I presume it starts with 이ㅎ... Yi H ... (Lee being the usual Anglicization of YI) but what would the other letters be?

Peter H. Lee was my first guide to the world of Old Korean poetry. I had seen part of Chhŏyongga in a book I read in high school, but I didn't see any other OK poems until I borrowed his Studies in Old Korean Poetry from the Berkeley library. I think I might have borrowed Flowers of Fire: Twentieth-Century Korean Stories during the following semester.

When trying to Google his Korean name last night, I was surprised that he had no English Wikipedia entry in spite of his importance in Korean literary studies. (I just checked - searching for his name in the Korean Wikipedia turns up nothing!) But I did find this page asking the question,

"What ever happened to Peter H. Lee?"

Gerry Bevers was a former student of Lee's:

I went to Wikipedia expecting to find an article on him, but there was nothing there, which bothered me since I feel he deserves to be recognized and remembered.

Bevers described his class with Lee:

I also remember his telling us that he learned English by memorizing a dictionary. He said that after he had memorized a page in the dictionary, he tore it out. I do not remember if he said he threw it away of if he said he put it on the wall and ceiling over his bed, but he did say that he tore the pages out.

English was just one of many languages that Lee learned. He was educated in Japanese during the colonial period and has been a professor of Japanese as well as Korean. He studied Japanology, Sinology, and German literature in Munich and French literature in Fribourg.

Back in 1990, I noticed that his 1959 OK poetry book was published by the Istituto Italiano per il Medio ed Estremo Oriente. I never guessed that he had learned Italian in Florence - or that a book of his German translations of Korean poetry was also published in 1959.

He studied spoken Chinese in Korea:

And Chinese - we had to learn to speak with the correct tones. I was not good at memorizing tones at the time [did he get better later?], but I had to memorize everything. Somehow I passed.

I presume that was after the liberation since I doubt spoken (as opposed to classical) Chinese was a middle and high school subject in wartime Korea under Japanese rule. But maybe I'm wrong.

In order for Koreans to make Korean literature known, first of all we have to make connections. That means we have to be able to point to some Western works or Chinese and Japanese works as parallels or contrasts [...]

I read quite widely from Greek and Latin literature to modern literature, so I was able to find comparable writers from ancient times [onward ...] Particularly when one is writing on twentieth-century Korean writers, you have to bring in other twentieth-century writers-Western, Chinese, and Japanese. That way the reader is better able to situate a given work: "Ah hah, this work belongs to this, it's like this," which makes the reader better prepared. I would strongly recommend this because our approach, when we wish to introduce and propagate Korean literary works, has to be comparative. If a Korean work just stands by itself, it's hard to attract the attention of Western readers because they won't have time to pick up that book unless it's distinguished in some way, by unique features or those it shares with well-known Western works.

I would add another reason for this strategy: Korean writers have never worked in a vacuum. Premodern writers knew Chinese literature. Here's Lee's description of the Korean who would have grown up to be postwar writers:

The zeal for reading was really strong at the time, particularly among the young intelligentsia. If you were on a tram or bus, you would seldom see a young man of my age group or students who were not reading. They all carried a pocket edition. These days I don't see that in Seoul. They read the newspapers and popular genres, but not Goethe, Andre Gide, or Thomas Mann, which we used to read on trams or buses. It was really amazing. The moment you got onto the tram, you saw that everyone was reading, except for commoners.

Ouch! I guess they were reading European literature in prewar Japanese translations. I can't imagine literary translation being a high priority in South Korea in the late 40s.

I'll close by quoting Lee's approach to reading literature:

Because when you pick up a volume of verse or work of fiction, you immediately grasp that a tremendous amount of suffering, hard work, and imagination went into making that book. You are even more aware of this when you yourself are a creative writer and you know how much time you spend writing a single poem. You have to go though ten or twenty revisions. You mumble a poem to yourself, even when you're walking or on the bus, because you're thinking of how to improve a particular line. That's how I feel whenever I pick up a book - I have respect for that author, I identify with him or her. My attitude is grave and solemn. I don't treat books lightly, but solemnly. It's something precious. That single book is imbued with an author's soul and blood. We cannot treat such a thing shabbily. That's the key attitude when dealing with a literary work; we respect the author who spent time, who suffered, who imagined, who wrote, and who wants us to take part in that experience.

