An approach to temperament
The behaviors of young children that are classified as temperamental can be named with constructs that remain close to the events (a Baconian strategy that minimizes error) or constructs that are considerably more abstract. The broader concepts are preferred by most scholars, but they require a large corpus of reliable information and a unifying idea. Because facts have been thin and theory weak, the names for temperamental characteristics have usually, and properly, referred to observed qualities, such as activity level, lability of mood, crying, and approach or withdrawal to novelty (Buss & Plomin, 1975). During the last two decades, however, biologists and ethologists have produced rich descriptions of behavioral ontogeny in mammalian species that hold the promise of providing an initial rationale for parsing the dimensions of temperament from above rather than from below.
One of the many theoretical perspectives that might be used for the classification of temperament rests on four assumptions. The first, which is noncontroversial, assumes that all animals possess structures and functions that permit them to cope with at least four universal adaptational problems: eating, sleeping, reproduction, and protection from harm. The second, only a little less obvious, states that evolutionary changes within the mammals have involved alterations in the degree of preparedness to react to particular incentives and to issue specific motor actions that serve the adaptation domains. During mating, for example, rodents are prepared to be especially sensitive to chemical signals; primates are more sensitive to visual cues. When harm is a possibility, kittens are prepared to freeze to a threat; primate infants become excited and emit distress calls.