2nd tone at the end of a sentence

[realmayo]

Here's another good example: 窦文涛, the host of 锵锵三人行, saying:  咱们讲公民，我现在越来越觉得，公民这个词......

 

gm2.mp3

 

I'd be interested if anyone would consider listening to the audio first and assessing how the two 公民s appear to be toned.

 

When I heard it I thought I immediately thought the first 公民 has mín rising as a normal second tone, but that the second 公民 had its mín as not rising at all.

 

But according to Praat (which I'm no expert in and therefore I might be misusing) both instances of mín are very far from the textbook 'rising', as shown in the attached picture.

 

The difference appears to be that the first first mín is longer than the second. But on first hearing it, I could have sworn that the first one was a normal rising second tone, and the second one wasn't.

 

How do they sound to others?

 

 

 

I should make clear that none of this is about so-called neutral or atonal syllables like 了. Although I wonder: is there a north/south difference with words like 春天, 公民： would northerners generally pronounce the second syllables neutral? Or partially neutralised? Or is something else going on?

 

 

On reflection, I think this is simply a case of 轻声 being much more widely practised in real life than it is in textbooks.

[Apollys]

What's interesting to me is that if we look at the first time the speaker said 民, we can divide the final up into two clearly different vowel sounds.  This segmentation then roughly divides the sound into a portion with a falling pitch and a portion with a constant pitch.  This is something that I've noticed in a very rough and general way throughout my Mandarin studies, viz, that Mandarin speakers may find it natural to speak different sounds with different pitch contours.  I've also encountered examples of rising tones pronounced as falling-rising, where the division occurs at the vowel change.

 

And there's a more basic example, which is the way native speakers pronounce third tones (in isolation).  Now, when I read about the idea of what a third tone is, the question came up, well, how does the sound that I'm making over time align with the frequency curve?  Is 雪 pronounced xu\ - e/, or do I say the whole sound "xue" in a low slightly falling tone and then use that final vowel sound to rise again?  My instinct was to pronounce the whole sound in the first half-third tone, then use the final sound to rise (naturally, this makes sense because it transposes most easily to natural speech when you chop of that rising tail).  However, the other day I was talking to my girlfriend and said a word in that style, and she immediately said that was weird; I should say it the way she does (e.g. as described before, "xu" downwards, then "e" upwards).

 

So, I seem to have gotten a bit excited... If we get back to the example at hand, I definitely hear a falling tone on the supposed "second"' tone in that example you gave in both cases, and I think you are absolutely correct that we should focus on the fact that unstressed tones are everywhere (especially given the growing two-character footprint in Mandarin), and we really have to categorize unstressed tones in a separate category from their stressed counterparts in order to accurately analyze spoken Mandarin.

[realmayo]

Another issue is whether these are in fact 'neutralised' or just 'semi-neutralised' ... that is, is how they are pronounced in this unstressed position influenced by the tone that they ought to have normally. For example in these circumstances would the second syllable in each of 公民 and 春天 have exactly the same tone?

[Apollys]

I would say unstressed is a completely different category from neutral.

 

I'm investigating some more examples and seeing a lot of second tones that show up as dipping tones - a slight fall then a slight rise.  (Will the real third tone please stand up?)  In unstressed or mostly unstressed positions though, IIRC.

[L-F-J]

For me;

 

公民 was pronounced gōngmín both times on the mp3.

 

The second syllables in 公民 and 春天 never have the same tone.

[陳德聰]

Well this thread has been thoroughly eye-opening (or perhaps ear-opening) for me. Thanks for that paper. Also I played around with praat today too, so I am now moving from the regional differences camp into the "this is a standard phonological phenomenon" camp.

But what interests me about this is that I still fully perceive the unstressed second tone to be a second tone. Like I genuinely thought it was rising in my own speech even after listening carefully to the YouTube video from OP, and noting that she does pronounce it as non-rising, haha I feel kind of silly now that the acoustic reality has slapped me in the face.

[realmayo]

I still fully perceive the unstressed second tone to be a second tone.

 

Yep definitely! Must be conditioning? Because if you took someone new to Chinese and played them the 公民 from the audio above, I bet they'd never say they heard any rising on the 民。

 

Or: do you mean you Praated yourself and found you were also neutralising the 民?

[realmayo]

Just one other, how do you hear 工人 and 非常 in 那个工人的技术非常好。

 

tmp2ceze1.mp3

 

[Apollys]

I hear 人 as falling and 常 as rising.

(Checking your Praat ss, I think it's likely that higher pitched part at the beginning is the ch, and the vowel sound itself is rising.  I didn't Praat it myself though.)

 

Edit: Okay I've gathered some mp3's over the past day that exhibit patterns like this in the second tone.  A couple of them may also have second tones that follow a "dip" frequency curve, first falling and then rising.

332467.mp3

332472.mp3

333764(昨,人).mp3

333791.mp3

334543.mp3

335104.mp3

G012-03.mp3

G025-03.mp3

R018-DLG.mp3

R030-DLG.mp3

R093-01.mp3

tmp8xnev2.mp3

[realmayo]

4 minutes ago, Apollys said:

that higher pitched part at the beginning is the ch, and the vowel sound itself is rising

Your analysis is correct (according to Praat).

[陳德聰]

On 2017/1/16 at 0:11 AM, realmayo said:

Or: do you mean you Praated yourself and found you were also neutralising the 民?

