Useful info about CI methods?

[grayson]

I have seen numerous people, both people I know and people I know of, develop accents that get them confused for natives. If you're an English speaker just take a trip to the Netherlands and you'll practically be tripping over such people.

Anecdata: this is not my experience, after 22 years living in the Netherlands. Sure, some Dutch natives manage this, but the overwhelming majority do not. If you only visit Amsterdam, a lot of the English you hear will, in fact, be spoken by natives (and non-natives from places other than NL, but that's generally accented in different ways).

There certainly are people who manage it -- my husband (Dutch native) sounds perfectly American when he speaks English -- but my experience living here has been that they are not terribly thick on the ground.

[Iversen]

(...) it's important to remember that the four-way classification of isolating vs synthetic vs polysynthetic vs agglutinative is an oversimplification... and was invented by IE speakers anyway. Strictly speaking, what is the difference between an agglutinative and an isolating language? Agglutinative languages essentially consist of strings of distinct morphemes whose order determines grammatical function... but wait, isn't that true of isolating languages? In one sense, the biggest clear difference between an agglutinative and isolating language might be the writing (...)

After Cainntear's very interesting post on the previous page the discussion has somehow moved around the question of language features that might make it easier to pick up a language without really studying grammar. And then you can ask: which elements are relatively easy to pick up from CI? People generally tend to think that synthetic languages with large morphological systems are harder to learn than those with simpler systems - but if the morphology is simple then this will probably have led to more diffuse systems based on the use of special pointer words or idiomatic phrases (and I count the rules concerning choice of prepositions after verbs as idiomatic in English as idiomatic rather than syntactic because they basically are chaotic). Relying solely on more or less comprehensible input (or combos of input plus dictionaries and/or translations) without grammar studies may seem more tempting for such languages, but the total sum of what you have to learn will ultimately be the same.

So how do you learn a morphological table? Well first look for patterns. In latin -ba- indicates imperfectum, and it is not harder to learn that than it is to learn the word "imperfectum". And then you add some endings, which mostly are the same as in other termpora except that there is a pesky -m in the first person singular where it's -o in the present, but by and large Latin endings are fairly regular. In the Perfect tense there are basically two competing systems, but it's still manageable. The real irregularities mostly arise when a verb has two (or more) different stems, and there can be all sorts of historical reasons for that, but you just have to live with the consequences. OK, look at that as learning two headwords from a dictionary instead of one.

So what makes Latin morphology harder to learn than Latin vocabulary in general? Well, maybe that those small clitics and fixes and god knows what are claimed by some eggheads not to have their own inbuilt meaning, but actually they have - the meaning of -ba - is "this is the imperfect tense". And thank the romans that they haven't made their endings so context depending as in some other languages.

The subdivision of languages in isolating vs synthetic vs polysynthetic vs agglutinative ones is basically a simplification that makes things more complicated if you trust it blindly. The reality is that languages whose 'blocks' don't change too much depending on the surroundings and haven't too many variations depending on choice of word root have a flatter learning curve than those that have complicated intercontingency rules. The -ba- of Latin is essentially NOT a synthetic language feature - it's more like one of those allegedly immutable thingies said to be characteristic of agglutinative languages. In contrast the verbal endings in Polish change a lot depending on the last consonant of the root, and that's really one of those things that can drive learners to madness, but at least there are dusty old grammarians that have systematized all that stuff and written it down in tables to make things easier for you. So be thankful that there are simple things to learn even in languages with lot of morphology, but don't refrain from looking in the "answers" just because some CI fanatic has said it wouldn't help you.

[Cainntear]

I'd like to correct my statement above—I was speaking rather loosely about the research. A more accurate statement would be "I've read studies show that, up through around age 12, some fraction of kids will naturally native native accents. Past that point, it's extremely rare, and usually involves either some kind of speech therapy, or rare natural talent."

It's weird to call "speech therapy" what most of us here would call "language learning". Phonology and phonetics is not some arcane secret you have to go to a specialist to learn about.

I cannot overstate how much I agree with this, and I have to thank both of you for making my own thinking much clearer to me.

I remember years ago watching a documentary about dyslexia, and one of the things being presented in it was people within education were looking to diagnose dyslexia early so as to be able to make interventions that would preempt emerging difficulties with initial literacy. However, the experts in childhood development were adamant that this was the wrong approach. Even if you can identify the things that make people predisposed to developing dyslexia, only a fraction of them will ever develop dyslexia.
The reason for this?
If you have the precursors, you are more prone to weaknesses in the teaching. Trying to develop interventions "for dyslexics" was basically ignoring the matter that if you genuinely teach everyone better, dyslexia is just less of a problem.

