I just came along this grammar book: A reference grammar of Japanese By Samuel Elmo Martin

http://books.google.ca/books?id=SszxbMtH...ru&f=false


there're so many things wrong with this book I don't even know where to start :lol:
can anybody find a grammar guide that's worse?

I bought this one a while back and find it pretty terrible too,

http://www.amazon.co.uk/Japanese-Compreh...176&sr=1-6

Although the one you have the link to beats it easily.

juniperpansy Wrote:I just came along this grammar book: A reference grammar of Japanese By Samuel Elmo Martin

http://books.google.ca/books?id=SszxbMtH...ru&f=false


there're so many things wrong with this book I don't even know where to start :lol:

Is this a joke? Martin's book is one of the foremost authorities on Japanese grammar. It's not intended for learning Japanese, but as a linguistic reference.

I don't think they ever taught it like that -- Martin did not write his book for the purpose of teaching students Japanese grammar. It's a linguistic reference book, intended to provide a linguistic description of Japanese. It is hard to understand if you don't have some grounding in linguistic terminology and explanations (the format is somewhat dense too; the layout could have been a lot better but is probably that way for reasons of space).

But faulting it for being complicated or not useful for learning Japanese is like criticizing RTK for not being a good kanji dictionary.

Why did he choose to use romaji? If it's intended for experts, hiragana and kanji should be no problem.

Romaji is the defacto standard in linguistics texts. You cannot get as precise details using kana without being verbose (eg talk about the k of ka).

Pretty much any linguistic work published in English uses romanization, no matter what the source language. Nowadays the romanization is often accompanied by kanji/kana in Japanese related publications, but in the 1970's when this book was published it was common to only use romanization.

Hiragana and kanji should be no problem for most people wanting to use Martin, but romanization shouldn't be any problem either.

I could understand for some specific cases but it seems a bit weird to me that you would write expert level analysis of a language and yet not even use the language's own script. If you're at the level where you can understand such a text, I expect that you should be able to at least read kana.

Well, remember that linguistics, in general, still considers a language as separate from a writing system. That is, in the typical linguistic description, kanji and kana are not part of the Japanese language, they're the writing system in which the language is usually represented. If you're doing a linguistic analysis of the grammar of the language, the script you use to represent it isn't relevant -- the material in Martin's book is no less valid because it's written with roman letters.

It's not that Martin didn't expect his readers to be able to read kana. There's just no compelling reason to break the standard linguistic convention and not use romanization. The pedagogical reasons for avoiding romanization are irrelevant because Martin is not a textbook.

As I said, today it's fairly common for both kanji/kana and romaji to be used, although it's not universal -- Tsujimura's "An Introduction to Japanese Linguistics", published in 2007, still uses romanization entirely, with no kanji/kana. (There are a few kana at the beginning of the chapter on phonetics, in a discussion of why kana cannot be used for phonetic analysis.) That's not because Tsujimura thinks her readers will not know how to read kana.

(Actually it's not limited to linguistics -- in my literature Master's thesis I used romanization.)

Let's not forget that sometimes linguists need to have a technical description of a language without actually being able to use the language. Being able to speak it is completely unnecessary and unrelated to some aspects of linguists. For that reason, even if kanji and kana are included, romaji should absolutely be included.

Yudantaiteki hit the nail on the head: in linguistics, writing is not nearly as interesting as speaking. There are over 6000 human languages on our planet, but only about 200 have a written system. The key aspects of language deal with how they are used to communicate verbally. (Consider when you read a newspaper, which is generally grammatically perfect. No one actually speaks like that.) Many linguistics books use IPA, but considering Japanese phonetics are very close to English phonetics (Japanese actually has less sounds), and this is a book written for English-speakers, I can't fault the use of romanization.

This kind of documentation is important as a reference, not to learn the language.

But what of this little gem? http://www.atrus.org/yokohama_dialect/sc...index.html

Jarvik7 Wrote:Romaji is the defacto standard in linguistics texts. You cannot get as precise details using kana without being verbose (eg talk about the k of ka).

How do Japanese linguists writing in Japanese get around this problem? (Do you not find Japanese texts which include bits romaji for specific descriptions effective?)

ydtt Wrote:Nowadays the romanization is often accompanied by kanji/kana in Japanese related publications, but in the 1970's when this book was published it was common to only use romanization. Hiragana and kanji should be no problem for most people wanting to use Martin, but romanization shouldn't be any problem either.

Quote:It's not that Martin didn't expect his readers to be able to read kana. There's just no compelling reason to break the standard linguistic convention and not use romanization.

What was the reason then for deciding to include examples in Japanese script? (I know that Martin's choice of romanization style without Japanese is a fairly significant inconvenience for many.)

