Cantonese pronunciation of 丫 in 南丫島

[chrisp]

My dictionary tells me that the pronunciation of 丫 in 南丫島 (Lamma Island) is "aa1", yet I've seen it romanised in a couple of sources as "nga1" when used in the Cantonese name for Lamma Island - Naam4 Nga1 Dou2. This confuses me because I can't see any evidence that the character 丫 is ever pronounced "nga1" on its own - is it some form of regional variant or more traditional/older form? Or some mutation that occurs in certain circumstances? I can't believe it's a typo as two different sources have it romanised with an initial "ng". I know the initial "ng" is often dropped in Cantonese but my dictionary gives forms with initial "ng" in tact.

[Hofmann]

丫 has always started with a glottal stop from Middle Chinese to Cantonese. ngaa1 is probably a hypercorrection. After people started dropping velar nasal initials, some people started adding it in weird places. Kind of like non-rhotic English speakers adding intervocalic r's.

[Michaelyus]

I'd say the hypercorrection is probably 'primed' by having the nasal before it in those compounds. The general instability of the ng- and null initials is at the preliminary stages of leading to a little bit of a literary/colloquial situation (as 愛 is undergoing, with singing being a de facto 'literary' register) or even allophonic.

[chrisp]

Interestingly, I've just seen that my Chinese-English Dictionary, published by The Chinese University Press, gives the pronunciation of 丫 as ngaa1, and not aa1.

[mcgau]

The problem may lie in the sound "ng-" , which is getting diminished in modern Cantonese, notably in Hong Kong.

Besides 丫 mentioned by Chrisp, 牛, 我, 亚 and a lot more seem to bear the same meaning regardless of having ng- vowel or not.

[Hofmann]

That is because they can be distinguished by their tone. The glottal stop initial is voiceless, which evolved to have 陰 tones. The velar nasal initial is a voiced sound, which evolved to have 陽 tones.