Investigating the impact of sample size on cognate detection

Abstract

In historical linguistics, the problem of cognate detection is traditionally approached within the frame-work of the comparative method. Since the method is usually carried out manually, it is very flexible regarding its input parameters. However, while the number of languages and the selection of comparanda is not important for the successfull application of the method, the sample size of the comparanda is. In order to shed light on the impact of sample size on cognate detection, an automatic method that closely mimics the comparative method is applied to a large dataset of varying sample sizes. The results clearly show that an increase in sample size is also accompanied by an increase in accurracy of the automatic method.