Pablo from Dreaming Spanish: Language can’t be explained

[Iversen]

If you have spent 800 hours in a classroom then you will probably have heard your native language far more than the target language,

If you spend 800 hours as a student in a monolingual classroom (as with the ALG 800 hours claim) you will not have heard your native language at all in theory, and if you hear it at all you will hear it far, far less than the target language.

I didn't say that it was a monolingual class - and I doubt that the majority of language classes have become monolingual since I left school aeons ago. That being said, my French teacher in the 'Gymnasium' (lycée, high school or whatever it's called in other countries) tried his best to teach a class of math(ematical)-phys(ics) students French by speaking French and only French (plus occasionally Italian or Spanish to me), but NO Danish and definitely abolutely NO English or German . He had three years to do it, but after 2½ years he one day shocked us all by suddenly speaking Danish, and the message was clear: we had (as a group) learned absolutely nothing during the first 2½ years, and he didn't want to witness a bloody massacre at the exam so from now on it was good old black school pedagogics. And during that last semester much less French was spoken, but most of us passed the exam -and a few might even have learned to speak the language. Since then I have been rather sceptical about the socalled natural method, at least in its pure undiluted form.

On the other hand grammar-translation is also deficient because it treats dead languages as passive, mostly written languages, and that way they stay dead as the dodo. In all other language classes I had in school (including the gymnasium) we spoke mostly Danish, the teacher sometimes said something in the target language, but the pupils only tried to utter something in it when they were forced to do so by the teacher, otherwise it was Danish über alles. That's where I got my ideas about class teaching, and I got it confirmed by asking people who attended evening school what precisely they did during those two hours once a week - they spoke mostly Danish of course just as in school, and then they gnashed cake (and spoke more Danish) in the pause. So basically you have the choice between not understanding anything during a monolingual class and understanding the mother tongue parts in oldfashioned black school teaching, but being shocked out of your wits if asked to say something in the target language. The important thing for me has always been to study at home, and then classes were just a place where you maybe would hear the target language spoken 10% of the time.

[dlt529]

So, a few things. One, we've known since about 1995 or so that input alone is not enough.

Not enough for what exactly? Conversational fluency? Or some higher standard?

This is a really great question with a complicated answer. Generally speaking, second language acquisition research rarely works with proficiency levels, and when they do they use the ACTFL or CEFR ratings, so it depends on what you consider conversational fluency. Usually, we track the development of specific structures, tracking to see how they're used and if/when/how they come into alignment with native-speakers. When I say input isn't enough, I mean it isn't enough to develop target-like accuracy in the language.

A great study that looked at this is Mackey (1999), generally accepted as the first empirical support that input alone is less effective than output. Although she looked at interaction by studying two groups: learners performing a task with a native speaker and getting corrections, a type of focus on form), other learners present but not interacting (no output). She found that the group that interacted (received input + output) made more gains than those who just observed the interaction. https://doi.org/10.1017/S0272263199004027 The Output Hypothesis was based on anecdotal observation of French immersion in the 1970s-1980s (Following Input-only methods), where students' french was still not targetlike after years of schooling in French, but Mackey was the first one to show that "active participation" (ie, output) was related to development.

In your original idea, with 800 hours of input vs 800 hours of instruction, if the instruction is good (meaning focused, lots of opportunities for input and interaction), the student will probably have a better command of the language (across all four skills) that someone who is exposed to 800 hours of input.

[TopDog_IK]

So, a few things. One, we've known since about 1995 or so that input alone is not enough.

Not enough for what exactly? Conversational fluency? Or some higher standard?

This is a really great question with a complicated answer. Generally speaking, second language acquisition research rarely works with proficiency levels, and when they do they use the ACTFL or CEFR ratings, so it depends on what you consider conversational fluency. Usually, we track the development of specific structures, tracking to see how they're used and if/when/how they come into alignment with native-speakers. When I say input isn't enough, I mean it isn't enough to develop target-like accuracy in the language.

