Bryan Caplan  

Kaplin's Simplifiid Speling

Medical Technology... Divisions on Immigration...

Greg Mankiw speculates that a world language will emerge, it will be either English or Chinese, and smart money says English will win. The interesting thing to me is that is it hard to become literate in either of the front-runners. Chinese is of course notorious for its lack of a phonetic alphabet, but English spelling is dreadful compared to say, German.

In my spare moments, I've worked out a whole system of phonetic English spelling. Imagine:

1. Dropping all silent and unpronounced double letters.

2. Dropping q (replace with kw) and x (replace with ks or z). Replace c with either k or s, and pronounce c like ch.

3. Standardizing vowels so that a single letter indicates a "short" vowel, and a double letter indicates a "long" vowel. Specifically:

Sound Old Spelling New Spelling
short A cat cat
long A hate haat
short E bet bet
long E beat beet
short I hit hit
long I write riit
short O Bob Bob
long O boat boot
short U up up
long U boot buut

Unfortunately for my ego, it turns out that I'm just re-inventing the wheel. A little google searching revealed the Simplified Spelling Society and at least one complete system for improved spelling. The latter highlights a number of sins of omission of Kaplin's Simplifiid Speling - like how am I going to spell "put"?

The main problem for both Chinese and English, however, is that this is a classic coordination game. Who goes first? Em Ii gooing tuu start bloging with simplifiid speling, muc les riiting ortiklz? When you put it this way, I think Chinese has an edge over English in the race to be the world language. Chinese spelling is so bad that they might actually pay the switching cost of a well-crafted phonetic alphabet. English, in contrast, is bad but livable, so I wouldn't be surprised if English becomes a dead language before its spelling changes.

Comments and Sharing

COMMENTS (31 to date)
Fabio writes:

Don't you know how many spelling bees will be destroyed if your diabolical orthography crusade ever succeeds? You sick twisted economist!!

diego writes:

What is the phonetic difference between a short I and a long E? Would there be any difference between hit/heet or bit/beet?

More interesting I think is the question of how spelling will evolve with the IM generation. Perhaps it will be an improvement over current spelling rules in the direction of Kaplin's Simplifiid Speling.

Mike Linksvayer writes:

By rule 1 the new spelling of cat should be kat.

John Thacker writes:

First off, English has a lot more than 10 possible vowel sounds, so that's one problem. (Among other things, what do you do with schwas?)

Secondly, Chinese has way too many sounds that are linguistically rare compared to other languages, plus the use of tones, which is also moderately rare and difficult for people not used to tonal languages. There are too many homophones if you don't use the tones. Chinese also requires distinguishing between sounds that sound identical to non-Chinese speakers. (non-aspirated t vs. aspirated, for example)

Japanese and Korean both made partial or full switches to phonetic systems, and both have sounds which are fairly common in world languages, or at least close to them.

Why can't we reform English pronunciation instead of spelling?

Of course, we have to agree on a way to pronounce "ough"...

Matt writes:

Brian Caplan, you are a nut. I have no other diagnosis.

Robert writes:

While I don't give good odds on anything becoming a world language, for the the Internet, if anything has a chance, I envision that it is something stealing a page from Chinese, being an ideographic written language bearing no particular relationship to any spoken language.

John S Bolton writes:

Children's books could have marks to indicate pronunciation, with slashes through the silent letters. Eventually, they could read the unmarked texts. Once you know the system, and have memorized the unsystematic exceptions, the lack of phonetic transparency is not a problem.
Language learning is for children, not adults. Our government indulges unreason by pretending that adults can learn a foreign language as well as children can.

Lex Spoon writes:

Among many other issues, notice that human languages grow, just like economies. Designing languages tends to not go very well because real languages are the result of many practical forces that even our best linguists do not always understand. Further, even if you had a magical language designer to get it all correct, do not expect your changes to stick over the years as the language grows with the culture.

For spelling in particular, note that the pronunciation of individual words changes over time. Thus, a literal spelling would somehow need to change over time.

For more fun, notice that English has many dialects depending on the community of people speaking it. Even if we posit a master linguist, imagine the political brawls over whose pronunciations get preferential treatment!

