Following Valdés’s definition (2000, 2005), heritage speakers are to some degree bilingual in the majority language of the country they live in and the heritage language. Concerning their linguistic competence, two positions can be distinguished: Incomplete acquisition (e.g. Bennamoun et al. 2013, Montrul 2008,2012,2016) and complete acquisition of a contact variety (e.g. Pires & Rothman 2009, Kupisch & Rothman 2016). Previous studies in the first branch have identified that certain grammatical domains are particularly vulnerable in HS grammars which include also tense-aspect phenomena (Montrul 2002,2009). In an elicitation study on tense/aspect and mood interpretation, Montrul & Perpiñán (2011) even find that HS perform below advanced L2 speakers (L1=English, L2=Spanish), maybe due to different metalinguistic competences. Research suggests that generally a great difference between the L1 and L2 feature configuration may hinder quick acquisition (Comajoan 2014, Salaberry 2008). According to the Feature Reassembly Hypothesis (Lardiere 2009, Hwang & Lardiere 2013), a full acquisition of the target system is generally possible, but the process is considerably slowed down if L1 and L2 differ in how features are organized.
In this talk, we apply the idea of comparing feature configurations to HS (Putnam & Sánchez 2013), using different language combinations: English/Spanish and German/Spanish.
In Spanish, past tenses are marked for (im)perfectivity (Zagona 2007). Following Domínguez et al. (2017), the Imperfect dormía is used for progressive, habitual and continuous events, i.e. bundling three features together, whereas the Preterit durmió is reserved for the perfective context. The English verb system involves a basic aspectual contrast found in the progressive form (Salaberry & Ayoun 2005), whereas German does not mark grammatical aspect at all (Heinold 2015). While both systems differ significantly from Spanish, the feature configuration of English resembles the Spanish one in the fact that it also presents two feature bundles, as opposed to German where all four aspectual notions (perfectivity, habituality, continuity, progressivity) can be expressed with the same underspecified verb forms.
We present data from three groups: (1) 20 HS in the USA, (2) 20 HS in Germany and (3) native Spanish speakers (N=15). Data were collected using oral interviews and an Aural Grammaticality Judgment Task (18 items). Based on statistical analyses, we find evidence for complete acquisition in both groups. In contrast to the performance of L2 speakers reported elsewhere (Diaubalick & Guijarro-Fuentes 2016,2017), we neither find a target-deviant convergence of lexical and grammatical aspect, nor an unexpected concentration of non-grammatical elements. Groups differ slightly in terms of their accuracy rates; Anglophone Spanish HS are almost native-like, German-speaking Spanish HS produce some occasional target-deviations. Generally, the accuracy rates on grammatical aspect were higher in oral production than the ones in the aural grammatical judgement task. None of these results, however, have reached significance.
We conclude that feature differences are less significant than the age of onset for Spanish HLA. Attested differences between the groups will be discussed in the light of the role of input, language activation and language combination, considering also the distinction between explicit and implicit knowledge (Rothman 2008).