So I decided recently, in order to round out my vocabulary, that I would read articles about one specific topic a week. In other words, this week I'm reading articles exclusively about astronomy, and next week I'll probably read articles about language or history.

What I'm wondering, though, is whether or not it's really necessary to create sentences for topic-specific terms. When dealing with words such as 低い or 思い or whatnot, obviously context is necessary to understand how these words are used. But let's say, while reading, I learn the word like 星間塵 (cosmic dust). Do I really need a sentence to help me remember how to use this word? Or, is adding a sentence like "星間塵による光の吸収の効果を考慮していなかった" going to help me remember the word any better? Wouldn't just adding the word 星間塵 on one side, and cosmic dust or the Japanese definition on the back be sufficient?

I'm interested in hearing everyone's opinions about this.

hmmm. Well in my personal experience sentences no matter short,small,one-word containing context in them have always helped me remember that context of how to use that sentence+understand it better. Sometime words are easily learn separate, but it's best to learn them in context. It's just like saying you know how to count from 1-100 perfectly fine and you know how to write+read+understand them. But when it's used in a sentences you'll be trying to decode what's going on and what the context is about. What I'm trying to get at is, it's better to learn in context than in a separate word. But learning separate words doesn't hurt, but be warned you might not always be able to understand it if it's used in context.

If you are studying words with few or no synonyms, context isnt all that important and you should probably just go with a production vocab card. I like to do a check for synonyms sometimes by using eng->jp in wakan and typing in a word or two from the definition to see what pops up.

I agree

When I'm working on just specifically defined nouns, such as scientific, philosophical, etc terminology I don't care to put in a sentence.
For adjectives and everyday language I do however.

I'm certain that it's not necessary, however, why wouldn't you put a sentence in?

Presumably, it's the same amount of effort, and you will pick up words that tend to go with the target words.

I don't think that sentences are always necessary for learning these kinds of terms, but they can still help a lot. I find that the context of a sentence or phrase helps a lot just by reinforcing collocations, grammar, and the context in which the word or term would be used. Also, doing just vocabulary cards for these kinds of words seems kind of similar to just trying to memorize the kinds of definitions you find in science textbooks or specialist dictionaries, which to me is a lot more boring than learning the terms in context.

I think a lot of scientific/technical vocabulary doesn't need it, however I usually just add the whole sentence just in case. To me it's easier to just deal with the sentence, than trying to figure out if I need a sentence or not.

Because most scientific/technical words are so well defined, it might be worth making "production" cards for them. Definition > word, or maybe even English > word.

On the other hand, science/technical stuff (at least at high levels) are usually in English anyway...

zer0range Wrote:I'm certain that it's not necessary, however, why wouldn't you put a sentence in?

I find it's not all that easy to find a sentence that's reasonably short, usefully demonstrative of the word, and where you're not likely to find yourself later failing the card because of some other word in the sentence than the one you were interested in. So in summary, "speed of creating the new card".

I don't know if the sentence method alleviates this kind of problem, but there are a lot more than definitions in language even in scientific context. For example, I can easily pick up on translations of advanced college math textbooks etc. done by non-mathematicians because they just sound like Japanese written by advanced learners who learned the language through word lists, textbooks and whatnot. Grammar is all right in a sense that they make sense, but collocations, word choice, sentence structures are always unusual. So the exact same thing in a scientific sense may not work the same way if you look at it linguistically. Actually sometimes you can tell if a Japanese person learned advanced math at a Japanese graduate school or in a foreign country from his or her wording in his math proofs in Japanese.

When scientific terms are used in text/speech for non-experts, the difference between literal translation and natural Japanese becomes larger for obvious reasons. It's like light travels faster than sound in English, but in Japanese 光は音よりも早く伝わる, i.e., light reaches the destination faster than sound in Japanese. Heat, electricity, etc. also collocate with 伝わる. They are not "conducted." Collocations and such are affected by everyday language to a larger extent, and definitions can be different than in strictly scientific context.

