Practice, Correction, and the Closed Feedback Loop

21 Mar 2010
Posted by dvlong

When people learn about Automatic Language Growth, they often listen and then quickly conclude that a program of input combined with practice makes sense. [When we began ALG at AUA in 1984, we already had a successful traditional Thai Program. The ALG program grew up alongside it, and then at one point for about 3 years, we also ran a combination course of the natural and structural program.  While it was appreciated by our clients, it was not beneficial as the results were always inferior] I want to talk here about Control Theory and why Practice and Correction of speech do not work. Studies of children learning their native language show that correction (some parents correct their children and some don't) doesn't have any positive effect but can often have a negative effect. Mostly, though it has no effect. Language development for children who are corrected and those who are not is equal. We can see why when we consider how a closed feedback loop works. This is also the reason why changes in language, or any other ability can be so difficult to alter.

In Control Theory with regard to langauge, there are three basic parts to the negative feedback loop. The brain, the mouth, and the ears. We all understand that these three are related. It's also important to understand how they relate as well. The diagram to the right shows the basic parts. If things were as simple as this, language couldn't work very well. There are some parts we need to understand better.  They are aspects of brain function but we should understand them separately.  There is the Reference, and a Comparator.  The Reference, or what we call a MIF, (mental image flash), is vital if our language is going to work at all. This reference is the equivalent of the thermostat setting on an air-conditioning unit.  Without a reference that is constant, there is no way for the system to work properly.  The ambient temperature is compared (see the small blue square) to the reference setting and the difference tells the Controller what to do.

So how does this relate to language acquisition?  Traditional language teaching has focused on the System Output, or Speech primarily.  The reason that it is ineffective is due to the fact that it's all working on output  which has not effect on input unless there is no MIF to begin with.  Without changing the input, the output will never change significantly. But the output originates from the MIF - Just as it's impossible to walk like a bonzo, so you cannot say a word you don't have a MIF for.

Let's apply this to a couple of issues that we see in language education currently.

People often say that their students are shy, and don't want to speak. But when left on their own with  friends, they'll talk non-stop, just not in the language the teacher is trying to help them use.  What's happening?  They act shy because they have no clear MIF from which to say anything.  Once those MIFs exist, they demonstrate the same willingness to speak as in their native language.

Practicing pronunciation and drills are a common technique used by teachers. In an effort to help their students say if correctly, the teacher may lead their student to practice saying it correctly 10 times.  Most people can do this, however after a few days, the students resort back to the same pronunciation they had before.  What the teacher needs to realize is that MIFs cannot be altered by speaking.  Speaking offers output for the ears to hear, and then compared with the MIF, corrections will be make on the other side - not the speaking side.  The problem must be solved before the MIF is created.  Once a faulty MIF is in place, it may not be possible to correct it.

When people begin to acquire a new language, they often listen to us explain our program and then imagine that when they begin speaking they will speak perfectly. This is not the case. There are aspects of language that can only develop after speech begin. Language operates in a feedback loop and without speaking the loop doesn't exist. But the reference point must be stable before the mouth starts working on it. The key is not that it's supposed to come out correctly at a certain point, but that the language must rest clearly in one's mind, before speaking can naturally begin.  Once the MIFs exist for a language's phonemes, vocabulary and grammar, sentences can be improvised and we are able to speak.

Si, probablemente lo sea

Si, probablemente lo sea

 David, can I summarize parts

 David, can I summarize parts of your post as follows?In order to learn a language, you need to establish all relevant mifs first. Only then should you start to produce output. The output produced will not be perfect right from the start, but will be successively corrected towards the (correct) mifs through the control cycle you’ve described. Eventually, you will gain native-like fluency.How can you establish the correct mifs? By experiencing the language over and over again.What happens if you speak too early (i.e., before the mifs are firmly established), or if you analyze the language? Because you lack the correct mifs, you will substitute them with mifs taken from somewhere else, e.g., your native language. Once incorrect mifs are established, it’s very hard to replace them with the correct ones. 

David, this seems like a good

David, this seems like a good model for the process. Here’s question for you. Some people seem able to “learn” a language (witness the founder of the AUA Thai program, Dr. Brown) with traditional methods of study, practice and repetition. Why do some people seem able to actively learn a language to near native fluency (such as Brown) while most don’t. Most like in 99.9 percent. Everyone churning away with dictionaries and actively thinking while “studying” believes they will be the one, will self-correct and move onwards and upwards. In my experience it’s rare. Why is that?

