Testing Inferences about Language Contact on Morphosyntax: A Typological Case Study on Alorese–Adang Contact

Abstract

When linguists make inferences about language contact, control data is required for reliable analysis. Historical data or reconstructions are typically used for that purpose. However, historical data is globally mostly unavailable, and reconstructions are laborious if comparing outcomes of language contact in a typological way. We assess a recent typological proposal (Di Garbo & Napoleão de Souza 2023) that argues that using a single language, called Benchmark language, can serve as a viable control for making inferences about contact. To evaluate this approach, we use as a case study a documented contact situation in Eastern Indonesia between the Austronesian language Alorese and the Papuan language Adang. We sample languages related to Alorese and use computational phylogenetic methods for ancestral reconstruction to analyse if using a single Benchmark produces deviating inferences about contact compared to using Bayesian ancestral reconstructions. Additionally, we test for isolation by distance effects by analysing if the probability of contact effects correlates with increasing geographic distance from the contact region. The results suggest that no isolation by distance effects exist in the data and that none of the single Benchmarks significantly deviate from the ancestral reconstructions.