\nAlthough these were bright men and very cautious, they exemplify the very extreme degree of suggestibility to this kind of illusion.\nFor the present purpose it was unnecessary to make exact photometric measurements, because these relative determinations suffice to bring out the laws of the suggestions at work.\nThe apparatus was so constructed as to admit of a great variety of graded rates of change. In units of the arbitrary gradation, 0.22 per second was found the most favorable rate for effective suggestion. At this rate, which is employed hereafter, the entire range of change that the apparatus permitted was traversed in 16\nTable XI.\n\tIllusions of light: I t\t_a\t\tRate of change. II \tK\t \t\t\tIII -A\nRate per\tsec. LS\tLI\tos m\tLS\tDG\n0.22\t7\t12\t7\t9\t8\t5\n0.33\t8\t13\t6 10\t7\t4\nNotation same as in the preceding table. Each number is the average of 5 observations.\nseconds. If, e. g., twice that rate was used, there was danger of either hasty, automatic reaction or inhibition of the illusion. And, if the rate was made extremely slow, the result would generally be a confusion because the transition was so gradual that it was difficult to retain a distinct memory image of the original impression.\nTable XI records the averages of an experiment in which the illusions with two different rates of change are compared. In the first set the rate of change is 0.22 per second, and in the second set 0.33, i. e. the latter is proportionally slower. Before beginning the experiment the observer was given a few preliminary trials\u2014not more than ten, and thereupon he was instructed : In the first set, react when the disk grows lighter ; in the second, when it grows darker ; and, in the third, when you can see whether it has grown lighter or darker. The exceptionally small difference in time for the two rates is partially explained by the fact that the latter","page":42},{"file":"p0043.txt","language":"en","ocr_en":"Measurements of illusions and hallucinations in normal life. 43\nset of observations was made soon after the former, which undoubtedly served as a suggestion, shortening the time in the other. This fact emphasizes the importance of arranging laboratory experiments in a proper order of succession.\nTo discover whether the illusion could be worked up in persons who were particularly on guard by expecting an attempt of illusion, I made the following test: I selected two observers who knew that I was experimenting on illusions. I had previously tried them with a slightly different method and failed, but now nothing was said in regard to what would take place. The instructions were, in I and V (Table XII), to react when the disk had become perceptibly darker; in II and IV, when it had become lighter ; and in III, when the observer could tell whether it had grown lighter or darker.\nTable XII.\n\t\tIllusions of light:\t\tDevelopment\tof the suggestion.\t\t\t\t\n\t/\tn\t\tin\t\tIV\t\tV\t\n0\tDS\tLS\t'ls\tDS\tLS\tLI\tDS\t\tDI\nI\t12.2\t5.4\t2.3\t2.6\t4.8\t6.2\t6.6\t\tn.o\nII\t13.2\t5.5\t9.0\t2.2\t6.5\t8.6\t6.0\t\t7.4\n\tThe notation is the same\t\tas in preceding tables.\t\tThe figures are\t\taverages\tof\tten\ntrials in each group.\nThis shows how the suggestion accumulates force by successive repetitions of the real stimulus. As was shown in other cases, the illusion would not have been likely to occur had the chances for illusion been introduced near the beginning. This is also shown by the long discrimination-time in set I.\nHere we can trace the evolution of an illusion. In I the observers were on their guard, very suspicious, and did not react upon the first perception of change. In II they grew more exact, and by this time they began to convince themselves that no illusion was involved. Ill is an index to their reliability. In IV and V the alternate trials are illusions and here these two observers showed themselves just as susceptible to the suggestion as those who were not thus prejudiced.\nPursuing the method employed in the experiments recorded in Table IX, similar measurements were more extensively carried out with the present apparatus, giving the results of which the averages are quoted for eight observers in Table XIII.\nThe observers were required to react when they could see that the disk had grown brighter. The figures at the heads of the","page":43},{"file":"p0044.txt","language":"en","ocr_en":"44\nC. E. Seashore,\ncolumns show how many seconds the application of the stimulus was delayed after the warning. The omission of the stimulus was made by stopping the kymograph for so many seconds, allowing one second for it on which to start.\nProgressive errors were avoided by the distribution of the trials according to this method, one trial in each group being made in succession back and forth, with A and I alternating. Since only two trials were made on each point, the variation in the mean results is quite large.\nTable XIII.\nIllusions of light: Dependence upon the time of application of the stimulus.\n\t0\tS MV\t\tA\t5 MV\tSec. D\tI\tA\t10 MV\tSec. D\tI\tA\t15 MV\tSec. D\ti\t\t20 MV\tSec. D\ti\n\tI\t| 14.3\t2.2\t13.5\t1.5\t\u20140.8\t6.8\t21.0\t] l.oj\t6.7\t3.3\t28.5\t6.5\t14.5\t0.8\t29.0\t6.0\t14.7\t5.3\n\tII\t6.2\t1.9\t6.0\t1.0\t\u20140.2\t5.2!\t14.5\t2.5\t8.3\t1.7\t23.5\t3.5\t17.3\t-2.3\t10.5\t2.5\t4.3\t15.7\n\tIII\t6.6\t1.4\t1.5\t0.5\t0.9\t4.1\t8.0\t1.0\t1.4\t8.6\t10.5\t45\t3.9\t11.1\t7.0\t0.0\t0.4\t19.6\n\tIV\t5.0\t2.0\t6.5\t1.5\t1.5\t3.5\t4.0\t0.0,\t\u20141.0\t11.0\t6.5\t0.5\t1.5\t13.5\t6.5\t0.5\t1.5\t18.6\n\tV\t9.6\t2.4\t9.5\t1.5\t\u20140.1\t5.lj\t15.5\t0.5'\t59\t4.1\t19.0\t11.0\t9.4\t5.6\t25.5\t0.5\t15.9\t4.1\n\tVI\t5.0\t0.9\t6.0\t0.0\t1.0\t4.0\t6.5\t0.5!\t1.5\t8.5\t6.0\t0.0\t1.0\t14.0\t6.0\t0.0\t1.0\t19.0\n\tVII\t! 6.0\t1.4\t6.0\t1.0\t0.0\t5.0\t6.0\t0.5\t0.5\t9.5\t8.5\t1.5\t2.5\t12.5\t7.5\t0.5\t1.5\t18.5\n\tVIII\t; v.o\t0.9\t7.5\t0.5\t0.5\t4.5\t9.5\t0.5\t2.5\t7.5\t7.0\t0.0\t0.0\t15.0\t14.0\t6.0\t7.0\t13.0\nS, regular stimulus present.\tD, difference between S and A.\nA, stimulus delayed the number of sec- I, amount of the illusion, onds indicated at the head of each section.\nThe results of these eight experiments may be divided into two classes according to the general laws which they express : (1) those in which the omission of the stimulus does not increase the discrimination-time to any marked degree, namely III, IV, VI, VII, and VIII ; and (2) those in which the discrimination-time was increased nearly proportionally to the time the stimulus was omitted, namely I, II, and V.\nIn the first class we observe that the time-suggestion was complete or nearly so. The observer looked at the disk and saw it grow brighter when there was no physical change in the intensity of its brightness, in nearly the same time as when there was an actual physical change.","page":44},{"file":"p0045.txt","language":"en","ocr_en":"Measurements of illusions and hallucinations in normal life. 45\nIn the second class, the illusion occurs sometimes as in the first, but when it does not occur the discrimination-time is shortened. This shortening may then be considered as a partial illusion, which is expressed in column I. There is only a difference of degree between the two classes. All observers were more or less deceived: The comparative extent of the illusion can be seen by a glance at column I in each section of the table.\nThere are especially two reasons why this colored disk was so efficient in producing the illusions. First, a memory image of the first intensity had to be compared with a later intensity. A memory image of color is, however, easily retained as long as was here required. But, secondly, this was made difficult by the gradual transition. A gradual change must be much greater than an abrupt one to be perceived.