10.2.26.23:57: WAS THE MOON GOOD IN TOKYO'S BRIGHT ERA?

If you thought "Target's Fish Man" was difficult, try the first two lines of the first Old Korean poem I ever saw back in high school in 1987, 處 容歌 Chhŏyongga 'Song of Chŏyong* (879 AD):

東京明期月良
夜入伊遊行如可

Here's a character-by-character breakdown:

My reconstruction is largely based on Kim Wan-jin's (KWJ) with some modifications: e.g., I project Middle Korean vowels back into Old Korean. The other reconstructions are by Ogura Shinpei (OS), Yang Chu-dong (YCD), Chi Hyŏn-yŏng (CHY), Kim Sŏn-gi (KSG), Sŏ Chae-gŭk (SCG), and Kim Chun-yŏng (KCY) as reproduced in Kim Wan-jin (1980: 225-226) and Yu Chhang-gyun (YCG 1994).

Peter H. Lee (2003: 73) translated these lines as

Having caroused far into the night

In the moonlit capital

And here is another translation by David McCann (1997; found in Lee and Ramsey 2000: 49):

In the bright moon of the capital

I enjoyed the night until late

Here's my more literal guess:

In the capital, under the bright moon,

Night deepens, I was carousing, and then ...

The meaning of the second line is uncontroversial. But I'm not sure about the first line.

Let's go through both lines word by word:

As a general rule, there is no way to be completely certain about the reading of any semantogram.

1.1-2: 東京 looks like Japanese 東京 'Tokyo', a name that wouldn't be used for the Japanese city of Edo until 1868. There are two schools of thought: one regards 東京 as Sino-Korean for 'eastern capital' and another regards it as a spelling for a native s-word for 'capital' (reconstructions vary, but none include 'eastern'). This s-word is the source of modern Seoul. (The YCD and CHY reconstructions of the s-word contain *-βr without a vowel as printed in KWJ 1980, but I suspect a vowel was accidentally omitted between *-β- and *-r.)

1.3-4: 明期 is generally translated as 'bright' modifying 'moon'. The k- of 期 represents the final *-k of the stem of the adjective. In theory, the Old Korean word for bright could have been a *-k stem other than the one ancestral to Middle Korean pʌrk-.

Lee and Ramsey (2000: 279) translated 明期 as a noun 'brightness' (cf. modern Korean palkki 'brightness'), so they might regard 明期月 as a compound noun 'brightness-moon'.

The -ɯy of 期 kɯy looks like a Middle Korean genitive suffix, and I wonder if at one time, the same morpheme followed both adjectives and nouns.

I don't reconstruct 期 as *-k-i since the Sino-Korean reading of 期 was kɯy. *ki was written as 只 or 支 in other poems.

1.5: A disyllabic reading of 'moon' is implied by another spelling 月羅里 with Sino-Korean -羅里 -ra-ri, presumably for *tʌlar-i. I use *l to represent a liquid sometimes lost in later Korean. The nature of the phonetic distinction between unstable*l and stable *r is unknown.

1.6: 良 was used as a phonogram for ra in later Korean (and in Japanese which presumably followed an earlier peninsular practice). ra reflects a Late Old Chinese/Early Middle Chinese reading *lɨaŋ with a vowel cluster *-ɨa- and final *-ŋ that did not exist in early peninsular language(s) or Japanese and was hence reduced to *ra. Here 良 represents the final *-r of *tʌlar 'moon' plus a suffix *-a whose context implies a locative suffix. 明期月良 by itself might mean 'on the bright moon' but since the narrator of the poem is a human in the capital, I have translated the locative suffix as 'under'.

2.1: All agree that 夜 was *pam, but a word not cognate to later Korean pam cannot be ruled out.

2.2-3: The -伊 *-i of 入伊 'enter, and' seems to correspond to later Korean -ə (Martin's 'infinitive') but cannot be cognate to it.