 

I mean I Praated myself (haha does that sound as funny to you as it does to me?) and discovered that despite my belief that I consciously pronounce every 2nd tone rising, I indeed do not. I genuinely thought that as a mid-to-late L2 learner I wouldn't have internalised that without being explicitly taught it.

Edited January 18, 2017 at 06:15 PM by 陳德聰
a million typos

[realmayo]

Praating oneself sounds a lot more pleasant than it really is!

 

16 hours ago, 陳德聰 said:

I genuinely thought that as a mid-to-late L2 learner I wouldn't have internalised that without being explicitly taught it.

 

I guess it just illustrates the inadequacy of the four-tone-diagram that we're all taught at the start, that internally we recognise how it fails to capture real speech, even if we don't actually know that we are aware of any discrepancy.

 

Just one more tiny snippet to add to the list: doing a listening comprehension, from a BLCU textbook; it's question 10 so the guy says: 十： blah blah blah. He definitely goes down and then up. I heard it clearly and only went to Praat for confirmation.

 

When I spent time trying to improve my tones, speaking them in isolation, this dip-on-the-start-of-a-second-tone was something I really struggled to eradicate. But a BLCU pro is doing the same thing!

 

audio shi.mp3

 

 

 

 

[L-F-J]

Again, for me, 工人 and 非常 are both obviously pronounced with a rising tone on the second syllable.

 

I guess your software is picking up on the second syllable starting off lower than the first syllable and is interpreting that as a dip. It seems to blur the tones throughout the utterance because it doesn't know where one syllable ends and another begins and is trying to make a continuous line.

 

In any case, I think you guys are splitting hairs and it really doesn't make one iota of difference. You obviously understand what the native speakers said, and so long as you try to hit the textbook tones when speaking, you'll be understood too.

[陳德聰]

Praat is a helpful tool that linguists often use for analysing speech, and while you could actually make the effort to smooth out the formants so it doesn't look choppy, it's not worth the effort if you can just identify for yourself where one syllable starts and ends based on the waveform.

 

10 hours ago, L-F-J said:

You obviously understand what the native speakers said, and so long as you try to hit the textbook tones when speaking, you'll be understood too.

 

I think we have already progressed far beyond "understanding" and "being understood" in this thread. I work as an interpreter, so while I do sometimes worry about the understanding and being understood issue, it is for a very different reason. But I think it's fascinating for purely academic reasons that the "textbook tones" exhibit much more variation in real life, and not just due to standard vs non-standard pronunciation (at least in my case I can be relatively sure it's not non-standard pronunciation).

 

I'm curious, could you record your pronunciation for us in a sentence? I'd like to Praat you.

[realmayo]

11 hours ago, L-F-J said:

for me, 工人 and 非常 are both obviously pronounced with a rising tone on the second syllable

 

I agree with you that 常 rises. But 人 doesn't rise.The pitch does not go higher. 

 

However it's clearly a second tone. So you 'hear' it as a rising tone. But it doesn't rise. I suspect that because your listening ability is so strong (I remember from another post, on those news-speed excerpts), you hear 工人's 人 as a second tone and know that it's a perfectly natural, normally-spoken 人, which you've heard pronounced that way thousands of times, and so figure that it must be a normal 人, which must be a rising tone ... except it isn't.

 

11 hours ago, L-F-J said:

I think you guys are splitting hairs

 

I'm sure you're right. But as 陳德聰 says, for some people this is all interesting in itself. Also, if you're listening hard to a perfect native speaker and the tones you hear don't match what you're expecting, it's comforting to get some explanation as to why: phew, my ears aren't deceiving me, they really are pronouncing it like this and that's normal.

 

For instance, until this topic was started I didn't know about Tone 2 sandhi, where basically in a three-syllable word, if the middle syllable is second tone, it can change to first tone (under certain circumstances).

 

As for 公民 and the other examples here, a regular Chinese speaker just explained this away as 轻声. I suspect it's true: there's a lot more 轻声 in regular speech than the textbooks say there should be. In San Duanmu's The Phonology of Standard Chinese he talks quite a lot about unstressed syllables:

 

in a natural speech corpus of Chinese (Duanmu e t al. 1998), about a third of all syllables are unstressed and reduced.

 

He says:

 

Stressed syllables carry lexical tones and are longer than unstressed syllables, which do not carry lexical tones. In addition, unstressed syllables undergo rhyme reduction. I will argue that stressed and unstressed syllables have different structures, in that stressed syllables are heavy and unstressed syllables, light.

 

Apart from the fact that full syllables have lexical tones and light
syllables do not, full syllables are louder and have greater duration and
amplitude than light syllables. For example, the rhyme duration of full
syllables is on the order of 200 ms, whereas that of light syllables is on the
order of 100 ms. In addition, light syllables show considerable reduction
in the rhyme (M. Lin and Yan 1980). In this regard, all full syllables are
phonetically stressed and all light syllables are phonetically unstressed.

 

 

 

[realmayo]

Also why is there all this focus on the second tone? What about the others?

 

Perhaps someone with a linguist background who is comfortable talking about z-scores could confirm, but my reading of a short paper here http://www.ling.upenn.edu/~laic/laic10poly.pdf is that: in two-syllable words, the tone of the first syllable is likely to be 'standard', while the tone of the second syllable is likely to be 'non-standard'; and furthermore, that the likelihood of the second tone being 'non-standard' is far higher than for the other three tones.

 

As in here:

 

 

Figure 4: Syllable differences for disyllabic words grouped by tone: F0.