Kendler and Kendler (1962) are remembered for their bold challenge to the behaviorist tenet, widely accepted in the middle of the 20th century, that the learning process functions in an identical manner across species and across the human life cycle. To the contrary, they argued, learning develops. Young children learn via associationist mechanisms, but by age 7, the learning process has been transformed into one involving internal mediating concepts that connect overt stimuli and responses.

Today it is apparent that the Kendlers overstated their case in claiming that young children do not form concepts. There is ample evidence to the contrary, and a different explanation for age differences on Kendlers' learning tasks is required. Following the Kendlers' work, the question of developmental changes in the learning process was largely put aside, as interest in children's learning declined in general. It is now, however, a somewhat different question than it was in the Kendlers' day. Rather than formation of S-R bonds, learning is now more likely to be defined as "change in understanding" (Schoenfeld, 1999). In this context, it is productive to resurrect the question largely abandoned following the Kendlers' work.

A study is described of 11-12-year-old children and young adults engaged in an identical learning task. Results support the proposal that learning comes to operate under increasing executive control in the years between middle childhood and early adulthood. Adults, it is suggested, are more likely to employ a meta-level executive allowing them to maintain dual representations, one of their existing understanding and the other of new information they are being asked to register. The executive allows these two representations to be maintained simultaneously and attended to flexibly. Without it, there exists only a singular experience of "the way things are" as a vehicle for understanding the world. New information may be assimilated to it, but no executive is available to monitor and manage the process.