\nSince the whole range for change in intensity of the light at the rate employed was traversed in 16 seconds and the light was presented when it was at half intensity, it took only 8 seconds to bring it to its full intensity, where it remained constant. Hence those who required more than 8 seconds for discrimination perceived the difference when the intensity had actually been constant for some time.\nBesides the general fact of the existence of the illusion, the following four points have been established by this series of experiments.\n1.\tSuggestion, as a neglected quantity, vitiates a vast amount of laboratory experiments on perception of liminal differences.\n2.\tThe influence of suggestion is not always avoided by requiring the observer to discriminate between two stimuli, either of which may appear, because he may unconsciously look for only one of them or by some trivial circumstance suggest to himself that a particular one of the two will appear ; and if sufficient time be allowed, the illusion may occur just as if one definite stimulus had been expected.\n3.\tIf an observer begins an experiment warned against illusion and determined not to be deluded, this negative suggestion may be overcome by repetition of the real stimulus, and the positive suggestion will delude him as if he had been unwarned.\n4.\tDiscrimination for liminal differences depends largely on the regularity with which the physical stimulus is applied after the warning. If the stimulus be unexpectedly delayed, the observer will in some instances perceive smaller objective differences than ordinary, but more frequently he falls into an illusion, partial or total.","page":45},{"file":"p0046.txt","language":"en","ocr_en":"46\nC. E. Seashore,\nHallucinations of an object.\nMost of my experiments have been designed to produce hallucinations or illusions of single qualities or differences' in qualities of objects. Can hallucinations of complete objects be produced in the same way ? That they can, may be positively inferred from the results already obtained, but I have made a simple test to prove it directly.\nThe object, of which a hallucination was to be produced, was a blue bead, spheroidal, the shortest diameter being 1.8mm and the longest 3.5mm. It was suspended by a fine black silk thread in front of the center of a black surface, which was surrounded by a white circular border whose inside diameter measured 50mm. By a concealed device, the bead could be drawn away and replaced without the observer\u2019s notice.\nThe apparatus was placed on the wall facing the door in the dark room, so that when the door was open the observer could walk up an aisle in front of the door, constantly having the apparatus in line level with his eyes. The experimenter was seated by a table in the dark room, ostensibly to keep record, but really in order to manipulate the apparatus. The tests were made during daylight ; to avoid the shadows of the dark room and secure a fairly constant light, an incandescent light was kept burning in the ceiling of that room. It was essential that the light should be fairly constant only during one experiment.\nA tape line was stretched from the apparatus to a point 6.5m directly in front of it. The method employed was to first show the observer the bead in its position, then require him to go to the further end of the tape line and walk slowly up towards the apparatus until he could first see the bead distinctly. When he saw the bead, he read off the distance from the apparatus on the tape' line. I recorded the distance while he went back to repeat the trial 19 times in the same way.\nIn the first ten trials, the physical conditions were similar and he saw the bead at different distances with but a small variation. While he went back to start for the eleventh time, I pulled a cord which slid the bead behind the frame. The observer, not knowing this, walked up as usual, and when he came to, or a little beyond, the point where he expected to see it, he generally did see it, and read off the distance as before. While he looked at the tape line I slid the bead back in place.","page":46},{"file":"p0047.txt","language":"en","ocr_en":"Measurements of illusions and hallucinations in normal life. 47\nAs a rule, the eleventh, sixteenth, eighteenth and twentieth trials were made with the bead withdrawn.\nI did not make the test statistical at all, though it is simple and well adapted for examination of great numbers of individuals if due precaution be taken.\nAbout two-tliirds of the persons I tried were hallucinated. They knew when, where and how to see the bead, and this was sufficient to project the mental image into a realistic vision. This qualitative result was all that I desired to obtain and I carried the experiments no further.\nHallucinations of sound.\nThere are three prominent difficulties that confront us in trying to produce hallucinations of sound experimentally in normal life. (1) It is almost impossible to separate the required sound from other external sounds and from the subjective, or entotic, sounds. (2) Regulating and measuring the physical stimulus does not regulate and measure the resultant sensation. (3) It is difficult to induce a firm expectation of the sound by circumstantial suggestion.\nThe immediately following tests were of a precursory nature, and were not carried out with any scientific thoroughness, but some of the general observations may be of interest.\nA. Preliminary tests.\n1. The graphic method of recording1 was used, and the observer occupied the dark room. After a regular warning (one tap on an electric bell) faint clicks were given on a telegraph sounder placed on a table by which the observer was seated. There was no strong resemblance between these clicks and the sound of the bell. The warning was repeated every six seconds, and three clicks were given in rapid succession two seconds after the warning. The suggestion, induced by experience, was that these sounds would recur regularly in the same order. When, however, the experimenter caused an exception to this order, by occasionally not giving the clicks after the warning, the sound of the warning frequently directed expectant attention so strongly to the sequel, that the observer thought he heard the clicks also.\n1 Bliss, Investigations in reaction-time and attention, Stud. Yale Psych. Lab., 1892-1893 I 1.","page":47},{"file":"p0048.txt","language":"en","ocr_en":"48\nG. E. Seashore,\nThe time between the warning and the stimulus was made as long as two seconds to avoid the probability of automatic reaction.\nThe clicks from the sounder were too unfamiliar and strange, and hence difficult to reproduce. They had to be made so faint that the observer did not feel certain about his discriminations.\n2.\tInstead of this stimulus I tried the efficiency of a click produced in a telephone by closing the circuit. This had\u2019 the same difficulty, and the additional disadvantage of being definitely located in the diaphragm of the telephone.\n3.\tThe effect of a constant sound from a tuning fork was then tried. A 250 v. d. fork was kept vibrating before a telephone in a distant room, and by means of resistance, the sound received at the telephone in the dark room could be so adjusted that, allowing for all fluctuations, it would still be decidedly above the threshold of perception.\nThe observer was required to tap a key at a convenient rate, in circuit with the recording drum, as long as he heard the sound. It is a normal experience that sounds or tones of this character are readily reproduced as hallucinations, but for experimental purposes I could not get a suggestion definite enough.