Participation in imagined worlds is a hallmark of the human species, and yet we know little about the context of its early emergence. The experiments reported here replicated and extended in 2 directions Lillard and Witherington's (2004) study of how mothers pretend to have snacks, across different ages of children (15-to 24-month-olds, Experiment 1) and to a different scenario (personal grooming, Experiment 2). Mothers' pretend behaviors changed little as infants aged, but there were some scenario differences. Most striking in this research was the consistency with which particular maternal pretend behaviors were associated with children engaging in pretense behaviors and smiling. The findings are discussed with reference to the child's emerging skills in joint attention and social referencing.Humans' ability to participate in worlds beyond the here and now is a hallmark of the species, foundational to planning, invention, art, and much else (Walton, 1990). Pretend play is one of the earliest clear instances of such participation, yet we know little about the context of its emergence. In an earlier study, Lillard and Witherington (2004) examined mothers' behaviors when pretending to have a snack of imaginary Cheerios and juice with their 18-month-olds versus when really having such snacks. Many behavioral changes across pretend and real scenarios were discovered, and several were associated with people ages 4 to adult correctly identifying pretense acts as pretense (Richert & Lillard, 2004) and with 18-month-olds' smiling at and participating in pretense acts, potential indicators of understanding (Lillard, 2004).The experiments reported here extend that work in two directions. The first experiment addresses whether mothers vary their pretend behaviors as children become more competent Copyright © 2007 pretenders, from 15 to 24 months of age, and whether different maternal behaviors are associated with children appearing to understand pretending across this age range. The second experiment examined whether the findings replicate in a different pretend context, personal grooming, to know if the pretend "signs" mark all pretense or only pretend snacking. The results of the earlier Lillard and Witherington (2004) work are reviewed first.Past literature suggested smiling as a possible indicator of pretense (Kavanaugh & Engel, 1998;Piaget, 1962;Wellman & Hickling, 1993). Children and adults use smiling as an indicator of pretend fighting (Boulton, 1993), and nonhuman primates apparently indicate play fighting with the "play face" (Eibl-Eibesfeldt, 1989;van Hooff, 1972). Lillard and Witherington (2004) found that when mothers were pretending with 18-month-olds, they did smile more than when really having snacks, and their smiles lasted longer on average. In addition, Lillard and Witherington found that pretend smiles were frequently judged to be "about" the mothers' own actions, namely because they were placed just after those actions. A child could plausibly reference the mother's pretend smile as an indicat...