Is pronunciation a separate skill?

[tiia]

I don't know what you do with axes, but I chop trees with them.

Yeah, I should definitely cover my whole education up to the end of high school in all my foreign languages too. I meant axis of course. Axe is French, it seems. No clue about Spanish or German now.

German: Achse.

[tommus]

I don't know what you do with axes, but I chop trees with them.

You can draw a 2D graph with axes. Axes is the plural of axis.

[Jiwon]

Correct, but the problem is that this isn't what is meant by "listening" in the four skills model. What you're talking about is what's usually termed "sound discrimination" or "phoneme discrimination". (Sorry for using jargon again, but it's important here.) To put it more simply, "hearing" is only part of "listening".

Yes, I was quite sure that the skill I was referring to was different from what listening comprehension, and hence the reason why I wrote "listening" and "hearing" with quotation marks. I do not major in linguistics, it is good to learn that phoneme discrimination is the correct terminology for what I described.

I think sound discrimination, rather than pronunciation, is underrepresented and undereducated in language pedagogy. Sure we do dictations, but that has more to do with orthography. You don't need to distinguish between "toast", "tost" and "tust" to write down "toast" because from your knowledge of vocabulary and orthography, you know the other two are very unlikely.

[Dragon27]

You can draw a 2D graph with axes. Axes is the plural of axis.

It is also interesting that "axes" and "axes" are pronounced differently.

[tarvos]

I don't know what you do with axes, but I chop trees with them.

You can draw a 2D graph with axes. Axes is the plural of axis.

I am aware, but Cavesa wrote axe, and I found that funny :p

[Cainntear]

Correct, but the problem is that this isn't what is meant by "listening" in the four skills model. What you're talking about is what's usually termed "sound discrimination" or "phoneme discrimination". (Sorry for using jargon again, but it's important here.) To put it more simply, "hearing" is only part of "listening".

Yes, I was quite sure that the skill I was referring to was different from what listening comprehension, and hence the reason why I wrote "listening" and "hearing" with quotation marks. I do not major in linguistics, it is good to learn that phoneme discrimination is the correct terminology for what I described.

I think sound discrimination, rather than pronunciation, is underrepresented and undereducated in language pedagogy. Sure we do dictations, but that has more to do with orthography. You don't need to distinguish between "toast", "tost" and "tust" to write down "toast" because from your knowledge of vocabulary and orthography, you know the other two are very unlikely.

Ah, sorry. I see what you're getting at now.

Hearing phonemes and pronouncing them is a bit of a chicken-and-egg situation. What is agreed on is that both rely on the same internal "phoneme map", and what is argued about is how to construct that phoneme map. Some say we can only build the phoneme map through exposure to the language, others say we can construct it in learning to speak.

I don't believe we can absorb it through exposure, due to the nature of what phonemes are. A phoneme is the "smallest unit of meaningful sound" and may exist in multiple forms (called allophones). For example, I imagine you had great difficulty learning the distinction between /l/ and /r/ in European languages, because in Korean, they're both allophones of a single phoneme -- your brain was taught from birth to consider them to mean the same thing. This means that you don't actually perceive any difference -- and how can you learn if your brain is ignoring the important data?

But you weren't talking about exposure -- you were talking about directed study of discrimination between phonemes. In theory, that sounds great, but I really don't know how you would go about it. You've already highlighted the problem with dictation, but that's part of a wider problem: language has lots of redundant information, so the context will tell you what you need to know, so it's very hard to force anyone to discriminate between phonemes. Sticking with the L/R phonemes, the word "brilliant" is clear and unambiguous even if you can't hear the phonemes. Even words with minimal pairs aren't usually a problem because of the rest of the sentence. In "I don't know the word", you won't mix up "word" with "world" for example.

So I don't see how we really can teach or learn much about phonemes from listening to start off with.

My personal belief is that the only way to teach the brain that a new phoneme exists is through pronunciation -- if you make the mouth form roughly the right shape every time, then your brain has to remember that the phoneme (or allophone) actually exists. I believe that once that's done, the brain will naturally begin to discriminate it in what you hear, because it then knows what to look for. It knows it has meaning, so it can't fail to notice it.

