Creating New Language Materials for an Online Pronunciation Training Platform.

This article presents CALST (Computer-Assisted Listening and Speaking Tutor), a multilingual online platform designed for beginner-level foreign and second language (L2) pronunciation training. CALST supports multiple target languages—including Norwegian, English, Greek, Italian, Spanish, and Catalan—and tailors exercises to learners’ native languages (L1) by using phonetic-phonological data from L1-L2map to identify new or unfamiliar sounds, consonant clusters, and prosodic features. The platform offers four exercise types focusing on segmental and prosodic pronunciation, with content development guided by detailed analyses of phonemic inventories, allophonic variations, and letter-to-sound correspondences to ensure relevance across diverse L1-L2 pairs. While CALST emphasizes intelligibility and comprehensibility over native-like accent, it allows learners to pursue near-native pronunciation if desired, and its scalable design invites language experts to contribute new language content without requiring technical expertise.