Variation-Based Distance and Similarity Modeling: A Case Study in World Englishes.

Area of Science:

• Sociolinguistics
• Quantitative Dialectometry
• Corpus Linguistics
• Computational Linguistics

Background:

• Comparative sociolinguistics and quantitative dialectometry provide foundational insights.
• Existing methods may not fully capture nuanced similarities between language varieties.
• Understanding dialectal variation is crucial for linguistic analysis.

Purpose of the Study:

• To develop and present a corpus-based method, Variation-Based Distance & Similarity Modeling (VADIS), for quantifying dialect similarity.
• To rigorously measure the correspondence in language users' choices between linguistic variants.
• To investigate syntactic alternations across international English varieties.

Main Methods:

• Employed a corpus-based approach, Variation-Based Distance & Similarity Modeling (VADIS).
• Analyzed three syntactic alternations in nine international varieties of English.
• Quantified similarity based on the correspondence of variant choices among language users.

Main Results:

• Probabilistic grammars demonstrated remarkable stability and similarity across the studied English varieties.
• A distinct clustering of 'native' (Inner Circle) English varieties was observed.
• 'Non-native' (Outer Circle) varieties, such as Indian English, exhibited greater heterogeneity.

Conclusions:

• VADIS offers a rigorous quantitative method for assessing similarity between language varieties.
• Linguistic structures show significant cross-variety stability, particularly in native varieties.
• The heterogeneity of non-native varieties suggests diverse pathways of language development.