Cainntear wrote:s_allard wrote:Finally, a concrete example that demonstrates how all this high-level theoretical discussion works. It seems here that our French speaker learning English will realize the voiced TH of "that" as "dat" and the unvoiced TH of "thin" as "tin". French speakers have a problem with the two sounds of the English TH because they do not exist in French. It is actually very interesting to note that European French speakers will approximate these sounds with "z" and "s" rather than "d" and "t". The latter sounds are more common in the English of speakers of Canadian French.
What we have to keep in mind is that these speakers actually hear the real "TH" sounds.
Do they?
When they attempt to reproduce these sounds,
Do they?
the approximate or rather incorrect variants come out. We see something similar when an English speaker learning French tries to say "la rue" and it sounds like "la roue".
Actually, most French speakers
can't hear the sounds except in a physical sense -- that is how phonemes work.
When a sound hits your ear, it isn't processed as raw data -- it is processed into phonemes. Your brain will map every received speech sound to a phoneme that it has in its acquired phoneme map. If you haven't acquired the phoneme, it's not part of your map, so you won't hear it.
Many people point at success in close listening as proof that this is not the case, but there is a marked difference between listening for something consciously and being able to recognise it unconsciously.
I don't understand this idea that fixing a pronunciation mistake will break something else.
Probably because you haven't made any effort to understand my argument.
My observation is that when a sound is corrected it will quickly and automatically spread to all the contexts
That is precisely the problem -- the correction of the pronunciation of "the sound" spreads to all contexts, but the core error is that the learner has a wrong notion of what "the sound" is. Even though they are consciously aware that they're talking about two sounds, their internal, unconscious model of the language still sees them as a single phoneme. A "correction" to that phoneme will affect all words using that phoneme.
Until the learner's internal model recognises that /t/ and /θ/ are different phonemes, the problem is uncorrectable. But you cannot just open up your brain, take all the words with that incorrect phoneme and split them into two. This is why it's an extremely difficult
especially since the spelling will be often helpful.
Unfortunately the spelling is of no immediate relevance to the learner's model of the language, which is why so many spelling mistakes reflect problems with the learner's pronunciation -- for example, learners of French who don't pronounce a difference between E, È and É generally end up forgetting which one goes where. This also holds for native speakers -- for example South American Spanish speakers commonly make mistakes with S and Z in writing because they make no distinction in speaking.
The student has to be first shown how to articulate the sound. Then a bunch of exercises to practice with. A tongue twister like "Is this the thing? - Yes, this is the thing." can be useful. Then away you go. Correcting pronunciation articulation takes time of course but that's the nature of the beast.
...if articulation was the problem.
But this brings us back to the core of the thread topic, because if we talk about
early speaking, articulation
is the problem and/or the solution.
People who speak early without proper guidance start to reinforce patterns using a faulty phoneme map, but if early instruction is given on articulation, it forces them to build up a serviceable phoneme map that can be refined later.
Some people claim that not speaking (i.e. listening only) will give them the opportunity to build up a phoneme map, but as I explained above, that's contrary to what we know about how the brain processes language. In fact, I'd say it's so counter-scientific that I'd describe it as a faith position.
What causes fossilised pronunciation problems is not speaking, or the lack of it -- it's a failure to teach/learn phonemics at an early stage.
The key is correction. These mistakes come from the interaction with the native language. If these pronunciations go uncorrected, the learner will probably keep making the same mistakes forever. And the student has to of course apply these corrections.
{Citation needed} There are uncountable numbers of papers out there showing the limited power of correction at a later stage.
But the main point here is that whether we call these fossilizations or bad habits, in my opinion, they are all the same thing. Bad habits are of course hard to change. If you've been speaking a certain way for 40 years, you're not going to change overnight. The way the term fossilization is used by some people, it seems that it's game over.
Again, I'll say
if you don't believe in fossilisation, say you don't believe in fossilisation. A fossilised error is an uncorrectable (or nearly uncorrectable) error -- if you do not believe there is such a thing, say it; don't just keep abusing the word.
It's a heck of a lot like saying that unicorns exist, and you've got one living down the road... just that it's a myth that unicorns have horns.
No, that's a horse, and unicorns don't exist.