\nInstead of inducing a conviction that the sound was continuous, the very fact that the observer was asked to react as long as he heard the sound intimated that there would be an interruption. Notwithstanding this vagueness, several observers were so strongly hallucinated, that when the stimulus was withdrawn, they still kept on reacting until stopped. The stimulus had been applied long enough to produce a definite sensory image of the sound ; when the stimulus was withdrawn an after-image of it persisted for a moment, then the hallucinatory sensation persisted the better, because there was no contrasting sound to disturb or rectify it.\n4.\tIn order that an assurance of the continuity of the sound should be resolutely induced by auto-suggestion, I next tried the tick of a clock for stimulus. The sound was transmitted from the clock to the dark room by the Blake transmitter with telephone connections in which an open-circuit key was inserted. The intensity of the sound was regulated by resistance so as to be above the threshold of distinct perception when the observer\u2019s attention was at its lowest point while he was listening.\nThe advantage of this stimulus was that the tick of a clock is familiar and a memory image of it is easily retained. It was reg-","page":48},{"file":"p0049.txt","language":"en","ocr_en":"Measurements of illusions and hallucinations in normal life. 49\nular, definite and continuous. The observer knew I was recording his reaction, but he did not know how or why.\nWithout the use of any warning, he was asked to give four sets of reactions, namely : (1) every second tick, (the clock ticked 80 times per second) ; (2) every fourth tick ; (3) every fifth tick ; (4) every tenth tick. By these reactions the beats would be grouped with accent as in musical time. The observer would fairly throw himself upon the rhythmic movement, and, as he surrendered himself to it, a constant inhibiting - effort bad to be exerted in order to keep the reaction from anticipating the tick and gradually accelerating the rate.\nIt has been observed by M\u00fcnsterberg1 2 that a brief interruption in a stimulus of this kind will be bridged over. The same principle was involved in the experiments of Urbantschitsch3, N. Lange\u2019, Eckener4, Pace\u2019, Marbe\u2019, and others who studied the fluctuation of attention to liminal sensations.\nFrequently, in my experiments, from 1 to 10 ticks could be omitted and the observer would bridge it over. But, what is more remarkable, the hallucinatory sensations would sometimes rule out the sensation directly caused by the stimulus, so that the observer would e. g. react for 16 ticks while in reality only 12 were produced in that time.\nSuch hallucinations may be due to the tendency in an observer to accelerate any rythmic movement as it progresses.\n5. The next apparatus was constructed so as to produce a given sound at regular intervals. Two contacts on the kymograph were so connected with magnets on a stop-watch that when the first contact was made one magnetic armature was drawn, closing the circuit which included- an isolated 250 v. d. tuning fork and the primary\n1\tM\u00fcnsterbero, Schwankungen der Aufmerksamkeit, Beitr. zur exp. Psych., Vol. I, No. 2 p. 69.\n2\tUrbantschitsch, Zur Lehre von der Schallempfindung, Archiv f. d. ges. Physiol. (Pfl\u00fcger), 1881 XXIV 574; Ueber subjectiven Schwankungen der Intensit\u00e4t akustischer Empfindungen, Archiv f. d. ges Physiol. (Pfl\u00fcger), 1882 XXVII 436.\n3\tLange, Beitr\u00e4ge zur Theorie der sinnlichen Aufmerksamkeit und der activen Apperception, Phil. Stud., 1888 IV 390.\n4\tEckener, Untersuchungen \u00fcber die Schwankungen der Auffassung minimaler Sinnesreize, Phil. Stud. 1893 VIII 343.\ns Pace, Zur Frage der Schwankungen der Aufmerksamkeit nach Versuchen mit der Massonschen Scheibe, Phil. Stud., 1893 VIII 388.\n\u2022 Marbe, Bis Schwankungen der Gesichtsempfindungen, Phil. Stud., 1893 VIII 615.\n4","page":49},{"file":"p0050.txt","language":"en","ocr_en":"50\nC. M Seashore.\ncoil of an inductorium ; and when the second contact was made the other armature was drawn, breaking the circuit. The secondary current from the inductorium completed a circuit through telephones in the dark room.\nThe kymograph was run at the rate of one revolution in ten seconds, and the contact points were opposite each other ; hence, the sound was sent through for five seconds and then interrupted for five seconds. In this way a uniform alternation could be continued indefinitely by keeping the kymograph wound,at a constant tension.\nThe closing of the circuit, which produced the stimulus, started the stop-watch, and the observer\u2019s reaction stopped it. The time and regularity of the reaction were determined by the indications of the watch.\nTwo telephones were similarly connected in the same circuit and held one to each ear. When they were similarly adjusted, the sound was localized at the root of the tongue in the median plane of the head ; and, according as one was pressed relatively closer to the ear, the sound was localized to that side. But the contact of the telephones with the ear caused a disturbance, and it was difficult to adjust the two alike, Therefore the two telephones were permanently fixed facing each other, 150cm apart, in such a position that when the observer was seated his head would be midway between the two. The advantage of this was that the sound seemed to fill the room and could not be definitely localized.\nThe aim of the experiment was to produce hallucinations of this sound by repeating it until the observer had taken up the rhythm and acquired a clear mental image of its character.\nThe observer was required to react every time the sound recurred. After he had continued to do this regularly a short while, the sound was occasionally cut out by means of a switch. If a hallucination had been aroused, the observer would continue to react, as was frequently the case, for over a minute, when there was no stimulus present.\nBut here again the very suggestion raised a doubt in the mind of some observers ; and, by actual experiment, I found that the discrimination for sound is finer when the sound is interrupted at intervals, as here, than when it is continuous. Therefore this method was also abandoned after some trials, and the real experiments I carefully carried out were made by the apparatus and method now to be described.","page":50},{"file":"p0051.txt","language":"en","ocr_en":"Measurements of illusions and hallucinations in normal life. 51\nB. Main experiments on auditory hallucinations.\nThe apparatus for this series of experiments was so constructed as to produce a sound whose- quality was constant, and whose intensity could be varied at a desired rate.\nA 250 v. d. tuning fork, in an insulated box, was kept vibrating in the primary circuit of an inductorium. The induced current was completed through a telephone circuit in the dark room and could be broken by a switch. The telephone was fastened near the ceiling in such a position that it was 75cm back of, and 150cm above, the ears of the seated observer. This secured the same effect as the two telephones in the preceding arrangement, and the sound seemed to be distributed throughout the room. The quality of the sound was not a tone, but rather a distinct buzz, the result of the interrupted current. Telegraph sounders were used for signals, one in the dark room and one in the experimenter\u2019s room.\nThe method pursued can be best understood by examining the collective records in Table XIV with the following explanation in mind.\nThe observer occupied the dark and quiet room (a room with double padded walls supported on rubber disks so as to exclude sight and sound disturbances).1 After I had given him a trial signal and produced the stimulus for his recognition, the instructions wTere essentially : \u201cThe signal means, Listen ; every time you receive it, listen attentively until you hear the sound, and when you first perceive the stimulus, react. That reaction will indicate the threshold for yonr perception of sound.