Language and Linguist Compass - 2007 - Romaine - Preserving Endangered Languages

Language and Linguistics Compass 1/1–2 (2007): 115–132, 10.1111/j.1749-818X.2007.00004.
Preserving Endangered Languages

Suzanne Romaine*
Merton College, University of Oxford
Abstract
Over the last few decades an increasing number of books, scholarly articles and
media reports have predicted that as many as 60 to 90% of the world’s some 6900
languages may be at risk of extinction within the next 100 years. This article
provides an overview of the current state of the world’s languages, explains some
causes and consequences of the loss of linguistic diversity, in addition to outlining
some of the range of efforts currently underway worldwide to preserve endangered
languages. We should think about languages in the same way as we do other
natural resources that need careful planning: they are vital parts of complex local
ecologies that must be supported if global biodiversity is to be sustained.
1. Introduction
Over the last few decades an increasing number of books, scholarly articles
and media reports have predicted an alarming decline in the number of
languages (e.g., Robins and Uhlenbeck 1991; Krauss 1992; Crystal 2000;
Nettle and Romaine 2000; Gibbs 2002; Abley 2003; Dalby 2003). Some
linguists think that as many as 60 to 90% of the world’s approximately 6900
languages may be at risk of extinction within the next 100 years. Nettle
and Romaine (2000: 2) estimate that about half the known languages in
the world have disappeared over the past 500 years. Crystal (2000: 19) suggests
that an average of one language every 2 weeks may vanish over the next
100 years. Krauss (1992) believes that only the 600 or so languages with the
largest numbers of speakers (i.e., more than 100,000) may survive. If this
is true, few of the approximately 6000 remaining languages will have a secure
future. No children are learning any of the nearly 100 native languages in
what is now the state of California. Only a handful of the hundreds of
Aboriginal Australian languages may survive into the next century. Similar
dismal statistics and gloomy prognostications emerge from various parts of the
globe.
The aim of this article is to offer a more sophisticated understanding of
what works and what does not in efforts to preserve endangered languages.
However, the first step in the solution to any problem is to acknowledge
its existence and understand its origins. Only by understanding the historical
© 2007 The Author
Journal Compilation © 2007 Blackwell Publishing Ltd
116 Suzanne Romaine
and social circumstances that have created the current threat to the world’s
languages can we hope to reverse it. This article will provide a detailed
overview of the scale and character of the problem of language endanger-
ment, consider some of its causes and consequences, and outline some of
the range of efforts currently underway worldwide to ameliorate the
situation.
2. Current State of the World’s Languages: Character and Scale of the Problem
of Language Endangerment
A lot is known about the character of the problem, if by that we mean
why and how languages become endangered. Nettle and Romaine (2000)
provide a global overview of some of the causes of the loss of linguistic
diversity. For most of the many millennia of human history, it seems likely
that the world was close to linguistic equilibrium, i.e., the number of
languages lost roughly equalled the new ones created. The reason why this
balance persisted so long is that there were no massive, enduring differ-
ences among the expansionary potentials of different peoples, of the kind
that might cause the sustained expansion of a single, dominant language.
Over the past 10,000 years, various events have punctured this equilibrium
forever. First, the invention and spread of agriculture, the rise of coloni-
alism, later the Industrial Revolution, and today globalization, electronic
technology, etc., have created the global village phenomenon. These forces
have propelled a few languages – all Eurasian in origin – to spread over the
last few centuries. Although I have placed these events in their chrono-
logical order in human history, they have not by any means played
themselves out to completion across the globe. For example, in Africa and
elsewhere hunter gatherers and semi-nomadic pastoralists are currently being
pushed out by the expansion of farmers, while China and other parts of
the world are just now undergoing the industrial revolution.
As large language communities expand, others contract. Over the last
500 years, small languages nearly everywhere have come under intense
threat. Speakers of large languages like English and Chinese find it difficult
to imagine the prospect of being the last speaker of their language, but
the last speakers of probably half the world’s languages are alive today.
Only two fluent speakers remain of the Warrwa language traditionally
spoken in the Derby region of West Kimberley in Western Australia. Only
about half a dozen elderly people on the island of Erromango in southern
Vanuatu can still speak Ura. Marie Smith Jones is the last person who still
speaks Eyak, one of Alaska’s 20 some native languages. Only two (Siberian
Yup’ik in two villages on St. Lawrence Island, and Central Yup’ik in 17
villages in southwestern Alaska) are spoken by children as the first language
of the home. Tefvik Esenc, believed to be the last known speaker of the
Ubykh language once spoken in the northwestern Caucasus, died in Turkey
in 1992. The disappearance of Ubykh is the final result of a genocide of
© 2007 The Author Language and Linguistics Compass 1/1–2 (2007): 115–132, 10.1111/j.1749-818X.2007.00004.x
Preserving Endangered Languages 117
the Ubykh people, who until 1864 lived along the eastern shore of the
Black Sea in the area of Sochi (northwest of Abkhazia). The entire Ubykh
population left its homeland when Russia conquered the Muslim northern
Caucasus in the 1860s. Tens (and possibly hundreds) of thousands of
people were expelled and had to flee to Turkey with heavy loss of life,
and the survivors were scattered over Turkey. And Turkey itself is a country
that until recently recognized no minorities and prohibited languages such
as Kurdish from public use.
Language shift is symptomatic of large scale processes and pressures of
various types (social, cultural, economic, and military) on a community
that have brought about the global village phenomenon, affecting people
everywhere, even in the remotest regions of the Amazon. Language shift
may be thought of as a loss of speakers and domains of use, both of which
are critical to the survival of a language. The possibility of impending shift
appears when a language once used throughout a community for everything
becomes restricted in use as another language intrudes on its territory.
Usage declines in domains where the language was once secure, e.g., in
churches, the workplace, schools, and, most importantly, the home, as
growing numbers of parents fail to transmit the language to their children.
Fluency in the language is higher among older speakers, as younger gen-
erations prefer to speak another (usually the dominant societal) language.
2.1 global distribution of linguistic diversity

