Retour vers liste

Détail de la contribution

Auteur: Jessica RETT

Co-Auteur(s): Hyams, Nina Winans, Lauren

The effects of syntax on the acquisition of evidentiality

Abstract/Résumé: Several recent studies have examined the acquisition of evidentiality as it is encoded in morphemes like aspect markers (Ozturk & Papagragou 2008, de Villiers, et al. 2009, Aksu-Koç, et al. in press), tense markers (Papafragou et al. 2007) and embedding verbs (Koring & de Mulder 2011). Because evidentiality is not the sole or primary semantic component of these morphemes, these studies provide only an indirect look at the acquisition of evidentiality. In English, evidentiality is syntactically encoded in copy-raising constructions (CRCs): CRCs with raised subjects (1) can only be felicitously uttered by a speaker with direct evidence for the proposition, while expletive CRCs like (2) are consistent with the speaker having either direct or indirect evidence. 1. John looks like he’s sick. CRC 2. It looks like John is sick. expletive CRC In this paper, we examine the acquisition of evidentiality in English-speaking children ages 2 to 5 by looking at their spontaneous production of CRCs. We show that children as young as 2 behave like adults in their ability to correlate the syntax of these constructions with the type of evidence they encode. We present the results of an adult experiment which provide the standard against which we compare the child data. Our adult experiment (n=90) tested the acceptability of raised and expletive CRCs in different evidential scenarios. Statistical analysis revealed significant main effects of syntax and evidence (p < .001), as well as a significant interaction effect (p < .001). There is a direct correlation between the syntax of these CRCs and the type of evidence available to the speaker: raised CRCs are only compatible with direct evidence, expletive CRCs are compatible with both direct and indirect evidence. These results are consistent with the predictions and analysis in Asudeh & Toivonen (2012). In our acquisition study we searched for all instances of CRCs with the predicates seem/look/sound. Importantly, the children showed an adult-like distribution of CRCs: while expletive CRCs were uttered in both direct (n=20) and indirect evidential contexts (n=15), raised CRCs were uttered exclusively in direct evidence scenarios (n=30). These results contrast in interesting ways with studies of morphologically-encoded evidentiality. They are contrary to the conclusions in Ozturk & Papagragou (2008), Papafragou et al. (2007), and Koring & de Mulder (2011) that suggest a relatively late mastery of indirect evidentials, possibly for TOM-related reasons. Additionally, this evidence of the early use of evidentials does not support the (neo-Whorfian) language>concept idea (Aksu-Koç, et al., in press), viz. early attention to evidential source does not depend on having a language with obligatory, morphologically encoded evidentiality.