Temperature (probably) does not shape sonority

The notion that phonological features of languages are shaped by environmental factors, such as temperature, has been a topic of intense debates in linguistic research. However, many previous studies on this topic have been limited by methodological biases and inadequate controls for genealogical and areal dependencies between sampled languages. This study aims to address some of these limitations by replicating a recent study on the relationship between temperature and sonority. We analysed data from ~2,800 languages, using a modified model that incorporates improved statistical controls for spatial and phylogenetic dependencies between languages. Our results show that the previously observed relationship between temperature and sonority is not robust and largely disappears when accounting for genealogical and areal factors. A focal point of our analysis is the large uncertainty around the posterior distributions, making any interpretation of mean values meaningless. By using a fine-grained analysis of phonetic features in the lexical data and improved statistical controls, we provide an additional perspective on the relationship between temperature and sonority, and underscore the need for continued methodological refinement in linguistic typology.