Language Log

« previous post | next post »

[This, a guest post by Lañitri Kirinputra, is the fourth and last in a series of four posts on Hokkien and related Southern Min / Minnan language issues.  The first was "Eurasian eureka" (9/12/16), the second was "Hokkien in Singapore" (9/16/16), and the third was "Hoklo" (9/18/16).]

Hokkien¹ Creationism

The romanization debate was hot around 2005-06. Should we go on using the one- and-a-half-century-old Pe̍h-ōe-jī (“POJ”) scheme? Or should we use a modified flavor of POJ? Or should we use another more Pinyin-like scheme? The POJ camp was going on and on day after day about how POJ had history, literature, yada yada yada.

“Brothers and sisters,” I said one day on a listserv, “let's face it. Only Presbyterians used POJ. POJ was small-time. Written Taiwanese was small-time. We ain't inheriting Written Taiwanese. Ain't nothing to inherit. We're creating Written Taiwanese.”

A tree falling in the forest. Nobody paid me no mind. Over the years, though, it came back to haunt me: how wrong I was that day. Not just me, but two generations of “Hokkien Creationists”, most of them too mild or confused to come out and say what they all believed, which was that there had been no Written Hokkien before them — or, if there was, it was invalid or incomplete or both.

You know about Pe̍h-ōe-jī

In case you don't, Pe̍h-ōe-jī (白話字) is a Hokkien romanization scheme² — the only one that's ever caught on at scale as a means of communication. By the mid 20th century, it was used by probably over 100,000 Hoklophones on both sides of the Taiwan Strait and around the Straits of Malacca. It faded in the second half of the 20th century with the rise of nationalistic compulsory education pretty much everywhere. Republic of China (“ROC”) offcials on Taiwan went the extra mile, banning the publication of POJ materials and raiding churches to confiscate existing publications in POJ.

The Taiwanese Creationists of the last 30-some years obviously knew about POJ. Many or most of us knew it too. The project to re-invent Written Taiwanese was no knock on POJ. It just kind of went without being said: no matter the merits of POJ, Hokkien had to at least have an alternate form that was based on Hàn-jī (漢字, Jp. Kanji, hereinafter “Hanji”). “A language like Hokkien” couldn't live on romanization alone. Till about 2005, there was a loud, proud POJ-only camp that argued against this idea, but let's put this issue aside for now. What I want to talk about is the Hanji-based Written Hokkien convention of pre-ROC-on-Taiwan times.

What people think they know about pre-ROC Hanji-based Hokkien

Anybody who sets out to learn or learn about Hanji-based Hokkien these days will get the impression that pre-ROC Hanji-based Hokkien was, at best:

1. Utterly unstandardized. To each man his own version. An unholy mess. Internally inconsistent.
2. Largely made up of sound-only Hanji, or what the Japanese call ateji (当て字): Hanji employed for their sound value, with no regard to their “underlying meaning”.
3. Well-represented in the experimental mess that was Hanji-based Hokkien in the 1980s and 90s.
4. Reasonably well-represented in a series of 19th and 20th century dictionaries and rimebooks that allow us to access a cleaned-up version of pre-ROC Hanji Hokkien without having to look at any actual writings.
5. Reasonably well-accounted for by the scholars of the late 20th century, who used the pre-ROC orthography as a starting point for the Neo-Hokkien they were creating, only replacing the parts that were unworkable.

This is what I used to think too, pretty much, before I found Tō͘ Kiàn-hong's (杜建坊) essays. From there I looked for and found dozens of pre-ROC Hokkien publications that had been uploaded to the internet. It was a rude awakening. Pre-ROC Hokkien was surprisingly consistent. It spanned 400 years. The later stuff showed a level of orthographic polish and de facto standardization comparable to the Cantonese of today. Most of the late 20th century orthographic mess was not to be found in the pre-ROC publications. (Most of the mess that the Taiwanese Creationist scholars and writers claimed to be cleaning up had in fact been introduced by their own selves!) In turn, the conventions in the pre-ROC publications were only partly represented in the dictionaries and rimebooks of their time. And while scholars of the last 30-some years often cited “popular” (民間 bîn-kan) usage, it turned out what they meant by “popular” most times was in fact select dictionaries and rimebooks instead of popular … usage.