That's not to say explicit focus on sound discrimination isn't helpful, but I believe it needs much less focus than pronunciation.

[Jiwon]

Ah, sorry. I see what you're getting at now.

Hearing phonemes and pronouncing them is a bit of a chicken-and-egg situation. What is agreed on is that both rely on the same internal "phoneme map", and what is argued about is how to construct that phoneme map. Some say we can only build the phoneme map through exposure to the language, others say we can construct it in learning to speak.

I don't believe we can absorb it through exposure, due to the nature of what phonemes are. A phoneme is the "smallest unit of meaningful sound" and may exist in multiple forms (called allophones). For example, I imagine you had great difficulty learning the distinction between /l/ and /r/ in European languages, because in Korean, they're both allophones of a single phoneme -- your brain was taught from birth to consider them to mean the same thing. This means that you don't actually perceive any difference -- and how can you learn if your brain is ignoring the important data?

But you weren't talking about exposure -- you were talking about directed study of discrimination between phonemes. In theory, that sounds great, but I really don't know how you would go about it. You've already highlighted the problem with dictation, but that's part of a wider problem: language has lots of redundant information, so the context will tell you what you need to know, so it's very hard to force anyone to discriminate between phonemes. Sticking with the L/R phonemes, the word "brilliant" is clear and unambiguous even if you can't hear the phonemes. Even words with minimal pairs aren't usually a problem because of the rest of the sentence. In "I don't know the word", you won't mix up "word" with "world" for example.

So I don't see how we really can teach or learn much about phonemes from listening to start off with.

My personal belief is that the only way to teach the brain that a new phoneme exists is through pronunciation -- if you make the mouth form roughly the right shape every time, then your brain has to remember that the phoneme (or allophone) actually exists. I believe that once that's done, the brain will naturally begin to discriminate it in what you hear, because it then knows what to look for. It knows it has meaning, so it can't fail to notice it.

That's not to say explicit focus on sound discrimination isn't helpful, but I believe it needs much less focus than pronunciation.

To summarize our argument so far, we seem to both agree on the fact that building a phoneme map (another linguistic term I am learning from you today. Fantastic!) is an extremely important step in language acquisition. We only disagree on which is better - training with phoneme discrimination or actual production/pronunciation.

I think it depends on what are the distinctive features for the phonemes in question. For some learning to pronounce is easier, while for others you have to learn to aurally distinguish in order to produce them correctly.

1. Consonants that are distinguished by place of articulation.
These are easier acquired by learning to pronounce them.
Quick examples I can think of are L/R, D/ð, Z/ð, S/θ
(And yes I didn't really know the difference between L and R when I first started learning English, but that was because none of the teachers emphasized any difference between the two. Funny because they were quick to pick on D/ð and S/θ.)

2. Consonants that are distinguished by manner of articulation.
These can also be acquired by learning to pronounce them, but I think for these you need to learn to distinguish them by sound as well.
This is because I think even in languages that use binary distinctions between voiced and unvoiced consonants, the distinction is not as binary and clear cut as they are thought to be. You will usually also use information on aspiration to determine whether the consonant is voiced or not. And unless you come from a linguistic background where you use 2x2 distinction of both voicing and aspiration, you have to learn to hear the difference.
I've written something about this in the thread about Korean pronunciation: http://forum.language-learners.org/viewtopic.php?f=14&t=4807

3. Vowels
I don't think vowel distinctions can be acquired by only learning how to pronounce them. Even if you learn that you lower your tongue for some vowels and raise it up for some others, you just have to learn to distinguish them by ear if you want to pronounce them correctly. This also goes for long/short vowels and nasal/non-nasal vowels.

Actually, I think the one effective way to train sound discrimination would be dictation or identification of minimal pairs without context. For example in English, try testing rich/reach/lich/leech or walk/work/wok/woke. For this same reason, I heard that our linguistic department has dictation in IPA as part of the final exam in their phonetics course.