My Conversation with Paul Romer

Here is the audio and transcript, Paul was in top form and open throughout.  Yes economic growth, blah blah blah, but we covered many related topics too:

COWEN: And you also think we should simplify the English language. Right?

ROMER: [laughs] Well, there’s two parts to that. One is, in writing and communication, there should be a very high priority on clarity. It’s hard to know what’s the mechanism that enforces that. There are variants on English, like the English used to write the manuals people use to service airplanes, where there’s a very restricted vocabulary. The words are chosen so that you can’t have any ambiguity because you don’t want somebody servicing a plane to get confused. So there are some things you could do on writing, word choice, vocabulary, exposition.

There’s a separate issue, which is that amongst the modern languages, English has the worst orthography, the worst mapping between spelling and sounds of any of the existing languages. And it’s a tragedy because English is becoming the universal second language.

The incidence of people who don’t learn to read is substantially higher in English than in other languages. People have known for a long time, it takes longer to learn to read in English because of the bad orthography. But what hasn’t gotten enough attention is that there’s an effect on the variance as well. There are more people who never get over this hurdle to actually learning to read.

If there were a way to do in English what they’ve done in other languages, which is to clean up the orthography, that could make a huge difference in the variation associated with whether or not people can learn to read English.


COWEN: Can a charter city work if we import good laws from the outside world but not the appropriate matching culture?

ROMER: You’ve zeroed right in on the connection. The real motivation that I had for charter cities was exactly this one that you can see in the US versus New Zealand. You can think of a charter city exercise . . .

This is actually the story of Maryland: We’re going to create laws, and we’re going to guarantee freedom of religion in Maryland, and it’s in the laws; it’s in the institution somehow. That didn’t turn out very well. Maryland had a Catholic elite but then large numbers of Protestant indentured servants or workers. And this kind of commitment to freedom of religion was not stable in Maryland at all.

The case that’s worth trying to copy is Pennsylvania, where William Penn recruited large numbers of people who actually believed in freedom of religion. The word charter comes from the charter that Penn wrote for Pennsylvania, but it wasn’t the document that mattered. What mattered was that there were a bunch of people in the founding population who were committed to this idea of a separation of church and state and religious freedom. And that’s what made it durable in Pennsylvania in a way it wasn’t in Maryland.

And this:

ROMER: …Moses was of this generation that was too enamored of the car, and this is where I think Jacobs had a better intuition. But the challenge, the dichotomy I would pose would be Jane Jacobs versus Gouverneur Morris.

Morris was the guy who drew the grid that laid out the rectangular street map for Manhattan.

We also discussed music, including Hot Tuna, Clarence White, and Paul’s favorite novel, dyslexia, what Paul has learned about management, and much more.  Self-recommending, if there ever was such a thing.


What might an inconsiderate overhaul of English orthography do to the etymologies of the English words?

What? E. Burke it's not clear what you're asking. Etymology is the study of the origin of a word, so an inconsiderate overhaul will merely make the etymology more convoluted.

Bonus trivia: orthogonal relations are big in relational databases, since you want information not to be repeated in records (for relational databases as opposed to 'flat' databases where you just pile everything into one big fat file, which is actually the modern trend). Which witch is which?

I think I ask just how stupid, just how inane, just how ridiculous (note: all three adjectives derive from Latin) Romer's proposed invitation to "rational orthography" for English might instead be thought.

I suppose we'd be forced to concede, ultimately, that rational orthography must further be imposed upon every foreign word incorporated into this West Germanic language--not only from numerous Indo-European tongues but from numerous non-IE languages, too.

We suffer today from enough "rational lunacy": may our fair language escape such pernicious designs for as long as possible--may such further rational lunacy even be opposed strenuously.

Poets: attack! (BTW: etymology is concerned not "merely" with a word's origin but with its history and the history of its cognates, inclusive of orthography.)

One suggestive poetical attack (excerpted in English):

Holy Science says:
“our tall cathedrals console
with hieratic truths,

and with lethal truths:
with our hieroglyphic math
do we conquer all.”

science and math lie:
these cannot tell truth entire,
Reason's halfwit slaves.

medieval the new:
days of rational belief
and mythical thought.

once poets restore
their tongues, then can they speak 'twixt
shadows and their things.