Speaking of politics, have you ever checked out how well France is doing with *its* language police? Not well. The French Language Academy can make all the proclamations it likes, but French speakers still say "hamburger" and "email". (This is good, IMHO -- these are very fine words, and I find it rather dark to try and "improve" a culture from above.)

Not to say that improved spelling is impossible. However, given the above, you need to be careful about your goals! I guess you should aim to improve the spelling of only core, frequently used words, because more outliers are constantly being added. Also, why not pick spellings that make sense with current rules? Instead of "riit", make it "riet" or "rite". That way, your spellings have a chance to catch on by natural selection.

As for the chicken and egg problem? Just go for it! Be a language rebel, just like the goes who say "personhole" instead of "manhole", who use words like "zie" and "zir" with a straight face, and who use singular "they" on purpose. It is the language version of greeting census workers with a shotgun. Yoo kaant take awae mie libertee!!


PS -- Noah Webster has already inserted a few spelling simplifications into American English. ;)

brtkrbzhnv writes:

diego: The words heat and hit (or heet and hit) are not pronounced identically in standard pronunciation. Phonetically, they're [hi:t] and [hɪt] respectively.

Wikipedia lists at least a dozen different vowel sounds for English (the precise number depending on the dialect and the manner of counting.) at Wikipedia. Anyway, I think you should learn from us Europeans and invent som fancy letters of your own, or at least import some of ours. For example, sought could be spelled såt and fur might be spelled fö(r).

Slavisa writes:

Brian, it's not spelling that hurts, it's grammar. You need a plan to drop all definite and indefinite articles, or at least some consistent rules of their use. Or consistent exceptions to the rules.

dearieme writes:

But if you standardised spelling, everyone would have to learn to pronounce English as I do. (It is universally agreed, isn't it, that my spoken English is the kat's, em, nitewayr? We certainly don't want the world speaking in a midwestern drone, do we? We'd all fall asleep the whole time.)

Famous J writes:

I think Mr. Thacker has it right on Chinese. The biggest reason Chinese will never become the world language is because it's tonal.

In general, people whose native languages are tonal have no problem picking up non-tonal languages (like English, say). The same is not true for native speakers of non-tonal languages. Sit in on a Chinese 101 class sometime and you'll see what I mean.

When you consider that between Spanish, English and Arabic, well over half of the planet doesn't speak a tonal language, that puts Chinese at a serious competative disadvantage.

Lauren writes:

Bryan wrote:

Chinese spelling is so bad that they might actually pay the switching cost of a well-crafted phonetic alphabet. English, in contrast, is bad but livable, so I wouldn't be surprised if English becomes a dead language before its spelling changes.

There is no such thing as "Chinese spelling." That you would use the term suggests you are struggling to make your point. You are stretching!

Chinese, like English, is a major language in the world, both in daily and internet use. It is a lingua franca of communication--pun intended (franca--French--was once the language of high intellectual communication and using other languages was spurned).

Writing, in any language or form, creates permanent records for communication. It is common for one group that has worked internally on its own standards to gripe about another about group's standards of the moment--for example, about details like spelling or how characters are written in other languages or its own.

What is economically interesting is that groups can talk to each other across time via writing, or preserve written records.

Writing across space during a given time is also interesting, but that problem has historically been solved in every time period.

Economists with an interest in language might want to consider statistically the relative sources or language and success of writing throughout the world.

Chinese as a language is very successful worldwide. So is English! Both are accretive; and I'd guess that that's the key.

If you want to focus on Asia, and if you want to think about matters like spelling or phonics, you should perhaps look to Korean (Hangul, a highly successful phonetic language for several hundred years) and Japanese.

If we're comparing major languages, alphabets, or phoneticizations, let me also add to your list Cyrillic, Hindi, Arabic, Hebrew, and native American written languages whose histories have also become more understood; and also southeast Asian languages from Vietnamese to Indonesian.

The internet has made competition in written language an open market. Are phonetic-based languages like English what will result from the fray? Phonetic-based languages are one contender.