I don't know if I was actually aware of how important this was to my thinking about adult language learning until right now, but it really fits with my whole philosophy.

I absolutely believe that learner differences are in reality differences in tolerance for weaknesses in teaching, and teaching without exhibiting weaknesses that learner A can cope with and learner B can't is not sacrificing learner A for learner B's sake -- both should learn better. Learner A has to do less work.

It was only the talk about pathologising accent problems that made me see the link: in initial literacy we pathologise people who find it harder to learn to read (dyslexics) and if we send L2 learners with pronunciation difficulties to SLT, we're pathologising them.

In both cases, teaching the skill properly eliminates the problem at least partially.

Whether we teach pronunciation or not shouldn't depend on whether we can achieve a perfect accent or not -- we can reduce the problems, so we should.

[Le Baron]

I have seen numerous people, both people I know and people I know of, develop accents that get them confused for natives. If you're an English speaker just take a trip to the Netherlands and you'll practically be tripping over such people.

Anecdata: this is not my experience, after 22 years living in the Netherlands. Sure, some Dutch natives manage this, but the overwhelming majority do not. If you only visit Amsterdam, a lot of the English you hear will, in fact, be spoken by natives (and non-natives from places other than NL, but that's generally accented in different ways).

There certainly are people who manage it -- my husband (Dutch native) sounds perfectly American when he speaks English -- but my experience living here has been that they are not terribly thick on the ground.

Most certainly. The people who claim this have probably only ever visited the Netherlands and never had an in-depth conversation. Or just repeat a tired old myth.

[Le Baron]

I think it's entirely justified for EMK to call accent training a form of speech therapy and not language learning. Actors get this sort of training for roles, people who are generally better than other people at voice manipulation, and even they struggle to maintain it at times.

Language learning is only learning how to use the language, not sounding like a native speaker.

[Cainntear]

I think it's entirely justified for EMK to call accent training a form of speech therapy and not language learning. Actors get this sort of training for roles, people who are generally better than other people at voice manipulation, and even they struggle to maintain it at times.

Language learning is only learning how to use the language, not sounding like a native speaker.

OK, but isn't that a bit arbitrary?

Coodnt aye sae that langwidj lerning is ownlee lerning how to yooz the langwidj an not righting lyk a naetiv speekr?
Or could I say possibly that a language learning is learning only how of using the language, not and sentence forming like a speaker of native?
Heck, we could even sing that language painting is only painting how to squirt the language and not using every word that a native speaker would.

I mean... that's patently absurd, right? We might forgive a small number of spelling errors, grammar errors or word choice errors, and we would forgive a sentence that combines a small number of errors of all three types, but we would surely draw the line long before accepting any of those as acceptible sentences in "learner English".

I think the teaching profession has done everyone a massive disservice by deciding that pronunciation doesn't need any attention until and unless it is so atrociously bad that it very seriously hampers communication.

I think that's where the difference of viewpoints between you and emk on one side and me and kulaputra on the other side lies.

Maybe kulaputra's post muddied it up by effectively treating language pathologies as though they don't exist, but my view is that if we *only* deal with pronunciation once we have "pathologised" it.
If we actually taught pronunciation, the problem would be lessened.
So the "problem" that needs an SLT is that the learner has failed to learn something that the teacher hasn't taught -- is that not getting things a little bit arse-backwards...?

[kulaputra]

I think it's entirely justified for EMK to call accent training a form of speech therapy and not language learning. Actors get this sort of training for roles, people who are generally better than other people at voice manipulation, and even they struggle to maintain it at times.

Language learning is only learning how to use the language, not sounding like a native speaker.

Do you disagree that pronunciation is part of language learning? If it is, what metric do you think should be used if not verisimilitude to native pronunciation?

Your statement is a bit like claiming "marathon training is not about running a sub 3 hour marathon". Technically true, and very few people will ever run that fast, but it's not a very useful statement. The point of marathon training is, however, to reduce your marathon time, and if that happens to get you to a sub 3 (or 2) hour marathon, who would disagree that would be an ideal result?

(Now I happen to think getting a near native or better accent is much less hard than running that fast, and we can disagree about that, but that's a secondary question.)

[Le Baron]

Do you disagree that pronunciation is part of language learning? If it is, what metric do you think should be used if not verisimilitude to native pronunciation?

No I don't disagree about it being a part of language learning, but that since it is a massive obstacle to actually sound like a native, that shouldn't necessarily be the goal. Verisimilitude is something different.

