The Zheng He dilemma: Language identification and automatic indexing

Proper names present special problems not only for theories of language but also for indexing and language identification, whether performed by human or mechanical agents, especially so when the names may also be understood otherwise than as names. The problem is compounded in the case of transliterated text and multilingual publications. If the indexer cannot understand the meaning of the text – whether name, phrase or sentence – a heuristic decision process cannot decide amongst various possibilities.