If you’re a follower of the Cantonese-centric part of my blog, you might have heard about the Cantonese Conversations project elsewhere. It is an initiative that Olly Richards and I, among others, took a while ago to create a set of Cantonese learning materials that are completely organic and natural, in order to fill the gap of listening/reading materials on the intermediate/advanced level. It is a package of video/audio recordings of native conversations, together with a written transcript, Jyutping transcription and my English translation.
It came out officially a few months back, and recently I’ve been working on putting out a revised version, particularly focusing on proofreading the Jyutping. And that got me thinking: if I assume correctly, the majority of users will be reading the Jyutping instead of the Chinese script. And there’s a troubling thing about Cantonese: not only does it lack standardisation, but it is also undergoing a number of pronunciation changes. What that means is if you listen to us native speakers, there will be multiple ways of pronouncing the same words. We hardly even notice the differences, even though they tend to sound drastically different to foreign ears.
In the Jyutping transcriptions inside the package, we aim to
- represent the recording as accurately as possible,
- expose the reader to different pronunciations, and
- limit the number of varieties to make sure readers can still recognise them as the same word.
And quite often, these goals came into conflict, and we had to try and strike a balance. So halfway through the revision process, I thought I’d take to my own blog to explain some of the most common divergent pronunciations, so that it is still possible to recognise the words under different guises. The following pairs of sounds are often called ‘lazy sounds‘ (懶音 laan5 jam1) by native speakers. Chinese teachers will tell you one of them is right and one is wrong, while younger people would say one is newer and one is outdated. The fact is that they’re both in use, so I suggest you use either one yourself, but make sure to get acquainted with both of them. Here we go!
n and l
The elephant in the room has got to be the most notorious ‘lazy sound’ ever: n and l. In 21st-century Hong Kong Cantonese, syllables traditionally starting with n are now pronounced with n and l basically interchangeably, with context helping to differentiate the meaning. Common ones in Cantonese Conversations include:
- 你 nei5/lei5 (you): we wrote nei5 for all instances, even though quite often the speakers actually say lei5.
- 諗 nam2/lam2 (to think): we opted for lam2, but personally I say nam2.
- 呢 le1/ne1 (particle): we chose the more common le1.
- 男;女 naam4/laam4; neoi5/leoi5 (male/female): in both cases, the l-variant is taking over.
- 年 nin4/lin4 (year): speakers kind of use whichever they feel like. We opted for the traditional nin4 here.
ng and no initial consonant
Similarly, syllables starting with ng are getting the ng dropped – probably a piece of good news for learners! But due to hypercorrection, a handful of syllables with no initial consonant are getting an extra ng, so make sure to recognise them as interchangeable when you hear them. Examples are:
- 我 ngo5/o5 (I): we opted for ngo5, which is still more common.
- 哦 ngo5/o5 (oh): since this is a relatively meaningless utterance, we picked ngo5, even though both are common.
- 屋 uk1/nguk1 (house/family): this is the case of hypercorrection where some speakers add an additional ng to the original uk1. You’ll see the original in the transcript.
The ‘ng’ syllable
Ah, probably one of the most challenging syllables for learners. There are a handful of words pronounced just ‘ng’, most notably 五 ng5 (five), which shows up in one of the conversations. It is commonly reduced to m5 in speech, but we kept the original ng5 in the transcripts.
However, the negative particle 唔 m4 (e.g. 唔係 m4 hai6) is always m4.
Two syllables merging into one
We Hongkongers are fast talkers, which is why many pairs of syllables that commonly used together have merged into a new one, and both versions are used to mean the same thing. And thanks to the lack of standardised writing, we write either version. The most notable ones are:
- 乜嘢/咩 mat1 je5/me1 (what). 咩 is originally the sound of a sheep bleating, but now represents ‘what’, with its tone slightly dropping at the end.
- 係唔係/係咪 hai6 m4 hai6/hai6 mai6 (is it/isn’t it): if you find ‘m’ being a standalone syllable funny, we might do too, which why we reduce the latter two syllables to one. But note that 咪 mai6 alone doesn’t replace 唔係 m4 hai6.
- To take it a step further: 係唔係呀 hai6 m4 hai6 aa3 → 係咪呀 hai6 mai6 aa3 → 下嘛 haa6 maa5 → 呀嘛 aa6 maa5 / 下話 haa6 waa5. No seriously, this is like Pokémon evolution on steroids. But in reverse.
- 即係/啫係/??? zik1 hai6/ze1 hai6/zei16 (filler particle, like English ‘like’): the first one is the original, the second one is reduced, and the third one isn’t even possible to write down: it’s like we glide both the vowel and the tone from ze1 to hai6. This filler is ridiculously common, and you’ll see zik1 hai6 everywhere in the transcript, even though you hear zei16.
- 但係 daan6 hai6/daai6 (but): since this is such a common conjunction, speakers tend to say it so fast that you hear one syllable. Rest assured that we always write out both, as they’re intended to be heard.
There are lots of other instances where the speaker slurs a consonant out of existence, so please try to follow the jyutping to decipher what exactly they’re skipping.
joeng2 and joeng1
咁樣 (like this) is usually pronounced gam2 joeng2, but sometimes inflected to gam2 joeng1. We transcribe whichever one the speaker uses.
min6 and bin6
Because of mind-boggling linguistic processes, direction/referential words such as 呢邊 (lit. this side) li1 bin1, 嗰邊 (lit. that side) go2 bin1 and 下面 (below) haa6 min6 have lots of different pronunciations. Namely, 呢邊 ‘here’ can be pronounced:
- li1 bin1
- li1 bin6
- li1 min6 (also written 呢面)
We write down whatever the speaker says. Just make sure you know they mean the same thing.
Speaking of which, the li1 above has a ton of different readings. For example, in 呢個 (this one):
- li1/ni1 go3
- lei1/nei1 go3
- ji1 go3 (also written as 依個)
laa3 and laak3 / aa3 and aak3
Particles – a classic headache in many East Asian languages. While I’m not going to explain the meaning here, I’d like to note that lots of speakers use these pairs kind of interchangeably, depending on the intended rhythm of the sentence. We write down the pronunciation as it’s said in the recording.
seng4 and sing4
This is a very particular case that comes up frequently: when 成 is used for the sense of ‘entire’, such as 成日 seng4 jat6 (all day long/all the time) and 成個 seng4 go3 (the entire), it has two interchangeable pronunciations, sing4 and seng4. But outside of this context, it is always pronounced sing4.
There are still many more variances within the Cantonese language, even just in Hong Kong. In fact, if you (are able to) read the Wikipedia article for ‘lazy sounds’, you’ll see many changes to the language, some of which are in the past (i.e. one of the varieties is already deprecated). What I hope to do with this article is make you aware of some common divergences that I’ve encountered while listening to the Cantonese Conversations recordings and going through the transcriptions, so that you aren’t confused while reading the Jyutping. I’m sure that with sufficient exposure, you’ll get used to these inconsistencies pretty quickly. And if there’s any other important pairs that you think should be added to this article, please go ahead and leave me a comment below!