Learning Languages in the Classroom and "in the Wild": Second language learning and embodied cognition

[Kraut]

https://www.psychologytoday.com/intl/bl ... n-the-wild


Learning Languages in the Classroom and "in the Wild"
Second language learning and embodied cognition

/.../

The reliance on explicit memory is also supported by patient foreign language teachers who are willing to wait and smile encouragingly, while we hunt for the right word.

Yet even the most superior conscious recall is too slow for everyday interaction – in ‘the real world’, transactions and interactions rely on automatic processes and few people are willing to wait while we fumble to retrieve our new words and order them just so. This pressure, however, gives learning “in the wild” an edge – to fit in, and keep up, naturalistic learners have no choice but to engage the same automatic processes and the same implicit memory that subserve native language use. Such engagement does not guarantee either accuracy or native-likeness, but it does ensure that both learning and retrieval of information rely on the same memory system.


/../

[Cainntear]

https://www.psychologytoday.com/intl/blog/life-bilingual/201501/learning-languages-in-the-classroom-and-in-the-wild


Learning Languages in the Classroom and "in the Wild"
Second language learning and embodied cognition

/.../

The reliance on explicit memory is also supported by patient foreign language teachers who are willing to wait and smile encouragingly, while we hunt for the right word.

Yet even the most superior conscious recall is too slow for everyday interaction – in ‘the real world’, transactions and interactions rely on automatic processes and few people are willing to wait while we fumble to retrieve our new words and order them just so. This pressure, however, gives learning “in the wild” an edge – to fit in, and keep up, naturalistic learners have no choice but to engage the same automatic processes and the same implicit memory that subserve native language use. Such engagement does not guarantee either accuracy or native-likeness, but it does ensure that both learning and retrieval of information rely on the same memory system.


/../

A few things I don't like...

It makes a strong dichotomy between declarative controlled practice and procedural loosely controlled practice, disregarding the whole spectrum in between -- the concept of slowly removing scaffolding to make practice increasingly procedural and decreasingly controlled shouldn't be a controversial topic; whether in language learning or any other field of learning.

Relatedly, it takes as given that the best way to learn something is to use it as though you've already learned it. I'm not going to rule out the possibility that this is true, but I can't see any justification to uncritically accept it as a truism.

It dismisses errors as unimportant, gives fluency primacy over accuracy, which assumes there is a binary choice. It seems to me that an early focus on fluency tends to become an impediment to accuracy in the long-term (through fossilised errors) whereas an early focus on accuracy may hurt fluency in the short term, but you can always continue to speed up as you continue to learn.

And there's where my biggest issue lies -- anyone promoting fluency over accuracy is expressing the belief that you have to make this choice, accepting that accuracy will probably remain outside the learner's reach forever. That means that they are making a choice on behalf of their students, denying them the possibility of ever being professional translators. But of course interpreters need to have both accuracy and fluency to do their jobs, and interpreters exist, so clearly it's not either/or. Anyone who says it is either/or is in effect stating that their chosen pedagogy is just not good enough.

[Kraut]

I think there is a wide range of how much grammar has to go into the teaching/learning of a foreign language - depending on many factors. I hardly need any grammar when I study Spanish- being a teacher of French. It was quite the opposite when I did Lithuanian, a language that abounds in declensions, has no articles, a loose word order, grammatical dual, that can only create meaningful utterances via word endings - even numerals need them. The vocabulary is completely opaque with the exception of modern borrowings. Comprehensible imput does not happen, if you haven't learnt grammar before, it is confusing input.
---------
The two major figures in the language learning business on the subject:

https://www.youtube.com/watch?v=9kjqiWOJh0Q
Steve Kaufmann
Language Learning Content or Grammar Focus? With @Olly Richards

[Cainntear]

Relatedly, it takes as given that the best way to learn something is to use it as though you've already learned it. I'm not going to rule out the possibility that this is true, but I can't see any justification to uncritically accept it as a truism.

Just to revisit this statement...

The article talks about "embodied cognition", so there is an attempt at justification, but again it seems a bit of a binary thing, as it seems to assume that the only way you can picture something is effectively in the middle of a role-play.

One of the biggest revelations for me when I started doing my first Michel Thomas course (the Spanish one) was how easy it was for me to fully experience the meaning of the sentences we were asked to translate.

People talk about "meaningful learning", and that was what I got from MT. Sentences like "I want it, but I don't have it" may not have a clear, explicit meaning, unlike "I have 2 sisters and an aunt", but I find it massively more meaningful.