I would have thought factors such as accessibility to linguists who don't read Japanese, book length/page space, and past typesetting limitations/costs would have been more significant than the idea that writing systems are less important.

Tzadeck has it right! My undergraduate linguistics specialty, syntax, was like that. Syntax is the study of grammars in general, so you don't really care about the parts of a language that have nothing to do with its grammar -- that is, structure and meaning. Syntacticians express sentences in a technical code so that even linguists who know no words of a language can understand what they're saying about its grammar.

Thora Wrote:

ydtt Wrote:Nowadays the romanization is often accompanied by kanji/kana in Japanese related publications, but in the 1970's when this book was published it was common to only use romanization. Hiragana and kanji should be no problem for most people wanting to use Martin, but romanization shouldn't be any problem either.

What was the reason then for deciding to include examples in Japanese script? (I know that Martin's choice of romanization style without Japanese is a fairly significant inconvenience for many.).

I have two answers for that one:

My Controversial Answer
In the 1970's, there was a general attitude that the Roman alphabet was the best way to write, and if the Japanese had any sense, they would switch over. Including written Japanese in books now is the linguist way of saying, "Sorry for that attitude. We're more PC now."

My Un-Controversial Answer
It's easier to look up Japanese examples on the internet and in real Japanese books if you have the actual words, rather than just their pronunciation.


I disagree with Yudantaiteki. I do, in fact, think that Martin did not expect his intended readers to be able to read kana and kanji.

Thora Wrote:(I know that Martin's choice of romanization style without Japanese is a fairly significant inconvenience for many.)

It really shouldn't be. If you're going to study Japanese linguistics, you're going to have to get used to reading Japanese sentences in romaji -- it's not something that should give an experienced person that much trouble, especially since translations are always provided as well.

Quote:I disagree with Yudantaiteki. I do, in fact, think that Martin did not expect his intended readers to be able to read kana and kanji.

You may be right.

One other reason for romanization is that linguistic articles are often used by comparative linguists or other people who don't necessarily know the language but need some information about it.

yudantaiteki Wrote:It really shouldn't be. If you're going to study Japanese linguistics, you're going to have to get used to reading Japanese sentences in romaji -- it's not something that should give an experienced person that much trouble, especially since translations are always provided as well.

I was asking why you think authors/publishers are more likely to include Japanese now (in addition to romanji)

The people I had in mind can manage romaji. ;-) (though probably not as well as people like yourself who spend the first 2-3 years reading Jordan's romaji.) With the various forms/hybrids of romaji out there, though, it can take a moment to infer the pronunciation, infer the word, and check the word against the English. Better to just glance at a Japanese sentence to get exact information immediately. Why muck around with English translations? (unless they're relevant to the point.)

Users of Martin's grammar aren't limited to linguists. Yale romaji isn't the standard or universally popular. Romaji is also harder to manually search/skim (Aethnen mentioned electronic searches). Incorrect translation is possible.

In my experience, people from various groups would prefer to have Japanese included (students, teachers, researchers, profs, translators.) Length is the obvious tradeoff. (I suppose digital texts with hypertext and customizable versions will make this issue moot.)

Until then, reading stuff like "zyuusuutuu no syootaizyoo"; "tyuuzyituni"; "sutahhu" is just unpleasant and, hopefully, unnecessary.

Quote:One other reason for romanization is that linguistic articles are often used by comparative linguists or other people who don't necessarily know
the language but need some information about it.

As Tzadeck and I mentioned. ( Aethnen too?)

Martin's grammar, though, isn't one of those linguistic texts Aethnen describes with the fancy codes and equations. It doesn't even present examples with romaji, English and labels neatly aligned for those who don't understand Japanese. So I tend to think the examples are aimed more at folks already familiar with the language (whether or not they can read it.)

Burritolingus Wrote:But what of this little gem? http://www.atrus.org/yokohama_dialect/sc...index.html

I'm normally just a lurker but that thing is hilarious. Skip to this page:

http://www.atrus.org/yokohama_dialect/sc...ew_09.html

Seriously, is this thing a joke??


EDIT: ok, i clearly missed the earlier thread

Ah I remember I saw this in my library, it was HUGE, It was almost as fat as a kanji dictionary dictionary a few rows underneath it, it is very heavy, linguistic text or not, having kana/kanji would really break up the flow . Doesn't look like it would be too much of a pleasant read but there does look to be lots of cool bits of info cluttered here and there. (i.e a reference)

I'd prefer if they had the romaji style where they have a line over the extended vowels, it looks silly seeing shit like "zyuusuutuu no syootaizyoo".

As a reference, if it were say.. $20AUD I'd buy it, but I don't think it's viable to actually buy it and study the whole thing.

Available from these sellers.
2 new from $334.30 4 used from $205.00


okay get bucked



and lol @ burritolingus