What is it about?

English is a weight-sensitive language, which means heavy syllables are more likely to attract word stress. At the same time, English words tend to be short, and stress in nouns and adjectives typically skips the final syllable. As a result, most common words in the language have initial stress. Not only is such a positional bias strong in English, but it's also prosodically salient, and easier to compute relative to weight-sensitivity. In this paper, I argue that the positional bias in English can be sufficiently strong to conceal weight-sensitivity in the language from second language learners—even when these learners can transfer weight effects from their native languages (Mandarin and Portuguese).

Featured Image

Why is it important?

This paper shows that confounding factors can hinder second language acquisition even when learners are seemingly target-like. In other words, this is case where we can objectively predict certain limits to language language acquisition. In addition, given the English lexicon, the paper suggests that it may take learners much longer than expected to fully acquire the stress patterns in English. Methodologically, the paper employs Bayesian data analysis to simulate native grammars. A comparison is then provided where a naïve initial state is pitched against positional and weight-based initial states. As a result, the paper also provides a statistical analysis that incorporates key theoretical assumptions in the field of second language research.


My main goals writing this paper were (a) to show that we can objectively determine certain limits to second language acquisition on the basis of lexical patterns, and (b) to demonstrate how a Bayesian data analysis can be more comprehensive, meaningful, and realistic in the field of second language research.

Guilherme Duarte Garcia
Ball State University

Read the Original

This page is a summary of: Language transfer and positional bias in English stress, Second language Research, December 2019, SAGE Publications, DOI: 10.1177/0267658319882457.
You can read the full text:



The following have contributed to this page