Getting a Wikipedia dump in said language and counting word occurrences and then learning the most common 5000 (or however many) words. There will be duplicates

Getting a Wikipedia dump in said language and counting word occurrences and then learning the most common 5000 (or however many) words. There will be duplicates (e.g. different forms of a single words) but who cares.