原 heard as falling-rising

[realmayo]

I remember in the past getting frustrated by how I heard quite a few 2nd tones as 3rd tones. I think I've worked out why it used to bother me -- and though may well be common knowledge, I thought I'd put it up here.

 

I just heard 原谅 yuánliàng and although I know that yuán is 2nd tone, I was sure I heard a dip at the start.

Someone who knows about linguistics and Chinese pronunciation may be able to confirm that [yuán] is a dipthong which begins with a [yu] glide.

 

I think I'm right in saying that the "rising-tone" isn't expressed until the [an] part of [yuán].

And if the [yu] part of [yuán] isn't rising - and perhaps even falls - it can cause the whole [yuán] to sound like something that isn't a 2nd tone, but one that first falls and then rises.

 

(Or maybe it's that the [y] has a sound that dips and the [uan] rises - I don't know much about linguistics... but I think it's less critical than that one sound goes down and the next one goes up.)

 

I'll attach an audio of the brief sentence, the word, and a pic of praat's analysis of the tones of the single word yuánliàng (with the [liàng] in the red-shaded section).

 

Does anyone else hear it the way I described?

 

sentence   sentence.mp3

yuánliàng   yuanliang.mp3

 

 

[Tomsima]

I think perhaps noone has replied yet as the opinion of a native speaker is really the only valuable insight into a pronunciation question like this. That being said, perhaps sharing my perspective might at least give you an additional insight. Firstly, tones are relative, and they can change a lot according to sentence context, choice of word stress, and of course the all important level of 標準普通話 of the speaker. With this in mind, the fact that the tone does dip slightly at the beginning may just be a feature of the speakers own habits, and may not necessarily be the 'best' second tone one could produce for 標準普通話. That being said, the tone rises at the end more than it really ever would in a third tone that comes before a fourth tone in a compound. And that for me is why I hear this as a second tone. I just heard someone say the word 短暫 just now, which is a 3/4 combination; the 短 is clearly low, and if you ask me doesnt rise at all at the end. But take that with a pinch of salt, as I'm not a native speaker.

[Publius]

13 hours ago, Tomsima said:

clearly low, and if you ask me doesnt rise at all at the end

Good point.

The 3rd tone is almost always realized as a low falling tone (21), except when at the end of a sentence, there's an additional rise (14) to form a full 3rd tone (214), which is how you first learned to pronounce the 3rd tone in isolation.

Rising is not the distinguishing feature of the third tone. A 3rd tone does not rise at all if it's followed by a 1st, 2nd, 4th or neutral tone.

 

[realmayo]

I 100% agree that it would be a misunderstanding of the third tone, to hear "falling-rising" before a fourth tone and think the "falling-rising" must be a third tone. That's not how I hear it these days, certainly.

 

The way I hear it - is a fall for the first part of the sound and a rise for the second part of the sound.

 

What I'm trying to say is: I don't remember anyone saying that the second tone can be pronounced as a falling-rising.

 

I took the next three examples of 原谅 that I found and analysed them: they all show the rise doesn't happen until late in the sound.

(The four sound sources are from four different textbooks, with different speakers.)

 

That makes me think that the so-called "on-glide" - the 'y' or 'yu' part - plays a stronger role than I'd realised in the past. That's to say, it doesn't rise, and one can hear it long enough to hear that it doesn't rise, and often even dips.

 

Here are the screenshots of the next three 原谅s that I analysed. The shaded pink parts are, very precisely from the start to end of the 原 sound. The other sounds aren't indicated so precisely.

 

 

 

 

 

 

[realmayo]

Actually I should qualify the above: it's possible that in fact, a 3rd + 4th together can sound like falling + rising + falling

 

That's because when a speaker falls at the end of a 3rd tone, and then moves up to the start of the 4th tone, that upwards gathering movement (which technically should't produce any sound) is I think often heard in natural speech.

 

A go-to example of 3rd + 4th would be 买卖: here's the analysis:

 

 

- the pink shaded part is 买

- note that if you chop the 卖 in half, the first half is a rising "ma" and the second half is a falling "ai"

 

[edelweis]

1 hour ago, realmayo said:

when a speaker falls at the end of a 3rd tone, and then moves up to the start of the 4th tone, that upwards gathering movement (which technically should't produce any sound) is I think often heard in natural speech.

 

In the sentence of your 1st post, yuanliang is preceded by a 1st tone jia. In order to pronounce the 2nd tone, the voice must start lower than the previous syllable. Is this not the same phenomenon?

