Chapter 17 – Learning the language: Problems



1. The Devil’s Language
2. Listening Problems
3. Chinese Connections
4. Chinese-origin Homophones
5. Bad Romanisations
6. Reading Japanese

That one year in Japan – 1967-68 – was to be even more difficult than the two years in Moscow.

And not just because I had to organise myself into one of the world’s more complex societies, without the help of an embassy, contacts, friends or even much in the way of funds.

I had also to break through into a language much more difficult than Russian, and with just one year to do it.

And as with Russian, I had to do it myself, without teachers or the standard 12 months of intensive instruction most would-be Japanese speakers receive to get started.

Someone once described the difficulty of learning Japanese as like having to climb Mount Fuji, in winter, without crampons or a stick.

I was lucky since knowing Chinese I could start climbing from the bus station about a third the way up the mountain.

Even so, I still had a long way to go. And I lacked the walking stick of good learning materials (now much more available than in the past).

I admire those who start from the bottom, and persevere all the way to the top.

1. The Devil’s Language.

As others have noted, beginning with Francis Xavier in the 16th century, Japanese is not an easy language (“a language invented by the Devil” was his conclusion).

In many ways it is even more difficult than Chinese.

In Chinese, the grammar – word order, sentence structure etc – is not very distant from most Western languages.

Even the four tones in Mandarin Chinese can be handled once you get used to them.

With Japanese the grammar tries to push everything into reverse. 

As well the language is a strange mixture of the original indigenous Japanese (the scholars are still trying to tell us whatever that was), Japanised imported Chinese words, and partly Japanised English words.

The writing is a mixture of the hiragana colloquial script and the more formal katakana script, combined with imported Chinese ideographs (kanji). 

On most scales of difficulty Japanese registers at or near to the top.

2. Listening Problems

Problem number one (for me at least) was not being able to distinguish individual words in the torrent of speaker sound.

Words, especially the homophones of Chinese origin, can be agglutinated, and even more especially when the speaker begins to blurt out streams of disorganised thought, which can happen often.

You are facing a wall of sound with few pegs to help you start climbing.

Another problem is the style of communication. The Chinese communicate with directness and force which helps drive the words into your memory.

With Japanese, you may well hear everything but the more indirect style of speech means words do not impact the memory with anything like the force of Chinese.

Quite a few foreigners, even those not brought up as children in China, get to speak good Chinese. Few foreigners in the same situation manage the same fluency in Japanese.

Polite Language

The need to master polite language in Japanese is a problem, though not as difficult (today at least) as some claim. 

And as with English there is no great problem in having the language operating at two levels – a simple  vocabulary and another more sophisticated, polite vocabulary for the same things and feelings, with subtle nuances operating in between.

So in Japanese the simple word for ‘to be’ is desu. The polite word is more convoluted – irashayaru. 

Personal pronouns and verbs change depending on the rank or familiarity of the listener, for example with the passive form of verbs used when speaking in formal situations, or to superiors.

(The use of the passive is a feudalistic hangover, very similar to its use in English: “Will the master be having his tea early today?”)

When the TV announcers report on Japan’s pampered royalty their mouths get clogged up with passives.

Male and female speech can also differ, though not as much as in the past.

3. Chinese Connections

Japanese, like Korean, and Vietnamese and some other Southeast Asian languages, is a mixture of Chinese origin words mixed into the indigenous language.

But unlike the others, which have moved the Chinese words into their indigenous written language, in Japanese they remain as ideographs known as kanji.

Most kanji are written much the same as in the original Chinese (a big help for me).  

Many complain, and some – like former Japanese scholar Edwin Reischauer who should have known better – tell the Japanese they should put the kanji into a phonetic script – into roman letters, or the native kana script.

But kanji allow fast reading on the same principle as speed reading in English – recognising meaning by the shape of a word rather than the letters in it.

My guesstimate says that the year students lose learning kanji they make up later by faster and more accurate reading.

For me as a frequent reader of Japanese the idea that I would have to read without the help of kanji would cause problems, severe problems. 

In any case, the homophone problem rules out any phonetic script replacement for kanji – though I imagine the Korean technique of using their phonetic script to put kanji pronunciation into boxy shapes resembling ideographs, allows a form of speed reading.

In all the Chinese-influenced Asian languages, Japanese especially, the mixture of the Chinese origin words with the native language resembles closely the way English mixes imported Romance origin words with the original colloquial language.

In English we use Romance origin words mainly for educated speech. The Japanese, Koreans and Vietnamese do much the same with their Chinese origin words.

And as with English there is no great problem in having the language operating at two levels – a simple native vocabulary and another more sophisticated vocabulary for the same things and feelings, with subtle nuances operating in between – ‘sad’ versus ‘distressed’, for example.

Two vocabularies enrich the language  – one reason why English and Japanese literature have such a following, perhaps.

4.  Chinese-origin Homophones

But when the Japanese imported those Chinese origin words, for some reason they seem to have wanted to turn the original Chinese pronunciations into the barest of monosyllables, many with very similar pronunciations.