The Lâm-koán Play Scripts

The oldest piece of Written Hokkien known to modern man is a 1566 edition of Nāi-kèng-kì (荔鏡記, Tale of the Lychee Mirror). The full title of this edition indicates it was a “re-publication” (重刊 têng-khan), so we know the tradition goes back even further than that. How much further is hard to say. As far as orthography goes, Nāi-kèng-kì falls within a genre of plays called Lâm-koán-hì (南管戲, “LKH”) because they're set to Lâm-koán music. The internal consistency of LKH orthography — and the consistency of the 1566 Nāi-kèng-kì orthography with the centuries of LKH orthography that came after — suggests that LKH-style Hokkien orthography was old hat by 1566.

LKH scripts were written³ in a high vernacular register⁴ of the dialect of inner city Choân-chiu (泉州, Mand. Quánzhōu), with lots of Classical Chinese elements thrown in. Nāi-kèng-kì also had a lot of lines written in Teochew,⁵ as well as some lines — belonging to the mandarins, of course — in a hybrid of Mandarin and Hokkien or Teochew.

LKH scripts were targeted at actors and backstage personnel. They were never meant for mass consumption. Still, for a written language to debut at such a non-mainstream venue might be par for the course. Timing-wise, LKH Hokkien could be seen as part of a wave of vernacular writing that came up in the Sinosphere mid-millennium, maybe as a trickle-down response to a greater emphasis placed on vernacular writing by the Mongols when they ruled China.

The Age of the Koa-á-chheh

Koa-á-chheh (歌仔冊, “KAC”) are a type of rhyming storybook, typically seven or eight dense pages, sometimes a lot longer or a lot shorter.⁶ KAC started with Siù-siōng Ông Chhau-niû sin-koa (繡像王抄娘新歌) in 1826. KAC published in Amoy and Shanghai (for the Taiwan Strait market, no doubt) tended to be re-tellings of well-known epics and folk tales, but Taiwanese KAC writers in the Japanese era branched off into all genres of fiction and even non-fiction.

KAC peaked as a cultural phenomenon in Taiwan from the 1920s to about 1960. At one point KAC were probably meant to be read aloud by a raconteur to an audience. By the 1920s and 30s in Taiwan they were being mass-marketed and sold direct to “the man on the street” for him to read through on his own. Two changes made this possible. First, new technologies made it cheaper to print things. Second, basic education in Classical Chinese (漢文) became way more widespread under Japanese rule — and for any fluent Hokkien speaker with a working knowledge of Classical Chinese based on Hokkien readings, KAC is an easy hack.

Pre-ROC-style Hokkien as a society-wide phenomenon

Orthographically speaking, LKH and KAC were not discrete phenomena with no relation to society. KAC clearly inherited the orthographic conventions of LKH — deliberately, in fact: a lot of early KAC were re-makes and extensions of Nāi-kèng-kì. In turn, mid-20th-century Taiwanese pop lyrics were inherited from KAC, at least sometimes. The orthography of Hokkien-language place names in Taiwan, southern Hokkien (福建, Mand. Fújiàn), and Southeast Asia also operated on the same logic. It was really one orthography, one unstandardized convention. To this we could add shop signs and menus, to some extent.

How does Creationist Hokkien differ from pre-ROC Hokkien, orthography-wise?

The tree I'm trying to bark up here is the threshold question of why or how is it okay to do away with and replace a long-running orthography (while pretending to uphold it). I'm not trying to argue the “merits” of one approach over the other, as if Hokkien had never had a Hanji-based orthography before and we were interviewing candidates. At the same time, though, the nature of the Creationist orthography tells us a lot about why Creationists felt the need to re-do Hokkien orthography in the first place.

Taking the ROC MOE recommendations as representative of Creationist orthography and LKH and KAC writing as representative of pre-ROC orthography, the main difference is that pre-ROC writers were Hokkien-centric. They used Hanji (and locally created Hanji-oids) to express Hokkien in writing. The life and times of Hanji in Classical Chinese or other languages was secondary. Creationists, on the other hand, are subconsciously Zhōngwén⁷-centric. For them, the central question of how to write any given morpheme is, “Which Zhōngwén morpheme is this cognate to?” They tend to not ask whether the morpheme has a “Zhōngwén” cognate — they have a religious⁸ belief that every Hokkien morpheme does, in practice, unless it's a modern non-Sinitic loan.

As a group, LKH writers — KAC writers, not so much — showed tremendous sophistication not only in recognizing Hokkien-Classical Chinese cognate pairs, but in “laying off” false cognate pairs. Creationists, though, flock to false cognate pairs because they believe a false pair is better than no pair at all.

