Dataset abstract

The dataset includes annotated spoken corpus data of N = 3685 utterances, created with a picture description task that elicited semi-spontaneous oral production data from N = 5 learners of Hindi at a Belgian institute. Following a longitudinal design, we gathered data during four observations. These found place in the second (Time1), third (Time2; Time3) and fourth semester (time4) of their Hindi course trajectory. The corpus data is annotated for (i) Learner, (ii) Time of elicitation, (iii) the use of -ko as a Differential Object Marker and (iv) -ko as another marker, as well as multiple features associated with ko-marking, including: (v) Specificity of the Direct Object, (vi) Animacy of the Direct Object, (vii) the (lemmatized) sentence verb, (viii) the (lemmatized) head of the Direct Object Noun Phrase, (ix) the (lemmatized) head of the Noun Phrase associated with other uses of -ko, and (x) the semantic role of these other uses of the ko-marker.

Article abstract

This article investigates the acquisition of Differential Object Marking (DOM) in Hindi as a foreign language based on a longitudinal study with five L1 Dutch speakers. A first aim of the study is to verify findings in previous cross-sectional studies. DOM in Hindi is governed by animacy and specificity and is particularly hard to acquire, with heritage speakers and advanced L2 learners omitting the direct object (DO) marker. A second aim is to investigate whether we can find evidence of either L1 influence, item-based learning or semantic mapping, testing explanations by Montrul et al. (2012, 2015, 2019), Narasimhan (2005, 2020) and Baten and Ponnet (2023). Data collected over four observation points confirmed high DOM omissions. Using a mixed effects logistic regression analysis, we found an increase in marking for human animate, specific DOs, and also for non-human animate and inanimate specific nouns. Our findings suggest that learners initially expand the optionality of -ko due to the complex syntactic-semantic constraints (possibly enhanced by L1 transfer) before gradually adapting to its constraints (possibly emerging through item-based learning).

MS Excel, 16.16.27

ELAN, 6.4

R, 4.2.2

RStudio, 2022.12.0