Are phonetic-based languages inherently the most productive, after controlling for population, politics, etc.? What measures of "production" are relevant for a matter like language? If phonetics are important, is English the most productive of the phonetic languages? If phonetics are not important, which languages should we focus on?

Lauren writes:

Famous J wrote, thoughtfully:

I think Mr. Thacker has it right on Chinese. The biggest reason Chinese will never become the world language is because it's tonal.

It's a common misconception that many Asian languages such as Chinese are ultimately handicapped relative to English because they are "tonal".

English is tonal. It is every bit as tonal as Chinese.

As a simple example you can try for yourself in English, try saying the word "automobile" out loud.

Every time you say it, you will say the last syllable in a lower frequency--that is, in a lower tone. Most multi-syllable words in English are tonal. It comes so naturally in English, and so naturally in Chinese, that no one thinks about it when speaking.

English is often even more tonally complex than Chinese because English uses tones to convey questions, exclamations, or vernacular. It's as hard to convey in writing in English as in Chinese, but try just a few different connotations for the spoken word "right": Right! Right? Riiight! Allright! a'right! a'ight. And all that just to mean OK in different contexts. I don't know that any linguist has ever bothered to count the number of tones, but I'm sure interested if someone has!

Mr. Thacker's point that both English and Chinese have a lot of inflectional sounds, usually associated with vowels, is indeed thought-provoking.

However, his comment that languages like Korean have "made partial switches" to phonetic bases is historically innaccurate. The phonetic basis of Korean happened many centuries ago.

The only switching being done today is because of the competition provided by the net. English has to compete as a language, both written and spoken.

David Youngberg writes:

Lex Spoon- I'm glad I'm not the first one to point out that languages are grown, not built. There must be a reason why we spell things the strange way we do: some of it are surely relics from forgotten decrees and some of it is obselete but some of it must be for good reasons.

Path dependence has locked us into this spelling pattern and while it will evolve over time, "revolutionizing" it is a pipe dream. It's not just a coordination game between "English speakers" and "Chinese speakers" but, most fundamentally, between one person and everyone they talk too. So tell you what, Bryan: You start spelling everything the "better" way and then see if all of us pick it up.

RenoWiggum writes:

Bryan writes:

Em Ii going tuu start bloging

Shouldn't that be "gooeeng"?

Wild Pegasus writes:

The problem with English spelling reform, at this point, is that English has so many dialects. If an American tries to speak to a northern Englishman or Newfoundlander, there's little chance the American will understand either one of them. However, the written language is practically identical everywhere. A phonetic reform raises the question, "Whose accent?"

Incidentally, this is probably what happened to Chinese: every Chinese language attaches the same meaning to each ideograph, but pronounces them differently. Thus, a person who is literate in one Chinese language is literate in them all, even if he couldn't understand a word of the speech.

- Josh

Michael writes:

N.B. You misspelled 'Mankiw'.

Kerry writes:

As usual, when Bryan steps outside of the realm of econ he starts talking the crazy talk. Spelling reform has a long and rich history of failure, and a cursory search reveals a lot of less akward systems.
Others have been commenting on alphabetical/syllabaric writing vs. character writing. As someone who has taught English to folks who read a character language and also learned a character language himself, I can say that English has a lot of advantages as a written language over Chinese. In America a child can read most of a newspaper by age ten or so. In China you need to have at least a late-middle school or early high school reading level to do the same thing. While English spelling is often obscure or arcane, it does have a number of patterns that can enable a reader to sound out words and thereby identify a previously unread word with a word in his spoken voabulary. In Chinese the reader is unable to do this in the majority of cases. The character structure may give a clue to the meaning if one knows a related character, but the pronuciation will be unknowable without a dictionary or instruction*. As such, the learner can get a lot farther faster with English, because there is a feedback system whereby one can try to pronounce and use previously unknown words one has read and likewise one can recognize in print most words one can speak. This means you need not study literature and conversation as separate and distinct entities, and can thus learn faster.

*There are some characters whose pronunciation is guessable from elements of the character. However these characters are in the minority, and if a character is unknown then it is not much help. It can be a good mnemonic device, though.

[Note to Kerry from the EconLog Editor: Please supply a functioning email address. A functioning email address is a requirement for posting on EconLog.]

