Similar results were obtained in a study by Cornelius and Caspi (1987). They studied adults between the ages of 20 and 78. These researchers examined relationships between performance on tasks measuring fluid cognition (letter series), crystallized cognition (verbal meanings), and everyday problem solving (e.g., dealing with a landlord who won't make repairs, filling out a complicated form, responding to criticism from a parent or child). Performance on the measure of fluid skill increased from age 20 to 30, remained stable from age 30 to 50, and then declined. Performance on the everyday problem-solving task and the measures of crystallized skill increased through age 70.

Likewise, the neofunctionalist position, advanced by Baltes and his associates (Baltes, 1987; Baltes et al., 1984; Baltes, Smith, and Staudinger, 1992; Dittmann-Kohli and Baltes, 1990) acknowledges that, although some aspects of cognitive functioning estimated via traditional tests may decline with age, stability and growth also exist, if to a lesser extent. The approach of Baltes and his colleagues also operates within the constructs of fluid and crystallized cognition, although a different emphasis is placed on the relative roles and meanings of these two kinds of cognition. Here, both aspects of cognition are considered as coequals in defining the developmental course of cognition. In general, Baltes argues that crystallized cognition has been too narrowly defined, and that its importance increases as one moves into adulthood and old age. In this sense, it may be inappropriate to associate a decrease in fluid cognition with an average decline in cognitive competence. Baltes and his associates see adult cognitive competence in terms of a dual-process model. The first process, called the mechanics of cognition, is concerned with developmental change in basic information processing that is genetically driven and assumed to be knowledge-free. With aging, there is a biologically-based reduction in reserve capacity (Baltes, 1987; Baltes et al., 1992). The second process, pragmatic cognition, relates the basic cognitive skills and resources of the first process to everyday cognitive performance and adaptation. Measures of pragmatic cognition within select domains are viewed as tapping skills more characteristic of adult cognitive life than are traditional psychometric measures of cognitive skills. Similar to empirical findings on the distinction between fluid and crystallized cognition, Baltes, Sowarka, and Kliegl (1989) showed that the mechanics of cognition tend to decline with age almost linearly, whereas the pragmatics of cognition tend to maintain relative stability throughout adulthood. For example, whereas linear declines were found in the speed of comparing information in short-term memory (i.e., aspects of cognitive mechanics), no age differences were registered for measures of reasoning about life planning (i.e., aspects of cognitive pragmatics). Cognitive skills are assumed to operate on content domains involving factual and procedural knowledge; they are regulated by higher-level, trans-situational, procedural skills and by higher-order reflective thinking (metacognition), all of which define the "action space" in which problem solving occurs within a given individual. According to this approach, successful aging entails limiting one's tasks and avoiding excessive demands. Baltes and Baltes (1990) use the concept of selection to refer to a self-imposed restriction in one's life to fewer domains of functioning as a means to adapt to age-related losses. It is assumed that by concentrating upon high-priority domains and devising new operational strategies, individuals can optimize their general reserves (Baltes, 1993). By relating adult cognition to successful cognitive performance in one's environment, this position acknowledges that not all tasks are equally relevant for measuring cognition at different ages (Baltes et al., 1984; Baltes et al., 1992).