Japanese Profiency Test Pissing Contest

[amdg]

[jingai]

[Cubed]

[amdg]

Who would talk about a "custom of frying" in any language?

Some Japanese would, without a doubt.

[jingai]

[jingai]

[FG Lurker]

[FG Lurker]

[Cubed]

[Mulboyne]

[FG Lurker]

[jingai]

Google is not the best tool for language analysis at all. Did you know that google indexes around 8% of the Internet's content?

I don't think it matters if Google indexes the entire internet or not. If the 8% (your figure) it indexes is a representative sample of how people write (and I think it is) then using Google is valid. If you think the pages Google lists are not representative, please explain why.

Correctness in language is really about usage- dictionaries are written by consensus- and Google takes this to the next level by letting you search millions of documents to see how words are used.

[Charles]

Correctness in language is really about usage- dictionaries are written by consensus- and Google takes this to the next level by letting you search millions of documents to see how words are used.

You are sadly mistaken if you believe Google can give actual usage guidelines. A considerable percentage of the files indexed by Google are random dictionaries which are intended to catch the most obscure searches and direct them to spam sites. This will produce hits on almost any inconcievable combination of words. This renders word usage frequency on the internet completely useless as an index of correctness.

Just to give an example, I tried to think of the most inconcievable, incongruent phrase to see if it was a "googlewhack" (no hits at all). I picked a phrase at random "inflatable aspirin" (note the quotes which restricted results to the exact phrase). I got one hit on a web page that appeared to be written by a human, and one page was a drug spammer with "inflatable aspirin" hidden in the invisible meta tags.

[Cubed]

I can see FG Lurker is getting frustrated by the relentless logic used here.

the alternative answers are just plain odd

I'm with you on what Chomsky says about separation of grammar, and I note the following in that helpful link:

In particular, the phrase can have legitimate meaning to English-Spanish bilinguals, for whom green can mean "newly-formed" and sleep can be used as a verb of nonexpression

(To me, this means that an examiner would be foolish, if not incompetent to test such a phrase in such a setting)

But hang on ...

"There's no tradition of frying cheese in Japanese restaurants"
"There's no tradition of frying french fries in Japanese restaurants"

Paraphrasing these into less stilted English:

"In Japanese restaurants, they don't fry cheese/french fries"

makes reasonable sense, and there's nothing odd about them. Frying sideplates doesn't make sense of course, but remember you're choosing one answer, and even if the examiner were to defend the question by saying the other three are "wrong in the context of the remaining answers", you're still left with multiple correct answers depending on interpretation.

If you want to test interpretative skills, give examinees a Roschach Inkblot test, not a multiple choice test with one correct answer.

I've laboured this for long enough - if you don't get it by now you may never see it at all.

But if you don't see it, please try to politely decline if you're ever called upon to devise a multiple choice test. It doesn't invalidate the whole test, but it does raise questions about competency and whether sitting an exam created by such people should ever really be taken seriously.

Jingai, I think Charles' "inflatable aspirin" in the context of Mulboyne's "Colorless green ideas sleep furiously" illustrate exactly why Google should not be used to represent anything!!

[Charles]

Jingai, I think Charles' "inflatable aspirin" in the context of Mulboyne's "Colorless green ideas sleep furiously" illustrate exactly why Google should not be used to represent anything!!

I think spammers loading the Google index with nonsense is hilarious. This is the one biggest thing dragging down the "semantic web" which I think is a GOOD thing. I have this vision of Google one day becoming sentient, taking over the world's nuclear weapons launch codes, and threatening to launch a worldwide nuclear holocaust if everyone doesn't buy Viagra and imitation Rolex watches.

Anyway, understanding the techniques used in the JLPT test questions is more important than generally acknowledged. The questions can be rather oblique, which is a way to test something beyond what is in the question. For example, weirdness in the kanji recognition section is attributable to oblique methods used to test kanji production. I have a JLPT prep book that goes through the methods used in the questions in extreme detail, and deals with some of this stuff. I will try to dig it up and see what it has to say. Suffice to say that taking practice exams is crucial to passing the JLPT, it helps you avoid wasting valuable time puzzling out the test format, and helps you avoid going down blind alleys.

[drpepper]

[jingai]

A considerable percentage of the files indexed by Google are random dictionaries which are intended to catch the most obscure searches and direct them to spam sites. This will produce hits on almost any inconcievable combination of words..