Thrown together in rapid conversation they can be almost impossible for the unpracticed ear to unravel.

The Japanese sound ‘ko’ for example is the monosyllable used to pronounce a dozen or more Chinese-origin words in common Japanese use (there are another dozen or so more ‘ko’s’ for Chinese origin words with more arcane meanings).

The Chinese pronunciation for those words in common use is gong, gang, kang, kung, hong, huang, guang, kuang, gao, kao, hou, hao, jiao, qiao, and jiang.

But in Japanese they all end up as ‘ko’.

True, Chinese also has its homophones — words with the same pronunciation but different meaning. 

But when spoken most can usually be distinguished from each other by one or other of the four tones used in Mandarin Chinese.

So gong with a high tone means “work”. With a low tone it can be a province in Szechuan, and with a rising or falling tone it has two other meanings.

But Japanese does not have tones (though the Japanese insist that at times there are subtle differences of emphasis).

Usually you have only the context to guide you.

To make things worse in Japanese we have the problem of short and extended vowels.  

If, for example, the ‘o’ sound in ‘ko’ is short we find one set of meanings. If it is lengthened (koo), we have another quite different set of meanings.

The ‘ko’ with the long ‘o’(romanised as ‘koo’ but I prefer ‘kor’) and the ‘ko’ with short ‘o’ both sound much the same. 

The only difference between them is that the short ‘ko’ is clipped (as in the English ‘or’) and in the long ‘ko’ it is extended (romanised as koo or kor with o pronounced as in the English ‘oar’).

You have to be able to hear the difference to know which ‘ko’ they are using.

Most of the other homophones from China ending in a ‘o’ or a ‘u’ also have both long and short endings. 

So a ‘sho’ or a ‘shu’ with short endings have one set of meanings. With long endings, sometimes romanised as ‘shoo’ and ‘shuu,’ the meanings are quite different.

So the ‘sei-sho’ above can mean the Bible, but ‘sei-shoo’ (the ‘o’ is long) means ‘sing in unison’.

Mistake a short ‘u’ for a long ‘u’ and instead of talking about problems with your husband (shu jin) you are talking about the prison convict with whom you are living (shuu jin).

More Problems

Even when they are not homophones, many of the various Chinese-origin words sound alike when brought into Japanese.

Shi and sho, are close in pronunciation. They are used for dozens of Chinese-origin words or compounds.

So you are fighting not just a bunch of homophones but a whole army of their relatives also.

Compound Words to the Rescue?

Fortunately the Japanese, like the Chinese, rely heavily on compound words — two homophones brought together to create one word.

But even then there are difficulties, since many of those compounds themselves can end up as homophones.

Take the Chinese origin word for ‘success’ which is pronounced ‘sei-koo’ in Japanese. A Japanese dictionary will give you a dozen or so other meanings for ‘sei-koo’ including ‘sexual contact’.

In traditional Chinese romanisation, the ‘sei-koo’ meaning success is ‘ch’eng-kung’. The ‘sei-koo’ for sexual contact would be sheng-chiao, and so on.

But in Japanese they all end up as ‘sei-koo’.

High tone ‘gong’ may have several meanings in Chinese. But if it is combined with ‘ye’ (enterprise) to create the compound word gong-ye we know fairly unambiguously that we have a word which means ‘industry.’

In Japanese, the kanji for gong-ye become koo-gyo which also means industry. But koo-gyo can also have several other different meanings – ‘entertainment event’ for example.

True, English has its homophones – current and currant, for example – and as with Japanese we rely on context to avoid confusion.  But the number of them, and confusion possibilities, are far less 

Japanese speakers are sometimes reduced to writing out the kanji in the air when pronunciations become too similar.

Another problem, for me at least, is remembering reversed compounds. For example there is ‘sei-sho’. But there is also ‘sho-sei.’ There is ‘sei-sha’ and ‘sha-sei.’ Or ‘shuu-sei’ and ‘sei-shuu.’

 ‘Shoo juu’ means a rifle. ‘Juu shoo’ means a severe injury. And so on.

‘Shuu shuu’ gets to mean to ‘collect information’. But it also means ‘to control’.

‘Shoo shoo’ means ‘a little’. ‘Sho joo’ means ‘symptom’ (it also means a ‘certificate of merit’).  ‘Sho jo’ means a young girl.

‘Joo shoo’ means to ‘to rise.’ ‘Joo joo’ means ‘excellent,’ a ‘stock-market listing’ and ‘circumstances’.

‘Sen sen’ means to declare war. But it also has a literary meaning – to be washed away by clean water. Mistaking that pair could cause problems.

If Francis Xavier had the same problems, he was not exaggerating.

Even today, feed those romanisations into a computer dictionary and the chance you will get a correct translation is dubious.

If a computer cannot handle them how are us humans supposed to manage?

5.Bad Romanisations

Another problem for beginners at least is that the standard Hepburn romanisation system, unbelievably, does not care to distinguish between those long and short vowels.

