There are conflicting accounts of how context variability affects childrens word learning. In some instances, toddlers andpreschoolers word learning appears sensitive to context changes (e.g., Goldenberg & Sandhofer, 2013; Vlach & Sandhofer,2011). In other cases, however, children show learning independent of context variability (e.g., Akhtar, 2005). There mayalso be instances where context variability promotes label retention (Twomey, Ma, & Westermann, 2017). Inconsistentfindings in this literature could be the result of task demands. Context dependencies may emerge when tasks are moredifficult, because children are unable to suppress irrelevant context features and focus on relevant inputs, which are factorsthat can contribute to the strength of context effects (Smith & Vela, 2001). We explored potential context effects in wordand fact learning using a design intended to reduce task load. Under these conditions, fact learning was affected by contextvariability, but word learning was not.

Asking questions about unknown things involves recognizing knowledge gaps, identifying information sources, and for-mulating appropriate questions. This active involvement propels development by individualizing the learning environment.To characterize active engagement in word learning, we investigated whether preschoolers ask questions about novel vo-cabulary and evaluate definition quality. In Study 1, preschoolers were asked to perform actions following instructionswith novel (transpose) or familiar (switch) verbs. They asked more questions about novel (M = 3.31 out of 9, SD = 3.34)than familiar verbs (M = .17, SD = .44), t(35) = -5.68, p ¡ .001. In Study 2, informative or uninformative definitions wereprovided. Preliminary data suggest that preschoolers only asked questions when faced with uninformative definitions (M= .65 out of 3). When faced with novel words, preschoolers not only elicit questions, but also determine whether theirinformation needs have been met.

We used a between-subjects selective trust paradigm to investigate whether 3-year-olds (N=28) and 5-year-olds (N=28)evaluate the quality of informants definitions for familiar and unfamiliar words. 3-year-olds did not choose the informativedefiner (silly=goofy) over the circular definer (silly=silly) for familiar or unfamiliar words. In contrast, 5-year-olds en-dorsed the informative definer for familiar (M=.71, t(12)=2.38, p=.04) and unfamiliar (M=.82, t(14)=3.41, p=.004) words.Additionally, 5-year-olds in the unfamiliar word condition chose to learn new information from the informative definer,such as asking about novel words (p¡. 001) and novel object functions (p¡.001). The unfamiliar word condition may haveelicited better performance than the familiar word condition because the contrast between the two informants was moreobvious. We are currently investigating whether 3- and 5-year-olds prefer an informant who uses familiar words to definenovel words (meager=small) to one that uses other novel words (meager=paltry).