Sewing Machines and Language Learning
Isn't foreign language learning due for a conceptual leap?
We live in a marvelous age: When I dial Fed-Ex, the computer converses with me easily about my shipping needs. An off-the-shelf Bluetooth earpiece/microphone understands "Call Home" the first time you attach it to your mobile. If your cholesterol is too high, they've got a drug for that. And don't even get me started on the internet and what it's wrought... And yet, if you look at a language learning book from 40 years ago and compare to what's on the shelves, there's often not a lot of difference. Sure, the book is cheap and comes with a CD instead of being extremely expensive and coming with a phonograph record. But there haven't been that many conceptual leaps forward from the old Language/30 dual rep cassette packs. There are two exceptions that come to mind: Michel Thomas and Pimsleur. Thomas apparently had a theory about how to teach just about anything and chose language to demonstrate it. And Pimsleur had a theory about memory and language that his programs implement (making him the audio forerunner of the SRS). What's interesting is that both Thomas and Pimsleur were thinking less about language and more about how we learn, and so their programs escape what seems to hold back many programs: They don't simply describe (or, like Rosetta Stone, merely present quasi-structured content) and rely on the learner to assimilate the teaching. They're actively engaged in making learning happen.
I think Thomas and Pimsleur are/were on to something, but I think we're going to find ways to take it further. I'm not sure what these will be, or when they will come, but there's one thing I'm fairly sure of: It's going to be psychologists and learning specialists, not language teachers, that show the way. Because what we need to get started in a language is not more knowledge, but a faster, more efficient way of getting the knack for a language's patterns as a native speaker (unconsciously) understands them.
Conceptual Leaps in Understanding - Recent Experiences
In my Uzbek, I recently got bored with continued efforts at mastering the content of the DLI phrasebook. So I've done two things lately: 1) skimmed a late 1800s grammar of Uighur and 2) started flipping through the Lonely Planet Central Asia phrasebook. I noticed two things: First of all, while I would have no hope of memorizing all the tables in the 1800s grammar, I had no trouble making sense of the more common paradigms with a couple hundred phrases memorized, many of whose patterns were familiar to me but not completely understood. Second, going back to the LP phrasebook sentence patterns started leaping off the page at me and even when I didn't understand why a particular sentence was put together a certain way I could certainly see the constituent elements. What I'm describing, of course, is nothing novel. It's what happens to any language learner at different points along his/her progression in learning. The question is: How do we speed it up?
Here comes the sewing machine!
Over at the Volokh Conspiracy, a law-blog, there have recently been a series of guest posts about, of all things, the sewing machine. In a nutshell, the sewing machine is a pretty mundane thing today, but at the time of its creation there were two big issues that came into play: 1) A new conception of what sewing is was needed in order to mechanize it. 2) A way was needed to recompense inventors from multiple teams through patents and licensing without causing invention and improvement to come to a halt because the possibility of infringement limited new inventors' wiggle room for improving on an as-yet imperfect idea.
Concept
Here's the conceptual part:
Invention, Refinement and Ownership
There's a second element here, and it's as big a concern as making a conceptual leap from our traditional understanding of language learning should such a leap appear to be needed or useful: Today we have Rosetta Stone, Pimsleur, Michel Thomas, Tell Me More, Assimil, Berlitz and more. They're all companies with trademarks, licensing arrangements and patents. But they're also companies with a wealth of experience creating language content. So, if the big breakthrough comes, do we leave all these companies, their talent and their experience behind while one company sits on the secret for learning languages well and quickly but can only bring out one or two language programs a year? What if the breakthrough is found at a public university? We have the language materials we have, both good and poor, for two reasons: 1) Language is an intriguing thing that draws people into studying it, hence university programs. 2) Language is big business.
I have a feeling that with advances in technology for studying the brain, plus an internet that lets people keep track of great amounts of information on arcane subjects, we're headed for some breakthroughs in learning in general and language learning in particular. It will be curious to see how things unfold when it's time to bring these to market.
A final note: Humanity worked out a way to stitch together clothes a hundred times faster than before once we stopped focusing on how we had sewn before and looked instead at what we were trying to achieve with sewing. So if in your own learning you stumble across an approach that seems bizarre but really works for you, keep at it and ask your friends to see if it works for them. And then, if you can, patent it!
