Mental sets are high-level representations that regulate lower-level selection of action. Despite the relevance of mental sets for coherent and purposeful action, relatively little is known about how they are selected and maintained in the face of internal or external interference. The central thesis of this proposal is that a better understanding of set-level processes can be attained by looking at interactions between mental sets and lower-level representations that constrain action in a bottom-up manner (i.e., locations of stimuli or response keys). Specifically, an automatic process is proposed that binds action-related lower-level representations (e.g., stimulus or response locations) to the mental set that is in control of the action. Bindings between set-level and low-level codes should usually be in the service of stable and coherent goal-directed action. However, such bindings may get in the way of efficient control when the association between low-level and high-level codes is ambiguous or highly variable. In particular, problems should occur when high-level control is impaired, as in some patient groups and, albeit to lesser degree also in the context of normal aging. Interference elicited from bindings between high-level and low-level codes may also lie at the heart of executive-control deficits observed in old adults. Therefore, in a majority of the proposed experiments I suggest to explore the hypothesis that old adults are much more negatively affected by ambiguous mappings between low-level and set-level aspects than young adults. Such a result would point to age-related difficulties with """"""""keeping apart"""""""" high-level representations that share low-level elements. The proposed project promises a bridge between lower-level perceptual/attentional processes and executive processes as well as answers to the important question of age differences in executive control.