I agree, rationalizing English spelling is a lost cause.

Other languages work fine with worse spelling. Chinese for instance does not even has an alphabet instead they use thousands of symbols and learn to memorize what each symbol means and words are learned in a completely different manner. English can be understood as a midpoint between Chinese and the Romance languages, where there is an almost perfect correspondence between sounds and symbols.

It wud graytlee improove kommuneikations by mayking spelling moor clowslee resembull foenetik speech

Watz dis? Ebonics or Esperanto?

Shouldn't you be responding in Spanish?

Well it would do wonders for the Scotts language which would soon be more unintelligible in print than it is when spoken, but what about West Virginia? Should the mountaineers have a different orthography than Boston?

The Shepardic Jews variety of Spanish (Ladino) is just like this. It's quite interesting and simpler than Spanish.

It wuud greitli impruuv kommiunikeishuns bai meiking spelling moor klouzli rezembal fonetik spiich.

At least do the exercise properly :)

Thank you. It has been too long since I did phonetics. :-)

Romer is telling tall tales and Tyler is playing the part of the credulous hayseed. English does not have the worst orthography, Chinese does.* Awkward spelling is not why many people are illiterate in English, rather it is the high percentage of ESL and, in the USA, ideological devotion to the known-to-be-ineffective See-Spot-Run ("whole word") teaching method.

Does Romer really believe that stuff or is he just pulling Cowen's leg?

*Even among Western European, Latin-alphabet languages, Danish spelling is worse.

Probably because every Danish language reform is that not every language called Danish has anywhere near the same pronunciation. This is even worse in Swedish, but then I don’t speak Danish at all..

Interesting that you bring up Chinese. My first thought was "clean up English spelling? You gotta be kidding!" But then I remembered that Chinese did exactly that with the simplified Kanji system. Japanese even did something similar by defining the Jouyou Kanji. So, maybe it is possible to clean up a modern writing system to make it easier to read and write.

Brazil has done it many times. "Paiz" (country) became "´país", "pharmácia" became "farmácia", "rhetórica" (rhetorics) bacame "retórica", etc.

I find simplified characters to be harder to understand than traditional because of lack of radical cues.

My understanding (from Language Log) is that only a fraction of Chinese people can read enough Chinese (characters, not Pinyin) to reach even the "read a newspaper" level of literacy. Writing "literacy" is even worse. Yet China seems to be doing moderately well.

'do to the etymologies of the English words'

So, budget and buffet - both of French origin, but one pronounced using English conventions (budget), while one is pronounced in French style. Except when it isn't - a buffet and a buffet mean two different things (not to mention different parts of speech that have no relationship at all), depending on how the word is pronounced. And more amusing is the fact that 18th century French gave us buffet while it is old French that gave us buffet. Which pronunciation preserves the etymology best, because using the same letters to create words that are pronounced differently is not exactly a good way to do it.

English orthography is the worst - however, it is unlikely to be improved, for any number of reasons.

Etymology being one of the more trivial - in part because a lot of 'Latin' words reached English through French, and in part because the Greek origins of 'television' are not Greek. Enjoy reading about hybrids -

English language taste and talent for literate satire having become so poor, native Anglophones may have to hope that soteriological contributions from the French (such as Jarry's conspicuous rendering of "merde" as "merdre") at the very least might help keep English speakers (and translators into English) alert to the Romans' "dog letter", the emphasis reminding us of the distinct literary contributions coming to us from native Roman satire, to say nothing of Latin orthography or semantics.

(Bonus trivia: dog crucifixion became an annual event [3 August, mutatis mutandis] for Romans eager to commemorate the Gauls' sack of Rome c. 387 BCE--the commemorations continued for almost five centuries.)

And yet we seemed to do without the Roman ablative without damaging English at all. And the ancient Greeks, whose literary contributions dwarf anything the Romans did, did not bother with it either.

Only in English do “House Cat” and “cat house” mean different things. Myself, I’ve never seen cat houses, only dog houses.

