3rd tone runs

[歐博思]

Also, I wonder if 幾種語言 being a question versus a statement would differently affect people's reading of 種.

[Apollys]

In which way were you thinking it would influence the pronunciation?  The example I had was a question (你会说几种语言？) and in the given audio sample the speaker pronounces 种 in the second tone.

[歐博思]

My hunch was questions would be grouped together more tightly, so more 3rds read as 2nds, which happens to match in this case.

[Yuliyyya]

From what I know, if you have consecutive 3 tone words together like 你只给我好好.... all 3rd tone words before the last one turns into 2nd tone,

 

so individually this would be:

你只给我好好 ni3 zhi3 gei3 wo3 hao3 hao3...

 

but when grouped together it would be pronounced as:

你只给我好好 ni2 zhi2 gei2 wo2 hao2 hao3... 

 

tone changes also applies to the word 不 bu4, normally a 4th tone but if it's followed by another 4th tone, bu4 becomes bu2

such as 不要 it is pronounced as bu2 yao4 not bu4 yao4

[Guest realmayo]

FYI:

 

买好酒

 

meaning 1: ‘finished buying wine’

[ [ 买 好 ] 酒 ]

can only be 2 2 3

 

meaning 2:  ‘to buy good wine’

[ 买 [ 好酒] ]

can be 2 2 3 or 3 2 3

 

via http://www-personal.umich.edu/~duanmu/ToneSyntax05.pdf

[Apollys]

Hmm, I wonder what eddyf would say about that paper's assertion that 我想买书 should be pronounced 2331.  He seemed quite confident that consecutive third tones should never occur.  (Also, next time I come across the audio I had for 我只想好好爱一个人 I will post it, since I think I recall that being pronounced with back-to-back third tones as well.)

[XiaoXi]

So, two different answers and no explanation for either... Guys please explain how I'm supposed to approach these in general. Or at least why you chose the pronunciation you did. 

 

That's just what sounds natural to us. You can't explain it any more than a basketball player explains how he can throw the ball into the hoop again and again. You just need to listen to a lot more. Eventually it will just be natural to you. You already did it once with your native language, just need to do it again. For now you can just rote practise the ways we've given you.

 

Hmm, I wonder what eddyf would say about that paper's assertion that 我想买书 should be pronounced 2331.  He seemed quite confident that consecutive third tones should never occur.  (Also, next time I come across the audio I had for 我只想好好爱一个人 I will post it, since I think I recall that being pronounced with back-to-back third tones as well.)

 

Yeah it sounds a bit weird to me. I'd say 我想买书 3231. You can't state two third tones in a row unless there's at least a slight pause in them. In that example there could be if it was 我想。。。买书。In this case it would be 2331.

[Apollys]

That's not an acceptably efficient learning process for me.  Learning a second or third language like it's your first is a terrible waste of time and resources.

[Guest realmayo]

It's best to be too fussy rather than too lax over pronunciation, sure. But that doesn't change the fact you're being too fussy (by trying to nail down a predictive theory you're comfortable with on a topic which remains controversial among Chinese linguists and scholars).

 

As for third tones, a rival way of looking at it is that the third tone is always the third tone and never becomes a second tone. That view requires you to see a third tone as comprising two parts, a low-or-falling and a rising part. And so a third tone can sometimes be pronounced in full, sometimes with just the rising part, and sometimes with just the low-or-falling part.

[Apollys]

Yes, that is how I have always thought of it.  I wouldn't consider it a rival view though, it's an equivalent view.  But I like it because it offers at least a slight motivation for the tone Sandhi.

[Guest realmayo]

It's rival because most people say instead that the third tone changes  to a second tone. But the alternative view is that the changed third tone sounds a little bit different to a second tone, and therefore it's not a second tone.

[Apollys]

I've always wished there were a convenient way to test this, sadly I don't think it would be worth the effort.

[Guest realmayo]

It's interesting enough for linguists I guess but hardly relevant if you want to speak the language fluently.

[Yuliyyya]

Hmm... now that I've thought about this carefully, I normally don't think about how the 3rd tone words really change it just come natural to me as I say them, but the 3rd tone change differs based on how you would group or pause between saying individual words.

给我打电话。gei3 wo2 da3 or gei2 wo2 da3 would be correct depending on how you pause in between

Example:

请给我打电话。qing3 gei2 wo3 da3 dian4 hua4

请 给我 打电话。

So when I read this, I would pause after I say the word 请 hence why I still say it in 3rd tone, and 给我 I would say in 1 group, because it just makes sense to me to group them like this.

