And the Ghosts Wailed: Which Writing System is Best?

When Cangjie created writing, Heaven rained millet and the ghosts wailed all night…

So goes the legend of the man who invented Chinese characters. The true story is (probably) more prosaic, but Heaven could have chosen a worse time to take notice. Spoken language is hard-wired into the human brain, but that is by no means true of writing. As late as 1950 most people worldwide were illiterate, and around 15% still are today. Most cultures became literate only through contact or force, and some still are not today. In all of human history, it has developed in only five places: concurrently in Mesopotamia and Egypt, China, the Indus Valley, and Mesoamerica. The latter two are extinct, and almost all of today’s scripts can ultimately trace their lineages to the banks of the Nile.

It is natural that the first stage in writing is typically simple tallying or drawing pictures. Obviously with such a small sample size one cannot statistically assert what is typical, but over time characters do tend to become more abstract. The extent to which this process proceeds varies, however. The Chinese got to logograms, characters representing whole words with no relation to their pronunciation, and stopped. The Egyptians forged on, ultimately favouring sound over meaning, until finally their Phoenician neighbours distilled their entire script to just two dozen or so sounds.

Everyone thinks his own native language makes sense, and obviously all scripts have unique aesthetic value. History also complicates things: unlike language, choice of script is almost always motivated by politics or culture. Scripts tend to work best for the languages for which they were designed: the Latin alphabet gels far better with Italian than with English.

I do not claim to be quantitative in my judgments of which scripts are “better” than others. But through my own background and some research over the past few days, I am at least broadly familiar with writing systems used today. So here it goes.

13: Chinese

One of the world’s most famous, and indeed beautiful, scripts is also its most nonsensical. Chinese is written in thousands of characters, each representing a single word, with no visible hint as to pronunciation. For many modern words two or three are glued together, often with little intuitive connection between them. They are so hard to learn that Mao tried to throw them out entirely, and was only dissuaded by (especially Mandarin) Chinese’s ridiculous surfeit of homonyms.

I will give honourary mention here to Simplified characters (Mao’s ultimate compromise, official in Mainland China and Singapore but not in Taiwan, Hong Kong, and elsewhere abroad). Though not as pretty or venerable as Traditional ones, they are easier to learn. Also, thanks to modern China’s economic clout and tech savvy, it is impressively easy to type in Chinese- in the most popular system, a user types Latin letters, and chooses from a selection of corresponding characters.

There are tens of thousands of Chinese characters, but one needs only a few thousand for practical literacy.

12. Japanese

Ancient Japanese greatly admired China and enthusiastically adopted their script, but quickly realised that it just did not work for their very different language. They ultimately arrived at a hybrid system: kanji (Chinese characters) are used for many words (especially nouns), while two kana scripts are used for grammar and phonetic renderings. One can avoid kanji entirely, but that is rare outside children’s books and instruction materials. Japanese is thus slightly less arcane than Chinese, although American military strategists in World War II did apparently consider the possibility that enemy fighter pilots’ brains would be so overtaxed by memorising characters that they would crash their planes (I do not think there is any evidence that pedantic frustrations were a motivating factor for kamikaze pilots).

11. Pure syllabaries

Japanese kana scripts are syllabaries, with each character representing a full syllable. Pure syllabaries, in which each character is entirely distinct rather than, for example, all syllables beginning with a k sound sharing a common element, are relatively rare. But for some reason they are weirdly popular among those inventing unique scripts for traditionally unwritten tongues: modern examples include Cherokee, Inuit, and some Eastern tribal languages.

10. Ge’ez

Most syllabaries are abugidas, in which characters including similar sounds tend to look similar. That term derives from the first four letters in Ge’ez, the script used by several languages in Ethiopia (compare to alpha-beta). I admit I do not know much about Ge’ez, but as the language for which is was designed is long-extinct, and as Ethiopia is an underdeveloped country with a high degree of linguistic diversity, there must inevitably be some issues with accurate phonetic rendering of modern speech.

9. Indian scripts

The many related scripts of India and Southeast Asia are abugidas, ultimately descended from the Brahmi script used to write Sanskrit. They are more intuitive than pure syllabaries, and are often tailored for individual languages (including non-Indian ones like Burmese, Khmer, and Thai), which is a plus. However, even Devanagari, the biggest hitter among them, is pretty much entirely unknown outside native speakers, and is really not made for representing other languages. Another strike against it is that, in spite of Indians’ enthusiasm for the Internet, it is apparently relatively difficult to use on computers. One minister was recently embroiled in a scandal for preferring his Hindi briefs rendered in Times New Roman.

The as-yet-undeciphered scripts of the Indus Valley Civilisation died out in the Bronze Age, and writing arrived in India for a second time via Persia (first as the Brahmi script, shown above). The distinctive horizontal line that connects Hindi writing today was a relatively recent development.