Bryan Caplan writes:

Thanks for pointing out the spelling mistakes. Mankow should of course be Mankiw, and going should be gooing. :-)

Brandon Berg writes:

John Thacker:
Chinese also requires distinguishing between sounds that sound identical to non-Chinese speakers. (non-aspirated t vs. aspirated, for example)

As far as I can tell, the aspirated and unaspirated 't' in Mandarin are exactly the same as 't' and 'd' (initial) in English, respectively. If there's a difference, it doesn't seem to be a terribly important one.

Ronnie Horesh writes:

I read somewhere that the Chinese language predisposes that culture to authoritarianism because, as Kerry says, it's difficult to learn it yourself, just by knowing a few basics: you have to be taught. English has its quirks, but you can go a reasonable way with the basic letter sounds and a few rules.

I'm not sure I agree with Lauren that English is as tonal as Chinese. We do vary the tones a lot, but we can at least understand monotonal speech (like synthetic computer voices) quite well. That would be impossible for Chinese.

Dylan writes:

Semplifyed speling can bee a bad thng. Complex spelling rules make for quick, robust reading and error checking. Bad spelling or transcription in a complex spelling environment is still easily understood; screwups by the careless or exceptionally dim in an easy spelling format seriously hinder comprehension. Or so I recall reading in a book on cryptography long ago.

Shradha Kaul writes:

English vs. Chinese? Give me English anyday! In spite of the almost non existent logic, English is very easy to learn and to write and to speak. Imagine the jarring Chines notes all around you and along with the mixed up alphabet.

Lets stick with English, shall we?

Lauren writes:

I'd like to focus on this question: Is there an evolutionary process leading toward a worldwide language?

If so, what factors influence the "winner", and which kind of language is most likely to win? Are there different competitions for written, spoken, and reading languages, and is there more than one long-run equilibrium?

In an era when computers and the net do a lot to level the playing field between languages, are either one of Chinese or English going to be the obvious "winner"? I'm not arguing for either one, though I'm going to take the devil's advocacy as my guide here. Like Mankiw, my personal guess is the winner will be some form of English. Maybe it all comes down to income, but if so, income is also endogenous. Regardless, I'm having a really tough time justifying a guess of iterating in on English as a worldwide language, or ruling out other major languages like Chinese.

Ronnie Horesh offered some fascinating hypotheses, first on possible side-effects of any potential winner:

I read somewhere that the Chinese language predisposes that culture to authoritarianism because, as Kerry says, it's difficult to learn it yourself, just by knowing a few basics: you have to be taught.

I strongly doubt this. First, one has to be taught written English in school (U.S., British, Australian, etc.), same as learning Chinese in China--mainland or Taiwan--or anywhere else, and for as many years. An interesting question might be to compare the number of months or years of school it takes to reach basic literacy of a given number of words--say some agreed-upon number of words readable by an average 7-year old.

In China and in the United States, I'm going to guess that it's about the same. Either way, kids have to sit down in kindergarten and first grade long enough to be taught. In both cultures there are gifted kids who learn to read on their own before starting school; but again, it would be interesting to make a comparison.

Chinese words, and particularly elementary ones, have the advantage of being very compact. Modern Chinese words all have two "syllables"--two characters. Speed-reading in Chinese is automatic--if you can read, you can speed-read. Chinese grammar is also very simple. The same paragraph in English and Chinese takes considerably less space in Chinese. The upshot may in fact be that it takes less classroom time to achieve basic literacy in Chinese than in English.

A common argument seems to be that once one learns the basics of phonetic-based languages, one can always read and learn more. That argument runs up against factual experience in the U.S.--that high school graduates are pretty illiterate when it comes to writing anything as long as a paragraph, and reading anything as demanding as a newspaper. Twelve years of school is apparently not authoritarian enough. I suspect that what kids have to be taught to go further, in any language, is the complexity of quality of writing and a taste for quality in reading, and that may take equally much authoritarianism on the part of schools and parents in any culture.

Years ago I'd have said that English would win the language survival war because you could type it. The advent of the net meant that typing was a critical skill!