2.4-5: 遊行 'play [and] go' is presumably a semantographic sequence for *no(l)ni- 'carouse'. I reconstruct a potential *-l- assuming that the word shares a root with Middle Korean no(r)- 'play' (the -r- disappears in certain forms: e.g., no-taka with the same suffix discussed below).

2.6-7: 如可 *-taka is used to write later Korean -taka > -taga 'and then' (Martin's 'transferentive'). 可 ka is a straightforward phonogram, but 如 is a metaphonogram whose reading is normally assumed to be ta, as in later Korean. However, according to my understanding of Alexander Vovin's hypothesis of nonlenition, the -k- of -taka did not lenite because it was part of an earlier cluster *-taCka. The *-C- of this cluster may have been *-p- since 如 corresponded to Korean -tap- 'be like'. If -taka were simply *-taka in Old Korean, this could have been written as 多可 with 多, used to write Old Korean *ta elsewhere. In later Korean, 如 represents ta, not tap, but this p-less reading might be due to cluster reduction:

Stage 1: 如 is *tap in 如可 *-tapka

Stage 2: 如可 *-tapka > 如可 *-taka

Stage 3: 如 is reinterpreted as *ta

There is one huge problem with this hypothesis. In the poem 風謠 Phungyo 'Wind Song' from the reign of Queen 善德 Sŏndŏk (632-647), 'come' (later Korean o-[n]da) is written four times as 來如 with a semantogram 來 'come' for the stem *o- This is normally reconstructed as *ota (KSG: *onta). Should this be reconstructed instead as *otap? I doubt it, since there is no reason to believe that word-final *-p was lost in Korean. Perhaps 如 was read as both *ta and *tap depending on context. The reading *ta would be derived from the first two-thirds of *tap.

2.27.0:25: Could the currently existing transcription of 'Wind Song' postdate the 7th century and the reinterpretation of 如 *tap as *ta? If so, then a 7th century scribe might have written *ota as 來多 with 多 *ta.

2.27.1:03: Lee and Ramsey (2000: 279) were also puzzled by 如 *-ta in words like 來如:

Were the ancient Sillans transcribing some other ending that they used instead to express the declarative? [E.g., *-tap or even something sounding like zyə, the Sino-Korean reading of 如?] Or, alternatively, has the declarative ending [-ta] familiar to us today changed in pronunciation so completely over the centuries?

*2.27.0:32: Peter H. Lee (2003: 73) described the background for this poem:

Its author, Ch'ŏyong, was one of the seven sons of the Dragon King of the Eastern Sea. Our source informs us that Ch'ŏyong married a beautiful woman. Seeing that she was so lovely, however, a demon of pestilence transformed himself into a man and attacked her in her room while Ch'ŏyong was away. When Ch'ŏyong returned and witnessed the scene, he calmly sang this song, which so moved the demon that it went away.

10.2.25.23:59: TARGET'S FISH MAN (PART 2)

Here's the solution to last night's problem (rhymes in bold):

The fisherman caught 'em

The ones at the bottom

It was also in part 2 of my cuneiform series.

And here's a character-by-character breakdown:

This exercise was meant to illustrate the extreme difficulty of deciphering a text in mixed script even when the underlying languages (Mandarin and English) are well known.

A future linguist trying to reconstruct English on the basis of these lines would not be able to reconstruct the unwritten -er- of fisherman unless other spellings like

魚兒男 FISH-er-MAN with 兒 er

飛師門 fei-shi-men (roughly pronounced 'fay-sher-mun')

were found.

If the linguist assumed the lines had an equal number of syllables, he might be at a loss to figure out where the seventh syllable was.

Perhaps the most difficult words are

之母 'em

底秋 bottom

之 and 底 are followed by phonetic clarifiers:

之母: read 之 with final 母 -m: e.g., as him, them, or 'em (rather than her or it or -'s; I should've included 'em as a gloss last night - sorry!)

底秋: read 底 with final 秋 -ottom < autumn: i.e., as bottom (rather than base)

Since the fisherman caught ones (plural), that rules out him as the reading of 之母. But there is no way to tell whether the author intended 之母 to represent them or 'em.