Things get worse in non-scientific context. You just can't expect translation would work. For example, you can learn a language by osmosis in English, but 言語を浸透によって学ぶ sounds so unnatural it doesn't make any sense. In Japanese you say, "自然に吸収する." It doesn't matter if "osmosis" is a scientific word.

Also, I think collocations and stuff exist not only between words, but also between phrases, sentences, and so on. Traditionally, translation into Japanese is done so it roughly follows sentence order in the original text, resulting in confusing Japanese. Have you ever felt like Japanese authors ramble in their writing and don't make points when you read a book written by a native Japanese speaker for the Japanese? Actually "make a point" is sort of an exotic idea to the Japanese. In traditional good writing in Japanese, authors amble zig-zag without explicitly making a point while implying opinions, points, etc. In other words, a point is always made between lines. What the author really means is "hidden," and the reader is supposed to "sense" it. This kind of cultural difference is less problematic in scientific text, but you can't get away with it.

I don't know if you can learn Japanese faster by putting sentences for scientific terms in your SRS, but I'm sure you're underestimating the amount of exposure you need if you think scientific terms require significantly less context to learn. You're supposed to learn if each word works the same way in a linguistic sense as in your mother tongue through exposure.

No, I don't think such terms need any sentences in your SRS. Of course, like magamo is saying, that doesn't mean exposure and context is unneeded, it just means it doesn't have to be in your SRS. One still has to read a lot of texts concerning it.

Magamo is right but what difference does it make if you srs that single word then just read heaps later on?

liosama Wrote:Magamo is right but what difference does it make if you srs that single word then just read heaps later on?

Apparently it depends on your learning method. The OP said,

mirina Wrote:Do I really need a sentence to help me remember how to use this word?

If he's relying heavily on SRS software to learn word usage, collocation, and whatnot, probably he needs at least a few sentences that illustrate how those words work. You would need a huge pile of cards to do that, but maybe it works for some people. If he was doing a sentence picking a la AJATT, he would only mine very very interesting ones that he really likes from the bottom of his heart. So this kind of question wouldn't come up in the first place, i.e., whether scientific or not wouldn't matter when an hardcore AJATTer puts a sentence into an SRS. The only deciding factor would be whether he really wants to remember the sentence.

An SRS can be used for various purposes from cramming kanji to memorizing favorite quotes to non-language studies. It all depends on how you use it.

By the way, I think some people are taking the 10,000 sentence thing in a very stupid way. I think the author of AJATT meant that if you're as picky as him when mining sentences, you will have been fluent by the time you have mined that amount of native sentences mostly because of a huge amount of exposure. If you're less picky, you will be less fluent at the 10,000 mark. The extreme case is when you put sentences from textbooks and stuff. I don't think those sentences count when it comes to the 10,000 sentence method.

If you're as hardcore as the creator of Supermemo, maybe SRSing is vital to your progress. But if you're anything like me, it's like a bonus when it comes to overall fluency in a second language. So without knowing how the OP is using an SRS, no one can tell if it makes a difference. I kind of think this is the same as bilingual dic vs. monolingual dic arguments. It doesn't matter much in the long run. I'm guessing "sentence method" and "always monolingual method" work better for advanced learners in general, but I could be totally wrong.

My take on the 10 000 sentence method is simply that if you add i+1 10 000 times, you learn 10 000 things. Since there's a huge overlap (you learn a new word and get more used to grammar in the same sentence), 10 000 means you learn a lot more than 10 000 things, especially if you stray from i+1 and go i+2 or more sometimes. Therefor, it shouldn't matter what sentences you pick and where you pick them as long as there's always new stuff in them, and I definitely think textbook sentences are a part of it. A huge part of it in fact since it's so hard to find i+1 sentences anywhere else.

What gives you fluency isn't the 10 000 sentences, that's where most people have gotten everything wrong, this is never what Khazu claims either. The immersion and exposure to native media is what gives you the fluency, the SRS just helps you not forgetting stuff and gives a solid base to work from.

Tobberoth Wrote:...[textbook sentences are] a huge part of it in fact since it's so hard to find i+1 sentences anywhere else.

