Quantifying Dialect Relatedness in Serawai (Bengkulu, Indonesia): A 200-Item Lexicostatistical Study

Abstract

This study quantifies the internal relatedness of the Serawai language in Bengkulu Province by comparing two named varieties Padang Capo (“o”) and Puding (“au”) and assessing whether they constitute dialects of a single language. Using a qualitative–quantitative lexicostatistical design, we elicited a complete 200-item Swadesh list from adult native speakers at both sites, normalised tokens, and applied conservative, rule-governed cognacy coding supplemented by analyst memos and a second-reader check; we also summarised recurrent phonological correspondences to interpret the numeric signal. Of 200 aligned items, 124 were cognate, yielding 62% lexical relatedness, with differences concentrated in limited, patterned vocalic and segmental alternations characteristic of dialectal separation. Interpreted against commonly used lexical-similarity bands and a standard basic-vocabulary retention model, the evidence situates the varieties’ most recent common stage at roughly the last 1.0–1.3 millennia, a heuristic window consistent with high mutual intelligibility. The findings support classifying “o” and “au” as close dialects within one Serawai language and, methodologically, convert impressionistic labels into a replicable baseline (complete elicitation, explicit coding rules, qualitative correspondence notes, and uncertainty-aware reporting). Implications include treating Serawai as a single language for educational materials, orthographic guidance, and public communication, and using the documented workflow to expand coverage across additional villages and neighbouring Malayic varieties; future work should add formal correspondence tables with instrumental phonetics and integrate sociolinguistic profiling and phylogenetic modelling to refine internal subgrouping.