Linguistic studies of non-native error patterns?

Do linguists ever study patterns of errors in English by non-native speakers and develop enough data to identify the native language of a speaker? I have known native speakers of many languages and have noticed that each group tends to make the same errors (Russians omit articles, Chinese make gender errors, etc.)

It’s not even necessary to study the errors in the speech of native speakers of language A who are speaking language B. Linguists can look at the various phonological, grammatical, syntactic, semantic, and vocabulary patterns in language A and in language B and predict what mistakes the native speakers of A will be likely to make when they speak language B.

What Wendell Wagner said.

There are a few times when linguists do the sort of thing that you mention. One of them is studying ancient texts, or even not-so-ancient texts, where you suspect or know that the author was not a native speaker of the language, and looking at the errors gives you some idea about what their native language was or how it worked. Think about late antiquity or early Medieval times, for instance, where people were still writing Latin, but the spoken language had moved on.

In general, there are a lot of interesting linguistic question tied into the ways that languages are produced by non-native speakers. The area that I’m particularly interested in, for instance, is what happens when you have enough non-native speakers of a language that they start to influence the original language. You can find cases in historical linguistics where a certain set of changes could only have arisen through the sorts of errors that non-native speakers would make. You can also find change in progress in the large immigrant communities of the major cities of Europe, where young children learning the language often only have access to people who are non-native speakers themselves.

One of the common “errors” my kids make speaking English to me is in response to negative questions. For example, if I ask

“Aren’t you hungry?”

they’ll answer “Yes” to agree they’re not hungry, while English idiom would be “No.” Is there a term to distinguish how negative questions are answered in a language? (French addresses the problem elegantly via a third word: Oui, Non and Si!)

septimus, read this Wikipedia entry for some information about the various sorts of yes/no patterns in languages:

Thanks, Wendell. Fascinating! So English once had not just 2 answer words, nor 3 like French, but a full complement of four with the the equivalent of French Si evolving as
yea soyes

Yes. At least, computational linguists do. The problem is called native language identification, or NLI for short. It’s a rather popular research problem lately, and there have been competitions at major computational linguistics conferences to see who can produce the most accurate system. See NLI Shared Task 2013, for example, which was held at a workshop at this year’s Conference of the North American Chapter of the Association for Computational Linguistics. I’ve got a basic understanding of the principles and have colleagues for whom this is a main area of study, so I can probably give further details if you’re interested.

Another thing to consider is the caliber of the teaching of English to the individuals who are making mistakes in English. Some of the mistakes they’re making may be because they were taught to make those mistakes. For an example of even native English speakers being taught incorrectly, consider the use of him and I as the object of a verb or preposition.

Yes, linguists do study patterns of errors by non-native speakers in English. I did it myself several times when I was at university. It can be a valuable real-world data source of how grammars, phonetics, etc. differ between languages. Actually the data is useful for lots of areas of linguistic study.

It also happens in reverse too. Non-English speaking people studying their non-native speakers for various purposes. It’s very interesting.

Thank you for that, more than I ever expected to discover. I am surprised that it is actually a kind of specialty. I am not a serious student of linguistics so details would probably go over my head but I’ll check this out.

Well, whattaya know. I learned more on the SDMB than in 5 years of high school French.

Another error I hear from Spanish speakers is a double negative, because that can be a word-for-word translation for a phrase that is grammatically correct in Spanish. Ex.: “I don’t have no one”. (No tengo a nadie --Carlos Santana). (Plenty of English speakers make the same error.[sup]1[/sup])


  1. The prescriptivist in me refers to this as an error. A descriptivist may call it a subculture dialect.