That's because most of your "i" consists of textbook things. In an extremal case where a learner hasn't learned any Japanese from textbooks and only picked up the language through informal conversation with native speakers, he would have a harder time finding the i+1 thing you speak of in textbooks because native speakers don't speak like that in real life (I mean, native speakers don't speak like a hypothetical person in a textbook in the sense that a real native speaker always takes context into account and don't choose phrases based on grammar and such. Hence, the hypothetical student in this extreme case would have a harder time finding "i+1" in context poor materials because his "i+1" doesn't mean "i+1" in a grammatical sense. This is especially so for the hypothetical person who only learned through informal conversation because, as far as I know, context rich informal dialogues are not what the average textbook would cover well.).

If the OP is doing some kind of immersion based learning, most likely his current "i" falls somewhere between all-textbook and all-native-material. Just because you're closer to the former doesn't mean everyone else is.

Also, the kind of "i+1" he needs may not be the same as what you need. Business Japanese textbooks might be good resources of "i+1"s for some people, but it might be pretty much useless for other people. Some people may have a huge pile of grammar rules in his "i," but some don't even when their overall fluency is almost the same.

Like you said, learning outside SRSing is the most important part. So less picky people will be less fluent when they reach the 10,000 mark because that means they have been exposed to a lesser amount of native material. So 10,000 sentences straight from a textbook would mean virtually nothing to the number 10,000 in the AJATT blog because, as you said, exposure to non-textbook Japanese is the crucial part of learning. Those sentences shouldn't be included when you use the number of sentences as a rough measure of the amount of native input. Um, isn't it obvious? I don't quite understand why you don't get this...

magamo Wrote:

Tobberoth Wrote:...[textbook sentences are] a huge part of it in fact since it's so hard to find i+1 sentences anywhere else.

That's because most of your "i" consists of textbook things. In an extremal case where a learner hasn't learned any Japanese from textbooks and only picked up the language through informal conversation with native speakers, he would have a harder time finding the i+1 thing you speak of in textbooks because native speakers don't speak like that in real life.

If the OP is doing some kind of immersion based learning, most likely his current "i" falls somewhere between all-textbook and all-native-material. Just because you're closer to the former doesn't mean everyone else is.

Also, the kind of "i+1" he needs may not be the same as what you need. Business Japanese textbooks might be good resources of "i+1"s for some people, but it might be pretty much useless for other people. Some people may have a huge pile of grammar rules in his "i," but some don't even when their overall fluency is almost the same.

Like you said, learning outside SRSing is the most important part. So less picky people will be less fluent when they reach the 10,000 mark because that means they have been exposed to less amount of native material. So 10,000 sentences straight from a textbook would mean virtually nothing to the number 10,000 in the AJATT blog because, as you said, exposure to non-textbook Japanese is the crucial part of learning. Those sentences shouldn't be included when you use the number of sentences as a rough measure of the amount of native input. Um, isn't it obvious? I don't quite understand why you don't get this...

That's a definite myth. A good textbook uses real world examples which can be overheard in any stratum of Japanese society and that is a fact. You probably have experience with bad textbooks which have colored your view of them. I've read good textbooks in Swedish myself so I know I'm correct, a dialogue example in a good textbook should sound more or less like the exact same dialogue on the streets of said country. If the textbook fails to do this, it's a fault of the textbook and not of textbook based learning.

And the i in i+1 isn't subjective. If you know no Japanese, it doesn't matter if you learn from a textbook, a japanese novel or a conversation with a Japanese person. Or would you claim that 食べる, one of the first words you learn in any textbook isn't used in real life? 映画を見に行かない isn't proper Japanese, it's just textbook Japanese?

Well, my year in Japan certainly disagrees with that notion.

The problem you're having is that you're making a distinction between native sentences and textbook sentences where there isn't any.

Nah, the problem is not whether it's taken from real examples. It's that a person without deep understanding of the language can't understand the context behind an example. Throw a random native sentence with a long explanation. Chances are your student wouldn't get the real meaning behind the sentence.

