Thread by @formosaphile

How many competing writing systems can one language have? Five? Ten? A hundred? I've been working all hours recently, so to switch modes I thought I'd do a long, very nerdy thread on Hokkien and how history has led to a proliferation of ways to write it. 🧵👇 1/34

But first, English. English spelling is famously a hot mess. In each of the words though, tough, trough, through, thought, thorough, bough, hough, hiccough, and lough the "ough" is pronounced differently (at least for me; some English dialects merge some of these). 2/34

Lots of people have proposed fixing this shambles, ranging from the modest changes introduced to American English by Noah Webster to the wholesale invention of new scripts like Shavian (the photo is from my copy of the 1962 biscriptal edition of Androcles and the Lion). 3/34

Webster aside, most of these proposals started off on the fringe and only got fringier. English orthography has changed little in the 267 years since Samuel Johnson's dictionary was published. But what happens when you have a language without an established standard? 4/34

Hokkien – a Chinese language spoken by tens of millions across Taiwan, Fujian, Penang, Singapore, and elsewhere – is the official language of nowhere. When writing a school essay or filling out government forms, Hokkien speakers use Standard Chinese, or Bahasa, or English. 5/34

It's also called different things in different times and places. Amoy, Taiwanese, Lan-nang, Holo, Hoklo, Minnan, Ban-lam, Tai-gi, Tai-gu, Tai-oan-oe, Fujianese and more besides. The names are a clue as to how dispersed and contested the Hokkien-speaking peoples are. 6/34

Hokkien has been routinely suppressed and denigrated. Speakers have been fined, beaten as schoolchildren, told over and over that their language is coarse and lowly and will stop them getting a job, or going to university, or that using it makes them suspect citizens. 7/34

Here's a screenshot from the 1996 film Naughty Boys and Soldiers (狗蛋大兵) which shows a common scene from school in the 1970s. The boy is wearing a board that says "I love to speak the National Language" (i.e. Mandarin) because he has been caught speaking Taiwanese. 8/34

Much of this will be familiar to speakers of marginalised languages around the world, be it Nahuatl, Irish, Tamil, Dungan, Paiwan, Sámi, Choctaw, or one of thousands of others. But none of these (as far as I know) has the wild variety of writing systems that Hokkien does. 9/34

As a Chinese language, perhaps the obvious place to start is with Chinese characters. After all, Hokkien and Mandarin are both descended from Old Chinese, which had a fully-fledged writing system well before it fractured into all the Sinitic languages we have today. 10/34

There are two major problems with using characters. The first is words that have no cognates in either Mandarin or Old Chinese. Perhaps 15% of the words in Hokkien (by frequency) fall into this category. The second is is historical rates of illiteracy in Fujian and Taiwan. 11/34

So what to do about the missing 15%? Four approaches are used: substitute characters from Standard Chinese with either a similar meaning or similar sound, inventing new characters, or filling the gaps with romanization (as in this image from A-chhûn, by Babuja A. Sidaia). 12/34

A good example of popular Hokkien writing in characters – dating back to at least the 1820s – is Koa-á-chheh (歌仔冊, "songbooks"). These were normally cheap chapbooks and the non-standard characters used varied from book to book (the one in the photo is from 1936). 13/34

As for illiteracy, nineteenth century Western missionaries wanted their converts to be able to read scripture, so they developed a romanization system called Pe̍h-ōe-jī. It remains the most successful romanization for Hokkien, with around 100,000 literate users at its peak. 14/34

Fun fact: Taiwan's first printed newspaper wasn't in Chinese characters. The Tâi-oân-hú-siâⁿ Kàu-hōe-pò (臺灣府城教會報, Taiwan Prefectural City Church News) was first published in 1885, and printed in Pe̍h-ōe-jī. Pictured is an edition from 1935 after the name changed. 15/34

Now, under the latter stages of the Japanese colonial government in Taiwan, and for much of the martial law era which followed, written Taiwanese Hokkien was either banned or strongly discouraged. Pe̍h-ōe-jī, which had never spread beyond a minority, went into hibernation. 16/34

