Technique to learn with sounds first, get the meaning second.

[Iversen]

While I definitely understand how many people can profit from starting with audio only and later adding the written part (with various ways to tackle this, as we're reading even in this one short thread), I simply don't believe in learning the sounds without their meaning. Listening to a bit of the language, as a beginner, should be followed immediately by the meaning. With the exception of learning to sing, perhaps, you don't necessarily need the exact translation for that (but it helps me).

Let's see it in graphic terms:there is a layer at bottom with 'raw' sounds (phones), a layer where these phones have been allocated to a standardized set of phonemes, a layer where you find words and phrases composed of phonemes, on top of that the 'getting the meaning' layer where you find the general meaning, and finally the stylistics and 'reading between the lines' layer, where you can draw conclusions about the meaning of it all without being restricted to the concrete message.

Maybe some persons can't see the point in learning sounds without lerning them in connection with specific words which have an meaning for them here and now. OK, then that forces them to learn a language through communicative acts (including one-way communication in the beginning if they believe in a silent period). In most cases that would make them follow a standard course with a teacher.

I don't advocate leaning sounds and postponing everything concerrning the meaning to later. I believe in finding ways to train the layers separately, which in the case of sounds and phonemes means listening to something without letting your desire to understand the meaning ruin the training. And in theory training sound discrimination would mean abolishing the wish to identify the official phonemes right away, but instead listening for raw sounds (and couple them to articulation as soon as possible). But this is not where I would start my listening exercises first. I would first read (!) a list of phonemes from a book so I knew what to listen for and THEN proceed to my trusty bloodhound listening, where you care zilch about the meaning but try to parse the stream of sounds into words and phrases through the phonemes. To skip the ugly 'my name is A', 'it is nice weather today' stage in language learning you need to be able to listen and make sense of the structuration of a stream of speech from native speech as soon as possible, and this skill can be trained separately.

Just as learning vocabulary can. But maybe not for those who need some kind of emotional involvement in dealing with language phenomena.

As for the lowest layer, the phones, I find them most relevant once you already have a working knowledge of the phonemic layer, i.e. when you want to study not only what people say, but also how they say it. And then you will of course be able to refine your first primitive ideas about how the new language should sound and get your allophones sorted out in the process. Whether active articulation should be part of the listening from the outset is questionable. Some have written that it should, but I am more inclined to think that you can listen first and then try out to pronounce the sounds as soon as possible after. Anyway, working with the raw phones is the element that should permit you to getting a more realistic view on the pronunciation, and you'll probably also be confused by the fact that different speakers pronounce different things differently - which is one reason that I prefer getting a solid foundation at the phonemic layer first.

And vocabulary learning and grammatical analysis should of course be done concurrently, not months later.

[leosmith]

I would first read (!) a list of phonemes from a book so I knew what to listen for

How do you do that if you don't know the script?

[Iversen]

How do you do that if you don't know the script?

I would just transcribe a couple of pages using a table from a language guide or a grammar or whatever, and that would work for alphabetical systems (and to some extent also for abjads, i.e. systems where you don't write the wowels). That exercise would of course not be enough to crack Chinese writing or Sumerian Cuneiform, but so far I have not planned to learn Mandarin, nor Sumerian. And I have not added a totally new alphabet to my collection for years. Everything I have studied recently has been in variants of the Latin, Cyrillic or Greek alphabets. But I first learnt Cyrillic and Greek writing many years ago by doing transcriptions so I know that the method works.

Most of the primitive lists in small booklets will of course reflect the language of the book they are in, and therefore the pronunciation 'help' in such books often is more more a distraction than a help. But just to get an overview over the phoneme inventary they are OK, and then you should start listening to get the 'sound' of the language. Listening to short snippets while looking at the corresponding text is one way of discovering where your first impression from the lists was inadequate and then you can proceed to weed out the misunderstandings. But I have also sometimes used speech synthethizers - which may seem provocative to people who believe that you must get the pronunciation right from the start, but I can actually start reading quite complicated texts long before my pronunciation has become top notch so I'm not worried. And the synthesizers are getting better all the time.