A great study that looked at this is Mackey (1999), generally accepted as the first empirical support that input alone is less effective than output. Although she looked at interaction by studying two groups: learners performing a task with a native speaker and getting corrections, a type of focus on form), other learners present but not interacting (no output). She found that the group that interacted (received input + output) made more gains than those who just observed the interaction. https://doi.org/10.1017/S0272263199004027 The Output Hypothesis was based on anecdotal observation of French immersion in the 1970s-1980s (Following Input-only methods), where students' french was still not targetlike after years of schooling in French, but Mackey was the first one to show that "active participation" (ie, output) was related to development.

In your original idea, with 800 hours of input vs 800 hours of instruction, if the instruction is good (meaning focused, lots of opportunities for input and interaction), the student will probably have a better command of the language (across all four skills) that someone who is exposed to 800 hours of input.

Thanks for your answer. I will take a look at this paper, as well as the papers you linked to in your last post.

Regarding the 800 hours, you say "across all four skills", would that include comprehension for normal native TV shows? 800 hours of immersion in a language like French, assuming the learner is a native English speaker, could probably get an immersion learner's comprehension pretty high. I wonder if the person coming from the classroom could really compete in terms of TV comprehension? I know my own experience with German is that reality TV shows, for example, take hundreds and hundreds of hours to reach near full comprehension.

[Cainntear]

So, a few things. One, we've known since about 1995 or so that input alone is not enough.

Not enough for what exactly? Conversational fluency? Or some higher standard?

Learning the language. Simple as that.

You cannot learn a language from input only -- you must engage in output.

[Cainntear]

So, a few things. One, we've known since about 1995 or so that input alone is not enough.

Not enough for what exactly? Conversational fluency? Or some higher standard?

Not even for conversational fluency. Understanding and producing language is not some guaranteed continuum where one follows from the other.

To expand on this and illustrate the point...

If someone says "He wants to eat strawberries", it does not require full perception to be understood. If you perceive it as "He one to eat estrobberriess" you can fully understand the message, as there is sufficient information there. A Spanish speaker is likely to make all these mistakes without an early focus on pronunciation as this is what happens when you fit English to Spanish phonetics rules.

With no indication that there is an error, there is no motivation for correction, so the incorrect perception persists. If your brain doesn't process all the input, it cannot result in correct output.

[Cainntear]

If you have spent 800 hours in a classroom then you will probably have heard your native language far more than the target language,

If you spend 800 hours as a student in a monolingual classroom (as with the ALG 800 hours claim) you will not have heard your native language at all in theory, and if you hear it at all you will hear it far, far less than the target language.

I didn't say that it was a monolingual class - and I doubt that the majority of language classes have become monolingual since I left school aeons ago.

The 800h thing is a specific claim of ALG teachers and not a general claim from across the field of language teaching.
I don't believe in ALG as an optimal (as should be clear from the above and pretty much everything I've ever said on these forums) but if we all keep talking about what we individually want to talk about rather than discussing the topic at hand, what's the point? The conversation is now devolving into something like the following.

1: Method A says they only need 800h in their classrooms. Wouldn't 800h of Method B be exactly the same?
2: No, 800 of Method B isn't the same as Method C.
3: Method B doesn't work.
1: Ah, but isn't Method B better than Method D? (Let's just ignore that Method A is not Method B, and that Method C is radically different from Method D...)
4: Sorry, but the problem with Method E is that it's not very good.
5: We weren't talking about Method E, we were talking about Method A.
6: Why would we talk about Method A? Method A is stupid.

I mean... I do agree with a lot of what you're actually saying in the message (and have said as much explicitly in the past) but if we're all talking about different things we get nowhere.

[dlt529]

In your original idea, with 800 hours of input vs 800 hours of instruction, if the instruction is good (meaning focused, lots of opportunities for input and interaction), the student will probably have a better command of the language (across all four skills) that someone who is exposed to 800 hours of input.

Thanks for your answer. I will take a look at this paper, as well as the papers you linked to in your last post.

