AI and ethnology

In 1956, the Eighth National Congress of the Chinese Communist Party decided that all ethnic groups living in the country should be represented in the National People’s Congress. At that time, ethnologists still classified ethnic groups according to the words they used for common everyday things such as house, sky and bird. But language affinities are in fact imprecise, and the term “ethnicity” itself is fuzzy and problematic. By analysing musical parameters instead, AI can work much more precisely – and is basically neutral. It is also subject to misuse, however, depending on the data fed into it. This section presents the results of the identification of tonal systems in the music of the Uyghurs and the Kachin, which differ significantly from one another.

AI sorts music of different ethnic groups

With this map, the AI sorts music from ethnic groups by detecting timbre differences. Each dot is a piece of music. The colours of the dots mean different ethnic groups. Yellow are the Uyghur people, the Kachin green. The map also considers other ethnic groups from neighbouring Yunnan, such as the Yi (light blue) or Khmer from Cambodia (dark green). Touch a point, listen to the music, and see which ethnic group it belongs to.

Touch the switches above the map to change the map’s background. You can choose visualisations of different timbre parameters by which the music is sorted. Yellow regions denote areas where the respective parameter is essential, while its influence is weak in blue areas.