EDIT: I just reread this message and came to think about the confused impression it might leave. One the one hand I recommend listening closely to speech to get the real sounds irrespective of how they relate to the phonemes, and on the other hand I write that you just have to read what phonemes there are in you target language and roughly how they sound, and then you can survive on that. The crux of the matter is the temporal order. To get your first bearings on a new language you need some kind of overview, like when you look at a town plan to find out where the main streets are and how you can get from where you are (A) to where you want to be (B). But once you are moving along in the streets you start looking at the traffic lights and signs and potholes and cars trying to run you down, etc etc. And then you'll also notice that the town plans aren't always precise..

[mentecuerpo]

Finally, for me pronunciation is never a separate facet of language learning, it’s there with every syllable you utter, every word you hear, every word you read, silently or aloud. Thus, never read, listen or speak a syllable of your TL without recognising it is always intricately tied to a sound/phoneme, and as such train yourself to read with the correct phonemes in your head, speak focusing always on your pronunciation and listen attentively to always compare to yourself (do you sound like that? why not?) and the standard form of the language (is it a regional accent?). Good luck!

Thank you. I think I get what you are recommending, I will try to be mindful with my pronunciation. Even on my fossilized English, I have more opportunities to correct it when I say it wrong, many times I receive immediate feedback from natives, who correct me on the spot because they want to get the meaning right. I am thinking on creating a log of my mistakes when corrected and work on those words.

With my new language, I am learning it as a personal scientific project, knowing what I know now about learning a new language, I think I can do a better job being more mindful about the process, working to get the phonemes right from the beginning.

My L2 English, I need to learn how to correct fossilized bad language habits, I learnt it very empirically, doing a lot of reading without getting the right pronunciation first, then the wrong sounds have been internalized in my brain language sound bank. The problem is not even a strong accent; it is the wrong sound! If I have the right sounds and and pronounce it with my Spanish accent, the quality of my English would improve so much.

There is a reason why phonetics is an important element of any given language, especially for us interest in oral communication. These principles have to be incorporated from the onset when learning a language.

[leosmith]

To clarify, the method I described in my earlier post was mostly designed to avoid pronunciation issues that happen if you start using the written language before being able to recognize or pronounce it. I believe that if you’re an Indo-European native speaker learning Indo-European languages you are less likely to notice the negative impact of activities such as reading before listening on your pronunciation. The comments I made regarding delaying meaning were sort of spur of the moment. It just so happens that mastering pronunciation in the beginning the way that I do is facilitated by less effort spent on meaning. I never thought of it exactly in that way before, but it’s true.

Regarding learning the alphabet along with basic pronunciation before you do anything else in the language, this normally doesn’t take very long. A dozen hours? A couple dozen? Of course this depends on the language. But after you do it, you have removed a major barrier to your pronunciation; the possibility of fossilizing errors is greatly reduced. So it’s well worth the time, or the discomfort you have in feeling a bit disconnected with the language. I really don’t think learning the alphabet, via native audio and native script, constitutes learning out of context, but I respect other’s opinions.

I think there is a myth that’s been circulating around language learning for the past couple decades that pronunciation isn’t important. I know it’s rarely said so directly, but it starts out with telling people not to try to achieve native-like pronunciation and ends up with people struggling to be understood. What I suggest in my first post is a very low price to pay for avoiding this situation imo, and the bigger the difference between your L1 and L2 pronunciation the more effective the solution.

[mentecuerpo]

...As us language learners know, a word's written form can impact our perception of how it sounds. Our perception of words is affected by our prejudice that derives from how that word is written. It causes us to hear phonemes that aren't present, and causes interference with how we perceive the sounds of a language. This is a result of our brains trying to make what we hear match up with what we think we should be hearing...

[ryanheise]

"The written form of any language does no more than represent these sounds visually."
-True about some languages, but by far not all of them. Mandarin and Japanese are the most typical examples of writing representing the meanings.

I agree with your comment, although a slightly tangential topic interests me here which is the extent to which learning to read first can interfere with a learner's ability to notice how native speakers actually pronounce individual words and connected words.

