threshold of a linguistic 'slippery slope', aka Linguistic Tip - Summary of Results

[Nicholas Ostler]

At the end of November last, I wrote:
> It is sometimes claimed that there is a critical minimum proportion of 
> speakers of a given language in a multilingual community for that 
> language to continue in everyday use. Such a claim makes sense in a 
> context where there is a background metropolitan language (typically 
> English, but it could as easily be Portuguese, Russian, Spanish or 
> Chinese) that is under no threat, and spoken by numbers approaching 
> 100% . The other, less widely spoken, language can only survive in 
> stable bilingualism with this background language if there is a fair 
> presumption, within a given community, that enough listeners are there 
> to understand it.
>
> The idea, then, is that there is a kind of tipping point, or a 
> threshold of the slippery slope, perhaps as high as 70%; if the 
> lesser-speaking community dips below this proportion, it will tend to 
> diminish further, until (without active policy measures) it might die 
> out altogether. But above this proportion, its numbers can vary up and 
> down with no long-term effect or trend visible.
>
> Is there a percentage figure one could give, and if so is there any 
> research that bears directly on this point?
Here are the answers to my question, as offered by FEL-listers and 
lgpolicy-listers.

A. In general, there was scepticism from linguists as to any percentage 
boundary.

Hartmut Haberland <hartmut at ruc.dk> wrote:
> I come immediately to think of Srivastava's matrix of +/-power vs-. +/- numbers,
> which allows (without quantification) for a minority language vto survive, as
> long as it has prestige and/or is associated with power. One might as well ask:
> how few speakers can a dominant language have before it loses its grip on a
> community? (See references below.)
>
> One also would have to distinguish between mother tongue speakers and users.
> Urdu, as far as I understand, is a minority language in Pakistan as far as
> native speakers are concerned, but not as far as second language speakers (or
> simply users) are concerned. Practically nobody speaks French (nor German) as
> their mother tongue in Luxemburg, but both languages have a very firmly
> entrenched position in society.
>   
Jean Aitchison wrote:
> ... my memory of discussions about this point is that what matters 
> is not so much percentages and proportions, but the FUNCTION of the remaining 
> uses of a declining language. IF a language is used only at home, then it might 
> die. IF it has at least one `official' use (education, etc) or `public' use (at 
> a market place, etc) then it has some hope of surviving.
E.Williams wrote (with many - excised - caveats):
> I'd be inclined to go back a few years to the social psychology stuff 
> & the concept of "ethnolinguistic vitality" rather than absolute 
> numbers or percentages.
> (I'm sending a list of  some older seminal papers  in this area)
> Density of personal networks & 'awareness of difference' between 
> groups (even a degree of antagonism?) seems to be crucial in language 
> maintenance.
> Examples include the Lozi in Zambia, and the Welsh in Caernarfon 
> (where there is that rare phenomenon in Wales, a large Welsh speaking 
> council estate.) There are of course many other similar cases.
Jane Freeland wrote:
> My own feeling on this is that it doesn't work in simple percentages 
> of speakers. Indeed, the criteria developed for judging language 
> vitality by the UNESCO Ad Hoc Expert Group on Endangered Languages, of 
> the UNESCO Intangible Heritage Unit, specifically warns against using 
> any single measure, but that we should attend to a spectrum of them, 
> and that in different communities different criteria will carry more 
> or  less weight.
> ... Another aspect of the problem, which we noted in our compilation, 
> is that in many multilingual communities, the threat to a small 
> language does not necessarily come from the mainstream language, but 
> from another, stronger, indigenous language.  This is certainly the 
> case with the Mayangna of Nicaragua, among whom Elena and I are 
> currently working, who are shifting to Miskitu at an alarming rate. 
> Indeed, I'm about to go back to Nicaragua for stage 2 of a pilot 
> project which looks at popular discourse on language and languages, to 
> try to get at the sociolinguistic/power issues underlying this.  As 
> Elena said in her paper at the Barcelona FEL, on paper the 
> circumstances for maintaining this language seem highly propitious, 
> but...
>
> Nancy Dorian suggests, in her paper in Whaley and Grenoble (1998), and 
> elsewhere, that languages spoken by very small numbers of people can 
> survive provided that they have a clear social function within the 
> spectrum of languages in the context, and suggests that we should 
> focus more research on what constitutes 'success' in often 
> unpropitious circumstances.  
Tony Woodbury wrote:
> There are cases where only two speakers remain in a community  and 
> they speak with each other. Such was the case in Sirenik with 
> Sirenikski Siberian Yupik. But there are also cases where two speakers 
> remain, but they haven't spoken to each other in the ancestral 
> language in years--I think Leanne Hinton has described such cases in 
> California. Therefore, there can be no threshhold in PRINCIPLE. 
