Japanese! The fastest language ever!

11 Jan. 2014

Spoken word

I’m 15, it’s the summer of 2007, and I’ve just arrived in Japan for a 6-week homestay program made possible by my stupidly awesome high school. At the end of the first week I get an email from my sensei back in America asking me how I’m doing, etc.

“Everything’s wonderful,” I reply, “but why do Japanese people talk so fast?!”

Eventually I got used to it – right around the time my summer ended.

But the question sat quietly in the back of my mind ever since. I chalked some possible answers up on the board:

- the syllable structure flows with lots of short-length vowels, making it possible to speak a thousand words a minute. English words don’t let you go that fast. For your own safety.

- the words themselves are just so long. Speak any slower and you’d forget what you were trying to say by the time you got to the middle.

- related to the above – the words might be long and full of syllables, but given context it’s easy for native speakers to fill in the blank once they hear the start of the word, letting the speaker move much faster without overwhelming the recipient with information.

- and finally, because maybe that’s just the way it is.

I never looked any deeper in to the question.

Until today! When one of my favorite reddits linked to this interesting article from 2012: why do Japanese people talk so fast?

It’s worth the read. Because it turns out Japanese people talk 27% faster than English speakers.[0]

Koichi has some cool theorizing on the reasons behind some of this in the article I linked. Something he brought up that I didn’t realize: Japanese has just 416 possible syllables, compared to 1,191 in Mandarin, and 7,931 in English.

My most important takeaway from the study: Japanese has the lowest “information density” of the eight studied languages. I.e. you have to use more syllables to convey the same amount of information.

As you might expect, Japanese speakers spoke the fastest – used the most syllables-per-second – but despite that fact, the language still manages to convey ~75% of the actual information conveyed by all eight other languages.

Perhaps this suggests there is simply a physical limit on syllabic rate and Japanese sits at that threshold. Or maybe the limit is on listening comprehension – 15 year old me would certainly agree with that. Either way, it’s seems that other languages allow a greater rate of information transfer – in fact, every other language transfers information between speakers at essentially the same rate. Except Japanese.

So to increase Japanese’s “information rate,” you could either speak faster or use fewer syllables to convey the same points. Speaking faster, well. My untested armchair hypothesis is that speaking faster is impossible and that Japanese already exists at the limit of rapid speech.

Use fewer syllables? Well, perhaps this is where some outrageous (to foreigners) Japanese contractions come in. To some degree, I think all languages do this – contract and manipulate their language (when speaking with the rest of your “tribe”) until the “information rate” is maximized.

Will not to won’t.

たべました to たべた

But even then, Japanese is fundamentally constrained by its available syllables. With only 416 combinations, mostly consonant-vowel pairs like “KA,” “KI,” “KU,” “KE,” “KO,” the language struggles to abbreviate and shorten words without mis-communicating entirely.

Japanese is missing many of the consonant-vowel-consonant options available to English – something that I found often caused my fellow high school students in Japan, who at the time were studying English, to struggle. “Sing”, for example, is not a valid construction in Japanese. You can’t just end with a “G” like that. Nor is my name for that matter. Conrad was written and pronounced “Con-ra-do.” A commenter gave the English example: “wolves.”

Written language

So here’s where it gets interesting. And even more hypothetical.

If you ignore Kanji for a moment and imagine written Japanese across all forms of media – newspapers, magazines, books, essays, signposts, ads, and more – I think the situation would be quite similar. Could we find another language that manages to write so much, yet convey so little?

I don’t know. But I think it would be tough. And it wouldn’t be English.

Why is this interesting?

Because written Japanese is actually incredibly information-dense. Reading the newspaper for non-native speakers is an exercise in painstaking translation and spot-the-hiragana.

Because Kanji.

Kanji. A writing tool appropriated from Chinese with its 1,000+ syllables and normal rates of “information transfer,” and then grafted onto Japanese.

One small character suddenly replaces two, three, or more characters in our hiragana-only script. At its heart, I think this is why I found studying kanji so difficult. I could speak the language, understand what was being said, and I even had a script to write it all down in. But it wasn’t enough.

The “information density” contained in a writing tool like kanji has to make up for the abysmal density of pure hiragana. And it does. When you can achieve some measure of kanji-literacy, reading becomes a breeze. Ambiguity disappears. Paragraphs shorten and become faster to read simply because your eyes don't have to travel so far. But until you’ve achieved that mystical level…damn. It’s hard to stop feeling angry at whoever decided to appropriate the Chinese script.

Of course, writing Japanese without kanji is like writing English without….what? What other language has a reasonable analogue? I’ve not been able to find one. Not in our Latin-derived languages. Not in Chinese. Not in Korean. Where else can you write down everything you want to say, and then go back and replace the majority of characters or letters with a symbol that represents groups of individual characters?

It would be like replacing the word “burger” with the McDonalds logo, that big curvy yellow “M.”

Except of course sometimes the big curvy “M” means cloud, or yellow, or sand dune. So sometimes you see it and say “burger,” and other times you see it and read aloud “sand dune.” It breaks the link between visual and audio learning, because you have to learn all three readings, even if you hear one every day and the others only once a month. And so once a month you find yourself reading something that doesn’t make sense, stop, pause, go back, think think think-

Ah. Of course. “The camel crossed the sand dune.”

Facetious, I know, but not wholly inaccurate.

And I think that’s what makes kanji so difficult for non-native speakers: a combination of depth (3,000+ to memorize), inconsistent visual and audio patterns (multiple readings), and sheer information density. But those same traits also make it worth studying.


[0] Note that the study’s sample size is neither particularly large nor particularly small. I’m taking their findings as the literal truth here instead of debating the study’s demerits. Given the surprisingly consistent data for all languages except Japanese, I think – regardless of the degree – it’s reasonable to conclude that Japanese is materially different.