Regarding the 800 hours, you say "across all four skills", would that include comprehension for normal native TV shows? 800 hours of immersion in a language like French, assuming the learner is a native English speaker, could probably get an immersion learner's comprehension pretty high. I wonder if the person coming from the classroom could really compete in terms of TV comprehension? I know my own experience with German is that reality TV shows, for example, take hundreds and hundreds of hours to reach near full comprehension.

So, I want to be clear I’m not saying immersion is a bad idea—800 hours of input can do great things for your intuitions about the language. It just won’t be enough to produce it.

For your question, of course the learner who spends 800 hours watching TV and nothing else will be better at understanding TV. Yhe point I was trying to make was that 800 hours of instruction will result in a more well rounded speaker on the four skills. If you’re only interested in listening or comprehension, input only is fine. Honestly, I would place a much higher value on input (and interaction, which includes input) than I would on output alone, but output is crucial to developing grammatical accuracy. The thing is, most people will want to speak eventually, and there is no serious academic evidence that waiting to produce the language helps to avoid errors. There is evidence that output helps at all levels to facilitate the process.

A major downside to input only or comprehension first approaches is how much of comprehension relies on very few words. A clever student only needs to understand the major nouns and roots (lexical words, not grammatical) to understand language. I’m not sure how many success stories I’ve seen on input ONLY with no previous grammar instruction. Many people who rave about dreaming spanish tried for years with “traditional” methods (probably grammar focused) before trying it, and we can’t discount all those years of study. They learned something, it just wasn’t useful for communication.

Anyway, let me know if you can’t access those papers and I can send you copies!

[TopDog_IK]

You cannot learn a language from input only -- you must engage in output.

Of course. Who says you can learn a language form input alone? Sounds like a strawman. Every immersion system I am aware of includes output, usually delayed until the learner has fairly high comprehension.

[Le Baron]

You cannot learn a language from input only -- you must engage in output.

Of course. Who says you can learn a language form input alone? Sounds like a strawman. Every immersion system I am aware of includes output, usually delayed until the learner has fairly high comprehension.

Yes, but what's missing here (and pretty much persistently with these 'I get input, then I gain comprehension, then I speak' views) is that output is a separate skill which isn't guaranteed by and provided for by comprehension of input. Physical formation of sounds and patterns and prosody and learning about how to use word choices and context etc in real time. There are quite a lot of people who grew up in a bilingual household whose passive comprehension can be excellent, 100% sometimes, but they can't speak properly. Chinese families in the west know this all to well. Why would that be if input delivers all the goods?
I grew up in a family like that. My older brother speaks flawless French, because my mother was still learning English properly herself when he was born and routinely fell back on French. I received a mix of French/English. My younger brother only understands to a certain level and can't speak it at all beyond elementary sentences, but we were both exposed to the same basic input.

Additionally there's something else going on when you 'acquire' a language in interactive circumstances. Me as a young lad asking my mother and aunts/uncles 'what does that mean, what does that mean?' is very different than purely blank exposure and input. Some of it is osmosis, quite a lot actually, but some needs explanation, and this is even normal for most people with their native language. Have we somehow forgotten that all our home and school life as a child is dominated by people telling us what so many things, words/expressions actually mean? Discussing and confirming things in use with peer groups over years and years? We may think we just worked them all out ourselves, but we didn't. This is not dramatically different in foreign language learning. It's just that impatience means we try to pack it all into a shorter time span, so it's a much slighter framework with gaps, only filled-in and strengthened over years of exposure and use.

I completely agree with input and exposure, but the idea that no explicit discovery or asking questions or checking uncertainties should happen is to me just so much nonsense.

[TopDog_IK]

Yes, but what's missing here (and pretty much persistently with these 'I get input, then I gain comprehension, then I speak' views) is that output is a separate skill which isn't guaranteed by and provided for by comprehension of input. Physical formation of sounds and patterns and prosody and learning about how to use word choices and context etc in real time.

Who is claiming otherwise? Delaying output doesn't mean that people don't have to practice speaking. This seems like a total strawman argument.