The Perversion of Krashen?

I have developed an interest in reading about foreign language acquisition. I’ve read reddit, language learners, lingq forums, various linguist blogs and anything I can find.

I find it amusing that there are so many language acquisition methods which have now rolled in extensive interesting comprehensible input (mostly reading) with references to Krashen and his hypotheses. In many cases the method seems to be the kind of thing Krashen was saying did not work well, but in light of his research they don’t change what they do but rather just roll in some reading in addition to what was done before.

In fact I am reading a document right now in which the whole ‘extensive reading’ is covered. There are targets for how much needs to be read per week, known-to-unknown word ratios are detailed (1 unknown per 50 words), ratios for reading time with/without looking up words…

It seems to me that Krashen has been pretty clear in many talks that what he is saying is: the learner consumes self-selected comprehensible input at a rate which is comfortable to them without set outcomes or expectations.

Correct me if I am wrong here, but this is what he saying, right? Not only that but he is saying this can replace many other activities to become the primary form of language acquisition?

In any case my wording here is biased because I already believe that comprehensible input* in the form of reading/listening is sufficient for language acquisition. I wasn’t really believing it months ago, but my own progress has shown me otherwise (thus far - its still very early days yet for me).

  • Comprehensible Input: Hard to define. What is N+1 anyway? Also I’ve been working with input which was certainly not “N+1” a few months ago but now is fairly comprehensible.

“1 unknown per 50 words”

“Comprehensible Input: Hard to define. What is N+1 anyway?”

Krashen’s 98% figure for comprehensible input is way off the mark in these times of online readers with auto dictionary lookup.

That is to say, comments like – “a text needs to be about 98% comprehensible in order for it to help the reader acquire new vocabulary” isn’t really applicable in 2016.

You can, and should, spend most of your vocab development period reading at around 80-95%, particularly with systems like lingq.

For example, there’s a difference between when you are reading at less than 80% known, 80-95% known, and 95%+. You can get to around 80% known pretty easily just by looking up all of the unknown words you first encounter, and this isn’t a bad idea, imo. This will get you about 5,000 words, most of which will be the high frequency words.

But then a lot of language learning is really all about tackling the 5,000-40,000 known words step. And for most of this journey you will be somewhere between 80-95% unknown.

For example, if you were to just import Oprah Winfrey chat show transcripts to lingq you will find the first 50 episodes can quickly take you to 80-85% known. Another 100-150 episodes will take you to about 85-90% (consistently). Then another 250+ episodes are required to take you to 95%+ (consistently). Such that each gain requires a (relatively) exponential increase in volume of input.

You will also find that the first step (the first 50 episodes) is no fun, and it is just a lot easier to use simpler, shorter material (or at least “mix” these in). Further, the second step (episodes 50-200) is not always a lot of fun either (even, if you really like Oprah). Even at 15% unknown, it is a real slog to read, no matter how you go about it. But, I think, it is helpful to try to read through a sentence or a paragraph and not (straight way) look up words here.

Also, this is where the real magic of transcribed audio comes in - being able to listen and read - it helps a lot to get through the slog, I think, and the brain seems to latch onto and focus on different things, plugging gaps, and reinforcing words – when you use listening and reading together.

The other issue, is that to read at 1 in 50 unknown, limits you from accessing native content early on, and limits all your content generally.

1 Like

I am a language teacher, but I HATE ALL METHODS of the language acquisition!
It’s interesting to read about different methods, but EVERYONE MUST WORK OUT HIS/HER OWN METHOD in order to learn a foreign language more quickly and more useful for themselves.
The people are different - someone learns with their eyes, someone with their ears, someone with their hands by writing etc, that’s why the methods must be different!..
But in any way the more time you devote to your new language, the more success you can achieve!
What are the most typical mistakes of the language students in my opinion as a teacher:

  1. Many stuidents are lazy in learning words, But I strongly belive that first 600-1000 words must be learnt by heart, otherwise it’s difficult to go ahead.
  2. A lot of learners ignore the Grammar, and without the elementary Grammar you can’t understand how to make up phrases from the words in a new language. So the basic Grammar must be learrt in the practical use.
  3. The most students are too impatient, they take right away too difficult texts and podcasts, maybe with 40-50% of unknown words - and then they are cruelly ‘struggling’ with such texts, they lose in this fight their motivation and give up learning a new language not understanding that they were guilty themselves because they overloaded themselves with the incomprehensible stuff!..
    In this point I agree with Krashen that ‘the comprehensible input’ (or in my term: ‘step-by-step’ method) is the key, cruicial moment in the language study!

You wrote,"I already believe that comprehensible input in the form of reading/listening is sufficient for language acquisition. "

I think that comprehensible input is not the sufficient condition but the necessary condition for language acquisition.
Although free voluntary reading, reading for pleasure, etc. are very important, you should write as much as possible in order to learn to write well. Some people think that if you want to write well, you should read only those books that are well written.

Thank you all for the insightful comments. Instead of replying to each I’m just going to write one here.

@iaing: What you say rings true with my initial experiences here using LingQ. I’ve lamented the lack of quantity in beginner material for Korean many a time now. So much like your Oprah example, I’ve been working through intermediate dialog transcriptions and I feel my vocab is increasing at a very good pace.

When I questioned “Comprehensible Input” though I was meaning that there is so much more than the vocab. For Korean (as example) I’ve now accumulated more than 70 things which can attach/transform/modify verbs to impart additional information. From indicating the verb is being recalled, quoted, intended to be done through to adding a nuance of uncertainty.

Not to mention all the expressions and phrases for which the whole is either greater or completely different to the sum of the parts so to speak.

At any rate I am not disagreeing at all with you - what you say makes a lot of sense.

@evgueny40 and @Yutaka:

Thank you for your comments too! I appreciate your insights.

There are two things that I’m unsure I entirely understand about Krashens point of view:

  1. He asserts that everyone learns language in the same way - that the brain acquires language in the same way for everyone. I think what he is saying is that the exact same kind of comprehensible input would work for everyone.

  2. He claims that practicing output through initial learning is not necessary. The way I think he means it is that if output were a mountain to be climbed, that the mountain gets smaller and smaller with volume of comprehensible input. Steve makes this claim too in a video about fluency in 5 days (if I recall! its been a long while since I watched it). The claim here is that the mountain can eventually be reduced to a small hill - that with sufficient input, free-flowing comprehensible output comes with astonishingly little effort.

These two claims seem to run counter to what the multitude (perhaps the majority?) of linguists and language teachers are saying.

Do I misunderstand Krashen (and Steve since he has also made these two claims in his videos) on how I interpret these claims? I’d like very much for these claims to be true but they seem quite audacious.

So I mentioned Steve, alongside Krashen there. I decided this morning I should look up the video(s) in which I recall him speaking on these topics to be sure I am not putting words in his mouth.

And to some extent I am. Here is the main video I recall:

He has done others on the topic, two of which are these:

So now that I have rewatched them, it would appear that what he is saying is:

(5 day to fluency ‘series’)

  • practice output alongside reading/writing until you feel you can make a go of 5 day immersion
  • then do the 5 day immersion and the promise is you’ll at least make a significant breakthrough in your target language

(Other video)

  • speaking is difficult. it will require a lot of practice and there is no way around it other than practice.

To a certain extent it feels slightly contradictory but I guess the premise for the ‘5 day fluency’ was that there would also be sufficient output practice. So it is preceded by many hours of talking to a tutor. Additionally the promise is somewhere between ‘a major breakthrough’ and ‘participating in everyday conversation comfortably’ (this I would feel is a fine goal).

So really for me it’s what I originally felt would be the case: that Korean is going to be a very tough slog, and that mountain of being able to speak is not going away so the sooner I start climbing it, the better.