The advantage of situational role-plays is that you can't slide into mechanical churning out of things like "I have a cat... I have a dog... they have hamsters..." but it's not the only way of doing it. Thomas's skill was in making consecutive prompts different enough from each other that you had to engage in the language and think about everything, not just think about the one bit that changes.

[Iversen]

My stance is that first you have to get as much knowledge about the foreign language as possible - and that means vocabulary, vocabulary, vocabulary ... and grammar. And grammar should be known in its total compass so that you know what there is in store for you, but you don't have to learn all the details from the beginnning. And it would be naive to expect school to give you all that knowledge by spending one or two hours a week in a hellish noise, surrounded by nose-picking bored fellow pupils who are more interested in their phones than in anything related to the teaching going on. At best school can help you to find out how and where to find the necessary information outside the class environment.

So do home study as a supplement, and try to make it efficient and as concise as possible. I do green sheets for the morphology, and anything that can't be squeezed into those sheets must wait. Vocabulary plus essential grammar plus some information about the things you haven't learnt yet should be enough to get the language started as a passive language. Add some training and - if possible - a concrete purpose, then it can become active on a low level. Which esssentially means that you can become fluent within a narrow range, but trying to be acribically correct at this stage would probably block you from also becoming fluent anytime soon.

The end goal should of course be to achieve a high degree of accuracy, but when are you in the best position to obtain this? When you are a struggling newbe or when you have reached the level where you actually can use the language and relax and have fun with it? The talk about fossilized errors has done a lot of harm because it has lured people into aiming for total correctness too early in the process. A fossilized error is simply an error which you didn't bother to correct when you had reached the level where you could have done it relatively easily, because you thought that it was enough to be able to use the language for practical purposes rather than keep on improving it. It is not something that by definition is hewn in stone.

As for situational role-plays: I have never hated anything as much as those attempts to force poor innocent pupils who can't even speak their target language yet to play amateur theatre in it. - as if amateur theatre wasn't bad enough in itself!

EngEngelsk.JPG

[mrwarper]

My stance is that first you have to get as much knowledge about the foreign language as possible - and that means vocabulary, vocabulary, vocabulary ... and grammar. And grammar should be known in its total compass so that you know what there is in store for you, but you don't have to learn all the details from the beginning.

And this is exactly why I don't feel much like discussing how to learn any language, etc. any more -- how hard can it be to understand that this is the core?

[...]The talk about fossilized errors has done a lot of harm [...] It is not something that by definition is hewn in stone.

This, too.

As for situational role-plays: I have never hated anything as much [...] - as if amateur theatre wasn't bad enough in itself!

I hear you, how I hear you

However, I have found a variant of this to be very useful in the classroom, I suppose because I excise all of the "amateur theatre" elements -- a conscious decision. Once I get everyone in a group to be completely at ease during classes (that is the key part, but I always work with small groups), I just incorporate micro-segments of this while engaged in otherwise normal conversation with students.

It is as easy as picking up casually one or two students and ask them to imagine for a second such -and -such situation (preferably something directly related to the conversation) and what they think people (people other than them if that helps) might typically say then -- just a sentence or two, on the spot. If they can't come up with anything, "that's OK, we'll try next time" and we move on to someone else, or some other topic -- but most of the time it is enough to get everyone involved and actively discuss how whatever the others said can be improved upon, or decide to demonstrate how they would "say it themselves", etc. In short, make everyone focus on the language, and then do a few useful things with it, instead of forcing some poor fellows into a 100% uncomfortable, absurd situation with everyone else watching -- that should qualify as cruel and unusual punishment.

[Steve]

It makes a strong dichotomy between declarative controlled practice and procedural loosely controlled practice, disregarding the whole spectrum in between -- the concept of slowly removing scaffolding to make practice increasingly procedural and decreasingly controlled shouldn't be a controversial topic; whether in language learning or any other field of learning.

My sense of things is that there is also something I've seen referred to as "automated declarative" that can mimic procedural to a limited degree. I suspect many people who improve but then hit long-term plateaus in spite of effort to improve weren't dropping scaffolding but were rather just speeding up use of declarative as much as they can. My experience in both music and language learning is that the scaffolding never dropped away until I intentionally tried to stop using declarative. I had to explicitly decide to use methods that focused almost exclusively on developing procedural into order to start using procedural to any degree. The change was then rather dramatic rather than gradual. I regret the decades of wasted practice and potential improvement I lost because of this.

