Sound class model (produced in a semi-geneative manner with help of LingPy) following Dolgopolsky (1964) as used in List (2014).