Can a non-native speaker speak better than a native speaker?

[reineke]

There is Nothing Like Native Speech: A Comparison of Native and Very Advanced Non-Native Speech
Britt Erman & Margareta Lewis
Stockholm University

“I’ve been here for 8 ½ years, my English should be more fluent than this. Yes … sometimes I really stumble on the words...on the words”

1. Introduction
The above quote shows that finding words can be hard even for someone who has lived and worked in the L2 community for a considerable time. Vocabulary is an area of L2 acquisition that has received increasing attention in the last couple of decades. The present study is part of the research program “High-level proficiency in L2 use”. The program seeks to provide answers to questions pertaining to what characterizes the very advanced L2 user, and involves several language departments at a Swedish university. This study compares vocabulary of different frequencies in the oral production of two groups of speakers of English, one non-native Swedish group and one native English-speaking group as a control. The non-native Swedish group has lived and worked in the UK (London) for an average of 7.3 years. The main aim of the study is to establish the rate of high-frequency and low-frequency words in the spoken data of these two groups.

Some of the questions asked in the interview concerned the Swedish participants’ knowledge of languages and in particular their knowledge of English. Questions relating to English included for example the age at which they started learning English at school (in Sweden), whether they found speaking English difficult when they arrived in England, and the extent to which they used English also
at home when in the UK. It is worth noting that all the Swedish speakers used English at work, and most of them had English-speaking partners at the time of the recording. Reading through the transcribed interviews it became apparent that the interviewees had rather varied perceptions of their knowledge of English, as the extracts below show. However, the general impression from these extracts is that the interviewees believe that their English is quite good, some even to the extent that English has taken over at the expense of their mother tongue, Swedish.

• An easy ride when it comes to languages. Watched English TV a lot when little. Always speak English with my English partner.
• It’s much more natural to use English when speaking about music. I just can’t find the Swedish word…
• English was one of my worst subjects in Sweden. Wasn’t good at English at first (was very shy) but then just started speaking to people.
• I was fluent when arriving in England.
• Sometimes I feel when I go back, I become so conscious about my Swedish. And obviously I can still speak Swedish...it’s no problem,
but ...
• … sometimes I could have difficulty of swinging back into ...into fluent Swedish. I mean, when it comes to the more advanced Swedish, I think. Because, I think, my Swedish stopped developing when I was 22 and I came here. And...and here I don’t ...I don’t
associate that much with Swedes.
• English is ...what I realize with English is [after living in France]... it got a lot more words than French. French is, I think, if you’re
good in French, you use grammar to show that you are educated.

In this last extract there is a hint that English is perceived as having a large vocabulary.

The aim of the present study is not to establish whether the London Swedes’ own perceptions of their knowledge of English has a bearing
on the results but to find out how the two groups differ in their use of vocabulary in this task, more specifically across two main frequency
ranges to be explained below.

Results for the 1−2000 frequency range

Our hypothesis that the LS group would be nativelike on measurements pertaining to this task given its everyday character is only partly supported.While the LS group is nativelike on tokens per hundred words, they use significantly fewer types compared to the NS group. This result gives support for the inclusion of types in vocabulary studies. The highly significant difference in T/T ratio in the LS group compared to the NS group indicates that they recycle more words in this frequency range.

Results for the 2000+ range

Our hypothesis that the LS group would be nativelike also in the 2000+ frequency range in view of the everyday character of this task was not confirmed by the results.The number of tokens per 100 words is significantly lower compared to the NS group, and the difference between the groups in the number of types per 100 words is highly significant, the p-value being close to zero. One possible explanation for this result is that the NSs use more specific vocabulary compared to the NNSs, which is in line with the results from several earlier studies (Ovtcharov et al. 2006; Lindqvist 2010; Erman & Lewis 2011).

In light of the fact that the 1–2000 frequency range covers between 80% and 90% of all spoken texts, and to judge by the results of the present study, variation in this frequency range obviously is a nativelike feature, which distinguishes native and advanced non-native speakers. It is proposed in the present study that reaching a nativelike level in types in the first 2000 frequency range should be included in what is considered advanced vocabulary. In other words, showing variation among the 2000 most common words should be a skill worth aiming for also for advanced non-native speakers.

On the basis of the results presented in this study it seems reasonable to suggest that a contributing factor to divergences shown between the LS and NS groups is the difference in exposure, which has an effect also on types of high-frequency words as well as in the range of productive vocabulary at large.

