**Background**
Bilinguals often outperform monolinguals in nonverbal tasks of cognitive control. For example, a meta-analysis of studies that compared the performance of bilinguals and monolinguals on conflict resolution tasks (e.g., the Stroop task) revealed a moderately significant bilingual advantage (Donnelly, Brooks, & Homer, 2015). At first blush, this finding augments our understanding of far transfer effects (how practice in one domain [language] results in changes to other [non-language] domains) and may inform educational policies and social practice. However, close inspection of the empirical data yield inconclusive results. For example, the effect reported in the above meta-analysis appears to be driven by data from a single study. Furthermore, the meta-analysis also revealed a large and significant main effect of ‘lab’ (research group). The inconsistent results, combined with publication bias (de Bruin, Treccani, & Della Sala, 2015), has led many scientists (e.g., Duñabeitia & Carreiras, 2015; Klein, 2016; Paap, Johnson, & Sawi, 2016) to question whether the bilingual advantage is real or merely an artefact of particular research practices. Moreover, significant bilingual advantages have not been observed in studies with large sample sizes (see Paap, Johnson, & Sawi, 2016, for discussion). To progress beyond the controversy and advance the science, in addition to carrying out more studies, we must come up with a theory that can account for the inconsistencies in the literature and explain when, how, and why learning two or more languages improve cognitive control. This is the focus of the current paper.
**Predictions**
We hypothesise that exposure to more varied language environments drive infants to explore (sample) further by constructing less detailed models of their environments and placing more weight on novel information. Getting by on less detailed models would allow the child to switch faster to novel stimuli and thus sample more from their environments. We will test our hypothesis by running a series of experiments that will seek to replicate with a larger sample—and thus support—Kovacs and Mehler’s (2009) finding of a bilingual advantage in infants (Experiment 1) and probe whether bilingual infants build, and get by on, less detailed representations of visual stimuli (Experiment 2), shift attention faster to a second visual stimulus (Experiment 3), and are (thus) less sensitive to the minute detail of a visual stimulus (Experiment 4).
Specifically, we predict that bilingual infants will: (a) be better at inhibiting a learned behaviour (Experiment 1); (b) respond more appropriately to more fragmented—or less detailed—visual stimuli (Experiment 2); (c) be more likely to abandon the visual processing of a stimulus to shift attention to a novel stimulus (Experiment 3); and (d) switch more frequently between two visual stimuli, spend less time visually processing a familiar stimulus, and thus be worse at remembering the details of a visual stimulus (Experiment 4). Although these infant studies can neither support nor disprove the theory that a bilingual advantage results from managing two or more languages during production in older children and adults, if we observe a significant group difference we will argue that it reflects experience-dependent adaptations that occur because of regular exposure to two or more languages. If we fail to observe a significant group difference, we will argue that there is currently insufficient evidence to conclude there is a bilingual advantage in preverbal infants.
**Please see our manuscript (**RSOS RR1 - DSOUZA ET AL.docx**) for more detailed information.**
