This is the third and final 🧵 on how linguistics can—and should—inform our understanding and appreciation of ancient Chinese poetry.
2/ We’ve been using the well-known Tang poem 送友人 Sòng yǒu rén by 李白 (701–762) (M. Lǐ Bái / Lǐ Bó, Cant. Lei5 Baak6) as an exemplar, and we’ll stick with it in this last installment.
3/ Earlier 🧵🧵 discussing rhyme and meter are at these two links.
4/ Rhyme and meter are near-universal features of poetry. They are found in poetic traditions around the world, across a wide variety of cultures, places, and times.
5/ But the particular ways they manifest in medieval Chinese regulated verse is closely connected to a feature of Chinese that not all languages possess: lexical tone.
6/ I tried to make the case in those earlier threads that we can’t fully appreciate the aural qualities of these poems, and of their poetic structures like rhyme and meter, without some understanding of the pronunciation of that time.
7/ The feature I want to talk about today, parallelism, isn’t like that. A modern reader who knows enough Classical Chinese to understand the meaning of the poem is going to be able to spot and appreciate parallelism without any specialized technical knowledge.
8/ But understanding how parallelism of the type we see in regulated verse is even possible, and fully appreciating the skill required to craft parallel lines, requires some linguistic explanation.
9/ The linguistic features of a language—the structural scaffolding that is its matrix—create possibilities for and impose constraints on literary expression. Classical Chinese is no exception.
10/ The two linguistic features of Classical Chinese of the medieval period that drive poetic parallelism are monosyllabicity and—you guessed it—tone.
11/ Before we dive in, let’s see our poem again. I give it in Chinese characters, followed by an English translation.
Green hills cross past the northern outer walls;
Clear waters circle ’round the eastern inner walls.
This is the place where we part,
Lone tumbleweeds on a journey of ten thousand miles.
...
14/
Like the drifting clouds are the traveler’s intentions;
Like the setting sun are the old friend’s feelings.
Waving your hand you depart from here,
To the the whinnies of our separating horses.
15/ The poem consists of eight lines grouped into four couplets. Two of those couplets, the first and third, exhibit parallelism. (These parallel couplets are called 對聯/对联.)
This parallelism operates simultaneously on *four* levels:
16/ The four levels of parallelism:
1️⃣ grammatical structure (i.e. sentence-level syntax)
2️⃣ word class (i.e. parts of speech or lexical categories)
3️⃣ meaning (i.e. semantic areas)
4️⃣ tone (i.e. pitch contour)
17/ The fourth of these—tone—is a bit different from the other three because, as we will see, it’s more of an anti-parallelism—a juxtaposition of complementary elements—like the paired nucleotides in the double helix structure of DNA.
Which is why parallel couplets are so cool.
18/ Let’s take a look at the first two lines of our poem, and investigate the four aspects of parallel construction that are operating in concert.
(I’ll over-simplify a bit to avoid getting too technical.)
19/ Here are those first two lines again, with word-for-word English translations:
青山橫北郭
Green / mountain / criss-cross / north / outer-city-wall
白水遶東城
Clear / river / wind-around / east / inner-city-wall
20/ 1️⃣ Sentence structure
Each line is a sentence with parallel grammatical structure: a Subject, a transitive Verb, and an Object: S-V-O.
青山 (S) 橫 (V) 北郭 (O)
白水 (S) 遶 (V) 東城 (O)
Not too interesting, I admit.
21/ 2️⃣ Word classes
Each pair of corresponding words in the two sentences belongs to the same part of speech:
青山橫北郭
白水遶東城
Adjective / Noun / Verb / Noun / Noun
22/ (I said we wouldn’t get too technical, so we’ll just sidestep the question of whether 青 and 白 are nouns or adjectives, and of whether adjectives even exist in Classical Chinese. "Adjective" is good enough for now)
23/ 3️⃣ Semantics
This is where things get more interesting. It’s not just that nouns match nouns and verbs match verbs. The parallelism goes beyond grammar, it’s also semantic.
The juxtaposed words must belong to the same domain of meaning.
24/ Here’s how we might characterize those domains for these two sentences:
青山橫北郭
白水遶東城
Color / Natural-Object / Path-of-Motion / Cardinal-Direction / City-Wall
25/ Before we get to 4️⃣ Tone, let's pause here to consider whether what we've seen so far is special or not from a linguistic perspective. How much do we see this kind of parallel structure across different languages?
26/ Those of you who have read a lot of my previous threads will suspect that I am going off on a tangent here mainly as an excuse to show a GIF of a little rascal impatiently drumming his fingers on the desk.
