Grammar-Translation, Translation Drills, Assimil Active Wave, Back-Translation, and Bruce Lee

[drmweaver2]

You lost me there, boss.

Sorry that I misunderstood you. It would be easier for me to follow you if you would express yourself at an appropriate level of language, i.e. objectively and matter-of-factly, and as clear as possible.

Okay. I'll simplify it.
Your argument makes no sense. Your premise is fallacious, your logic is flawed and therefore, your conclusion is invalid.
Simple enough?

Other than that, I followed you quite easily.

[neumanc]

You lost me there, boss.

Sorry that I misunderstood you. It would be easier for me to follow you if you would express yourself at an appropriate level of language, i.e. objectively and matter-of-factly, and as clear as possible.

Okay. I'll simplify it.
Your argument makes no sense. Your premise is fallacious, your logic is flawed and therefore, your conclusion is invalid.
Simple enough?

Other than that, I followed you quite easily.

If you don't want to discuss seriously, you shouldn't take part in discussion forums. That's all I will say to that. As always, serious contributions are welcome.

[drmweaver2]

If you don't want to discuss seriously, you shouldn't take part in discussion forums. That's all I will say to that. As always, serious contributions are welcome.

Wow.
Your initial post is logically fallacious on its face, starting from a bad premise.
Your response to my reply was insulting.
I clarified my initial reply in as simple a language as possible - as you requested.
I have been entirely serious.

And you essentially accuse me of not being serious and invite me to not respond?

Look in the mirror.

I'll reword my initial reply just in case what seems to be impenetrable isn't: It is absolute rubbish for you to assert there is something detrimental about memorization in language learning. I can't say it more simply than that.

You have a nice day.

[StringerBell]

My assumption is that doing translation drills will never be as effective than having thoroughly internalized 3,000 ready-made sentences that can be used and mixed with each other in daily life.

This is something I'm currently thinking about, since I'm halfway through Jak Uczyć Się Języków (How to Learn Languages) by Luca Lampariello. In this book, he explains the methods he uses to learn languages and why he thinks they are efficient and necessary. When he's in the beginning stages of learning a new language, he relies on bi-directional translation to help him focus on the structural patterns of his target language and how it differs from his native language.

According to him, this is a really important way to activate procedural memory (how to assemble the components of the language to speak it) as opposed to solely relying on declarative memory (memorizing individual words). According to him, one of the biggest mistakes that language learners make is to only focus on declarative memory and completely ignore procedural memory, which is why it's really common for people to have hundreds or thousands of words memorized but then not be able to formulate sentences when they want to speak the language.

So far I don't agree with everything in the book, but this idea has gotten me wondering more about the benefits of doing translation, and whether it actually makes a difference compared to just speaking more earlier on. My initial belief was that memorizing a lot of common, ready-to-use phrases was the most efficient way to go, but now I'm not really sure.

[neumanc]

My assumption is that doing translation drills will never be as effective than having thoroughly internalized 3,000 ready-made sentences that can be used and mixed with each other in daily life.

This is something I'm currently thinking about, since I'm halfway through Jak Uczyć Się Języków (How to Learn Languages) by Luca Lampariello. In this book, he explains the methods he uses to learn languages and why he thinks they are efficient and necessary. When he's in the beginning stages of learning a new language, he relies on bi-directional translation to help him focus on the structural patterns of his target language and how it differs from his native language.

According to him, this is a really important way to activate procedural memory (how to assemble the components of the language to speak it) as opposed to solely relying on declarative memory (memorizing individual words). According to him, one of the biggest mistakes that language learners make is to only focus on declarative memory and completely ignore procedural memory, which is why it's really common for people to have hundreds or thousands of words memorized but then not be able to formulate sentences when they want to speak the language.

So far I don't agree with everything in the book, but this idea has gotten me wondering more about the benefits of doing translation, and whether it actually makes a difference compared to just speaking more earlier on. My initial belief was that memorizing a lot of common, ready-to-use phrases was the most efficient way to go, but now I'm not really sure.

Very interesting. I didn't read Luca's book, but I would be interested in hearing about his ideas more in depth. I don't think he's quite clear in his YouTube videos (maybe intentionally, because he has something to sell?). However, speaking of the difference between procedural and declarative memory: I absolutely agree that using your procedural memory is good for you. However, doesn't translating and back-translating sentences rely on declarative memory, i.e. conciously searching for words and consciously applying rules? On the other side, isn't speaking, repeating, and reciting sentences from memory as advocated by Alphonse Chérel something that relies on, and fosters, procedural memory (i.e. knowing how to do something, but not necessarily why)? So maybe Luca advocates in fact memorization (although calling it bi-directional translation)? I would very much like to know more about this subject.