> Conceivably, the two Sirenikski ladies would have CONTINUED 
> indefinitely had they not died.
>
> On the other hand, we've seen cases of "tip" occurring in a 
> generation, and that I've seen myself in Chevak. When I got there in 
> '78 everyone spoke Cup'ik; now, people under 30 do not speak Cup'ik. 
> So even 100% can fail to offer a toehold. ... 
> for all I know, there are local generalizations to be made (e.g., what 
> might hold for Latvian might also hold for Lithuanian and Estonian); 
> although even so, things can go differently despite all other things 
> seeming to be equal: There is a famous case of two villages on 
> opposite sides of the Kuskokwim River in Alaska, one of which went 
> from Yup'ik to English in a generation, while the other hung proudly 
> on to Yup'ik. The villages were of the same size and practiced the 
> same religion (Russian Orthodox). Further analysis showed that the 
> villages simply opted for different survival strategies--one took up 
> overland and river trade with (somewhat) distant whites; the other 
> opted for river and back-country fishing, hunting, and trapping. So 
> really it was a matter of overall worldview. (I think you can guess 
> which one retained Yup'ik.)
>
> Statistically speaking, you don't need thresholds to prove a point--I 
> should think that TRENDS toward loss would also be convincing.
>
Abderrahman El Aissati wrote:
> ... there are a few known models that help predict language shift (the
> 'generations' model (Gonzo and Saltarelli), the code-switching model
> (Meyers-Scotton), etc.) but none can give the exact amoount of speakers
> needed to maintain a language. I think it is more appropriate to quantify in
> terms of functions of the language (hence the idea of a parametric model):
> if a language serves religious purposes, it takes one very religious person
> to use it (Classical Arabic and Latin), if a language symbolises a break
> away from traditional patterns/style of life, it needs two people to speak
> it (examples of colleagues in Morocco who use French)...
> Now these lges have a metropolitan backing, but the real issue is that they
> have/seerve important functions. If you can find a minority lge that can
> serve a high (symolic or otherwise) value, then you can predict that the
> community which uses it will continue using it (or probably raise its status
> like Hebrew) as long as those functions are needed! If this sounds kind of
> circular, then I'm afraid this is the nature of language and use...
mark abley wrote:
> Fascinating question. It neatly sums up why speakers of Welsh in 
> northwest Wales are right to be wary. I'm always wary of precise 
> figures because of the subtle differences between individual cases, 
> but 70 per cent seems a reasonable enough guess to me. Three quick 
> points:
> 1. I think this rule tends to be more applicable to bilingual 
> communities than truly multilingual ones. Much of India and some parts 
> of Africa don't seem to fit the rule too well -- i.e., the "background 
> metropolitan language" perhaps has less weight when several other 
> languages are in use, not just one.
> 2. In rare cases a minority language may be able to gain in strength 
> so as to reach a point where 70 per cent (or some such figure) is not 
> so much "maintained" as "achieved". Latvia springs to mind. I wonder 
> if East Timor (Tetum) may prove to be another.
> 3. An area for further research -- by sociolinguists, not by me! -- 
> would take off from this perceived threshold. Where a language is 
> spoken across a fairly wide area, or in many communities, obviously 
> the 70 per cent (or whatever) figure applies to each community, not to 
> an entire language -- e.g., Inuktitut remains strong in the Canadian 
> Arctic even though there are places in the Western Arctic where it has 
> dipped well below 70 per cent and is weakening fast. But what happens 
> when some communities abandon their ancestral language, while others 
> strongly maintain it? For a language to remain vibrant, is there 
> likewise a minimum percentage of communities (not just of individuals 
> within a single community) that need to maintain it?
P-J Ezeh wrote:
> ...If my own experience among the Orring of southeastern
> Nigeria (language: Korring) is anything to go by, I
> will say that the only factors that may lead to
> glotticide are economic and political, especially the
> later. There are recent examples locally here and
> farther afield, for example under the erstwhile Soviet
> Union, where minority laguages were threatened
> directly by political policies. In the Nigerian case
> there wasn't a direct political threat but the
> minority speakers reached a point where they began to
> see their distinct linguistic identity as a liability
> in the new political order that destroyed their
> original social structure and merged them with a
> numerically superior people with a different language.
>
> This has been the subject of at least two conference
> papers that I have had the honour to give, the last
> being at that of the last Pan-African Anthropoloigal
> Association in Yaounde. Perhaps what is even more
> interesting is that, contra the popular assumptions,
> the threat comes from not the culture of an external
> world political power but a local rival group who for
> a complex configuration arising from British
> colonization has assummed some advantageous position.
>   

