📄 Vernacular universals and language contacts: Evidence from varieties of English and beyond

Summary

This item is a book review, not an empirical study. Walcir Cardoso evaluates an edited volume by Filppula, Klemola and Paulasto (Routledge, 2009), abbreviated VULC, which originated in a 2006 World Englishes symposium hosted by the University of Joensuu at its Mekrijärvi research station in Finland. The book asks two questions: how contact-induced change relates to vernacular universals (VUs), meaning features that recur across non-standard varieties, and whether the two can be told apart.

Its fifteen chapters move from theory to case studies to methodological comparison. The contributors fall into three groups:

• Some attribute shared non-standard features to universality (Chambers, Schreier, Tagliamonte).
• Others stress cross-dialectal influence (Britain & Fox, Mufwene).
• A third group doubts that the dichotomy holds at all (Trudgill, Thomason).

Cardoso concludes that most chapters favour cross-dialectal influence combined with social and learnability factors. He singles out Thomason's diagnostic criteria for attributing a change to contact. He criticises the bias toward morphosyntax, the English-only scope and the messy references, but still recommends the book as the entry point to the topic. No dataset, model or treebank is involved.

Key points

• The review appeared in Canadian Journal of Applied Linguistics 12 (2009). The book has vi + 385 pages: an introduction plus fifteen chapters in four parts.
• Only two chapters take a phonological perspective (Schreier; Britain & Fox). The rest focus mainly on morphosyntax.
• Thomason (chapter 15) offers four criteria for estimating whether contact, rather than VUs, shaped a given change. She also argues that overlapping causes (VUs, cross-dialectal influence, ease of learning) make a strict split hard to establish empirically.
• Other factors raised across chapters include high vs. low contact language status (Trudgill), productive and perceptual ease, gender and age.
• The criticisms are that other languages appear only as illustrations (e.g., Thomason, Odlin), and that the per-chapter reference lists overlap heavily (Chambers 2004 appears almost everywhere) and are formatted inconsistently. The reviewer suggests one unified bibliography.

Interesting ideas

• I1. An observed feature may come from several overlapping causes at once, so forcing a single explanation can be misleading.
• I2. Before claiming that a change has a particular source, apply an explicit checklist of evidential criteria, as in Thomason's chapter.
• I3. Features that recur across many non-standard varieties are systematic rather than random noise.
• I4. Check whether evidence is spread evenly across levels of grammar before generalising, since conclusions drawn mostly from morphosyntax may not carry over to phonology.
• I5. Merging scattered, duplicated reference lists into a single uniform store removes overlap and formatting drift.

For Mova

How these ideas could serve Mova, our small open language model — its files, stages and tests:

• I1 (learn): when diagnosing disagreements between witness parsers or deriving repair rules, allow several co-occurring causes (dialect, tokenization, true ambiguity) instead of assigning one label per error.
• I2 (tree): during UD dialect conversion, require explicit tests before deciding whether a parse divergence reflects an annotation convention or a genuine parser mistake.
• I3 (tokens): let the tagger and lemmatizer treat recurring vernacular patterns as regular variants with their own analyses, rather than as errors to normalise away.
• I4 (answer): audit test suites for balanced coverage across phenomena so that a high score in one area is not read as general competence.
• I5 (world): compile knowledge from WordNet, FrameNet and Wikidata into one deduplicated, uniformly formatted layer instead of keeping per-source copies.