Hacker News Reader

There is extensive research going back many decades that strongly implicates the role of perception in the difficulty that second language learners have with pronunciation.

Basically, learners have a tendency to hear the second language through the phonological 'filter' of the first language. They assign incoming speech sounds to categories based on the cues that are used in their first language rather than the second language, so they actually perceive the sounds in a different way from native speakers. This is a concept known as categorical perception. For example, Japanese learners of English tend to perceive English /l/ and /r/ as the same sound and cannot discriminate them reliably in perception experiments. Therefore, we can speculate that they have trouble producing the sounds differentially because they don't actually hear a difference between them.

So, the first step in overcoming pronunciation issues is not necessarily to teach tongue positions, but to train perception.

A similar problem is when the other language has a sound that is very similar, but not exactly the same as in your language. For example, if the other language has the same set of wovels, except some of them sound slightly different. As a foreign learner, you are likely to pronounce (and probably also hear) it as exactly the same as in your native language.

That is, just like it is difficult to distinguish between "their X" and "their Y", it is also difficult to distinguish between "your X" and "their X". Maybe even more difficult, because no one tells you that the difference exists.

> For example, Japanese learners of English tend to perceive English /l/ and /r/ as the same sound and cannot discriminate them reliably in perception experiments.

I wonder what would happen if you took a group of Japanese people, taught them how to make the /l/ and /r/ sounds, by telling them how to place the toungue etc., and them told one of them to choose and make one of these sounds, and the other one to guess which one it was.

I mean, just like it is easier for one non-native English speaker to understand another non-native English speaker, because they speak more slowly and carefully, maybe it would be easier for one Japanese person to understand another Japanese person's pronounciation of /l/ and /r/, because it would be more exaggerated.

It’s also entirely possible to never figure out the proper tongue positions when you are learning your native language. Instead of an “accent” you then have a “speech impediment”.

I grew up with a couple, I had some special instruction in high school to correct the fact that I said my R’s with the same mouth positions used for W’s; I also had a persistent issue with saying “tr” as a mushy kind of “ch” on one side of my mouth that I only finally corrected in, like, my mid-forties.

And yeah, once I looked up how to properly pronounce that “tr” it didn’t take too long to fix it.

Yes, that's exactly what I was thinking of when I said that certain sounds are probably harder to get right than others - R and S/TH being the two most common I believe. I'm not sure if this is the same thing as not recognizing how to pronounce a foreign language, or if it is more related to other speech impediments, such as stuttering.

I really don't think that's it to be honest. I'm talking about the difference between an expert level speaker of a second language and a native speaker of that language.

Getting basic phonology right happens way before you get to that problem.

Do you know an example of someone who has learned a second language and sounds indistinguishable from a native speaker to a native speaker? I don't, but would be curious to hear about counterexamples. I suspect maybe people who speak professionally (e.g., actors as GP suggested) might be candidates, but I don't know any...

You are right, I probably over-sold it a bit. Yes, becoming indistinguishable from a native speaker is a pretty high bar, though having a fairly non-descript accent I think is achievable.

For cases of non-native speakers mastering an accent, one anecdote I heard of was Hugh Laurie's audition tape for the role of Dr House. The story goes that the producers of the show considered it extremely important to have the character be played such that he was quintessentially American, and were very happy with the audition tape from Hugh Laurie, who convinced them on this front - only to meet him later and be shocked at his native RP accent.

Sure, it's somewhat rare but not that uncommon.

The Swedish researchers Kenneth Hyltenstam and Niclas Abrahamsson have a series of studies looking at immigrants in Sweden who began learning Swedish as adults and who are judged to be native speakers by true native speakers who heard their speech.

In one study, they gave a very thorough battery of linguistic tests to these adult learners, and for some of them, the only test they failed to perform equal to native speakers was on their knowledge of proverbs and idioms.

The place you're most likely to find such near-native speakers is in immigrant communities.

Cool, that literature is pretty much exactly what I've been looking for.

Without having looked at the precise methodological details, it seems to broadly support my point, though:

"Our findings lead us to the conclusions that the rare nativelike adult learners sometimes observed would all turn out to be exceptionally talented language learners with an unusual ability to compensate for maturational effects and, consequently, that their nativelikeness per se does not constitute a reason to reject the critical period hypothesis."

https://www.researchgate.net/publication/232004616_The_robus...

(By critical period and maturational effect they refer to the hypothesis that age determines, or influences, whether you'll be able to attain a native-like level.)

Setting aside their interpretation of the results, which is a bit heatedly contested (and I believe they were focused more on syntax rather than speech), there are really two questions to ask that I think are relevant to the discussion:

1) Is it POSSIBLE for the speech of an adult language learner to be indistinguishable from a native speaker?

2) If (1) is true, are such instances common?

The reason (1) is important to ask is because a strict critical period hypothesis may rule this out by saying once the brain has matured beyond a certain point (age ~12, for example, or possibly younger for phonology), you can no longer acquire a language to the extent of a native speaker.

But there is plenty of evidence that (1) is true. In the Hyltenstam and Abrahamsson literature, they are able to find quite a few learners whose speech is indistinguishable from that of native speakers. They then run a battery of more strict tests (syntax/knowledge tests IIRC) against these learners that reveal slight differences.

However, (2) is certainly false. It's somewhat rare to find an adult learner who truly sounds like a native speaker. That, however, isn't evidence in favor of a strict critical period/maturational effects hypothesis, since there can be many reasons for failure.

My intuition is that there is much less support for a strict critical period hypothesis these days. Most people now talk about a 'sensitive period', which reflects that there are a broad range of factors that favor successful acquisition in children but that adults have decreased sensitivity to.

I know one guy who has a hobby/obsession to learn perfect English pronounciation. (English English, not American English.) Seems to me he did a great job. I am not a native speaker, so my opinion probably doesn't matter much here, but at least I can say he sounds quite different from other people who have English as a second language.

Knowing how much time and money he spent to achieve this goal, it seems to me like a complete waste of time and money. Apparently, his opinion differs. But his example makes me believe that the goal is attainable... but also explains why most people don't do it.