THE TONIC HYPOTHESIS AND THE PATTERNING OF BEHAVIOR *

One useful way to think about problems of behavioral organization is in terms of the question of pattern formation. A pattern in this sense is defined as the separation of behavior into component activities on the one hand and the examination of linkages between these components on the other. We can think here in terms of an analogy to a musical performance, where the performance must be defined both in terms of individual notes and pauses and in terms of the simultaneous and sequential articulation of these individual notes and pauses. The question for analysis in terms of behavioral mechanisms is the extent to which definable underlying processes are circumscribed in terms of their “temporal” and “lateral” aspects, This problem is often summarized in terms of the issue of control specificity.S* It is quite clear that certain causal processes in behavior may extend over longer time periods than the individual activities that they influence. For example, seasonal changes in hormonal state in a stickleback fish can produce long-term consequences, as measured by the relative probabilities of various activity patterns such as fighting and courtship.13 When viewed from a long time scale perspective, the hormonal shifts that occur during the onset of breeding season in sticklebacks can increase the probability both of fighting and of courtship behavior, as shown in FIGURE 1. Short-term interactions between fighting and courtship are often negative rather than positive, however. This means that the occurrence of fighting behavior reduces the likelihood that courtship activities will be observed in close proximity on the one hand, and that occurrences of courtship behavior decrease the likelihood that fighting will be observed within the next short time frame, on the other. Clearly, one of the major tasks of the behavioral scientist is to come to grips with the relations between causal processes that operate over quite disparate time scales. We can, for convenience, call the relatively long-term processes “tonic.” The fact that processes exist that extend beyond the articulation of any individual form of behavioral expression carries with it the logical conclusion that these processes exist during the performance of several different classes of activity. A question that follows, therefore, is whether certain classes of causal factors may influence several different classes of overt behavioral expression.