27/ That's not the case! So let's just get that GIF out of the way now, and then we'll proceed as if it never happened.
28/ Okay, where were we?
Oh, yes: Is this kind of parallelism unique to Classical Chinese?
Well, no, not generally speaking. All languages employ parallelism for rhetorical or literary effect, so far as I know.
But there is something special at work here.
29/ It’s possible to achieve similar levels of parallelism in English or Mandarin poetry, but it’s harder.
Why should it be harder? Grammatical and semantic matches don't seem that hard to construct.
30/ To get at this, let's start by answering a different question:
What kind of feature of the classical literary Chinese language makes it possible to construct such cleanly parallel sentences under the formal constraint of having an equal number of syllables in each line?
31/ The answer is: monosyllabicity. In Classical Chinese, a very large percentage of vocabulary words is one syllable long. It's a higher percentage than in any modern Chinese language.
32/ To be sure, there are exceptions. There are some bisyllabic words (sometimes called "binomes").
But those exceptions are mostly limited to a few well-defined domains: onomatopoeia, insect and plant names, and sound-symbolic/ideophonic/mimetic words.
33/ These areas of vocabulary are not trivial, but they are somewhat marginal and it’s easy to avoid them when compiling parallel couplets—or juxtapose two of them against each other because they are in the same semantic domain.
Some binome examples: 菡萏 蝤蠐 窈窕
34/ The key point is this:
For most Classical Chinese vocabulary, single-syllable words are the norm. And that means that you can easily construct parallel lines *that end up being the same length*. Which is a formal requirement of regulated verse.
35/ It’s not so simple in English.
In Classical Chinese, all color words are one syllable.
But in English we have “yellow” and “purple” as well as “red” and “blue”.
36/ Likewise, in Classical Chinese, all natural object words are one syllable.
But in English we have “sky” and “lake” as well as “mountain” and “river”.
37/ So in English it’s harder to maintain parallelism in a couplet while keeping line lengths the same. Doing so would probably constrain word choice to the detriment of expressive power.
I guess that's arguable. But if you've tried your hand at translating into English,
38/ you've probably run up against this problem.
You can see it in my English translation, which strives for parallelism but doesn’t quite achieve it.
Green hills cross past the northern outer walls;
Clear waters circle ’round the eastern inner walls.
39/ I cut “around” down to “’round” to match the single syllable of “past”, but “waters” and "circle" are two syllables against “hills” and "cross", each one syllable. So my first line has 10 syllables and my second has 12.
Argh.
40/ Sure, I could have matched the length of “waters” by changing “hills” to “mountains”. But that shifts the image that the line paints. I wanted “hills”.
[But I dunno, maybe “mountains” is better anyway? Well, this is a tangent we probably don't want to take.]
41/ The point is that the varying syllable lengths of basic English vocabulary create complications that aren’t present in the Chinese. And that is one reason parallelism is an achievable goal for Tang poets, achievable without curbing expressive power.
Monosyllabalism FTW!
42/ Now, you might think I’m saying these poets weren’t very skilled.
"It’s so simple to construct parallel lines in monosyllabic Chinese, a child could do it as easily as Lǐ Bái!"
☝️ What you think I'm saying, but I'm not.
43/ That’s because I’ve left something out. We’ve only looked at three of the four components of parallelism!
The skill of the poet in composing these parallel lines, as well as their impact on the ear, only become apparent when we factor in the last piece: tonal opposition.
44/ You’ll recall from the thread on meter that the rhythmic feel of this poem derives from the alternation of two types of tone: long and level (平) vs. short and contoured (仄).
45/ This alternation creates a specific acoustic impression and drives the metrical pattern of each line. The reversal of those tonal categories in the paired lines of a couplet is a formal constraint that is essential to the genre.
46/ We saw that opposition at work in that previous thread on meter:
47/ 4️⃣ This is the fourth aspect of parallelism, which has to be integrated with the other three: the paired words, the ones that match each other in grammar and meaning, must have *opposite* tonal values.
Think about what that means if you're the poet.
48/ Lǐ Bái wasn’t free to pick any two color terms that conveyed the image he wanted to paint for us. He had to choose one that was Level-Tone and one that was Oblique-Tone.
49/ Same with the cardinal directions.
Same with the natural objects.
Same with the verbs of movement along a path.
And same with the types of city wall.
It's no mean feat.
50/ Now we can appreciate the incredible skill of the poets working within these formal constraints. The monosyllabicity of Classical Chinese makes frequent use of parallelism achievable; the formally constrained tonal prosody turns it into a technical and artistic challenge.