5. Conclusion and discussion
As is clear from our results, our hypothesis, that the LS group living and working in the L2 country would be nativelike on both frequency ranges studied in view of the fact that the participants are invited to talk about themselves, was in the main contradicted by the results. In only two out of six measurements (one for each frequency range) did the LS group score like the NS group. More specifically, they produced a nativelike number of tokens per 100 words in the high frequency range (1–2000), and were nativelike on the T/T ratio in the frequency range beyond 2000 (2000+). The most interesting result cutting across the two frequency ranges is that the LS group produced significantly fewer types compared to the NS group.

https://www.google.com/url?sa=t&rct=j&q ... _4Pgz_mJnh

[Cavesa]

Yes, but only in exceptional cases. There are novel writers, who write in a non native language, I guess those are solid candidates.

Writing is a very different matter from speaking.

I assume you understand the concept of "fast thinking" vs "slow thinking", but here's a brief summary of the idea for readers who aren't:
Fast thinking is "second nature" -- decisions we can make unconsciously because we've practised the skill to such an extent. Fast thinking is related to (but not necessarily the same as) the idea of "procedural memory".
Slow thinking is actively analysing data to consciously reach a decision. Slow thinking is related to (but not necessarily the same as) the idea of "declarative memory" -- conscious memory of facts and ideas.

Speaking and writing both normally use a mixture of fast and slow thinking, but the balance in the two is different. Speaking, being real-time, relies very heavily on fast thinking and procedural memory. Writing is a much more deliberate process and uses much more slow thinking and declarative memory than speaking does. English is loaded with words that clearly show that writing is regarded as a conscious process -- eg sentence and paragraph composition; talk of writing as a craft; the term wordsmith for a poet or author -- and I'm sure many languages will be the same; indeed many writers find the word "writer" itself insulting, as their real job isn't putting the words on the paper, but coming up with the words in the first place -- no-one calls a sculptor a "chiseller", after all.

Some writers go very deeply into "slow thinking" their writing and can give reasons for a staggering number of word choices in their books. It stands to reason, then, that lack of spoken fluency is no barrier to a non-native writing a book that is acceptable to native speakers. Being able to consciously construct prose is a great intellectual feat, but it's not "better" that being able to speak at full speed with little hesitancy or errors, just a very different skill.

Someone with poor spoken skills can supplement

True, not all the writers speak as well as they write, you're right. However, I'd definitely say those exceptional foreigners speaking better than most natives (I said they were exceptions right in the previous post) are much more likely to come from this group than almost any other.

Or teens, who have moved abroad and received majority of their education in the language can end up speaking better than uneducated natives.

That goes back to artificial notions of "better" though. If natives don't speak that way, is it really the language?

Not all the natives speak equally well. Yes, "better" is a rather subjective word. But when I look at those teens, who speak their second language as well as their native one, and then at many natives with horrible skills (in general people with low education, not too high intellect, without interest in reading, without interest in speaking about more complex things than their immediate needs), the winner is obvious.

The fact someone has been using a language from birth doesn't make them a great speaker of it. And those teens moving abroad and entering the local schools (and reaching higher education in the language) speak like educated natives (=better than average speakers), but they simply don't fit the bill neurologically.

Our comparison of who is a "better" speaker is indeed based on not too clear criteria (what is and isn't a mistake, that is a topic that can light up very wild discussions even between people from different towns). But it is absolutely obvious that the idea that natives simply don't make mistakes (and basically just create new features of the language) and are perfect is wrong. Many natives speak their languages badly. Not recognising a certain standard as the norm and some differences from it as mistakes, that is just a very wrong way to legitimise ignorance and stupidity. Assuming that one's birth gives a certain right to make mistakes and still be superior to the foreigners or immigrants, that is pure discrimination of the second and third group.

As I said, it is rather rare for non natives to speak better than natives. But there are non natives that speak equally well as educated natives, and there are natives, who are simply bad at using their native language (usually due to their own neglect).

[Cainntear]

The fact someone has been using a language from birth doesn't make them a great speaker of it.

Hypothetical question:
Who is the "better" language speaker of these two?
person A speaks exactly like their "non-standard" parents,
person B speaks something halfway between their "standard" teacher and their "no-standard" parents.

On the one hand, person A has achieved notionally "perfect" acquisition of a presented model; on the other, person B has achieved partial acquisition of two, and synthesised something that lets them operate in two different contexts.

I would say it's impossible to pick one of the two as "better" -- they are merely different.

Not recognising a certain standard as the norm and some differences from it as mistakes, that is just a very wrong way to legitimise ignorance and stupidity.

