the input hypothesis and the fallacy of antimoon and AJATT

This is the first post in a three part series: The Input Hypothesis 2 and The Input Hypothesis 3 are also available for your reading pleasure.  🙂

For the first time since sixth form I’ve actually been browsing through some papers on linguistics.  Taking a trip down memory lane to A-level English language, a lot of the material we covered was based on the Book of Chomsky.  Well, “the Books”, actually – he’s output a phenomenal amount of material on language acquisition over the course of his lengthy and prestigious career, and has for decades been seen as the authority on the subject.  A lot of his research can be summed up thus: adults fundamentally acquire language in a different way to children – this difference supposedly means that adults find it a lot harder to learn a second language.  In fact, either Chomsky or one of his offshoots (hereafter referred to as “Chomskites”[1]) posited that it was impossible to adults to learn a language to complete fluency, and that even trying just got harder with age.

I was (more recently) introduced to one Stephen Krashen.  I was surprised to find out that he doesn’t contest children and adults acquiring language in different ways (I do) – but virtually everything else is a different kettle of fish[2].  The biggest and most important thing he’s said is though is that true, native-level bilingualism is possible even for second languages learners, and there are a variety of people that support that statement.  Get the plates Chomsky, cuz you just got served!

Krashen’s theories are split up into five hypotheses, the most well-known of which is the Input Hypothesis.  This has formed the basis of Antimoon and AJATT.  The hypothesis broadly states that humans can only acquire language in one way – from exposure to comprehensible input and understanding.  This tallies with how children learn language – one step at a time.  Material for children is carefully graded so that whatever they do, it’s not too much of a stretch.

Clearly however, the Input Hypothesis is completely at odds with traditional language learning, wherein one of the first things students are told to do is speak – to say new words and phrases without hearing them first.  This is what almost all textbooks tend to do too – they give exercises often for a student to construct their own sentences.  This is highly error prone – compounding the problem is that learners who learn in this way tend to internalize their mistakes.  This is known as fossilization, and what we end up with is a situation in which virtually every learner of a foreign language is easily distinguishable from native speakers by virtue of their inaccuracies.  This is why both Antimoon and AJATT don’t recommend much speaking when learning a language.

I think this is something of a fallacy though – this conclusion is drawn from speakers who learn using non-input based methods.  Children never undergo fossilization.  They make all kinds of mistakes, all the time, and do so also when talking to each other – so why do we not live in a world when everyone’s all like “me want nana!”?  I’m pretty damn well sure that it’s because they have constant exposure to their language – even if they don’t conform right away, they have no choice but to, given the constant corrections thrown their way.  What happens if they don’t receive constant corrections?  They keep making the mistake[3].  Clearly then, fossilization isn’t limited to just adult learners, but can be countered with sufficient feedback.

Given that children don’t internalize their screwups, and since both AJATT and Antimoon are based on massive native input which mimics the way children learn language, – I think it’s pretty safe to start speaking whenever a learner feels like it.  If making mistakes doesn’t harm children in the long run (given sufficient feedback), then it shouldn’t hurt learners.

I say this, because I think it really is important for learners to be speaking sooner.  There is a difference between active and passive vocabulary.  If you’re familiar with Heisig, you’ll know that he advocates writing Chinese characters for practice, even if the eventual goal is only to read them.  This is because active recall (writing) tends to produce stronger memories that passive recall (recognition).  This is documented in more papers than there are stars in the sky.  Also, if you think about when you were at school – whenever was an assignment just to read?  Never, that’s when – it was to read something and write an essay on it.  Or, in the case of maths, to do a load of problems on the subject matter.  Yes, you make mistakes, but that is part of the learning process, and so long as someone is there to correct you everything runs smoothly.

Khatzumoto has said previously that his speaking was considerably weaker than his listening, in both Japanese and Cantonese, and I’d say that that’s a direct consequence of placing a disproportionate amount of emphasis on input rather than output.  Incidentally, Krashen has stated in papers that he doesn’t think language production is harmful at all to learning – quite the opposite – he just cautions against doing so too early.  I agree that we only acquire language through constructive input; the niggly point is that we improve our language ability through output.

[1] Actually, that’s a lie.  They won’t be referred to again.  I just thought “Chomskites” sounded funny.

[2] Why a fish?  Why a kettle?  Why is there more than one?! One of those eternal mysteries of the English language, I guess.

[3] This actually happened to a friend’s child – they thought it was cute that she was forever referring to herself as “me” (instead of “I”), even when she way 8? 10? years old.  Cute perhaps, but not funny for her in the long run.


13 thoughts on “the input hypothesis and the fallacy of antimoon and AJATT

  1. Pingback: critical thinking « thousand mile journey

  2. Pingback: All Cantonese All The Time « thousand mile journey

  3. Pingback: All Cantonese All The Time « thousand mile journey

  4. >I think this is something of a fallacy though – this conclusion is drawn from speakers who learn using non-input based methods. Children never undergo fossilization.