[Xmmm]

What's the difference between memorizing 3000 sentences out of a book, and watching 500 hours of TV where you're exposed to a stream of roughly 2,000,000 words in the target language?

For Italian, I found all these drills completely unnecessary. I just watched several hundred hours of TV and bang, I could speak. I've done only 40 hours of conversation, but it's enough that I've eliminated almost all of the basic grammatical errors, have pretty good fluency, and mostly now it's just a question of lacking vocabulary and having constructions that are too simple and unsophisticated. I'm not sure I could even get more sophisticated by drilling or memorizing sentences. I think I just need to read (which I'm doing).

For Russian, I found these sorts of drills do give a leg up in conversation. FSI and Glossika more than Assimil -- but maybe I was using Assimil wrong. However -- a leg up to what? I mean, essentially the drills allow me to occasionally express something cleaner and in a more sophisticated form than I would otherwise be capable of. But ten seconds later the conversation has switched, and I'm back to sounding like a halfwit. Deep down inside I feel the real answer is more input -- 1000 hours, 2000 hours, whatever it takes.

So, even though I'm doing Modern Russian end-to-end because Speakeasy guilt tripped me -- I'm skeptical about how fantastic the results are going to be.

Another thing I've noticed is drilling individual vocabulary words with a production emphasis is very helpful. This production flash card review will make the word very available in conversation, activating it as it were. You can put a really simple sentence together, pop in the right word, and you sound a heck of a lot better than you really are. For example:

Prior to the US Thanksgiving Day holiday, I had a tutoring session in Italian. I had been reading Il grande gioco, where English and Russian spies are constantly dressing up as Muslim pilgrims to infiltrate various parts of Afghanistan. And so in some sentence or other I was able to say i pellegrini instead of "you know, the people who came on the boat from England." I felt very raffinato.

I thinking I'm trying to say: "It might be better for a learner to develop a vocabulary of 10,000 words and let the complicated sentences emerge over time, rather than drill complicated sentences endlessly with 3000 words."

[neumanc]

It is absolute rubbish for you to assert there is something detrimental about memorization in language learning.

We absolutly must have misunderstood each other. I am advocating memorization in language learning. However, one should memorize the right things. The best I could think of are complete sentences embedded into memorable stories--just like Assimil lessons. Everything else like isolated vocabulary, dictionaries, etc. seems to be less effective. And please let's not get personal. After all, we all are here for the same reason--to learn from each other and to help each other in language learning. If I insulted you, I am sorry.

[David1917]

My Assimil sessions are about 30 minutes each, the back-translation of dialogues is immensely useful, as is traditional grammar-translation. Bear in mind, and I may have mentioned this in the thread where this was brought up - but in a GT manual you are given a vocabulary and expected to manipulate it grammatically in the exercises. So it's not really off-the-fly or that different from translating a dialogue you are already familiar with.

I'd also like to propose a distinction between internalization and memorization. This is admittedly a very fine line, but worth mentioning nonetheless. I would define memorization as being able to recite completely from memory, and internalization as being very familiar with. You see a movie 2-3 times and you can anticipate "this part is great," but you see it 10 times and you're annoying the people around you by reciting actors' lines before they're actually spoken.

With Assimil, I would say you are more likely to internalize the dialogues upon successive reviews, and so when you reproduce them via translation, that's when you lock in the procedural aspect of producing language. I have to compare it to music - heard a song a few times, can play bits and pieces by ear, but you don't memorize it until you have a look at the sheet music and play through it enough times that you can perform without it. At various points in my day, a line or 2 of various Assimil dialogues will float through my head, but I certainly cannot produce any of them in their entirety from memory.

Here is the Assimil scheme I use: Each lesson corresponds to a theoretical day of the week, as is the intention of the course. I review all of the lessons for that "day" in each session, after having previewed the 7-10 subsequent lessons. Let's say I'm on lesson 75. I'll first do lessons 82-75 with the various stages of shadowing. 75 I will either stop after each line to focus on any troublesome points, and possibly even write it down (I'm doing this with French since the spelling is a major issue, whereas I do not for German because I have a decent grasp on the written form.) Then comes the review stage, where I go back in 7's to lessons 68, 61, 54, 47, 40, 33, 26*, 19, 12, 5. By this point, Lesson 5 is obviously incredibly simple and familiar, and due to the way the actors speak in the introductory lessons, it's a nice way to reinforce/revisit pronunciation by slowing down a bit.

