There are also converging lines of evidence that automatization skill is a key aspect of cognition. Sternberg (1977) found that the correlation between performance on an analogy problem and measure of general cognition increased with practice. The first stage of Ackerman's (1987; Kanfer and Ackerman, 1989) model of automatization also is related to cognition. Theorists such as Jensen (1982) and Hunt (1978) attribute the correlation between such tasks as choice reaction time and letter matching to the relation between speed of information processing and cognition. An alternative explanation is that some of the correlation is due to the effects of automatization of processing.

The skill to deal with novelty and the skill to automatize information processing are interrelated. If one is able to automatize well, one has more resources left over for dealing with novelty. Similarly, if one is well able to deal with novelty, one has more resources left over for automatization.

The contextual subtheory. According to the contextual subtheory, intelligent thought is directed toward one or more of three behavioral goals: (a) adaptation to an environment, (b) shaping of an environment, or (c) selection of an environment. These three goals may be viewed as the functions toward which cognition is directed. Cognition is not aimless or random mental activity that happens to involve certain components of information processing at certain levels of experience. Rather, it is purposefully directed toward the pursuit of these three global goals, all of which have more specific and concrete instantiations in people's lives.

Most intelligent thought is directed toward the attempt to adapt to one's environment. The requirements for adaptation can differ radically from one environment to another—whether environments are defined in terms of families, jobs, subcultures, or cultures. According to the triarchic theory, and the contextual subtheory in particular, the processes, experiential facets, and functions of cognition remain essentially the same across contexts, but the particular instantiations of these processes, facets, and functions can differ radically. Thus, the content of intelligent thought and its manifestations in behavior will bear no necessary resemblance across contexts. To understand cognition, one must understand it, not only in relation to its internal manifestations in terms of mental processes and its experiential manifestations in terms of facets of the experiential continuum, but also in terms of how thought is intelligently translated into action in a variety of different contextual settings. The difference in what is considered adaptive and intelligent can extend even to different occupations within a given cultural milieu.

Shaping of the environment is often used as a backup strategy when adaptation fails. If one is unable to change oneself to fit the environment, one may attempt to change the environment to fit oneself. Shaping, however, is not always used in lieu of adaptation—it may be used before adaptation is tried. In science, the distinction can be made between those who set the paradigms (shape) and those who follow them (adapt) (see Sternberg, 1999).

Selection involves renunciation of one environment in favor of another. Selection is sometimes used when both adaptation and shaping fail. Failure to adjust to the demands of a work environment, or to change the demands to fit one's interest, values, expectations, or skills, may result in a decision to seek a new job. But selection is not always used as a last resort. It may reflect an intelligent person's recognition that a situation is not suitable and that no attempt to change oneself would improve the fit.

Adaptation, shaping, and selection are functions of intelligent thought as it operates in context. It is through adaptation, shaping, and selection that the components of cognition, as employed at various levels of experience, become actualized in the real world. This is the definition of practical cognition used by Sternberg and his colleagues (e.g., Sternberg, 1997a; Sternberg and Wagner, 1986).