Your statement is a bit like claiming "marathon training is not about running a sub 3 hour marathon". Technically true, and very few people will ever run that fast, but it's not a very useful statement. The point of marathon training is, however, to reduce your marathon time, and if that happens to get you to a sub 3 (or 2) hour marathon, who would disagree that would be an ideal result?

Even though I wouldn't claim that and didn't, which means my statement isn't like that at all. I would say marathon training is probably good for anyone, but that referring to it as anything other than marathon training - such e.g. I am a marathon runner - is false.

(Now I happen to think getting a near native or better accent is much less hard than running that fast, and we can disagree about that, but that's a secondary question.)

I happen to think that getting a native accent is generally not possible. What people get is somewhat like how impressionists gather together certain elements, but that when you put them next to the person they are impersonating they sound a lot less like them. And yet I still think it's worth having a go at this.

[Cainntear]

I have seen numerous people, both people I know and people I know of, develop accents that get them confused for natives. If you're an English speaker just take a trip to the Netherlands and you'll practically be tripping over such people.

Anecdata: this is not my experience, after 22 years living in the Netherlands. Sure, some Dutch natives manage this, but the overwhelming majority do not. If you only visit Amsterdam, a lot of the English you hear will, in fact, be spoken by natives (and non-natives from places other than NL, but that's generally accented in different ways).

There certainly are people who manage it -- my husband (Dutch native) sounds perfectly American when he speaks English -- but my experience living here has been that they are not terribly thick on the ground.

Most certainly. The people who claim this have probably only ever visited the Netherlands and never had an in-depth conversation. Or just repeat a tired old myth.

Or perhaps even a modern exaggerating twist on one.

It is received wisdom that Dutch people speak good English -- and it's true; it's observable fact.

People's views of what means to be "good" at a language include accent, so even though they have probably observed dozens of Dutch people who are readily identifiable as Dutch from their accents while speaking English, they're erroneously (and unconsciously!) taking "Dutch people have native-like accents" as the logical consequence of the truthful observation that "Dutch people are really good at English".

But it could be something else. Perhaps they're unconsciously confusing the fact that they have fewer problems understanding Dutch people than (for example) Spanish people with the idea that they somehow sound more native-like...?

Cos the reason that Dutch people are easier to understand isn't that they lose their own accent -- maybe it's just that their non-native accent is somehow more "compatible" with the English accent, so it's easier for us to tolerate the differences...?

For example, one of the major categories of difference between languages is the PTK vs BDG distinction. In English, the primary difference is in voicing, and aspiration is only a secondary difference. Dutch has no aspiration, and the only distinction is voicing. We can normally hear a Dutch T in English and understand that it's a T, not mistaking it for a D. In both languages it's a "voiceless alveolar stop" -- the secondary thing of aspiration is so inconsequential that it isn't part of the name it's known by. However, the Spanish one is a "voiceless dental stop", and that change of character seems (to me -- subjective opinion) to be more significant, because non-specialist natives (i.e. non-linguists and non-language-teachers) appear to have more difficulties with that. There's a bit of ambiguity because we have dental phonemes in English too (/ð/ and /θ/), so a Spanish /t/ has features that remind us of both /t/ and /θ/ in our own phoneme map, and we have to look for more clues from the context to determine what the speaker's intent was.

[Cainntear]

Do you disagree that pronunciation is part of language learning? If it is, what metric do you think should be used if not verisimilitude to native pronunciation?

No I don't disagree about it being a part of language learning, but that since it is a massive obstacle to actually sound like a native, that shouldn't necessarily be the goal. Verisimilitude is something different.

OK, but that doesn't mean that sounding good shouldn't be a goal.

Your statement is a bit like claiming "marathon training is not about running a sub 3 hour marathon". Technically true, and very few people will ever run that fast, but it's not a very useful statement. The point of marathon training is, however, to reduce your marathon time, and if that happens to get you to a sub 3 (or 2) hour marathon, who would disagree that would be an ideal result?

Even though I wouldn't claim that and didn't, which means my statement isn't like that at all. I would say marathon training is probably good for anyone, but that referring to it as anything other than marathon training - such e.g. I am a marathon runner - is false.

Hmmm.. I kind of think he's right, though.

Consider this...
When we say "learning a native-like accent is impossible" is almost always used as a justification for not putting more than a very superficial bit of effort into teaching and learning pronunciation.

The analogue to that would be to say that "running a sub 3 marathon is a rare skill and most of you will not achieve it" (which is true) and using it as a justification for not actually teaching techniques about running speed.

Learning a native-like accent is something the majority of people don't achieve... so don't try to pronounce well.
Running a sub-3 hour marathon is something the majority of people don't achieve... so don't try to run fast.