Statistical learning of phonotactics by children can be affected by another statistical learning task

Area of Science:

• Child language acquisition
• Speech perception and production
• Statistical learning

Background:

• Children typically show a high-frequency advantage in producing phonotactic sequences.
• This advantage is often demonstrated using statistical learning paradigms.
• Prior research indicates familiarity with a sequence enhances its production accuracy.

Purpose of the Study:

• To investigate the high-frequency advantage in phonotactic sequence production.
• To explore potential exceptions to the high-frequency advantage.
• To examine the influence of prior statistical learning on subsequent learning.

Main Methods:

• Twenty-nine children aged 4-5 participated in two experiments: phonotactic statistical learning and prosodic contour statistical learning.
• The order of the experiments was randomized for different participants.
• Phonotactic sequence production accuracy was measured after familiarization with varying frequencies of sequences.

Main Results:

• Children who completed the phonotactic learning first showed the expected high-frequency advantage.
• Children who completed the phonotactic learning second exhibited an inverse effect, favoring low-frequency sequences.
• This suggests the order of statistical learning tasks can influence sequence production outcomes.

Conclusions:

• The typical high-frequency advantage in phonotactic learning is not universal and can be modulated by prior learning experiences.
• The order of statistical learning, specifically the preceding task (prosodic contours), can impact the high-frequency advantage.
• Further research into multistream statistical learning may illuminate the mechanisms behind this observed exception.