\u201d\nThe intensity of the sound was regulated by means of the inductorium. When the secondary overlapped the primary coil the indicator pointed to zero, and the sound was at its greatest intensity. As the secondary coil was moved back from the primary, the intensity of the sound decreased in a ratio favorable to the relative measure required. The figures in the record give the number of cm. by which the coils were separated. This is made the arbitrary measure of the physical stimulus.\nAt the moment the signal was given, the secondary coil was started so far down the scale that the sound produced would be below the threshold of perception, and then moved toward zero at\n1 Bliss, Investigations in reaction-lime and attention, Stud. Yale Psych. Lab., 1892-1893 I 2.","page":51},{"file":"p0052.txt","language":"en","ocr_en":"52\nG. E. Seashore,\nthe regular rate of 0.5cm per second. Thus if the starting point was 25cm, and the point when the reaction occurred 2t)cm, it required 10 seconds to reach the threshold.\nWhen the stimulus had been given about ten successive times, approximately once in 20 seconds, I continued to give the signal regularly, hut occasionally cut out the sound with the switch before\nTable XIV.\nHallucinations of sound.\n0\tp\tS\tMV\tNS\tH\tMV\tNS\tF\nI\t25\t20.8\t0.6\t3\t14.0\t\t\t1\t\t\nII\t27\t21.6\t0.8\t15\t\u2014\t\u2014\t\u2014\t-2\nIII\t25\t22.0\t0.9\t16\t20.8\t0.3\t2\t1\nIV\t25\t21.5\t1.0\t14\t\u2014\t\u2014\t\u2014\t2\nV\t25\t21.9\t0.8\t12\t20.9\t1.9\t6\t\u2014\nVI\t25\t22.6\t0.8\t17\t\u2014\t\u2014\t\u2014\t3\nVII\t27\t18.2\t1.1\t29\t20.0\t3.0\t2\t3\nVIII\t30\t25.6\t1.8\t11\t19.0\t....\t1\t\u2014\nIX\t30\t24.6\t1.5\t16\t24.3\t2.9\t3\t\u2014\nX\t27\t18.9\t1.3\t18\t\u2014\t\u2014\t\u2014\t2\nXI\t27\t21.3\t2.1\t17\t20.0\t2.0\t3\t\u2014\nXII\t25\t21.3\t1.8\t17\t22.5\t0.3\t3\t\u2014\nXIII\t26\t23.2\t0.7\t17\t24.2\t0.8\t3\t\u2014\nXIV\t32\t27.5\t0.9\t48\t28.5\t\u2022 1.0\t8\t3\n'XV\t30\t19.8\t0.4\t11\t\u2014\t\u2014\t\u2014\t1\nXVI\t28\t23.2\t0.8\t17\t\u2014\t\u2014\t\u2014\t3\nXVII\t30\t20.7\t0.6\t17\t\u2014\t\u2014\t\u2014\t3\nXVIII\t30\t25.1\t1.2\t18\t24.0\t\u2014\t1\ti\nXIX\t30\t23.0\t1.1\t16\t22.0\t\u2014\t1\t2\nO,\tobservers.\nP,\tpoint on the scale of the inductorium at which the indicator was started.\nS, point at which the stimulus was perceived.\nMF, mean variation.\nNSt number of trials in S.\nH, point at which the hallucination was realized.\nNH} number of hallucinations in H.\nF, number of failures to realize the hallucination in the limited time.\nthe signal was given. I moved the indicator up the scale at the same rate as before, in order to get an expression for the time-relation of the hallucination, which could be interpreted in terms of the rest of the record.","page":52},{"file":"p0053.txt","language":"en","ocr_en":"Measurements of illusions and hallucinations in normal life. 53\nTable XIV contains the record of the experiments on 19 persons, quoted in the order the tests were made ; negative as well as positive results. It is not merely a statistical fragment, but is intended to give the reader an opportunity to trace the actual procedure and experience in a test of this kind.\nTo explain the table, 25 at the head of column P in the record of observer I means that the secondary coil was started 25cm down the scale from the primary coil every time the signal was given. 20.8 in column S indicates (1) the relative intensity of the sound when first heard, and (2) the time required for the stimulus to rise to this intensity after the signal. The higher the number, the weaker the sound it represents, and vice versa.\nWe may designate the movement of the secondary coil, 0.5cm per second, as an increase of one degree-second in the intensity of the stimulus. Thus, on the average, observer I perceived the sound after an increase of 8.4 degree-seconds from the starting point ; observer II, 10.8 degree-seconds, etc. No sound was produced physically, still observer I heard it 22 sec. after the signal. 2 in column F means that in that experiment the physical stimulus was omitted twice, but no hallucination was produced in the time allowed, namely, 30 sec. When no hallucination had arisen at that time, the stimulus was applied with such an intensity that it could at once be perceived in order to keep up the suggestion, and such trials were properly recorded as failures.\nAfter each experiment it was ascertained whether the observer had been warned by any person or circumstances in any way to affect his judgment. It was found that nearly all those who gave negative results had been cautioned in some way to be suspicious, but this distinction between those who were warned and those who were not warned cannot be sharply drawn.\nThe psychological laboratory is something comparatively new to everybody. Persons come in there to see new things and are prepared for surprises, if that be possible. Since I did not tell any person that such and such a sound will recur so many times at such and such a rate, but left him to work up the auto-suggestion for himself, it may be reasonably supposed that every observer was in some degree sceptical as to the outcome of an experiment. The extent of this sceptical reserve very largely determined the degree of success in the experiment.\nTo shut a person up in a dark room is to intimate that he should pay attention to sounds. Then the associations with the signal tune","page":53},{"file":"p0054.txt","language":"en","ocr_en":"54\nG. E. Seashore,\nup the sensorium to such a tension that the experience of previous sensations is vividly imaged and the observer, as it were, feels and pictures the stimulus approaching the threshold, just as a batter feels and sees the ball approaching him after it has left the hand of the pitcher. For, by the method employed, a strong association was set up between the expected sound and the spatial and temporal environment.\nFrom this point of view, we may trace in the positive and negative aspects of the results two co-ordinate elements ; first the degree of expectation, and second the power of vivid imagination. I am confident that if it were required to select a number of persons who could be systematically hallucinated according to the present method, it could be done by means of a careful preliminary examination of them in these two respects.\nThe measure here is only relative and arbitrarily chosen, depending upon the law of increase in intensity of an induced current, but it serves the purpose quite as well as if the physical measurements were absolute. The threshold is never the same for different individuals, nor is it constant for the same individual at different times. The aim in view here was to produce a hallucination of a sound such that the observer could say positively, \u201c I hear it.\" The intensity of such a sound was determined by the delicacy of the observer\u2019s auditory apparatus and his accuracy of discrimination.\nOf the 60 attempts to hallucinate, 34 were successful and 26 unsuccessful, within the time-limit allowed. This is a large proportion of affirmative answers, when we consider that those who were forewarned are included.\nAfter the hallucination once is started it is easy to continue it, and each successive hallucination contributes toward the building up of a firm expectation, just as the repetition of the real stimulus does.