Close Encounters of the Word Kind: Attested Distributional Information Boosts Statistical Learning

Statistical learning, the ability to extract regularities from input (e.g., in language), is likely supported by learners’ prior expectations about how component units co-occur. In this study, we investigated adults’ experience with sublexical their native language influences performance on an empirical learning task. Forty German-speaking adults completed a speech repetition task which they repeated eight-syllable sequences two experimental languages: one containing disyllabic words comprised of frequently occurring German syllable transitions (naturalistic words) and other made unattested (non-naturalistic words). The participants demonstrated both naturalistic non-naturalistic stimuli. However, was superior for sequences, indicating that had used existing distributional knowledge faster more accurately than words. This finding supports theories statistical as form chunking, whereby co-occurring become entrenched long-term memory.