Talk:List of endangered languages in Central America

A couple of observations

Hello there, just a little bit of input to help keep this page alive... (1) Lenca in El Salvador is certainly extinct. Being distracted with other tasks it's not convenient for me to look up a source for this affirmation right now, so I'll wait until I do and then change the main page. As for Honduran Lenca (which may not be the "same language" as ES Lenca - I think the question remains open because of the even scantier documentation of Honduran Lenca), some writers claim it is also extinct while others hold out some hope that there may still be a few speakers around somewhere, in which case nobody knows how many (certainly very few). Not sure how that sort of information would be registered in the table, but anyway that's what I gleaned from a bit of Internet research on Lenca I did a few months ago.

(2) Pipil is a language I know a lot about, certainly enough to know that it is impossible to say how many speakers there are. The "20" given in Ethnologue has been around for a few years and competes with other slightly more optimistic estimates. I am part of a local project which works with maybe a dozen speakers, and since this is not a census project one can only extrapolate from its experience that there must be quite a few more "out there somewhere" - which would surely bring us to more than twenty, despite the fact that the Ethnologue "20" is already old and numbers are presumibly dropping as we lose the elderly.

(3) On the other hand, there is a big problem here: how to define "speaker". With any language, in addition to full speakers and total non-speakers there are sure to be some semi-speakers (however you wish to define these). With large overall speaker figures you can probably afford to ignore that issue for most purposes. But if you're talking about 20 speakers, what if fifteen (or all twenty) are actually faulty enough to be called semi-speakers? Or does 20 mean (as it should in theory, I guess) 20 full speakers plus an unspecified number of semis? In that case, by just a little nudging of the threshold applied to differentiate between full and semi you could maybe boost your "speaker" figures - maybe jack them up to 100 or 200? My point is that with very low speaker figures, probably most speaker "statistics" are prone to be potentially misleading or just not fully meaningful.

(4) Often in practice, in the case of Pipil for example, there is a further important problem: the (probable or known) existence of what I call crypto-speakers - people who will not admit publicly to being speakers (and also people who will not say who else is a speaker, for that matter). If this phenomenon is widespread in a given language community (or even crypto-language-community?), standard-type surveys may be quite unreliable. In the Pipil case I take the existence of some crypto-speakers (and almost by definition, we may never know how many!) as possible evidence suggesting (though not proving) that there are really more speakers than the stats say - good news for the survival of the language. The bad news is that crypto-speakers are presumably more likely than "confessed" ones (speakers who have "come out"?) to fail to form part of speaker networks and hence participate in a language community as such. Though this is not necessarily always the case either: there are probably not just individuals but families (i.e. households) that are (to some degree) crypto-X-speaking - we have turned up a couple of examples like that with Pipil. "Typical" scenario: F, a field workers, visits an indigenous family who say they can't speak X. Finally M, the mother, admits to "remembering" a few words and phrases. F pleads with M to say some of them, and finally she does. G, the grandmother who has sat there silently until now, corrects M several times as M searches her memory. But what really amazes F is that other members of the family now start to join in the discussion of what is correct - including the grandchildren! How on earth can they know, if the language is "not spoken" in the family? And what if F had taken on faith the first answer that nobody spoke the language, and moved on? I realise none of this is going to be much help for the data tables, but it brings you a bit closer to the realities of life with an endangered language...

(5) The real reason I started writing these observations was to say that I don't understand how the "speakers" column in the tables works. What for example does "20 (1987) 196,576 (1987)" mean for Pipil? I assume this is saying 20 speakers, and a population of 196,576. If so I think we need a more explicit way to tabulate that. In the entry for ES Lenca, "36,858 in El Salvador (1987)", how is one to know that this (I assume!) refers to the population, not to the number of speakers (which as I said above is actually 0) (apart of course from the clue that if it were the latter Lenca would not be quite so "endangered"!)? --A R King 07:26, 1 March 2007 (UTC)

OF course it's quite common for an ethnic group to retain many aspects of its separate identity even while losing the language, due to language shift.. lack of transmission of the native language to younger generations, as they learn a dominant language in school etc. I believe I pulled the pop. figures from Ethnologue. :-) --Ling.Nut 15:02, 1 March 2007 (UTC)