I was reading the Wikipedia page on hypercorrection, and I noticed this particle section on Cantonese:

In Cantonese, some speakers omit the initial [ŋ]. For instance, the character 牙 (Jyutping: ngaa4, meaning "tooth"), ends up being pronounced "aa4". Prescriptivists tend to consider these changes as substandard and denounce them for being "lazy sounds" (Chinese: 懶音; Jyutping: laan5 jam1). However, in a case of hypercorrection, some speakers have started pronouncing words that should have a null initial using an initial [ŋ], even though according to historical Chinese phonology, only words with light tones (which correspond to tones 4, 5, and 6 in Jyutping) had voiced initials (which includes [ŋ]). Because of this hypercorrection, words such as 愛 (Jyutping: oi3, meaning "love"), which has a dark tone, are pronounced by speakers with an [ŋ] initial, "ngoi3".

This lead to me to think about the pronunciation of 啱. Looking at the entry on Wiktionary, it only has one pronunciation (ngaam1). It is in the first tone, but still has an ng- initial, and is the only counter-example I could find to the above paragraph. How accurate is the above paragraph from Wikipedia? If it is accurate, why is it that only tones 4, 5, and 6 that have ng- initials, and why does 啱 not follow that rule? I am also hoping that the etymology of 啱 will help explain this. I am assuming that 啱 is not the "original character", and is a newly invented character due to uncertain origins. The Wiktionary entry does have some speculations, but I also am wondering if other (possibly definitive) theories exist.

2 Answers 2


So it is a general rule of thumb that any Cantonese syllable with the voiced initial consonants ng-, n-, l-, m- plus tones 1/2/3 is not "Classical", as it violates the defining voicing-based tone split of the four tones that happened in the transition from Middle Chinese to Modern Chinese.

This means that 啱 is either not derived from Middle Chinese (so from a Tai-Kadai or an Austroasiatic or other non-Sinitic source), or it has gone through a more complex history that means it's only indirectly traceable to a Old or Middle Chinese lexeme (it thus might have a regular Old/Middle Chinese complement, which is often considered its 本字).

In the world of Chinese word etymology, things have very seldom been certain. A lack of good corpus data and earliest attestations for these dialectal characters is a major reason for the current state of affairs. Not much can be said "definitively" at all, though we do know it is not in the Kangxi Dictionary.

The theories on English Wiktionary seem possible, perhaps even reasonable (though unsourced at this point; I have found a source that points to the connection with 諳 deriving from a certain Mr 楊子靜, perhaps through his 廣州話分韻詞林), but that is all that can be said about them really.

The other colloquial word that comes to mind with ng- with tones 1/2/3 is , although that has a variant with the null initial too.

  • Another example is 鉤 ngau1, but the explanation for this is that it originally had a g- (IPA: /k-/) initial that shifted to ng- for some reason.
    – Claw
    May 11, 2021 at 17:17
  • @Claw Yeah, the 鉤 shift is usually attributed to avoidance of homophony with a vulgarity in Cantonese.
    – Michaelyus
    May 12, 2021 at 8:08

that only tones 4, 5, and 6 that have ng- initials

well, it's incorrect.

for the initial "ng-", there're several syllables have the tone from 1 to 6, such as:

nga1 丫 nga2 啞 nga3 亞 nga4 牙 nga5 瓦 nga6 訝


ngaau1 咬 ngaau2 拗 ngaau3 坳 ngaau4 肴 ngaau5 咬 ngaau6 樂


ngai1 繄 ngai2 矮 ngai3 翳 ngai4 危 ngai5 蟻 ngai6 藝


ngau1 勾 ngau2 嘔 ngau3 漚 ngau4 牛 ngau5 偶 ngau6 吽


about the quoted wiki page, it's doubtful.

  • 1
    I noticed that for the characters you listed, ones with tones 1, 2, or 3 don't have a ng- initial (according to Wiktionary). For example, 繄 is listed as ai1 (or ji1), 矮 is listed as ai2, and 翳 is listed as ai3. However, as you suggested, what is stated in the Wikipedia page may not be the truth, but I am confused as to how to accurately define whether a character should have an ng- initial or not based off of how the pronounciation has evolved from Middle Chinese. Many sources will list both readings (for instance, the link you provided listing the ng- initial readings as "X 的異讀字"). Mar 30, 2019 at 3:36
  • Another thing to note is that Wiktionary does not list any of the charactes with tones 4, 5, or 6 as having a zero initial reading, but again, I'm not sure how accurate Wiktionary is, which is partial why I would like some clarity on how these readings are determined. This note is also provided for characters with tones 1, 2, or 3: The zero initial /∅-/ is commonly pronounced with a ng-initial /ŋ-/ in some varieties of Cantonese, including Hong Kong Cantonese. Mar 30, 2019 at 3:39
  • 1
    If you're talking about the English Wiktionary, there are several very smart contributors there who know a lot about Chinese. You can ask a question about individual words or characters in the "Tea Room" and about etymologies in the "Scriptorium". If you're talking about the Chinese Wiktionary, I'm not familiar with it but they may have equivalent mechanisms for discussing characters, words, and etymologies. Apr 1, 2019 at 12:07
  • 1
    Also, the CUHK website itself states for every single example up there, that the ng- with tones 1/2/3 are 異讀字, and that there is a more common/standard variant without ng- at the front.
    – Michaelyus
    Apr 2, 2019 at 12:12

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.