You should get out more, Eastern Nevada is beautiful

I know that English orthography is bad. But the worst? I think not. I've been reading Language Log for years. I you hunt around there you'll find plenty of articles on the difficulty of both reading and writing Chinese characters, for example this.

Exactly, and after using computers many highly educated Chinese regularily cannot remember how to draw many characters they can read.

I tried to find that poem that repeats in classic calligraphy the same Chinese character 1000 times but with slight changes and it means 1000 different things, but failed. The closest I got was this:

Bonus trivia: I saw a Filipino with a Greek character tattoo. Upon inspection, I concluded it was his own name phonetically pronounced with Greek characters. Pretty original. Either that or it's a word I never heard of. I hope he did not think I was too weird for following him and his girl in a mall while trying to read his forearm and look nonchalant.


Chinese and Japanese are the worst, but Chinese is especially egregious in this regard because at least Japan made moves to phoneticize the language without resorting to romanization, albeit imperfectly. HanYu PinYin is a poor hack for native speakers of the language while create different problems at the same time. Some interesting Chinese language facts:

- It's believed there are approx. 50-60,000 Chinese characters, of which the average Chinese university graduate (i.e. above average) will have a working on-call knowledge of 15,000. Chinese language scholars = 20-25,000. Chinese language experts and researchers at the highest level = approaching 40,000+. NO ONE knows them all.

- Several characters are extent whose meaning and pronunciation have been lost to time. Some of these have been reassigned new pronunciation and definitions, while others are the subject of continued study by linguistic "archeologists" within China. The field is larger nowhere else than in China.

- PRC simplification of the character system in the 1950s made the situation much worse and created more problems than it solved.

- TiBi WangZi (forgetting how to write) is a direct negative of the failure to adopt an indigenous phonetic writing system within China. It is a HUGE problem and growing, to the extent that some classrooms in China have teachers that are now monitoring students notebooks to ensure they are actually not simply writing down important information in English.

- Technical/research/engineering writing undertaken in China that is published on its own in Chinese or without an accompanying English translation is considered largely a joke and disregardeable out of hand.

- If it is not currently, China will be the largest English speaking country within the next 10 a 2nd language.

'on the difficulty of both reading and writing Chinese characters'

Arguably, ideographs do not have orthography. How they are pronounced is independent of the ideograph. Though of course, ideographs do represent the writing down of a language - without most of the concerns of orthography.

'The English word orthography dates from the 15th century. It comes from the French orthographie, from Latin orthographia, which derives from Greek ὀρθός orthós, "correct", and γράφειν gráphein, "to write".

Orthography is largely concerned with matters of spelling, and in particular the relationship between phonemes and graphemes in a language. Other elements that may be considered part of orthography include hyphenation, capitalization, word breaks, emphasis, and punctuation. Orthography thus describes or defines the set of symbols used in writing a language, and the rules regarding how to use those symbols.

Most natural languages developed as oral languages, and writing systems have usually been crafted or adapted as ways of representing the spoken language. The rules for doing this tend to become standardized for a given language, leading to the development of an orthography that is generally considered "correct". In linguistics the term orthography is often used to refer to any method of writing a language, without judgment as to right and wrong, with a scientific understanding that orthographic standardization exists on a spectrum of strength of convention.'

That lack of connection between meaning and pronunciation is probably the strongest advantage of ideographs, though there are many more disadvantages when viewed from the perspective of using an alphabet.

"Most natural languages..." All?

I've now read the whole conversation. They do get around to Chinese and Japanese, but those remarks are disconnected from the remarks about English orthography.

I know that English orthography is bad, though just how bad, I don't know. But the worst? I think not. I've been reading Language Log for years (it's just around the corner), and I've read a number of article on the difficulty of both reading and writing Chinese characters, Japanese too. (I'd provide links, but those tend to get filtered out, if they don't actually keep a comment from being posted.)

I must admit, I recall seeing a table which had English mid-ranked on orthography, not the worst. So likewise confused...

Correct. To call English the worst is most definitely Euro-centric as it only looks bad compared to highly phonetic languages with very regular spelling.