Can learning the written form first interfere with the ability to notice correct pronunciation even in the case of Japanese, where as you've correctly pointed out, some parts of the writing (i.e. Kanji) represent meanings, not sounds? I tend to think yes, because a learner who learns to read Kanji first will still go through a translation process where, through the aid of furigana (common reading aids that translate kanji into a more phonetic syllabary), the learner will construct from this a phonetic model of how to produce the sound, and here is where learners will make assumptions influenced by their L1. For example, the learner will translate a Kanji into a sequences of consonants, vowels, consonants, vowels, but may fail to notice the prosody when native speakers actually speak, cases when certain vowels are unvoiced, and a range of other modifications to sounds that occur only in the context of sentences.

Don’t worry that you don’t get the meaning, that’s the idea, you will match the meaning to the sound on part 2, once you have learned the sound. I don’t think it is necessary to learn meaning of the sound simultaneously, it is not practical.

The technique might be useful for some people, but totally frustrating to others. I cannot view the meaning as something secondary. And as I am now observing again on my bf, learning French, lots of people simply find in uncomfortable to learn something they cannot see the use of right away. So, learning the pronunciation without the meaning is unnecessarily frustrating and I still cannot see what advantage it brings. Just the sounds without meaning are not practical at all.

I can completely get your point that it could be frustrating for many learners, although I'll accept your challenge to think of some sort of advantage that this might bring

This is where my above tangent might swing back into relevance. If understanding the written alphabet or syllabary first is one thing that could interfere with your ability to notice correct pronunciation, can merely understanding the "meaning" also be another potential interference? I really only asked myself this question as a thought experiment, but I came up with the answer of "yes", perhaps it could, for similar reasons as above.

One example where I fell victim to knowing the meaning and letting that interfere with correct pronunciation is that Japanese intonation for questions is completely different from English. I remember a long time ago before I really knew any Japanese asking a friend how to say "How are you?". She said "O Genki desu ka?" (お元気ですか?). I tried to say it back "O Genki desu ka?", and I felt I had copied every syllable perfectly, but she immediately corrected me saying "NO! O Genki desu KA?" and used her hand to point out the shape of the intonation for the KA. Apparently in Japanese questions, the intonation at the end goes down and then up. I felt surprised that even though I had a musical background and thought I was a natural at noticing melodies, I couldn't hear the melody of this sentence due to my L1 interference.

I dare say that if I had no idea what the meaning of the sentence was first and then she had asked me to mimic the sounds, I would have got the melody/intonation right. And from there, if she had then revealed the meaning afterwards, I would have noticed on my own that that's how intonation for questions works in Japanese.

One other potential advantage of learning the pronunciation before the meaning has been advocated by Alexander Arguelles in his shadowing technique. He says that one of the reasons he likes to practice mimicking the pronunciation of sentences before he learns the meaning is that to hear a new word, to say that word many times, to then become curious about the meaning of that word, and then to "finally" find the answer as to what it means can be a great boost to your memory.

Finally, I want to mention the YouTuber called mimicmethod who advocates for learning to mimic pronunciation first without understanding the meaning, in order to simply get a "feel" for the sounds of the language, or what he calls the "flow" of the language. While this could potentially be a bland, boring way to start out learning a language, he seems to make it fun by mimicking music/singing.

I don't necessarily claim to be an advocate of any of these techniques, although I think there are at least some reasons why learning pronunciation before meaning could potentially be advantageous.

But at the end of the day, it is up to the individual what is enjoyable or motivating for them, and if such a technique is going to turn someone off learning a language, then they probably shouldn't use that technique despite any advantages it might have.

[mentecuerpo]

"The written form of any language does no more than represent these sounds visually."
-True about some languages, but by far not all of them. Mandarin and Japanese are the most typical examples of writing representing the meanings.

I agree with your comment, although a slightly tangential topic interests me here which is the extent to which learning to read first can interfere with a learner's ability to notice how native speakers actually pronounce individual words and connected words.