100 pages of explanations and real examples from corpus about 食べる doesn't teach you usage, collocation, etc. as well as the same amount of exposure would. Do you really think good textbooks present "映画を見に行かない?," "映画見に行かない？," "映画、見に行かない？," "映画に行かない？," "映画見に行こうよ," "映画を見に行きません？," and various others in a way they would appear in real life? I doubt that. Each sentence is natural, but the way it is presented may not.

It's like a per-made sentence deck wouldn't work as a native input source regardless of whether it's based on real examples. Since the number 10,000 is just a rough measure of the amount of native input, any kind of pre-made deck would give a false impression of progress if you include them in the 10,000.

It does depend on the textbook, but I also disagree with this extreme separation of "textbook" and "real" Japanese -- here are some random sentences I'm pulling from my textbook just flipping through:
先生は、今晩電話を入れてくださいとおっしゃいましたよ。
定食って、なんですか。
この窓、開けていい？
次の授業、二時半までですねえ。
値段があんなに高くては困るでしょう？
去年は、まだよく会いました。
I don't see what's wrong with any of these sentences; there are some places where you might say that certain native speakers might not use the structures (i.e. a guy might be less likely than a girl to use the でしょう？) but I can't see any of those sentences as being just flat-out things that native speakers would never say.

Quote:100 pages of explanations and real examples from corpus about 食べる doesn't teach you usage, collocation, etc. as well as the same amount of exposure would.

But if the student can't even understand the sentences from their textbook, how are they supposed to learn anything from "exposure"? Exposure to what?

yudantaiteki Wrote:It does depend on the textbook, but I also disagree with this extreme separation of "textbook" and "real" Japanese -- here are some random sentences I'm pulling from my textbook just flipping through:
先生は、今晩電話を入れてくださいとおっしゃいましたよ。
定食って、なんですか。
この窓、開けていい？
次の授業、二時半までですねえ。
値段があんなに高くては困るでしょう？
去年は、まだよく会いました。
I don't see what's wrong with any of these sentences; there are some places where you might say that certain native speakers might not use the structures (i.e. a guy might be less likely than a girl to use the でしょう？) but I can't see any of those sentences as being just flat-out things that native speakers would never say.

Can you guess what kind of person would say them to whom in which context with what kind of tone of voice? Like Tobberoth and I said, exposure is the key. So they shouldn't count when you're measuring the amount of exposure to native material. In short, I think the 10,000 should be seen as a rough measure of "real context" you exposed to.

Oh, and neither the AJATT guy nor I have ever said you shouldn't use a textbook to learn a language (especially for understanding). I don't recommend you use grammar and such to produce your own sentence though.

yudantaiteki Wrote:Exposure to what?

Context.

magamo Wrote:Nah, the problem is not whether it's taken from real examples. It's that a person without deep understanding of the language can't understand the context behind an example. Throw a random native sentence with a long explanation. The chances are your student wouldn't get the real meaning behind the sentence.

100 pages of explanations and real examples from corpus about 食べる doesn't teach you usage, collocation, etc. as well as the same amount of exposure would. Do you really think good textbooks present "映画を見に行かない?," "映画見に行かない？," "映画、見に行かない？," "映画に行かない？," "映画見に行こうよ," "映画を見に行きません？," and various others in a way they would appear in real life? I doubt that. Each sentence is natural, but the way it is presented may not.

It's like a per-made sentence deck wouldn't work as a native input source regardless of whether it's based on real examples. Since the number 10,000 is just a rough measure of the amount of native input, any kind of pre-made deck would give a false impression of progress if you include them in the 10,000.

I agree with you, what we're disagreeing on isn't where exposure should come from, it's where sentences in your SRS should come from. I'm not saying someone who has exposed themselves to nothing but textbooks would speak correct natural colloquial Japanese, I'm saying that a sentence with 食べる form a novel is no more fitting in an SRS than a sentence from a textbook, which is why it isn't important where sentences come from when claiming a number like 10 000. You can have 30 000 sentences and still speak bad Japanese, but it won't be because all of them are from a textbook, it will be because you haven't had proper exposure.