The global distribution of linguistic diversity in the modern world is
strikingly uneven. Some indications of this can be obtained by looking at
statistics about the number of languages and the peoples who speak them
in various parts of the world. According to the Ethnologue database compiled
by the Summer Institute of Linguistics, 6912 languages are spoken by a
world population of 5,723,861,210 (Gordon 2005; see also the Ethnologue
website: http://www.ethnologue.com/). Note that these are only estimates
and that this population figure is lower than most other official statistics
which place the world’s population at about 6.34 billion. This is because
the Ethnologue lacks population estimates for about 5% of the languages.
These figures also do not include endangered dialects of healthy or relatively
healthy languages (see Wolfram and Schilling-Estes 1998 for discussion of
dialect endangerment). No one knows exactly how many languages and
dialects there are, and there is still very little research on most of the world’s
languages. Linguists have tended to work on the familiar and easily acces-
sible languages spoken by large numbers of people. While thousands of
linguists have probably worked on French or English over the last 100 years
or so, there are thousands of other languages that have received little attention,
and thousands still that have received none at all. The Ethnologue singles
out more than 3000 languages in need of surveys. Skutnabb-Kangas
(2000: 27) comments that we know more about the number of pigs in
118 Suzanne Romaine
Denmark, their size and weight, than we do about the world’s languages.
Denmark is a bacon-exporting country and pigs are a critical economic
resource. Few people see languages as a resource. More often, linguistic
diversity is seen as a problem.
Nevertheless, if we accept the Ethnologue figures as rough working
estimates, we see huge disparities between the sizes of the populations speak-
ing the world’s many languages. If all languages were of equal size, each would
have about 828,000 native speakers (or 917,000 using the population
figure of 6.34 billion). Yet, only about 5% of the world’s languages have
at least 1 million native speakers. These 347 languages account for 94%
of the world’s population (and this figure is increasing). In contrast, the
remaining 95% of languages are spoken by only 6% of the world’s people.
Figure 1 displays the inverse relation between world population and
number of native speakers per language. Nearly 80% of the world’s popu-
lation speaks a total of only 75 languages. Each of these has 10 million or more
speakers. Only eight languages in the world have more than 100 million
speakers; they are spoken by about 40% of the population. These include
Mandarin Chinese (873,014,298 speakers), Spanish (322,299,171), English
(309,352,280), Hindi (180,764,791), Portuguese (177,457,180), Bengali
(171,070,202), Russian (145,031,551), and Arabic (136,411,737).
These very large languages are all spoken in more than one country, such
as English, for instance, with large groups of speakers in the UK, USA,
New Zealand, Canada, Australia, and South Africa. For similar historical
reasons, Spanish and Portuguese spread through colonial conquest and
settlement, and are now widely spoken in Latin America, parts of Africa,
the Pacific, and North America. Today an Indo-European language, either
English, Spanish, Portuguese, (or French, not among the top eight, with
Fig. 1. Inverse relation between number of languages and size of population.
its 65 million speakers) is the dominant language in every country in North,

Central and South America. Languages such as Bengali and Hindi have
been brought by post-colonial diaspora populations to the UK. Likewise,
Mandarin Chinese, now spoken by more people than any other language
in the world, is found not only in mainland China, but also in Singapore,
Taiwan, and Hong Kong (as are other Chinese languages such as Wu,
Cantonese, etc.).
The spread of these languages is even more extensive if figures for
second-language speakers are included. Although English is not the largest
language in the world (Mandarin Chinese has more first-language speakers),
it is one of a small handful of what may be called global languages in
terms of geographic spread and number of users worldwide. Non-native
speakers outnumber native speakers. As recently as the sixteenth century,
knowledge of English was nearly useless outside of the British Isles, but
nowadays, one can go practically anywhere in the world and hear English
spoken. The language of the ‘global village’ is indeed English.
Most of the world’s languages, however, do not show the same geographic
spread as these big eight. Much of the rest of the world speaks languages
with fewer than 10,000 speakers. The median number of speakers for the
languages of the world is only 5000–6000, and nearly 85% of languages
have fewer than 100,000. Eighty-three percent of the world’s languages
are spoken only in one country. Moreover, most languages do not even claim
a territory as large as a country. Leaving aside the world’s largest languages,
spoken by more than 90% of the world’s population, the remaining some
6500 are confined to 5% of the world’s most marginalized peoples, who
have generally been on the retreat for several hundred years. The majority
of these smaller languages may be at risk. The Ethnologue estimates that there
are 204 languages with fewer than 10 speakers, 344 languages with
10–99 speakers, and 548 languages with fewer than 99 speakers. These
smallest of small languages comprise nearly one-tenth of the world’s
languages.
These small languages are unevenly shared among countries and across
continents, as shown in Table 1 displaying the number and percentage of
languages and speakers by area of the world. Certain parts of the world
such as the Asia-Pacific region are hotbeds of linguistic diversity in terms
of number of languages, while others such as Europe are more uniform.
Indeed, if some horrific catastrophe wiped out all the languages of Western
Europe tomorrow, we would lose relatively little of the world’s linguistic
diversity. Europe has only about 3% of the world’s languages, and most of
the largest European languages are also widely spoken outside Europe.
More importantly, however, most of the languages of Europe are struc-
turally quite similar, because they are related historically. If we were to lose
the same number of languages in Papua New Guinea or South America,
the loss would be far more significant, because the divergence between
languages there runs much deeper (see Nettle and Romaine 2000: 33 –39).
120 Suzanne Romaine
Table 1. Number/percentage of languages/speakers by area of the world (from