Can learning the written form first interfere with the ability to notice correct pronunciation even in the case of Japanese, where as you've correctly pointed out, some parts of the writing (i.e. Kanji) represent meanings, not sounds? I tend to think yes, because a learner who learns to read Kanji first will still go through a translation process where, through the aid of furigana (common reading aids that translate kanji into a more phonetic syllabary), the learner will construct from this a phonetic model of how to produce the sound, and here is where learners will make assumptions influenced by their L1. For example, the learner will translate a Kanji into a sequences of consonants, vowels, consonants, vowels, but may fail to notice the prosody when native speakers actually speak, cases when certain vowels are unvoiced, and a range of other modifications to sounds that occur only in the context of sentences.


I can completely get your point that it could be frustrating for many learners, although I'll accept your challenge to think of some sort of advantage that this might bring

This is where my above tangent might swing back into relevance. If understanding the written alphabet or syllabary first is one thing that could interfere with your ability to notice correct pronunciation, can merely understanding the "meaning" also be another potential interference? I really only asked myself this question as a thought experiment, but I came up with the answer of "yes", perhaps it could, for similar reasons as above.

One example where I fell victim to knowing the meaning and letting that interfere with correct pronunciation is that Japanese intonation for questions is completely different from English. I remember a long time ago before I really knew any Japanese asking a friend how to say "How are you?". She said "O Genki desu ka?" (お元気ですか?). I tried to say it back "O Genki desu ka?", and I felt I had copied every syllable perfectly, but she immediately corrected me saying "NO! O Genki desu KA?" and used her hand to point out the shape of the intonation for the KA. Apparently in Japanese questions, the intonation at the end goes down and then up. I felt surprised that even though I had a musical background and thought I was a natural at noticing melodies, I couldn't hear the melody of this sentence due to my L1 interference.

I dare say that if I had no idea what the meaning of the sentence was first and then she had asked me to mimic the sounds, I would have got the melody/intonation right. And from there, if she had then revealed the meaning afterwards, I would have noticed on my own that that's how intonation for questions works in Japanese.

One other potential advantage of learning the pronunciation before the meaning has been advocated by Alexander Arguelles in his shadowing technique. He says that one of the reasons he likes to practice mimicking the pronunciation of sentences before he learns the meaning is that to hear a new word, to say that word many times, to then become curious about the meaning of that word, and then to "finally" find the answer as to what it means can be a great boost to your memory.

Finally, I want to mention the YouTuber called mimicmethod who advocates for learning to mimic pronunciation first without understanding the meaning, in order to simply get a "feel" for the sounds of the language, or what he calls the "flow" of the language. While this could potentially be a bland, boring way to start out learning a language, he seems to make it fun by mimicking music/singing.

I don't necessarily claim to be an advocate of any of these techniques, although I think there are at least some reasons why learning pronunciation before meaning could potentially be advantageous.

But at the end of the day, it is up to the individual what is enjoyable or motivating for them, and if such a technique is going to turn someone off learning a language, then they probably shouldn't use that technique despite any advantages it might have.

Thank you for your contribution to this blog, ryanheise.

It is very educational your comments about the Japanese language (Kanji first will still go through a translation process where, through the aid of furigana) and the meaning attached to it. Very interesting from the linguistic point of view.

I have watched Alexander Auguelles videos but I did not pay much attention to what your pointed on his shadow technique.
Thank you for pointing out the youtuber mimic method, is this the same person who has built an online method, the mimic method?

[DaraghM]

I recently conducted a language learning experiment which ties in with some of the discussions here. Over the years, I’ve noticed my slowest progress seemed to be in acquiring a listening and speaking capability in a language, compared to understanding a written text. As I’d always learnt both together, it was hard to tell if it was the listening part that was slowing my progress. The first part of the experiment consisted of pure listening sessions to native speech, without any study of the written language, it’s grammar or vocabulary. The language I chose was Romanian, a Romance language with a case system. The source for my listening was Romanian internet radio. As I already know French, Spanish and some Italian, it wasn’t a completely blind experiment. What I discovered surprised me. Even with no studying of the text, I was starting to understand some words after a week, and some aspects of the grammar after a month. Proper nouns and some Romance and Russian cognates were my sign posts. After two months, I could cobble together some basic Romanian sentences, although with unknown grammar errors. It almost felt like text based study made no significant difference to my listening ability. The amount of listening per day wasn’t as significant as the number of days. I wonder if sleep consolidation is a key factor in listening.

