During China’s last imperial dynasty the Manchu language was the official tongue of 450 million Chinese. Today, the number of people speaking Manchu has been whittled down to no more than 100, most of them elderly.
The potential loss of the language that once dominated the mighty Qing era (1644-1911), has prompted Chinese researchers to attempt to save it by using artificial intelligence.
“If we don’t accelerate our pace and increase our efforts to protect it, the Manchu language in its original form will disappear completely when these elderly die,” Wang Di, a researcher at the provincial academy of social sciences in the northeast province of Heilongjiang, wrote in the latest issue of The Border Economy and Culture, a monthly Chinese academic journal.
In a recent study, Wang and her team deployed artificial intelligence to identify and synthesise the sounds of the endangered language.
“We first analysed the morphological structure and syllables of Manchu and established annotation rules. Then, with the help of field work, we built a language bank and studied the phonetic identification of Manchu and how to convert it into international phonetic symbols,” she said.
“Lastly, we applied the hidden Markov model [a possibility-analysing tool] to synthesise short Manchu phrases in the hope to provide some reference for the rescue and preservation of the endangered Manchu language.”
Wang said that dozens of languages in China are dying out and that researchers should be developing something akin to Google’s Endangered Languages Project, which gets users to upload audio, text and video files of languages that are dying out.
Part of the reason for the erosion of Manchu is that the remaining Manchu population of about 10 million has integrated with other ethnic groups, and borrowed lots of foreign expressions to keep up with the times.
Zhao Aping, a Manchu language-researcher, sounded the alarm as early as 2007 about its disappearance: “The spoken Manchu language is a living fossil, and we will witness the disappearance of this ‘living fossil’.” Researchers were particularly worried that China would not have enough Manchu users to decode the huge volumes of official documents from the Qing dynasty.
So far, Wang and her fellow researchers have made strides in collecting the data needed to preserve Manchu, but she warns that the digital reproduction of the language is still in the early stages.
Nonetheless, wrote Wang, time is of the essence: “No moment should be wasted.”