data compiled in the Ethnologue, table 1)
Area Number of Percentage of Number of Percentage of

languages languages speakers speakers
Africa 2092 30.3 675,887,158 11.8

Americas 1002 14.5 47,559,381 0.8
Asia 2269 32.8 3,489,897,147 61.0
Europe 239 3.5 1,504,393,183 26.3
Pacific 1310 19.0 6,124,341 0.1
Total 6912 100.0 5,723,861,210 100.0
Overall, the risk to rare languages is greater than the risk to more common
ones (Nettle and Romaine 2000: 62–67).
Papua New Guinea alone contains 11.9% of the world’s languages, but
only 0.1% of the world’s population and 0.4% of the world’s land area.
The overall ratio of languages to people is only about 1 to 5000. If this
ratio were repeated in the USA, there would be 50,000 languages spoken
there (Nettle and Romaine 2000). Around 80% (N = 5542) of all the
world’s languages are found in just 20 nation-states, among them some of
the poorest countries in the world. They include Papua New Guinea
(820), Indonesia (742), Nigeria (516), India (427), USA (311), Mexico
(297), Cameroon (280), Australia (275), China (241), Democratic Republic
of Congo (216), Brazil (200), the Philippines (180), Malaysia (147), Canada
(145), Sudan (134), Russia (129), Tanzania (129), Nepal (125), Vanuatu
(115), and Myanmar (113). As the list reveals, considerable diversity is
found in some of the most highly developed, industrialized nations. Canada
and the USA account for 456 languages or 6.5% of the world’s linguistic
diversity. If we add Australia with its 275 languages as the other large
predominantly Anglophone nation in the list, this yields a total of 731
languages, amounting to 10.5% of the world’s linguistic diversity.
2.2 assessing endangerment

Having examined some substantial disparities in the geographic spread
and size of languages, we are in a better position to address the question
of how much of the world’s linguistic diversity is in danger. Estimates of
the number of threatened languages vary a great deal from 50 to 90%
depending on the criteria used to assess risk. UNESCO’s World Atlas
of the World’s Languages in Danger of Disappearing (2001) estimates that
50% of languages may be in various degrees of endangerment. (See
www.unesco.org/culture/en/endangeredlanguages/atlas for an online ver-
sion of the Atlas currently under development; see also the UNESCO
Red Book of Endangered Languages at http://www.tooyoo.l.u-tokyo.ac.jp/
Redbook/index.html.)
However, more research is needed before we can start understanding

the importance of various factors in promoting or hindering language
maintenance. The pulse of a language quite clearly lies in the youngest
generation. Languages are at risk when they are no longer transmitted
naturally to children in the home by parents or other caretakers. UNESCO
suggests that languages being learned by fewer than 30% of the younger
generation may be at risk. Not surprisingly, we have very little informa-
tion about the number of languages no longer being transmitted naturally
to children, but the statistics we do have from several parts of the world,
for example, Canada and Australia, as well as reports from fieldworkers on
the ground are quite alarming. In Canada perhaps only three languages
out of around 50 (Cree, Atikamek, and Inuktitut) are viable in the long
term. (See, for example, Norris and Jantzen’s 2002 work on assessing the
continuity of Aboriginal languages using census statistics, and the Task Force
on Aboriginal Languages and Cultures 2005.) The Australian continent is
a linguistic graveyard, with 90% of Aboriginal languages near extinction
(McConvell and Thieberger 2001). Only some 50 languages are widely
spoken today; of these only 18 have at least 500 speakers. These 18
languages account for roughly 25,000 of the remaining 30,000 speakers
of Aboriginal languages today. No Aboriginal language is used in all
spheres of everyday life by members of a sizeable community.
If we use size as a proxy for degree of endangerment, estimates of the
scale of the problem will still vary, depending on how many speakers are
thought to be needed for a language to be viable. Table 2 shows the
percentage of indigenous languages in different continents with fewer than
some number of speakers. The languages in Australia and the Pacific and
the Americas are mostly very small; over 20% have fewer than 150 speakers,
and nearly all have fewer than 100,000. Africa, Asia, and Europe, in
contrast, have a fair number of medium-sized ones (100,000 to 1 million
speakers), in addition to some giant languages. Such languages are probably
safer in the short term at least. If Krauss (1992) is right in thinking that
Table 2. Percentages of languages according to continent of origin having fewer

than indicated number of speakers (from Nettle 1999: 114)
Continent/region < 150 < 1000 < 10,000 < 100,000 1 million
Africa 1.7 7.5 32.6 72.5 94.2

