Evidence suggests that non-mathematical knowledge affects the encoding of mathematical word problems, by promoting intuitive representations of problem situations. Finding the optimal (least costly) solving strategy to certain problems thus requires switching from an intuitive representation to an alternative, counterintuitive one. We investigate whether modulating the cognitive load attached to a problem alters the ability of adults and 4th-5th graders to perform such recoding when presented with problems whose optimal solving strategy calls for a counterintuitive representation. In two experiments, we manipulate either extraneous cognitive load with a dual-task paradigm or intrinsic cognitive load by varying numerical values. We predicted that increasing extraneous load would hinder participants’ ability to perform recoding, by burdening the executive resources necessary for cognitive flexibility, whereas increasing intrinsic load would foster recoding, by encouraging participants to switch to a more optimal strategy to reduce the computational cost of solving the problem. Our findings are discussed with regard to these hypotheses, shedding light on the interactions between cognitive load and flexibility in arithmetic reasoning.
