At least half of the world’s 7,000 languages are endangered today, that is to say, they are set to lose all of their remaining speakers by the end of this century because younger generations are no longer learning these languages. There are countless factors that lead to language endangerment, and no endangerment process is alike another one. This notwithstanding, we can notice three important general features of every language endangerment case. First, linguistic diversity is geographically clustered. We have a number of language hotspots and in these language endangerment is particularly frequent. Second, language endangerment occurs always in speech communities that are dominated by more powerful communities. Third, we find these communities frequently in borderlands. These general features are interrelated. High diversity of language coincides with smaller speech communities which are often the result of a different economic organization. These communities become then ‘assimilated’ or ‘integrated’ in the course of nation building processes. In my talk, I illustrate these mega-trends on the examples of Japan, Korea and China. I will in particular address the language endangerment of Ryukyuan in Japan, of Jeju in Korea and of Yi in the People’s Republic of China. I argue thereby that language endangerment is a ‘good problem’ in the sense that addressing issues of language endangerment inevitably leads to discussing, considering and addressing other problems as well. These other problems involve the relation between powerful and weaker communities, the fostering of interculturaltolerance and solidarity, but also a reconsideration of the relations between the center and periphery.