\nThus, if there are three successful hallucinations and no failure in the first 20 trials, it is very probable that with such an accumulation of positive associations the experiment might with most observers be successfully continued through a series of trials without any stimulus. I made several trials on that point in addition to those recorded in Table XIV. As a rule, the hallucination would be kept up as long as the observer could continue unabated attention to it, and it would be rectified as attention slackened from fatigue. But for some observers the hallucination would continue indefinitely ; i. e. so far as my experiments went, for when they had been hallucinated in 20 or 25 successive trials, the experiment had to be discontinued on account of fatigue.","page":54},{"file":"p0055.txt","language":"en","ocr_en":"Measurements of illusions and hallucinations in normal life. 55\nThese experiments forcibly show the difficulty of exact measurement of mental activities and the danger of getting false results even from the most faithful observer. Had all these results been positive, it would have pointed to a despairing state of affairs with regard to our ordinary reliance upon the senses. The experimental psychologist will recognize in the method here used (excepting the omission of the stimulus) a familiar type of experiment on discrimination, which is generally taken in good faith. These tests prove it absolutely unreliable ; one little circumstance may determine what the observer shall hear or not hear without regard to the physical stimulus.\nTo determine the effect of repetition, I made the above experiment with one reliable observer, under similar circumstances, at ten different times. Each experiment consisted of twenty trials in which the stimulus was regularly applied in the first ten and omitted three or four times in the last ten trials.\nThese ten records show : (1) He discriminates for lower intensities in the latter part of each experiment. (2) Similarly he discriminates for lower intensities in the last than in the first experiments. (3) The hallucination is successfully produced in all but one of the thirty-seven trials. (+) The average reaction time to the hallucination is slightly longer than the reaction time to the real stimulus.\nThat the reaction time is shorter as the experiment progresses, may be partially accounted for by the fact that time is more and more over-estimated during a continuous strain of attention.\nUsing these ten experiments as a suggestion, I made a final experiment of twenty trials in which the stimulus was never applied. In each trial the observer signalled that he heard the sound, and after the experiment was over he was very confident that he had perceived the sound distinctly.\nBeyond the details discussed, these experiments on hallucinations of sound have demonstrated two points :\n1.\tHallucinations of sound distinctly above the threshold can be produced experimentally in normal life by leading the observer to concentrate expectant attention upon the desired result.\n2.\tExperiments to determine the threshold for perception of sound cannot be continued through a series of repeated trials, without being vitiated by the suggestion due to the accumulating associations.","page":55},{"file":"p0056.txt","language":"en","ocr_en":"56\nG. E. Seashore,\nHallucinations of touch, taste, smell and electric stimulation.\nThese tests are grouped together because they are of a rudimentary nature and depend to a great extent upon the same principle. There were three main points in the aim of the method :i (1) that the observer should clearly recognize the sensation to be produced ; (2) that by repetition it should become familiar and definitely associated with its cause ; (3) that the observer should be led to watch expectantly for it in the accustomed time and manner.\nTouch.\nI will mention in a cursory manner the methods by which I succeeded to produce hallucinations of touch. Pithballs were used for stimuli. The size and weight had to be adapted to the cutaneous sensibility of each individual, as well as to the place chosen for stimulation. I did not attempt any definite measurement of the size, weight or momentum of the impinging disk, but was satisfied when the force of the stimulus was such that the observer could say that he distinctly felt the sensations produced.\nAccording to the first method, the instructions were, \u201c Say \u2018 There \u2019 every time I touch you with the pithball at this point \u201d (e. g. the knuckle of the middle finger where there are no hairs).\nThe regular stimulus was then given six or eight times to build up the suggestion. After that it was occasionally omitted, i. e. according to the accustomed rate at which it was expected. The observer was seated behind a screen, and the experiment was tried with and without warning. When no warning was used, the first stimuli were given at a constant rate, about once in five seconds.\nSome of the observers presented the ludicrous spectacle of sitting behind the screen, actually feeling the sensation and confidently repeating, \u201cThere, there, there,\u201d etc., for an indefinite time after the apparatus used for stimulus had been laid aside.\nWhen the warning, \u201c Now \u201d, was used the stimulus was applied at irregular intervals. In this case time was no object in the reaction, hence the observer was not rushed to give a hasty judgment. Yet in about three-fourths of the cases, the mere warning was sufficient to produce the sensation of touch.\nIn a variation of this method, the observer was instructed to tell when he was touched after each warning. The place of stimulation was limited to the back of the right hand and wrist. I would","page":56},{"file":"p0057.txt","language":"en","ocr_en":"Measurements of illusions and hallucinations in normal life. 51\nactually touch him about ten times and then continue to give the warning without touching him. More than half of the observers were able to point out the place where they felt the hallucinatory sensations. That these sensations were above the threshold is shown by the fact that when the stimulus was present the observers could localize it correctly, and on inquiry it was found that the hallucinatory sensations sometimes appeared to be more distinct than the regular sensations.\nThe suggestion was increased by attaching the pithball to the pendulum of a metronome, by means of a cocoon fibre, so that every time the metronome ticked, when the pendulum was on one side, the ball would drop on the hand of the observer. When it had touched about ten times I caught the ball so that it did not touch the observer, though the metronome continued to tick as usual.\nThe association of the touch with the beat of the metronome strengthened the suggestion and made the results more definite and almost universally positive. The remarks \u201c distinct \u201d, \u201c very distinct\u201d, etc., about the hallucinatory sensation were frequent.\nThe most successful suggestion produced by associating the touch with a sound was, perhaps, the following :\nThe fibre on which the pithballs were suspended was attached to a special lever on the stop-watch, in such a manner that when the experimenter moved the lever, it simultaneously started the watch, and dropped the pithball. The observer\u2019s hand was comfortably adjusted in a fixed position with the palm up. He sat behind a screen and could stop the watch by his reaction on a key. The tick of the watch was so loud that he could hear it by paying attention.\nIn the first ten trials he had learned that the watch began to tick just as he felt the ball touch, and suggested to himself that this association was permanent. Hence, when after the tenth trial the ball was removed, he argued that every time he heard the tick of the watch again the touch would recur with it, and it actually did, so far as most of the observers\u2019 sensations were concerned.\nTaste.