Crazy idea: Using the Koch-Farnsworth method for ear training from day one

General discussion about learning languages
User avatar
RyanSmallwood
Orange Belt
Posts: 187
Joined: Sun Jul 19, 2015 12:15 pm
Languages: Native: English
x 779

Re: Crazy idea: Using the Koch-Farnsworth method for ear training from day one

Postby RyanSmallwood » Wed Jun 15, 2022 1:46 pm

I've mentioned this a bunch of times on here already, but this is another reason I'm a big fan of using programs like audacity to remove pauses from FSI courses and use them for input/listening. FSI courses start at natural speed from the beginning, and slowly build up each new word for introducing it in a dialog, have phonology drills to help you notice the new sounds, and hearing the grammar exercises in rapid succession is another way to hear lots of simple sentences with slight variations and build automaticity.

It of course has the drawbacks always worth mentioning of being a bit old and not so interesting, but IMO its one of the best ways to start off a solid listening foundation and an intuitive understanding of grammatical structures. The lessons are usually a more manageable duration when you remove the pauses too, so its not too difficult to study or review at least 1 lesson a day.
6 x

User avatar
Le Baron
Black Belt - 3rd Dan
Posts: 3513
Joined: Mon Jan 18, 2021 5:14 pm
Location: Koude kikkerland
Languages: English (N), fr, nl, de, eo, Sranantongo,
Maintaining: es, swahili.
Language Log: https://forum.language-learners.org/vie ... 15&t=18796
x 9392

Re: Crazy idea: Using the Koch-Farnsworth method for ear training from day one

Postby Le Baron » Wed Jun 15, 2022 2:38 pm

The chief problem is really that training for listening is actually different than just listening. In the video that BeaP posted (which I'm just using for reference) the tutor shows you how to manage your way through one thing. This is great, but having things pointed out one time doesn't mean you'll recognise something else next time; it's not ideally transferable in that way. And every language learner must know this. It's why even when you've unravelled something and understood it, you just run into more rapid speech and it's incomprehensible all over again. This was evident in the other thread where someone posted a snippet of speech and it seemed like half-gibberish until it was pointed out (by zenmonkey). Then all was 'clear'. There are thousands of these things and they are often random or even the same thing done differently. So you need this guidance multiple times. Plus long-term exposure. That above all.

'Successfully' understanding some audio book because you listened to it 100 times is rather obvious. You've familiarised yourself with a certain thing. This is like knowing your way around an area because you walk there all the time. It doesn't mean you'll know your way around all other areas just because you mastered that one! It means you can do basic transferable skills like walking, looking, recognising cul-de-sacs etc. The analogues are everywhere: training for a parachute jump? Not the same as a parachute jump. Training for fires as a fireman using controlled fires? Not the same as a real fire. And every fire has its peculiarities.

Unfortunately training for things is all we can do and it's worth doing well. Personally I don't believe in magic 'short cuts' in language learning (and there's no shortage of them on offer, plus 'I've made an app for...' and 'I've worked out mathematically that..'). Some short cuts yes, but reducing foreign speech recognition from months to 'mere hours'? I don't think so.
7 x

Cainntear
Black Belt - 3rd Dan
Posts: 3469
Joined: Thu Jul 30, 2015 11:04 am
Location: Scotland
Languages: English(N)
Advanced: French,Spanish, Scottish Gaelic
Intermediate: Italian, Catalan, Corsican
Basic: Welsh
Dabbling: Polish, Russian etc
x 8666
Contact:

Re: Crazy idea: Using the Koch-Farnsworth method for ear training from day one

Postby Cainntear » Wed Jun 15, 2022 3:28 pm

Here's the problem: in Morse, you're dealing with a very simple code.
If we describe it in terms of language, there are only two "phonemes", and the only distinction between them is "vowel length". It's hard to miss the distinction between them.

With a spoken language, though, you're going to be hearing all sorts of phonemic variations at once, and with the limited input you are receiving there is no actual impetus to notice these. There are very few situations where a beginner is going to encounter true minimal pairs, and even there, the context will almost always disambiguate them anyway.