I read a book long ago on Shannon's Information Theory. I can't put my fingers on it now, gone for years. But one table I DO recall is a table of the information content of various languages (written). English was the MOST redundant. (I don't recall how many, if any, non-Indo-European languages were included.) Would a change in orthography change that? For the better? What IS better? less or more redundancy/efficiency (signal information density)? Orthography, it seems to me, is a trivial "problem". You may recall the linguist TC mentioned yesterday who studied the senses (shape is not a sense in any sense, but still) by testing the consistency of the verbal response to a given stimulus. (I suspect the study is rubbish, since context and culture would be extremely difficult to account for, imho) but still the same approach => test of the correspondence between a word and its pronunciation would make for a trivial study comparison between languages. (I'm not sure how tone, etc. would impact such a study...) Hey, if anyone can point me to the definitive listing of English orthography, I'd like to know. I constantly see English language articles including characters (and accents marks) not present in the English alphabet. I don't want to seem too naïve here, but can we really expect to hit a moving target? Talk about naïveté. There are, it is said, 44 phonemes in English. Assuming we want to have a 1:1 mapping, that would require 18 new glyphs. Sure we could use accent marks - but those would almost certainly conflict with the foreign usages already "part" of English. I'm pretty sure I could (eventually) learn to read English with 18 new characters, but where's the evidence that that would be an improvement? (How often is the IPA spelling useful to you? (if it were such a wonderful thing (to map 1:1) why aren't hyperlinks between words and their IPA spelling (or pronunciation) pervasive on-line?) Is it really necessary that spoken and written language is a 1:1 mapping? Why? Is there a there there? Color me dubious. But speaking of windmills: let's talk about adoption of the metric system...LOL

Speaking of IPA, those who really believe English orthography needs change, why aren't you using IPA in your writings? Seems an obvious place to start. If not you, who? (Think of the benefits, specially for ESL-ers.) (aside from the fact that Unicode doesn't even contain all of the IPA glyphs (as such).) Oh, I just learned that IPA is the norm teaching English (and French) in China & Russia. So, not so dumb.

Your memory may have played you false. Every computer-interface and web designer knows that you have to leave a lot more space for messages in languages other than English, because most of them have lower information density. The much richer vocabulary of English helps a lot here. Instead of decorating a few English nouns and verbs with a slew of other words, we English speakers simply emit le mot juste.

The need to leave more space is not predicated on English having high information density, and even less so on it having be higher than most languages. Even if all languages are equal in information density it will be unlikely that for any given phrase your language's version will happen to be the longest (if you support 10 languages of roughly equal density you'll on average have 1/10 of the phrases having their longest translation be in your language).

"What mattered was that there were a bunch of people in the founding population who were committed to this idea of a separation of church and state and religious freedom. And that’s what made it durable in Pennsylvania in a way it wasn’t in Maryland."

A devastating Alt-Right conclusion that deserves consideration.

Political institutions matter less than their building material. Without populations culturally/genetically committed to the ideals, all the ink-on-paper in the world won't save you. Illiberal populations cannot be simply converted to liberalism by institutions.

Yes. And what is to stop countries from interfering with successful charter cities that conflict with their leading cities by reneging on the rules or else changing the demographics of these new cities when it suits them? China's relentless pursuit of Sinification in its Western areas and Tibet is evidence of how deeply it understands the importance of culture. Similarly the Russians were not only able to impose Russian over a large area, but the USSR school system was able to teach a surprisingly homogenous curriculum in the whole country with the promotion of a spoken language that is much more homogeneous among average speakers than English is in the USA. So this idea works against homogenization that countries prefer. This charter city idea basically assumes that we can "sabotage" the existing culture by creating an independent region that will do very well and create interest groups to subvert the dysfunctional culture of the host country. Not a crazy idea, but one likely to meet failure often, or worse, violent conflict.

Not really an Alt-Right position, I would say. Seems more like a standard Burkean bit of wisdom. Oh, the attitudes and cultural traditions of a group of people might matter in the kind of polity they build? Well, I'll be cow-kicked!

The notion that people make countries and healthy institutions don't just grow out of the Magic Dirt is practically crimethink in the US. It's a pretty significant admission by Romer but not surprising in light of his project to replace Meso-Americans living on the Caribbean coast with high-IQ whites, northeast Asians and Hindu from Silicon Valley.

