What are the reasons to justify that some text is in X language?

Question

Let us say that I am in a library alone and I have a text that I think that is in X language, for example, this fragment of the 9th chapter of the 2nd part of the novel 1984 by George Orwell, that I think that is in English:

It was curious to think that the sky was the same for everybody, in Eurasia or Eastasia as well as here. And the people under the sky were also very much the same—everywhere, all over the world, hundreds of thousands of millions of people just like this, people ignorant of one another’s existence, held apart by walls of hatred and lies, and yet almost exactly the same—people who had never learned to think but who were storing up in their hearts and bellies and muscles the power that would one day overturn the world. If there was hope, it lay in the proles! Without having read to the end of THE BOOK, he knew that that must be Goldstein’s final message. The future belonged to the proles.

There is nothing in the book or the library that say explicitly that it is in English.

There are no English speakers, and all I have at my disposal is a lot of books that are said to be written in English, plenty of dictionaries of English, and manuals of English grammar.

What argument can I provide to justify that this text is in English? My idea is that I can say that this text is in English because:

• The words appear in the dictionary of English.
• I can find examples of those words in other books that are said to be in English, with similar usages.
• The text follows the syntax and the morphology that the manuals advice.
• If I translate the words into my language, according to the English dictionary, they seem to present some coherent message.

I am dealing with something like that right now, but in Latin. I have tested each word and they seem to be properly declined, the meaning makes sense, and I have found examples of the oddly written words in other Latin texts. However there are other texts that criticise this text for being in "defective" Latin, or not being in actual Latin at all. Yet I tested the way I said and it passes the proof, so I think that the criticism is just more from a capricious point of view, calling it "defective" because it is not stylistically laudable, but not because it is missing anything linguistically.

So, I am looking for concise tests, or theories that talk specifically about how to know that something is in X language. As I said, there is no community to consult, and no native speakers, just texts, so there has to be a way to know that with only texts.

Honestly, I think that my proposed method is correct, however I have not found any book or article explicitly proposing a test to know if some text is in X language. Can you recommend some theory or some text about this or about the limits of language (when is a language still a language?)? Do you think that my argument is valid and solid and that it wouldn't need references to others?

Thanks for your attention.

P.S. I used the fragment of 1984 because the novel itself insists in saying that they are being forced to speak in "Newspeak", and that they are being conditioned to stop being able to distinguish well meaning and logic; so in my imaginary circumstance, I think that it might be probable to the reader of the book to wonder if it is actually written in English.

Answer 1

You are making the erroneous assumption that "X is in language Y" is a claim that can be verified as true or false. For many combinations of text and language, of course it can be verified, but for many others it cannot.

Note that I am not just restating your question: I am not talking about whether there is a test for verification (your question) but whether verification is meaningful. If a an otherwise English text contains one foreign word, is it English? What about if it contains an English word that has not been recorded since the 17th century? If it contains a word invented by the author of a science fiction novel? What if it uses a grammatical structure used in only a small number of dialects of English?

From your description, I think you have just such a case as this: it is not a matter of whether you have a technique to identify the language: it is a disagreement about what counts as Latin and what doesn't.

Answer 2

There are no objective absolute criteria for language-identification. There may, however, be subjective criteria, and there may be comparative criteria. Subjective criteria depends on one or more individuals declaring that some property of a language is, definitionally, either in or out. The criteria could be stipulated by an individual author (of a normative grammar or similar work), by a quasi-official board (The National Language Academy; The Swahili Council), or by some guy on the street who declares 'There is no word "ain't" in English'. There is no basis for accepting some such criteria, but if you do, it might allow you to filter texts for being English vs. non-English. Linguists typically consider this to be a pointless and unscientific enterprise.

Comparative criteria are somewhat more useful (still, not all the way to useful). Asking "Is this English" as an absolute will probably give a useless subjective answer, but asking "Is this English, or is it Dutch" ought to give a fact-based answer. Similar comparative questions would be "Is this Modern English, or Middle English?", or "Is this US English, or British / Indian / Australian English?". Or, you can ask "Is this English as written by a native speaker of some dialect of English; or is this English as written by a second-language speaker?". You might also ask "Is this English as written by an under-schooled or careless author who is a native speaker; or is this English as written by a careful, well-educated native speaker?". But that question should lead you to ask "Is this person posing as low-register native speaker, or does this represent how they actually talk". I am not suggesting that you can be certain about anything pertaining to a writer's performance, I am saying that the most you can hope to to accomplish is say something about social facts surrounding an author, based on their writing. Asking "Is this English?" is scientifically meaningless.

Answer 3

Like in most cases of categorisation, Prototype Theory helps. There are exemplar texts of a language. I'm not actually an expert of Latin, but perhaps Cicero, Seneca, the Vulgate, and Aquinas could be considered to be exemplars of Latin, showing how the language has changed. Close to each exemplar text are the texts written in very similar style, by native speaker contemporaries. Texts by second language speakers are further out, as is Church Latin, and then those halfway to becoming Italian, Spanish, French, Romanian etc. Way out on the distant periphery is English, whose grammar is thoroughly non-Latin, but which shares a lot of vocabulary.

With Prototype Theory, the task is not to establish rigid defining borders, but to find the most helpful exemplars, and then to organise the data around them. If your text seems to be declined perfectly, that suggests it's quite central. But there could still be semantic, or stylistic issues which mark it out as non-native Latin, late Latin, etc.