Preference for Distinct Variants in Learning Sound Correspondences During Dialect Acquisition

Area of Science:

• Phonological acquisition
• Second dialect learning
• Sociolinguistics

Background:

• Sound correspondences (SCs) are learnable phonological patterns in second dialect acquisition.
• SCs can involve similar or distinct variants, but the effect of variant similarity on SC learning is understudied.
• Existing hypotheses suggest either a preference for distinct (salience hypothesis) or similar (learning bias hypothesis) variants.

Purpose of the Study:

• To investigate how sound similarity influences the learning of sound correspondences (SCs) in an artificial language learning experiment.
• To test whether learners prefer similar or distinct variants within SCs.
• To examine the role of mapping complexity in SC learning.

Main Methods:

• An artificial language learning experiment was designed to assess SC learning.
• The degree of similarity between dialect variants was measured using multidimensional phonetic and phonological metrics.
• These metrics were cross-validated with typological evidence.
• Learning of one-to-one, two-to-one, and one-to-two SCs was analyzed.

Main Results:

• No significant effect of variant similarity was observed in the learning of simple one-to-one SCs.
• A preference for the most distinct dialect variant emerged in the learning of SCs with complex mapping structures (two-to-one and one-to-two).
• This preference for distinctness was conditional on the complexity of the mapping structure and the magnitude of the sound distance.

Conclusions:

• The study confirms a preference for distinct variants in SC learning, but only under specific conditions.
• This preference is triggered by complex mapping structures and requires a sufficient degree of acoustic distance between variants.
• Findings contribute to understanding the factors influencing phonological pattern learning in second dialect acquisition.