Let us suppose that we have a text that in its majority follows the orthographic or grammatical rules of a language X, but 10% of the words have orthographic mistakes, and 10% of the sentences have grammatical mistakes. Is this text still a text in the language X?

I am not talking about popular mistakes or new trends and fashions, nor of "varieties" nor "dialects" with communities endorsing or practising that way of writing; I am talking about mistakes that are happening only in that text, and that probably none else will adopt later, and 99% would easy recognise those details as mistakes. However, the reader can still understand the text and infer easy what it is being tried to say.

I am writing a paper in linguistics in which this issue is very relevant and I have not found any academic text (maybe because I am not sure about how to phrase this problem) that talk about this. What is the view of linguistics on this issue? At what point is a text not longer in language X? Could you recommend me some linguistic (or philosophy of language) theories or discussions that talk about this (the identification of a text as pertaining to certain language)?

  • 6
    How could you even say that there are errors, if you (you, the OP) did not assume that the text is in X in the first place?
    – DaG
    Aug 5 '19 at 13:09
  • 1
    This reminds me of raeding wrods with jubmled lettres :)
    – Andrew T.
    Aug 5 '19 at 15:02
  • "Mistake" is by definition done by a reasonably competent L1 speaker in the context of language acquisition, where "Error" describes a learners unwitting corruptions of the language--arguably still in that language. I take it that is not your field of interest? en.m.wikipedia.org/wiki/Error_analysis_(linguistics)
    – vectory
    Aug 9 '19 at 20:22

A useful philosophical framework for discussion is the classic dichotomy between competence and performance, promulgated by the theory of generative grammar, and the difference by I-language and E-language – see Chomsky 1986 Knowledge of language. Questions of text are in the domain of performance and E-language domain, whereas by definition there are no errors in competence.

Errors of speling and grammar change not a text from being "In X" to "Not being in X". Instead, they make the text be "In X, with errors". Massive errors of a specific type could change a text in Norwegian to Swedish in which case you would say that the text is "In Swedish". Such a shift cannot arise from true error, it would arise from an I-language matter such as that the author does not actually speak Norwegian or didn't learn either of the literary standard dialects of Norwegian, but did learn Swedish.

The errors might be so numerous that the text is utterly incomprehensible as being in any language, but such errors would not be due to the writer's performance, instead they would come from external scrambling -- not impossible with electronic transmission. The most common form of that is the conversion of a text into gibberish Chinese, by misparsing the Unicode.



If the sentence is intended to be in that language, then I'd say that it is in that language. Intent matters a lot in communication - for example, the utterance "Hallo!" would be a valid sentence in multiple different languages, but any particular usage of it "is" in the language which the speaker intended, inferred from the surrounding context.

It's worth noting that in the domain of formal languages e.g. Chomsky type-3 regular languages with formal grammar, it would literally be the case that a sentence is clearly "in language X" or "not in language X" if it has some mistake; it's just that this definition and these rules aren't directly applicable to general linguistics and natural language processing.


Yes, of course.

In corpus linguistics we often have to deal with texts digitised using OCR and containing some amount of OCR errors. We are completely aware that we cannot get rid of the OCR errors completely with reasonable effort, but we treat them as texts in the language they are supposed to be.

  • If I'm not mistaken, the Linguistics Stack Exchange has an English only policy. Please replace "COR" with "OCR" to comply.
    – Jetpack
    Aug 7 '19 at 2:12
  • Been there, done that Aug 7 '19 at 8:54

Drawing from math I want to say, it depends.

First of all, a text is not generally in that language, but evidence that may allow reconstructing that language. The syllogism calling evidence the fact, the text a language should only be allowable if the language can be reconstructed from the text completely. Of course most texts are too short for that. It may be compared with text that is believed to be in the same language. This biases the comparison; For the worse, I guess, if it can obscure severe mistakes. That is, there is a difference between systematic and spontaneous effects.

Secondly, mutual understanding defines language, so your question seems tautologic in the binary sense. In a gradual sense, your "10%" are arbitrary when not derived from a reasonable, quantifiable measure.

I had just been reading on point-set topology and find that the imagary of a limit of a set and a neighborhood of a limit are a fitting metaphor. One step in the wrong direction, one small mistake too much, can render the message incomprehensible; but there is a known sequence of inferences (the reconstruction) that approach the intended meaning. This is pseudo "abstract non-sense" and my abstract algebra is weak; Started with Closed-form expression and read up-to Open Set. The imagery is Fig 1. Extended with Fig 2.

Fig 1: Imagine the point *p* is a text, *V* is a language; *V* is not a *neighborhood* of *p*, but *p* is an element of *V*, [ref](https://en.m.wikipedia.org/wiki/File:Neighborhood_illust2.svg)

Fig 2: *p* is not the interior, but the boundary of *V*, [ref](https://en.m.wikipedia.org/wiki/File:Runge_theorem.svg)

Another problem is that your 99% are supposed to be competent speakers, but this property cannot be established without the sought metric. It leads to a slippery slope well known in the categorization of e.g. the development from Latin to Italian, if Dante is the proto-type for not-Latin. And it leads to arguments about how you can be sure that you understand the devious text correctly (which is a general problem).

For a German English speaker, it sounds as if you described Dutch, though the error rate is closer to 100% percent and my understanding thus rather limited unless I know what's being said. You might need to weigh the defective features for an accurate percentage, in which case the idioms, grammar morphology and verbal base should appear much closer to 10% defective (the numbers arw figurative, orders of magnitude). I know you said no varieties but I like to think of Dutch as Rudi Karell's accent when speaking German, and any real written Dutch, read with my own intonation just appears as an immense speech impediment. (No offence intended, I love it). The whole "dialect with an army" idea applies. But note that boundaries by definition can be approached from both sides.

  • Can someone fix the pics? I do not know what I did wrong.
    – vectory
    Aug 6 '19 at 6:11

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.