Sound class model (produced in a semi-generative manner with help of LingPy) following the ASJP (Wichmann et al. 2016) as used in List (2014).