I spent about 40 years as a musician with most of my playing in a more improvisational style where I saw songs as a chord progression that I would layer individual notes onto. Mentally, this was quite intense as I was constantly processing chords, scales, notes, and what would go with what. The actual physical playing of bass, keyboards, and guitar was in procedural memory as I could play notes or chords more or less by deciding what to play. However, the decision of what notes to play was almost completely declarative by explicitly thinking about music theory and choices related to those rules of what does and does not fit together. Reflecting on this, I had automated my explicit use of declarative to choose what to play. I played exclusively in major and minor scales. My attempts to add other things such as blues or jazz scales didn't get very far because of the prohibitive amount of time needed to assimilate dozens of scales in all keys and learn many chord substitution rules. It also meant I ended up on a plateau for a few decades because the only way to get off was to spend a lot of time at it.

After learning a lot more about language learning, especially the general principle that you only get good at what you actually practice doing, I tried this on piano. The main skill I really wanted was to improvise a song in various styles including jazz or blues types of notes and various rhythmic patterns. Instead of taking the standard recommended approach of memorizing scales and chord substitutions, I decided to just start practicing improvising on songs. I literally closed my eyes and started playing notes having to rely on my ears to determine if I liked what I was playing. It was the only way I could think of to stop myself from thinking of everything in terms of music theory. Within a few weeks, some unexpected things happened. The first is that I started to "feel" and "hear" songs. I was starting to rely on a sense of what sounded and felt "right" to me rather than intellectually thinking about if it fit the correct rules. The second is that I found it took little mental effort. Previously, playing anything required my total attention (to be mentally processing the rules). Now, I can hold a simple conversation while I'm playing. My piano improvisation had been limited by how fast I could mentally processes musical rules. Now it is more limited by the physical speed of my fingers. I can just "think" what I want something to "feel" and "sound" like and it just sort of appears. I've found that my natural experimentation has been naturally leading to many scale and chord substitutions with fancy names. I've also found that I can look up a scale, practice it a few times to get a general sense of what it sounds and feels like, and then start using it to some degree. Basically, I have started building a procedural music creating structure in my brain that is akin to the procedural structure I have for speaking English. I also was not spending a lot of time doing this, maybe 10 to 15 minutes of playing per day with some huge gaps of weeks where I didn't play much.

This transition was something that only took a week or so to do. This was not a gradual transition from declarative to procedural. It was not a gradual dropping away of a scaffold. It was an intentional change where I forced myself to stop using declarative and it occurred very quickly. Now, my use of declarative is a minimal amount to set the general direction of what I'm playing. There is huge difference between the automated declarative I used to use and the procedural that I now use. When I try to go back to play lead guitar (which is still pretty much automated declarative for me), it is so painful I don't like it. Instead of free and easy and relaxing, I feel like I'm grabbing a couple ton load and trying to drag it. This is similar to the transition I went through when I started spending more of language learning time listening, following, and reading. It wasn't a slow gradual dropping of scaffolding. It was a rapid acquisition and improvement of more procedural skills when I didn't give myself the opportunity to use declarative memory.

Here's a YouTube link of me learning to improvise a new song I'd only just heard for the first time a few minutes before. In other words, no practice and starting from scratch. Before this recording, I had figured out the chords and melody in perhaps 5 to 10 minutes of listening (no notes or chords provided). I then immediately started recording and just started trying to play the song. First couple minutes are just getting the feel of the song. At about minute 2, I clumsily start to move to other things. From about 2:45 or so on, I'm starting to feel the song and able to some things I really like. Spending a few hours on this song over a few weeks compared to 5 minutes, many more complex and intricate rhythms, chords, and arpeggios would start to arise and I could start to polish those. https://www.youtube.com/watch?v=8COjb-35KIk

Bottom line, back when I was using automated declarative memory to drive my piano playing, it was mentally intense and the result of practicing at a level of an hour or so per day. Now primarily using procedural memory and practicing an average of at most 10 minutes per day, I can play more complex, intricate, and musically interesting things than I ever could before. It's also a lot more relaxing.

My sense (at least for me and the way my brain works) is that the idea of a smooth transition from explicit learning using declarative memory creating a scaffolding that will slowly drop away as procedural memory emerges is overly simplistic. My sense is that there is a huge difference between explicit and implicit learning methods with regard to the brain structures that are created and what those brain structures allow me to do. For more complicated things (reading a foreign language or improvising music), my experience is that explicit methods tend to lead to automated declarative functions whereas implicit methods tend to lead to procedural functions. Twenty years ago, I could easily tell you what I was doing for pretty much every note I played when I was improvising. Today, I cannot verbally describe the more interesting things I play except if I go back and try to analyze what I did. My ability to improvise on piano is now primarily procedural and my ability to analyze and communicate what I play is primarily declarative.