Asia 5.5 21.4 52.8 81.0 93.8
Europe 1.9 9.9 30.2 46.9 71.6
North America 22.6 41.6 77.8 96.3 100
Central America 6.1 12.1 36.4 89.4 100
South America 27.8 51.8 76.5 89.1 94.1
Australia/Pacific 22.9 60.4 93.8 99.5 100
World 11.5 30.1 59.4 83.8 95.2
122 Suzanne Romaine
only languages with over 100,000 speakers are safe, then up to 90% of
the world’s languages may be at risk.
Size does not tell the whole story, but it may be the best surrogate at
the moment for vulnerability to the kinds of pressures leading to language
loss. A large language could be endangered if the external pressures on it
were great (e.g., the South American language Quechua, with millions of
speakers), while a very small language could be perfectly safe as long as the
community was functional and the environment was stable (e.g., Icelandic,
with fewer than 300,000). Small size has been a stable characteristic of
languages in Australia and New Guinea for millennia, many of which
were quite healthy until a couple of centuries, generations, or decades ago.
However, small languages can disappear much faster than large ones, and
forces have now been unleashed in the world that small communities find
very difficult to resist, while larger groups may have the resources to do
so. In present circumstances size may be quite critical in determining
survival. All our estimates are guesses, but even if the viability threshold
is set at the lower level of 10,000 speakers, 60% of all languages are already
endangered. The situation is slightly better in Africa (33%), Asia (53%),
and Europe (30%), but much worse in North and South America (78%
and 77%) and Australia and the Pacific (93%).
Not only is the geographic spread of languages unequal, but so is their
status. The functions a language is used for may also tell us something
about its long-term viability. Fewer than 4% of the world’s languages have
any kind of official status in the countries where they are spoken. A small
minority of dominant languages prevail as languages of government and
education. English is the dominant de facto or official language in over
70 countries; French has official or co-official status in 29 countries. The fact
that most languages are unwritten, not recognized officially, restricted to
local community and home functions, and spoken by very small groups of
people reflects the balance of power in the global linguistic marketplace.
3. Language Preservation:What Is Being Done and What Can and Should Be Done
The prospect of the loss of linguistic diversity on such a large scale has pro-
mpted both communities and scholars to propose programs of intervention
to preserve and revitalize languages (see Hinton and Hale 2001; Grenoble
and Whaley 2006: Chapter 3 for examples). Because much typically needs
to be done quickly with too few resources, setting realistic priorities is
paramount. There is no one-size-fits-all solution for revitalization and
preservation. The immediate need is to identify and stabilize languages
under threat so that they can be transmitted to the next generation in as
many of their functions as possible. This means assessing which functions
are crucial to intergenerational transmission and have a reasonable chance of
successful revival and continuation. Every group must decide what can best
be done realistically for a particular language at a particular time.
3.1 reversing language shift

Fishman’s (1991, 2001) plan of action, referred to as reversing language
shift (RLS), is based on an eight-point scale called GIDS (Graded Inter-
generational Disruption Scale) that assesses the extent to which transmis-
sion has been disrupted in order to assist communities in targeting their
efforts to specific activities needed at each point of the scale. Fishman has
been quite insistent about the necessity of proceeding from the bottom
up, and of securing intergenerational transmission at home (Stage 6) before
proceeding to higher levels, such as use in schools, media, government,
etc. One of the most frequent mistakes activists make is to attempt to
reverse language shift by promoting the minority language in the domains
now dominated by the majority language. Fishman (1991: 54 –5) com-
pares the distribution of Basque and Spanish in 1957 and 1987, noting that
over this 30-year period there has been a decrease in the use of Basque
at home and in private domains, while there has been a slight increase in
the use of Basque in public, formal domains such as government, education,
etc. In Wales too, where Welsh has been introduced in public domains
where it used to be excluded, for example, public administration and
education, its previous dominance in other critical domains such as the
home, and even the chapel in some areas, has been weakened. The family
is no longer the main agency of language reproduction (Romaine 2006).
King (2001: 26) distinguishes between RLS and language revitalization,
which can be understood as not necessarily attempting to bring the language
back to former patterns of familial use, but rather to bring the language
forward to new users and uses. In doing so, however, we must not deceive
ourselves that the efforts directed at the latter will restore intergenerational
transmission. There may be an increase in users and uses of language
without intergenerational transmission, but the observations made by
Dauenhauer and Dauenhauer (1998: 97) are shared by many communities
whose languages are facing loss of intergenerational transmission:
[I]t is unrealistic to expect the Native languages of Southeast Alaska to recover
fully and thrive as they did sixty to a hundred years ago. But they can continue
to be used in many ways, both oral and written, that are of enduring spiritual
value to the individual and the community, even if these new uses are far more
limited and restricted than they would have been in the past.
3.2 revitalization through immersion

Communities around the world have increasingly looked to schools and
other teaching programs as a way to revitalize their languages. Immersion
models of various types are widely used to promote indigenous and minority
languages. Some programs are total immersion, such as the Hawaiian and
Blackfeet ones, modelled on the Maori ‘language nest’ (Te Kohanga
Reo), while others may be partial. Language nests typically aim to provide
124 Suzanne Romaine
a sheltered environment in which young children are exposed to the

