Escudero, P. 1 , Mulak, K. 1 & Vlach, H. 2
1 MARCS Institute, University of Western Sydney
2 Department of Educational Psychology, University of Wisconsin, Madison
Word learning studies have typically focused on participants' ability to learn novel word-referent pairings, presented unambiguously, in a single trial. However, in the real world, words are encountered amidst myriad possible referents. Recently, a cross-situational learning paradigm has been used to better model the ambiguity in real-world situations. In a typical experiment, participants are presented with multiple visual referents on a screen as the corresponding spoken words are presented in random order. To learn word-object pairings, learners need to track co-occurrence probabilities of words and referents across trials. Previous studies on cross-situational learning have involved words that are phonologically very distinct (e.g., 'regli', 'colat'). In the current experiment, we examined whether 22 17-month-olds could learn monosyllabic words that form phonologically minimal pairs, differing in either a consonant ('bon'-'don') or vowel ('deet'-'doot'), near-minimal pairs ('don'-'deet'), and non-minimal pairs ('bon'-'deet'). Infants were presented with a typical cross-situational learning paradigm, with two words and two referents in each trial. During testing, participants again saw pairs of visual referents but heard the spoken word for only one of the images (a preferential looking paradigm). Using a Tobii X120 eye-tracker, the looking time to the named image was recorded. We found a main effect of pair type, F(3, 63) = 2.968, p = .039, Partial Eta Squared = .124. One-sample t-tests revealed participants' looking to the named image was above chance only for vowel minimal pairs, t(21) = 2.197, p = .039. This finding runs contrary to the results of our adult study with the same stimuli and task, where adult participants successfully learned all word pairs and mean accuracy was lowest for vowel minimal pairs. Implications of our findings for the role of vowels and consonants in early word learning, as well as why we observe differences in performance across development, will be discussed.