I'm too tired to continue my cuneiform series, so I'm going to put it aside for a couple of days. In the meantime, here's a puzzle to work on that will make you understand what deciphering Old Korean is like. These are two rhyming lines of modern English written in Chinese characters:

的魚男考特之母

的一們在的底秋

What are those lines, and how did you figure out what they are? All the clues you need are below. No knowledge of Chinese is necessary. In fact, a knowledge of Chinese can be an impediment!

I'll post the answer tomorrow. Many hints are in the previous post.

10.2.24.2:00: HOW WAS CUNEIFORM DECIPHERED? (PART 2)

In part 1, I asked,

How would one know if Akkadians read [the Sumerian graph] LUGAL as ...

- lugal (a loanword from Sumerian)

- a cognate of m-l-k words in Arabic and Hebrew

- a word unrelated to lugal or m-l-k

- two or more of the above

While Googling for the Akkadian reading of LUGAL, I found this article.on 'kunogenesis' by 池田 潤 Ikeda Jun, a professor specializing in Semitic languages (paragraphing mine):

“Kunogenesis” is a neologism coined by myself. It means the emergence of kun [native Japanese translation-based phonetic] values (Phase 1) and phonetization [= desemanticization] thereof (Phase 2).

In Phase 1 of the Japanese writing system, semantic association played an essential role. It created new logograms by involving three elements: the graphic shape of a Chinese character, its meaning, and a sequence of one or more phonemes in Japanese which has the same meaning as the Chinese character. As this type of logographic-kun (with a rigid semantic link) was the original and authentic use of kun, it was called “seikun” (lit. genuine kun) in Japanese.

In Phase 2, scribes disregarded the meaning of the Chinese character and using the rebus principle applied the kun values to other homonyms and then even to any homophonous sequence of phonemes, thus achieving real phonetization of the logogram. Japanese kunogenesis culminated in creation of new phonograms based on existing logograms. Such extended use of kun was called “kungana” (phonetic kun) in Japanese. Let us call it phonographic-kun.

Let me illustrate Ikeda's points with English instead of Japanese.

Phase 0. 王 represents Mandarin wang 'king'.

Phase 1. 王 becomes a logogram for the English word king, a translation equivalent of Md wang. king 'king' is the English seikun of 王.

Phase 2. 王 becomes a phonogram for the English syllable /kɪŋ/ without any regard for semantics. Hence shocking could be written as <sho王>, even though 'shocking' has nothing to do with monarchs.

Phase 2 has already occurred in nonstandard English spellings like <b4> for before. The logogram <4> for four has become a phonogram for the English syllable /for/, even though 'before' has nothing to do with 'four'.

I'm not comfortable with Ikeda's use of the term kun because kun readings and 'kunogenesis' are not unique to Japanese. The equivalent to kun on the Korean peninsula is 훈 hun (the Sino-Korean reading of the graph 訓 for kun). I've been trying to come up with a neutral English term. For now, I'll use metaphone, short for metaphrase (Greek for 'translation') + phone.

Here's another English example of metaphonograms:

Phase 0. 秋 represents Mandarin qiu 'fall; autumn'.

Phase 1. 秋 becomes a logogram for the English words fall and autumn, both translation equivalents of Md qiu. fall and autumn are the English orthometaphonetic readings (= seikun) of 秋.

Phase 2. 秋 becomes a metaphonogram for the English phonetic sequences [fɔl] and [ɔtm]. Hence follow could be written as <秋low> and bottom could be written as <b秋>.

Imagine you are an archaeologist in a future era where English has been completely forgotten (4gotten?). You find an English text written with sinographs as metaphonograms. Can you figure out that

- 秋 had two (ortho)metaphonetic readings

- 秋 was read as [fɔl] in 秋low

- 秋 was read as [ɔtm] in b秋

What if you knew German and assumed that the metaphonetic reading of 秋 was cognate to German Herbst? Then would you assume

- 秋 had only one metaphonetic reading herbst

- 秋low was read as herbstlow

- b秋 was read as bherbst

But German has nothing like herbstlow 'follow' or bherbst 'bottom' and no other data suggests that English had an initial bh-. (2:15: And the actual English cognate of Herbst is harvest which doesn't mean 'fall/autumn'!)

