I know it will get better with time, but applying tone sandhi in practice seems unnatural when I’m struggling to remember the next word in the sentence. In my textbook, one of the practice sentences for drilling is, in the context of confirming a sequence of directions:
我出了這個飯店、往北走、對不對?
“After leaving this hotel, I walk north, right?”
When it comes time for me to speak this sentence out loud in a normal speaking speed (and as part of a longer practice conversation, which is already placing burdens on my short-term memory), I first remember that 往 supposedly has the third tone but seems to be spoken in almost a neutral tone. Then I struggle to remember the word for north… oh yeah, it’s 北 from 台北, so that’s third tone, then I speak 北, and then I am remembering the next word 走 – oh yeah, that’s a third tone too, so the 北 (which I already spoke out loud) should have been spoken in the second tone.
I’m finding that becoming able to speak at full-speed requires some kind of sentence planning forethought that I’m not used to. It’s not enough to remember how to say the next word in the sentence, because its pronunciation may be affected by the word after it. It seems to require phrase-level (not word-level) planning in the mind before the vocal tract is engaged. So when planning the sentence in real-time in my head, I have to be able to recall 往北走 as a complete conceptual phrase (with tone sandhi already applied) that can then be recited at full-speed.
But at the same time, it’s also necessary to remember the original third-tone value of 北 (and maybe also 往) for correct pronunciation in other contexts.
I suppose it could be rationalized as being similar to the pronunciation change between the words “perfect” and “imperfection”. They both contain the word “perfect”, but as part of the word “imperfection”, it simply has to be memorized that in this context, the stress must be on “fec” and not on “per”.
From someone who went from zero Chinese to business level Chinese with no classes (most my Chinese came from my wife who doesn’t speak English), I can tell you that tone sandhi is not something you should worry about. In my opinion, it’s not something that you remember to do, it just comes naturally when you are speaking a sentence. Most Taiwanese don’t even realize that tone sandhi is a thing…
Yeah God bless you if you can really remember that kind of thing, process it in real time, and say it. Try Taiwanese where each tone has its own sandhi It comes naturally eventually.
What’s controlling the tones is your larynx, and when you are speaking fast, if you want to read the third tone absolutely like how the textbook describes it in a row, you are asking your larynx to fine tune tension control several times in a very short period. That’s why people just get lazy and let it sort of glide over. There is absolutely nothing wrong with saying every tone exactly right, especially if you slow it down a little, it would just sound like you are emphasizing a point.
The sandhi rules are a lot more complicated in Taigi, and while I learned that, I realized it’s better to treat them not as individual characters, but learn them as a group. So instead of 往‧北‧走, then think about how to apply sandhi rules, think of the whole thing just as 往北走 with the sandhi built in. It is easier to do that when you are just listening and speaking, not just reading something as text first.
The course I’m using is an old-fashioned FSI course which emphasizes various kinds of drills. I used FSI for another language and eventually gained fluency in it, so for me the drill-drill-drill approach of FSI works.
Now, one of the drill types is a substitution drill. In this case, you are given the sentence pattern:
我出了這個飯店、往東走、對不對?
Then you are prompted with each of 東、西、北、南、左、右、etc. and must repeat the same sentence while substituting only the direction word. I find this drill useful and realistic (and there are more complex drills also, like transformation drills, requiring you to restate a concept using a different sentence structure, in real-time).
But due to tone sandhi, what I feel like is happening in my head is not the substitution of the single direction word like 東、西、北、南、左、右, but instead a substitution of the entire phrase (with sandhi applied) of 往東走、往西走、往北走、往南走、往左走、往右走. That seems to be the most effective “chunking” strategy that allows me to do the real-time substitution in my head and to speak the sentence at full speed. And by practicing repeating these phrases (往東走、往西走、往北走、往南走、往左走、往右走) over and over in isolation and with correct tone sandhi, then when it comes time for real-time substitution, I can do it.
Yes, exactly. That’s the conclusion I finally came to after stumbling through the sentence word-by-word and realizing, “hey, this isn’t working – I have to chunk it into a conceptual phrase 往北走”. Of course, the textbook doesn’t tell you this; it just expects you to keep practicing until you can say it at full speed (including substitutions and transformations).