language by fluent elders and other caretakers. In Hawaiian immersion
schools Hawaiian is used across the curriculum from pre-school onwards,
and English is introduced as a subject from the fifth grade (around age 10)
for 1 hour a day. Most of the students attending are English speakers and
are learning Hawaiian as a second language. The language nest/immersion
model contrasts with more conventional language teaching where the
language is taught as a subject for a limited number of hours with fewer
opportunities for high levels of academic or informal engagement with
the language in use. In immersion there may be little, if any, focus on
language learning per se in the form of direct teaching of grammar and
vocabulary. Language is acquired through the meaningful interaction
required to learn academic content in various subjects.
Other variants of the model may rely on bilingual immersion com-
bined with a third language taught as a subject. In parts of the Basque
Country, Basque and Spanish are used for instruction during primary
education, and English is taught as a subject beginning in kindergarten.
The Kahnawake Survival School in Quebec, dedicated to preserving
Mohawk language, culture, and history, combines total and partial immer-
sion at various levels.
Successful immersion programs obviously rely on the availability of
fluent teachers and teaching materials and are therefore less easily imple-
mented when a language has only a small handful of speakers. In California
and elsewhere in native North America, learners have benefited from a
Master–Apprentice program, which brings together a fluent elder and a
learner, who use the language for everyday activities (Hinton 2002). Another
variant brings together elders and learners of various ages in immersion
camps for several days or longer, often during the summer, to engage in
cultural activities in which native languages are used to varying extents.
Where transmission has ceased altogether or been interrupted for a
long period, prospects for revitalization rely on documentation and /or
reconstruction to ‘reclaim’ what some have called ‘sleeping languages’.
One example is the Kaurna language once used by Aboriginal people in
what is now the area of Adelaide in South Australia. Although it has not
been spoken for more than a century, some people are now using the
language for limited activities such as greetings, songs, and naming activities
(Amery 2001).
Revitalization activities of these various types, however, will not save
languages without firm community foundations for transmission. There is
an important distinction to be made between learning a language in the
artificial environment of the classroom and transmitting it in the natural
environment of the home. Schools in Ireland have achieved most of what
can be expected from formal language education, namely, knowledge of
Irish as a second language acquired in late adolescence. They have not led
to its spoken use in everyday life, nor its intergenerational transmission.
Nowhere have language movements succeeded if they relied on the school

or state to carry the primary burden of maintenance or revival. Indeed,
Grenoble and Whaley (2006: ix) note that ‘an honest evaluation of most
language revitalization efforts to date will show that they have failed’.
Moreover, in most communities revitalization and shift proceed in
tandem because not all community members agree on what can and should
be done. Language revitalization movements tend to affect only a small
minority of individuals, usually a small group of urban intellectuals
initially, and they do not always succeed in gaining widespread popular
support. The movement to revive Irish, for example, began among the
educated middle classes in Dublin, a place usually perceived as alien and
interfering by the remaining native speakers in the remaining Irish-speaking
areas in the west. New varieties of language often emerge in immersion
schools that are different to the varieties traditionally learned at home. Lack
of secure home and community foundations for transmitting minority
languages means that these new varieties may eventually replace traditional
varieties, but until they do their authenticity will be contested. In some
cases disputes have erupted over control of schools and linguistic resources.
Language can easily become politicized when it is no longer unselfcon-
sciously reproduced within families. Language choices become scrutinized
as an index of one’s authenticity and degree of commitment to the cause
of language revitalization. On the Scottish island of Skye, Macdonald
(1997: 238) observed that to the local Gaelic-speaking population, Gaelic
was part of a local identity rooted in everyday practice rather than as part
of a politicized package of language, heritage and culture advocated by
those outside the community. Most of those who opt for the new Gaelic
medium programs are those who speak very little Gaelic at home. One
40-year-old man who grew up speaking Gaelic at home said: ‘I speak the
Gaelic here with my parents and when I go up to the [hotel bar], but I
speak it not because I have to but because this is what we speak. I like the
Gaelic. But if it is going to become something artificial, then well, I won’t
feel like speaking it at all. I don’t want Gaelic to be kept alive by making
it artificial . . . For myself, I’d prefer if it died’ (MacDonald 1997: 218).
3.3 documentation
Documentation and revitalization activities go hand in hand, although
some linguists, such as Newman (1998), view endangered languages as a
‘hopeless cause’, while regarding documentation as an urgent scientific
task. A substantial literature is emerging on documentary linguistics in an
effort to establish a set of best practices (Himmelmann 1998; Lehmann
2001). A number of national funding bodies, such as the US National
Science Foundation and the German Volkswagen Stiftung, and international
agencies, such as UNESCO, have undertaken documentation projects
and set up archives to serve as repositories for data. HRELP (Hans Rausing
126 Suzanne Romaine
Endangered Languages Project) at the University of London’s School

of Oriental and African Studies (SOAS) has placed its emphasis on
documentation.
Estimates of the costs of providing adequate documentation range from
around $55,000 per language to fund 2 years of work by a linguist to
provide a basic dictionary and grammar to $2 million per language to
fund an in-depth study over 15 years, complete with audio-visual archives
and other comprehensive materials. Crystal (2000: Chapter 4) proposes a
figure of $585 million ($65,000 per year for 3 years’ work to provide basic
documentation for 3000 languages). Depending on one’s perspective, this
figure represents a lot of money or not very much at all. For example,
the USA pledged $350 million to aid victims of the 2004 Indian Ocean
tsunami; about the same amount is spent to wage war on Iraq for a day
and a half (Monbiot 2005). Even assuming that this sum was available for
language documentation, there is a shortage of people who are able and
willing to do the work. This is to say nothing of the political and logistical
impediments of getting to places where the most endangered languages
are spoken.
The enormity of the task as well as the rate of technological change
pose formidable challenges to scholars working to record endangered
languages, relying on ever more sophisticated forms of technology. Some
of the new information technologies that we think of as ‘advanced’ are
actually proving to be far less durable than the older technologies of print
and stone that technology gurus assured us they would replace. As old
storage mechanisms become obsolete, information must be continuously
moved to the latest medium or the data will be lost. Yesterday’s cutting
edge technologies all too quickly turn into today’s obsolete technologies.
What is salvaged will inevitably be small compared to what will be lost.
Although salvage operations aimed at recording a language for preser-
vation in books are worthwhile endeavors, and may be all that can be
accomplished for some severely eroded languages, they do not address the
root causes of language decline, and without further action they do not
contribute substantially to language maintenance efforts in the long term
(Romaine, forthcoming a). Grammars and dictionaries are artificial envi-
ronments for languages. They reflect only a fraction of the diversity of a
language in its everyday use and cannot capture the ever-changing nature
of language. Focusing on documentation in books is like concentrating
our efforts on preserving the spotted owl by building a museum where
we can display stuffed owls but doing nothing to preserve the bird in its
natural habitat or guarantee that it can reproduce itself.
Despite these realities, much of the professional linguistic literature on
language maintenance and preservation has been concerned with preserving
the structures of individual languages in grammars and dictionaries, or has
directed its attention to educational programs in threatened languages. There
is a tendency to reify languages as artifacts when we should be focusing
on living communities and language ecologies. Language ecology refers