You might get back on the right track if you notice that

- 秋 is also used to write 'descend'

- the first syllable of German fallen 'to fall, descend' resembles the first syllable of German folgen 'to follow', the translation equivalent of English 秋low

You then hypothesize that 秋 was read as *[fal] or *[fol], and you reconstruct 秋low as *fallow or *follow.

One correct metaphonetic reading down, one to go. You think there might be a second reading since b秋 probably wasn't *bfall with an un-English cluster *bf-. Since English generally only allows b- to be followed by a vowel or a liquid (bV-, bl-, br-), you suspect the second metaphonetic reading begins with a vowel or a liquid.

Then you find a rhyming poem with the lines

The fisherman caught 'em

The ones at the b秋

and conclude that 秋 must have a second reading sounding like -aught 'em.

She went to school in the 秋.

秋 leaves are red.

I would read 秋 as fall in the first sentence and autumn in the second, but reversing the readings would also be OK.

Next: How could the Akkadian reading(s) of the Sumerian graph LUGAL be rediscovered using similar techniques?

10.2.23.2:03: HOW WAS CUNEIFORM DECIPHERED? (PART 1)

I only have the faintest idea after reading a five-page chapter on decipherment in CBF Walker's Reading the Past: Cuneiform (1987). Four pages are spent on Old Persian, whose cuneiform script is a syllabary much simpler than Sumerian or Akkadian. Old Persian cuneiform only has logograms for five words: 'king', 'country', 'earth', 'god', and 'Ahura Mazda'.

On the other hand, Sumerian and Akkadian cuneiform are full of logograms, polyphony (multiple readings of a phonetic character), and other complex features: e.g., Sumerian ugnim 'army' was written as

KI.SU.LU.ÚB.GAR

with graphs literally meaning 'place where the water bag is put' (Cooper 1996: 42). How do Sumerologists know that KI.SU.LU.ÚB.GAR didn't simply represent ... kisulubgar?

Here's a simpler question: how do they know the readings of logograms? One could make reasonable guesses about the readings of Akkadian logograms on the basis of other Semitic languages, but that assumes those readings had cognates.

The Akkadian word for 'king' was written with a logogram LUGAL for Sumerian lugal 'king'. How would one know if Akkadians read LUGAL as ...

- lugal (a loanword from Sumerian)

- a cognate of m-l-k words in Arabic and Hebrew

- a word unrelated to lugal or m-l-k

- two or more of the above

(I'll deal with the real answer to that question tomorrow.)

Worse yet, one cannot use cognates to guess the readings of Sumerian logograms, since Sumerian is an isolate without any relatives.

The final page in Walker's (1987: 52) decipherment chapter barely touches such questions. The decipherment of Old Persian was just one step on a "torturous path". Even bilinguals are not clear keys: e.g., the Old Persian and Babylonian spellings of Darius are not equivalent:

OP <da-a-ra-ya-va-u-ša> for Daarayavauš

B <da-ri-ia-a-muš> for Dariamuš (the B graphs for da and a are not the same as in OP!)

I'm curious because I think it is impossible to be certain about the readings of some Old Korean and Khitan logograms at present. How do we know that Old Korean 春 'spring' was read as pom, as in modern Korean, rather than as, say,

- a loanword *tshyun (< Late Middle Chinese *tshün 'spring')

- an ancestor of pom like *pam (reconstructed by Kim Sŏn-gi as reproduced in Kim Wan-jin 1980: 218) or even *palum (resembling Old Japanese paru 'spring')

- a word without any later Korean cognates

We think that Khitan large script 'five' was read *tau since it was used to write the Chinese syllable 討 *thaw. The use of its small script equivalent for the first syllable of 'hare' also points to *tau:

<tau-li-a> *taulia (cf. Classical Mongolian taulai 'hare')

Note that a knowledge of Mongolic alone cannot lead to the right reading of or , since *tau has lost the -b- and -n- still preserved in CM tabun (though CM is newer than Khitan!). And Khitan *taulia is not completely identical to CM taulai 'hare'.

