Integrating Sight and Sound: The Role of Proficiency in L2 Multimodal Perceptual Development

Aims and Objectives: This study examines L2 categorization of the Spanish vowel inventory by English-speaking learners of Spanish. It focuses on how input modality and proficiency shape the perceptual weighting of acoustic and visual-articulatory cues. Method: Participants completed a multiple-alternative forced-choice vowel identification task in four conditions: audiovisual, audio-only, visual-only, and mismatch audiovisual. The unimodal conditions tested access to acoustic and visual-articulatory information separately, whereas the audiovisual conditions tested vowel categorization under cue convergence and cue conflict. Data and Analysis: The data were analyzed using Bayesian hierarchical multinomial logistic regression to evaluate the probability of selecting a specific Spanish vowel as a function of the input modality and language proficiency. Findings: Learners were highly accurate when auditory input was available, and adding visual cues did not improve identification in the audiovisual condition. Visual-only responses were less accurate overall, but participants showed systematic choice preferences based on articulatory similarity, particularly in the /e/–/i/ and /o/–/u/ contrasts. Higher proficiency was associated with greater accuracy across conditions. In mismatch trials, learners generally favored the auditory vowel, but higher-proficiency learners showed increased reliance on the visual vowel. Originality: This study extends L2 speech perception research by examining whether English-speaking learners of Spanish can use visible articulatory cues to identify Spanish vowels. Rather than treating visual speech as a general aid to comprehension, it tests when visual vowel information supports categorization, when it is overridden by auditory input, and how proficiency changes learners’ weighting of auditory and visual cues. Significance and Implications: The results show that L2 vowel perception is primarily driven by acoustic-phonetic cues, while increasing proficiency is associated with greater sensitivity to the visual-articulatory information available in vowel contrasts. These findings support a multimodal account of L2 phonological development, in which proficiency, visual salience, and articulatory similarity shape how learners map speech cues onto L2 vowel categories.