to the social environment and domains in which a language is used. There
is a distinction to be made between documenting language data and
sustaining a language. What is being saved or preserved? Nora and Rich-
ard Dauenhauer put it well when they write that ‘Preservation [ . . . ] is
what we do to berries in jam jars and salmon in cans. [ . . . ] Books and
recordings can preserve languages, but only people and communities can
keep them alive’ (quoted in Lord 1996: 68). As Mühlhäusler (2002) argues,
we cannot preserve languages, but we can try to preserve language ecologies.
4. The Ecology of Language

The preservation of a language in its fullest sense ultimately entails the
maintenance of the group who speaks it, and therefore the arguments in
favour of doing something to reverse language death are ultimately about
preserving cultures and habitats. Languages can only exist where there is
a community to speak and transmit them. A community of people can exist
only where there is a viable environment for them to live in and a means
of making a living. Where communities cannot thrive, their languages are
in danger. When languages lose their speakers, they die. Extinctions in
general, whether of languages or species, are part of a more general pattern
of human activities contributing to radical alterations in our ecosystem.
In the past, these extinctions took place largely without human interven-
tion. Now they are taking place on an unprecedented scale through our
intervention, in particular, through our alteration of the environment.
Nettle and Romaine (2000) see the extinction of languages as part of the
larger picture of worldwide near total ecosystem collapse.
Although we are still in the early stages of understanding the ramifications
of the loss of diversity in ecosystems, species, cultures, and languages,
there is a growing body of factual evidence and supporting theory pointing
to an extinction crisis in the realms of both biological and cultural–
linguistic diversity. A new wave of interdisciplinary studies is yielding a
holistic view of diversity and considering how these two worlds of differ-
ence, biological and cultural/linguistic, are related and what common
factors are at work to diminish them, or conversely, to sustain them (Maffi
2001; Harmon 2002; Romaine, forthcoming b). Nettle and Romaine
(2000: ix) use the term ‘biolinguistic diversity’ as a key concept in a holistic
approach to the understanding of diversity. It refers to the rich spectrum
of life encompassing all the earth’s species of plants and animals along with
human cultures and their languages.
Figure 2 shows the areas of greatest linguistic diversity, which coincide
with areas that are high in biodiversity (Nettle and Romaine 2000: 43).
In other words, the areas identified as biodiversity hot spots by conservationists,
such as Myers, Mittermeier, Mittermeier, da Fonseca, and Kent (2000) are
also hotbeds of linguistic diversity, for example, New Guinea and tropical
128 Suzanne Romaine
Fig. 2. Geographic distribution of languages of the world (adapted from Nettle and Romaine
2000: 33, figure 2.1).
Africa. These regions richest in biolinguistic diversity are inhabited by