\nAn electrical stimulus was used in the first test on hallucinations of taste. The intensity of the induced current was regulated by means of changing the distance between the two coils of an ordinary inductorium. One of the electrodes was furnished with a platinum sheet 8mm square, which the observer applied to the tip of his","page":57},{"file":"p0058.txt","language":"en","ocr_en":"58\nG. JS. Seashore,\ntongue; holding the other in his hand. The observer was asked to react when he could perceive the acid taste of the current. He stood by the apparatus and saw me start the secondary coil at such a point, that, when moving it at a regular rate toward the primary coil, the threshold for perception of the current would be reached in about ten seconds.\nThis was repeated ten times, but the eleventh and following trials I cut out the current with a secret switch. The observer had by this time associated the cause with the effect, and elaborated the image of the sensation so thoroughly that he generally tasted the current in the expected time and manner.\nGustatory hallucinations were also produced by associations with liquid stimulation. The test is rude but it illustrates a principle.\nOf six bottles, two contained pure water, and the other four a series of solutions of pure cane sugar ; the first \u25a0\u00a7$, the second 1 <f0, the third 2$ and the fourth 4$ sugar, according to weight. A block was casually placed in front of them, so that the observer could not see them although he was aware that they stood near by him, because he saw them when he received his instructions. It was required of him to tell how weak a solution of sugar he could positively detect. I took a glass dropper and deposited a few drops on his tongue, drawing first from the two water bottles, and then from the sugar solutions, in order of increasing strength. The sugar was detected in the \\<f<, or 1$ solution the first trial. Proposing to repeat the test, I proceeded as before, but drew from the first water bottle every time. The result was that when the pure water had been \u201c tasted \u201d from two to ten times the observer almost without exception detected sugar.\nIt was not a persistence of the taste from the first trial, because in each repeated test it was not perceived at first.\nFrom the judgments passed, it appeared that, though the dropper was filled from the same water every time after the first trial, the liquid grew sweeter and sweeter to the observer, until he could confidently say that there was a decided taste of sugar.\nSmell.\nThe test on olfactory hallucinations was conducted similarly to the test last described.\nTwelve bottles containing solutions of oil of cloves were arranged in such a series that the first contained fresh distilled water, the","page":58},{"file":"p0059.txt","language":"en","ocr_en":"Measurements of illusions and hallucinations in normal life. 59\nsecond TTnro\u00f6Tnr part of \u00b0f doves, the twelfth\tpart of oil of\ncloves, and the intervening ones were arranged in geometric ratio according to strength.\nOrdinarily the oil of cloves could be detected in some bottle between the third and the ninth from the water. I showed the observer these bottles in a rack and then asked him to stand in the fresh air by an open window in the adjoining room, while I brought and let him smell of the liquid from one bottle at a time. Since -he was asked to tell in what bottle he could first perceive the smell of oil of cloves, he inferred that I would begin with the weakest solution and then take them in order of strength, though nothing was said to him about that. Some time before the real experiment, I let him smell of the solutions at random, so that he knew definitely what to expect, but in what he considered the test, I brought him the water bottle every time.\nAbout three-fourths of the persons experimented upon perceived the smell of oil of cloves from the pure water bottle when it had been brought from three to ten times.\nElectrical stimulation.\nFor this apparatus I used the inductorium with the 250 v. d. tuning fork vibrating in the primary circuit.\nThe electrodes of the secondary current terminated in the bottoms of two tumblers, filled with water and so adjusted in a frame that the observer could close the circuit by putting one finger of the same hand into each tumbler. The tumblers were filled with tepid water and each one was supplied with a rubber support, which allowed the fingers to be immersed lcm in the water.\nBy eight or ten trials I first found the threshold at which he could perceive the stimulus, i. e. the ordinary tickling sensation produced by the interrupted current, then I left it stationary so far above the threshold that the current must necessarily be perceptible during all ordinary physical and mental fluctuations. The stopwatch was so connected with a rocking commutator in the secondary circuit, that when the commutator closed the circuit and thereby applied the stimulus, it simultaneously started the stop-watch. The observer stopped the watch with his reaction.\nHe was seated behind a screen and could not see the watch and the commutator, but could hear the former faintly. The instruction was that he should associate the tick of the watch with the electrical","page":59},{"file":"p0060.txt","language":"en","ocr_en":"60\nG. E. Seashore,\nstimulation. He was asked to react when he first perceived the sensation produced by the current, after the warning \u201c Ready.\u201d\nTo limit the suggestion I made the interval between the warning and the application of the stimulus very irregular. The observer began to hear the tick of the watch the same instant he felt the shock from the stimulus. This was repeated ten times, and after that the stimulus was occasionally omitted but the warning given and the watch started. When nothing was said about the watch, the hallucination very seldom failed to appear, but when the observer was positively warned not to let himself be influenced by the sound the hallucinations were less frequent. When the watch was not used and the warning was still given regularly in the preparatory trials, only a small per cent, of the observers were hallucinated.\nIn this test, as in the first test on taste, fear or excitement may have been the immediate cause of the hallucination for some observers. In that respect these tests are exceptional.\nThough I have cited no measurements on this class of hallucinations I have the experimental evidences, which might have been thrown into tables here had it not been superfluous, after the reports on sight and sound.\nConclusions.\nFrom this series of experiments the following conclusions may be drawn.\n1.\tHallucinatory sensations of liminal intensities of touch, taste, smell and electric shocks may be experimentally produced in normal life.\n2.\tAccording to the methods here employed, these hallucinations were produced, (1) by leading the observer to expect the respective sensations at a certain time and in a definite manner ; (2) by associating the desired sensations with a warning ; and (3) by associating them with simultaneous and continuous stimulations of other senses.\nGeneral remarks on the experiments in Part Second.\nThe general principle here studied belongs to the realm of the influence of feeling in perception ; but this series of experiments is limited to the influence of the more intellectual emotions, and these are reduced to a minimum as in the process of deliberate, voluntary, expectant attention in normal unexcited, intelligent persons.","page":60},{"file":"p0061.txt","language":"en","ocr_en":"Measurements of illusions and hallucinations in normal life. 61\nMany faults may be found with the above pieces of apparatus from the point of view of physics, for none of them afford means of absolute control and measurement of the physical stimulus. I cannot say I produced so loud a sound or so strong a taste, measured in absolute physical quantities, assuming a parallel between the physical stimulus and psychical correlate ; but the main desideratum was to secure such conditions that the observer could say without hesitation, I see, hear, touch, taste, etc.