Koch's rationale was to prevent the beginner learning strategies that would hamper them later in their development, but here the learner is encouraged to develop a strategy that leads to understanding now with difficulties later, so Koch's technique does not serve the same end here.
5 x

Cainntear
Black Belt - 3rd Dan
Posts: 3469
Joined: Thu Jul 30, 2015 11:04 am
Location: Scotland
Languages: English(N)
Advanced: French,Spanish, Scottish Gaelic
Intermediate: Italian, Catalan, Corsican
Basic: Welsh
Dabbling: Polish, Russian etc
x 8666
Contact:

Re: Crazy idea: Using the Koch-Farnsworth method for ear training from day one

Postby Cainntear » Wed Jun 15, 2022 4:28 pm

Dragon27 wrote:But the analogy here is not with the kind of slower speech that the nature documentary narrator employs in order to not overwhelm the listeners with info, the "slow code" speed is like a deliberately drawn-out speaking speed aimed specifically at the beginners in the audio-materials for the language learners. It should be natural, not necessarily fast, but not so slow that the learner will be inclined to cheat (i.e. translate in their head, instead of developing a direct understanding of spoken phrases).

Why should it be natural? There is an argument that language input for beginners should exaggerate the things that are important in the target language, to force us to notice them.

The problem is that in practice, a great many teachers are going to do the opposite, and make the language easier for the learner by making the pronunciation more L1-like.

(I've known more than a few TEFL teachers who work or have worked in Romance-speaking countries who switch to syllable-timed speech when speaking English to learners, for example.)
3 x

Online
Dragon27
Blue Belt
Posts: 616
Joined: Tue Aug 25, 2015 6:40 am
Languages: Russian (N)
English - best foreign language
Polish, Spanish - passive advanced
Tatar, German, French, Greek - studying
x 1375

Re: Crazy idea: Using the Koch-Farnsworth method for ear training from day one

Postby Dragon27 » Wed Jun 15, 2022 4:42 pm

Cainntear wrote:Here's the problem: in Morse, you're dealing with a very simple code.
If we describe it in terms of language, there are only two "phonemes", and the only distinction between them is "vowel length". It's hard to miss the distinction between them.

Language comprehension is obviously much more complex, since we train our ability to recognize entire (phonetic) words and phrases, just going phoneme by phoneme is not sufficient. In Morse code, as far as I know, it's still, more or less, letter-by-letter recognition, given the speed (20-30 wpm is already fast for a great deal of people, and it's just one word every 2-3 seconds), although some people say that they do start recognizing "words" as a whole at very high levels and speed (but these are rare). So most people copying at decent speeds recognize letters, while people, who are only able to copy slow code, count individual dits and dahs. At least, that's my impression, I might be wrong.
1 x

User avatar
luke
Brown Belt
Posts: 1243
Joined: Fri Aug 07, 2015 9:09 pm
Languages: English (N). Spanish (intermediate), Esperanto (B1), French (intermediate but rusting)
Language Log: https://forum.language-learners.org/vie ... 15&t=16948
x 3631

Re: Crazy idea: Using the Koch-Farnsworth method for ear training from day one

Postby luke » Wed Jun 15, 2022 7:33 pm

Le Baron wrote:Personally I don't believe in magic 'short cuts' in language learning. Some short cuts yes, but reducing foreign speech recognition from months to 'mere hours'? I don't think so.

It's in the "some short cuts" that a good teacher or learning material can help out.

For listening, a concise description of common or typical vocal reductions can be helpful. E.G., I've seen videos for learning English with titles like: "going to" => "gonna", "want to" => "wanna". For the learner, understanding that in informal conversation these are common is helpful. Systematizing the learning of these sorts of vocal reduction would be one of the "short cuts".

Back to the "touch typing" analogy, which I think is a better fit with respect to language learning. The juj exercise shows its utility as the learner types words like justice, just, junction, jubilee, adjust, etc. Carrying that simple typing exercise further, the ded, frf, sws, etc exercises all begin to help the automation of typing words like "dedicate", "friend", "switch", etc. Beyond letter sequences that mimic the exercises, the student's fingers get trained on where important vowels are, which are used in almost every word.