The second part of the experiment was studying another language, but as text only, with no audio. I didn’t overlap the two parts. The language I chose for the second part was Latin, the original Romance language with a case system. Again the time period was for two months. My progress in Latin was extremely rapid. Without the listening component, I was making much faster progress than any previously studied language. I covered all six cases, in a number of declensions, and a large swathe of the verbs, with the major exceptions of the subjunctive, most passive constructions and gerund. The technique I used was the two way grammar translation method, which really focused my learning. I must admit the Latin experiment was much more enjoyable than the Romanian experiment, even though it was a lot more intense. The amount of time spent in Latin per day did make a difference, as did the number of days, but the amount of days to a lesser extent.

The whole experiment has made we wonder if two separate, but related, processes are involved in language learning. A slower listening acquisition, similar to Krashen’s theories, and a faster complex skills acquisition process for written texts, more closely aligned with DeKeyser.

[mentecuerpo]

I recently conducted a language learning experiment which ties in with some of the discussions here. Over the years, I’ve noticed my slowest progress seemed to be in acquiring a listening and speaking capability in a language, compared to understanding a written text. As I’d always learnt both together, it was hard to tell if it was the listening part that was slowing my progress. The first part of the experiment consisted of pure listening sessions to native speech, without any study of the written language, it’s grammar or vocabulary. The language I chose was Romanian, a Romance language with a case system. The source for my listening was Romanian internet radio. As I already know French, Spanish and some Italian, it wasn’t a completely blind experiment. What I discovered surprised me. Even with no studying of the text, I was starting to understand some words after a week, and some aspects of the grammar after a month. Proper nouns and some Romance and Russian cognates were my sign posts. After two months, I could cobble together some basic Romanian sentences, although with unknown grammar errors. It almost felt like text based study made no significant difference to my listening ability. The amount of listening per day wasn’t as significant as the number of days. I wonder if sleep consolidation is a key factor in listening.

The second part of the experiment was studying another language, but as text only, with no audio. I didn’t overlap the two parts. The language I chose for the second part was Latin, the original Romance language with a case system. Again the time period was for two months. My progress in Latin was extremely rapid. Without the listening component, I was making much faster progress than any previously studied language. I covered all six cases, in a number of declensions, and a large swathe of the verbs, with the major exceptions of the subjunctive, most passive constructions and gerund. The technique I used was the two way grammar translation method, which really focused my learning. I must admit the Latin experiment was much more enjoyable than the Romanian experiment, even though it was a lot more intense. The amount of time spent in Latin per day did make a difference, as did the number of days, but the amount of days to a lesser extent.

The whole experiment has made we wonder if two separate, but related, processes are involved in language learning. A slower listening acquisition, similar to Krashen’s theories, and a faster complex skills acquisition process for written texts, more closely aligned with DeKeyser.

DaraghM, great experiment.

On your studies of the Romanian language by a strong emphasis of “audio only.” How much did you gain from this in your listening comprehension skill? Did it help you with your pronunciation?

It looks to me that you were not doing pronunciation drills, but it is possible that it helped you too with your pronunciation skills.
We need to experiment and find what works best for each individual learner and depending on what are the goals.
If your goal is oral communication then listening comprehension and reproducing accurately the phonemes of the language will be a necessary component of your time dedicated to learning your target language.

If you are interested in learning how to read and write the language, then the reverse translation, with transcript will work for sure. Reading books, grammars, standalone language method books, anything with text is enough.
If the language learner wants to learn to understand, speak, read and write, then a combination of the above maybe effective. I prefer audio with text as the initial part of my studies.

The problem I see with your study is that one of the languages involve was Latin a dead language. It makes sense to learn Latin with text only. But you will not know for sure, what impact learning with text only made on your Latin pronunciation.

(I don’t even think the catholic priest in the Vatican can speak it right, and if they do, who is to judge if they are using the pronunciation correctly. I think Latin is the only language which your native accent can influence it freely and you then can call it based on your mother tongue: Spanish/Latin, American English/Latin, German/Latin. There is no standard to compare Latin. I am sure that if a Latino priest gives mass in Latin, it will be with Spanish accent, no doubt. And that will be ok because it is Latin.)