Labels in infants' object categorization: Facilitative, or merely non-disruptive?

Chan, K. C., Westermann, G. (2018). Labels in infants' object categorization: Facilitative, or merely non-disruptive? Poster presented at the 21st Biennial International Conference on Infant Studies, Philadelphia, USA. 


Previous research has claimed that both labels and novel communicative signals (such as beeps used in a communicative context) can facilitate object categorization in 6-month- olds, and that such effects can be attributed to their communicative nature (e.g., Balaban & Waxman, 1995; Ferguson & Waxman, 2016). Yet, other research has shown that young infants can form categories simply based on visual information in the absence of any auditory signals (e.g., Behl-Chadha, 1996; Quinn, Eimas & Rosenkrantz, 1993). Further, the auditory overshadowing hypothesis suggests that auditory input disrupts concurrent visual processing (and thus, categorization) in infants, with more familiar auditory signals interfering less (Sloutsky & Robinson, 2008). Therefore, two questions arise: (1) do communicative auditory signals truly facilitate categorization in 6-month-olds when compared to silence; and (2) can the comparable influence of labels and novel communicative signals be explained by familiarity with these auditory stimuli? The present study addressed these questions by familiarizing 6-month-old infants (N=28) on a sequence of animals (dinosaurs or fish) in silence and in the presence of pre-familiarized tone sequences. These stimuli have been used in previous studies to argue that labels and tones used in a communicative sense during familiarization enable infants to form categories (Ferguson & Waxman, 2016). Nevertheless, we found that infants in both conditions looked longer at an out-of-category object during test (silent: t(11) = 3.25, p = .005, d = 1.02; familiar tones: t(12) = 2.31, p = .040, d = 0.64; Figure 1), indicating successful categorization even in the absence of labels or communicative tones, and in the presence of familiar tones. These results are consistent with the auditory overshadowing hypothesis in showing that infants can form categories in silence, and that labels do not facilitate, but merely do not disrupt, categorization. Furthermore, our results also suggest that the effect of communicative tones is one of familiarity rather than communicative function. The findings in the present study warrant a reinterpretation of the results of previous studies using the same stimulus set, such that labels and familiar auditory signals do not facilitate, but, due to familiarity, also do not disrupt category formation in infants, whereas less familiar auditory signals disrupt categorization. An important implication of the findings of the present study is that it is vital to include appropriate control conditions in studies of linguistic effects on object categorization.