Not all academics toe the Progressivist/Whig line. I had an exchange with Robin Hanson pointing out that his concept of a buy-in for citizenship would lead very rapidly to global segregation along ethno-cultural lines and his response was pretty much, "Yes, exactly!"

I agree on the crimethink part. I just think the idea that people of different cultures, classes, or ethnic groups aren't entirely interchangeable is an insight far older than the alt-right. Credit where its due!

'The notion that people make countries and healthy institutions don't just grow out of the Magic Dirt is practically crimethink in the US.'

Sure it is, because no one in America has ever read these words - 'We the People of the United States, in Order to form a more perfect Union, establish Justice, insure domestic Tranquility, provide for the common defence, promote the general Welfare, and secure the Blessings of Liberty to ourselves and our Posterity, do ordain and establish this Constitution for the United States of America.'

'And that’s what made it durable in Pennsylvania in a way it wasn’t in Maryland."

A devastating Alt-Right conclusion that deserves consideration. '

Or ridicule, as it turns out that Pennsylvania and Maryland are part of the same nation, one with a notable devotion to the idea of religious freedom. You get to choose who is being more more ridiculous, though, whether Romer or the 'alt right.'

"The incidence of people who don’t learn to read is substantially higher in English than in other languages. People have known for a long time, it takes longer to learn to read in English because of the bad orthography."

This is a strong assertion. I was curious to read about it, what Mr. Romer tells is indeed a fact.

Apparently, our ability to read is controlled by our ability to split whole words into the individual sounds that compose them. In the link above there are several research quotes that point at the importance of the ability to hear and manipulate words for reading.

On one hand, English and its spelling makes difficult to learn to read. On the other hand, reading is not and innate skill, while walking or speaking are innate. So, do people who never learn to read is because "English is damned complicate" or simply the normal variations of mental abilities hinder them by not having enough skill to recognize sounds? or Both?

"And it’s a tragedy because English is becoming the universal second language." Here Mr. Romer loses the perspective. Come on, the English alphabet has only 26 characters, there's no ñ, ç, ä, è, ó, ê and many more. For people that already know the latin alphabet/script, it's easier to learn English as a second language compared to people who learned other scripts.

Spelling reform is the true mark of lunacy and megalomania, from George Bernard Shaw to the Chicago Tribune.

One little remarked effect is that when it is radical enough to stick it makes all pre reform books much harder to read. Qin Shihuangdi and Mao Zedong were certainly aware of this.

That's a negative portrait of a spelling reform.

I'll give a positive one. Remember civilizations/cultures that have lasted long have had spelling reforms because language is changed by people after some time, see France, Spain, China. If there's no spelling reform it means......the civilization was quite short-lived.

'Spelling reform is the true mark of lunacy and megalomania'

Oddly, no American ever seems to say that about a certain Noah Webster. Possibly because they recognize that 'recognize' is closer to how one pronounces recognise before Webster came along.

(But then, let us not get into a discussion of how to pronounce the letter 'z'.)

It payz to advertize.

Well, mocking Webster is like shooting fish in a barrel, since he is a notable contributor to the disaster that is English orthography.

Chequemate, right?

Clarity in communication? All one needs is a copy of Strunk and White, a copy of Harbrace, and, most importantly, a desire for clarity in communication. It's that last requirement that trips up many economists and ideologues; they have no desire for clarity in communication.

"Eschew obfuscation": words for our times!

This. Academics are notorious for using thick sentences and complex words for no reason other than that they think that is how academics should speak.

Consider: "After cogitating on the material, and with due consideration for its contents and meaning, upon reflection I find myself having material concerns that inform my need to reject the conclusions therein."


"I read your paper, and I disagree."

It's not English's fault that some people have a continual need to display their intelligence through dense prose.

There's a certain brand of academic who would always choose the former over the latter. Especially in an academic paper. And most especially in fields with a high ratio of bullshit to facts.

As for changing English... English is an emergent language. It isn't designed or controlled. And that is it's strength. Lord save us from control freaks who think they know better and want to force change on evolved systems 'for their own good.'

"I read your paper, and I disagree." Yes, that's good American.