indigenous peoples who represent around 4% of the world’s population
but speak at least 60% of its languages. The explanation for these linkages
is not yet fully understood but will need to be sought in a sophisticated
ecological theory that takes account of peoples’ interactions with their
environment. Nettle (1999), for instance, refers to the notion of ‘ecological
risk’ as a significant influence on the formation and persistence of linguistic
groups. This factor refers to the amount of variation people face in their
food supply over time, which in turn is related to other variables such as
climate, diversification of productive and income-generating activities,
food storage, mobility, and patterns of social exchange. In areas where
rainfall is continuous throughout the year and communities are able to
produce their own food supply, they are not so dependent on their
neighbors for subsistence. Distinct languages may be more likely to evolve
and be maintained in small, self-sufficient communities. The greater the
ecological risk, the more people must develop larger social networks to
ensure a reliable food supply. Because language norms spread through
social networks, the average size of a language group increases in proportion
to ecological risk.
Not only do biodiversity and linguistic diversity share the same geographic
locations, they also face common threats. For example, an increasing
number of languages die each year, as the homelands of small indigenous
communities are being destroyed, or the communities are assimilated into
larger nation–states, some of which are actively seeking to exterminate
them. Because the historical and current causes of the threats facing the
Table 3. Extinction predictions for languages and species (based on data from
Sutherland 2003)
Fish 5%
Plants 8%
Birds 11%
Mammals 18%
Languages 32% (50 – 90%?)
earth’s languages, cultures, and biodiversity are the same, the solutions are
also likely to come from the same place: empowering local people and
communities. The measures most likely to preserve small languages are the
very ones that will help increase their speakers’ standard of living in a
long-term, sustainable way.
Despite the efforts of international organizations such as Terralingua
(www.terralingua.org), founded in the USA in 1996, and the Foundation
for Endangered Languages (www.ogmios.org), founded in the UK in 1995,
to raise awareness of endangered languages, linguists have been rather
slower in realizing the threats posed to language than ecologists have been
in recognizing the impending extinction in biodiversity. Yet available evi-
dence suggests that the threat to languages is much greater than to some
of the most common species of plants and animals. Table 3 compares some
projected rates of extinction for languages with those for species, using
figures from Sutherland (2003), whose estimates for the threat to languages
are rather smaller than the ones offered above. Sutherland acknowledges
that his rates are conservative. The rates for the various species are from
the IUCN (International Union for Conservation of Nature and Natural
Resources) Red List Criteria and Categories for endangered species (see
www.iucnredlist.org). However, even if we accept Sutherland’s lower figures
for language endangerment, languages are at greater risk than species in
any of the individual classes of animals, and all plants.
Because language plays a crucial role in the acquisition, accumulation,
maintenance, and transmission of human knowledge concerning the nat-
ural environment and ways of interacting with it, the problem of language
endangerment raises critical issues about the survival of knowledge that
may be of use in the conservation of the world’s ecosystems. The 49 members
of the Great Andanamese tribe, the last survivors of the pre-Neolithic
population of Southeast Asia, face a serious threat of extinction, but all
survived the huge tsunami unscathed after it hit the Indian Ocean and
Bay of Bengal in December 2004 because they knew exactly which trees
would not be swept away (Abbi 2006). Folk traditions and other forms
of knowledge passed down orally for generations are always only a gen-
eration away from extinction.
Furthermore, the issue of endangered languages cannot be separated
from people, their identities, their cultural heritage and their rights. When
130 Suzanne Romaine
we lose sight of people and the communities that sustain languages, it

becomes easy to argue, as have a number of critics, that there is no reason
to preserve languages for their own sake. However, maintaining cultural
and linguistic diversity is a matter of social justice because distinctiveness
in culture and language has formed the basis for defining human identities.
We should think about languages in the same way as we do other natural
resources which need careful planning to ensure their survival: they are
vital parts of complex local ecologies that must be supported if global
biodiversity, as well as human cultures and even humanity in general, are
to be sustained.
Short Biography
Suzanne Romaine has been Merton Professor of English Language at the
University of Oxford since 1984. Her research interests lie primarily in
historical linguistics and sociolinguistics, especially in problems of societal
multilingualism, linguistic diversity, language change, language acquisition,
and language contact in the broadest sense. She has conducted extensive
fieldwork in Europe (first on the language of working-class schoolchildren
in Scotland and subsequently on patterns of bilingualism and language loss
among Panjabi speakers in England) as well as in the Pacific Islands region
(first in Papua New Guinea on the language of rural and urban school-
children, and most recently in Hawaii). Her most recent book, Vanishing
Voices. The Extinction of the World’s Languages (Oxford University Press,
2000), co-authored with Daniel Nettle, won the British Association of
Applied Linguistics Book of the Year Prize in 2001.
Endnote
* Correspondence address: Suzanne Romaine, Merton College, Univer-
sity of Oxford, Oxford OX1 4JD, UK.
Works Cited
Abbi, Anvita. 2006. Endangered languages of the Andaman islands. Munich, Germany: Lincom
GmbH.
Abley, Mark. 2003. Spoken here. Travels among threatened languages. Toronto, ON: Random
House of Canada.
Amery, Robert. 2001. Warrabarna Kaurna! Reclaiming an Australian language. Lisse, The
Netherlands: Swets & Zeitlinger.
Crystal, David. 2000. Language death. Cambridge, UK: Cambridge University Press.
Dalby, Andrew. 2003. Language in danger. How language loss threatens our future. Harmonds-
worth, UK: Penguin.
Dauenhauer, Nora M., and Richard Dauenhauer. 1998. Technical, emotional, and ideological
issues in reversing language shift: examples from Southeast Alaska. Endangered languages.
Current issues and future prospects, ed. by Lenore A. Grenoble and Lindsay J. Whaley, 57–
99. Cambridge, UK: Cambridge University Press.
Fishman, Joshua A. 1991. Reversing language shift. Theoretical and empirical foundations of
assistance to threatened languages. Clevedon, UK: Multilingual Matters.
——. ed. 2001. Can threatened languages be saved? Reversing language shift, revisted: A 21st
century perspective. Clevedon, UK: Multilingual Matters.
Gibbs, W. Wayt. 2002. Saving dying languages. Scientific American 287.78 – 86.
Gordon, Raymond G., Jr. ed. 2005. Ethnologue: Languages of the world, 15th edn. Dallas,
TX: SIL International. Online version <http://www.ethnologue.com/> accessed May 28,
2006.
Grenoble, Lenore A. and Lindsay J. Whaley. 2006. Saving languages: An introduction to
language revitalization. Cambridge, UK: Cambridge University Press.
Harmon, David. 2002. In light of our differences. How diversity in nature and culture makes
us human. Washington, DC: Smithsonian Institution Press.
Himmelmann, Nikolaus P. 1998. Documentary and descriptive linguistics. Linguistics 36.161–95.
Hinton, Leanne, and Ken Hale, eds. 2001. The green book of language revitalization in
practice. San Diego, CA: Academic Press.
Hinton, Leanne with Matt Vera and Nancy Steele. 2002. How to keep your language alive: A
commonsense approach to one-on-one language learning. Berkeley, CA: Heyday Books.
King, Kendall A. 2001. Language revitalization processes and prospects: Quichua in the Ecua-
dorian Andes. Clevedon, UK: Multilingual Matters.
Krauss, Michael. 1992. The world’s languages in crisis. Language 68.4 –10.
Lehmann, Christian. 2001. Language documentation. A program. Aspects of typology and
universals, ed. by Walter Bisang, 83–97. Berlin, Germany: Akademie.
Lord, Nancy. 1996. Native tongues. Sierra Magazine 81.46 – 69.
Macdonald, Sharon. 1997. Reimagining culture. Histories, identities and the Gaelic renaissance.
Oxford, UK: Berg.
Maffi, Luisa ed. 2001. On biocultural diversity. Linking language, knowledge, and the environment.
Washington, DC: Smithsonian Institution Press.
McConvell, Patrick, and Nicholas Thieberger. 2001. State of indigenous languages in Australia
– 2001. Second Technical Paper Series No. 2. A report compiled for Environment Australia,
Department of Environment and Heritage, Canberra.
Monbiot, George. 2005. The victims of the tsunami pay the price of war on Iraq. The
Guardian. January 4, 2005.
Mühlhäusler, Peter. 2002. Why one cannot preserve languages (but can preserve language
ecologies). Language endangerment and language maintenance, ed. by David Bradley and
Maya Bradley, 34 –39. London: Routledge Curzon.
Myers, Norman, Russ Mittermeier, Christina G. Mittermeier, Gustavo A. B. da Fonseca, and
Jennifer Kent. 2000. Biodiversity hotspots for conservation priorities. Nature 403.853 – 8.
Nettle, Daniel. 1999. Linguistic diversity. Oxford, UK: Oxford University Press.
Nettle, Daniel, and Suzanne Romaine. 2000. Vanishing voices. The extinction of the world’s
languages. Oxford, UK: Oxford University Press.
Newman, Paul. 1998. ‘We has seen the enemy and it is us’: The endangered language issue as
a hopeless cause. Studies in the Linguistic Sciences 28.11–20.
Norris, Mary Jane, and Lorna Jantzen. 2002. From generation to generation: Survival and main-
tenance of Canada’s aboriginal languages within families, communities and cities. Ottawa:
Minister of Indian Affairs and Northern Development.
Robins, Robert H., and Eugenius M. Uhlenbeck, eds. 1991. Endangered languages. Oxford, UK:
Berg.
Romaine, Suzanne. 2006. Planning for the survival of linguistic diversity. Language Policy
5.443 – 475.
——. forthcoming a. Linguistic diversity, sustainability, and the future of the past. Endangered
and minority languages and language varieties: Defining, documenting and developing, ed.
by Kendall King, Natalie Schilling-Estes, Lyn Fogle, Jackie Lou, and Barbara Soukup.
Washington, DC: Georgetown University Press.
——. forthcoming b. Biodiversity, linguistic diversity, and poverty – some global patterns and
missing links. Language and poverty, ed. by Wayne Harbert, Sally McConnell-Ginet,
Amanda J. Miller, and John Whitman. Clevedon, UK: Multilingual Matters.
132 Suzanne Romaine
Skutnabb-Kangas, Tove. 2000. Linguistic genocide in education – or Worldwide diversity and

