For a long time now, I have been wondering, how did people back then learn Mandarin Chinese characters? Today, we have the convenience of learning new characters by just looking at the hànyǔ pīnyīn, which is the official romanisation system for Standard Mandarin Chinese in mainland China, and is also used in teaching Mandarin Chinese in Singapore, and to some extent, Malaysia and Taiwan. Created in the 1950s, the hànyǔ pīnyīn is by far, the most extensive teaching method for Mandarin Chinese characters today.

Rewinding back several decades, we find another transliteration system called the bopomofo, also known as the zhùyīnfúhào (注音符號). Created in the 1910s, it was used in mainland China up until 1958, but in Taiwan, it is still in use today. Consisting of 37 characters and four tone marks, the bopomofo attempts to transcribe all possible sounds in Mandarin Chinese. In fact, this is the main input method for Chinese characters in Taiwan, including dictionaries, books, and other documents.

However, these methods were all created in the 20th century. How did these characters get taught centuries back, or even, a couple of millennia back?

The answer seems to do something related to the bopomofo. You see, the bopomofo works by having separate characters for the initial consonants and rhymes and medials. Combining these rhymes and medials would yield the full combination of bopomofo finals, or how every single way a Chinese character or syllable could end. The tone markers, denoting one of the four Mandarin Chinese tones, would be placed last. So for, say, bottle in Mandarin Chinese (瓶子, pinyin: píngzi), it would be transcribed in bopomofo as ㄆㄧㄥˊ ㄗ. Remember this rhyming thing mentioned here. It will come in useful later on.

Early Chinese dictionaries used to employ this “sounds like” approach to teaching Chinese characters, and is also known as the dúruò 讀若 method. By showing a character with a similar pronunciation, this was perhaps the earliest known method used to teach new Chinese characters. Surviving examples showing this method include the Erya (爾雅) from the 3rd century BCE, literally translating to “approaching elegance”, using the 爾 (you) as a phonetic loan character for 邇 (approach, near).

Several centuries later, the fanqie method arose, theorised to originate from the introduction of Indian phonetic knowledge brought over to China in the 1st century CE, when Buddhism spread to China. While surviving sources and texts showing fanqie include the Yupian and the Jingdian Shiwen from the 6th century, perhaps the most renown surviving text was the rhyming dictionary known as the Qieyun (切韻), publishing in the Sui dynasty at the turn of the 7th century CE. The Qieyun was so significant, it became an “authoritative source” for literary pronunciations when Classic Chinese poetry boomed in the Tang dynasty. In fact, it is one of the most important sources linguists turn towards when trying to reconstruct Middle Chinese sounds as they were spoken back then.

So how did the fanqie work?

Simply put, a character’s pronunciation was essentially dependent on two characters, known as the “onset” and the “final”. The starting consonant was determined by the “onset”, while the rest, including the tone, medial glide, nuclear vowel, and the code, were determined by the “final”. Does it sound familiar? Because this is a rather similar method employed by the bopomofo.

The most famous example displaying this is the first entry of the Qieyun, with the character 東 (east). In the Qieyun, it is written as 德紅反. This means that 東 is pronounced with the onset of 德 [tək], and rhymes with 紅 [ɣuŋ], giving the pronunciation as [tuŋ].

But what about the 反 character? It does seem to be a usual marker used in the Qieyun, alongside 切 (hence the method called 反切), but their origins are relatively obscure.

And here is the confusing bit — several characters can be used for the same initial or final sounds, and no character was used to spell itself. So we have such cases arising:

The main implications from this is that both 東 and 德 share the same initial consonant. While this method could determine the categories of initial and final sounds, it could not reconstruct the precise sound value used in Middle Chinese back then.

You may have noticed that 東 had a suggested pronunciation of [tuŋ] in Middle Chinese, compared to dōng used in Modern Mandarin Chinese today, made from the characters dé and hóng according to the Qieyun. This shows how much Chinese pronunciation has changed since the publication of the Qieyun, indicating significant sound changes over time as languages evolved and are influenced by regional dynamics.

The issue about the Qieyun is, it did not directly record Middle Chinese as a spoken language, but more rather how characters should be pronounced when reading classic literature that used Middle Chinese. Additionally, no one knows for sure which variety of Chinese the Qieyun recorded. Some linguists point towards Chang’an, others suggested that it was an amalgamation of various regional pronunciations, while the late Chinese scholar Zhou Zumo (1914-1995) suggested that the Qieyun showed a north-south regional compromise between literary pronunciations from the Northern and Southern dynasties. In any way, the Qieyun, employing the fanqie method, could help historical linguists reconstruct the sounds of the more ancient forms of Chinese, giving a better picture of how Chinese sounded like back then, and how inaccurate many historic Chinese dramas were in pronouncing classical forms of Chinese.