B. Sympathy for the  idea, such as it was came from those campaigning 
for small languages, usually in a European context.

Giorgio Cadorini wrote:
> My personally – not proved – impression is that probably a critical 
> percentage will exist, but it will change following:
> --- the density of the speakers in the larger language community (if 
> they are 15%, but living and working one next to the other, the 
> language could be not in danger at all – maybe it was the case of Jews 
> in Eastern Europe until XXth century);
> --- the differences in the way of living: nomad Romanis in settled 
> Europe preserve their language until XXth century.
> Then also speakers forming a high percentage could lose their language 
> very quickly because of mass change. 
Andrejs Veisbergs <anveis at lanet.lv> wrote:
> this is an interesting issue, sometimes superficially discussed also in Latvia.
> The general idea here is that it is somewhere between 50% and 75%. But of
> course there are no studies, it is just a psychological issue.
> In Latvian case the idea is linked also with the pre 2world war situation with
> its 75% Latvian majority. This is then taken as a normal situation towards
> which one should aspire. 50% was the result of the soviet policies (Latvian
> speakers sank to 52%) and Latvian felt that if it goes below, it would mean
> Latvians cease to be majority / titular nation which theoretically under the
> soviets could lead to.... noone knows really what. I would say a purely
> psychological issue, but then psychological factors are sociolinguistically
> very important. One shouldalso take into account the levelof language
> development, city/country distribution, mobility, inertness, etc. But I would
> feel!!! that 70% is about the tipping point in the long term.
>   
Hassan Ouzzate wrote:
> From my experience of language situation in Agadir, your figure does not
> seem at all unrealistic. I would tend to confirm your description. 