In English it would probably be '"I've read your paper. I disagree."

The diminishing vocabulary of English, perhaps a result of an intellectually impoverished educational system and the phenomenon of "texting", can be daily observed. Unless something dramatic occurs to alter the course of events we can look forward to a discourse that has been reduced to the grunts and squeals of baboons.

After doing some Spanish or German on Duolingo one evening I tried out the Esperanto course just to see how silly it is. The egg was on my face instead. What a great invention that was.

Too bad it did not continue to grow. (Story is that wars, Nazis, and Communists killed it.) That would have solved a lot of problems with language learning.

I still play with it from time to time just for fun. There's something about the language that the brain just enjoys.

From the perspective of history, the dying Polish-Lithuanian Commonwealth is a hilariously bad place to invent a "universal" language.

English is the global lingua franca. Merely the fact that you don't have to convene a committee to determine the gender of inanimate objects puts it ahead of the closest competition.

Esperanto? Huh, try Volapuk!

The only sensible way to revise English orthography is so to arrange things that everyone in the world will speak with the clear, euphonic elegance of my own pronunciation of the tongue. Can anyone seriously expect every English speaker in the US, India, and so forth - including England - to change their accents? There would need to be elocution lessons for everyone but me. And even that would work for only a short time, until pronunciations started drifting apart again.

It's easy to think of a few simplifications that could easily be made, but few would contribute much help, would they? Do away with "ph" for "f" - OK, but is it worth it?

Would "k" for "ck" be an improvement? Yes, but to what end? Maybe you could then reserve "ck" for the "ch" in "loch". But then few English speakers can pronounce "loch" anyway.

I'm open to persuasion, but I've never seen a proposed scheme marked by a sense of proportion, or even of sanity.

..."clean up the orthography..."

But then we will miss our wonderful Spelling Bees!

I don’t think Romer was being that serious when he talked about reforming English. Tyler kind of took an off the cuff suggestion too seriously.

Maybe there is a connection between English'es orthography and its acceptance as a universal second language?

"There is actually a growing number of economists who are thinking about culture as the label for describing some of these more subtle effects that operate in our beliefs, in our preferences, our norms about right and wrong, and how those interact when we work together in groups. These are the things that we didn’t understand when we were kind of naïve about predicting growth in the developing world, and that we still don’t understand very well."
I love it when economists talk about something other social scientists have known for decades and act as if it were a new discovery. They should call this "econosplaining."

Good idea, especially as it would doubtless be pronounced 'conosplaining.

Romer's comments about language seem wrong as a number of people above have noted and his tastes in literature are banal. Makes you wonder about the other stuff.

COWEN: Should China and Japan move to romanized script? It’s very hard to learn how to read Japanese, especially a newspaper. You may not be able to do so until you’re 10 years old. And Japan has good schools.

I asked a couple of Japanese teachers about this years ago, and they said the brightest students do read novels and newspaper articles before their peers just as in Western countries. There aren't high percentages of 10 year olds reading a newspaper in the West, either.

Actually, only when you are about 20 years old you are really able to read a newspaper in Japan. That is because you only learn all the basic 2,100 different letters by the time you finish high school. So it takes about 12 years to learn how to read Japanese. That is, to read the non-specialized literature, in academic fields in Japanese they make up a lot of new letters, so there exists well over 10,000 different letters of Japanese and most people do not know more than a fraction of all letters: most Japanese, including people with PhDs, do not know how to read most Japanese characters.

It is easy to distinguish in Japanese material made for the wide public and children because it contains furigana. Children's books and manga for example, has the text written in proper Japanese and in simplified furigana scrip besides it.

English has bad orthography because it was so open (generally via invasion) to the entry of words from different cultures - from a Germanic base with Nordic loans then sprinkled with Norman French and later Parisian French plus tons of Greek & Latin loans.

Show me a country with a long-term monolithic non-invaded culture, and you have great orthography.

That said, can you really argue that English is so bad because “ch“ has 3 pronouniations versus tens of thousands of signs you need to learn in writen Chinese?

Poland, Polish orthography is very good, except it bears little resemblence to anything you would recognize.

Comments for this post are closed