你可以给我打电话吗? ni3 ke2 yi3 gei2 wo3 da3

Again if I were to break it up and say it in terms of pauses it would be

你 可以 给我 打电话 吗？

So within each "group" all consecutive 3rd tone words except the last one would change to 2nd tone.

This one is an awkward one as in I said this phrase to myself over and over how the tones would go whether it's gei2 wo2 da3 or gei2 wo3 - pause - da3 dian4 hua4 but the way I say it is as written as above hence why 我 was pronounced as 3rd tone due to how I pause.

我可以买你喜欢的那个。wo3 ke2 yi2 mai3 ni2 xi3 huan de na ge

我 可以买 你喜欢 的 那个。

If you were to read it like this exaggerating the pauses and break them into groups I think you'll find that they follow the 3rd tone rule change pretty straight forward, so far...

The 3rd tone change just depends on how you pause in betweeen each phrase, so there's no real right or wrong just say it however comes natural to you. But certain phrases will always make sense to group together such as 可以 because it is one phrase.

我只想好好爱一个人。wo3 zhi2 xiang3 hao2 hao3

我 - 只想 - 好好 - 爱 一个人 。

哪里可以买自行车? na2 li3 ke2 yi2 mai3

哪里 - 可以买 - 自行车？

na2li3 ke2yi2mai3 zi4xing2che1?

你会说几种语言？

你会说 - 几种 - 语言？ni3hui4shuo1 ji2zhong3 yu3yan2?

[eddyf]

I would say the guy who wrote that paper seems to have done his homework so I guess there's some linguistic justification for claiming that two 3rd tones can happen in a row. But I still think that either there has to be a slight pause between them, or the second syllable has to start lower than the first syllable ends. Two connected low tones at a constant pitch just sounds weird to me.

And even if there are some scenarios where two consecutive 3rd tones are allowed, are there any scenarios where they are required? If not, then maybe we are better off not using them, assuming it will be wrong most of the time.

Btw I agree that this is one of those topics where learning it intuitively through exposure will actually be easier than trying to work out all the rules and use them. Our brains have built-in language learning instincts. You have to let them do their thing.

[陳德聰]

Was just talking about this with the boyfriend, who just like me pronounces 你會說幾種語言? as "nǐ huì shuō jízhǒng yǔyán", that is, with two consecutively falling/low 3rd tones (neither are fully realised). I then was like okay, what about if I said it "nǐ huì shuō jízhóng yǔyán"? We both came to the conclusion that if someone said that in conversation, the difference is not enough to even notice, let alone make a natural/unnatural pronunciation judgment, but both of our pronunciation habits are to have two consecutive low 3rd tones.

I did a really crappy research paper on how focus affects tone sandhi in my undergrad where some of the background literature and a very limited sampling of my random friends showed that trying to elicit focus on a 3rd tone that would have otherwise undergone sandhi usually confounds the sandhi process. With that as a hypothesis, the difference between 幾種語言 in a statement vs. question could very well produce different sandhi results.

With regard to 我想買書, isn't the position of the paper that [買書] operates as a unit? Thus it would make sense that sandhi does not operate across this boundary due to the pairing. I would say 2231, unlike what is proposed by that paper and what is proposed by XiaoXi, but you also have to take into account potential region-based differences too. A cursory glance shows that the paper talks about speakers of backgrounds with Min languages... It's possible that speakers' language backgrounds will affect the way they interact with tone sandhi in "Standard Chinese". Also 2331 is not something I'd say, but it doesn't perk my ears up when I hear someone say that. It sounds perfectly acceptable. 3231 on the other hand, actually sounds a bit bizarre to me, but again, it was already stated in #7 that there is enough variety in speaker judgments that we could just say "no set rules".

But there are definitely things that sound weird to the point of not being okay. 我想買本書 to me can be pronounced acceptably as 23231 and 32231, but not 22331. Probably because [買本書] is a unit, in which 買本 pairs up to necessitate a 23 combo. 22231 would also just be kind of silly. Hasn't this been discussed a lot on the forum? I feel like there have been tonnes of papers posted in the past.

[Apollys]

Nice, the good news is the examples you gave line up with my intuition!  And it's nice to know there's room for flexibility, which accounts for the confusion that I've had thus far given the variance I've encountered so far in how people deal with third tone chains.

[Apollys]

A question inspired by Yuliyyya's post: I'm wondering if you guys have any insight into the pronunciation of 好好 as 33 (23) vs 35, I've always heard it as 35 so far, but I've seen it written as 33 also.

[eddyf]

I actually learned it as 31, possibly with 儿化音 (好好儿).

[Flickserve]

Apollys,

I think what you can do is do some sampling of different native speakers and examine their tones.

Stick the MP3 of the sentence into a program called praat which analyses the tones and can give a graphical output of the tone.