C. And there was some suggestion for refinement:
Cunliffe D J (Comp) wrote:
> The below is a quotation from "Spreading the word, the Welsh language
> 2001" by J Aitchison and H Carter which looks at the results of the 2001
> census in Wales (page 134/5). It is talking about the Welsh speaking
> heartland.
>
> "These language bases act therefore as a core reserve sending out pulses
> of speakers which keep the language alive over a broader territory. This
> condition assumes that the core, if not monolingual, returns proportions
> of speakers well over 70 per cent, and preferably over 80. A condition
> where those proportions are substantially lowered, even though there is
> a wider spread bilingualism with proportions nearer 50 per cent, can be
> interpreted as no more than a stage in language death. Without the
> resource of first language speakers brought up in the minority language
> in a well-delineated heartland, then gradual decline and elimination is
> certain."
>
> What I thought was interesting was the idea that it might not be
> necessary to have this 70% threshold across the whole region/country, if
> a heartland exists.
Emily McEwan-Fujita wrote:
> Dear Nicholas,
>
> ... the phrase 'linguistic tip' was coined by Nancy Dorian in her 1981
> book _Language Death: The Life Cycle of a Scottish Gaelic Dialect_: "In
> terms of possible routes toward language death it would seem that a
> language which has been demographically highly stable for several
> centuries may experience a sudden 'tip' after which the demographic tide
> flows strongly in favor of some other language." (p. 51)
>
> There _is_ a slightly different kind of claim floating around that is
> based not on percentages, but on actual numbers of speakers. This claim,
> frequently repeated by linguists and the media, is that a language needs
> 100,000 speakers to survive. This claim originated as a series of cautious
> speculations on the part of the linguist Michael Krauss, and is not a
> scientific theory. I trace the origin and transformation of this claim,
> and its application to Scottish Gaelic by British journalists, in my
> article titled "'Gaelic Doomed as Speakers Die Out'? The Public Discouse
> of Gaelic Language Death in Scotland." It is forthcoming next year in
> _Leasachadh na Gaidhlig: Revitalising Gaelic in Contemporary Scotland_,
> Wilson McLeod, ed., Dunedin Academic Press (Edinburgh). I'd be happy to
> send you a copy via e-mail.
>
> Based on my analysis of this "100,000 speakers" claim, I would tend to
> take with a grain of salt any claims of an actual percentage of speakers
> within a community needed to maintain a language as a means of everyday
> use. Not least because journalists' favorite way to use the 100,000
> speakers claim in Scotland is as "scientific proof" of the supposedly
> imminent (or even completed) death of Scottish Gaelic, and as an argument
> against public funding for Gaelic revitalization efforts.  The best suggestion I received for the origin of the 70% figure put it
> down to simple arithmetic:
>   
D. Here is the one suggestion of where the 70% figure came from, putting 
it down to simple arithmetic:

Wilson McLeod wrote:

> This is not entirely responsive to this query, but my understanding of 
> the 70% threshold (as explained by Kenneth MacKinnon) is a logical 
> one: that 70x70 =49. If only 70% or lower of a given community know 
> Xish, then the chances of any two people in that community who are 
> unfamiliar with each other will each know Xish is less than 50%. This 
> means that it is more likely than not that they will simply use Yish.
>
> This is obviously very far from an iron-clad rule, let alone an 
> observed trigger of language shift in real situations, but it does 
> provide a logical explanation for choosing this specific 70% threshold.

E. One offered alternative, non-quantitative criteria:

Jane Simpson wrote:
> Peter Sutton has claimed long-term stability for very small communities in
> Cape York with limited contact with colonisers.  I speculate that, in the
> colonial language situation, it helps to have distributed communities -
> i.e. if you live in a small community but can frequently visit a
> neighbouring community and speak the same language, this is a confirmation
> of the usefulness of your language, and an incentive to keep speaking it.
>   
E. There was one socio-caveat

Adrienne Redd wrote:
this term the "tipping point" is a concept well known to (at least 
American) sociologists and is used with regard to (usually black-white) 
racial mix. The idea is that blacks may be comfortable with up to 80% 
whites but that whites may only be comfortable with up to 10% blacks and 
after that point "white flight" takes place and whites may flee a given 
neighborhood. This is a highly charged term and no longer used by 
sociologists, so be aware of its other connotation.

F. And there does seem to be some bibliography on it
Angela Kluge wrote:
> Grimes, Joseph E.  1986.  Area norms of language size.  In:  Elson, Benjamin
> F. (ed.).  1986.  Language in global perspective:  Papers in honor of the
> 50th anniversary of Summer Institute of Linguistics 1935-1985.  Dallas, TX:
> Summer Institute of Linguistics, 5-19.
>   
(Thanks to Angela, I now have a pdf of it, and could send it to you.)