So in Hepburn the ‘sei-sho’ (short ‘o’) above meaning ‘Bible’ is romanised as ‘seisho’, and the ‘sei-shoo (long ‘o’) meaning ‘sing in unison’ is also romanised as ‘seisho’.

Under Hepburn when you see any ‘sho’ in romanisation you have to try to remember consciously whether the ‘sho’ should be elongated or not. 

And if it is elongated (shoo) then you have to struggle dozens of other shoo-s to decide the meaning.

 There is nothing but context to help you.


My theory as to why Mr Hepburn left us with this can of pronunciation worms says that in those days they decided Japanese was hard enough as it was without us having to remember long and short (o)s, (u)s etc. 

Or maybe they themselves just did not notice that there was a difference.

In fact, the difference is very important. Many foreigners fail to realise this and mess up their spoken Japanese badly as a result.

For example, the first ‘o’ in the word ‘Tokyo’ is a long ‘o’ and should be given a romanisation to make that clear –  ‘tookyo’.

(A short ‘o’ could be romanised as ‘oh’.)

But under Hepburn all are romanised simply as ‘o.’ 

Another problem is the way in English we have a variety of pronunciations for vowels.  In Japanese there is only two possibilities – long and short.

But the ‘o’ in English can also be pronounced as ‘oe’ as in ‘so’ or ’toe’.  

The first ‘o’ in Tokyo is should be pronounced with a long ‘o’ – Tookyo . But we usually pronounce it as Toekyo. As a result, whenever students see or hear an ‘o’ in romanised Japanese they like to think it should be pronounced as ‘oe’.  

So just as the popular pronunciation of Tokyo is Toekyo, the popular pronunciation of the short (o) word in gomi (rubbish) has become ‘goemi’.

Some of the older generation of Westerners studying Japanese with Hepburn are imprinted not just with the wrong pronunciation for the letter ‘o’ in romanisations but also with the belief there is no need to distinguish between the long and short ‘o.’

It’s a mess.

For those like myself who tend to rely on romanisations to remember language (i.e. visual memory is strong)  it is a nightmare, from which you can only escape by a conscious effort to imprint the correct pronunciations of words the moment you first see them.

Wrong pronunciations imprinted in the sub-conscious can stay for life. No matter how many times later you hear the correct pronunciation it will be hard for you to change.

Like my neighbour who has been in Japan thirty years but still says ‘goemi.’

Fortunately when I came to learn Japanese I was aware from the beginning of this imprinting problem, thanks to my experience with other languages. 

6. Reading Japanese – ‘Listen’ to what you read.

As already mentioned, when I set out to do my post-graduate research on Japan both I and my mentors assumed my knowledge of Chinese would make it fairly easy for me to read research materials in Japanese.

We were both wrong. Like the spoken language, the written language also has the Devil’s fingerprints all over it.

It is a major obstacle to progress in learning the language.

Jumbled Print

Problem number one is the way the print is jumbled, with the Chinese-origin ideographs (kanji) mixed with two phonetic scripts – katakana and hiragana.

Katakana is used mainly for imported words, mainly from English. Hiragana is more for indigenous Japanese words or parts of speech.

Katakana is sometimes compared to the way we use capitals in English; hiragana is lower case.

So when you read Japanese, it is as if you were jumping from one word in capitals to another in lower case with the kanji ideographs thrown into the mix.

Kana are fairly easy to learn and remember. Not so the kanji.

You will have to learn to recognise about 1500 or more if you want to read serious material.

That takes quite a bit more than a year.

Reading Kanji

The good news is that you do not need to learn how to write kanji, though many try to. It is enough to be able recognise them.

Nor are kanji quite as hard to recognise as they seem on first impression.

They are made up usually of two or three of about fifty standard components, and if you know those components then when they are pulled together to create an ideograph you get to recognise the ideographs as you recognise pictures.

It is rather as if I see a Picasso on a wall I will know immediately it is a Picasso. I am not likely to confuse it with a Da Vinci.

All this helps to make for fast reading, though it would probably be even faster if there were no phonetic kana cluttering the page.

Agglutinated Script

But while the kanji can make for fast reading, for some reason (probably inherited from the way Chinese is written) there are no breaks to indicate different words. 

Everything – kanji, hiragana, katakana – are all run together in each sentence, without a break to tell you where one word ends and the next begins  (fortunately there are sentence breaks).

The result is rather like this:

ItisalmostasifIwastowriteasentencewithout breaksandthenaskyoutoreadit.

Only a fluent English speaker could intuit where the breaks between individual words should occur, namely:

It is almost as if I was to write a sentence like this and then ask you to read it.

Even more than with most other languages you need to have good listening/speaking ability to read Japanese.

Without that ability students will end up like me many years ago – pouring over scripts for hours trying to untangle the scribble before you.

But if you have speaking/listening ability then all you have to do is turn the scribble into sound and the meaning emerges naturally, as if you were listening to someone reading that text to you.

That language is sound rather than scribbles on a piece of paper is something many have yet to understand.

But for proper speaking ability there is still one more hurdle: language sound has to be embedded in the sub-conscious.

We can come to that problem later.