51/ The feel of this tonal opposition is, I think, simply impossible to capture using formal English linguistic structures.
I don’t just mean because English is non-tonal.
52/ I don’t think there is any kind of binary of English pronunciation that could serve as a proxy. (Maybe you could imagine attempting translations in Russian that exploit the opposition of “hard” vs. “soft”, i.e. non-palatalized vs. palatalized syllables?)
53/ It’s also impossible to capture the feeling in the reading traditions of modern Chinese languages.
Why?
54/ Because there are no modern varieties of Chinese in which all Middle Chinese Level-Tone words are pronounced long and level, and all Middle Chinese Oblique-Tone words are pronounced short and contoured.
57/ I’ll leave you to contemplate the other parallel couplet in “Seeing off a friend”. Think about both the grammatical category of each word and the meaning category of each word.
浮雲游子意,
落日故人情。
58/ I won't say more about it other than to point out that my translation once again falls a bit short because it has uneven line lengths:
Like the drifting clouds are the traveler’s intentions;
Like the setting sun are the old friend’s feelings.
59/ Even if you elide the middle vowel of “traveler”— “Like the drifting clouds are the trav’ler’s intentions”—you’ve still got one more syllable in “intentions” than in “feelings”.
60/ I can kind of get away with it if we think of terms of feet rather than syllables, because the “in” of “intentions” is unstressed.
And it can’t really ever be, because the linguistic underpinnings of this Chinese poetic device aren’t part of how English is structured.
62/ I want to end this thread by pointing out that my discussion of the first three aspects of parallelism isn’t anything special. You’ll find lots of similar explications in textbooks and scholarly articles, and in classrooms all over the world.
63/ But it’s not often you find scholars thinking or talking about how the opposed tonal patterns in the two lines of a parallel couplet create a particular effect on the ear, a clash of opposites that is juxtaposed with the harmonic blending of syntactic and semantic categories.
64/ Similarity and difference, identity and opposition, are at play together: sound and sense interweaving in patterned complexity.
65/ Last thing before I finish: a shout-out to @tommazanec, who has always paid attention to the phonological aspects of these poems, and who explains parallelism so nicely in this article. Check out his innovative English translation experiment. eastasian.ucsb.edu/wp-content/upl…
66/ I haven’t forgotten my pledge to make a recording to illustrate some of the acoustic features that I’ve discussed in these three threads. It may take a while, but I’ll get it done, I promise!
/end
• • •
Missing some Tweet in this thread? You can try to
force a refresh
It may seem obvious from the title — Chinese Characters across Asia: How the Chinese Script Came to Write Japanese, Korean, and Vietnamese — but that’s just part of the story. There’s a lot more happening in these pages.
2/
Before I got any further, let me just say that although this is a sophisticated book but it’s most definitely NOT a technical book. It’s meant to be readable and understandable by just about anybody, without being dumbed down or inaccurate.
3/
3/ Recall that in Part 1, we established that you are a linguist in the year 3022 (that's you in the image). And you are working on reconstructing Cantonese as spoken 1,000 years in your past! You’ve got an excellent textual source to help you, a dictionary.
At long last, Part 2 of this thread. We’re thinking about how much we could reconstruct of late 20th-century spoken Cantonese from a vantage point 1,000 years in the future ... if this dictionary were our only available source of information.
2/ Here’s the setup: The year is 3022, you’re a linguist, and you’ve stumbled across a precious document: a dictionary of Cantonese. The existence of the language was already known, but no direct documentary evidence was known to be extant: until now.
3/ You undertake a systematic analysis of the dictionary data. This is the book you eventually proudly publish: a reconstruction of the ancient language Cantonese from ten centuries ago!
In a thread I posted a few days ago, I explained that the Mandarin name Yālù and the Korean name Amnok not only refer to the same river, but are in fact historically the same name.
2/ One of the great things about sharing these ideas on Twitter is that more knowledgeable people point out mistakes or provide additional information.
I got some very informative feedback/pushback on the Manchu etymology: the “twist” in that thread.
1/ This is the river that divides the Korean peninsula from continental East Asia. It runs along the current border between North Korea and the People’s Republic of China.
What is its name? Depends on which side of the river you are on.
2/ When I first learned that the Yālù River and the Amnok River were the same river, I assumed that these Mandarin and Korean names must be different, unrelated names.
YALU ≟ AMNOK
3/ Later, after I’d become more sophisticated about Chinese and Korean language history, I realized that they are historically the same name: the Mandarin and Korean pronunciations of 鴨綠/鸭绿 meaning ‘duck green’.