Thread by @ZevHandel

1/ Does all this historical linguistics stuff I keep posting about have any actual application—aside from being inherently interesting? 🙊

A thread on how linguistics can—and should—inform our understanding and appreciation of ancient Chinese poetry.

Lets' go! ⏩

2/ I’m a linguist, so I find discussions of language and script inherently of interest. But they can also inform our understanding of many other humanistic aspects of our past, so in this next thread I’d like to offer a glimpse of the utility of my field for analyzing poetry.

3/ I say “thread”, but it will probably have to be a series of threads 🧵🧵🧵, because a linguistically informed introduction to even a single short Tang poem is a surprisingly intricate endeavor.

(Even when over-simplified for Twitter.)

4/ The great Qing dynasty phonologists in China and the great Indo-European historical linguists in Europe in the 18th and 19th centuries were motivated by philological goals: to better understand, appreciate, and interpret ancient texts—their form and their content.

5/ In recognition of that tradition, I’d like to demonstrate how an understanding of Chinese language structure can inform our appreciation of Tang poetics.

6/ And along the way we’ll debunk the myth that some modern Chinese reading traditions of classical poetry adequately capture the acoustic structures of Tang poetics.

(I’m looking at you, southern Chinese languages!) 👀

7/ Our subject will be

送友人 (Mand. Sòng yǒu rén | Cant. Sung3 jau5 jan4 )
by the great Tang poet 李白 (701–762)
(Mand. Lǐ Bái / Lǐ Bó | Cant. Lei5 Baak6)

a specimen of 五言律詩 (five-character-per-line regulated verse).

8/ Before diving into details, let’s take note of something peculiar about the contemporary academic study of Chinese poetry.

Poetry is an inherently acoustic medium. 👂 The impact of many poetic devices is directly connected to how they strike the ear rather than the eye.

9/ This essential acoustic property is reflected in the fact that many poetic traditions, including that of China, are intimately tied up with oral recitation. 🗣➡️👂

10/ It is impossible to imagine academics seriously investigating the form and function of ancient poetry without considering its contemporary pronunciation. How could Beowulf's poetics be analyzed or appreciated without considering how it sounded at the time of its composition?

11/ And yet, amazingly, ancient Chinese poetry is typically approached through modern character readings that bear only the most tenuous connection to its composer's and readers'/listeners' ancient pronunciations.

12/ We know *why* this is the case: the nature of Chinese writing easily lends itself to anachronistic systems of pronunciation.

13/ When we see in Beowulf that the word for ‘sword’ is written “sweord” and for ‘three’ is written “þrēo”, we can tell immediately that they must have had different pronunciations. But “劍” and “三” offer no such visual clues, and no obstacle to reading them as jiàn and sān.

14/ So one can certainly understand why ordinary people would read ancient Chinese poetry in modern pronunciation. After all, historical Chinese phonology is a highly esoteric field, and we wouldn’t expect non-specialists to have access to it.

15/ Yet it should shock us a bit that scholars of medieval and ancient Chinese poetry do not typically pay much attention to reconstructed pronunciations. Yes, one can drily appreciate the algebraic quality of the ancient rhyme and prosodic patterns, but ...

16/ this is a far cry from knowing how these things actually sounded to the ear. And lacking that understanding, how can we appreciate the way those aural qualities interacted with the meaning and emotional impact of the poetry?

17/ It is often remarked that modern southern traditions of poetic recitation are closer to ancient pronunciations, and therefore superior to, recitation in Mandarin. Yet this is true only to a limited extent at best.

18/ There are no modern Chinese reading systems that are not significantly changed from the northern Chinese literary pronunciations of over a millennium ago.

19/ I’ll show how this poem by Lǐ Bái has features that can only be appreciated through Middle Chinese recitation, by way of comparison with Cantonese and Mandarin.

(It’s my intention to eventually produce a short video with audio recordings to supplement this thread. 🎤)

20/ Classical Chinese poetry makes use of many poetic devices that are found in poetic traditions around the world and are familiar to all of us: rhyme, assonance, consonance, sound symbolism, metaphor, allusion, rhythm, and metered structures (feet, lines, stanzas, etc.).

21/ But I want to focus on some aspects of the Chinese poetic tradition that are directly connected to linguistic features of medieval Chinese—features that are not found in all human languages and therefore don’t play a role in all poetic traditions.

22/ These linguistic features are *tone* and *monosyllabicity*.