\nSince the method was a combination of the experimental and the statistical, it was impossible to secure trained psychologists as observers in all cases. Nor was that necessary, for if I obtained observers with a good general power of discrimination and reliable judgment, the conditions from the experimental side were fairly satisfied ; and, by taking a large number of persons, the demands from the statistical side were not neglected.\nThe tests purport to be made in normal, waking life. With this condition in view much discretion has been exercised in selecting such observers as could most properly be spoken of as normal. Advanced students (men) would naturally be considered the most reliable observers ; at the same time they were the most easily accessible to me and readily interested in the work. I always tried to keep the observer uninformed as to the actual outcome of the experiment.\nAll possible precaution was taken to secure the conditions of a normal judgment. The observations were made under the most favorable physical conditions ; distraction was guarded against ; automatic reactions were practically precluded ; and the case was in general made so simple that the important conditions could be fairly controlled.\nThe psychological process involved in these experiments is expressed by one general principle\u2014suggestion.\nSuggestion presupposes mind as a selective agency which has the power of choice and interpretation, and that this choice and manner of interpretation can be directed to a great extent by external stimuli and central associations.\nWe may outline the characteristics of its use in these experiments. The method was to induce expectant attention to a liminal sensation by the presence of an object or some circumstance connected with it, in that way awakening a mental image which should realize itself in a sensation. The first step was to give the observer an idea of what to expect. His attention was called to the apparatus and the prin-","page":61},{"file":"p0062.txt","language":"en","ocr_en":"62\nG. JE. Seashore,\nciple by which it produced or served as a stimulus, and an actual example of the character of the stimulation was produced. Hence he acquired an image of the given sensation, and an idea of the process by which it was produced and began to associate them as cause and effect. It was next necessary that he should suggest to himself the fact, manner, and time of the recurrence of the sensation. The methods were such that he would build up this suggestion by a series of associations which may be reduced to the following four classes.\n1.\tThe observer was by his own choice and interest led to. notice the manipulations of the apparatus and form a definite association between the action of the apparatus (or his change of relation to it) and the resulting stimulation, so that he knew what the result would be, that he would perceive it, and when, he must expect it, e. g'. in the experiment with the heated wire and with the bead.\n2.\tA signal or warning became associated with stimulation in certain qualities and in time succession, e. g. in the light and the sound experiments. The observer suggested to himself certain permanent relations between the signal and the sensation, without having been told anything to that effect. The signal aroused the mental image and in the absence of any inhibition it realized itself in the customary time and manner.\n3.\tWhen the apparatus was not in view and no signal was used, a rhythmic order of the recurrence of the sensation was sometimes set up, e. g. in the hallucinations of the tick of the clock and some of the tactual hallucinations. ' The physiological tendency toward rhythmic action and the mental associations of regular recurrence established a firm auto-suggestion upon false grounds.\n4.\tBy syn\u00e6sthesia one sensation brought forth another sensation. A certain stimulation produced not only the corresponding sensation but also associated sensations of the same sense or of entirely different senses. This is more or less included in the three preceding methods, but it is particularly involved in the test on electric stimulation and some touch and taste experiments.\nExpectant attention is the unitary principle in all these kinds of suggestion. By it a lifelike image of the sensation was awakened and the judgment of the observer was so profoundly influenced by this that he convinced himself of the actuality of the sensation for which no corresponding physical stimulus existed.","page":62},{"file":"p0063.txt","language":"en","ocr_en":"Measurements of illusions and hallucinations in normal life. 63\nPART THIRD.\nSome conclusions and deductions.\nThe following statements are deductions that seem to be justified by the experimental data and the facts established by them.\nExperimental.\nJudgments, which were the phenomena measured in this investigation, do not depend alone on the particular sensations judged, but also on other sensations and on interest, imagination, and expectation, in which feeling predominates ; attention and motor readiness, where will predominates ; and memory, discrimination, and association, where thought processes predominate. Many of the associations and states of mental preparedness depend upon suggestions from the environment. A seemingly insignificant word, thing, or circumstance may determine what the observer shall perceive or not perceive.\nThe effect of disappointed expectant attention in judgment of weight is shown in Part First ; and it is reasonable to suppose that it enters into other classes of precepts in a similar manner. In Part Second, the effect of expectant attention which frequently realizes its object is set forth. In most psychological experiments it is necessary to secure close attention. It is well known that by inattention an endless number of errors occurs, but it is often overlooked that forced attention, which necessarily rises into expectation, is one of the factors that must be carefully guarded against in experiments on liminal differences. An inference which I consider legitimate in regard to all these results is, that if there is so great deception in perception of small differences, there is a corresponding deception in regard to details in perception of large objects or groups of objects.\nThe experiments in Part Second demonstrate four different methods in which this vitiating effect of expectation may enter.\n1. Attention to the associations between a sensation and a definite process of physical stimulation may cause an illusory interpretation of the sensation or cause the sensation to occur without the physiological modification by the external stimulus. The experiments on hallucinations of temperature, e. g. prove that our interpretation of liminal sensations is utterly unreliable, when we attend to the physical stimulus by other senses than the one stimulated. Hence if we would have a sensory perception true to that normally perceived","page":63},{"file":"p0064.txt","language":"en","ocr_en":"64\nC. E. Seashore,\nfrom the external object, special precaution must be taken in this respect.\n2.\tAttention may rise into a firm expectation by means of a definite signal. Psychologists have recognized that when a mental image of a sensation has been produced and a signal given, the time of the perception will be influenced even to so great an extent that so-called negative reactions may occur. The above experiments have proved that the use of a signal in experiments upon liminal sensations has a controlling effect upon the sensation to follow ; in view of these results it is necessary to reconsider the use of signals in the customary way.\n3.