My understanding is that individuals differ in their capabilities for using working, declarative, and procedural memory, and not to mention in their interests, desires, and needs for a particular skill. This strongly suggests to me that the optimal mixture and order of explicit and implicit learning methods will vary person to person and that using a highly suboptimal order could slow someone way down. It also suggests to me that any transitions between declarative, automated declarative, and procedural will be at least somewhat different from person to person.

[Cainntear]

It makes a strong dichotomy between declarative controlled practice and procedural loosely controlled practice, disregarding the whole spectrum in between -- the concept of slowly removing scaffolding to make practice increasingly procedural and decreasingly controlled shouldn't be a controversial topic; whether in language learning or any other field of learning.

My sense of things is that there is also something I've seen referred to as "automated declarative" that can mimic procedural to a limited degree. I suspect many people who improve but then hit long-term plateaus in spite of effort to improve weren't dropping scaffolding but were rather just speeding up use of declarative as much as they can. My experience in both music and language learning is that the scaffolding never dropped away until I intentionally tried to stop using declarative. I had to explicitly decide to use methods that focused almost exclusively on developing procedural into order to start using procedural to any degree. The change was then rather dramatic rather than gradual. I regret the decades of wasted practice and potential improvement I lost because of this.

I pretty much agree with you, I just think we diverge at one late moment.

My second major epiphany in high school over languages was when I realised that as I'd built up a quicker and quicker ability to process the rules, there was a second part of my brain that had learned to do it automatically. It was like a second voice in my head popping out and shouting "I know this! I know this!", but I was ignoring it while I thought through the answer. Once I knew I was doing it, all I had to do was learn to listen to it rather than blocking it out.

When I went to university and started learning about neural networks in artificial intelligence, it all made sense to me. The brain works in parallel -- it's not just processing one possible thought at a time. When it knows the initial intention and it knows the correct output, it will try to wire up the optimal path to get the output based on the intention. The computational model is an oversimplification of the biological one, but there's an important truth at it's core: we need to know both the input and the output to build up the process. Extensive listening leaves us in doubt about the input (intended meaning) and we have to infer that from the output (language) we receive... but as we haven't built up the circuitry, our inference is going to be uncertain at best, and wrong at worst. Uncontrolled practice gives us a situation where we've got the intended meaning, but we don't know the correct target output, and no-one else can help us with it because they don't actually know our intended meaning -- only we do.

When working with a teacher, if the teacher doesn't actually know our intended meaning, they have to infer it from our output, and if they infer wrong, their "correction" will only make things worse.


Anyway, as for music, I had an epiphany with that too. I used to sing in a choir, and sometimes played in the backing band. There was a mixture of people with formal training who couldn't improvise and people without formal training who couldn't read music. I myself had moved from being a play-from-sight-and/or-memory player to a mostly by-ear-and-improvising player.
What struck me was that when any semi-competent musician plays a wrong note, it's not usually totally wrong -- a beginner musician might hit a discordant note, but anyone intermediate and above is likely to hit something harmonically related to the current chord and the intended note. This shows that the musician has already internalised the rules procedurally as well as learning them declaratively -- the wrong note is just improvising by any other name.

Again, learning to improvise is just learning to trust the procedural part of your brain and accept that it knows what it's doing.

[Le Baron]

And this is exactly why I don't feel much like discussing how to learn any language, etc. any more -- how hard can it be to understand that this is the core?

I feel like I agree (I've said similar before) and yet I think about the difference between what people who frequent a forum like this do and teaching languages to people whose lives don't revolve around languages.

Those people who scorn the 'natural method' have at least one valid point: that recreating the 'natural' situation, immersive acquisition, in order to learn a language similar to how L1 was learned, is not a two-hour a week thing. The difference between acquisition in an immersion situation and living your normal life outside the target language, whilst recreating a form of semi-interactive immersion, is palpable.

The old view where people were grounded in essential grammar (or rather a lot of it) and learned to read in the language was seen as being useless for or divorced from creating speakers of the language. So the shift in the late-sixties to seventies was to active learning and acquisition through use. And grammar importance probably got a bit lost.

However I wonder if the methods a lot of dedicated language learners follow, large-scale input, is ever suitable for teaching languages as a subject. Your average student can't do that much input alongside everything else. In the older system you at least could understand grammar properly enough to read and when you can read you have the tools/potential to learn a lot of words... And when you learn a lot of words, as we know, you sort of 'know' the language.