They affect *rhyming*, *rhythm*, and *parallel couplets*.

Before we get into all that, here is the full poem in Chinese characters, followed by a translation and pronunciations in Mandarin and Cantonese.

23/

送友人

青山橫北郭，白水遶東城。
此地一為別，孤蓬萬里征。
浮雲游子意，落日故人情。
揮手自茲去，蕭蕭班馬鳴。

24/
Seeing off a friend, by Lǐ Bái (701–762)

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.

25/

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 whinnies of our separating horses.

26/ Here’s the pronunciation in Mandarin:

Qīng shān héng běi guō, bái shuǐ rào dōng chéng.
Cǐ dì yī wéi bié, gū péng wàn lǐ zhēng.
Fú yún yóu zǐ yì, luò rì gù rén qíng.
Huī shǒu zì zī qù, xiāo xiāo bān mǎ míng.

27/ And here in Cantonese:

Cing1 saan1 waang4 bak1 gwok3, baak6 seoi2 jiu5 dung1 sing4.
Ci2 dei6 jat1 wai4 bit6, gu1 pung4 maan6 leoi5 zing1.
Fau4 wan4 jau4 zi2 ji3, lok6 jat6 gu3 jan4 cing4.
Fai1 sau2 zi6 zi1 heoi3, siu1 siu1 baan1 maa5 ming4.

28/ Let’s look first at the rhyme scheme. The poem consists of eight lines organized into four couplets. The last characters of each couplet rhyme. These four rhyming characters are:

城
征
情
鳴

Let's look at their pronunciations in detail:

29/
Key: Mandarin | Cantonese | *Baxter's Middle Chinese

城 chéng [tʂʰʌŋ³⁵] | sing4 [sɪŋ²¹] | *dzyengA [dʒēŋ]
征 zhēng [tʂʌŋ⁵⁵] | zing1 [tsɪŋ⁵⁵] | *tsyengA [tʃēŋ]
情 qíng [tɕʰiŋ³⁵] | cing4 [tsʰɪŋ²¹] | *dzjengA [dzjēŋ]
鳴 míng [miŋ³⁵] | ming4 [mɪŋ²¹] | *mjaengA [mjǣŋ]

30/ Based on the Mandarin readings, you might think the first two characters in -eng form one rhyme, and the last two characters in -ing form a distinct second rhyme. But that’s not the case: all four are intended to rhyme together here, as the Cantonese readings in -ing suggest.

31/ You’ll notice that in Baxter’s Middle Chinese reconstruction, the last character 鳴 has a different vowel from the other three ("ae" instead of "e"). But Baxter’s Middle Chinese is a 6th-century reconstruction.

32/ By the 8th century, when Lǐ Bái was writing, these vowels had merged, as indicated by annotations in rime books that words like 鳴 should be “used together” (同用) with words like 情 when selecting rhyming words for poetry. We'll need to adjust the reconstruction accordingly.

33/ This vowel merger that took place over a millennium ago is the reason why the words 情 and 鳴 rhyme today in Mandarin and Cantonese.

But this is all by way of preliminary explanation. We're about to get to the first key point about Chinese tone and poetry:

34/ In traditional Chinese poetry, tone is part of rhyme.

Rhyming is based on identity of pronunciation from the main vowel through to the end of the syllable.

*And that pronunciation includes the pitch pattern carried by the vowel, i.e. the tone.*

35/ So for two words to rhyme for Lǐ Bái, their tones had to be identical. Note that all four words have the 平 ("even" or "level") tone of Middle Chinese, labeled “A” in Baxter’s reconstruction and marked with a macron “¯” in my phonetic representation, e.g. 征 *tsyengA [tʃēŋ].

36/ Note that the four words do not have the same tone in Mandarin. Nor do they have the same tone in Cantonese. Or, in fact, in any of the scholastic reading traditions used in China today (so far as I am aware).

37/ The auditory effect of the original rhyme—including the repetitive alignment of tone—is absent when this poem is read aloud in anything other than a reconstructed Middle Chinese pronunciation.

38/ Gotta stop here for now.

Hahah, this was the simple one! The next two—poetic meter based on tonal prosody, and parallelism of lines in couplets—are more complex (and more interesting) aspects of poetic structure. I’ll tackle those in future threads.

39/ And, at some point, I’ll get some recordings up so you can *feel* in your body the effects that are hard to grasp based on my written descriptions alone.

/end of part 1