Like I've said before, SRS isn't a substitute to exposure so there's no point in trying to make it one. Use the SRS to learn what 食べる means (with a textbook sentence or any other), use exposure to get the hang of when it's good to use and how it should be positioned in your output.

Oh yeah, just to clarify, I don't think sentences in an SRS is or can ever be an estimate of your exposure to native material. I have less than 3000 sentences in my deck and I have had LOADS of exposure to real Japanese, probably far beyond what Khazu had at 5000. All it can be is an estimate to how many words you've found important enough to learn actively, add grammar and colloquialism etc to that if that's what you're actively using your SRS for.

magamo Wrote:Can you guess what kind of person would say them to whom in which context with what kind of tone of voice? Like Tobberoth and I said, exposure is the key. So they shouldn't count when you're measuring the amount of exposure to native material. In short, I think the 10,000 should be seen as a rough measure of "real context" you exposed to.

The problem with this idea is that real contexts don't always give learners enough information to know why certain phrases or styles are being used in certain contexts. This is especially true of fictional works, but even in real life situations it can be hard to say exactly why certain things are being used the way they are. I don't think you can say that as soon as you move beyond textbooks things start to "count" because you know how they're being used.

I understand your concern with textbook material, but I think it's a mistake to drawn a sharp line between "textbook" and "real" Japanese, particularly when you're talking about people in the early stages of learning. A textbook example sentence like "定食って、なんですか。" shows you how to use a certain pattern; of course just looking at that example doesn't enable you to speak like a native, but it gives you a basis for learning from real context, and it gives you something that you can actually use -- Xってなんですか is a useful structure to know, and it's the kind of thing you can use in a real context even if you've never been "exposed" to it in native materials before.

tobberoth Wrote:I don't think sentences in an SRS is or can ever be an estimate of your exposure to native material.

I hope not, since I've never used SRS, that would put me at 0.

Tobberoth Wrote:Oh yeah, just to clarify, I don't think sentences in an SRS is or can ever be an estimate of your exposure to native material. I have less than 3000 sentences in my deck and I have had LOADS of exposure to real Japanese, probably far beyond what Khazu had at 5000. All it can be is an estimate to how many words you've found important enough to learn actively, add grammar and colloquialism etc to that if that's what you're actively using your SRS for.

That's what I meant by "less picky = less fluent when you reach a certain number"... Sorry, I guess I worded it poorly.

@yudantaiteki

I don't know what you're talking about, but no one said you shouldn't use textbooks no matter what. Tell me where I said something like that in this thread. Also, here's from the AJATT blog's FAQ:

AJATT Wrote:I also recommend using beginner-oriented books. Even (in the beginning) books with English translations in them — books designed to help Japanese people learn English are especially cool.

Would you stop punching a straw man in the face? Your posts often give an impression that someone is claiming students shouldn't use textbooks at all.

When people say textbook examples aren't real, I think most people mean something similar to "A picture, CV and report about personality, etc. won't substitute for real interaction with the person." They might be useful to get a general idea about what kind of person she is, but you won't be able to fully understand her through reading them. If you talk about what kind of person she is by generalizing the information you get like "She's a very nice person! (because I got the impression from the pic!)" as if you've known her for ages, you're heading for a disaster. It would lead to misunderstanding. The same goes for language. You should have been interacting with it for ages in real life. If you use a grammar rule/vocabulary/whatever you learned from "a picture, CV, report, etc.," you're heading for the same kind of disaster.

When I say, "Don't use grammar when speaking. Don't generalize grammar rules when writing. Don't translate to speak," I mean something like "Don't talk about her as if you've known her for ages when you don't know her well." I know it's kind of impossible to follow this especially for people who like systematic leanring, but I'm of the opinion that students should at least try to avoid relying heavily on grammar when they produce their own sentences. Grammar etc. should be the last resort.

You don't say, "magamo is a bad guy in every imaginable way!," right? You don't know what kind of guy I am. You don't even know if I'm really a he. I don't think you should do the same thing to a language. I'm just saying you shouldn't talk based on superficial information. Generalizing it is even worse.

