The structure of executive functions in preschoolers is controversial. Miyake and colleagues found that, in adults, inhibition, shifting, and updating are correlated but distinguishable processes; this finding was sometimes replicated with schoolchildren. Based on schoolchildren data, Im-Bolter, Johnson, and Pascual-Leone proposed a four-component model, with mental attentional capacity and inhibition as basic resources, and shifting and updating as executive processes that partly rely on those resources. Studies on preschoolers propose either a single factor or two factors. We tested 118 children (36 to 73 months old) using measures of mental attentional capacity (Mr. Cucumber, Direction Following Task, and Backward Word Span), inhibition (Bear/Dragon and Day-Night Stroop), shifting (Dimensional Change Card Sort) and updating (Magic House). Confirmatory factor analyzes showed that the best-fitting model comprises two highly correlated factors -- one loading the mental attentional capacity measures, the other loading inhibition, shifting, and updating measures. Performance in inhibition, shifting, and updating tasks showed qualitative changes with increasing attentional capacity. In preschoolers, shifting and updating seem to emerge tied to inhibition -- but only when sufficient capacity enables children to perform executive control tasks. This seems consistent with Im-Bolter's et al.'s proposal of two basic resources.