There is such a thing as respectful disagreement, but to describe the norms of someone else's area of expertise as "stupidity" is every bit as ignorant as you claim my view to be. It is well established now in linguistics and education that while there is practical usefulness in a standard, it is not useful to present it as "correct", and colloquial usage as "mistakes".

It is clear and undeniable in the data, both historical and current, that the kids who do best at school are the kids who enter school with a language model similar to that of the teacher. Kids whose English/French/whatever is constantly corrected are less likely to interact in the classroom and withdraw from the learning process. They perform poorer throughout their school careers.

This reinforces the prejudice that speaking a form further from the standard is assumed to be the result of being uneducated, when the truth is that they are uneducated as a result of speaking a non-standard form.

I'm sure you'll be aware of a country called Czechoslovakia, and I'm sure you'll be aware of a language called Czechoslovakian. Now I realise that language was only one part of a complex issue, but it was an issue. The Czechoslovakian language was really just Czech and Slovaks speaking Slovakian were told they were speaking bad Czechoslovakian. Then look what happened.

Assuming that one's birth gives a certain right to make mistakes and still be superior to the foreigners or immigrants, that is pure discrimination of the second and third group.

This is not a question of superiority. If I was a natural blond, it wouldn't make me superior to someone who bleaches their hair blond. If was a natural redhead, it wouldn't make me superior to someone who dyes their hair red. I've actually got dark brown hair, and it doesn't make me superior to someone who dyes their hair dark brown, I just happen to have natural brown hair and they don't.

Speaking native English is often assumed to be an advantage, but the fact is that at a certain point it becomes a disadvantage: in international discussions, non-natives understand each other than natives understand non-natives. But of course you know this, don't you? You've even given this as the reason for not studying for English certificates -- you have no interest in speaking more like a native when you're only likely to be speaking English with other non-natives.

[tarvos]

Probably, given certain contexts, it can happen! But it's kind of pointless to state that one variant of a language is better than another, or to impute some kind of intelligence to a certain variant. I'm inclined to agree with Cainntear here.

I'll give a nice example. A week ago I had a Spanish class with a Chilean teacher. Her native tongue is Chilean Spanish. (This is a variety of Latin American Spanish, but it's still Spanish, despite what people from the RAE or Spain might say). However, there are certain items in my vocabulary that she doesn't know (which indicates that I may have, especially in certain contexts, a richer vocabulary than she does, even though Spanish is not my native tongue). For example, I used the word "chapurrear" (an ibericism in the context of speaking a language in a sort of broken fashion), and also the word "triunvirato" (triumvirate), which is a general word, but perhaps a very specific or formal one. Sometimes I will say something that is a very Spanish word to use - and she doesn't understand, but that's not because her Spanish is bad, in Spain you just use a different word.

And to think - I'm not Spanish. But I could still teach my Chilean teacher words.

But no one in their right mind would claim me as a higher authority on Spanish than her. Nobody. Even though I'm a fluent speaker.

[dampingwire]

I would certainly agree that someone using a non-standard language variant but using it in a consistent way should not be "marked down" in the "quality of native language use" stakes. (... if you switch between 'aks' and 'ask' I think it's going to confuse me, so ...)

But when you encounter someone who makes a quite ambiguous statement and you probe a little (because it actually matters that the correct information is imparted) and you find that they can rephrase it more clearly, might you be tempted to think that their use of language could do with some improvement? That would perhaps be harsh for an isolated incident, but not all such incidents are isolated.

Deciding whether you find someone "difficult" to converse with is fairly easy. Determining the underlying cause (accent, insufficiently accomplished language learner, insufficiently polished language use, not making the effort to communicate effectively, etc.) is probably harder.

However, since the purpose is to communicate, I'm sure that you can generally tell after any reasonably lengthy conversation whether you felt overly taxed in trying to decipher the other party's intent. If that's your measure of "native fluency" then I know a few non-native speakers of English that beat a few other native speakers of English that I've met.

[aokoye]

I'm very bemused by the idea that bilingual children whose school has a medium of instruction that isn't their L1 are somehow an exception. In 2011/2012 there were over 90 languages spoken at home by children across the largest school district in my state. Nearly 10k students spoke a language other than English at home and it's fairly safe to assume that there are more students who are L2 English speakers who weren't accounted for in that. Portland (the city this district is in) isn't especially large (647,805 people as of 2017) nor is it a city that anyone in their right mind would consider diverse with regards to race or national origin.

Needless to say, I know and have known a ton of L2 English speakers who speak English "just as well" as L1 English speakers. This is also an issue of, "just because you think they're an L1 speaker doesn't mean they are."

[Cavesa]

The fact someone has been using a language from birth doesn't make them a great speaker of it.