Japanese linguists developed their own system for indicating Taiwanese pronunciation based on katakana, but it never became fully-fledged orthography – the primary use was in dictionaries and language textbooks as ruby text. Image is from the 1932 dictionary 台日大辭典. 17/34

Similarly when the KMT arrived in Taiwan after World War II they adapted Zhuyin Fuhao (Bopomofo) to use for Taiwanese and Hakka (the other major Chinese language in Taiwan). For Taiwanese this involved adding some new characters (in green here) and tone markings. 18/34

Meanwhile, across the strait in Fujian a new romanization system was developed in 1982 at the University of Xiamen, based on Hanyu Pinyin (the standard Mandarin romanization). Those of you familiar with Pinyin will see the resemblance in this dictionary entry for "tiger". 19/34

So this takes us up the late 1980s, and the end of martial law in Taiwan. Once restrictions are lifted there's a creative explosion in the small realm of written Taiwanese – dozens and dozens of new writing systems proposed within the next decade. 20/34

Most are new romanizations, and a lot of these are quite similar – some use "ts" for one sound, others use "ch" etc. Some of the more unusual ones are interesting in their own right, like Liim Keahioong's Taiwanese Modern Spelling System (TMSS) and its descendants 21/34.

Most romanizations use diacritics to indicate the seven tones of Hokkien, for example in Pe̍h-ōe-jī: a, á, à, ap, â, ā, a̍p. But for TMSS Liim took the same route as Gwoyeu Romatzyh, the Mandarin romanization developed by Chao Yuen-ren: tonal spelling. 22/34

Tangent: I started learning Mandarin twenty years ago through Bopomofo, then Pinyin, and now I've entered über-nerd territory by switching to Gwoyeu Romatzyh (and it's really helping with vocabulary retention). Image: a topical example from today's revision in @plecosoft 23/34

Like Gwoyeu Romatzyh TMSS has no diacritics, just the 26 letters of the Latin alphabet. MLT (a system based on TMSS) represents the seven tones like this: af, ar, ax, ab, ar, a, ap. This photo is from a book written in another related system, Phofsit Daibuun (普實臺文). 24/34

Another interesting orthography is called Crkunl. It blends romanization and other symbols in an attempt to eliminate any ambiguity (even though nobody else seems to struggle with ambiguity in romanized Hokkien). Here's a scan with a list of the symbols. 25/34

OK, we've seen systems based on Chinese characters, Bopomofo, the Latin alphabet, and katakana. But there's one other major writing system in East Asia; a source of great pride to many people who use it, it even has its own holiday. Step forward Hangul. 26/34

There's a lot to like about Hangul. It's an elegant system for writing Korean, both much easier to learn and more suited to the language than the Chinese characters Koreans used before it was invented. I think it's an attractive-looking script too. 27/34

A few attempts have been made at developing a Taiwanese Hangul. Some just take inspiration from the featural way Hangul are constructed (see the image for an example), but use an entirely new set of components to make the characters. 28/34

Ohers borrow the system wholesale, changing the pronunciation of some components to suit Taiwanese. Just for fun, show this image to someone who speaks Korean and ask them to guess what it says (or even which language it is in). 29/34

It's Taiwanese, of course, and it's Matthew 6:1: "Beware of practicing your piety before men in order to be seen by them; for then you will have no reward from your Father who is in heaven." (I may be an apostate, but I still prefer the Revised Standard Version.) 30/34

All told there are over a hundred published systems for writing Hokkien. Some have had official or quasi-official backing, which paradoxically has often made the community less likely to use them, given the way different governments have treated Hokkien. 31/34

This was the case with TLPA, which was promulgated by the KMT government in 1998 and positioned as transription rather than an orthography. This was seen as deliberate sabotage by many in the Tâi-bûn (written Taiwanese) community, who by and large favoured Pe̍h-ōe-jī. 32/34

Even today there is no accepted standard that's used everywhere. Some people use exclusively Chinese characters, some use mixed characters and romanization, and some use romanization (mostly Pe̍h-ōe-jī or the closely related Tâi-lô). 33/34

So why so many systems? In a nutshell: scattered communities, political repression, and the dominance of other languages in the places Hokkien is spoken. If you'd like to find out more, see Henning Klöter's fantastic book Written Taiwanese (2005). 34/34