 David, the problem I am

 David, the problem I am finding is that studying Thai in England as a hobby, with no Thai connections, I have to talk early. Talking Thai is the only way I have found to get interactions with Thai people interaction in Thai. I understand the concepts of the ALG method but I have to talk. My time is limited, dropping Thai and humour into brief encounters starts to reap rewards.My question would be if you are self-aware of the potential problems and why they occur do you think it is possible to offset them? 

Thank you for the reply

Thank you for the reply David, but I can’t seem to see my comment on the page, this probably looks a bit disjointed  .Does this mean that I can’t learn a language properly unless I already need it?My thought on the self-awareness thing was simply that being aware of the process allows me to control my internal state, and my own feedback loop.When I speak I just say what comes out naturally, I don’t force it, which means of course that after eight weeks there is a lot of English mixed in.So I go in order a coffee in Thai I have a Thai voice in my head for the greeting, the ordering of the coffee, I don’t adjust that voice on the basis of what comes out of my mouth, but what comes back at me in Thai is what interests me.In one restaurant get myself a table, order some food, order a beer in Thai, I have a relatively massive exposure to these sounds as I seek out Thais giving lessons on Youtube, podcasts etc. and let them wash over me, I don’t suddenly replace the voices in my head with mine because I know not to. The waiter asks me if I want chang beer in Thai, I reply in English now, Elephant? (I don’t have a good feel for this word in Thai yet) he laughs and mimics a big trunk and says chang chang back. I have that sound and association now, I have got it from him by using Thai.There is of course much more.I can’t see any other way for me to get Thai experience in my position. My Chinese connections came from me speaking Chinese, at some point in the not to distant future it looks increasingly likely that my work may involve me using Chinese, but this gravitational pressure comes from picking up Chinese.Your 800hrs figure to get 70-80% understanding sounds fine to me if I am in Thailand with the time to attend a course and swimming through all the Thai around me, anywhere else lets say I use video for input I guess that even videos of ALG lessons are not as affective as actually being there so that assuming I am using random video it may take 2400hrs for me to reach that point. I doubt I have the time for one hour of video that takes me to 6.7 years!!! Communicating with people is the fun but and I am not getting any younger. This made worse by the fact that I mostly have time just for audio.I am not trying to find the best, method, just the best possible compromise in the circumstances, else I am robbed of my Elephant and all the other little experiences to-date.I can never learn exactly like a Child because I have already been infected by one language meme already (nothing can replace that first kiss of thought).

Yes Andrej, this is a clear

Yes Andrej, this is a clear summary. Substituting MIFs is where the adult language problems come from.

I think this is true in most

I think this is true in most fields of study – the curriculum is written by the experts – in the case of language that means people like Dr. Brown. For whatever reason, these people have special ability. In Brown’s case he was able to remember without effort, all the things he needed to in order to reproduce language and it interested him. It was one thing he did easily. There were other things much more difficult for him that could take him weeks where you or I might not even stop to think about it and see our way clearly. In his case, social skills came very difficult for him. Many felt he was anti-social or that he didn’t like them. Of course this was not the case, but it was an area of difficulty for him. The point is, a curriculum for social studies developed by someone like Brown, would be much more engaging and interesting I think, that the sort of stuff we get in school today. Why? Because he would be developing it from the point of the average or “normal” person. Instead, we use experts who don’t begin to understand where the majority of the people are with regard to their subject. In the case of language, this has been devastating – so today we have an entire machine that is largely irrelevant – but people are sold on it never-the-less, not because it works but because they don’t have any choice that makes sense. We’re all products of this system in many ways.

My guess is that you are

My guess is that you are learning Thai because you want to use Thai. If you needed it, you would have Thai connections, and these could provide you with the input required.In answer to your question, no. Self awareness is the problem. Being aware of the language and how you’re relating to the sounds, grammar, or whatever is exactly what adults do that children don’t. Children generally seem to be only aware of the happenings – they might stop to look at a word here or there, or play with sounds, but there’s a vast difference between that and the focus we adults try to apply. When a child touches something, he doesn’t begin to analyze and understand why his finger actually feels. He’s relating to the experience in a much deeper, non-technical way. Until adults learn to de-focus the language and focus on the happening, the language will always be center stage – and this is why adults are poor language learners. We’ve lost the theme and it’s become a boring, classroom subject or in any event, a whole lot of work. Why we’ve turned something we do perfectly into a classroom subject that must be studied is really beyond me. There are so many good things to learn in that format – language isn’t one of them.Can a person produce words they don’t know without focusing on them directly? I don’t think so. The only way I understand this can happen is when we’ve listened/experienced enough to form the basic understanding of things at an subconscious level.

Post new comment

The content of this field is kept private and will not be shown publicly.
This question is for testing whether you are a human visitor and to prevent automated spam submissions.