The past few months, I've been learning a language here in Singapore that's been noted for its crazy mind-bending use of
tone sandhi. I thought I'd write a little about it in this post, since it's a
phenomenon that some linguists may not be familiar with (given the tendency for many to run away at the first 'hearing' of anything tonal). At the end of this post, I'm also going to throw in a little puzzle set that I created, just to give people a chance to see the sorts of data some linguists work with. I'm hoping it'll appeal to all the puzzle solvers out there.
Tone sandhi in Mandarin Chinese
Experienced learners of Mandarin will already be familiar with the phenomenon, exemplified by the initially confusing and dreaded rule that specifies that Tone 3 becomes Tone 2 before another Tone 3. This prevents you from saying two Tone 3s, one after the other. For example, the word for 'you' in Mandarin is 你
nǐ (with Tone 3) when said on its own and the word for '(to be) good' is 好
hǎo (also Tone 3). However, when you put them together to get the ubiquitous Mandarin greeting 你好, written as
nǐ hǎo in
Pinyin, you find that 你 is now pronounced
ní with Tone 2. (This makes it homophonous with 泥
ní 'mud', but most speakers can work out from context that you're not talking about the quality of earth.)
Importantly, the rule applies whenever two Tone 3s occur next to each other in the same phrase, regardless of the actual meaning of the words. Using another example, 很
hěn, an intensifier with the meaning of 'very', remains as Tone 3 in phrases like 很多
hěn duō 'a lot' and 很快
hěn kuài 'very fast', since 多
duō has Tone 1 and 快
kuài 'has Tone 4. But if you want to say 很好
hěn hǎo 'very good', you would have to pronounce 很 as
hén, with Tone 2.
Ask a native speaker of Mandarin why on God's less-than-green earth they would say 你好 or 很好 this way, and they'll probably just say that 'it sounds nicer'. There's also actually no physiological, or aesthetic, reason preventing you from producing two Tone 3s in a row. The thing is,
tone sandhi rules are language-specific: some tone languages do allow sequences of similarly low (and creaky) tones to occur next to each other, while others may disallow sequences of two falling tones, which Mandarin does allow.
Of course, if you're only interested in learning a tone language that does have tone sandhi, it doesn't really help to ask
why it happens, or for instance, why Tone 3 becomes Tone 2 and not Tone 4. You just need to accept that it does happen and that it happens the way it does. And then you need to learn how to apply the tone sandhi rules in actual speech so you don't sound completely moronic.
Tone sandhi vs Tone change
On the other hand, if you're in the business of describing tonal languages, tone sandhi is something that pops up again and again. It can sometimes be a little tricky to talk about, since there's still some disagreement as to how to what the term 'tone sandhi', sometimes called 变调
biàndiào in Mandarin, should include. At least, it is generally accepted that 'tone sandhi' differs from 'tone change', or 变音
biànyīn, which describes similar kinds of tone alternations that are restricted to specific words, largely due to historical reasons. For example, 好 when pronounced
hào with Tone 4, means 'to be fond of' (example taken from Chen 2000: 31) - here you can see the connection with 好
hǎo '(to be) good', which indicates a likeable quality. However, this correspondence between Tone 3 and Tone 4 is specific to 好, and changing Tone 3 on another word to Tone 4 is not likely to yield a similar change in meaning.
In contrast, tone sandhi rules, which can also be the products of historical changes in a language, are more 'general', in the sense that they almost always apply regardless of the meaning of words as long as the necessary sound environment condition is present. However, there are instances when tone sandhi rules are not strictly observed - even native Mandarin speakers may sometimes fail to observe the rule described above when confronted with new compound words consisting of Tone 3 + Tone 3.
A tone sandhi puzzle
In the process of learning this tonal language in Singapore, which I'm calling 'Language X' for the moment, I came up with a little puzzle involving tone sandhi. It's similar to the problem sets we give out to undergraduate linguistic students, except I've simplified it a little so you don't need a lot of linguistic knowledge to solve it. I've used the letters A-G to indicate the tones, as well as some symbols known as
Chao tone letters which give a visual representation of the tones. The 'stopped' tones refer to tones on words that end in the consonants
k and
h.
You can view a draft of the puzzle below. Now this may not be the easiest puzzle to cut your linguistics teeth on, but I hope it gives you a taste of the sorts of data linguists work with, and the kind of analytic skills required to describe languages.
The solution will come in mid-June!
[I may have to post less frequently than I already do this coming month because I'm busy revising my Masters thesis to get it published.]
Reference
Chen, Matthew Y. 2000. Tone sandhi: Patterns across Chinese dialects. Cambridge: Cambridge University Press.