2018 Sat Poster 6447

Saturday, November 3, 2018 | Poster Session II, Metcalf Small | 3:15pm

Early comprehension of canonical and non-canonical word orders in Mandarin
J. Zhu, A. Gavarró

Previous studies show that as early as 19 months infants have an abstract representation of the canonical word order whether the target language is VO or OV (Franck et al. 2013 for French; Gavarró et al. 2015 for Hindi-Urdu). However, Candan et al. (2012) found some variation in early acquisition: English children were able to parse canonical sentences before their second birthday, earlier than Mandarin children; the authors attributed the contrast to the more rigid word order of English in the input, that would boost acquisition. Here we pursue this investigation with a comprehension experiment with Mandarin infants.

We tested 24 typically-developing Mandarin infants with a mean age of 17;5 months (SD = 2.2) both on canonical SVO sentences (1a) and non-canonical sentences involving the topic ba construction (1b,c). The conditions in our experiment included only pseudo-verbs (illustarated in (1)). The infants were presented with two simultaneous videos, one representing the target event, the other the same event with theta-role reversal. Following the method in Franck et al. (2013), the recording of gazing time took place in four windows: the baseline and three consecutive exposures to the target sentence. The results (see Table 1 and Fig. 1) showed that, in both the SVO and the SbaOV condition, infants preferred looking at the scene with the first NP being the Agent, the same as adults. The proportion of looking time to the target video showed above chance performance. No significant difference was found in the baseline window. However, in the OSbaOV condition, they looked longer at the scene with the first NP as Patient during the first (t(23) = 3.35, p = .003, d = .65) and the second presentation (t(23) = 2.08, p = .049, d = .57) of the test sentence. The proportion of looking times to the scene with the first NP as Agent differs significantly between the SO and OS conditions during the first (Wilks’ Lambda = .709, F2,22 = 4.52, p = .023, η2 = .291) and the second presentation (Wilks’ Lambda = .535, F2,22 = 9.57, p = .001, η2 = .465) and differences between SVO and OSV, and SOV and OSV were both significant (p = .001 and p = .031, respectively), while the difference between SVO and SOV was not statistically significant.

From these we conclude that Mandarin-speaking infants were interpreting as adults and thus sensitive to the target word order from age 17;5 at the latest. This contrasts with the results from Candan et al. (2012), who only found evidence for comprehension at almost age 3. The result also questions the hypothesis that more rigid word orders are more easily acquirable, as they argued for English. On the other hand, the results are consistent with the findings of Lassotta et al. (2014), who found comprehension of non-canonical word order in French 21-month-olds. Additionally, if we consider the presence of the ba construction in child-directed speech, we find it to be scarce, with the implications this has for frequency-based accounts of language acquisition.