Pre-ROC writers made heavy use of sound-only Hanji (ateji). Creationists are allergic to sound-only Hanji. Their main complaint about pre-ROC Hanji-based Hokkien is that it uses too many sound-only Hanji. It's true that sound-only Hanji water down the graph-to-meaning relationship of Hanji… Pre-ROC writers didn't care too much, but Creationists strongly object.

Pre-ROC writers also made heavy use of meaning-only Hanji, analogous to Japanese kunyomi (訓読み). Creationists are also allergic to most meaning-only Hanji. Their objection seems to be that meaning-only Hanji water down the graph-to-sound relationship of Hanji. This is true if we look at things Zhōngwén-centrically or Middle Chinese-centrically, but that's exactly how most Creationists look at things, subconsciously.

Creationists' rejection of both sound-only and meaning-only Hanji at the same time brings to mind the Mandarin saying Yòu yào mǎér pǎo, yòu yào mǎér bù chī cǎo (又 要馬儿跑、又要馬儿不吃草): wanting the horse to run, run, run but never graze. You can't have it both ways. If a Hokkien morpheme doesn't have a Chinese cognate, assuming we can't coin new Hanji-oids at will, we then have to use either a sound-only Hanji or a meaning-only Hanji. The only alternative would be to inject non-Hanji elements like POJ into the text — another idea that a lot of Creationists recoil from.

Cornered by their own prejudices, Creationists typically respond by making sure every Hokkien morpheme has a Chinese cognate! If a solid cognate can't be found, they just take the most likely one — no matter how unlikely — and make believe it's The One.

Creationists hate meaning-only Hanji because of their cognancy fetish. Every meaning-only Hanji is like a white flag saying, “Hey, look, they couldn't find a Chinese cognate for my morpheme.”

Another dimension of difference is that some Creationists expect to be able to read Hokkien fluently based on their ability to read Mandarin fluently. This is based on the old belief that Hokkien, Mandarin, etc. are not really different languages. For this reason, paradoxically, meaning-only Hanji based on Mandarin — not the usual Classical Chinese — were introduced.

For example, the Creationists brought in 才 from Mandarin to write Hokkien chiah (meaning "only if; just now", etc.). In LKH and KAC, this word was written with the sound-only Hanji 即 over 95% of the time — none of the famous pre-ROC lack of uniformity. What's more, 才 itself was not only a meaning-only Hanji, but, internally to Mandarin, it was also a sound-only Hanji. All this was forgiven in order to make Hokkien more Mandophone-friendly.

Again, the Creationists brought in 的 from Mandarin to write Hokkien possessive marker and nominalizer ê, a word that was consistently written either 个 or 兮 pre-ROC. Internally to Mandarin, 的 was also a sound-only Hanji.

To zoom back out a little, another pattern of difference between pre-ROC orthography and Creationist orthography is that the former favors simpler, lighter, more common Hanji while the latter favors complicated, stroke-heavy, rarer Hanji. The pre-ROC orthography was like Japanese and Classical Chinese in tending to use light, simple Hanji for function words. The Creationist orthography is like Mandarin or Cantonese orthography, with dense, convoluted Hanji sometimes being assigned to the most common, most unassuming function words.

It's an ROC thing

By and large, Creationist Taiwanese is clearly a creature of the ROC. And the way it emerged — as if it was the frst real Hanji-based Hokkien orthography ever, as if there had been no Hokkien orthography before ROC times — is the classic pattern in the modern ROC-on-Taiwan where everything that came before Chiang Kai-shek is ignored, and the phrase “… in the history of Taiwan” redirects to “… in the history of Taiwan post-1949”.

Ironic, if we consider the political views of most Creationists.

“Correcting” popular usage

In 17 years of learning and using Hokkien the Creationist way (but with a wicked linguist's twist), I came across these recurring “errors” that cropped up when the general public wrote in Hokkien. It turns out that some of these “errors” were just instances where pre-ROC Hokkien had something written one way, and the Creationists just decided it should be written a different way.

For example, to this day the general public tends to write gín-ná (meaning CHILD) as 囝仔. The scholars say it should be 囡仔. Their disciples tend to correct 囝仔 as if it was a garden variety error. In fact, 囝仔 goes back at least as far as the earliest KAC. It's the use of 囡仔 that's questionable on several fronts. It reflects not some deeper linguistic truth but rather an obsessive-compulsive need to bring Written Hokkien in line with the footnotes of famous essays and dictionaries scribed in Hángzhōu and points north.