Hypothetical question:
Who is the "better" language speaker of these two?
person A speaks exactly like their "non-standard" parents,
person B speaks something halfway between their "standard" teacher and their "no-standard" parents.

On the one hand, person A has achieved notionally "perfect" acquisition of a presented model; on the other, person B has achieved partial acquisition of two, and synthesised something that lets them operate in two different contexts.

I would say it's impossible to pick one of the two as "better" -- they are merely different.

This is not the difference. A good speaker is one, who may speak perfectly a local dialect or non standard language, but still has huge vocabulary, and is capable of using the standard language, when appropriate.

What I am against is glorification of limited, ignorant, stupid people. Just because they are native, they are not good speakers of the language. If they don't have vocabulary for anything beyond their everyday life, if they are unable to use either standard or non standard variant, they are simply bad speakers of the language.

And there is such a thing as making mistakes. You are turning this into a discussion about people speaking dialects, or choosing non standard language in informal situations (which is absolutely ok). Making mistakes is different from that.

I am talking about people speaking with obvious mistakes. Or would you also claim that a native person with a significant health condition making them use only 100 words is a better speaker of the language than an advanced non native? That is laughable.

Not recognising a certain standard as the norm and some differences from it as mistakes, that is just a very wrong way to legitimise ignorance and stupidity.

There is such a thing as respectful disagreement, but to describe the norms of someone else's area of expertise as "stupidity" is every bit as ignorant as you claim my view to be. It is well established now in linguistics and education that while there is practical usefulness in a standard, it is not useful to present it as "correct", and colloquial usage as "mistakes".

It is clear and undeniable in the data, both historical and current, that the kids who do best at school are the kids who enter school with a language model similar to that of the teacher. Kids whose English/French/whatever is constantly corrected are less likely to interact in the classroom and withdraw from the learning process. They perform poorer throughout their school careers.

It is laughable to all anything concerning humanities as "established fact". Established facts come from research in science. When it comes to humanities, there are only opinions, some of which are prevalent. That should just be clear.

I am not talking about colloquial usage. That is part of speaking a language well. I am talking about mistakes. And the fact that deviations from the standard (in most situations) are considered a mistake, that is an extremely important and valuable defence mechanism of any culture.

Those kids, who don't get corrected (because using even well established colloquial language at school is simply a mistake), grow up in adults, whose emails are barely understandable. Sure, I see it all the time, that people you do not correct are definitely more confident. But in many cases, it just gives them more confidence to make mistakes or to not improve their skills. Perhaps this philosophy of not correcting people much is behind so many badly writing natives of various languages? Just a thought.

This reinforces the prejudice that speaking a form further from the standard is assumed to be the result of being uneducated, when the truth is that they are uneducated as a result of speaking a non-standard form.

I'm sure you'll be aware of a country called Czechoslovakia, and I'm sure you'll be aware of a language called Czechoslovakian. Now I realise that language was only one part of a complex issue, but it was an issue. The Czechoslovakian language was really just Czech and Slovaks speaking Slovakian were told they were speaking bad Czechoslovakian. Then look what happened.

Nope, there has never been a Czechoslovakian language. Even if you google the correct spelling, you'll get the information it was just a socio-political idea, not the reality. Czech was dominant, Slovak was important only in a part of the country, true. But nobody was thinking of themselves as of a native speaker of "Czechoslovak", and no Slovak was considered just a bad speaker of Czech, if that is what you you were hinting at. The linguistic struggles were only about dominance of one language over the other in the public space, never about "good or bad Czechoslovak".

Assuming that one's birth gives a certain right to make mistakes and still be superior to the foreigners or immigrants, that is pure discrimination of the second and third group.

This is not a question of superiority. If I was a natural blond, it wouldn't make me superior to someone who bleaches their hair blond. If was a natural redhead, it wouldn't make me superior to someone who dyes their hair red. I've actually got dark brown hair, and it doesn't make me superior to someone who dyes their hair dark brown, I just happen to have natural brown hair and they don't.

Speaking native English is often assumed to be an advantage, but the fact is that at a certain point it becomes a disadvantage: in international discussions, non-natives understand each other than natives understand non-natives. But of course you know this, don't you? You've even given this as the reason for not studying for English certificates -- you have no interest in speaking more like a native when you're only likely to be speaking English with other non-natives.

You don't understand. Being a native or non native is basically a neurological difference. Something happening in the childhood and you have no control over it. No more than over your colour. But your language skills are definitely something you can affect.

Claiming that natives simply speak perfectly, because they are native (=because of the area of their cortex the language is stored in), that is the same thing as assuming they are better because of their colour.

