| − a + b | responded to by | rnot 1 + r2, |
| − a + g | " " | rnot 1 + rnot 2, |
| − a + l | " " | rnot 1 + r12, |
| − a + q | " " | rnot 1 + rnot 12, |
| − a + v | " " | rnot 1 + r22, and |
| − a + B | " " | rnot 1 + rnot 22, as shown in Scheme III. |
Scheme III
| a | opp. | b | g | l | q | v | B | |
| of a | (opp. of b) | (opp. of l) | (opp. of v) | |||||
| r1 | 6 | 1 | 1 | 1 | 1 | 1 | 1 | |
| rnot 1 | 6 | 1 | 1 | 1 | 1 | 1 | 1 | |
| r2 | 1 | 1 | 2 | |||||
| rnot 2 | 1 | 1 | 2 | |||||
| r12 | 1 | 1 | 2 | |||||
| rnot 12 | 1 | 1 | 2 | |||||
| r22 | 1 | 1 | 2 | |||||
| rnot 22 | 1 | 1 | 2 | |||||
In this series of twelve experiences a connects with r1 six times and the opposite of a connects with rnot 1 six times. a connects equally often with three pairs of mutual destructives r2 and rnot 2, r12 and rnot 12, r22 and rnot 22, and so has zero tendency to call them up. − a has also zero tendency to call up any of these responses except its opposite, rnot 1. b, g, l, q, v, and B are made to connect equally often with r1 and rnot 1. So, of these elements, a is the only one left with a tendency to call up r1.
Thus, by the mere action of frequency of connection, r1 is connected with a; the bonds from a to anything except r1 are being counteracted, and the slight bonds from anything except a to r1 are being counteracted. The element a becomes predominant in situations containing it; and its bond toward r1 becomes relatively enormously strengthened and freed from competition.
These three processes occur in a similar, but more complicated, form if the situations a + b, a + g, etc., are replaced by a + b + c + d + e + f, a + g + h + i + j + k, etc., and the responses r1 + r2, r1 + r7, r1 + r12, etc., are replaced by r1 + r2 + r3 + r4 + r5 + r6, r1 + r7 + r8 + r9 + r10 + r11, etc.—provided the r1, r2, r3, r4, etc., can be made singly. In so far as any one of the responses is necessarily co-active with any one of the others (so that, for example, r13 always brings r26 with it and vice versa), the exact relations of the numbers recorded in schemes like schemes I, II, and III on pages 172 to 174 will change; but, unless r1 has such an inevitable co-actor, the general results of schemes I, II, and III will hold good. If r1 does have such an inseparable co-actor, say r2, then, of course, a can never acquire bonds with r1 alone, but everywhere that r1 or r2 appears in the preceding schemes the other element must appear also. r1 r2 would then have to be used as a unit in analysis.
The 'a + b,' 'a + g,' 'a + l,' ... 'a + B' situations may occur unequal numbers of times, altering the exact numerical relations of the connections formed and presented in schemes I, II, and III; but the process in general remains the same.
So much for the effect of use and disuse in attaching appropriate response elements to certain subtle elements of situations. There are three main series of effects of satisfaction and discomfort. They serve, first, to emphasize, from the start, the desired bonds leading to the responses r1 + r2, r1 + r7, etc., to the total situations, and to weed out the undesirable ones. They also act to emphasize, in such comparisons and contrasts as have been described, every action of the bond from a to r1; and to eliminate every tendency of a to connect with aught save r1, and of aught save a to connect with r1. Their third service is to strengthen the bonds produced of appropriate responses to a wherever it occurs, whether or not any formal comparisons and contrasts take place.
The process of learning to respond to the difference of pitch in tones from whatever instrument, to the 'square-root-ness' of whatever number, to triangularity in whatever size or combination of lines, to equality of whatever pairs, or to honesty in whatever person or instance, is thus a consequence of associative learning, requiring no other forces than those of use, disuse, satisfaction, and discomfort. "What happens in such cases is that the response, by being connected with many situations alike in the presence of the element in question and different in other respects, is bound firmly to that element and loosely to each of its concomitants. Conversely any element is bound firmly to any one response that is made to all situations containing it and very, very loosely to each of those responses that are made to only a few of the situations containing it. The element of triangularity, for example, is bound firmly to the response of saying or thinking 'triangle' but only very loosely to the response of saying or thinking white, red, blue, large, small, iron, steel, wood, paper, and the like. A situation thus acquires bonds not only with some response to it as a gross total, but also with responses to any of its elements that have appeared in any other gross totals. Appropriate response to an element regardless of its concomitants is a necessary consequence of the laws of exercise and effect if an animal learns to make that response to the gross total situations that contain the element and not to make it to those that do not. Such prepotent determination of the response by one or another element of the situation is no transcendental mystery, but, given the circumstances, a general rule of all learning." Such are at bottom only extreme cases of the same learning as a cat exhibits that depresses a platform in a certain box whether it faces north or south, whether the temperature is 50 or 80 degrees, whether one or two persons are in sight, whether she is exceedingly or moderately hungry, whether fish or milk is outside the box. All learning is analytic, representing the activity of elements within a total situation. In man, by virtue of certain instincts and the course of his training, very subtle elements of situations can so operate.