One challenge with self-study is keeping up a good pace. It’s tempting to keep working on one set of drills for weeks at a time until I have completely mastered them and can say them with full fluency, but that level of time investment then slows down the overall progress through the entire course.
I was just thinking that I didn’t have this issue with Mandarin but do have it Taiwanese.
Eventually even Taiwanese starts to feel more natural by getting used to hearing common words like kóng (講） as second tone at the end of a phrase or sentence or first tone before another word.
Yes exactly this. This is how I used to practice by remembering tone combinations. If you know how 美國， 莓果，and 沒過 should sound then you can mimic that for any word with same tone combination.
I totally get your frustration, but just trust that with time–as you said–it becomes second nature. I’m by no means fluent, though I should be, but tone sandhi isn’t an issue at all (yet I’m still completely bewildered by Taiwanese).
I liken it to how in English we don’t think about changing a to an before a vowel sound, or whether plural s is pronounced as /s/ or /z/. After years of speaking it just comes naturally.
Do you usually say “he isn’t” or “he’s not”? For the life of me, I have no idea what I usually say; it just naturally comes out. Same with Mandarin’s tone sandhi.
Yet there are times when two characters can be read in the same reading with or without sandhi to give completely different meanings. It’s a part of the grammar.
Tone Sandhi exists across tonal languages because it’s the natural way to produce the sounds. Imitation of native speakers will get you much farther than trying to remember the rules. I have very advanced Chinese and a background in linguistics and had to take a moment to remember which tones do what with tone sandhi when I saw your post. That’s how irrelevant it is to me. Other than being told about its existence and being drilled on the tone changes in 1:1 classes, it has had little impact on my Chinese experience…
For me it’d be the equivalent of trying to remember when to tap your ts and when not to when speaking General American, or when to insert linking rs in the UK or elsewhere. If you are thinking about how this word is spelled or does the next word starts with a vowel when you speak, you are probably struggling.
It’s pretty obvious how 3>2 tone in Mandarin makes pronunciation easier, but it’s harder to understand how the tone sandhi in Taigi came about. I mean, 1>7, 2>1, 3>7, 4>8, 5>7, 6th tone is gone, 7>3, 8>4, but some of those are different if you’re in the south or north. If I’m even remembering correctly. I guess it’s mostly tones going toward the center of the tone space but having the 4 and 8 just swap (in one dialect) seems…cruel. lol
At Maryknoll, back in the day, right after they dropped three two-inch thick textbooks on your desk and said “After you finish these we can really get started”, they had you count to 100 multiple times. It was helpful in getting used to the tone changes by ear.
This still the best way to get Taigi tonal sandhi in the head. Saw it first on @greves ’ youtube video.
1 陰平
2 陰上
3 陰去
4 陰入
5 陽平
6 陽上
7 陽去
8 陰入
-ptk
-h
-ptk
-h
Oo-oo
khóo-khóo
tshàu-tshàu
siap-siap
phoh-phoh
tâm-tâm
bān-bān
ku̍t-ku̍t
po̍h-po̍h
The following won’t at all help memorize Taigi’s tonal sandhi system, it’s just my guess on the logic/mechanism behind the system. Here I’m only talking about the most common Taigi tonal sandhi, the first character in a doublet changes its tone. There are three pitches high (33), mid (22) and low (11) in the tone scale.
My personal observation about Taigi’s tonal sandhi is that it’s almost a pitch accent system. By that I mean there if the original tone is just one note, the sandhi shifts down one scale.
That corresponds to 1(33)>7(22), 7(22)>3(11)
For the entering tones, it’s flipped. So 4(2)>8(3), 8(3)>4(2)
When you can’t go any lower, then you move to a glide downward, simulating it’s getting lower.
3(11)>2(31)
When the original tone is a glide through pitches, such as 2(31), 5(12) then the sandhi changes it to the starting toneme, aka the first pitch. Here I’m talking about Tsuân-tsiu rules, so 5>3.
2(31)>1(33), 5(12)>3(11)
If you go with the Tsiang-tsiu rules, 5(12)>7(22), then I guess my rule starts to fall apart.
It comes naturally eventually.