The only gulf is ever between the written/spoken forms. So it revolves in a sort of circuit and it's really the case that you need a bit of both approaches.

[mrwarper]

I feel like I agree (I've said similar before) and yet I think about the difference between what people who frequent a forum like this do and teaching languages to people whose lives don't revolve around languages.

Lucky I (used to) frequent a forum like this and I teach people whose lives absolutely don't revolve around languages. And I am not alone, so others might chime in as well.

The difference between acquisition in an immersion situation and living your normal life outside the target language, whilst recreating a form of semi-interactive immersion, is palpable.

The difference is opportunities 24/7 vs. what you can get, 'multiplied' by what you would make out of those opportunities. I have seen people not learn to order a drink after many years of immersion, I have seen people make their own extra opportunities because they were (predictably) not having enough with what came their way. Everyone is a bit different, so this is impossible to measure except in such hypothetical terms.

The old view [...] the shift in the late-sixties to seventies was to active learning and acquisition through use. And grammar importance probably got a bit lost.

Yep, the bad old pendulum movement that Wilfried Decoo explained (and documented) so well. The only thing I would have to add, a language learning industry will always fuel this dynamic because selling materials is as major a goal as maybe actually getting people to learn.

However I wonder if the methods a lot of dedicated language learners follow, large-scale input, is ever suitable for teaching languages as a subject. Your average student can't do that much input alongside everything else.

Students can and do range from A1 to C2. Simplifying, there's a big tipping point at B2, and it is that they should have all grammar under their belts by then. That means from then on they are supposed to be independent learners, which harmonizes quite well with the CEFR definitions, and 'all' they have to do from that point is basically what we do here (don't we all do variations of the same core stuff?), so the sooner they learn what to do and start doing it, the better for them, and the more time they have to experiment and adopt a personal variation that fits them best. And that's what I try to teach them.

For my sub-Cs, mock 'independent learning' is something I make them start doing in fun and in tolerable doses to get the hang out of it while they wait for the day when it is 'the only thing' they will need to keep doing after they can leave the nest. Meanwhile, I keep them busy enough trying to get appropriate command of grammar bits more or less as laid out in the CEFR-oriented textbooks by old-fashioned study and explanations coupled with showing how it is all used and put in action in real materials they like (books, movies, etc.). If all goes well --not often the case--, the proportion of total time devoted to teaching and checking whether they are learning to work by themselves vs total should roughly go from 0% to 100% as they move from A1 to the end of B2.

As for the B2+s (those who come to me for the first time with a B2 certificate), ideally I could explain how to learn independently and get them going in a month or two perhaps, but in practice this is where shortcomings dragged from previous stages of their personal processes become more prominent. So in the end getting them ready to go is a mixture of teaching how to learn independently --which is usually new to them, and thus as much of a shock as often a struggle-- and 'hole plugging' in varying degrees. Oh, and a lot of showing how the 'driving license' approach is nearly useless at the C levels.

I also try to teach some other 'off-beat' stuff such as phonetics so people have it easier to get decent pronunciation if they decide to make some effort in that direction.

Teaching like this is highly rewarding when I am allowed to teach what students really need and keep up some rhythm. However I often meet rather passive people who I would not even call students, and we may end up doing 'homework' in class, because it is the only way to get any work done at all. And it is getting worse...

There bust be many factors, but I would highlight a mix of plummeting standards and ever growing absurd curriculum requirements: when I finished high school you could be sure of being functional at a B2 level if you took advantage of your language classes, and some command of foreign languages was a reasonable recommendation for those engaging in the Erasmus program and the like, as it should be. Fast forward to the present, and now a B1 certificate in some foreign language is a formal requirement to get any degree, while we have loads of undergraduates enrolling in parallel language classes to get that. This clearly implies that high school levels have gone to the dogs, and a high percentage of people who engage in classes because it's required makes it obviously more difficult to find the students who have a more sustainable interest in learning.

Regarding certification exams, the trends above naturally result in people shifting exam targeting towards those which are perceived as easier, or that openly aim at the lower end of the same CEFR band as others (a B1's a B1, right? ; )

The only gulf is ever between the written/spoken forms. So it revolves in a sort of circuit and it's really the case that you need a bit of both approaches.

Transfer between the two gets better the further you get, but it should be obvious that both will always need some practice, except maybe for those learners with equally unbalanced needs, f.e. people who only want to read technical literature or be conversational and maybe able to watch TV.