human rights? Mahwah, NJ: Lawrence Erlbaum Associates.
Sutherland, William J. 2003. Parallel extinction risk and global distribution of languages and
species. Nature 423.276 – 9.
Task Force on Aboriginal Languages and Cultures. 2005. Towards a new beginning. A foun-
dational report for a strategy to revitalize First Nation, Inuit and Métis languages and
cultures. Report to the Minister for Canadian Heritage. Ottawa: Aboriginal Languages
Directorate, Canadian Heritage, Government of Canada.
UNESCO. 2001. World Atlas of the world’s languages in danger of disappearing. Paris:
UNESCO.
Wolfram, Walt, and Natalie Schilling-Estes. 1998. Endangered dialects: A neglected situation in
the endangerment canon. Southwest Journal of Linguistics 14.117–31.

Language and Linguist Compass - 2007 - Romaine - Preserving Endangered Languages

Uploaded by

Copyright:

Available Formats

Language and Linguist Compass - 2007 - Romaine - Preserving Endangered Languages

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Language and Linguist Compass - 2007 - Romaine - Preserving Endangered Languages

Uploaded by

Copyright:

Available Formats

Language and Linguistics Compass 1/1–2 (2007): 115–132, 10.1111/j.1749-818X.2007.00004.

Preserving Endangered Languages

2.1 global distribution of linguistic diversity

Fig. 1. Inverse relation between number of languages and size of population.

its 65 million speakers) is the dominant language in every country in North,

Table 1. Number/percentage of languages/speakers by area of the world (from

Area Number of Percentage of Number of Percentage of

Africa 2092 30.3 675,887,158 11.8

2.2 assessing endangerment

However, more research is needed before we can start understanding

Table 2. Percentages of languages according to continent of origin having fewer

Africa 1.7 7.5 32.6 72.5 94.2

3.1 reversing language shift

3.2 revitalization through immersion

a sheltered environment in which young children are exposed to the

Nowhere have language movements succeeded if they relied on the school

Endangered Languages Project) at the University of London’s School

on living communities and language ecologies. Language ecology refers

4. The Ecology of Language

Africa. These regions richest in biolinguistic diversity are inhabited by

we lose sight of people and the communities that sustain languages, it

Skutnabb-Kangas, Tove. 2000. Linguistic genocide in education – or Worldwide diversity and

You might also like