Back to our friend, Mr. @SpanishInput, who brought us The importance of n-grams or why you still don't understand spoken..., a computation search for common n-grams, then curated to find and explain the typical "vocal reductions" could be a big help for listeners. Perhaps that could be an extension of the work he has already done.

Whether the student is encouraged or discouraged from using vocal reductions is not the key. Knowing that they are common and what they are is the key.
4 x
: 124 / 124 Cien años de soledad 20x
: 5479 / 5500 5500 pages - Reading
: 51 / 55 FSI Basic Spanish 3x
: 309 / 506 Camino a Macondo

User avatar
SpanishInput
Yellow Belt
Posts: 97
Joined: Sun Sep 26, 2021 3:11 pm
Location: Ecuador
Languages: Spanish (N), English (C2), Mandarin (HSK 5)
x 469

Re: Crazy idea: Using the Koch-Farnsworth method for ear training from day one

Postby SpanishInput » Wed Jun 15, 2022 10:27 pm

Hi! Thanks for all the replies!

So this idea was already brought up years ago? Wow, it seems there's nothing new under the sun. Or maybe I read it here years ago, forgot about it, and now I'm fooling myself thinking this is something new.

So, let me give a bit more clarification of what I have in mind: The teacher/tutor/exchange partner could introduce just a very limited number of words/groups of words per session, and come up with a text that repeats these ad nauseam. Gianfranco Conti calls this "flooding" in his book "Breaking the sound barrier". I'll quote him:

Explicit modelling. Aural input is carefully designed and used to model speaking and assist L2 acquisition. This means simplifying and ‘flooding’ language with numerous occurrences of the words, lexical and grammatical patterns we teach, in the same way as a caregiver adjusts their language to the needs of young children. Activities require thorough and extensive processing to promote recycling and noticing of language forms.

Conti, Gianfranco ; Smith, Steve. Breaking the Sound Barrier: Teaching Language Learners How to Listen (p. 22). Kindle Edition.


Professor Conti has advocated for the transcription technique in his blog, "The Language Gym".

Now, coming back to the speed part:

When. You. Speak. Word. By. Word. You're. Actually. Altering. The. Way. Words. Sound.

So, when I was talking about "natural speed" I'm not talking about talking fast and slurring everything. I'm talking about connected speech spoken clearly at a rate that is within the range of what's considered normal for a native speaker. For example, a documentary narrator. To prevent the learner from getting overwhelmed, the teacher/recording could leave, let's say, one full second between sentences. Maybe a bit more. Lemme attach an example of what I have in mind. In this example, the recording only introduces very, very common words and word combinations, together with some well-known proper nouns that are cognates in English. Of course, this would make a lot more sense with corresponding visuals:

1A-LLF_mixdown_Mono.mp3


And of coooourse this should go hand in hand with some basic teaching of the phonetic/phonological system of the new language. As Master Yoda said, "you must unlearn what you have learned". You need to break free from what you "know" about the /t/ and the /d/ phonemes, for example, because in L2 they're going to be totally different. That's why I'm throwing in the transcription technique. If the student, for example, needs to listen to each sentence, type it and after pressing Enter he can see every single mistake he made highlighted in red, he's going to notice gaps both in his general knowledge of L2 and also in his knowledge of L2's phonology. He'll start to slowly re-think what can be considered a /t/ and what can be considered a /d/, for example, and will re-calibrate his mental representation of these phonemes for L2. But he really needs near-real-time feedback for this.

This is an example from Anki. Anki lets you create "type in" flashcards, and automatically highlights your mistakes. I believe this highlighting is really necessary, because it can be really hard for learners to notice what they typed wrong. Reddit is full of posts of people wondering what they typed wrong in Duolingo.

anki003.PNG


In the example above, you're bringing the fact that "qué" has a silent U and an accent mark to the student's attention. If you come from English, you might just ignore Spanish accent marks and assume they're just some kind of optional decoration.
You do not have the required permissions to view the files attached to this post.
4 x

User avatar
Le Baron
Black Belt - 3rd Dan
Posts: 3513
Joined: Mon Jan 18, 2021 5:14 pm
Location: Koude kikkerland
Languages: English (N), fr, nl, de, eo, Sranantongo,
Maintaining: es, swahili.
Language Log: https://forum.language-learners.org/vie ... 15&t=18796
x 9392

