Heritage Language Acquisition under Feature Reassembly: The development of Spanish viewpoint aspect morphology by heritage speakers in the United Kingdom

Abstract

This thesis examines the suitability of the Feature Reassembly Hypothesis (Lardiere, 2009) as applied to heritage language acquisition by Putnam and Sánchez (2013) for understanding heritage speakers’ grammatical knowledge using the specific case of the development of Spanish viewpoint aspect morphology by heritage speakers in the United Kingdom. Spanish heritage speakers in the United Kingdom have not been widely studied (Guardado, 2018), but heritage speakers present a unique challenge to generative theories of linguistics because the outcome of the first language acquisition process often results in grammars that diverge substantially from the input (Lohdnal et al., 2019). Under a feature-based account, the continuous and habitual readings of the Imperfect would be expected to be more vulnerable for heritage speakers due to the cross-linguistic differences in expression of these meanings between English and Spanish, whereas the progressive reading of the Imperfect and the association between perfective and Preterite morphology would be less vulnerable (Arche, 2014; Domínguez, Arche & Myles, 2011, 2017). The specific research questions addressed in the thesis pertain to how viewpoint aspect is expressed in Spanish by both the heritage speakers and baseline speakers, how any observed divergence from the baseline manifests among the heritage speakers, and whether input quantity or language activation measures are better predictors of heritage speakers’ grammatical outcomes. This thesis constitutes the first investigation of feature reassembly or viewpoint aspect in Spanish heritage speakers in the United Kingdom. It is also the first study to examine Arche’s (2014) derivation of viewpoint aspectual interpretation in heritage speakers. In order to answer these questions, the Cat Story narrative retelling task, and an adapted version of the semantic interpretation from the Spanish Learner Language Oral Corpus project (SPLLOC; Domínguez et al., 2013) were deployed to assess the heritage and baseline speakers’ use and knowledge of viewpoint aspect distinctions. Additionally, the heritage speakers’ speech rates in Spanish and English were calculated using an additional narrative retelling task, the Loch Ness story, which was also used in the SPLLOC project (Mitchell et al., 2008), and the heritage speakers’ current and lifespan use of and exposure to English, Spanish and additional languages was calculated using Unsworth’s (e.g. 2013) Bilingual Language Experience Calculator (BiLEC), yielding measures of input quantity, input quality, and language activation. Twenty-five child heritage speakers of Spanish between the ages of five and fifteen participated in the project, as did sixteen of their parents. The baseline varieties of the parents are Colombian Spanish, Mexican Spanish, two distinct varieties of Argentinian Spanish, and various regional varieties of Peninsular Spanish. The Peninsular varieties include speakers from the Communities of Andalucía, Madrid and Valencia, but crucially none from Galicia or the Canary Islands. The participants’ usage rates of all verb forms in the narrative retelling task, accuracy rates in the semantic interpretation task, speech rate, and BiLEC measures were analysed using non-parametric Wilcoxon rank sum and signed rank tests, as well as Spearman’s rank correlations. The results show that whilst individual heritage speakers converge on the baseline grammar, a number of heritage speakers do not and that as a group the heritage speakers often differ from the baseline in specific ways. One of the main divergences observed was the use of the Present in lieu of the Imperfect in habitual and continuous contexts, which was associated with Spanish speech rate and relative output in Spanish, two measures of access to Spanish. Since these contexts are the ones in which greater vulnerability is expected under feature reassembly, and this association is predicted under Putnam and Sánchez’s model, this manifestation of divergence provides substantial support for their model. However, another way in which the heritage speakers diverged from the baseline was observed for heritage speakers for whom the baseline variety allows the extension of the Present Perfect to express a restricted sub-set of perfective meaning. Several of these individuals used the Present Perfect to express perfective meaning in the Cat Story task, whereas the baseline speakers consistently used the Preterite for these contexts. The source of this divergence was internal to the properties of the Spanish baseline, therefore this divergence is not a result of feature reassembly but pertains to earlier in the acquisition process. Even so, the use of the Present Perfect or Preterite within this group of heritage speakers was strongly associated with the language access measures. An implication of this finding is that Putnam and Sánchez’s model can be extended from only explaining divergence from the baseline grammar as a result of feature reassembly, to explaining divergence that obtains in the initial stages of feature selection and feature assembly as well.

More information

Identifiers

ORCID for James Corbet: orcid.org/0000-0003-4660-9551

ORCID for Laura Dominguez: orcid.org/0000-0003-2701-2469

ORCID for Jaine Beswick: orcid.org/0000-0002-1866-939X

Catalogue record

Export record