But many Khitan large script readings are unknown. As 'five' and 'hare' demonstrate, one should not simply assume that 'white' (reading unknown) was read as *caɣan, as in Classical Mongolian. It could be an entirely unrelated word, just as *liauqu 'red' is not cognate to CM ulaɣan 'red'. The reading of 'red' can be confirmed by its small script spelling

<l-iau-qu>

Were the readings of Sumerian and Akkadian logograms confirmed through alternate phonetic spellings: e.g., was the logogram LUGAL 'king' ever spelled phonetically as LU.GAL? Are there cuneiform logograms which have no alternate spellings and hence no definite readings?

"Two previous attempts [to read a a tablet for making beer], by J.D. Prince in 1919 and M. Witzel in 1938, had produced less than satisfactory results. A line that now even a first year Sumerian student will translate "you are the one who spreads the roasted malt on a large mat (to cool)," was translated "thou real producer of the lightning, exalted functionary, mighty one!" by the first author, and "stärkest du mit dem Gugbulug(-Tranke) den Gross-Sukkal" ["strengthen thou with the Gugbulug (drink) the large Sukkal"] by the second."

10.2.21.22:27: CUNEIFORM AND KOREOGRAPHY

This passage in CBF Walker's Reading the Past: Cuneiform (1987: 16) reminded me of the complexity of deciphering Koguryo place names like 滅烏 'destroy crow':

[...] Akkadian-speaking scribes used Sumerian signs to express Akkadian terms, e.g., Sumerian udu-meš for Akkadian immerū 'sheep', or mixed the two, e.g., Sumerian gal = 'great', but gal-u = Akkadian rabū 'great'. (For clarity Assyriologists write Sumerian in normal script or capitals and Akkadian in italics.)

The example also reminds me of Old Korean and Japanese combinations of semantograms with phonograms:

OK 二尸 'two' (written as Chinese 二 'two' plus Chinese 尸 'corpse' for the end of the native Korean word presumably corresponding to the -l of later Korean tul 'two')

Jpn 二つ futatsu 'two' (written as Chinese 二 'two' plus the hiragana つ)

(22:42: Cf. the use of -nd in English to clarify that 2 is read as seco- rather than as two in 2nd.)

OK -尸 and Jpn -きい are clarifiers like the -u in Akkabian rabū 'great'. The Japanese use of clarifiers is almost certainly a carryover from earlier Korean peninsular writing.

Thus the Sumerian sign á 'hand' corresponds to Akkadian idu 'hand'; hence the sign comes to be used for the [Akkadian] syllable id, and also for it, iṭ, ed, et, and eṭ. (Walker 1987: 16)

Over twenty different signs can be read /du/ [in Sumerian], and the sign KA has several times the half-dozen readings mentioned above [zú 'tooth', kir₄ 'nose', inim 'word', gù 'voice, sound', and dug₄ 'to say']. (Cooper 1996: 43).

Similarly, although a small amount of polyphony exists in the early [Akkadian] syllabaries, it is only in the first millennium that it becomes rampant: KUR can be read mad, nat, lad, šad, sad, kur; UR is ur, lik, tàn, taš/s, tís; SAL is s/šal, rag, mim, mám, and so on. (Cooper 1996: 47; cf. the polyphony of Tangut phonetic elements.)

Japanese writing is also complex*, but at least its users are alive to confirm how to read graphs.

Unfortunately, no literate Koguryo are around anymore to tell us how to read 滅烏 'destroy crow'. Worse yet, we have no running texts in the languages of Koguryo. Although 烏 is a relatively frequent character in Koguryo names (implying that it stood for a simple syllable and/or had multiple readings), Song (2004: 466) lists 滅 in only two Koguryo names, the place name 滅烏 and the personal name 寂滅 'tranquil destruction', a Chinese term for nirvana (and hence not a native Koguryo word).

*The simple-looking graph 一 'one' is difficult to read in Japanese. It can represent ichi, itsu, iC (C = consonant determined by folllowing consonant), hito, kazu, etc.