Page:Philosophical Review Volume 3.djvu/235

] in so far as the method of minimal changes is to be classed under the general heading of j. o. d., the experiments on the extent, form, and time of movement leave the question untouched. From the very nature of the case, where the 'experimentee' himself produces the norm, and then seeks to bring a second sensation or complex of sensations into a given relation with the norm, no method of gradations could be employed, and therefore none of those precautionary measures, which, in the opinion of the writer, make the m. of m. c. as valid as the error methods, were available. Then, too, in experiments of this kind on movement, aside from the great complexity of the perceptions involved, there comes in a disturbing factor which is not generally present in the comparison of sensations of the special senses, viz., between the norm and the compared stimulus there intervenes a third complex of sensations due to the return of the limb to its initial position. If, in getting the j. o. d. between two tones, a third tone very different from either should intervene, it would seriously affect the judgment.

The writer must confess to a feeling of surprise to find the so-called method of doubled stimuli lumped together with the method of mean gradations under the general heading of "method of estimated difference." Beside Merkel, the inventor of this method, no one, so far as I know, has taken the method seriously,—certainly no accepted conclusions have been drawn from its use. As Wundt first suggested, and as Fullerton and Cattell assert, the estimation of one stimulus as double another is a matter of association and experience. A complete analysis of the method of mean gradations would take too much space; it may, however, be pointed out that this method is generally regarded as a valid one, and that its validity has been tested by searching investigations which were independent and mutually corroborative (Neiglick, Lehmann, Ebbinghaus). So far as association and experience play any part in it whatever, they play a disturbing part, and the more such influences are removed, the more accurate and constant the determinations become. The difference between the two methods is a vital difference,—in the method of doubled stimuli we find a direct comparison of sensations,—in the method of mean gradations we find a comparison of differences of sensation.

Of especial worth in this treatise are the conclusions drawn from a comparison of the results of the several psychophysical methods in the same line of investigation, and from acute evaluation of the data.