I think the thing that doesn't make sense to me about all of these language preservation efforts is the idea that a language has to have people actively speaking it.

That's a gigantic lift, that takes a ton of effort. Even if you teach a new generation they likely won't keep using it in their day to day.

Why not just catalog it? Get the most experienced speakers to record hundreds of hours of audio of it, if it had a writing system catalog that too.

Problem solved. Language preserved. Otherwise you're fighting a losing battle trying to keep a language actively being used when only a few thousand people know it still.

Because you need to have native speakers in order to properly analyse the language. Let’s say you have a corpus and you’ve found two words that seem nearly synonymous: what’s the difference between them? A native speaker might be able to tell you about such subtle differences, but it can be impossible to know otherwise. Or, let’s say you’ve found an affix with an unusual distribution — if you want to test your theory of where it is allowed, you’ll need to have a native speaker to double-check if you’re using it correctly. Or even at a more fundamental level, one of the first things any book about field linguistics mentions is that it’s nearly impossible to transcribe audio without going through it with a native speaker, because you won’t necessarily be able to perceive all the sounds at first. (Disclaimer: I’m not a linguist myself, but I’ve seen this advice mentioned several times.) Problems like this are why language preservation generally ends up as a race against time until the last native speaker dies.

It sounds like the main argument you're making is that having a record wouldn't catch every single edge case. But maybe that's okay? If you get 90% of the value of preserving the language and you miss the 10%, it's probably still worth doing vs. trying to artificially keep a language alive.

If you read about this language, it emerged from Old Norse. Do you think there was some kind of "Preserve Old Norse" movement?

I just don't see how an effort like this could possibly succeed. If you make kids learn it, the chances of them retaining fluency in it are pretty low I'd think? They'd have to be using it a lot. And I think most people would probably rather their kids learn English as a second language.

I guess a solution would be to teach the kids English and Elfdalian instead of English and Swedish, and then when they need to talk to people from other regions they just use English as the lingua franca.

> It sounds like the main argument you're making is that having a record wouldn't catch every single edge case. But maybe that's okay? If you get 90% of the value of preserving the language and you miss the 10%, it's probably still worth doing vs. trying to artificially keep a language alive.

The problem is, it probably won’t be 90/10. I worry that it may be more like 50/50 or even 10/90, though to be sure I’d have to ask a linguist who’s actually worked in the field.

> If you read about this language, it emerged from Old Norse. Do you think there was some kind of "Preserve Old Norse" movement?

There’s a difference between evolution into a new language and language death. The former makes no difference to language diversity: the old language is just as interesting as the new one. (Or, in many cases, it increases diversity, because one language can split into multiple daughter languages.) The latter decreases language diversity, as a spoken language is irretrievably lost.

> I just don't see how an effort like this could possibly succeed. If you make kids learn it, the chances of them retaining fluency in it are pretty low I'd think? They'd have to be using it a lot. And I think most people would probably rather their kids learn English as a second language.

I don’t disagree. Language revival rarely if ever succeeds. This is why it’s important to do as much linguistic work on each language as possible while native speakers are still alive — simply making records of the language is not enough.

>may be more like 50/50 or even 10/90

Maybe for some languages that are at the point of only having 2 speakers left. But at that point, you likely couldn't revive the language anyways. For this language, they absolutely should be able to record it in its current state in an extremely comprehensive way.

>There’s a difference between evolution into a new language and language death.

But Old Norse did die. It happened to sprout several new languages, most of which are not in danger of dying out like Danish and Swedish. The original Old Norse is gone though.

You seem to only be valuing a language based on if it's a net plus or minus to the "Language Diversity" which as you're using it seems to mean the number of distinct languages that are currently spoken natively by currently alive humans. Old Norse dying was -1 but sprouted many new languages so +n. Where Elfdalian dying would just be -1.

I would think a language would be of value based on the content of the language itself, for linguists to study.

>simply making records of the language is not enough.

I would consider doing linguistic work on the language with native speakers to be part of making records of the language. I don't think we're in disagreement here.

The original idea I was pushing against was trying to artificially keep a language alive by trying to keep large population groups speaking it natively. It goes against how language has always worked.