But I forgot that computer power could--and has!--solved that problem, bringing Chinese typing down from the level of specialized skill requiring thousands of dollars worth of equipment to something kids can do.

Computers have also reduced some of the intensity of the competition by gradually improving automated translation. I suppose that will serve to prolong any movement toward a single worldwide language. Mankiw could lose his 500 year bet because it's too short a period!

Too, a worldwide language may not be a stable equilibrium. Spoken languages lapse into dialects over large areas in every country. Written languages, though, don't seem to be as fluid, though the appearance of a whole new level of written English through the abbreviations kids use to IM and text-message each other suggests otherwise.

Ronnie continues:

I'm not sure I agree with Lauren that English is as tonal as Chinese. We do vary the tones a lot, but we can at least understand monotonal speech (like synthetic computer voices) quite well. That would be impossible for Chinese.

I'm meeting with a student from China this afternoon, and I'm going to ask this question.

My guess, though, is that synthesized/computerized speech in Chinese is easier to understand--and program!--because there are only a finite number of possibilities for each syllable. There are a fixed number of spoken syllables, each with a maximum of four tones (and most with fewer in practice). Combined into two-syllable words, the number of words is still as large as in English, but is easier to program a computer to replicate because each syllable does not also require additional inflection depending on the length of the word (as in my previous example of the word "automobile") or context. (In both languages, though, there are some tonal differences conveying things like questions or according to tense.) The monotonic synthesized speech used by English-speaking computer automations is an interim measure precisely because English has so many possible tones that it's so-far impossible to record them all in a way that makes sense to listeners!

Great thoughts, Ronnie. The Hayekian question underlying your first observation is mighty intriguing.

(who speaks, reads, and writes a little Chinese, by the way)

John Thacker writes:

The phonetic basis of Korean happened many centuries ago.

The invention of hangul, yes, as a completely phonetic alphabet was several centuries ago. But the use of hanja took much longer to go away, certainly, and you'll still sometimes see it in certain situations in the ROK (though not in newspapers), though it was banned in the DPRK.

Responding to another comment, certainly English has accent and intonation, just as, say, Japanese does. However, getting the wrong intonation or accent is much, much less likely to cause confusion in English or Japanese than in Chinese. There are many fewer cases of changing the meaning by changing the accent in both, and in both there are a limited number of standard intonations and accents for words and phrases that stretch across syllable boundaries. (E.g., accent the second syllable in most words, change it to the third when it becomes an adjective, rise at the end of a question, etc.) Pitch accent (Japanese) or stress accent (English) are completely different, then, because the various moras, syllables, and words are not independent.

(Of course, Japanese has its own problem of homophones in complex words by borrowing characters and pronunciations from Chinese but losing the tonal information.)

Most Chinese languages (not Shanghai dialect/language, which is like a pitch accent) have independent tones for each syllable or mora, which is an entirely different situation.

marko writes:

Thea real problem with "reforming" spelling is that for an indefinite time period people would have to know both. All the new books would be printed in "New English" and all of the old books would be, well, in old. So, kids would have to study both. But, OCR software, coupled with some smart software for "respelling" could probably do the job of respelling all of the books written in "Old English" in several years.

And this is probably the reason why it was relatively easy to switch in Serbia (my country) from Old Slavic spelling to Serbian spelling (completely phonetic) in mid 19th century - there were simply to few books in the old language to make that switching too costly.

Roger Snowden writes:

Looks like Dutch, with a Candian twist. How aboot that?

Richard Russell writes:

Does no one remember the efforts of George Bernard Shaw and G. H. Wells to reform English spelling? They do make sense and are not that difficult. The direction was to create type characters equivalent to each phoneme but the total of letters then became 48 or so. That does not seem to me to be a huge obstacle to childern since their language learning programs are in full gear and we adults finally die off anyway.

On a slighter scale, since we use a Latin model with German additions, does any one know the extra letters that Emperor Julian The Apostate introduced into Latin that could be reused for our purposes?

Kent writes:

I have to agree with Lex Spoon. You can no more design an orthography than you can design a culture.

Comments for this entry have been closed
Return to top