Re: Crazy idea: Using the Koch-Farnsworth method for ear training from day one

Postby Le Baron » Wed Jun 15, 2022 10:33 pm

I just listened to the audio and it sounds like the audio from various language courses I have from the 70s and early 80s. Revolving around small subsets of repeated words and structures to create familiarisation. Isn't this the norm?
1 x

Cainntear
Black Belt - 3rd Dan
Posts: 3469
Joined: Thu Jul 30, 2015 11:04 am
Location: Scotland
Languages: English(N)
Advanced: French,Spanish, Scottish Gaelic
Intermediate: Italian, Catalan, Corsican
Basic: Welsh
Dabbling: Polish, Russian etc
x 8666
Contact:

Re: Crazy idea: Using the Koch-Farnsworth method for ear training from day one

Postby Cainntear » Thu Jun 16, 2022 6:27 am

SpanishInput wrote:So, let me give a bit more clarification of what I have in mind: The teacher/tutor/exchange partner could introduce just a very limited number of words/groups of words per session, and come up with a text that repeats these ad nauseam. Gianfranco Conti calls this "flooding" in his book "Breaking the sound barrier". I'll quote him:

Gianfranco Conti wrote:Explicit modelling. Aural input is carefully designed and used to model speaking and assist L2 acquisition. This means simplifying and ‘flooding’ language with numerous occurrences of the words, lexical and grammatical patterns we teach, in the same way as a caregiver adjusts their language to the needs of young children. Activities require thorough and extensive processing to promote recycling and noticing of language forms.

Conti, Gianfranco ; Smith, Steve. Breaking the Sound Barrier: Teaching Language Learners How to Listen (p. 22). Kindle Edition.

But notice the lack of mention of phonemes there.

Whole word recognition does not necessarily require noticing the characteristics of phonemes.

And of coooourse this should go hand in hand with some basic teaching of the phonetic/phonological system of the new language. As Master Yoda said, "you must unlearn what you have learned". You need to break free from what you "know" about the /t/ and the /d/ phonemes, for example, because in L2 they're going to be totally different. That's why I'm throwing in the transcription technique. If the student, for example, needs to listen to each sentence, type it and after pressing Enter he can see every single mistake he made highlighted in red, he's going to notice gaps both in his general knowledge of L2 and also in his knowledge of L2's phonology. He'll start to slowly re-think what can be considered a /t/ and what can be considered a /d/, for example, and will re-calibrate his mental representation of these phonemes for L2. But he really needs near-real-time feedback for this.

This will help to an extent, inasmuch as the learner will be more likely to start to notice differences in aspiration and voicing, but it will never force them to notice the difference between dental and alveolar articulation, for example, because that's a distinction between the two languages and there is nothing within Spanish that requires the distinction too be noticed. Failure to notice doesn't simply make the learner's accent worse, but it obscures the relationship between the fully articulated and lenited allophones.
1 x

User avatar
zenmonkey
Black Belt - 2nd Dan
Posts: 2528
Joined: Sun Jul 26, 2015 7:21 pm
Location: California, Germany and France
Languages: Spanish, English, French trilingual - German (B2/C1) on/off study: Persian, Hebrew, Tibetan, Setswana.
Some knowledge of Italian, Portuguese, Ladino, Yiddish ...
Want to tackle Tzotzil, Nahuatl
Language Log: viewtopic.php?f=15&t=859
x 7030
Contact:

Re: Crazy idea: Using the Koch-Farnsworth method for ear training from day one

Postby zenmonkey » Thu Jun 16, 2022 10:36 am

Le Baron wrote:I just listened to the audio and it sounds like the audio from various language courses I have from the 70s and early 80s. Revolving around small subsets of repeated words and structures to create familiarisation. Isn't this the norm?


Exactly, thinking about this I was ... wait... aren't Pimsleur lessons organised like this?

bread.
ate bread.
He ate bread.
He ate bread and invited her out for drinks.
4 x
I am a leaf on the wind, watch how I soar


Return to “General Language Discussion”

Who is online

Users browsing this forum: No registered users and 2 guests