DISCRIMINATION, SENstat.r., or DIFFKR• 1:NM% I. SENSITIVITY. The liana. applied to an in tellectual function of ext reme importanee---our ca paeity to pronounce upon the likeness or differ mice of sense-presentat ions. The phrase "must not be taken to denote a faculty of eompa rison. in t he of a peculiar eonseious process existing a long.ide of the various contents. It merely ex proses, in the first instance, the general foot that %to have different experienees rind ew.peri once them dilTerently; in Other words, it eovers the introspection (q.v.) of different contents and the report of their difference. . . But we also use it to indicate our experience of like contents, mid our report of their (Knipe). The differential sensitivity would, then, be exercised if we were called upon to Judge of the likeness or difference of two reds, or of the a of two violins. ln such eases. we a I experience the colors and tones in the same or in different ways; they make the same or a different impression upon our mind: while we then (b) formulate this experience in words, saying 'like' or 'different.' as introspection dic tates. The former process, of introspective dis crimination, is termed the direct, the latter, of rendering the experience in words, the indirect differential sensitivity. There are, then, two problems before us. We must ask, first. as to the conditions under which we are able to discrimi nate, and the limits within which discrimina tion is valid: and, second, as to the adequacy of language to reflect or reproduce the results of introspection.
(1 ) A very little consideration suffices to show that an accurate comparison of sense-contents cannot be undertaken in any and every case of difference; but that certain favorable conditions are necessary to a reliable judgment. Suppose, e.g. that we have two reds before us—a saturated red. and a reddish gray of the same brightness. (See SATURATION.) It is impossible, with all imaginable care in comparison, to assert that the saturated red is three or four or any definite number of times as red as the less saturated color. The one impression is that of a 'good' red; the other that of a dull or gray red; more we cannot say. or suppose that three pressures are given, a light, a moderate, and a heavy. It is impossible to assert with ally confidence that the moderate pressure is three times as much heavier than the light as the heavy is than the moderate: we can say that the former difference is 'a good deal greater' than the latter. hut that is all. Now experiment has proved that the conditions favorable to exact discrimination are realized in two, and only in two, instances: and it is, accordingly, with these that investigators of the differential sensitivity have occupied them sel•es. The first instance is that of 'difference
determination.' the second is that of 'difference comparison.' In the former procedure. the psy chologist aims to determine the 'just noticeable difference' between two like contents: in the lat ter, he seeks to effect a subjective equality of stimnlus-differenees. An illustration of differ ence determination appears under the title MTION (9.v.) : it is found that two tones are diseriminable. in the middle region of the music al scale, if their pitch-numbers differ by 0.2 of a vibration. This value 0.2 is the just notice able difference of tonal quality. As illustrations of difference comparison, we may take the follow ing: given two tones, it is required to find a tone which lies, for sensation, midway between them: and, given a black and a white, it is re quired to find a gray which appears to be equi distant from both. Let the tones have. e.g. the pitch-numbers 200 and 400: then the middle tone has (approximately) the pitch-number the 4timulus-differenee 300-200 is equal. in sen sation, to the stimulus-difference 400-300. In other words, we have the required middle tone when we have the arithmetical mean between the two tonal stimuli. With hrightnesses the ease is different. Let the unit of black have the photo metric value 1. and the unit of white the photo metric value Oil: .o that a white disk (360') stands to a black disk as 360 x tlo to 36u X or as 21,010 to 301. The middle gray is not the arithmetical mean, a disk of Isu white and 180° black (value 1.0,!Ntl, This much too light. The middle point is found rather with a disk of some 41 of white and 31!) of black; i.e. with a stimulus whose photometric value is approximately 27sS. the geometrical mean bet \leen the limiting Idaek and White. It is clear, then, that the course of the differential sensitivity is not the same for tones and for brightnesses. By an extension of experiment to the qualities of the other senses, and to the sensa tion-attributes other than quality (see SExsA 'flux ). various laws of the differential sensitivity have been established. I an example, see WEBER'S LAW.) A valid determination of these laws is possible only on the condition that the sources of experimental error are strictly con trolled by the investigator. Attention, expecta tion, and practice must be at their best: habit uation and fatigue must be avoided (see these terms). The 'constant errors' of time and space must also be ruled out: as many experiments must be taken in the time-order oh as in the order La, and as many with stimulus a to the right as with the same stimulus to the left; and the results of the double series averaged.