In the above example, *26 is most likely the 2nd wave lesson. So I'll stop here to do the translation and really lock in the structural elements. I'm still not really trying to memorize at this point - just fully understand how/why it's working the way it is. I expect reinforcement from continued review and, more importantly, from other sources.

I don't think any one method is better than another, nor is any one course. They all complement each other if well-chosen. In this case, I think the internalization of dialogues and structures, coupled with a deeper analysis of same via the active wave, plus a healthy dose of grammatical translation work and pattern drills to push towards automaticity all are important to create a well-rounded regimen to become better at conversing in a language.

[Cainntear]

My issue is how vague and fluid the definition of ALM is.

My first understanding of the term was grammar rules and phrase/sentence-translation exercises, but then I saw someone describe it as being a method for learning to read the classics, and actually an "authentic materials" approach because it actually took in passages of proper classical Greek/Latin really early on. Now there's talk of back-and-forth translation in the mix, so I don't see the point of even talking about "ALM". (Which is not a criticism of you, as you've been very clear on what you're talking about.

Third, by way of explanation, it seems to me that both here on the LLORG and on the HTLAL, there have been a few brief discussions of how one should proceed with Assimil Active Wave. However, I have not come across a clear statement of the benefits of this type of practice, of its possible weaknesses, or of any alternatives. As there has been not much discussion, I suspect that most users of the Assimil courses simply ignore the Active Wave but that, were they to be queried on the matter, they would likely reply: “Oh, yes, I always complete the active wave, it works just fine!”

I've only completed one Assimil (Catalan) and I didn't do the active wave. In fact, I just had to pull an Assimil off the shelf to check what the active wave actually is. I think I was put off by the idea of going right back to lesson 1 -- surely after 50 lessons of Assimil, lesson 1 is going to be stupendously easy? I get that receptive skills trail productive skills, but surely not by that much. I suspect that the nature of the active wave isn't a pedagogical decision as much as a practical one -- "the medium makes the method", as Wilfried Decoo said, and the content of the active wave is led by trying to keep down the amount of paper needed in the book (and until and unless Assimil realise this, they'll be unnecessarily hampering the e-Méthode product by continuing as though paper was still of concern).

Ultimately, it was the absence of both (a) an audio prompt and (b) an opportunity to compare my vocalized translation with a recorded model, that put me off this type of practice. In other words, the translation activity in itself may very well be beneficial; however, to my mind, the limitations of the Assimil recordings prevent its being exploited properly.

Yup -- I hear ya.
I would hope that the e-method interface makes this process more user-friendly. Is there anyone here who has tried it and would like to comment?

2. Another alternative are the oft-decried audio-lingual sentence-pattern drills, the Glossika mass sentence files, and the possibilities of SRS applications such as Anki. In these types of exercises and applications, the user is provided with an L2 auditory prompt and a short pause in which to respond. Some exercises involve L1-L2 translation practice. Recorded materials are available up to and including the lower-intermediate level.

I would say that they are not "an" other alternative, but "other alternatives". My biggest beef with audio-lingual drills is how little they rely on recall. If you have a pattern drill or a substitution drill, almost all of the sentence is given to you, and you are never asked to recall the target structure from memory.

I think that's the key thing to think about in any prompt-response teaching/learning format -- recall, and the variation between prompts.

At one extreme we have audio-lingual drills where there is very very little grammatical variation (low demand on memory); at the other extreme we have SRS where there is no attempt to relate one prompt to the next and you could have things that are entirely unrelated (high demand on memory).

Pimsleur and MT sit in the middle and are very different beasts. Pimsleur's variations are moderately minimal (Do you have reals? No I have dollars.) and tend to be quite predictable (so areasonably low memory load). Thomas, on the other hand, managed his variation quite deftly -- he would focus on one new point and keep the variation in it reasonably low, but integrate it with other, better known language that would vary more; and it was very difficult to predict what he was likely to ask next, which meant that the student always needed to use memory at all times.

3. Another alternative involves question-answer exercises. In this type of practice, the user is listens to an L2 recording of a dialogue or narrative for which a transcript is available. He then listens to, and responds to in the pause provided, a series of questions in the L2 on the content of the recorded passages. Finally, he listens to a recorded model response and repeats it in the pause provided. To a large extent, this type of practice is similar to the Pimsleur and Michel Thomas programmes: prompt, think, reply, compare.

One of my issues with L2 QnA is that you're stuck with a choice between two suboptimal solutions:
1) The student must reply to the prompt in full sentences mirroring the question in order to practice grammar. However, in natural conversation, we only reply in full sentences if we are not mirroring the question. Unnatural answer-in-sentences seems to be tiring, boring and off-putting to most students.
2) The student responds naturally in sentence fragments. Obvious downside -- no practice of sentence syntax.

