Fig. 8
From: The Austronesian and the Micronesian Comparative Dictionaries as CLDF datasets
Agreement between numbers of forms per language as stated in the legacy HTML pages and as computed from our CLDF data. The y-axis is scaled logarithmically, so more than two thirds of the languages show no difference (although this might be due to splits and mergers canceling each other out). The most extreme outlier – Maranao, with 88 words less in our data was cross-checked manually, and indeed had the expected amount of duplicates.