Coulmas, ed. 1984: Linguistic minorities and literacy: Language policy issues in
developing countries. (esp. article by Srivastava, quote by Hartmut Haberland above. He 
quoted the paper (and reproduced the Srivastava's diagram, which he says is the really
interesting thing) in his paper in:
A language policy for the European Community : prospects and quandaries / edited
by Florian Coulmas. Berlin: Mouton de Gruyter, 1991. (Contributions to the
sociology of language: 61))

Williams' references on Ethnolinguistic Vitality and Group Identity
> Clement, R., (1980), 'Ethnicity, Contact and Communicative Competence 
> in a second language' in 'Giles, H et al (eds) Language: social 
> psychological perpectives, Oxford, 1980
> Giles, H., Bourhis R.Y., & Taylor, D., 1977, 'Towards a Theory of 
> Language in Ethnic Group Relations' in Giles, H., et al (eds) 
> Language, Ethicity and Intergroup Relations
> Giles, H., & Byrne, J.L., 1982, 'An Intergroup Approach to Second 
> Language Acquisition' in JMMD 3,1
> Giles, H., Rosenthal, D., & Young, L., 1985, 'Perceived 
> Ethnolinguistic Vitality: the Anglo- and Greek- Australian Setting' 
> inJMMD 6,3/4
> Johnson, P., Giles, H., & Bourhis, R.Y., 1983, 'The Viability of        
> Ethnolinguistic Vitality: A Reply' in JMMD 4,4
> JMMD, Vol 3, No 3, 1982: Special Issue on Language and Ethnicity
> Labrie, N. & Clement, R., 1986, 'Ethnolinguistic Vitality, 
> Self-Confidence and Second Language proficiency: an Investigation' in 
> JMMD 7,4
> Saint-Blancat, C., 1984, The Effect of Minority Group Vitality Upon 
> its Sociopsychological Behaviour and Strategies in JMMD 5,6
> Smolicz, J.J., 1983, 'Modification and Maintenance: Language among 
> School-children of Italian background in South Australia' in JMMD 4,5
> Taylor, D.M., (1980) 'Ethnicity and Language: A social psychological 
> perspective' in Giles, H, Robinson, W.P. & Smith, P.H. (eds) Language: 
> Social Psychological Perspectives, Oxford
> Ward, C., & Hewston, M., 1985, 'Ethnicity, Language and Intergroup 
> relations in Malaysia and Singapore: A Social Psychological Analysis' 
> in JMMD 6, 3/4
Aurolyn Luykx wrote:
> There was a panel on "tip" at the AAA last year, though I think it must have been in San Francisco, because I went to Atlanta and didn't see it. If you can get hold of the conference program or the book of abstracts, those would list who the presenters were.
Unfortunately, this seems to be beyond the event horizon of 
http://www.aaanet.org/pubs/index.htm

Adrienne Redd wrote:

 Leisy Wyman, author  of (2004) Language Shift, Youth Culture, and 
Ideology - a Yup'ik example (Unpublished dissertation) examined /just 
this question/ in following two cohorts of Yup'ik Eskimo youths over 
five (!) years while she lived and taught in a Yup'ik village in Alaska. 
Leisy's email address is Leisy at aol.com <mailto:Leisy at aol.com>. I believe 
she lives on the West Coast of the U.S., so she will be nine hours 
earlier than you are. 

Thanks to all of the 25 who responded. I have endeavoured to include all 
the substantive points they made, but many more wrote saying how 
interesting they found the issue. I hope they had the patience to read 
this far.

Nicholas Ostler

-- 
Foundation for Endangered Languages
Registered Charity: England and Wales 1070616
172 Bailbrook Lane, Bath BA1 7AA, England
+44-1225-852865  nostler at chibcha.demon.co.uk
http://www.ogmios.org

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/lgpolicy-list/attachments/20060105/58b677ef/attachment.htm>