| Empirical research carried out in educational contexts (e.g., Sanz, 2000) as well as laboratory studies (e.g., Nayak, Hansen, Krueger, & McLaughlin, 1990) show that prior language experience facilitates L3 acquisition, and that level of bilingualism is closely related to level of proficiency in a third language (e.g., Sanz, 2007). Within The Latin Project framework, (Sanz, Bowden and Stafford), the present study investigates whether Cummins' Threshold Hypothesis (1979), which proposes that cognitive advantages of child bilingualism emerge only when a specific proficiency level in both languages is attained, can be extended to adult language learners. Furthermore, the study investigates whether external conditions, i.e., amount of explicit information on how the language works provided by the feedback, affect the position of the threshold. Cognitive Capacity measurements (MLAT, a WM sentence span test, and L1 & L2 PSTM tests) are included in an attempt to provide explanations for the results obtained.;The study compares performance by 151 Basic, Intermediate, Advanced and Native-like speakers of L2 Spanish during the initial stages of development of an L3 under two computerized treatments which combine input-based practice with feedback [+EF vs. -EF]. Effects are measured via pre/post/delayed comparisons on accuracy and latency in written interpretation, aural interpretation, grammaticality judgment, and production tasks.;In line with previous studies, the findings show that level of bilingualism is closely related to L3 proficiency level (e.g., Munoz, 2000) and that even partial bilingualism enhances subsequent language learning (e.g., Thomas, 1988), although these benefits appeared under less challenging conditions i.e., when the rules were provided, or when the tasks were less demanding. Also, regardless of proficiency level, explicit feedback enhanced language development, although the effects were severely affected by attrition.;Importantly, as hypothesized, the study has identified the presence of one or two thresholds for the positive effects on cognition of adult bilingualism, with their positions changing depending on external conditions. Under more demanding the conditions, (i.e., -EF and more cognitively demanding tasks), only one threshold emerges and is delimited by near-native proficiency. The appearance of these thresholds is explained with aptitude (as measured by the MLAT) and WM capacity (as measured by a sentence span test), which suggests that for adults learning an L2 in formal contexts, achieving a high level of proficiency results in a cognitive edge. |