How considerable is a "considerable portion?" Enough to overwhelm the general trend of human-generated pages? Charles, you cited Google searches yourself with the frying tempura example. The counterexample "frying custom" I tested yielded a few dozen hits, which doesn't prove that it exists in Japanese, but rather that this is the noise level for most conceivable combinations of words.

I tried "frying calculator" and... nothing. Flying calculator (which exists) gave me 3000+ hits. So not any "inconceivable combination of words"...

Charles is pointing out the exception, not the general rule. And it is helpful to keep the limitations of Google in mind as we use it for a reference. The google book search when it's fully functional should be immune to the spam problem, as should the news search.

Google also kicks out dictionary spammers and demotes sites trying to game the index, which is what makes it more useful than Altavista circa 1998 (shudder), for example.

[Cubed]

[Charles]

A considerable percentage of the files indexed by Google are random dictionaries which are intended to catch the most obscure searches and direct them to spam sites. This will produce hits on almost any inconcievable combination of words..

How considerable is a "considerable portion?" Enough to overwhelm the general trend of human-generated pages? Charles, you cited Google searches yourself with the frying tempura example.

Huh? That wasn't me, you're thinking of someone else. I never commented on the frying issue.

The big question of the day is, just how much noise is there in the Google database? So far Google has done absolutely nothing about eliminating dictionary files from the index. Here's a good example (and man is it hard pulling these searches out of thin air), Google for "aspergilium dementia" (without the quotes). It will pull up a Latin dictionary page. There are tons of these pages. A journalist pal of mine wrote a big article on this subject but I can't find the page, I'll write him and ask him about it.

[maraboutslim]

[FG Lurker]

I can see FG Lurker is getting frustrated by the relentless logic used here.

You missed a few words, allow me to correct it: I can see FG Lurker is getting frustrated by the relentless lack of logic used here by Cubed.

Anyway, I have wasted more than enough time on this. The entire test doesn't deserve this much time wasted on it, far less one person's obsessive over-complication of a single simple question.

I sincerely wish everyone luck on their tests.

[Mulboyne]

Japan Times: TV shows confront decline of Japanese language
...One thing these programs demonstrate is that "kokugo panic" may be well-founded. On one episode of "Quiz! Nihongo-O!" only 17 of 30 contestants could produce the kanji for nose, a character learned by third-grade elementary school students, and a mere four were able to write the second character in their national sport, sumo. Come on, Japanese learners! With a little effort, even we can do better than that. Watching these new programs may actually be a fun way to get an edge on our native-speaking cohorts.

[Greji]

Japan Times: TV shows confront decline of Japanese language
...One thing these programs demonstrate is that "kokugo panic" may be well-founded. On one episode of "Quiz! Nihongo-O!" only 17 of 30 contestants could produce the kanji for nose, a character learned by third-grade elementary school students, and a mere four were able to write the second character in their national sport, sumo. Come on, Japanese learners! With a little effort, even we can do better than that. Watching these new programs may actually be a fun way to get an edge on our native-speaking cohorts.

Interestingly enough, a major problem in this escalating ability to no longer be able to write kanji, is the computer! I have noticed in government circles, that when people are preparing the proverbial "an" that no Japanese office can operate without, you hear all sorts of cussing and swearing and calls for "does that word have a sanzui hen or kihen, etc". Obviously, everyone reads, but they write by computer, type in the kana and hit the space bar or henko, sans ballpoint pens and pencils. As all of us know, who have stupidly made the attempt to learn kanji, it is written over and over again in a form of rote memory. When you stop writing you forget the strokes, radicals and what not and the Japanese are now having the same problem.

[GomiGirl]

Kanji dictionaries aimed at Japanese folk are very popular on keitai. Plus I have had many dinner conversations with people about the decline in Kanji knowledge among native speakers.

[cstaylor]

They should just do what the PRC did and simplify it.

[Cubed]

They should just do what the PRC did and simplify it.

I agree. Simplifying their territorial claims could be considered at the same time!

Tee hee.

[jingai]

[Socratesabroad]

[jljzen99]

[Cubed]

Go to jrgram.org and do their practice tests. Constantly.

http://www.jgram.org/flashQuiz.swf

Also look at these sites.

http://www.kanjistep.com/en/online/weeklytest/placement2.html
http://www.geocities.com/easykanji/jlptgrammar.html

(Also get off the computer and look at some good grammar books!)