Quierro comprar una guitara: Lexical Encoding of the Tap and Trill by L2 Learners of Spanish

The Spanish language has two rhotics, the voiced alveolar tap / / and the voiced alveolar trill /r/. Depending on context, these phones are either in free variation or complementary distribution; the only exception is in intervocalic position where the presence of the tap or trill signals a difference in meaning, e.g. pero /pe o/ ‘but’ vs. perro /pero/ ‘dog’ (Hualde, 2005). American English, on the other hand, has a single rhotic which is realized most often as a voiced alveolar approximant and thus is different from either of the Spanish rhotics (Ladefoged & Johnson, 2010). American English-speaking learners of Spanish therefore must learn to produce both the tap and trill in intervocalic position in order to differentiate between minimal pairs. Previous research on the acquisition of these phonemes, however, has found that even advanced speakers often fail to distinguish between the tap and trill in production (Face, 2006; Rose, 2010b). Initially it may appear possible to attribute this difficulty in production to a difficulty in perception, since Rose (2010a) found that Spanish / / and /r/ were assimilated by naïve American English listeners to the same English category / / in the majority of cases; 96.2% of Spanish /r/ tokens and 57.7% of / / tokens were identified as English / /, making English / / the modal response for both Spanish rhotics. Nevertheless, Rose (2010a) found that learners at all levels were highly accurate at discriminating the / -r/ contrast, with a mean discrimination accuracy between 86.7% and 94.4% across proficiency levels in an AXB task. Even American English speakers without experience with Spanish discriminated between these phonemes with 80.2% accuracy on average. Given that learners’ difficulties with Spanish rhotics do not appear to be due to a problem in perception, previous researchers have cited a difficulty in articulation as an explanation for learners’ failure to distinguish these phonemes in production (Face, 2006; Rose, 2010b). The tap in Spanish “is produced with a single rapid contact of the tip of the tongue against the alveolar ridge,” while the Spanish trill “is produced with several such rapid contacts, generally two or three” (Hualde, 2005, p. 181). Although the tongue tip strikes the alveolar ridge in the production of both the tap and the trill, the articulatory gestures required to realize these two segments are quite distinct. The trill is a more complex segment which requires precise control over the positioning of the articulators and the amount of air flow (Solé, 2002). It is among the last segments acquired by native speakers (Jiménez, 1987), and its complexity has led to substantial dialectal variation in its production since even a small change in any of the articulatory gestures of the trill greatly alters the sound produced (Widdison, 1998). Therefore, an inability to reproduce the articulation of the trill could explain the lack of distinction between the tap and the trill in production by many L2 learners. This is substantiated by the type of