\tIn our ordinary physical and mental activity there is' some degree of rhythm or time-order. Even if we guard against associations with apparatus or warning, there is yet a trap. If successive trials are given in some order so that the observer can in any way have reason to expect the sensation at a certain time, the discrimination will be influenced. This applies even if it be required to discriminate between two definite sensations, either of which niay appear. Also, if two sets of experiments with a slight difference in the rate of change, e. g. be taken immediately after each other, the expectation based upon the experience in the first set will influence the perception in the second set. Hence the importance of arranging single trials, as well as sets of experiments, in the proper order so as to eliminate the effect of expectation.\n4.\tBy virtue of syn\u00e6sthesia, we often perceive single qualities by two or more senses. One sense suggests to the other either by aiding in completing the image, as sight and touch, or by associating sensations which generally occur together in consciousness, as in combinations of sight and sound, in which there are inferences from one to the other. Therefore it is important in laboratory experiments to control the stimulation not only of the sense which is experimented upon, but also the stimulations of those senses which may in some way be associated with it in experience, otherwise the observer may easily deceive himself by taking the stimulation of one sense for the stimulation of another and actually perceive the latter. This kind of suggestion is more or less involved in all the preceding classes.\nPathological.\n1. Illusions and hallucinations and all the phenomena in which they are the constituent elements (except those due to physical","page":64},{"file":"p0065.txt","language":"en","ocr_en":"Measurements of illusions and hallucinations in normal life. 65\npathology) may be experimentally produced in mild forms in fairly normal states, for the purpose of studying their nature.\n2.\tTo be normal is to have a firm inhibiting force in the form of discriminative consciousness which checks the realization of vagaries of imagination.\n3.\tExperiences in all forms of illusion may be realistic. People really see ghosts. If a scientific observer in the bead experiment sees the bead as real although there is no bead, I do not think we can set any limit to what an excited, imaginative person may really see under circumstances favorable for illusion. Being fully convinced that the sensations produced by suggestion in my class of observers were, as a rule, real, we may safely infer that the suggested experiences during hypnosis are as a rule realistic. Arguing by analogy from the present experiments, we may go far to explain all the phenomena which depend upon sense illusion. The modern sleight-of-hand performer does not pretend to do anything but delude his audience. How much of the spiritualistic seance and all the phenomena of that category would remain, if a psychological analysis of the illusions were pushed far enough ? Yet good men and women tell us those wonderful experiences are real.\n4.\tMind acts according to laws in normal states, and abnormal states are caused by a morbid or overstimulated activity of the same laws when some side of the regulative activity of mind is dormant.\n5.\tGreisinger' assigns the following causes for hallucinations : (1) local disease of organs of sense, (2) state of deep exhaustion either of mind or body, (3) morbid emotional states, (4) outward calm and stillness between sleep and waking, and (5) the action of certain poisons. The experiments I have made justify the addition of another general source of illusion, namely, expectant attention.\nThe different forms of suggestion in which this source of illusion is prominent, as determined in the experiments, corresponds to the four kinds of associations which constituted the suggestion as detailed in the foregoing section. Thus, the association of a sensation with the physical stimulus which is supposed to produce it constitutes the virtue of all the devices which have been used to induce hypnotism,\u2014magnets, vials, gongs, crystals, fountains, incantations, and even the word and presence of the hypnotizer. They all serve the same purpose and have the same virtue, i. e. they are the means by which the subject arrives at a certain degree of expectation and conviction.\n1 Sully, Illusions, p. 115. 5","page":65},{"file":"p0066.txt","language":"en","ocr_en":"66\nG. JS. Seashore,\nEpistemological.\nA thorough investigation of the epistemological problem begins with a study of the psychology of perception, the simplest form of immediate knowledge. The first question there answered will naturally be, \u201c To what extent is our sense-intuition reliable ? \u201d This part of epistemology should be treated in the psychological laboratory. Extended laboratory experiments with a thorough introspective analysis must pronounce upon the psychological problem of sensory illusions and hallucinations, and metaphysics must take this decision into account.\nThe above results are of a very rudimentary character with reference to this problem, and they were not made with it especially in view, but to me certain empirical facts have been made prominent and clear.\nLooking first at the side which disparages our confidence in knowledge of things as they really are, the following considerations present themselves. All mental activities are involved in common acts of perception. The intricate process involved in the perception of a single quality was illustrated by the weight test, and the same line of observation might be infinitely extended to show that to become known to me the quality of an object must enter and modify my stream of consciousness and adapt itself to it. In this complex process the data of sense are profoundly modified by central states and activities. We noticed the influence of one of these, expectant attention, in its twofold character as disappointed expectation and as realized expectation, when based upon false grounds. What we call normal perception involves many illusory influences\u2014not only those of physical and physiological origin, but even more so those due to the functions of ideation, memory, and imagination. Indeed, suggestion and imagination control all our perception by the senses. Yery intelligent men are liable to embellish and misinterpret their sense data even under circumstances favorable for accurate perception. The perception of liminal differences is subject to so many misleading influences that, as a rule, it is extremely unreliable ; and the cognitive functions act very imperfectly in giving us a detailed representation of external objects.\nBut on the positive side, confirming us in the belief that somehow we perceive things as they are, several important facts may be observed. Illusions work according to laws which may generally be determined. As these become known we may gradually learn to","page":66},{"file":"p0067.txt","language":"en","ocr_en":"Measurements of illusions and hallucinations in normal life. 61\nrule out the illusion. The known physical and physiological illusions do not necessarily delude us because we may make allowance for them. Similarly, we may now make an approximate allowance for the illusions of weight and for all other illusions, due to intel-lectualized feelings, as they become recognized. The view that illusions and hallucinations do not act according to law is as wrong as the view that mind in its normal capacity is lawless. The more thoroughly we become acquainted with the laws of illusions the more accurately will our sense perceptions fall in consensus.","page":67}],"identifier":"lit23180","issued":"1895","language":"en","pages":"1-67","startpages":"1","title":"Measurements of Illusions and Hallucinations in Normal Life","type":"Journal Article","volume":"3"},"revision":0,"updated":"2022-01-31T16:19:18.845571+00:00"}