> Maybe for some languages that are at the point of only having 2 speakers left. But at that point, you likely couldn't revive the language anyways. For this language, they absolutely should be able to record it in its current state in an extremely comprehensive way.

Not at all. Even for languages with many more speakers you can get these sorts of problems.

For a very concrete example, consider this Japhug text [0]. I consider myself fairly knowledgable about phonetics, and I can read the transcribed text out loud. But listening to the accompanying recording, I can barely hear the distinctions between many of the sounds, or figure out where the word boundaries are. Such a recording can only be transcribed by carefully going through it with a native speaker.

Or, for a report by a more experienced linguist, here’s e.g. R.M.W. Dixon (2010) on the subject:

> All texts should be transcribed in the field. In the early stages the linguist should not attempt transcription without a native speaker by their side. As the linguist gets to know the language better they can attempt an initial transcription on their own, but should always then go over it in detail with a consultant. In the case of a language I had been working on for more than thirty years (Dyirbal) I could get a new text 95 per cent right, but there were always a few points that I missed the proper meaning or full significance of, and had to have them pointed out by an expert consultant. One thing one should never do is just record texts in the field and try to transcribe them later on, back at base (whether one is working on phonetics, phonology, grammar, discourse, or whatever); this is a sure recipe for an incompetent analysis.

[0] https://pangloss.cnrs.fr/corpus/show?lang=en&mode=pro&oai_pr...

> But Old Norse did die. It happened to sprout several new languages, most of which are not in danger of dying out like Danish and Swedish. The original Old Norse is gone though.

It did not die. There is a difference between language death and language evolution. In the latter process, there is no one point at which you can say, ‘Old Norse is dead and gone’; it is simply the case that speakers of the language found Old Norse texts less and less intelligible until they could no longer be understood. But the Old Norse language itself is hardly gone: many of its features are preserved in the derived languages, and Old Norse may still be reconstructed by comparison between its daughters.

Perhaps a more illustrative example would be Old English. Is Old English ‘dead’? Possibly, but it never ‘died’: it just gradually evolved into Middle English and then Modern English (and Scots).

> You seem to only be valuing a language based on if it's a net plus or minus to the "Language Diversity" which as you're using it seems to mean the number of distinct languages that are currently spoken natively by currently alive humans.

Not at all. I agree with you that languages are valuable because they give a unique perspective to both their native speakers and linguists at large. A natural consequence of this viewpoint is that one extra language is a net good, but I don’t think a language is valuable only because of this.

> I would consider doing linguistic work on the language with native speakers to be part of making records of the language. I don't think we're in disagreement here.

Ah, in that case we are indeed in agreement. I had assumed that by ‘making records’ you had meant only making audio recordings etc. I should have clarified what you meant exactly.

> The original idea I was pushing against was trying to artificially keep a language alive by trying to keep large population groups speaking it natively. It goes against how language has always worked.

I don’t think it’s a good idea to ‘artificially’ keep languages alive if there’s no reason to do so — but if a linguistic community wants to keep a language alive, then why not?

Probably the human tendency for loss aversion. No one today cares about the exact preparations and steps needed to prepare an offering of sheep intestines to Zeus other than a some PhDs in classical studies, but I could imagine ancient Greeks thinking that their youth deciding to worship Roman or Christian gods was a bad thing that took away from their cultural heritage. But after that event happened, no one today really cares or notices the loss of the precise minutia of ancient Greek cultural practices and values.

If another random species of lemur died off and we never sequenced its genome, you might not care. But many scientists would. That arrowhead you discount as useless is another data point to an anthropologist. One man’s trash is unfortunately another man’s dissertation and a window into the past.

Edit: an apostrophe

Yes, these niche languages are essential for comparative linguistics and sometimes more generally.

Why not? As far as "looking for something to do", preserving a dying language seems a lot more interesting than a lot of other things people do with their spare time.

The incuriousness in this thread is atypical of the board in my opinion.

At one point I used to think that language diversity was actually a net bad thing for humanity, because it decreases our ability to communicate and so on. These days language technology and machine translation are solving that communication issue and we may choose to preserve dying languages for the sake of understanding how language itself works, how humans migrated in the past, and so on.

Some languages open your understanding of the other ones from the same family tree.

Learn one, and you are partially learning the other.