[Mig]

Without so much of linguistic expertise, as I am learning tones by heart, I noticed that 2nd tone is raising with a long vowel at the end. This is the way I am learning slowly to distinguish 2nd compared with other tones. Hope this make sense

[realmayo]

On 5/7/2019 at 8:16 PM, edelweis said:

In the sentence of your 1st post, yuanliang is preceded by a 1st tone jia. In order to pronounce the 2nd tone, the voice must start lower than the previous syllable. Is this not the same phenomenon? 

 

It could be. Maybe I'm particularly sensitive to how I hear second tones because I remember being grumpy when doing some pronunciation work with a tutor a few years ago that I was consistently not picking them right. I can still clearly remember that little room and me asking myself how on earth am I hearing third tones instead of second tones!

 

Also until a few weeks ago I'd barely heard any Chinese for about a year. But I had been drilling hundreds of minimal-pairs of pronunciation practise in another tonal language that I thought I'd try to learn. Coming back to Chinese, I feel like that kid in the Mel Gibson film, I see ghosts everywhere / hear tones everywhere.

 

Clearly for a Chinese syllable, is it only a certain part of its vowel that needs to behave like we all assume tones are supposed to behave. I'd love to know if it's consistently the same part.

 

I'm also very surprised about how much pitch change is present in unexpected places. Listening to mǎi, for example, the initial ['m' + schwa] sound definitely changes pitch.

Or I was listening to a qíng in 情况 - the rising tone in the qíng didn't start until after the regular 'i' sound had finished and we were into the '-ng'. But it sounded like a 100% normal 2nd tone to me the whole time - until I broke it down into bits.

And of course the initial sound for yuán, before the rise, which started all this off.

 

None of this would change how I'd go about practising my pronunciation - I'd try to model it on native speech, rather than on native speech that I'd analysed to bits.

 

 

 

[Tomsima]

1 hour ago, realmayo said:

I had been drilling hundreds of minimal-pairs of pronunciation practise in another tonal language that I thought I'd try to learn

 

Interesting, mind sharing more? 

 

2 hours ago, realmayo said:

hear tones everywhere

 

tfw when brain starts processing english intonation as if its a tonal language, really strange. must be even more so for you, as it sounds like the language you've been learning has more tones than in mandarin

[Michaelyus]

Native(-ish) intuition, no professional articulatory phonetics background.

 

The recording in post #1 feels natural and unremarkable to me. Relatively careful. The rising nature of the second tone on "yuán" is clear and noticeable, and I can't imagine confusing it with a third tone (although I have been primed to notice it from the title).

 

There are two issues that have to be addressed here: 

1) f0 vs tone perception

2) local tone vs global intonation

 

I will warn you though that this analysis is not necessarily conducive to better tonal perception and production, because in general people do not have the skills to describe pitch changes in language. It's difficult enough processing the sound signal into the mathematical form and then into a graphical form!

 

The first issue then:

a) Consonants affect f0. Voiced consonants (like Mandarin Pinyin m-) are at a generally low f0, and (often) lower than the vowel that follows.

b) Voiceless consonants (by definition[!]) don't really have a f0.

c) Hence, the pitch change on the vowel is identifiable, whereas consonants are mentally processed separately. I take this to mean that the tone is identified from the f0 of the vowel segment, and not the consonant segments.

d) However, the vowel itself also affects f0 (high vowels, like /u/ have a higher intrinsic f0 than low vowels like /a/), even in Mandarin (although the effect is apparently a lot weaker for /i/ and /o/).

 

With the issue of intonation:

a) Like the vast majority of spoken languages in the world, Mandarin has declination, where f0 reduces across the utterance.

b) Downstep also occurs in Mandarin, where certain tones, like the low tone / tone 3, cause a step change in the f0.

c) Anticipatory effects also occur. This gives rise to a whole different way of segmenting sound waves and spectrograms into phonetic units, the "target approximation" model.

d) These "coarticulatory" effects are perceived and judged by listeners as a holistic whole; remove the sound from its context, and it becomes significantly more difficult to identify its tone.

e) Naturally, declarative, interrogative intonations are different. Different accents are different: intonation is among the biggest drivers in accent perception cross-linguistically.

 

In summary, we know way too little about how intonation and tone interact to even characterise native speech, much less build a model. But if your perception and native listeners' is off, seeing the f0 pattern in front of you might allow you to see where your problems lie, if you bear in mind a litany of caveats.