We live in a marvelous age: When I dial Fed-Ex, the computer converses with me easily about my shipping needs. An off-the-shelf Bluetooth earpiece/microphone understands "Call Home" the first time you attach it to your mobile. If your cholesterol is too high, they've got a drug for that. And don't even get me started on the internet and what it's wrought... And yet, if you look at a language learning book from 40 years ago and compare to what's on the shelves, there's often not a lot of difference. Sure, the book is cheap and comes with a CD instead of being extremely expensive and coming with a phonograph record. But there haven't been that many conceptual leaps forward from the old Language/30 dual rep cassette packs. There are two exceptions that come to mind: Michel Thomas and Pimsleur. Thomas apparently had a theory about how to teach just about anything and chose language to demonstrate it. And Pimsleur had a theory about memory and language that his programs implement (making him the audio forerunner of the SRS). What's interesting is that both Thomas and Pimsleur were thinking less about language and more about how we learn, and so their programs escape what seems to hold back many programs: They don't simply describe (or, like Rosetta Stone, merely present quasi-structured content) and rely on the learner to assimilate the teaching. They're actively engaged in making learning happen.
I think Thomas and Pimsleur are/were on to something, but I think we're going to find ways to take it further. I'm not sure what these will be, or when they will come, but there's one thing I'm fairly sure of: It's going to be psychologists and learning specialists, not language teachers, that show the way. Because what we need to get started in a language is not more knowledge, but a faster, more efficient way of getting the knack for a language's patterns as a native speaker (unconsciously) understands them.
Conceptual Leaps in Understanding - Recent Experiences
In my Uzbek, I recently got bored with continued efforts at mastering the content of the DLI phrasebook. So I've done two things lately: 1) skimmed a late 1800s grammar of Uighur and 2) started flipping through the Lonely Planet Central Asia phrasebook. I noticed two things: First of all, while I would have no hope of memorizing all the tables in the 1800s grammar, I had no trouble making sense of the more common paradigms with a couple hundred phrases memorized, many of whose patterns were familiar to me but not completely understood. Second, going back to the LP phrasebook sentence patterns started leaping off the page at me and even when I didn't understand why a particular sentence was put together a certain way I could certainly see the constituent elements. What I'm describing, of course, is nothing novel. It's what happens to any language learner at different points along his/her progression in learning. The question is: How do we speed it up?
Here comes the sewing machine!
Over at the Volokh Conspiracy, a law-blog, there have recently been a series of guest posts about, of all things, the sewing machine. In a nutshell, the sewing machine is a pretty mundane thing today, but at the time of its creation there were two big issues that came into play: 1) A new conception of what sewing is was needed in order to mechanize it. 2) A way was needed to recompense inventors from multiple teams through patents and licensing without causing invention and improvement to come to a halt because the possibility of infringement limited new inventors' wiggle room for improving on an as-yet imperfect idea.
Concept
Here's the conceptual part:
The fundamental problem with these many independent inventions of the eye-pointed needle was primarily conceptual, not mechanical. The early efforts at using machines for sewing attempted to replicate the motions of the human hand in sewing fabric, i.e., driving a needle with a thread through a piece of fabric and then pulling the same needle back through to the other side of the fabric. In 1804, for instance, Thomas Stone and James Henderson received a French patent for a sewing machine that replicated hand-sewing motions by using mechanical pincers. Unsurprisingly, their machine was unsuccessful and saw only “some limited use.” As with the invention of the typewriter in the late nineteenth century, sewing-machine inventors needed to make a conceptual break between human-hand motion and mechanical motion.There's a key point here for language learners. Some language learning methods like Berlitz and Rosetta Stone try to provide a faster, better organized way of learning a new language the way you learned your first one. Pimsleur and Michel Thomas are more like a 19th century sewing machine: You've still got thread put through cloth by a needle (vocabulary used in sentences shaped by grammatical patterns) but the process for learning and using the material is different from the way people learn a language "naturally," just as a sewing machine does not in fact replicate hand motions. Not yet clear: As we learn more about how people learn and conceptualize language, will we ultimately discover that we need something closer to a natural method? Or something even more "unnatural"?
Invention, Refinement and Ownership
There's a second element here, and it's as big a concern as making a conceptual leap from our traditional understanding of language learning should such a leap appear to be needed or useful: Today we have Rosetta Stone, Pimsleur, Michel Thomas, Tell Me More, Assimil, Berlitz and more. They're all companies with trademarks, licensing arrangements and patents. But they're also companies with a wealth of experience creating language content. So, if the big breakthrough comes, do we leave all these companies, their talent and their experience behind while one company sits on the secret for learning languages well and quickly but can only bring out one or two language programs a year? What if the breakthrough is found at a public university? We have the language materials we have, both good and poor, for two reasons: 1) Language is an intriguing thing that draws people into studying it, hence university programs. 2) Language is big business.