    Children are given massive input for up to 2 to 3 years before they speak.

    >Khatzumoto has said previously that his speaking was considerably weaker than his speaking

    My writing is weaker than my writing…

    • >>I think this is something of a fallacy though – this conclusion is drawn from speakers who learn using non-input based methods. Children never undergo fossilization.

      >Children are given massive input for up to 2 to 3 years before they speak.

      And therein lies the fallacy – give adults that same massive input and they too will be able to achieve fossilisation-free native-level fluency. If you don’t give them that input then don’t try to compare apples and pears.

      >>Khatzumoto has said previously that his speaking was considerably weaker than his speaking

      >My writing is weaker than my writing…

      WHY did no-one point this out beforehand?!

  5. If you think about it, all of our speaking is considerably lower than our listening. You may know what what 10,000 words mean, but your everyday speaking vocabulary is only half or less. Though, this isn’t a particularly difficult word, when was the last time you said “splendid”. Or maybe a word like ludicrous. You certainly know what those words mean, don’t you? Yet, you don’t use them much at all. That’s what he meant by that. If you listen to him speak on videos on youtube, you’ll see his Japanese is perfectly fine. I advocate the silent period. Give yourself a month or two, and you’ll find that after, there are words and phrases that just flow out of your mouth, much like they would in your native tongue. Speaking too early is just unnatural. You think of grammatical rules, unsure if you’re right or wrong instead of developing the sense of what sounds right and what doesn’t. With the silent period and input hypothesis I can hear bad Spanish as well as I can hear bad English. My Spanish just flows out as well, like it’s supposed to.

    • Heyy… thanks for the thoughtful comment 🙂

      I think everyday vocab is ~2000 words, although it depends who you are and what you tend to talk about.

      I personally didn’t get much out of the silent period, for reasons i would still like to figure out. Actually my Chinese progressed a lot since practicing grammar etc. more… there is much I have to write on the subject if I find the energy to 🙂

      How long have you been at the Spanish for?

  6. Hello Eldon Reeves,

    i am always happy when I find someone who is interested in the same things as I am .
    It ‘s a L2 learning ,since I am not playing any instrument (consequence of electronic music listening 🙂 )
    Hence, I have decided to write a note here on your site.
    Here is my personal experience with “input” in school and when I did it by myself.
    I have been “learning” English language for the last 15 years .I was learning it in my primary and secondary school and at the college as well (although I studied pharmacy) . Someone would think that it is a lot of time learning English language . To be honest , it is , but it turned out ineffective.
    In primary school , emphasis was on grammar (fill in the gaps) and reading from the textbooks. Rarely did we listen to something interesting . It was followed by 4 years of secondary school where we learned the same things again. It was again the same at the college.
    If it wasn’t for video games, MTV , my interest in different cultures and internet I would be still at the intermediate level. there is no substitute for the interest in the language and consistency .
    If you learn 1h/day , 365 days that is “only” 365 hours per year .
    Let’s compare. In my primary school : 45 minutes (3/4h) per week or 1,5 h per week
    for 5 years. Minus holidays and the time when I wasn’t in the school. That is less than 70h per school year . It is less than 350h for primary school ( 5 years) . I suppose that I haven’t received more than 1500 h of formal English learning during the last 10 years.
    That is really bad if you want to learn a language.
    I had a similar situation with my German language. Studied it for 5 years and almost forgot everything.
    Everything changed when I learned that I am moving to Germany. I learned in two months more than in those 5 years and now I am here in Germany and I think that
    I will be able to speak it without big problems until the end of 2011.
    I am at least , listening to it every day .
    All the best.
    See you .

    • Hey Mr. Dax

      Sorry for missing your comment! Not sure what I can add – your English seems good enough, so just crack on with the fun stuff in Germany in a similar way and I guess it should all be good. Good luck with the 2011 deadline!


  7. I know this article is hella old, but damn it, I have opinions.
    I remember learning about language acquisition during A levels too. I disagree with the ‘no output’ thing too – like you said, children make constant grammatical errors when first speaking, which last for years. Studies have shown that constantly correcting their errors has no effect – children will reach the same level of basic grammar competency whether they had their inaccuracies pointed out or not. Same with learning as an adult – you’l make mistakes, but they’ll naturally solve themselves with more input and practise.

    • Yers, it’s very interesting… since writing the above article, I’ve noticed my Cantonese includes a variety of features I don’t think I’ve ever had explicitly drawn to my attention. At some point, it just… happened.

      Anything special you noticed with the Russian?

  8. Pingback: The Input Hypothesis and Various Fallacies #3 | Thousand Mile Journey

  9. Look up The Loom of Language. Great book-and old…

    Kids generally don’t mind asking any question-adults often do. That has a huge effect on learning.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s