Abstract
Communication is a core human capacity that enables the exchange of ideas, emotions, and experiences. Its success relies on the ability to perceive, interpret, and produce linguistically meaningful sounds.Second language (L2) acquisition is a multifaceted process shaped by cognitive, perceptual, and social variables. A persistent challenge for L2 learners involves the recognition and production of sounds that, although present in the first language (L1), differ in their phonetic realization and/or functional distribution. These difficulties are often exacerbated by limited exposure to the target language and the absence of explicit, informative feedback.One approach that has gained prominence in educational contexts---particularly with the integration of technology---is High Variability Phonetic Training (HVPT). Current research suggests that this method can be highly effective. Training learners with a broader array of pronunciations and diversifying the language samples they encounter helps promote a more generalized and adaptable categorization of sounds. This wider spectrum of input has been shown to support the transfer of newly acquired skills across different contexts, ultimately enhancing overall communicative competence.The present dissertation first assesses the effectiveness of HVPT over the more traditional Low Variability Phonetic Training (LVPT) in the L2 acquisition of Spanish stop consonants.It then examines how the formation of new phonetic categories contributes to the realignment or shift of perceptual boundaries in the perception of stop consonants. Lastly, this work investigates how the development of novel phonetic categories can influence the production of stop consonants in both the L2 and the L1. The results indicate that phonetic training improved participants�� ability to identify and categorize Spanish stop voicing. While both the HVPT and LVPT groups showed perceptual gains, no significant differences emerged between the two in the perception task. In contrast, production outcomes revealed that HVPT resulted in more target-like realizations of Spanish stops than LVPT, although both training types proved effective.Notably, training also impacted L1 production: English voiceless stops exhibited reduced aspiration, indicating a subtle but measurable shift toward Spanish-like VOT values��evidence of phonetic drift.Taken together, these findings support the effectiveness of phonetic training in promoting the formation of novel phonetic categories and provide insights for models of L2 phonological acquisition and the perception�Cproduction relationship in second language learning.