I have a feeling that with advances in technology for studying the brain, plus an internet that lets people keep track of great amounts of information on arcane subjects, we're headed for some breakthroughs in learning in general and language learning in particular. It will be curious to see how things unfold when it's time to bring these to market.
A final note: Humanity worked out a way to stitch together clothes a hundred times faster than before once we stopped focusing on how we had sewn before and looked instead at what we were trying to achieve with sewing. So if in your own learning you stumble across an approach that seems bizarre but really works for you, keep at it and ask your friends to see if it works for them. And then, if you can, patent it!
5 Comments:
It's all well and good to say 'we need a breakthrough' but nobody ever had a breakthrough by sitting around and saying they needed one.
Also, you're making the assumption that there's a breakthrough to be made, and that the breakthrough can be made with modern technology. The modern sewing machine could not have been made before the assembly line was. And even if you hand-made a prototype, it couldn't have been done in the bronze age.
One last thing: You point to books and say 'they haven't changed!' but that's the same as pointing to the same old needle and saying 'it doesn't work any different!' The needle had to be changed to get the job done, and books probably do, too. By that, I mean that a static bunch of information is nice as a reference, but as a learning tool, it's vastly inferior to truly interactive learning devices.
Take Anki/Supermemo/etc (SRS software) for example. While comprised of sets of static information, the interaction the present is optimized for rapid knowledge acquisition. If someone were to sell an Anki deck that provided a fast-track to learning a language, many would buy it. (Many would pirate it as well, but they do that to books, too.) The decks I've seen so far are made by amateurs and make no attempt to organize the information in an optimal order. This is partly because nobody sees any money to be made in it (Anki is free, the existing decks are free, etc) and partly because advanced learners and fluent speakers have no interest in putting together a proper deck for free.
So maybe we don't even need this 'breakthrough'. Maybe it's already there and nobody is taking advantage of it yet.
Good article. There is a difference between learning quickly and learning effectively.
I have used Assimil, and I can say that the content is too limited. A wide variety of content helps make connections.
I guess I could write all day about this, but I won't.
In response to your last 2 sentences: There are few of us working with the TV Method. But it is not patentable.
Keith,
Thanks for the pointer to the TV method. I'll have to check it out.
William,
Thanks for the thorough comment. You raise a lot of good points that actually pave the way to thinking our way to the next breakthrough. You're especially on target with this: The decks I've seen so far are made by amateurs and make no attempt to organize the information in an optimal order. If it's the language acquisition people who make the breakthrough, the "technology" may well be in the sequencing of information for delivery, not how it is delivered. In fact, I think this is likely the case.
As for the failure of fluent speakers to put together decks, I don't think they/we can help as much as one would think. Looking back, I have no idea what parts of my learning were laying the foundations for my language breakthroughs and which were wasted time. All I know is that what I have done cumulatively has brought me to where I am today. This is why I think learning specialists, not language teachers, will solve the next riddles. This is also why I object to those books and articles where a language learner talks about all his stumbles, then says that that's all nice but here's the way he'd learn today: It's doubtful the method he's pieced together would work were it not for all the background knowledge he's picked up.
As a final note, it's distinctly possible the breakthrough is out there, just as the sewing machine was out there for twenty years before it caught on. Maybe someone will figure out what Michel Thomas was really doing and take it to the next level for example. At that point, we come to my second point: If you've got the breakthrough, how does it get marketed profitably so that it doesn't get lost.
Language education and music education have much in common -- enough so that this draft paper on an alternative approach to music education might be of interest:
www.igetitmusic.com/papers/JIMS.pdf
It argues that (a) musical information has an inherent structure; that (b) this inherent structure is largely hidden by traditional musical instruments and notation, and that (c) exposing this inherent structure in novel notations and instruments could make music education more efficient.
As with the sewing machine example cited above, the key to this "breakthrough" is conceptual. And just as the sewing machine was not commercially feasible until after the Industrial Revolution brought down the cost of making and assembling finely-machined interchangeable parts, so the ideas in the above paper would not be feasible without electronic music synthesis, notation software, and electronic (MIDI) controllers.
Your comments welcome, at www.igetitmusic.com/blog.
Jim,
Nice paper. I would quarrel with footnote 2 and the plus ça change reference. If you're pushing JIMS, the better quote is that there's nothing GNU under the sun.
On a serious note, it would be interesting to try to conceptualize fundamental elements of language the way you have music. The thing is, most language learners come to it with quite a bit of background in the structure and terminology of their own language, albeit very imperfectly acquired, whereas most novice music students have few ideas about music beyond enjoying listening or humming. If only there were a way to approach first language education with, eg, the diagrams they use in generative grammar.
Post a Comment
Links to this post:
Create a Link
<< Home