Tobberoth Wrote:Like I've said before, SRS isn't a substitute to exposure so there's no point in trying to make it one. Use the SRS to learn what 食べる means (with a textbook sentence or any other), use exposure to get the hang of when it's good to use and how it should be positioned in your output.

Why do these have to be two separate processes? If you take your sentences from novels/theses/how-to books/movies/songs that you've read/seen/heard, most of the time you will remember:
-who said the sentence
-what the context was
-what the mood was
-what the significance was

It will also make the sentence much more memorable because of the above associations. Wins all around.
Sentences from a textbook won't give you any of that.

Of course, if you're someone who accumulates content en-masse without actually seeing it in context, then you might as well be ripping it out of a textbook. That's why I cringe every time I see the term "sentence-mining."

magamo Wrote:Would you stop punching a straw man in the face? Your posts often give an impression that someone is claiming students shouldn't use textbooks at all.

Well, when you say things like "he would have a harder time finding the i+1 thing you speak of in textbooks because native speakers don't speak like that in real life" you shouldn't be surprised if people reply with arguments along the lines of "don't draw a sharp distinction between textbook and real Japanese" (yudantaiteki) and "that's a definite myth. A good textbook uses real world examples" (Tobberoth).

I'm not particularly trying to argue with your underlying point, I'm just saying your reply looks like more of a non-sequitur to me (not to mention more confrontational) than yudantaiteki's did...

JimmySeal Wrote:

Tobberoth Wrote:Like I've said before, SRS isn't a substitute to exposure so there's no point in trying to make it one. Use the SRS to learn what 食べる means (with a textbook sentence or any other), use exposure to get the hang of when it's good to use and how it should be positioned in your output.

Why do these have to be two separate processes? If you take your sentences from novels/theses/how-to books/movies/songs that you've read/seen/heard, most of the time you will remember:
-who said the sentence
-what the context was
-what the mood was
-what the significance was

It will also make the sentence much more memorable because of the above associations. Wins all around.
Sentences from a textbook won't give you any of that.

Of course, if you're someone who accumulates content en-masse without actually seeing it in context, then you might as well be ripping it out of a textbook. That's why I cringe every time I see the term "sentence-mining."

I think people with your type of thinking is greatly overestimating the impact SRSed sentences has on your use. I have no sources to base it on but my own experience, but it seems to me that seeing something in one context, even if you memorize it, won't help you at all. You need to see it in twenty or even a hundred different contexts to even have a shot. That one sentence which you got from a specific context with a specific mood or whatever is fine, but it's just one aspect out of a hundred. The minuscule difference it has over a textbook sentence can IMO be ignored. And that isn't even taking into account how much more specific, idiomatic or complex a native use in a novel can be. The textbook is made to be simple, direct and easy to understand, it doesn't contain a lot of context which you might miss out on in a native source you're not fully comprehending. Therefor, IMO, learning a word from something simple and direct is a lot more effective than trying to learn how to use it properly in context form one source when that obviously won't even be close to enough anyway.

pm215 Wrote:

magamo Wrote:Would you stop punching a straw man in the face? Your posts often give an impression that someone is claiming students shouldn't use textbooks at all.

Well, when you say things like "he would have a harder time finding the i+1 thing you speak of in textbooks because native speakers don't speak like that in real life" you shouldn't be surprised if people reply with arguments along the lines of "don't draw a sharp distinction between textbook and real Japanese" (yudantaiteki) and "that's a definite myth. A good textbook uses real world examples" (Tobberoth).

I'm not particularly trying to argue with your underlying point, I'm just saying your reply looks like more of a non-sequitur to me (not to mention more confrontational) than yudantaiteki's did...

Ah, thanks. And sorry, yudantaiteki. I was thinking something like native speakers don't speak a language in real life without taking context into account. I meant "Native speakers don't choose words/phrases because grammar/explanations tell them it's ok" by "Native speakers don't speak like that." So the point was that the hypothetical student in the extreme case I was talking about right before the quoted sentence would have a harder time finding "i+1" in context poor materials because to him an "i+1" sentence doesn't mean an example that teaches a grammar point he doesn't know. Obviously I suck at getting my point across...