Yes, the non natives speaking the language extremely well are rare, I've never claimed otherwise. But there are such people. The non natives who immigrated as teens are such an example. And of course they are better speakers of the language than people, who may be natives physiologically, but speak badly with poor vocabulary and tons of mistakes.

Yes, the non natives in general speak differently from the natives. I have never claimed it was normal for non natives to speak better than the natives or exactly as them. It is just true, that some exceptions get to an extremely high level in the non native language, while some natives (mostly by choice) speak really badly. It is obvious, who is a better speaker, if we compare such examples, and the differences visible on the MRI won't change that.

[Iversen]

This discussion is running in two directions: one question is whether non-natives can acquire the same skills in a language as a native, the other is whether all native speakers actually are 'good' speakers of their language. And you can doubt this last claim even though you recognize that few non-native speakers even reach the level of a mediocre native speaker of a certain language.

I have heard speakers of Danish who couldn't say a complete sentence without stuttering - as if they constantly forgot they were saying and then corrected themselves. Others have very slurry speech - even more slurry than normal Danish -so that even other Danes have problems understanding them. There are people who miraculously have avoided to pick up even very common words, and others wo seem unable to construct a complicated sentence with several layers of subordinates. Such people are definitely not good speakers, but it is possible that they still are able to distinguish wellformed sentences from grammatical errors (as postulated by Chomsky, who clearly wanted to cut down on the tedious field work in linguistics). Or in other words: they might know things about their language which they for some reason are unable to use in practice.

But even if this is the case, it is questionable what the norm for 'good speakers' actually is.

Take one of those foreigners who have learnt to speak a language, but who doesn't speak any of the recognized dialects or sociolects. It could in principle be an unofficial 'school standard' in their home country, but it is more likely to be some kind of idiolect - especially when it comes to pronunciation. So such a person could be a more fluent speaker, have a more distinct pronunciation and know more words and expressions than at least some of the persons I mentioned above, but because he/she doesn't fit into one the recognized boxes he/she would still be categorized as a learner who just got stuck somewhere along the way, whereas any statement about a linguistic question by a native person who couldn't say a full sentence without stuttering would be accepted as a godgiven truth. It is not quite fair, but we accept it because we expect even bad native speakers to know more about their language than any learners, even though they don't actually speak it particularly well. And then we forget about the few exceptions.

If people do express themselves in a consistent way it is in principle possible to claim that they are good at whatever variant they are speaking - it's just not the standard we were expecting. And they could with some justification claim that it is unfair to measure them with an irrelevant yardstick. And if a child that speaks some kind of dialect in school is penalized for not speaking the official standard language that would definitely be a valid criticism. But then it's also about time that we drop the idea that any native speaker from any geographical or sociological subgroup is capable of delivering the final and incontestable verdict about each and every question concerning his/her native language.

And if we don't expect perfect knowledge AND perfect performance from native speakers, then we shouldn't expect more from language learners.

[Cainntear]

But then it's also about time that we drop the idea that any native speaker from any geographical or sociological subgroup is capable of delivering the final and incontestable verdict about each and every question concerning his/her native language.

We can't drop what we don't hold.

No-one of any consequence today* would claim that any given native speaker can "deliver a final and incontestable verdict" on correct language usage, and no-one in this thread has made any such claim.

Punching strawmen will never earn you a black belt.

* Chomsky's notion of grammaticality has been roundly proven incorrect.

[Iversen]

Actually I agree with everything Cainntear writes above. I'm just not as sure as he is that all traces of Chomsky's stance on grammaticality tests from around 1956 have been weeded out - although it now only would be shared by people we may deem as inconsequential (I hope).

Besides I have seen cases (outside this thread) where a native speaker claimed something that might be true in that person's own idiolect, but definitely not in in a broader context. I have even been caught in that trap myself: I once claimed in an article that a certain kind of sentence knot was impossible in Danish, and just a few days later I saw an example on a piece of paper pinned to a door. It is still not acceptable in my idiolect, but I should have known that there were people in my own town that didn't shy away from using such a construction. Unfortunately I have forgotten what exactly the problem was, but if I find out I'll add the example I found to this message.


FOUND - but not where I expected it:
I denied that this was possible:
Ham mener jeg ikke var i stand til nære en sådan interesse (literally: "him think I not was capable of harbouring such an interest")

but then I found this in a folder named "Værd at vide om musikmiljø Århus":
Ham troede jeg var død ("him thought I was dead" -meaning: I thought he was dead)
(quoted from the magazine Sprint 1981 no. 2)