8. Arabic (& Co.)

The ascendance of the Arabic alphabet is an excellent example of how, in writing, politics trumps linguistics. As Islam spread, speakers of Indo-European, Turkic, Austronesian, Bantu, and other languages all adapted the alphabet for their use. Persians were the earliest adopters and also the most influential: because Persia was the cultural acropolis of the Islamic world, newer converts often adopted their version of the alphabet rather than Arabic directly. Urdu’s use of the Persian alphabet is arguably the main thing setting it apart from Hindi. In the past two centuries, the process has run in reverse: myriad Muslim peoples, most prominently the Turks, have ditched Arabic in favour of Latin.

Arabic is written right to left, which is somewhat unusual in a Latin-influenced world but is by no means a deficit. It also boasts spectacular calligraphy, developed by necessity to compensate for Islam’s strictures on graven images. However, it does lose points for being an abjad (again, named for its first letters), which is an alphabet that does not represent vowels, or does so only via diacritics. This makes some sense for Semitic languages’ unique grammar of consonant triples, but for other languages it is confusing, which is why Arabic is the lowest-ranking alphabet.

7. Hebrew

Hebrew is another Semitic abjad, with the same baggage as Arabic. However, it has a slight edge for two reasons: first, its letters come in only one form, compared to two in the modern Latin alphabet or up to four in Arabic. Second, although it was an awkward fit for historical Jewish languages like Yiddish (which is Germanic), it is well-suited for Hebrew. Modern Hebrew has an unfair advantage because it is pretty much a constructed language, but accidents of history count.

6. Christian Near East Alphabets

While Muslim converts adopted the Arabic script (and often the language itself), Christian holdouts stuck to their own ways. Armenian and Georgian developed under the dual influence of Greek and pre-Islamic Persian scripts. Coptic, the liturgical language of Egyptian Christians, has some claim to evolving directly from ancient hieroglyphs.

These are true alphabets, but I rank them below European peers for two reasons: first, they are unknown outside minor language communities. Second, Caucasians, at least, have a weird proclivity for making half their alphabets permutations of a single letter. Armenian has 11 that all look like some variation/ inversion of u (ա, պ, տ, ու, ո, դ, ը, ղ, հ, կ, հ), while Georgians really went wild on Ω.

All alphabets are ultimately descended from Ancient Egyptian, but Coptic is probably the most direct.

5. Mongol (& Co.)

The traditional alphabets of the steppe peoples of Central Asia rank highly for being true alphabets in with distinct letters. They are very distinctive for being written top-down; this was traditionally true for Chinese characters as well, but for Mongol, Manchu, etc. it would be impossible to write them any other way. They look cool, but today fight for survival even on their home turf: Mongolia itself officially uses Cyrillic.

The Forbidden City is one of the few places where one sees evidence that until just over a century ago, China’s de jure official language was Manchu.

4. Greek

The granddaddy of all Western writing is very well-known today, despite only being used for one minor European language. It is an excellent alphabet, whose main defects stem from being a victim of its own success. Because it is so prestigious, Greeks never changed it even as their language shifted significantly over millennia, which causes some confusion when reading either Ancient or Modern Greek. Greeks also were only too happy to spin off adaptations for up-and-coming cultures, for example Latins or later Slavs, both of whom ultimately overshadowed them.

3. Cyrillic

Developed by 9th Century Greek Orthodox monks as part of their mission to civilise the Slavs, Cyrillic has always been the somewhat less-successful younger brother of Latin. It is, of course, well-tailored to Slavic (though Catholic Slavs opt for Latin; among Serbo-Croatian speakers this causes some awkwardness). Most letters either have direct equivalents in, or are exactly the same as, Latin, but Cyrillic is burdened with a lot of redundant vowels. Thanks to Russian influence it spread to many non-Slavic Eurasian languages, but it is telling that, upon getting free of the USSR, many have switched to Latin.

2. Latin

The Latin alphabet became the international standard because of the worldly dominance of those who used it, but the world could have done worse. It is easy to learn and straightforward to use. Of course, different languages pronounce letters, especially vowels, differently, but the alphabet can readily be made to represent pretty much any sound (the International Phonetic Alphabet used by linguists is mostly Latin). Today, Indonesians (who switched in 1901), Turks (1927), and Vietnamese (1910) find it just as natural as Italians and Spaniards do.

1. Hangul

In 1443, King Sejong of Korea commissioned his top scholars to create a new writing system to replace Chinese characters. They came back with Hangul, a set of letters packaged into syllable blocks. Obviously it is tailored to Korean, but it is extremely impressive. Syllable blocks are all organised in the same way, and letters’ shapes are based on the shape the mouth or tongue makes when pronouncing them. Idiosyncrasies of the Korean language are all accounted for logically and intuitively.

Easier than A, B, C.

Sejong was enthusiastic. People observed of the new letters, “A wise man can acquaint himself with them before the morning is over; even a stupid man can learn them in the space of ten days.” But the Confucian elite were not so pleased. After the king’s death they went right back to Chinese characters; Hangul would not become the national standard until after World War II.

Hangul is the only script that was invented from scratch, structured according to reason rather than cultural tastes and historical accidents. Getting there was complicated. But the Koreans’ achievement is singular, almost on par with a nation successfully adopting Esperanto. Language is irrational, writing even more so- at least some people got it right.

Former linguistics student; current investment bank analyst who sometimes thinks about something other than spreadsheets