The other issue is the frustration of giving an answer you think is correct, hearing a different answer, and not knowing whether you made a mistake or just used a different form from the course author. You can't learn from your mistakes if you don't know you've made one!

[neumanc]

My Assimil sessions are about 30 minutes each, the back-translation of dialogues is immensely useful, as is traditional grammar-translation. Bear in mind, and I may have mentioned this in the thread where this was brought up - but in a GT manual you are given a vocabulary and expected to manipulate it grammatically in the exercises. So it's not really off-the-fly or that different from translating a dialogue you are already familiar with.

I'd also like to propose a distinction between internalization and memorization. This is admittedly a very fine line, but worth mentioning nonetheless. I would define memorization as being able to recite completely from memory, and internalization as being very familiar with. You see a movie 2-3 times and you can anticipate "this part is great," but you see it 10 times and you're annoying the people around you by reciting actors' lines before they're actually spoken.

With Assimil, I would say you are more likely to internalize the dialogues upon successive reviews, and so when you reproduce them via translation, that's when you lock in the procedural aspect of producing language. I have to compare it to music - heard a song a few times, can play bits and pieces by ear, but you don't memorize it until you have a look at the sheet music and play through it enough times that you can perform without it. At various points in my day, a line or 2 of various Assimil dialogues will float through my head, but I certainly cannot produce any of them in their entirety from memory.

Here is the Assimil scheme I use: Each lesson corresponds to a theoretical day of the week, as is the intention of the course. I review all of the lessons for that "day" in each session, after having previewed the 7-10 subsequent lessons. Let's say I'm on lesson 75. I'll first do lessons 82-75 with the various stages of shadowing. 75 I will either stop after each line to focus on any troublesome points, and possibly even write it down (I'm doing this with French since the spelling is a major issue, whereas I do not for German because I have a decent grasp on the written form.) Then comes the review stage, where I go back in 7's to lessons 68, 61, 54, 47, 40, 33, 26*, 19, 12, 5. By this point, Lesson 5 is obviously incredibly simple and familiar, and due to the way the actors speak in the introductory lessons, it's a nice way to reinforce/revisit pronunciation by slowing down a bit.

In the above example, *26 is most likely the 2nd wave lesson. So I'll stop here to do the translation and really lock in the structural elements. I'm still not really trying to memorize at this point - just fully understand how/why it's working the way it is. I expect reinforcement from continued review and, more importantly, from other sources.

I don't think any one method is better than another, nor is any one course. They all complement each other if well-chosen. In this case, I think the internalization of dialogues and structures, coupled with a deeper analysis of same via the active wave, plus a healthy dose of grammatical translation work and pattern drills to push towards automaticity all are important to create a well-rounded regimen to become better at conversing in a language.

Thank you, David1917, for this very informative post. I very much like your systematic approach, I might as well try it out! Possibly, what you are doing (minus the shadowing part) comes closest to what Chérel actually had in mind. At least, I now can see that there might be something more to the active wave than I had imagined. I especially like the distinction between internalization (familiarizing yourself with the lesson content up to the point where bits and pieces can be remembered easily) and memorization (learning the dialogues by heart word for word). It's very true, memorizing whole dialogues or even sentences, if they are longish, is difficult to achieve and most probably not necessary for language learning. After all, you aren't supposed to recite the dialogues while speaking your target language, but to compose your own sentences by means of combining and re-combining the bits and pieces of the sentences drawn from memory. Maybe, I am overdoing it with Assimil and should revise less and/or study less intensively and/or more systematically? In any case, this would result in quicker progress through the lessons and, especially, leave something to be gained from the active wave.

What I've been doing lately is the following: (1) Shadowing through the new lesson a few times in order to learn the pronunciation. (2) Repeating each line aloud several times, after having read it aloud, as often as necessary, so that I can do it flawlessly, fluently, and swiftly. (3) After that, looking on the right hand side (the teaching language) and recalling the target language sentences from memory. In most cases, this will be possible to a very large degree. (This observation led me to the conclusion that memorization might have been what Chérel actually had in mind.) (4) Revising whole bunches of older lessons (i.e. 22 to 28, the next day 29 to 35, and so forth, eventually restarting at the beginning of the book) while repeating steps two and three. (5) At irregular intervals shadowing through large parts of the course in order to reinforce the correct pronunciation.

The advantages of this approach are that it's absolutely free of frustration and will result in internalization, if not memorization, of the course. The disadvantages are that it takes quite a long time to accomplish, thereby possibly hindering progress. Furthermore, I'm possibly depriving me of the advantages of doing back-translating exercises.