KAC and pre-ROC-style Written Hokkien nearly went extinct in the 1960s, as the ROC labored to replace Taiwanese cultures and languages with ROC culture and Mandarin. It's amazing that pre-ROC usages like 囝仔 are still in the collective memory to this day.

What conspiracy?

For the record, I kind of doubt any scholar of Hokkien would've collaborated on the eclipse of pre-ROC Written Hokkien on purpose. Taiwanese people born in the 1940s and 50s were the first generation of people anywhere to get clusterfucked full-strength by the ROC propaganda machine from the first day of school on. They were misled themselves from the start. The other side of the strait was not good times either. Let's leave it at that.

Meanwhile, the age-old tendency of landlord-scholars to only cite other landlord- scholars probably goes a long way toward explaining the “popular usage” bait-and- switch described earlier.

Why does any of this matter?

So now the ROC state is sponsoring one Creationist orthography while the pre-ROC orthography that got us from the mid 16th to the mid 20th century and was used to name cities and rivers is cast away while we make believe that all its worthwhile elements have already been incorporated into the Creationist orthography. Why does this matter? As I see it —

First, the aesthetic downgrade. Admittedly subjective.

Second, the Creationist orthography is in effect seeded with thousands of false clues regarding the linguistic history of Hokkien.

Third, the pre-ROC orthography is Hokkien-centric. The Creationist one is not.

Fourth, the sound-only Hanji used in the pre-ROC orthography are essentially Hokkien manyōgana (万葉仮名) — the precursor to hiragana. An aesthetically Hanji-compatible alphabet or syllabary is exactly what Hanji-based Written Hokkien needs going forward. The Creationist orthography moves away from this toward a Middle Chinese-centric phonological paradigm.

Fifth, most languages are written the same way they were written last year, and the year before that, and so on. Why should Hokkien be the exception? What justifies a reset? This could be a non-rhetorical question.

For anybody that's just interested in Hokkien orthography for its own sake, I say look where you may not have known to look.⁹

---

Notes

1. Come, let's use “Hokkien” and “Taiwanese” interchangeably here.
2. Near-identical schemes also called “Pe̍h-ōe-jī” or the equivalent were introduced for Teochew, Hakka, and other languages too, to be precise.
3. LKH is alive as a performance art, but no new plays have come out in over 100 years.
4. This might sound like an oxymoron to many modern minds. The high register of modern Hokkien (or Cantonese) is non-vernacular, and the vernacular register is low, all of this almost “by definition”. A decent modern example of a high vernacular might be the language of Vietnamese pop. It's worth keeping in mind that Hokkien and Vietnamese were “the same kind of animal” for much of their history. Many of the “obvious” differences are artifacts of the last 120 years or so of politicking.
5. Hokkien and Teochew may not have been as distinct from each other 450 years ago as they are today, either in terms of “raw” mutual intelligibility or in terms of social perception.
6. So how are koa-á-chheh (歌仔冊) and koa-á-hì (歌仔戲) related? Short answer: they're not.
7. What is Zhōngwén (中文) anyway? Apparently: a combination of Modern Standard Chinese (Mandarin), Classical Chinese, and all languages past or present spoken within the borders of modern China that either use(d) a Hanji-based orthography or have (had) a diglossic relationship with the Hanji-based written form of some other language. This concept didn't “crystallize” till after the rise of Zhōnghuá (中華) ideology in the late 19th or early 20th century. Zhōngwén is usually translated as “Chinese” in English, but sometimes it's translated as “Mandarin” because that's its core element. “Zhōngwén” is where the “zh” in all them language codes comes from, by the way.
8. This is more than a figure of speech. The core of Hoklophone religious belief is patrilineal ancestor worship. Almost all Hoklophones trace their patri-lineage back to some mythic once-upon-a-time in the Chinese heartland. No surprise that so many Hoklophones try to trace all or almost all the words of their language back to the same place.
9. You can read a good many koa-á-chheh here: http://cdm.lib.ntu.edu.tw/cdm/search/collection/kua-a-tsheh
   You can get to know the pre-ROC Hokkien orthography word by word or Hanji by Hanji at my Tumblr: http://sui.tawa.asia/

---

Illustrations

A page of a KAC.

A page of POJ.

Lee Kuan Yew's notes for one of his early speeches
in Hokkien, in POJ with Mandarin glosses.

September 20, 2016 @ 5:50 pm · Filed by Victor Mair under Topolects, Writing, Writing systems

Permalink